Gastrointestinal Cancer

47

University of California San Francisco (UCSF) criteria and liver dysfunction predict hepatocellular carcinoma (HCC) recurrence after surgery for HCC: A large multi-centre study

Conner Blackmore^1,2, Ian Lockart^2,3, Yuen Kang Joseph Yeoh², Ciara Flynn², Gregory Dore^3,4, Mark Danta^2,3, Jacob George^5,6,7,8, Basheer Alshiwanna^1,2,9, Maryam Alavi⁴, Behzad Hajarizadeh⁴ and Miriam Tania Levy^1,2,9

¹Liverpool Hospital, Sydney, Australia; ²Faculty of Medicine, UNSW, Sydney, Australia; ³St Vincent's Hospital, Sydney, Australia; ⁴The Kirby Institute, Sydney, Australia; ⁵Storr Liver Centre, Sydney, Australia; ⁶Westmead Institute for Medical Research, Sydney, Australia; ⁷Westmead Hospital, Sydney, Australia; ⁸Faculty of Medicine, University of Sydney, Sydney, Australia; ⁹Ingham Institute for Applied Medical Research, Sydney, Australia

Background and Aim: Primary liver cancer is the third leading cause of cancer-related death worldwide, with hepatocellular carcinoma (HCC) accounting for 80% of primary liver cancers. Liver resection is a curative treatment for early HCC (more than 2cm or when inaccessible to locoregional ablation), however there is a paucity of literature on predicting the likelihood of HCC recurrence following resection. We evaluated factors related to recurrence following primary HCC resection with curative intent.

Methods: We retrospectively reviewed the electronic medical records of patients with HCC who underwent primary resection at three tertiary referral hospitals in Australia between January 2008 and May 2022. Baseline and follow-up characteristics, including liver disease, patient, and tumour characteristics were collected. The incidence rate of HCC recurrence following curative resection, and the factors associated with recurrence risk was evaluated.

Results: A total of 242 patients underwent surgical resection with a median follow up of 36.6 months (IQR: 13.8-60.5). The overall survival rate at 3 years was 54%, with total disease-free survival at 3 years 29%. Underlying HCC aetiology was hepatitis C virus (HCV) in 73 (30%), hepatitis B virus (HBV) in 98 (41%) and non-viral in 71 (29%). Clear histological margins and complete post-operative radiological response was achieved in 190 (79%) patients with 1 (0.4%) 90-day mortality. Recurrence occurred in 94 (39%) with median time to recurrence of 18.8 months (IQR: 8.6-35]). The incidence rate of recurrence was 10.5 per 100 person-years (95% CI: 8.6 – 12.9). Multivariate Cox regression analysis identified an increased risk of HCC recurrence was independently associated with: MELD score > 7 (aHR: 2.36; 95% CI: 1.20-4.63; p = 0.012), Alpha-fetoprotein (AFP) > 400 (aHR: 4.76; 95% CI: 2.07-10.91; p < 0.001), albumin-bilirubin (ALBI) grade > 2 (aHR: 1.71; 95% CI: 1.00-2.90; p = 0.049) and if tumour was outside UCSF criteria (aHR: 2.09; 95% CI: 1.08-4.04; p = 0.029). HCC aetiology was not associated with recurrence risk.

Conclusion: HCC recurrence in those having liver resection is more likely in those with worse liver function and those with tumours outside the UCSF criteria. Enhanced post-surgical interval surveillance could be considered for patients with this risk profile.

49

Patient-derived organoid culture: Advancing personalized chemotherapy for gastrointestinal peritoneal metastases

Harleen Kaur¹, Josephine Wright^1,2, Laura Vrbanac^1,2, Timothy Price³, Emma Bradshaw³, Alexander Stavropoulos¹, Dion Gouskos¹, Christopher Lauder³, Marcus Troschler³, Peter Hewett³ and Susan Woods^1,2

¹The University Of Adelaide, Adelaide, 5000, Australia; ²South Australian Health and Medical Research Institute, Adelaide, 5000, Australia; ³The Queen Elizabeth Hospital, Woodville South, 5011, Australia

Background and Aim: Peritoneal carcinomatosis from gastrointestinal tumours are considered a poor prognostic factor. The development of cytoreductive surgery and intraperitoneal chemotherapy including HIPEC & PIPAC have improved median overall survival for these patients by 21-32 months. The advent of patient-derived organoid culture (PDOs) has provided a breakthrough in personalized medicine, allowing researchers and clinicians to recapitulate the complexity and heterogeneity of individual tumours in vitro. We aimed to determine whether PDO cultures could be established in a time frame to be useful in predicting and guiding targeted drug treatments and to provide compelling evidence for the integration of PDOs into clinical practice.

Methods: Tumour samples were collected from patients prior to receiving HIPEC/PIPAC treatments. PDOs were established and drug panel testing ex vivo was performed. Cell viability was measured in quadruplicate following treatment with various HIPEC/PIPAC drugs (oxaliplatin, mitomycin, gemcitabine, nab-paclitaxel). Non-linear regression curve fits were used to generate dose response curves in GraphPad Prism with 95% cell viability used to measure resistance.

Results: PDOs were successfully generated from 24 patient samples (n=39) with a 74% culture success rate. Establishing organoids, immunohistochemistry and cell viability data following drug testing were completed within 8-10 weeks and reports on drug sensitivities were provided to the treating clinicians. This resulted in a recommendation for treatment change for three patients undergoing PIPAC as organoid cultures suggested insensitivity to the first-choice PIPAC chemotherapy.

52

A worldwide ascertainment of biallelic NTHL1 patients finds first individuals developing colorectal cancer without polyposis

Weilun Gao^1,2, Chuyi Liao³, Omar Salehi¹, Janakan Selvarajah¹, Dietje Elisabeth Fransen van de Putte⁵, Patrick R Benusiglio⁶, Jose van Montfoort⁷, Bernadette van Nesselrooij⁵, Petra Cohn-Hokke⁹, Nicolas Pachter¹⁰, Krzysztof Bernatowicz¹¹, Kevin Monahan¹², Karl Heinimann¹³, Giovanni Innella¹⁴, Lars Joachim Lindberg^15,16, Anne Marie Jelsig^15,16, Karin AW Wadt^15,16, John Gásdal Karstensen^15,16, Dan Buchanan^2,8, Richarda de Voer⁴ and Finlay Macrae^1,2

¹The Royal Melbourne Hospital, Melbourne, Australia; ²The University of Melbourne, Melbourne, Australia; ³Western Health, Melbourne, Australia; ⁴Department of Human Genetics, Research Institute for Medical Innovation, Radboud university medical center, Nijmegen, Netherlands; ⁵Department of Genetics, University Medical Center Utrecht, Utrecht, Netherlands; ⁶Department of Medical Genetics, Pitié-Salpêtrière Hospital, APHP, Sorbonne University, Paris, France; ⁷Erasmus University Medical Center, Rotterdam, Netherlands; ⁸Colorectal Oncogenomics Group, Department of Clinical Pathology, The University of Melbourne, Melbourne, Australia; ⁹Netherlands Cancer Institute, Amsterdam, Netherlands; ¹⁰Genetic Health WA, King Edward Memorial Hospital, Subiaco, Australia; ¹¹Adult Genetics Unit, Royal Adelaide Hospital, Adelaide, Australia; ¹²St Mark's The National Bowel Hospital, London, United Kingdom; ¹³Institute for Medical Genetics and Pathology, University Hospital of Basel, Basel, Switzerland; ¹⁴Faculty of Medicine, University of Bologna, Bologna, Italy; ¹⁵Danish HNPCC Register, Copenhagen University Hospital, Hvidovre, Denmark; ¹⁶Department of Clinical Genetics, Copenhagen University Hospital, Hvidovre, Denmark

Background and Aim: NTHL1 associated tumor syndrome (NATS) is a rare germline autosomal recessive disorder which predisposes to early onset colorectal cancer, breast cancer, endometrial cancer and meningiomas. Currently only 49 families have been identified through published literature worldwide. Our multicenter collaborative identifies 21 additional carriers and aims to further elucidate intestinal and extraintestinal manifestations of this syndrome.

Methods: A protocol with ethics application was approved at Royal Melbourne Hospital to collect deidentified clinical information of biallelic NTHL1 carriers from multiple familial cancer centers worldwide. This information was disseminated through the International Society for Gastrointestinal Hereditary Tumours (InSiGHT) with the formation of a NTHL1 collaborative. Invitations were sent with clinical data forms to collaborators collecting key phenotype and genotype information including genetic testing, ethnicity, pedigree, tumour types, age of diagnosis and endoscopy data. Collected information was then extracted for genotype and phenotype correlation, descriptive statistics and presented in a visual timeline to assist interpretation.

Results: 21 (13F/8M) carriers from 18 families were collected from 13 institutions. Predominant ethnicity is Caucasian (19/21; 90.5%) with one individual from Australia. The c.244C>T, p. (Gln82*) was the most common likely pathogenic/pathogenic variant with 12 individuals homozygous and further 9 individuals compound heterozygous for this variant. Median age of first cancer diagnosis was 53 years (range 32-66). Adenomatous polyposis was the most frequent tumour phenotype, with (19/21; 90.5%) individuals showing polyposis on colonoscopy. On average 69 polyps developed before age 56.4 years with a predominance for adenomatous polyps. A mixed polyp phenotype including hyperplastic and serrated polyps was demonstrated in (14/19; 73.7%) individuals with polyposis. 2 individuals from the same family (F14P1, F14P2) had no history of colonic polyps including at the time of colorectal cancer diagnosis ages 43 and 55 respectively. 13 individuals were diagnosed with colorectal cancer (median 56 years, range 43-77), 7 individuals with breast cancers (median 53, range 32-64). Other isolated tumors include skin cancers, endometrial cancers, meningiomas and gynecological cancers (figure 1).

88

Extreme elevation of Ca19-9 caused by intraductal papillary mucinous neoplasm

Phillip Te¹, Louise Blake¹, Kimberly Cukier¹ and Damian Dowling^1,2

¹Barwon Health, Geelong, Australia; ²School of Medicine, Deakin University, Geelong, Australia

Introduction: Carbohydrate antigen 19-9 (Ca19-9) is a serum tumour marker used to aid in diagnosis and management of pancreatic and bile duct malignancy. Ca19-9 has low specificity for malignancy at low level elevation, however markedly elevated levels are generally considered indicative of malignancy. Benign diseases causing markedly elevated levels of Ca19-9 are exceptionally rare. Intraductal papillary mucinous neoplasms (IPMN) are benign cystic lesions of the pancreas with malignant potential and are not typically associated with raised Ca19-9. To date, there appears to be no previous cases in the literature of marked elevation of Ca19-9 due to IPMN. We present a case of a sudden extreme elevation of Ca19-9 levels in a patient due to IPMN.

Case Report: A man in his 70s presented for outpatient review with a 2-week history of right sided abdominal pain. His medical history included Ischaemic Heart Disease with prior coronary artery bypass grafting and pacing wires, Type 2 Diabetes Mellitus complicated by diabetes-related nephropathy and cheiroarthropathy, overweight (BMI 26.4kg/m²) and diverticular disease. Initial investigations revealed an elevated lipase of 436U/L (13-60U/L) and Ca19-9 of 54KU/L (<37KU/L), with normal full blood count, biochemistry, and liver function. CT imaging of his abdomen and pelvis was suggestive of IPMN. The patient’s pain resolved, and he underwent surveillance of the IPMN with Ca19-9 testing. This revealed a chronically elevated, but stable Ca19-9 level of 49-52 (<37KU/L). 17 months post initial diagnosis of IPMN, there was an extreme elevation in his Ca19-9 level of 3949 (<37kU/L). The immediate concern was of malignant transformation of the IPMN, however further imaging with PET/CT of his abdomen and pelvis revealed the unchanged IPMN without metabolic activity. Endoscopy and endoscopic ultrasound revealed the known singular pancreatic tail cyst with no other lesions, and fine needle aspirate was non-diagnostic of malignancy. MR imaging was contraindicated due to pacing wires. Other causes of elevated Ca19-9, including renal disease, poorly controlled diabetes and liver disease were considered. However, the patient’s renal function, diabetic control and liver function remained stable over the time in which the patient had high Ca19-9, which made these causes less likely. Subsequent testing of his Ca19-9 two weeks later showed a reduction to 1257KU/L, then returned to his baseline mildly elevated state three months post initial elevation (Figure 1).

94

Value based assessment of same day upper and lower endoscopy for patients referred for colonoscopy with a positive immunochemical feacal occult blood test

Ayesha Shah^1,2, Naomi Moy¹, Amanda Whaley¹, Teressa Hansen¹, Kate Virgo¹, Uwe Dulleck^3,4,5, Natasha Koloski^1,2,6, Mike Jones⁷ and Gerald Holtmann^1,2

¹Princess Alexandria Hospital, Metro South Health, Woolloongabba, Australia; ²University of Queensland, Brisbane, Australia; ³University of Canberra, Australia, Canberra, Australia; ⁴University of Munich, Munich, Germany; ⁵Queensland University of Technology, Brisbane, Australia; ⁶The University of Newcastle, Callaghan, Australia; ⁷Macquarie University, Sydney, Australia

Background and Aim: A number of individuals who participate in population-based colorectal cancer (CRC) screening return a positive immunochemical faecal occult blood test (iFOBT) but do not have an identifiable lesion found at colonoscopy to account for their positive iFOBT. Thus, it has been argued that an upper gastrointestinal (GI) endoscopy might be warranted for iFOBT+ patients with a negative colonoscopy. An upper GI (UGI) endoscopy performed at the same time of the colonoscopy can be delivered with marginal added costs. This study aims to determine the additional costs from a health service perspective to perform a dual endoscopy and colonoscopy and the costs per life year saved.

Methods: Administrative data were used to determine the incremental costs for endoscope and procedure room utilisation. Patients referred for evaluation of iFOBT+ underwent same day esophagogastroduodenoscopy (EGD) and colonoscopy and significant upper and lower GI lesions were included. In all patient’s GI symptoms were assessed utilising the Structured Assessment of Gastrointestinal Symptoms. Relevant clinical data including symptoms were documented. Costs data are presented as purchasing power adjusted US$ (OECD 2020 dataset).

