Techniques and Innovations in Gastrointestinal Endoscopy最新文献

Background and Aims

Several criteria have been developed for screening of Barrett's esophagus (BE). Previous studies are limited by analysis in highly selective populations. This study evaluated the performance of 4 criteria in identifying patients with BE from a primary care population.

Methods

This was a post-hoc analysis from a previous retrospective study reporting on BE screening rates of eligible patients in a large health system. Patients were included if they were screened in a primary care population for BE, were at risk for BE per 1 of the accepted criteria, and underwent screening with esophagogastroduodenoscopy. Sensitivity for BE was calculated for 4 screening criteria. These included criteria from the American College of Gastroenterology (ACG), American Society of Gastrointestinal Endoscopy (ASGE), European Society of Gastrointestinal Endoscopy (ESGE), and the Kunzmann model.

Results

In a primary care population, 1077 patients were screened for BE with an esophagogastroduodenoscopy. BE was found in 105 (9.75%) patients. Positive predictive values and negative predictive values of ACG/ESGE, ASGE, and Kunzmann's criteria for BE on pathology were 9.9% and 90.45%, 9.12% and 89.22%, and 10.84% and 92.29%, respectively. The area under the receiver operating characteristic curve for ACG/ESGE, ASGE, and Kunzmann's criteria was 50.49%, 52.23%, and 54.04%, respectively. There was no statistically significant difference in the presence of BE on pathology among patients who did or did not meet ACG/ESGE, ASGE, or Kunzmann's criteria for BE screening.

Conclusion

Current BE screening criteria need optimization in a primary care setting. Use of a combination of criteria might capture the highest number of BE patients.

Colorectal cancer is the third most common cancer in the United States and the second leading cause of cancer-related death. Colonoscopy with polypectomy reduces the incidence and mortality of colorectal cancer, and with the growing number of advanced endoscopic resection techniques, select malignant polyps can now be managed endoscopically. Optimal management of malignant colorectal polyps involves careful assessment of the polyp based on morphologic and surface features to determine the risk of submucosal invasion. Lesions with features of deep submucosal invasion typically require surgical resection, although in the absence of these features, superficially invasive lesions are often amenable to endoscopic resection with techniques such as endoscopic mucosal resection, endoscopic submucosal dissection, and, more recently, endoscopic full-thickness resection. These resection techniques should be performed by endoscopists trained in advanced resection modalities.

Background and Aims

Estimation of the solid component in walled-off pancreatic necrosis (WOPN) during endoscopic ultrasound (EUS) is an important parameter to guide the selection of stent for drainage but is often not objectively quantified.

Methods

We aimed to study the interobserver agreement (IOA) among endosonographers assessing debris in the pancreatic fluid collection and their choice of stent. EUS videos of 15 patients having pancreatic fluid collection with concurrent magnetic resonance imaging (MRI) assessment of debris were independently reviewed by 40 endosonographers for the percentage of debris (10% increments) and their choice of stent (plastic or metal). The Fleiss kappa (κ) coefficient was used to assess IOA. Post-hoc analysis was done using wider debris intervals (20% or greater increments).

Results

Poor agreement was observed for the percentage of debris (κ = 0.188), which did not improve (κ = 0.196) even after including only expert endosonographers with more than 10 years of experience (n = 33). There was fair agreement (κ = 0.266) with reference MRI on the percentage of debris. On post-hoc analysis, the IOA (κ = 0.56, substantial agreement) and agreement with MRI (κ = 0.59, substantial agreement) improved as the debris intervals widened (<10%, 10%-50%, >50%). The agreement for stent selection (plastic vs metal) was poor (κ = 0.174) and did not improve with case volume (κ = 0.153 among respondents with >25 EUS-guided drainage/year; n = 21) or years of experience (κ = 0.195 for >10-year experience; n = 33).

Conclusion

IOA between endosonographers regarding estimation of debris in WOPN and subsequent stent choice for drainage is poor. The experience of endosonographers did not improve IOA. Studies to standardize the EUS criteria for debris assessment in WOPN and subsequent therapeutic approaches are warranted.

Colonoscopy remains one of the most common procedures performed by gastroenterologists and is critical for early detection and management of precursors to colorectal cancer (CRC). Although CRC remains one of the deadliest malignancies, earlier detection of precancerous polyps is directly associated with increased patient survival. As such, quality metrics for colonoscopy, such as polyp detection and mucosal visualization, are key parameters that are directly tied to patient outcomes. Over the past 2 decades, artificial intelligence and machine learning (AI/ML) tools have been tested and developed to augment colonoscopy performance and in 2021 resulted in the first-ever FDA-approved computer-aided detection (CADe) tool. This narrative review begins by reviewing the evidence behind the use of CADe that led to FDA approval. Next, the review discusses the current evidence and technological approaches for computer-aided diagnosis for optical in situ histopathological differentiation of colorectal polyps, including narrow-band imaging, blue light imaging, and endocytoscopy. Studies are ongoing to develop systems to predict the depth of submucosal invasion and to assess endoscopic disease activity among patients with inflammatory bowel disease. The applications of AI/ML to quality improvement are explored, including real-time assessment of bowel preparation, detection of cecal intubation, and automated polyp reporting and surveillance recommendations using natural language processing. Despite initial cost concerns, models have suggested that CADe systems could result in long-term cost savings and are generally accepted by patients and gastroenterologists. There is some reservation in adopting computer-aided diagnosis systems among gastroenterologists due to medico-legal concerns. Future directions for AI/ML in colonoscopy include health system improvements, such as automating note writing, optimizing procedural scheduling, and predicting sedation needs.

Background and Aims

Pancreatic fluid collections (PFCs) are common complications of acute pancreatitis and can cause pain, difficulty feeding, and infection. Endoscopic ultrasound-guided drainage has become the standard of care, with lumen-apposing metal stents (LAMSs) replacing double pigtail plastic stents (DPPSs) as the preferred device. Coaxial placement of DPPSs through LAMSs is hypothesized to lower the risks of adverse events related to LAMSs. We conducted a retrospective study to evaluate the safety and efficacy of this strategy.

Methods

We conducted a retrospective study of consecutive patients with PFCs undergoing endoscopic cystgastrostomy with LAMSs and DPPSs or LAMSs alone at two U.S. academic tertiary care centers from January 