Results: Incremental cost for cleaning the endoscope was $60, with no additional costs for repairs required due to service contracts. To operate the procedure room, it cost $23/min and an added UGI endoscopy increased the average time by 8 minutes. 779 patients with a mean age of 61 (± SD, ±8.9) years were included. Colonic lesions were detected in 659/779 patients as compared to 673 patients with at least one relevant UGI lesion (P>0.7). Detected malignancies included 18 colorectal and 5 gastric cancers. Other UGI pathologies found included: 22 gastric ulcers, 29 Barrett’s oesophagus, 11 duodenal ulcers, 1 duodenal adenoma, and 3 oesophageal varices. No specific symptoms were associated with any malignancy. Absence of colonic lesions was not a predictor of UGI pathology (p>0.8). Considering direct costs, additional endoscopy increased costs by $244. Cost for clinical assessment, patient booking, recovery and patient’s worktime lost remained unchanged. Overall, the incremental cost per UGI cancer diagnosed was $25,812. This compares to $79,400 per CRC within the population based FOBT CRC program when considering total direct costs of $1,427 for colonoscopy.

Conclusions: iFOBT+ is associated with the presence of either upper or lower intestinal pathology. EGD performed at the same time of colonoscopy reveals significant UGI pathology in FOBT+ patients while incremental costs to diagnose UGI malignancies are small compared to the cost to identify CRC.

123

DNA hypermethylation markers of upper gastrointestinal adenocarcinoma: Preliminary results of systematic review

Thang Dao^1,2,3, Zexi Allan^3,4, Samith Alwis⁵, Ebtihal Mustafa^3,4, Niall Tebbutt^6,7, David S Liu^4,8,9,10, Stephen Wong^3,4 and Nicholas J Clemons^3,4

¹Department of Medicine, The University Of Melbourne, Parkville, Australia; ²Department of Medicine, St Vincent's Hospital Melbourne, Fitzroy, Australia; ³Division of Cancer Research, Peter MacCallum Cancer Centre, Parkville, Australia; ⁴Sir Peter MacCallum Department of Oncology, University of Melbourne, Parkville, Australia; ⁵Department of Surgery, Austin Health, Heidelberg, Australia; ⁶Department of Surgery, University of Melbourne, Parkville, Australia; ⁷Department of Medical Oncology, Austin Health, Heidelberg, Australia; ⁸Division of Cancer Surgery, Peter MacCallum Cancer Centre, Parkville, Australia; ⁹Upper Gastrointestinal Surgery Unit, Division of Surgery, Anaesthesia, and Procedural Medicine, Austin Health, Heidelberg, Australia; ¹⁰General and Gastrointestinal Surgery Research and Trials Group, The University of Melbourne Department of Surgery, Austin Health, Heidelberg, Australia

Background and Aim: Methylated tumour DNA in tissue and bodily fluid is a promising biomarker for different types of cancer. This study systematically reviews the literature on the available DNA hypermethylation markers of upper gastrointestinal adenocarcinoma. Identifying markers will be used to construct a novel methylomic test in the Western population aiming to improve diagnosis and staging in patients with upper gastrointestinal adenocarcinoma.

Methods: The databases Pubmed, Scopus, and Web of Science were systematically searched between 2010 and 2023. English-language articles that collected patients’ tissue samples to explore DNA hypermethylation markers of gastric, junctional, or oesophageal adenocarcinoma, with specified CpG loci or methylation primers were included. Patient characteristics, genes studied, methodology, CpG loci or methylation primers, and sensitivity and specificity information were collected.

Results: Preliminary results included 88 out of 149 articles reviewed, comprising 8421 patients (median age: 53 to 74 years, 73.8% men). Of these, East Asian cohorts comprising 7610 patients were studied in 64 publications. Samples used were mostly of primary tumour but also included plasma (18 studies) and peritoneal fluid (2 studies). Majority of methylation markers were detected using methylation-specific polymerase chain reaction, while pyrosequencing and bisulfite DNA sequencing were less common. There were 127 genes and CpG sites explored for methylation markers. The genes most investigated were RASSF1A, p16, RUNX3, CDH1, and MGMT. The most sensitive markers for oesophageal adenocarcinoma (12 studies) were in the genes miR124-3, NDRG4, miR129-2, and ZNF569 (82.5% to 88.6%); the most specific markers were in CDKN2A, MLH1, RRAD, miR129-2 (89.7% to 100%). The most sensitive markers for gastric cancer (71 studies) were in the genes CYP26B1, KCNA4, BMP3, LINE3, and LINE4 (90.0% to 96.9%); the most specific markers were in EBF3, E-cadherin, ADAMTS9, and MINT2 (96.7% to 100%). For junctional cancer (five studies), the most sensitive markers were in cg09177106, sFRP1, and SLC46A3 (77.8% to 83.3%); the most specific markers were in Dkk3, SLC46A3, and cg09177106 (94.4% to 100%).

Conclusion: The included studies reported the DNA methylation markers in predominantly East Asian middle-aged to older male patients with upper gastrointestinal adenocarcinoma. Several DNA hypermethylation markers with high sensitivity and specificity were identified. These markers will be evaluated in patient samples from the Western population to test their clinical utility in this setting.

126

Rare case of esophageal epidermoid metaplasia

Arvinf Rajandran, Christopher Graddon, Anthony Sakiris and Sneha John

Gold Coast University Hospital, Southport, Australia

Introduction: Esophageal epidermoid metaplasia (EEM) is a rare esophageal lesion characterised by a dense granular layer with overlying hyperorthokeratosis resembling the epidermis of skin. Given the low prevalence, data to guide management also remains sparse. Characterisation of EEM and their relationship to squamous dysplasia and squamous cell cancer (SCC) is thus important to determine management.

Case: We describe a 73-year-old female with a background history of gastroesophageal reflux disease (GORD) and longstanding lymphocytic esophagitis that was diagnosed in May 2021 after presenting with an episode of food bolus. Index gastroscopy showed significant esophageal narrowing which required serial endoscopic dilatations up until mid 2022. This was in addition to topical steroids and PPI for the lymphocytic esophagitis. Gastroscopy in May 2023 showed features of Barrett's and an area concerning for squamous dysplasia, however biopsies were non dysplastic. A repeat gastroscopy 3 months later again detected an area of nodularity in the esophagus at 30-32cm from the incisors although biopsies only showed reactive epithelial changes without malignancy. Surveillance gastroscopy in March 2024 showed three areas between 30 and 34cm of patchy 4-10mm mucosal variance characterised by nodularity. This time however, biopsies of this areas showed EEM and low-grade squamous dysplasia. Patient underwent an endoscopic mucosal resection (EMR) of the lesion.

Discussion: EEM (or esophageal leukoplakia) appears to share some similarities with a more common condition known as oral leukoplakia which is an oral, potentially premalignant condition associated with tobacco use, alcohol intake and HPV infection. EEM is a preneoplastic condition associated with squamous dysplasia and esophageal squamous cell carcinoma (ESCC). One study highlighted squamous neoplasia occurring before, concordant with or after EEM. Genetic alterations in esophageal epidermoid metaplasia specimens' that predispose them to oesophageal squamous cell carcinoma have been reported. Other associated conditions include lichen Planus, Barrett’s oesophagus, GORD, tobacco and alcohol use. There is also an association with oesophageal dysmotility, distal constriction, and stasis. In this case, the patient had a long-term history of chronic esophagitis in the form of lymphocytic esophagitis and GORD.

The diagnosis is made clinically on endoscopic appearance of the lesion (colour, thickness and texture) and confirmed histologically (hyperkeratosis, hyperorthokeratosis, hyperplasia and dysplasia). Hyperkeratosis is not specific for EEM and is also found in pill esophagitis, caustic ingestions, and sloughing esophagitis. The presence of a dense granular area with associated hyperorthokeratosis is specific for EEM. Given the pre neoplastic potential, endoscopic resection and close follow may be warranted. Larger lesions with dysplasia may benefit from endoscopic resection and or ablation such as with RFA. Surveillance gastroscopy, initially at 6 months from the diagnosis and then yearly with 4-quadrant biopsies every 1–2 cm can be considered for lesions that are not easily resectable, diffusely panesophageal and do not harbor dysplasia. In our patient, EMR was performed as per patient preference with a plan for further endoscopic surveillance.

131

Adrenal tumours in patients with pathogenic adenomatous polyposis coli (APC) mutations: a retrospective study

Lyman Lin¹, Victoria Beshay² and Finlay Macrae³

¹St Vincent's Hospital Melbourne, Fitzroy, Australia; ²Peter MacCallum Cancer Centre, Melbourne, Australia; ³Royal Melbourne Hospital, Parkeville, Australia

Background and Aim: Adrenal tumours may be associated with familial adenomatous polyposis (FAP). In the literature, most studies use the clinical definition of FAP (more than 100 adenomatous polyps found in endoscopic studies). However, not all patients that meet clinical criteria for FAP carry pathogenic mutations in the adenomatous polyposis coli (APC) gene, as there is genetic heterogeneity responsible for FAP with the polyposis sometimes explained by genetic and environmental factors other than pathogenic APC mutations. Reciprocally, not all the patients with pathogenic APC variants will fulfil the clinical criteria of FAP. This study aims to investigate the characteristics of adrenal tumours in patients with pathogenic or likely pathogenic APC variants.

Methods: This is a retrospective cohort study of consecutive patients that carry a constitutional pathogenic or likely pathogenic APC variant as reported in ClinVar or the InSiGHT LOVD database. They were tested by the molecular pathology laboratory at a tertiary cancer centre. Benign variants, likely benign variants or variants of unknown significance or where no APC variant was identified were excluded. Patients without available radiological reports in our database were also excluded. The CT and MRI scans were conducted from January 2009 to January 2023.

Results: 90 patients were included with a median age of 27 (IQR: 19.75 – 37.5) when confirmed to carry APC variants. The prevalence of adrenal mass was 26.7% (24/90) among patients with pathogenic or likely pathogenic APC variants. 34 adrenal tumours were found among these patients with a median maximal diameter of 17mm (IQR: 12.5-23). 25 (73.5%) of the tumours were radiologically assessed as adenomas, 3 (8.8%) were myelolipoma and 6 (17.6%) indeterminate. 14 (58.3%) patients had unilateral tumours, 9 (37.5%) bilateral, and 1 (4.2%) unknown laterality. We observed no genotype-phenotype correlation for adrenal tumours among the pathogenic or likely pathogenic APC mutations. There were only four patients who had adrenal hormone testing. Two were tested as part of a secondary hypertension screen, and both had Conn’s Syndrome. The other two were tested due to systemic symptoms of weight loss and fatigue in the context of their adrenal tumours. They both had mild abnormalities in hormone testing, not considered responsible for their clinical presentation. No adrenal malignancy was detected.

Conclusion: The prevalence of adrenal tumours among patients with pathogenic or likely pathogenic APC mutations in our cohort is at least two to three times higher than that reported in the general population based on international population-based studies. The hormonal functions of patients with pathogenic APC variants and adrenal tumours should be investigated further to better understand whether there are any clinical endocrine implications of the increased prevalence of adrenal tumours in these patients.

135

Investigating the role of E. coli Nissle biofilm formation in colorectal cancer tumour colonisation

Rebekah de Nys^1,4, Thais Martins de Lima^1,4, Sadia Hasan^1,4, Julia Leeflang^1,4, Kate Barratt^1,4, Georgette Radford^1,4, Sara Foschini⁴, Tharindie Silva^1,4, Laura Vrbanac^1,4, Josephine Wright¹, Daniel Worthley^1,2, Jeff Hasty³ and Susan Woods^1,4

¹South Australian Health and Medical Research Institute, Adelaide, Australia; ²Colonoscopy Clinic, Spring Hill, Australia; ³University of California at San Diego, San Diego, United States of America; ⁴The University of Adelaide, Adelaide, Australia

Background and Aim: Advanced colorectal cancer (CRC) patients often exhaust current treatments, prompting the need for innovative, new approaches. Genetically engineered Escherichia coli Nissle (EcN) strains are currently undergoing clinical trials for the treatment of refractory advanced cancers and phenylketonuria. Our research demonstrates EcN preferentially colonizes tumours over healthy tissue post-oral administration in our orthotopic CRC mouse model, as well as in CRC patients. We are working with world-leading synthetic biologists to genetically modify EcN to release anti-cancer therapeutics, direct to the tumour. However, tumour-colonisation rate of EcN varies across CRC model subtypes/tumour sizes and concerns remain regarding biocontainment and treatment efficacy. Addressing these challenges relies on first elucidating how EcN preferentially colonises the tumour microenvironment (TME). Our existing in vivo studies show EcN colonizes the luminal surface of colorectal tumours and exists in a community of other bacteria- that is, in a biofilm. We aim to investigate the role EcN biofilm formation plays in colorectal tumour colonization, to generate EcN strains with enhanced tumour specificity and therapeutic release.

Methods: All organoid experiments were performed using CRISPR/Cas9 engineered mouse colorectal canser ΔAKP (Apc^Δ/Δ, Kras^Δ/Δ, p53^Δ/Δ) and normal colon organoids. Microbe-organoid co-culture assay was performed by culturing organoids with EcN-GFP for 4 hours before treatment with gentamycin. EcN-GFP that infect organoids are protected from gentamycin treatment and proliferate. EcN organoid-colonisation is visualized by fluorescence microscopy. Biofilm and growth assays were performed using tumour tissue from our ΔAKP orthotopic mouse model, normal colon tissue from control mice and organoid lysates. Lysate was generated by mechanical homogenization in 1× PBS, cleared by centrifugation and normalised to protein concentration. Biofilm and growth assays were performed using M9 minimal media supplemented with tissue or organoid lysate or equivalent volume of vehicle (V). To determine if CRC cells produce a protein that promotes EcN biofilm formation, CRC organoid lysate or V was digested by Proteinase K or denatured by boiling.

Results: Our organoid-microbe co-culture assay allows us to observe EcN tumour-homing behaviour in real time. We observed clear EcN colonisation of CRC organoids, but not NC organoids. Our in vitro biofilm assays show increased EcN biofilm formation in the presence of CRC tissue and organoid lysate in comparison to NC tissue and organoid lysates, without increased bacterial growth (Figure 1). Organoid lysate protein denaturation and digestion did not affect biofilm formation (data not shown), suggesting that CRC cells may produce small molecules rather than proteins that promote EcN biofilm formation.

160

Cost-effectiveness analysis of comprehensive strategies for Barrett’s esophagus screening

Tomonori Aoki¹, Norma Bulamu² and David Watson³

¹College of Medicine and Public Health, Flinders University, Adelaide, Bedford Park, Australia; ²Flinders Health and Medical Research Institute, College of Medicine and Public Health, Flinders University, Adelaide, Bedford Park, Australia; ³Department of Surgery, Flinders Medical Centre, Adelaide, Bedford Park, Australia

Background and Aim: Whilst Barrett’s esophagus (BE) surveillance is well established, screening for BE is generally not supported, even though clinical outcomes are likely to be better if cancer/dysplasia in BE is detected at an early stage. Potential tools for screening include endoscopy, less-invasive non-endoscopic devices, and non-invasive risk stratification models. We aimed to develop and evaluate a cost-effectiveness model that would allow different strategies for BE screening to be evaluated for potential clinical application in the Australian context.

Methods: Analysis was conducted using two hypothetical cohorts of the general population aged ≥50 years (BE prevalence of 1.9% and 6.8%). Four categories of screening tools were included i) risk stratification based on non-weighted clinical factors according to US/European BE surveillance guidelines, ii) weighted risk stratification using algorithmic models, iii) less-invasive non-endoscopic devices such as Cytosponge-trefoil factor 3 (TFF3), and iv) endoscopy. Using a decision-analytic model, the cost per BE case identified and the incremental cost-effectiveness ratio (ICER) were calculated across six potential strategies for BE screening: i+iv, ii+iv, iii+iv, i+iii+iv, ii+iii+iv, and only iv.

Results: The cost per BE case identified was lowest in the weighted stratification (followed by Cytosponge-TFF3 and then endoscopy) strategy at AUS$14,129 and AUS$5,180, at 1.9% and 6.8% BE prevalence, respectively. The most cost-effective strategy was the weighted stratification (followed by Cytosponge-TFF3 and then endoscopy) strategy at a BE prevalence of 1.9%. Further, the Cytosponge-TFF3 (followed by endoscopy) strategy was more cost-effective (ICER=AUS$20,681) at a BE prevalence of 6.8%. ICERs were also sensitive to the less-invasive devices’ costs and the diagnostic accuracy of weighted stratification.

168

Butyrate impacts polyp initiation in familial adenomatous polyposis: results of a randomised, double-blind, placebo-controlled crossover trial

Finlay Macrae¹, Julie Clarke⁷, Trevor Lockett¹⁰, Karen Harrap⁹, Virginia Bird¹, Brooke Flanders¹, Alex Boussioutas⁶, Mark Appleyard⁸, David Williams¹¹, Sophie Zaloumis³, Suresh Sivanesan¹, Aysha Al Ani², Arun Gupta², Ralley Prentice⁵, Don Cameron⁴, Allan Spigelman¹², Digsu Koye³ and Patrick Lynch¹³

¹Departments of Colorectal Medicine and Genetics, Royal Melbourne Hospital, Parkville, Australia; ²Gastroenterology, Royal Melbourne Hospital, Adelaide, Australia; ³School of Population and Global Health, University of Melbourne, North Ryde, Australia; ⁴Department of Gastroenterology, Royal Children's Hospital, Herston, Australia; ⁵Department of Gastroenterology, Monash Medical Centre, Parkville, Australia; ⁶Department of Clinical Genetics and Genomics, Alfred Hospital, Melbourne, Australia; ⁷CSIRO Health and Biosecurity, Adelaide, Australia; ⁸Department of Gastroenterology, Royal Brisbane and Women's Hospital, Herston, Australia; ⁹CSIRO Health and Biosecurity, Brisbane, Australia; ¹⁰CSIRO Health and Biosecurity, Parkville, Australia; ¹¹Gastroenterology and Hepatology Parkville, Australia; ¹²Surgery, St Vincent's Hospital, Sydney, Australia; ¹³MD Anderson Cancer Center Department of Gastroenterology, Hepatology and Nutrition, Houston, USA

Background and Aim: Butyrate may reduce risk of colorectal cancer (CRC) and can be delivered to the colon using butyrylated starch (HAMSB). This trial evaluated the effects of HAMSB on polyp burden in participants with Familial Adenomatous Polyposis (FAP).

Methods: The study was a randomised, double-blind, placebo-controlled crossover trial in FAP. Participants ingested 40g of either HAMSB (Ingredion, Bridgewater NJ, USA), or low amylose starch (LAMS; Ingredion, Bridgewater NJ, USA) for six months, followed by the alternative product for six months, and then a six month washout. Participants underwent video-recorded colonoscopies at baseline, six, 12 and 18 months to assess polyp burden, and for biopsy collection. At baseline, two colonic tattoos were placed: tattoo 1 where polyps were cleared at each scope; and tattoo 2, where polyps were left in situ for the entire study if safe. The primary endpoint was global number of colorectal polyps. Secondary endpoints included the number of small (<2.4 mm), medium (2.4-9 mm) and large (>9 mm) polyps in the large bowel and the overall total and number of small, medium and large polyps in tattoo areas 1 and 2. A subset of 14 participants collected faecal samples for analysis. Generalised linear mixed models were used to estimate the ratio of mean polyp counts in intervention compared to placebo periods controlling for relevant variables and accounting for missing data. Based on concordance analyses only the assessment of global polyp counts and sizes for the CIA were used in the study analysis, whereas the burden of polyps in tattooed areas 1 and 2 were counted by two gastroenterologists independently.

Results: 72 participants were randomised (33 female) with 49 participants completing the study. In the intention to treat analysis HAMSB reduced the global number of colonic polyps by 10% (95% CI: 0.77-1.06) and number of small polyps by 12% (95% CI: 0.71-1.1). The total number of polyps in tattoo 1 was reduced 22% (95% CI: 0.58-1.05, P=0.106) with a 28% reduction in number of small polyps (95% CI: 0.5-1.03, P=0.074). Polyp burden in tattoo 2 was not consistently affected by treatment. In the per protocol population the effect on global and small polyp numbers were slightly stronger. The mean global count was 13% reduced when HAMSB was ingested compared to LAMS (95% CI: 0.73-1.05, P=0.152) and the mean count of small polyps was 21% reduced (95% CI: 0.62-1.0, P=0.051), medium polyps was 13% increased (95% CI: 0.67-1.91, P=0.655) and large polyps was 15% decreased (95% CI: 0.14-5.27, P=0.860). There was evidence of carryover of treatment effects between periods. HAMSB increased faecal butyrate concentration compared to LAMS with high esterified butyrate levels indicating bacterial release of free butyrate was saturated.

Conclusion: HAMSB delivers significant quantities of butyrate to the colon of FAP participants and ingestion tends to reduce the growth of small polyps without causing regression or growth of existing polyps. The carryover effect suggests protection extends after supplementation has ended. Although we could not establish an effect at a 0.05 level of significance, we were encouraged by the consistency of the observations in the direction of protection. Future biopsy analysis may determine if the reduction in polyp burden was due to cellular apoptotic or proliferative responses to colonic butyrate. HAMSB supplementation has potential to reduce the risk of CRC development in FAP patients and this may extrapolate to protection from sporadic CRC in the wider community where low dietary fibre intake is a major factor contributing to the high incidence of this disease.

Clinical trial: https://www.anzctr.org.au/ Number: 12612000804886

170

Autophagy and sex differences in gastric inflammation and microbiota

Isidora Simovic¹, Karla Vinasco¹, Nadeem Kaakoush² and Natalia Castano Rodriguez¹

¹School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Kensington, Australia; ²School of Biomedical Sciences, UNSW Sydney, Kensington, Australia

Background and Aim: Over 50% of the global population is estimated to be infected with Helicobacter pylori, the leading cause of gastric cancer (GC). Yet, most infected individuals remain clinically silent, inferring the contribution of host (i.e., genetic) and environmental (i.e., diet, smoking) factors, as well as the overall non-H. pylori microbiota in gastric carcinogenesis. Autophagy is an intracellular degradative pathway that carries critical roles in mediating host innate immunity, inflammation, and tumour suppression. Anti-microbial autophagy, also known as xenophagy, can be directed towards invasive pathogens such as H. pylori. The germline mutation ATG16L1 rs2241880 (A > G; Thr300Ala) leads to a loss of function-like phenotype, causing defective autophagy execution. We have previously shown rs2241880 to significantly increase the independent risk of H. pylori infection and gastric carcinogenesis in Han Chinese, Australian Caucasian, and Dutch populations. We now aimed at investigating the underlying biological mechanisms contributing to rs2241880-related GC pathophysiology.

Methods: Gastric microbiota surveying was performed on 10 GC patients and 136 controls from a high-risk Han Chinese population using 16S rRNA (V4) gene amplicon sequencing to assess the influence of rs2241880 on microbiota diversity and composition. Further, the host gastric transcriptome was assessed in a population subset using RNA-Seq. Our in vitro model utilized CRISPR/Cas9 to generate knock-in AGS daughter cell lines representing all three rs2241880 genotypes (AA, AG, GG). Both edited and non-edited epithelial cells were challenged with H. pylori GC26 (vacA+ s1m1, cagA+) where inflammatory, autophagic and lysosomal activity were evaluated.

Results: We observed an opposing sex specific rs2241880 influence on gastric microbiota diversity; in females, richness was negatively associated with rs2241880 carriage (coefficient: - 6.14, p-value: 0.002), while conversely, in males, we observed no effect on richness but a positive association with both evenness (coefficient: 0.08, p-value: 0.01) and Shannon’s index (coefficient: 0.4, p-value: 0.01). In males, rs2241880 was independently associated with enrichment of Capnocytophaga (p-value: 0.02) and a decrease in Haemophilus (p-value: 0.02). In females, rs2241880 was independently associated with enrichment of both Rothia (p-value: 0.0003) and Lautropia (p-value: 0.0008). rs2241880-carrying gastric epithelial cells exhibited reduced autophagic and lysosomal activity during acute H. pylori infection. In parallel, an aberrant inflammatory response was observed with exacerbated IL-8 but reduced TNF-α and IFN-β production.

Conclusion: Gastric microbiota surveying revealed sex differences in the influence of rs2241880 on diversity. Further, host genetic-microbial interaction analysis revealed sex-specific taxon associations. ATG16L1 rs2241880 carriage elicited an abnormal inflammatory response coupled with a disrupted autophagic response to H. pylori infection in gastric epithelial cells. Here, we provide mechanistic insight into how ATG16L1 rs2241880 influences H. pylori-mediated gastric carcinogenesis via dysregulated inflammation and altered gastric microbiota diversity and composition.

178

Impact of diverticulosis and polypectomy on risk of developing metachronous colorectal cancer

James Fiori¹, Steven Kim¹, Marina H Wallace², Samantha Rankin³ and Oyekoya Ayonrinde^1,4

¹Department of Gastroenterology and Hepatology, Fiona Stanley Hospital, Murdoch, Australia; ²Department of Colorectal Surgery, Fiona Stanley Hospital, Murdoch, Australia; ³Department of Clinical Services, Fiona Stanley Hospital, Murdoch, Australia; ⁴Medical School, The University of Western Australia, Crawley, Australia

Background and Aim: Metachronous colorectal cancer (CRC) occurs in up to 3% of patients following surgical resection¹. Diverticulosis has been proposed as a risk factor for missed lesions and therefore metachronous CRC by rendering colonoscopies technically more difficult and impairing views of affected mucosal areas². Another putative cause of metachronous CRC is tumour seeding into the mucosa when polypectomy is performed during the colonoscopy when CRC is diagnosed (index colonoscopy). Previous studies have identified identical molecular signatures between primary and metachronous tumours in cases where recurrence occurred at the site of previous polypectomy, and viable tumour cells have been shown to be present in the working channel of colonoscopes after tumour biopsy³. There is a lack of data regarding the impact of synchronous polypectomy and diverticular disease on metachronous CRC risk. As such, we aimed to determine whether diverticulosis and performing polypectomy at index colonoscopy increased the risk of developing metachronous CRC in an Australian cohort.

Methods: Single centre retrospective case-control study of adults who underwent surgical resection for CRC over a two-year period (January 2016 to December 2017). Colonoscopy details collected were the location of CRC, whether polypectomy was performed, location of polyps removed, and the presence and location of diverticulosis. Clinical records up to 5 years post-surgery were reviewed to identify cases of metachronous CRC, defined as recurrence occurring between 6 months and 5 years post resection. The proportion of metachronous CRC in patients who underwent polypectomy during the index colonoscopy was compared to those who did not, and between those who had diverticulosis present and those who did not.

Results: Our study population comprised 225 patients with median (IQR) age 71 (60-77) years. There were 8 (3.6%) metachronous CRCs; 1 (0.4%) at the site of endoscopic mucosal resection for a 15mm sessile serrated lesion, 3 (1.3%) anastomotic site CRCs, and 4 (1.8%) at other sites within the colon. Most metachronous CRCs occurred in the rectum (38%). There was no significant difference in the proportion of metachronous CRC in patients who underwent polypectomy at the index colonoscopy compared with those who did not (1.9% vs. 5.1%, p=0.283). Diverticulosis was reported in 33.3% (left-sided 24.4%, pancolonic 7.1% %, unstated location 1.8%). There was no significant difference in proportion of metachronous CRC between those with diverticulosis and those without (p=0.274).

Conclusion: The risk of developing metachronous CRC was not increased by polypectomy or diverticulosis at index colonoscopy.

References

1. Hassan, C., Wysocki, P. T., Fuccio, L., Seufferlein, T., Dinis-Ribeiro, M., Brandão, C., & Ponchon, T. (2019). Endoscopic surveillance after surgical or endoscopic resection for colorectal cancer: European Society of Gastrointestinal Endoscopy (ESGE) and European Society of Digestive Oncology (ESDO) Guideline. Endoscopy, 51(03), 266-277.

2. Troelsen, FS, Sørensen, HT, Erichsen, R. Risk of a post-colonoscopy colorectal cancer in patients with diverticular disease: A population-based cohort study. Endoscopy. 2024; 56: 471-481.

3. Backes, Y., Seerden, T. C., Gestel, R. S., Kranenburg, O., Ubink, I., Schiffelers, R. M., & Moons, L. M. (2019). Tumor seeding during colonoscopy as a possible cause for metachronous colorectal cancer. Gastroenterology, 157(5), 1222-1232.

204

Accuracy of prose CT reports for staging of pancreatic cancer

Lin Li¹, Hasna Kazi¹, Charles Pilgrim^1,2 and John Zalcberg¹

¹Cancer Research Program, School of Public Health and Preventive Medicine, Monash University, Melbourne, Australia; ²Department of Surgery, Alfred Health, Melbourne, Australia

Background and Aim: The current standard of care for radiological reporting of CT scans for pancreatic ductal adenocarcinoma (PDAC) is the traditional narrative-style prose report, a method prone to omission of relevant negatives and poor consistency in reporting of anatomical characteristics important to an operating surgeon. Standardisation of definitions for resectability is outlined in the International Consensus Guidelines. From these guidelines, 63 radiological features that determine the resectability of PDAC have been derived; these provide the basis for determining the adequacy CT scans for pancreatic cancer. This study aimed to assess the comprehensiveness of the current prose reporting against the Guidelines.

Methods: 150 CT scan reports from patients with PDAC were collected from 19 hospitals in 5 states across Australia from July-December 2023, as part of the SCANPatient clinical trial. All patients had been presented at HPB multidisciplinary meetings held in respective institutions. Their comprehensiveness was assessed by determining how many of the 63 benchmark features defined in the Guidelines were adequately addressed in the report text.

Results: Less than half the relevant fields were appropriately addressed in the standard prose report (41%, 10.7 of an average 26 relevant fields, as not all fields were relevant for each case). Less than 35% of reports adequately addressed the tumour-vessel relationship with the SMA or SMV. The coeliac artery was addressed in just 30 reports, and terminology used in the prose reporting was frequently non-standardised and vague.

214

Exploring the biomarker potential of circulating small extracellular vesicles in pancreatic cancer

Arunima Panda¹, Ilaria Casari¹, Abir Halder², Walid Abu Shawish^2,3, David Greening⁴, Danielle Dye¹, Krish Ragunath^1,2 and Marco Falasca^1,5

¹Curtin Health Innovation Research Centre, Curtin Medical School, Curtin University, Perth, Australia; ²Department of Gastroenterology and Hepatology, Royal Perth Hospital, Victoria Square, Perth, Australia; ³Department of Gastroenterology, Alfred Health, Melbourne, Australia; ⁴Molecular Proteomics, Baker Heart and Diabetes Institute, Melbourne, Australia; ⁵Department of Medicine and Surgery, University of Parma, Via Volturno 39, Italy

Background and Aim: Pancreatic cancer (PC) is a highly aggressive malignancy characterised by a lack of early symptoms, delayed diagnosis, and prone to resistance to conventional chemotherapy, leading to high fatality rates. CA19-9 is the only available biomarker used in routine clinical practice that has a sensitivity of ~ 80% in symptomatic patients but is not useful as a screening tool. It is crucial to focus on early detection and effective treatment options to address the challenges in managing PC. Circulating small extracellular vesicles (sEVs) play a role in cell-to-cell communication in cancer and contribute to tumour progression and metastasis. This study aims to identify specific proteins in sEVs that can be used as biomarkers to improve the diagnostic landscape for PC patients (Fig. 1).

Methods: Extensive proteomics and phenotypic analysis of sEVs was performed from peripheral blood samples in PC patients (n=19), chronic pancreatitis (n=3), and healthy participants (n=10). Ethical approval was obtained for using human blood samples for this study (Royal Perth Hospital RGS4208, Curtin Ethics HRE2021-0757). The identified proteins from proteomics-based mass spectrometry had a 40% stringent cutoff threshold that was applied to further refine the list.

Results: The proteomic landscape of PC, pancreatitis, and healthy individuals reveals distinct molecular signatures, providing valuable insights into the disease pathogenesis. Our analysis, based on a stringent selection criterion (ANOVA p-value < 0.05 and Z-score normalization), yielded a clustered heatmap of proteins that distinguish each group. ALDH1A1, SAA2, and GGT1 were present in sEV derived from PC patients compared to healthy and chronic pancreatitis-derived sEVs. Further, the presence of certain proteins in both PC and chronic pancreatitis samples including, C8B, C8G, ADH4, CPA1, ACAA2 and PIPOX suggest a shared molecular response to pancreatic pathology.

229

A study of incidence of young colorectal cancers in Geelong: Are the numbers increasing?

Daniel Yee Lee Ng, Sarah Taylor, Phillip Te, Kabir Ahmad, Sina Alexander and Jonathan Watson

Barwon Health, Geelong, Australia

Background and Aim: Colorectal cancer (CRC) is a significant cause of mortality and morbidity in Australia, estimated to be the 4^th most diagnosed cancer and the 2^nd most common cause of cancer-related death. Whilst the implementation of the National Bowel Cancer Screening Program (NBCSP) in 2006 has led to a decline in CRC mortality rates in those aged >50 in Australia, growing evidence demonstrates a rise in incidence of CRC in younger populations aged <50 years. This study aims to observe the incidence of CRC in the 30-49 year age group in the Geelong region over a 5-year period between 2017-2021. Secondary objectives include comparing this incidence with national data and stage of disease at time of diagnosis.

Methods: Data was collected via diagnostic coding and electronic medical file review. All patients aged 30-49 years who had undergone colonoscopy at Barwon Health and were diagnosed with CRC between 2017-2021 were included in the study. Datapoints collected include indication, age at time of colonoscopy, gender, pathology/histological diagnosis, stage of disease (where available) and survival for up to one-year post-colonoscopy (where available). Patients with known hereditary CRC syndromes were excluded.

Results: A total of 10,102 patients aged between 30-49 years underwent colonoscopy at Barwon Health between January 2017 and December 2021 (2,138 in 2017; 2,191 in 2018; 2,568 in 2019; 1,605 in 2020; 1,600 in 2021). 21 patients were diagnosed with colorectal adenocarcinoma (3 in 2017; 5 in 2018; 5 in 2019; 2 in 2020; 6 in 2021). 8 patients were diagnosed with metastatic disease (2 in 2017; 2 in 2018; 1 in 2019; 3 in 2021) and 7 with nodal disease (1 in 2017; 1 in 2018; 1 in 2019; 2 in 2020; 2 in 2021). 4 patients were diagnosed with neuroendocrine tumours (2 in 2018; 2 in 2020). 1 patient was diagnosed with lymphoma in 2021; making a total of 26 colorectal cancers diagnosed within this 5-year period. Age-specific rates (per 100,000) were calculated for the population of Geelong and found to be lower in comparison to national CRC rates published by the Australian Institute of Health and Welfare. There are several potential reasons for this. Firstly, the presented data only represents cancers diagnosed within the public hospital system and it is possible other patients were treated at private hospitals and thus not captured by this data collection method. Secondly, it does not include patients that proceed to surgery without colonoscopy. Thirdly, it does not include patients with metastatic disease that are diagnosed with liver biopsy without proceeding to colonoscopy.

232

Clinical utility of the faecal immunochemical test as a colonoscopy triage strategy in patients with iron deficiency

Jennifer Pham¹, Geraldine Laven-law¹, Rachel Parker¹, Esme Jasko², Jean Winter¹, Peter Bampton³, Robert Fraser^1,4, Charles Cock^1,4 and Erin Symonds^1,4

¹Flinders University, College of Medicine and Public Health, Flinders Health and Medical Research Institute, Bedford Park, Australia; ²Department of Gastroenterology and Hepatology, The Queen Elizabeth Hospital, Central Adelaide Local Health Network, Woodville, Australia; ³SA Group of Specialists, Kurralta Park, Australia; ⁴Department of Gastroenterology and Hepatology, Flinders Medical Centre, Southern Adelaide Local Health Network, Bedford Park, Australia

Background and Aim: Iron deficiency (ID) including ID with anaemia (IDA) can be caused by bleeding from gastrointestinal lesions such as colorectal cancer (CRC) and advanced pre-cancerous neoplasia (APCN). Most guidelines recommend bi-directional endoscopy in men and post-menopausal women with IDA, but recommendations for younger women is unclear. The latest (August 2023) United Kingdom National Institute for Health and Care Excellence (NICE) guidelines suggest that the faecal immunochemical test (FIT) can be used for IDA patients of all ages for triaging to colonoscopy. It is important to determine the performance of FIT in young individuals (<50 years) with ID/IDA. We evaluated the diagnostic performance of FIT for advanced colorectal neoplasia (ACN) in young and older South Australians undergoing colonoscopy for ID/IDA.

Methods: Adults (aged ≥18 years) referred for colonoscopy due to ID or IDA were prospectively invited to complete a two-sample FIT (OC-Sensor, Eiken Chemical Company, Japan) within 6 months before colonoscopy. Patients with incomplete colonoscopies and procedures with poor or unknown quality were excluded unless ACN was present. Patients with inflammatory bowel disease were excluded. ACN was defined as CRC and APCN (adenomas ≥10mm, with high-grade dysplasia or villous change, ≥5 adenomas, sessile serrated lesions ≥10mm, and/or serrated lesions with dysplasia). FIT was considered positive at ≥10 Hb/g faeces. Multivariable logistic regression was used to determine demographic and clinical factors associated with ACN. Odds ratios (OR) are reported with 95% confidence intervals (CI) and p-values <0.05 were considered statistically significant.

Results: 173 ID/IDA individuals were included, 46 (26.6%) aged <50 years. 34 individuals had ACN (19.7%), including 6 (3.5%) with CRC and 28 (16.2%) with APCN (6 aged <50 years). No cases of CRC were found in individuals aged <50 years. FIT positivity rate for all ages was 29.5% (table). Using a positive FIT as a triaging tool, 70% of colonoscopies could be avoided or delayed while detecting all cases of CRC and 50% of APCN. The number of individuals needed to scope to detect CRC was 9 compared to 29 without using FIT. A positive FIT was the only factor associated with ACN, independent of age, sex, and colonoscopy history; OR 5.3 (95% CI 2.1–13.5, p<0.001).

Conclusion: A positive FIT can detect CRC in ID/IDA patients, independent of age. FIT may be used to triage patients for colonoscopy to detect causes of ID/IDA which include ACN.

Table: Performance of FIT for detecting ACN in ID/IDA patients.

233

The diagnostic performance of the faecal immunochemical test for detecting advanced colorectal neoplasia in patients with iron deficiency: a systematic review and meta-analysis

Jennifer Pham¹, Geraldine Laven-law¹, Erin Symonds^1,2, Molla M Wassie¹, Charles Cock^1,2 and Jean Winter¹

¹Flinders University, College of Medicine and Public Health, Flinders Health and Medical Research Institute, Bedford Park, Australia; ²Department of Gastroenterology and Hepatology, Flinders Medical Centre, Southern Adelaide Local Health Network, Bedford Park, Australia

Background and Aim: The faecal immunochemical test (FIT) is used in colorectal cancer (CRC) screening programs for asymptomatic individuals. There is increasing interest in using the FIT to triage patients with lower gastrointestinal symptoms and iron deficiency anaemia (IDA) for colonoscopy. Some studies have shown that FIT has a high accuracy for CRC (90% sensitivity and 87% specificity) in symptomatic patients, but there is limited evidence on FIT’s performance for iron deficiency (ID), including IDA and non-anaemic ID (NAID). The aim of this systematic review was to evaluate the diagnostic performance of FIT for advanced colorectal neoplasia (ACN), including CRC and advanced pre-cancerous neoplasia (APCN) in ID patients.

Methods: MEDLINE, Embase, and Web of Science were searched for studies published after 2010 that used a quantitative FIT to detect ACN against the reference standard colonoscopy in adult patients (≥18 years) with ID. Random effects meta-analyses were used to determine the diagnostic performance of FIT for CRC and APCN, reported as pooled diagnostic odds ratio (DOR) with 95% confidence intervals (CI). When studies reported multiple FIT thresholds, the primary FIT threshold and largest cohort were used. Summary receiver operator curves (SROCs) were generated with pooled sensitivities and specificities.

Results: Nine studies were included in the final analysis. One study included a NAID cohort and seven reported IDA as a separate group. 8/9 studies (FIT thresholds ranging 4-20 μg Hb/g faeces) reported data for CRC with a DOR of 44.2 (95% CI 26.3–74.4, Figure). 8/9 studies (FIT thresholds ranging 4-35 μg Hb/g faeces) reported data for APCN with a pooled DOR of 4.2 (95% CI 3.1–5.7). SROC analysis showed an area under the curve (AUC) of 0.93 for CRC, and 0.70 for APCN. FIT detected CRC and APCN in ID patients with 90.7% and 49.3% sensitivity, and 81.0% and 82.4% specificity, respectively. Using five studies, FIT had a sensitivity of 88.0% and specificity of 83.4% for detecting CRC in patients with IDA at a FIT threshold of 10 μg Hb/g faeces.

323

Managing colonoscopy workloads: using faecal immunochemical tests to identify individuals who do not have colorectal cancer

Graeme Young¹, Erin Symonds² and Geraldine Laven-law¹

¹Flinders University, Adelaide, Australia; ²Flinders Medical Centre, Adelaide, Australia

Background and Aim: Efficient deployment of colonoscopy resources is desirable. A proven approach is to select those most likely to have colorectal cancer (CRC) or advanced precursor lesions (APL) by screening using a faecal immunochemical test (FIT), in which a positive test is followed up by colonoscopy. However, current FIT configurations miss one-quarter to one-half of CRCs. Recent evidence suggests that those with very low faecal haemoglobin (f-Hb), measured by quantitative FIT (qFIT) have a low chance of CRC (Int J Cancer 2023;152:1536). Thus, qFIT could be used to rule out or delay colonoscopy in symptomatic, surveillance, and/or average-risk screening populations. Our aim was to identify the f-Hb levels that identified a very low risk for CRC.

Methods: f-Hb levels were determined from a one-sample FIT completed in a surveillance population (n=32,485 tests; OC-Sensor, Eiken Chemical Co., Japan). FIT accuracy for advanced neoplasia (CRC and/or APL) was estimated at different f-Hb positivity thresholds in those who had undergone colonoscopy within 12 months after FIT (n=4,110).

Results: The colonoscoped population included 94/4,110 (2.3%) with CRC, 603/4,110 (14.7%) with APL, and the remainder with non-significant or no pathology. For f-Hb levels below 20 μg/g faeces (the Australian screening threshold; a in Figure), the population positivity rate (and hence colonoscopy workload) rose much faster than the sensitivity for advanced neoplasia. At the test limit of detection (LoD; 1.8 μg/g faeces; b in Figure), sensitivity for advanced neoplasia increased to 63.7%, however, 19.1% of CRC would have been missed, and 72.2% would not have proceeded to colonoscopy. By reducing the threshold to 0.6 μg/g faeces, fewer than 10% of CRC would have been missed and 78.3% of advanced neoplasia would have been detected, while avoiding colonoscopy in 48.6% of the population. f-Hb was undetectable in 59.4% of the population, including 5/94 (5.3%) with CRC and 108/603 (17.9%) with APL.

330

Evaluation of pancreatic volume and morphology as a prediction of postoperative pancreatic insufficiency

Fiona Jones

St. Vincent's Hospital Melbourne, Melbourne, Australia

Background and Aims: Pancreatic exocrine insufficiency (PEI) and pancreatogenic diabetes (T3cDM) are common following pancreatic surgery with respective incidence of 46-100% and 3-40%, depending on resection type. Improved survival following pancreaticoduodenectomy (PD) and distal pancreatectomy (DP) necessitate the identification of factors that may modulate long-term sequelae of pancreatic dysfunction, and the effect on quality of life. There is limited data on the relationship between pancreatic volume and post-operative outcomes. Pre-operative risk stratification may allow more comprehensive pre-operative counselling, with improved post-operative testing and treatment. The aims of this study were to evaluate the relationship between pre- and post-operative pancreatic volume (PV) and morphology with the development of PEI and pancreatogenic diabetes.

Methods: We carried out a retrospective study of all patients who underwent pancreatic surgery at St. Vincent’s Hospital Melbourne over a 6 year period from 2016 to 2022. Clinical and demographic data were collected by retrospective electronic chart review. Patients were contacted directly, where required, to assess diabetic status and the requirement for pancreatic enzyme replacement therapy. Multiplanar 5mm slice arterial, portal venous and non-contrast CT images of the abdomen pre and post-surgery were analysed. Pancreatic volumetry measurements were obtained using GE HealthCare Volume Viewer by two specialist gastrointestinal radiologists.

Results: A total of 106 patients were included in the analysis (M=56, F=50). 73.5% patients underwent PD with the remainder undergoing extended DP or DP. Mean pre-operative PV was 68.02 (SD 33.31) and mean post-operative PV was 32.22(SD 19.75); p<0.001. Mean reduction in PV was 35.79 (SD 27.23). There was a significant reduction in pancreatic duct diameter following pancreatic surgery (pre-op 4.12, post op 3.13, p<.002). There was a significant correlation with PV loss post-operatively and parenchymal attenuation on the pre-operative CT. Patients requiring pancreatic enzyme replacement therapy had significantly reduced post operative pancreatic volume (29.04 vs 38.97, p=0.032). In-patients without a pre-existing diagnosis of DM there were 8 cases of de-novo DM (10.7%). No significant association was identified between post-operative pancreatic volume and new onset DM or increase in DM medical therapy.

337

Combined phosphorus-32 implantation and chemotherapy: a comparison with standard therapy using a propensity-score weighted landmark analysis and an assessment of its impact on vascularity in locally advanced pancreatic cancer

Amanda H Lim^1,2, Darshan Nitchingham¹, Jana Bednarz^2,3, Madison Bills⁴, Laxmi Lanka⁵, Berry Allen⁶, Alvin Tan⁶, Rohit Joshi⁷, William Hsieh⁴, Benjamin Crouch⁴, Joshua Zobel¹, John-Edwin Thomson⁸, EuLing Neo⁸, Romina Safaeian¹, Edmund Tse^1,2, Christopher Rayner^1,2, Andrew Ruszkiewicz^2,9,10, Jayden Wong¹¹, Nimit Singhal¹², Dylan Bartholomeusz^1,4, Frank Weilert¹³ and Nam Nguyen^1,2

¹Department of Gastroenterology and Hepatology, Royal Adelaide Hospital, Adelaide, Australia; ²University of Adelaide, Adelaide, Australia; ³SAHMRI Women and Kids Theme, South Australia Health and Medical Research Institute, Adelaide, Australia; ⁴Department of Nuclear Medicine, Royal Adelaide Hospital, Adelaide, Australia; ⁵Department of Radiology, Waikato Hospital, Hamilton, New Zealand; ⁶Department of Nuclear Medicine, Waikato Hospital, Hamilton, New Zealand; ⁷Medical Oncology, Lyell McEwin Hospital, Elizabeth Vale, Australia; ⁸Department of Hepatobiliary Surgery, Royal Adelaide Hospital, Adelaide, Australia; ⁹Surgical Pathology, SA Pathology, Adelaide, Australia; ¹⁰Centre of Cancer Biology, University of South Australia, Adelaide, Australia; ¹¹Department of Oncology, Waikato Hospital, Hamilton, New Zealand; ¹²Department of Oncology, Royal Adelaide Hospital, Adelaide, Australia; ¹³Department of Gastroenterology, Waikato Hospital, Hamilton, New Zealand

Background and Aims: Pancreatic cancer is highly lethal. Poor intra-tumour vascularity contributes to its limited response to chemotherapy. The combination of standard chemotherapy and endoscopic ultrasound (EUS)-guided phophorus-32 (³²P) microparticle intra-tumoural implantation has revealed encouraging results in locally advanced pancreatic cancer (LAPC). However, comparative studies are lacking. Therefore, we compared chemotherapy and ³²P implantation with standard therapy using a propensity-score weighted analysis (PSWA). We also aimed to assess changes in pancreatic tumour vascularity following ³²P implantation, using contrast enhanced-EUS (CE-EUS).

Methods: We conducted a retrospective cohort study comparing LAPC patients with combination therapy at 2 centres with standard therapy patients from a single centre, from August 2017 to January 2023. Landmark analysis was used to address immortal time bias. PSWA was applied to reduce the impact of selection bias. The primary outcome was overall survival at 24 months after first-line treatment initiation, with treatment effect expressed as restricted mean survival time (RMST; average event-free survival time). In a sub-cohort of patients with combined therapy, CE-EUS was performed just prior to, and at 4 and 12 weeks after implantation. Time intensity curve was analysed for 90 seconds after IV contrast bolus to ascertain peak intensity and intensity gain.

Results: 104 patients were considered. The landmark date was designated as 3 months after initiation of first-line chemotherapy. After excluding patients who died before the landmark, 101 patients were included in the PSWA (35 vs. 66 chemotherapy only). The RMST within 24 months after chemotherapy initiation is estimated to be 112 days longer for patients with combination therapy (459 days, 95%CI 393-536) compared to chemotherapy only (347 days, 95%CI 308-392). The restricted mean local progression free time within 24 months is estimated to be 112 days (95%CI 36-187) longer and the probability of downstaging is 22.3% higher (95%CI 5.12-39.5, p=0.03) with combination therapy. Eighteen and fifteen patients continued to have follow-up CE-EUS at 4 weeks and 12 weeks respectively, post-implantation. Baseline (pre-implantation, post-chemotherapy) median intensity gain of contrast enhancement within the tumour increased from 32.15 (IQR 18.08-54.35) to 46.85 (IQR 35.05-76.6; p=0.007) and 66.3 (IQR 54.7-76.3; p=0.001) 4 weeks and 12 weeks post-implantation respectively.

Conclusion: This is the first comparative study between chemotherapy and ³²P implantation and standard therapy in patients with LAPC, demonstrating survival, disease control and downstaging benefits. The increased microvascular flow with combined therapy likely allows more delivery of chemotherapy to the tumour and thus, could explain its encouraging outcomes.

354

Synchronous nervous system and gastric diffuse large B cell lymphoma: a case report

Thant Zaw¹, Ajish Radhamma^1,2, Sharon Avery¹ and Montri Gururatsakul^1,2

¹Cairns Hospital, Cairns North, Australia; ²James Cook University, Smithfield, Australia

Introduction: Gastric diffuse large B-cell lymphoma (DLBCL) is a prevalent form of gastric lymphoma known for its aggressive nature, usually responsive to treatment. However, managing synchronous central nervous system (CNS) DLBCL can be particularly challenging due to limited treatment options. This is a case report of a patient with synchronous CNS and systemic (gastric) DLBCL.

Case report: A 58-year-old male patient presented with a constellation of perplexing symptoms, including progressive subtle abnormal movements and unexplained weight loss. He denied any significant past medical history, and gastrointestinal symptoms. MRI revealed multiple cerebral lesions, some displaying haemorrhagic features, subsequently PET scan illuminated hypermetabolic activity within both supratentorial and infratentorial cranial regions, coupled with gastric entanglement (Picture 1). Upper gastroscopy found a non-bleeding cratered gastric ulcer, measuring 10mm (Picture 2), biopsy results confirmed a malignant cellular infiltrate dominated by diffuse large B-cell lymphoma, distinguished by a staggering proliferation index of 90%. He was commenced on chemotherapy regimen, with the aggressive induction chemotherapy in anticipation of autologous stem cell transplantation. Regrettably, the treatment course was marred by a series of serious severe complications following the second cycle of chemotherapy, with severe infection, underwent an above-knee amputation and multiple debridement. Despite exhaustive interventions, the patient's clinical trajectory went downward rapidly, prompting a shift towards palliative care centred on comfort and symptom control, and he peacefully passed away.

360

A rare but important cause of biliary tree dilation: Case report and review of the varying manifestations of intraductal papillary mucinous neoplasm of the bile duct

Hao Ting Liao¹ and Arjuna Somasundaram²

¹Gold Coast University Hospital, Gold Coast, Australia; ²Royal Brisbane and Women's Hospital, Brisbane, Australia

Introduction: Intraductal papillary mucinous neoplasms of the bile duct (IPMN-B, IPMN-BT) are a rare entity in western countries but have higher prevalence in eastern countries, attributed to the endemic nature of hepatolithiasis and clonorchiasis [1,2]. They are considered a biliary variant of intraductal papillary mucinous neoplasms of the pancreas (IPMN-P) though have a much higher malignant potential than the pancreas variant, at up to 41-83% [3]. This case report demonstrates the imaging findings of IPMN-B across various imaging modalities and highlights the importance of early diagnosis due to its premalignant nature.

Case report: A 70-year-old female presents with deranged LFTs and imaging findings of marked intra-hepatic and extra-hepatic duct dilatation. Following investigation with endoscopic retrograde cholangiopancreatography (ERCP), endoscopic ultrasound (EUS) and intra-operative cholangiogram in the setting of laparoscopic cholecystectomy, suspicion for IPMN-B was raised on an MRI liver with gadoxetic-acid (Primovist). Repeat ERCP and spyglass confirmed the presence of a villiform mass in the right main hepatic duct, later histologically confirmed as IPMN-B.

Conclusion: Intraductal papillary mucinous neoplasm of the bile duct are rare phenomena in western countries. This case study highlights direct and indirect findings of IPMN-B on several modalities, as well as the importance of recognising IPMN-B early.

References

1. Park, SG, Baek, DH, Kim, GH, Heo, J, Song, GA, Ahn, SJ, et al. Intraductal papillary mucinous neoplasms of the bile duct treated with argon plasma coagulation. Korean Journal of Pancreas and Biliary Tract. 2017; 22(1): 39–45. https://doi.org/10.15279/kpba.2017.22.1.39

2. Park, HJ, Kim, SY, Kim, HJ, Lee, SS, Hong, GS, Byun, JH, et al. Intraductal papillary neoplasm of the bile duct: Clinical, imaging, and pathologic features. American Journal of Roentgenology. 2018; 211(1): 67–75. https://doi.org/10.2214/ajr.17.19261

3. Nakayama, Y, Tomino, T, Ninomiya, M, Minagawa, R, et al. Recurrent intraductal papillary neoplasm of the bile duct due to intraductal dissemination: a case report and literature review. Surg Case Rep. 2021; 7(1): 238. https://doi.org/10.1186/s40792-021-01318-0

391

Effects of exosomes derived from pancreatic cancer and stellate co-cultures on insulin signalling pathway factors: Potential mediators of pancreatic cancer-related diabetes

Chamini Perera^1,2, Patrick Wu^1,2, Dinuki Perera^1,2, Tanzila Khan^1,2, SM Zahid Hosen^1,2, Alpha Raj Mekapogu^1,2, Zhihong Xu^1,2, David Greening^3,4,5,6, Suresh Chari⁷, Ron Pirola^1,2, Jeremy Wilson^1,2 and Minoti Apte^1,2

¹UNSW, Sydney, Australia; ²Ingham Institute for Applied Medical Reserach, Sydney, Australia; ³Baker Heart and Diabetes Institute, Melbourne, Australia; ⁴Baker Department of Cardiovascular Research, Translation and Implementation, La Trobe University, Australia; ⁵Central Clinical School, Monash University, Melbourne, Australia; ⁶Baker Department of Cardiometabolic Health, University of Melbourne, Australia; ⁷Department of Gastroenterology, Hepatology and Nutrition, MD Anderson Cancer Centre, University of Texas, Houston, USA; ⁸Metabolic Signalling Group, Curtin Medical School, Curtin Health Innovation Research Institute, Curtin University, Perth, Australia

Background: Pancreatic cancer (PC)-related diabetes (PCRD) is thought to be a paraneoplastic phenomenon driven by the developing cancer. Pancreatic stellate cells (PSCs) are present around PanINs (earliest lesions of PC) and play a key role in cancer progression. However, their role in PCRD is unknown.

Hypothesis and Aim: Factors carried by exosomes from cancer cells and PSCs impair β cell function and insulin signalling in peripheral cells (adipocytes/hepatocytes) leading to PCRD. We previously demonstrated that mouse PSC+PC-exosomes significantly impaired insulin secretion by mouse β (MIN6) cells and insulin signalling in hepatocytes (AML12). Thus, in this study, we aimed to i) assess the effects of PSC+PC-exosomes on insulin signalling in mouse adipocytes (3T3L1), ii) identify the exosome cargo that may mediate the effects on insulin signalling and iii) validate the role of selected proteins in insulin signalling using AML12 cells.

Methods: Exosomes isolated from co-cultured mouse PSCs and PC (KPC) cells (PSC+KPC-Ex, n=3-4 preparations) or acinar cells (controls) and from co-cultured human PSCs and AsPC1 (cancer cells). 3T3L1 cells (n=4 preps) incubated with PSC+KPC-Ex and assessed for insulin signalling pathway factors [(Insulin Receptor (IR), IR substrate (IRS), Akt, Glut4)] by immunoblotting. Proteomic analysis of exosomes and parent cells performed (LC-MS-MS) followed by Gene Set Enrichment Analysis (GSEA). Of the specific proteins identified, G6PD expression in KPC cells modulated using siRNA techniques; exosomes isolated, incubated with AML12 cells and effects on insulin signalling assessed.

Results: PSC+KPC-Ex significantly decreased pIR/IR and pAkt/Akt ratios in adipocytes (Fig 1A-B), compared to acinar-Ex (p<0.05). PSC+KPC-Ex i) carried proteins known to modulate insulin secretion [Glucose-6-phosphate dehydrogenase (G6PD), Gluose-6-phosphate isomerase (G6PI), Lactate Dehydrogenase-A (LDHA) and Aldolase B]; and ii) were enriched in proteins known to regulate insulin signalling (Clusterin, IGFBP3, Regucalcin), compared to their parental cells. Intriguingly, human PSC+AsPC1-Ex also demonstrated enrichment of Aldolase B (associated with insulin secretion) and Clusterin and Regucalcin (related to insulin signalling). Furthermore, inhibition of G6PD in KPC-Ex resulted in significantly increased IR expression (p<0.01) in AML12 cells compared to Acinar-Ex treated AML12 cells (Fig 1C).

393

Transcriptomic analysis of exosome cargo derived from pancreatic cancer and stellate cells: prospective mediators of pancreatic cancer-related diabetes

Helen Binang^1,2, Wilson Wong³, Tanzila Khan², Anandwardhan Hardikar³, Zhihong Xu^1,2, Marco Falasca⁴, Jerry Greenfield⁵, Ron Pirola^1,2, Jeremy Wilson^1,2, Chamini Perera^1,2 and Minoti Apte^1,2

¹Pancreatic Research Group, South Western Sydney Clinical Campuses, School of Clinical Medicine, Faculty of Medicine and Health, UNSW Sydney, Sydney, Australia; ²Ingham Institute for Applied Medical Research, Sydney, Australia; ³Western Sydney University, Sydney, Australia; ⁴Curtin University, Perth, Australia; ⁵St Vincent Clinical School, Faculty of Medicine and Health, UNSW Sydney, Sydney, Australia

Background: The diagnosis of pancreatic cancer (PC) is preceded in 34% of patients by a diagnosis of diabetes 3-5 years earlier [pancreatic cancer-related diabetes (PCRD)], which could be a harbinger of asymptomatic PC. Pancreatic stellate cells (PSCs) produce PC stroma and interact with cancer cells to promote disease progression.

Hypothesis: PSC-PC cell interactions promote PCRD by secreting factors (proteins, RNA, lipids) carried by exosomes that impair islet cell function and peripheral insulin signalling.

Aim: To identify the RNA cargo within exosomes derived from PC cells and PSCs cultured alone or together.

Methods: Cells that were cultured alone or together included: i) mouse PSCs and KPC cells (PC cell line) (PSC+KPC) (n=4/group); ii) Cancer-associated human PSCs (CAhPSCs) and AsPC1 (human PC cell line) (CAhPSC+AsPC1) (n=3/group). Plasma samples were obtained from KC mice fed a high fructose diet (HFrD) that induces impaired glucose tolerance and accelerated cancer progression (KC-HFrD; KC-control n=4/group). Exosomes were isolated using ultracentrifugation, RNA was extracted and sequenced on the Illumina sequencing platform.

Results: Compared to KPC-exosomes, PSC+KPC-exosomes showed significant upregulation of 742 and downregulation of 5986 mRNAs. Compared to AsPC1-exosomes, CAhPSC+AsPC1-exosomes showed upregulation of 782 and downregulation of 6855 mRNAs. Interestingly, 18mRNAs upregulated in mouse PSC+KPC-derived exosomes were also found to be upregulated in human CAhPSC+AsPC1-derived exosomes, while 2490 downregulated mRNAs were common to both species. Of the 18 upregulated and 2490 downregulated mRNAs in co-culture-derived exosomes, 7 were upregulated and 227 were downregulated in KC-HFrD plasma exosomes compared to KC-controls, showing an overlap of the expression of these mRNAs in both in-vitro and in-vivo settings. Moreover, high expression of 1 of the 7 upregulated mRNAs (PICALM) and low expression of 46 of the 227 downregulated mRNAs were associated with poor survival of PC patients. The downregulated mRNAs, DPH5, LPIN1, TRIM27, PRKAG2, PTCD3, PGAP1, ABCC5, HIPK2, CMTM4, and U2AF2 are of particular interest because they have been reported to be mediators of diabetes and/or PC.

Conclusion: We have shown for the first time that the exosomal cargo from PSC-PC interactions may have diabetogenic and tumorigenic effects. Detailed characterisation of these mRNAs may identify novel biomarkers/therapeutic targets for PC.

406

Distal colonic crypt hyperplasia is not adversely impacted by emu oil and saireito in a murine model of colitis-associated colorectal cancer

Stephanie Thomson^1,2, Tahlia Kennewell³, Gordon Howarth^1,4,5 and Suzanne Mashtoub^1,2,4,6,7

¹Gastroenterology Department, Women's & Children's Hospital, North Adelaide, Australia; ²College of Medicine and Public Health, Flinders University, Bedford Park, Australia; ³Future Industries Institute, University of South Australia, Mawson Lakes, Australia; ⁴School of Biomedicine, Faculty of Health and Medical Sciences, University of Adelaide, Adelaide, Australia; ⁵School of Animal and Veterinary Sciences, Faculty of Sciences, Engineering and Technology, University of Adelaide, Roseworthy, Australia; ⁶Discipline of Surgery, The Queen Elizabeth Hospital, Woodville, Australia; ⁷School of Medicine, University of Western Australia, Perth, Australia

Background and Aim: Ulcerative colitis is an inflammatory bowel disease characterised by chronic inflammation of the rectum and colon, which can lead to the development of colitis-associated colorectal cancer (CA-CRC). Previously, we demonstrated therapeutic potential of the nutraceuticals Saireito, a traditional Japanese Kampo medicine, and Emu Oil, derived from emu fat, in reducing overall tumour burden in a murine model of azoxymethane (AOM)/dextran sulphate sodium (DSS)-induced CA-CRC (Chartier et al. 2021). We aimed to investigate the mechanism for this decrease in tumorigenesis by examining the impact of Emu Oil and Saireito, alone and in combination, on colonocyte kinetics in a mouse model of AOM/DSS-induced CA-CRC.

Methods: Female C57BL/6 mice (n = 10/group) were intraperitoneally injected with either saline or the carcinogen AOM (7.4mg/kg) on day 0, followed by ad libitum access to either plain water or three cycles comprising 7 days of DSS followed by 14 days of water. Mice were gavaged thrice weekly for 9 weeks with water (80μL), Emu Oil (80μL), Saireito (1g/kg), or a combination of Emu Oil and Saireito (160μl; 80μL Emu Oil + 80μL Saireito). Sections of proximal and distal colon were stained with haematoxylin and eosin for assessment of crypt depth, colonocyte count, and colonocyte diameter. Data were presented as mean ± SEM. p<0.05 was considered statistically significant.

Results: Crypt depth in the distal colon increased in mice with AOM/DSS-induced CA-CRC (151 ± 12μm) compared with healthy controls (94 ± 3 μm; p<0.001). Administration of Emu Oil (149 ± 10μm), Saireito (152 ± 8μm), or a combination thereof (151 ± 10μm) did not further increase crypt depth in CA-CRC mice (p>0.05). Treatments had no impact on crypt depth in healthy mice (p>0.05). Colonocyte count in the distal colon increased in CA-CRC controls (33 ± 1.9 cells/crypt) compared with healthy controls (21 ± 0.6 cells/crypt; p>0.001). There was no further impact (p>0.05) on distal colon cell count in CA-CRC mice treated with Emu Oil (32 ± 1.9 cells/crypt), Saireito (32 ± 1.4 cells/crypt), or Emu Oil/Saireito (36 ± 2 cells/crypt). Colonocyte count was unaffected in healthy mice (p>0.05). There was no significant change in average colonocyte diameter in the distal colon of CA-CRC controls compared with healthy controls (p>0.05). Similarly, mean colonocyte diameter remained unchanged among treatment groups (p>0.05). In the proximal colon, crypt depth tended to increase in CA-CRC controls (161 ± 15μm) compared to normal controls (116 ± 8μm), although this just failed to attain statistical significance (p=0.066). There was no significant difference in colonocyte count or mean colonocyte diameter between control groups (p>0.05). Emu Oil, Saireito, or Emu Oil/Saireito treatment did not impact proximal colonic crypt depth, colonocyte count, or mean colonocyte diameter in either healthy or CA-CRC mice (p>0.05).

Conclusion: Increased crypt length in the distal colon of mice with CA-CRC was primarily attributed to a process of crypt cell hyperplasia, as opposed to hypertrophy, evidenced by an increase in colonocyte count with no change in colonocyte diameter. Treatment with Emu Oil, Saireto, or combined Emu Oil/Saireito in experimental CA-CRC did not significantly impact crypt depth, colonocyte count, nor colonocyte size in the proximal or distal colon, suggesting that the decrease in tumour burden described previously may have occurred through a different process. Future studies should investigate the impact of Emu Oil and Saireito on expression of tumour biomarkers such as p53, APC and K-ras, which are intrinsically associated with CRC development.

412

Use of high-resolution colonoscopy to accurately monitor disease progression in a murine model of colitis-associated colorectal cancer

Suzanne Mashtoub^1,2,3,4,5, Sisanda Mhlanga^1,2, S George Barreto^4,6, Paul Hammond^1,7 and Gordon Howarth^1,2,8

¹Gastroenterology Department, Women's & Children's Hospital, North Adelaide, Australia; ²School of Biomedicine, Faculty of Health and Medical Sciences, University of Adelaide, Adelaide, Australia; ³Discipline of Surgery, The Queen Elizabeth Hospital, Woodville, Australia; ⁴College of Medicine and Public Health, Flinders University, Bedford Park, Australia; ⁵School of Medicine, University of Western Australia, Perth, Australia; ⁶Flinders Medical Centre, Bedford Park, Australia; ⁷Discipline of Paediatrics, University of Adelaide, Adelaide, Australia; ⁸School of Animal and Veterinary Sciences, Faculty of Sciences, Engineering and Technology, University of Adelaide, Roseworthy, Australia

Background and Aim: Colonoscopy is a gold standard technique to survey and diagnose various diseases of the colon and rectum, including inflammatory bowel diseases (IBD) and colorectal cancer (CRC). Numerous mouse models of IBD and CRC have been developed to understand disease pathogenesis, monitor response to novel therapies and for pre-operative assessment, though often require euthanasia at various timepoints for monitoring. The azoxymethane (AOM)-dextran sulphate sodium (DSS) model of colitis-associated CRC accelerates the progression of tumour development, inducing colonic tumours that share the histopathological characteristics of human CRC. Herein, we review techniques for monitoring colitis and tumour progression in AOM/DSS-induced colitis associated-CRC in mice.

Methods: Female C57BL/6 mice were injected with AOM (7.4mg/kg; intraperitoneal) followed by ad libitum access to three DSS (7 days)/water (14 days) cycles. Colitis and tumour progression were monitored by colonoscopy on days 20, 41 and 62 using the Coloview miniendoscope system (Karl Storz, Tuttlingen, Germany), which includes a rigid straight forward colonoscope (1.9mm diameter) with operating sheath, air pump for insufflation, light source, camera and video monitor. The colonoscope was advanced to the splenic flexure at approximately 4cm and videos recorded upon withdrawal. Videos were assessed for colitis severity including colon thickening, vasculature pattern, fibrin, mucosal surface granularity and stool consistency, and colonic tumours counted. On the day of euthanasia (day 63), a total colectomy was performed and longitudinally opened to visualise and photograph tumours (Canon, Tokyo, Japan). Tumour number and size were determined using Olympus Soft Imaging Solutions software (Tokyo, Japan).

Results: In healthy control mice, there was an unobstructed circumferential view of the colonic mucosal surface which consistently appeared translucent (thin), shiny and smooth, without evidence of prominent vasculature or disrupted patterns; visualised stool was formed and mobile. In AOM/DSS mice, signs of mucosal inflammation were present from day 20, including less translucency hence mucosal thickening, loss of vasculature pattern, occasional bleeding, adherence of loose stool and evidence of small tumour development (approximately 10 per mouse). By day 40, mice demonstrated mucosal surface granularity with complete loss of translucency, fibrin development, bleeding and diarrhoea, often impairing circumferential view; this required intermittent and careful colonic clearance with saline. On day 62, visualisation was largely impaired in the setting of ongoing diarrhoea and near-impassable tumours, prompting careful navigation around tumours and colonic clearance using saline. On the day of euthanasia, direct visualisation of longitudinally opened colons revealed tumours concentrated in the rectum and distal colon; tumour numbers directly correlated with colonoscopic assessment, highlighting that colonic tumours were confined to the anatomical location distal to the splenic flexure.

Conclusion: High-resolution colonoscopy represents an important, accurate and cost-effective technique for repeated monitoring of colitis activity and CRC development in mice. Advanced assessment can also be employed using temporal biopsies to analyse immunological and molecular parameters. Moreover, this technique lends itself to the evaluation of other colorectal diseases, including diverticular and fistulating disease, to promote the development of potential new therapies.

413

Participation, positivity and polyps: bowel cancer screening in people aged 40 to 49 years

Erin Symonds^1,2, Geraldine Laven-law², Charles Cock^1,2, Molla M Wassie², Maddison L Dix² and Graeme Young²

¹Department of Gastroenterology and Hepatology, Flinders Medical Centre, Bedford Park, Australia; ²College of Medicine and Public Health, Flinders University, Bedford Park, Australia

Background and Aim: Colorectal cancer (CRC) screening in Australia is provided through faecal immunochemical tests (FIT) to individuals aged 50-74y. Recent guideline changes mean that Australians aged 45-49y can now request FITs from the National Bowel Cancer Screening Program (NBCSP), and those aged 40-45y may request tests from their GP. As NBCSP participation is lowest in those aged 50-54y, it is unclear what the participation will be in younger ages, or what the colonoscopy findings will be after a positive FIT. This study compared FIT participation in people aged 40-49y to older age groups, and determined positivity rates and yields at colonoscopy.

Methods: As NBCSP FITs have not yet been provided to individuals <50y, we analysed data (2011-2019) from a surveillance colonoscopy program (SCOOP) that also provides FITs as an interval screening modality. Individuals were provided two-sample FITs (Eiken Chemical Company, Japan) between colonoscopies, using the same brand and haemoglobin positivity threshold as used in the NBCSP. Positive FITs were followed up with colonoscopy. Colonoscopy outcomes were assessed for advanced neoplasia (CRC, advanced adenomas and high-risk serrated lesions). Statistical analyses were with Chi-square tests and logistic regression, with p<0.05 considered significant.

Results: 15,726 FITs were provided, including 1,424 to ages 40-49y (51.7% female), 4,662 to 50-59y (49.8% female) and 9,640 to 60-74y (47.4% female). Both FIT participation and FIT positivity were lowest in those aged 40-49y compared to the older age groups (Table, p<0.01 for both observations). Within the age groups 50-59y and 60-74y, FIT participation was significantly associated with female sex and the number of FITs previously completed, while in the 40-49y cohort, there was no association with sex, but participation was more likely with a higher socioeconomic status and with previous FIT completion (p<0.05, Table). Yield of advanced neoplasia at good quality follow-up colonoscopy (n=750) was similar across age groups, with FIT positive predictive value of 10.0% for 40-49y, 10.1% for 50-59y, and 12.7% for 60-74y (p>0.05).

Conclusion: FIT participation and positivity rates are lowest in people aged 40-49 years, but incidence of advanced neoplasia at follow-up colonoscopy is comparable with older age groups. Participation increases with prior exposure to FITs, which may be challenging with the NBCSP as the younger cohort need to request a screening test. Appropriate education and endorsement are needed to support CRC screening engagement in younger individuals.

Table. FIT positivity and predictors of participation by age group.

456

Post-colonoscopy colorectal cancer rate at a tertiary referral hospital in New Zealand

Jonas Harder, Mehul Lamba and Teresa Chalmers-Watson

CDHB, Christchurch, New Zealand

Background and Aim: Colorectal cancer (CRC) is the third most common cancer in the world. Screening colonoscopy is one of the most common tools to identify suspicious lesions. Post-colonoscopy CRC (PCCRC) is one of the emerging metrics to assess quality of colonoscopy service. The aim of our study was to determine the PCCRC rate and perform root-cause analyses at a tertiary referral hospital in Canterbury, New Zealand and to examine factors associated with development of PCCRC.

Methods: All patients undergoing colonoscopy from 2018 to 2023 were matched with CRC cases diagnosed in Canterbury, New Zealand. PCCRC-3Y cases (CRC developing within 6-36 months of a cancer negative colonoscopy) were identified and adjudicated by two gastroenterologists. Detailed case-note analysis was performed. The data was analysed by descriptive statistics.

Results: A total of 963 patients with CRC were diagnosed from December 2018 to December 2023. Of these, 20 patients developed PCCRC at an average interval of 602 days (95% CI 461-742) after a cancer-negative colonoscopy. The PCCRC rate from 2019-2020 was 2.18% (95%CI 0.77%3.59%). 16/20 cases were likely missed lesions despite adequate examination. Nine cases were interval type PCCRC (detected before interval screening), 7 type B (detected after recommended surveillance) PCCRC and 4 type C (detected when no surveillance was recommended). Eight of these 20 cases were stage III/IV cancers. Eight cases were metachronous CRCs. Seven cases demonstrated mismatch repair deficiency, of which 5 were sporadic cases with BRAF V600E mutation.

Conclusion: In our tertiary referral colonoscopy cohort, PCCRC-3Y rate was 2.18% (95%CI 0.77%-3.59%). Eighty percent of all PCCRC cases were likely due to missed lesions despite adequate examination.

483

Lymph node metastases in early gastric cancer: the Japanese Gastric Cancer Treatment Guidelines can be safely applied for endoscopic submucosal dissection in the west

Edward Young^1,2, Louisa Edwards^2,3, Andrew Ruszkiewicz^1,3,4 and Rajvinder Singh^1,2

¹Lyell McEwin Hospital, Northern Adelaide Local Health Network, Elizabeth Vale, Australia; ²Faculty of Health and Medical Sciences, University of Adelaide, Adelaide, Australia; ³Royal Adelaide Hospital, Central Adelaide Local Health Network, Adelaide, Australia; ⁴SA Pathology, Adelaide, Australia

Background and Aim: There is considerable disparity in the incidence of gastric cancer between eastern and western countries, with many eastern countries now participating in nationwide gastric cancer screen programs.¹ Consequently, endoscopists in high-incidence regions are more frequently identifying and treating endoscopically resectable early gastric cancers. This has led to the development of risk-stratification models to identify appropriate lesions for endoscopic resection, with the most recent 6^th Edition of the Japanese Gastric Cancer Treatment Guidelines (2021) now considering en bloc endoscopic resection to be indicated for all endoscopically intramucosal adenocarcinomas that are well differentiated and non-ulcerated, ulcerated but well differentiated and ≤3cm, or non-ulcerated undifferentiated lesions ≤2cm.² After resection, this guideline defines endoscopic curability according to the eCura system, with eCura A and B lesions being considered endoscopically cured and not requiring additional therapy.³ We sought to assess whether these criteria can be safely applied in a western population where data is lacking.

Methods: Data was retrospectively recorded for all patients who underwent any form of gastrectomy in four Australian Public Hospitals between 2000 and 2021. Demographic data, lesion characteristics (size, differentiation, invasion depth, lymphovascular invasion and ulceration) as well as the presence and number of lymph node metastases was recorded. Those given neoadjuvant chemotherapy were excluded from the study.

Results: A total of 1,465 patients were included in this study, including 558 patients who underwent gastrectomy for gastric adenocarcinoma without neoadjuvant chemotherapy (median age 70, 64.2% male). Of these, only 5.4% (n=30, CI 3.8-7.6%) were T1a and 18.4% (n=101, CI 15.4-21.9%) were T1 overall. Based on the Japanese Gastric Cancer Treatment Guidelines, 11.5% of lesions (n=64, CI 9.1-14.4%) met absolute criteria for endoscopic resection, with 7.8% of these (n=5, CI 3.4-17%) having positive lymph nodes at gastrectomy. According to the eCura system, 9.9% (n=55, CI 7.6-12.6%) of lesions would have been considered eCura A or B based on their histology. None of these eCura A or B lesions had positive lymph nodes at gastrectomy.

Conclusion: The eCura system for endoscopic curability could have been safely applied in this western population. Even in western countries, patients with early gastric cancer meeting Japanese guidelines for endoscopic resection should where possible undergo en bloc endoscopic submucosal dissection. Lesions that meet eCura A or B criteria histologically should be considered endoscopically cured. These patients require close ongoing endoscopic and/or imaging surveillance but do not require additional treatment after endoscopic resection.

References

1. Bray, F, Ferlay, J, Soerjomataram, I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018; 68: 394-424.

2. Japanese Gastric Cancer Treatment Guidelines 2021 (6th edition). Gastric Cancer 2023; 26: 1-25.

3. Hatta, W, Gotoda, T, Oyama, T, et al. A Scoring System to Stratify Curability after Endoscopic Submucosal Dissection for Early Gastric Cancer: "eCura system". Am J Gastroenterol 2017; 112: 874-881.

522

Evolving management strategies in gastrointestinal immunotherapy related adverse events

Alice Kerkham¹, Su Win², Salayman Mousa³, Calvin Chan¹, John Park² and Jeff Chang¹

¹Nepean Hospital Gastroenterology Department, Sydney, Australia; ²Nepean Hospital Oncology Department, Sydney, Australia; ³University of Sydney, Sydney, Australia

Background and Aim: The use of Immune Checkpoint Inhibitors (ICIs) have been demonstrated to significantly improve progression free-survival in a number of malignancies. Immune related adverse effects (IRAEs) are, however, an increasingly prevalent unwanted outcome, and now a widely recognised entity, with several guideline-based management strategies published. Gastrointestinal IRAEs (GIIRAEs), including colitis and hepatitis are among two of the most reported IRAEs with combination ICI use accounting for 15% and 10% of all IRAEs respectively. We aimed to report the rates and severity of GIIRAEs for patients who had received combination immunotherapy with nivolumab and ipilimumab, and analyse the investigations, treatment regime and response in a single tertiary referral centre.

Methods: A retrospective cohort analysis was conducted of all patients receiving ipilimumab and nivolumab for melanoma, renal cell carcinoma and mesothelioma between the dates of 2019 to 2023. The local oncology database and electronic hospital records were reviewed. Baseline demographics, incidence, grade and treatment of GIIRAEs were collected.

Results: A total of 99 patients on combination ICI were identified. 26 patients (26%) had a GIIRAE of whom 15% (15/99) had diarrhoea and 11% (11/99) had hepatitis. 47% (7/15) of patients with diarrhoea had radiological or endoscopic confirmation of colitis (5) or enteritis (2). Of the 5 patients with Grade 3-4 diarrhoea, 3 (60%) required escalation to infliximab, all of which only had mild colitis evident at colonoscopy. 7/11 (64%) of patients had Grade 3-4 hepatitis, with only one patient requiring escalation to steroid sparing agent, despite only 57% (4/7) being commenced on higher dose intravenous steroids as recommended by current guidelines. 40% (10/26) of Grade 1-2 IRAEs were managed conservatively. Only 50% (13/26) of all identified cases had involvement of a gastroenterologist.

Conclusion: Management strategies for GIIRAEs continue to evolve, and our cohort highlights the heterogeneity in investigation and management practice over the course of four years. Based on our findings, symptom-based assessment of severity for diarrhoea associated IRAEs appears to be a better predictor for the need for escalation of therapy rather than severity of investigation findings. For hepatitis associated IRAEs, a less aggressive approach to steroid therapy compared to current published guidelines may be adequate. Further prospective studies with a larger cohort would be of value to confirm our observations.

525

Aetiological insights into early-onset colorectal cancer and adenoma tumourigenesis through genomic tumour mutational signature profiling

Dan Buchanan¹, Peter Georgeson¹, Alysha Prisc¹, Eric Joo¹, Khalid Mahmood¹, Romy Walker¹, Natalie Diepenhorst², Julie Toner¹, Finlay Macrae¹, Christophe Rosty^1,3, Ingrid Winship¹ and Mark Jenkins¹

¹University of Melbourne, Parkville, Australia; ²Monash University, Parkville, Australia; ³Envoi Pathology, Herston, Australia

Background and Aim: The incidence of early-onset colorectal cancer (EOCRC; diagnosed <50 years of age) is increasing in Australia and globally, the cause of which is unknown. The aim was to characterise the tumour mutational signature (TMS) landscape of EOCRC and early-onset-adenomas (EOAds) to identify the spectrum of mutational processes.

Methods: Non-hereditary, mismatch repair proficient EOCRCs and EOAds from the ANGELS study (n=92) and the Colon Cancer Family Registry (n=100) underwent tumour and matched germline whole exome sequencing. Single base substitution (SBS) and indel (ID) TMS were calculated with COSMIC v3.4 definitions, with individual signatures considered present when observed at ≥10% within a tumour.

Results: Of the 170 EOCRCs (mean age at diagnosis 38.5±7.9 years, 58.5% females), 39.2% were diagnosed between 18-35yrs, 39.2% between 36-45yrs and 21.6% between 46-55yrs with 31% located in the proximal colon, 37.4% distal and 31.6% in rectum. The most prevalent TMS with associated known aetiologies were SBS1 (94.7%), SBS3 (40%), ID2 (98.2%) and ID1 (79.4%). For TMS without an associated or known aetiology, SBS89 (27.1%), SBS17b (1.8%), ID5 (77.1%) and ID14 (27.6%) were the most prevalent. The prevalence of SBS89 decreased significantly with increasing age at diagnosis (p<0.0001) and SBS89 was associated with the BRAF p.V600E somatic mutation (p=0.0001). SBS88, associated with colibactin DNA damage from pks⁺E.coli bacteria, was enriched in EOAds compared with EOCRCs (13.6% versus 2.4%, p=0.03).

Conclusion: EOCRCs characterised by SBS89 mutational processes were associated with the earliest age at diagnosis and the somatic BRAF p.V600E mutation. EOAds demonstrated evidence of bacteria-related tumourigenesis.

538

Genomic profiling of Lynch syndrome-related colorectal cancers elucidates the differing pathways of tumourigenesis: implications for preventing post-colonoscopy colorectal cancer

Peter Georgeson^1,2, Daniel Vo^1,2, Romy Walker^1,2, Khalid Mahmood^1,2,3, Jihoon Joo^1,2, Mark Clendenning^1,2, Julia Como^1,2, Sharelle Joseland^1,2, Susan Preston^1,2, Marci Chai^1,2, John L Hopper⁶, Aung K Win⁶, Alex Boussioutas^10,11, Christophe Rosty^1,2,4,5, Finlay Macrae^7,8,9, Ingrid Winship^8,9, Mark Jenkins⁶ and Dan Buchanan^1,2,8

¹Colorectal Oncogenomics Group, Department of Clinical Pathology, Victorian Comprehensive Cancer Centre, The University of Melbourne, Parkville, Australia; ²Victorian Comprehensive Cancer Centre, University of Melbourne Centre for Cancer Research, Parkville, Australia; ³Melbourne Bioinformatics, The University of Melbourne, Melbourne, Australia; ⁴Envoi Specialist Pathologists, Brisbane, Australia; ⁵University of Queensland, Brisbane, Australia; ⁶Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, The University of Melbourne, Melbourne, Australia; ⁷Colorectal Medicine and Genetics, The Royal Melbourne Hospital, Parkville, Victoria, Australia; ⁸Genomic Medicine and Family Cancer Clinic, Royal Melbourne Hospital, Parkville, Melbourne, Victoria, Australia; ⁹Department of Medicine, The University of Melbourne, Parkville, Victoria, Australia; ¹⁰Department of Gastroenterology, The Alfred Hospital, Melbourne, Parkville, Australia; ¹¹Central Clinical School, Monash University, Melbourne, Australia

Background and Aim: The emergence of gene-dependent cancer risks in Lynch syndrome carriers and the potential for alternate pathways of tumourigenesis has implications for cancer prevention, with development of post-colonoscopy CRCs proposed as being pathway-related. We aimed to utilise genomic evidence to identify differing pathways of tumourigenesis in colorectal cancers (CRCs) from Lynch syndrome carriers.

Methods: From whole-exome sequenced CRCs, including 21 MLH1, 12 MSH2, 14 MSH6, and 13 PMS2 carriers and 148 non-hereditary mismatch repair proficient (MMRp) CRCs, we calculated tumour mutational burden (TMB), tumour mutational signatures, microsatellite instability (MSI), loss of heterozygosity (LOH), and somatic mutations in APC, CTNNB1, KRAS, TP53, and RNF43. The mutational contexts of somatic mutations in APC were utilised to infer the likely pathway to carcinoma, whether via the proficient adenoma pathway, deficient adenoma pathway or non-adenoma pathway.

Results: MSI levels were significantly lower in MSH6 carrier CRCs compared with CRCs from the other MMR gene carriers (MANTIS p=0.001, MSIsensor 2x10^-7, MSIseq 5x10^-6). CRCs from PMS2 carriers showed higher levels of mutational signatures ID1 (p=3x10^-14) and lower levels of ID2 (p=0.0006) and ID7 (p=0.001) compared with CRCs from MLH1, MSH2 and MSH6 carriers. TMB and neoantigen load were not significantly different between CRCs from the four MMR genes. The non-adenoma pathway was relatively prevalent in MLH1 (62%) and MSH2 (58%) carriers compared to MSH6 (14%) and PMS2 (23%) carriers, suggesting higher risk for non-detection and/or rapid progression to carcinoma.

Conclusion: CRCs from Lynch syndrome carriers demonstrated distinct genomic differences based on the defective MMR gene. Genomic evidence illustrated gene-specific genomic differences and gene-specific pathway differences with implications for CRC prevention.

552

Novel findings related to risk factors, clinicopathological characteristics and molecular features of colorectal cancers from people with serrated polyposis syndrome and evidence for its genetic aetiology

Dan Buchanan¹, Marci Chai¹, Eric Joo¹, Khalid Mahmood¹, Sharelle Joseland¹, Julie Arnold², Nathan Atkinson², Amanda Whelan¹, Andrew Metz³, Alex Boussioutas⁴, Finlay Macrae¹, Mark Jenkins¹, Ingrid Winship¹, Susan Parry² and Christophe Rosty⁵

¹University of Melbourne, Parkville, Australia; ²New Zealand Familial Gastroenterological Service, Auckland, New Zealand; ³Royal Melbourne Hospital, Parkville, Australia; ⁴Alfred Hospital, Melbourne, Australia; ⁵Envoi Pathology, Herston, Australia

Background and Aim: Serrated Polyposis Syndrome (SPS) is characterised by multiple serrated polyps in the colon and rectum, a familial component and increased risk of colorectal cancer (CRC) in the patient and their first-degree relatives (FDRs). Determining the risk factors for CRC development and their molecular phenotype will aid in CRC prevention. Similarly, identifying the genetic basis to SPS will support prevention and early-detection strategies for SPS.

Methods: 807 participants who met the 2010 or 2019 WHO criteria for SPS were recruited to the Genetics of Colonic Polyposis Study. Clinicopathological characteristics were collated from participants’ medical records. DNA mismatch repair (MMR) protein expression was assessed using immunohistochemistry and tumour DNA from 74 CRCs was tested for CpG Island Methylator Phenotype (CIMP), MLH1 methylation and for somatic BRAF p.V600E mutations. CRCs were categorised into three molecular subtypes: 1) MMR-deficient (MMRd)/BRAF p.V600E/CIMP-positive; 2) MMR-proficient (MMRp)/BRAF p.V600E/CIMP-positive, and 3) MMRp/BRAF-wildtype/CIMP-negative, with subtypes 1 and 2 considered serrated pathway CRCs and subtype 3 considered adenomatous pathway CRCs. 501 of the 807 SPS cases, including 24 relatives from 11 families, were selected from the Genetics of Colonic Polyposis Study for germline whole genome or exome sequencing (WGS/WES) based on one or more of the following criteria: young age at SPS diagnosis, high serrated polyp count, CRC diagnosis or family history of SPS. Predicted pathogenic variants (PPV) were defined by in silico prediction tools and gnomAD allele frequency <0.05%.

Results: Of 807 SPS cases, 132 (16.4%) developed 177 CRCs. CRCs were predominantly in females (66.7%), located in the proximal colon (54.9%), were MMRp (64.7%) and occurred at the time of SPS diagnosis (69.7%). SPS cases with CRC had older age at SPS diagnosis (54.6 ± 16.8yrs) when compared with cases without CRC (41.8 ± 14.9yrs, P=1.8e-15). CRC development was associated with a higher total polyp count (4^th v 1^st quartile, OR=4.0, 95%CI=2.3-7.2, P=1.2e-06), and the presence of extra-colonic cancers (OR=2.2, 1.1-4.4, P=2.4e-02). The presence of at least one traditional serrated adenoma (TSA; OR=4.9, 2.6-9.1, P=7.85e-07) or conventional adenoma (CA; OR=3.9, 2.3-7.1, P=2.23e-06) was associated with an increased CRC risk. Serrated pathway CRCs, comprised of molecular subtypes 1 (32.4%) and 2 (21.6%), were the most common (54%) compared with adenomatous pathway CRCs (46%). The adenomatous pathway CRCs were more common in males (P=2.49e-02), had a younger age at diagnosis (P=3.13e-04), and more commonly observed in the distal colon and rectum (P=7.87e-05), when compared with serrated pathway CRCs. Of the 501 SPS cases with germline WGS/WES (median age at SPS diagnosis=30 (IQR=25) years, median serrated polyp count of 34 (IQR=16), 66% were females, 97% white European, and 15% developed CRC. Pathogenic variants were identified in APC, MLH1, MSH2, MSH6 and FLCN genes. Of the reported SPS candidate genes in the literature, only PPVs in RNF43 (n=2), and CFTR (n=9), occurred in more than a single person with SPS. Fifteen PPVs segregated with FDRs with SPS in the 49 familial SPS cases identified representing novel candidate SPS genes.

Conclusion: This large study of SPS identified that CRC in SPS is heterogeneous with regards to phenotype, age at diagnosis and anatomical location and displays a complex aetiology related to multiple risk factors and molecular pathways of tumourigenesis with nearly half of CRCs in SPS developing via an adenomatous pathway. Candidate SPS genes from the literature did not validate in our study. The analysis of familial SPS cases identified multiple genes segregating PPVs potentially representing novel SPS predisposition genes.

555

Exploring unwarranted clinical variation in the management of early rectal cancer: Findings from an Australian hospital network

Anthony Whitfield^1,2, Anthony Sakiris^3,2, Julia Gauci^3,2, Clarence Kerrison^3,2, Sunil Gupta^3,2, Oliver Cronin^3,2, Timothy O'Sullivan^3,2, Brian Lam^3,2, Francesco V Mandarino^3,2, Varan Perananthan^3,2, Hunter Wang^3,2, Puma Sundaresan^3,2,4, Toufic El-Khoury^3,2, James Toh^3,2, Nimalan Pathmanathan^3,2, Eric Y Lee^3,2, Nicholas Burgess^3,2 and Michael J Bourke^3,2

¹Blacktown Hospital, Sydney, Australia; ²Westmead Clinical School, University of Sydney, Sydney, Australia; ³Westmead Hospital, Sydney, Australia; ⁴Sydney West Radiation Oncology Network, Sydney, Australia

Background and Aim: Radical resection in the rectum is associated with significant morbidity and poses a mortality risk, particularly in elderly or comorbid patients. Early or T1 rectal cancer is often curable and can be managed with local endoscopic resection techniques with reduced morbidity and negligible mortality. To establish a pathway for appropriate treatment of these early lesions through the multidisciplinary team (MDT), an early rectal cancer working party (ERCWP) was formed to evaluate management trends and outcomes within our hospital network. The ERCWP was multi-speciality comprising of gastroenterologists, colorectal surgeons and radiation oncologists.

Methods: A retrospective review was performed across four institutions that make up our network to identify all patients with a new diagnosis of localised (T1-T4,N0,M0) rectal or recto-sigmoid cancer between January 2015 and August 2023. Patients with confirmed adenocarcinoma staged as T1N0M0 or T2N0M0 were included in the analysis. Available electronic medical records from the time of index resection and all encounters for the following 12 months were reviewed. Patients were excluded if their treatment was completed at an outside institution. Patient outcomes were stratified by location of tumour (rectum/recto-sigmoid) and mode of resection (Endoscopic resection, Local surgical excision or Radical surgical resection). Major adverse events included any one of the following: ostomy, deep incisional site infection, abscess, wound dehiscence, pneumonia, sepsis, septic shock, acute renal failure, deep venous thrombosis, pulmonary embolus, stroke/cerebrovascular accident, myocardial infarction, ventilator >48 hours, still in hospital >30 days, cardiac arrest, reoperation, readmission and/or mortality.

Results: A total of 111 localized (T1/T2,N0) and 153 locally advanced (T3/T4,N0) lesions were newly diagnosed during the study period. 95/111 patients with localized disease were included (M=62/95, 65.3% F=33/95, 34.7%). Mean age was 67.8 years in the endoscopy group and 67.0 years in the radical surgical resection group. Mortality was 2.2% in the group of patients who had surgery (1/46). There was no mortality in patients who received endoscopic resection. Major adverse events were recorded in 16/46 (34.8%) of patients who underwent surgery. There were no major adverse events in patients who received endoscopic resection. Mean length of stay was 1.47 days in patients who had endoscopic resection and 10.2 days in patients who had radical resection. Mean length of stay was prolonged where there was a treatment related major adverse event (14.8 days, mean difference 9.8 days, p=<0.001). 78.5% (51/65) of the cohort with T1 cancer was low risk (well or moderately well differentiated and no lymphovascular invasion). Of the 16 patients with T1 rectal cancer who underwent definitive surgery, 15 (93.8%) were suitable for cure by en-bloc endoscopic excision. Following establishment of the ERCWP in 2020, there have been no cases of radical resection in the rectum for T1 cancer. For T1 cancer in the rectum, major adverse events were 5/34 (14.7%) pre ERCWP formation and 1/13 (7.7%) post ERCWP formation (p=<0.001).

569

The tumour microenvironment influences checkpoint expression by gamma delta T cell subpopulations in hepatocellular carcinoma

Paul Armstrong¹, Andreea Atanasescu², Tom Gallagher³, Emir Hoti³, Lydia Lynch⁴ and Cliona O'Farrelly²

¹University College Dublin, Ireland; ²School of Biochemistry & Immunology, Trinity Biomedical Sciences Institute, Trinity College Dublin, Dublin, Ireland; ³St. Vincent's University Hospital, Dublin, Ireland; ⁴Princeton University, USA

Background and Aim: γδ T lymphocytes are important tumour surveillance cells in healthy liver, exhibiting antitumour activity through diverse mechanisms. They could provide important immunotherapeutic targets for the treatment of HCC and liver metastases. Here we characterise γδ T cell populations in HCC and use an in vitro model of a hepatoma tumour immune microenvironment (TIME) to measure its effect on checkpoint expression by γδ T cells.

Methods: Ten HCC resections were studied: 3 with well differentiated and 7 with moderately differentiated tumours. Samples taken from three sites of each HCC resection tumour, tumour adjacent, and distal uninvolved liver, were processed immediately to provide single cell suspensions for analysis by flow cytometry. Results were analysed using Graphpad Prism.

Results: Significant populations of γδT cells were present at each site from 10 HCC resections. Higher populations of γδ2 T cells were present in tumour tissue (p=0.03) when compared with⁺γδ1 and γδ3 (p=0.02) T cells. PD1⁺γδ2 T cells were enriched in tumour compared to uninvolved liver tissue (p=0.03). Well differentiated tumours were enriched with PD1⁺γδ1 (p=0.03) and PD1⁺γδ3 subtypes (p=0.016) compared to moderately differentiated tumours. The adjacent and distal liver areas of well differentiated samples were enriched with a PD1⁺γδ3 population (p=0.001 and p=0.01 respectively) compared to moderately differentiated samples. Using an in vitro model of the primary hepatic tumour immune microenvironment (TIME), we found that the hepatic TIME significantly influenced expression of inhibitory immune checkpoint (PD-1, CTLA-4, LAG-3) by γδ T cell populations (Figure).

578

Polyp landscape in patients with colorectal cancer

Zainul Azhar, Hellen Kuo, Cathal McGowan and Nicholas Tutticci

QEII Jubilee Hospital, Coopers Plain, Australia

Background and Aim: Anecdotally, a proportion of patients with colorectal cancer (CRC) have no or very few polyps at the time of diagnosis or first follow up. There is little data on the proportion of CRC patients with this phenotype. We aimed to determine the synchronous polyp burden in CRC patients from a high volume colonoscopy center.

Methods: Consecutive CRC diagnosed at index colonoscopy and age-sex matched index colon controls from January 2014 to December 2023 were extracted from a colonoscopy quality database. As polyp clearance is not routinely performed at the time of CRC diagnosis, an accurate synchronous polyp burden can only be identified combining the polyp count index and first surveillance (clearing colon) within 12 months. The CRC patient polyp count was compared with the control index colon polyp burden, anticipating full polyp clearance in most cases. Left sided hyperplastic polyps were excluded from the polyp count.

Results: CRC was identified at 264 index colonoscopies the majority of which were distal (59.5%) predominately from the rectum (27%). The median age was 60 and half (48.1%) were female with 55% of patients stage ≥T3 disease at diagnosis. The most common CRC colonoscopy indications were positive FOBT (30%) or symptoms (29%). The median polyp count across all CRC cases was four. In the CRC cohort, 21.6% (57) had 0-1 polyps identified with 10.2% (27) having 0. In contrast to the matched control group which had 0-1 polyps in 55.3% of cases (32.5% 0 polyps; 21.0% 1 polyp). A low (0-1) synchronous polyp count in CRC patients was more common in females (26.7 vs 16.8% p=0.049) and in the distal colon. (23.57% vs 18.63%, p = 0.35).

579

Liver and peripheral blood immune populations differentially associate with outcome in advanced hepatocellular carcinoma treated with combined transarterial chemoembolization plus immune checkpoint blockade

Paul Armstrong¹, Lindsey Clarke², Stephen Stewart³, Austin Duffy² and Cliona O'Farrelly⁴

¹University College Dublin, Ireland; ²St. Vincent's University Hospital, Dublin, Ireland; ³Mater Misericordiae University Hospital, Dublin, Ireland; ⁴School of Biochemistry & Immunology, Trinity Biomedical Sciences Institute, Trinity College Dublin, Ireland

Background and Aims: The complexity of the liver immune environment is likely to influence and reflect tumour growth. Changes in the relative proportions of lymphoid and myeloid populations in the liver and peripheral blood may therefore be of useful prognostic value in patients with hepatocellular carcinoma (HCC). Here we aimed to evaluate liver and peripheral blood lymphoid and myeloid populations as prognostic biomarkers in patients with advanced HCC treated with immune checkpoint inhibitors (ICI) plus transarterial chemoembolization (TACE).

Methods: Patients with HCC (Childs Pugh A/B7; Barcelona Clinic Liver Cancer Stage B/C; ECOG 0/1; sorafenib-naïve or experienced) were enrolled in a clinical trial (UCDCRC/19/01 EudraCT no. 2019-002767-98) of tremelimumab plus durvalumab and subtotal TACE performed on week 6. The presence of neutrophils and lymphocytes were investigated using routine Hematoxylin and Eosin (H&E) stained section slides. Each patient had H&E slides from tumour, and background liver assessed by two experienced histopathologists independently. Neutrophil: lymphocyte ratio (NLR) was defined as the neutrophil count divided by the lymphocyte count in each x40 HPF. Peripheral blood NLR and platelet: lymphocyte ratio (PLR) were calculated from baseline bloods (pre-treatment) and again after the third dose of Durvalumab (on treatment). Primary outcome was defined as 6 month progression free survival (PFS) and secondary outcome was overall survival (OS).

Results: 13 patients took part in the study: 10 had adequate pretreatment pathology samples for assessment; 6 patients exhibited PFS at 6 months. Peripheral blood NLR and PLR, both associated with overall survival when measured either pre-treatment (p =0.019, p =0.02 respectively) or on treatment (p=0.008, p=0.02 respectively). In uninvolved liver, the presence of neutrophils in the parenchyma correlated with poorer one year OS (p=0.048). Patients with higher tumour infiltrating lymphocyte (TIL) numbers had poorer 6-month PFS (p=0.019). We also saw a strong association between higher TILs and higher peripheral AFP (0.017, r =0.8). Higher intratumoural NLR associated with longer PFS (p=0.008). The median PFS of patients whose intratumoural NLR was ≥0.08 was 8.7 (7-9.7) compared to 3.8 (3.1-5.3) in those whose NLR was <0.08.

Conclusion: Here, fewer TILs and a higher intratumoural NLR correlated with improved clinical response in advanced HCC treated with ICI plus TACE. However, increased neutrophils in background liver, and increased peripheral blood NLR and PLR correlated with poorer overall survival. These somewhat paradoxical findings support the hypothesis that neutrophils and lymphocytes are susceptible to polarization; and can have protumoural or antitumoral roles depending on their phenotypic state.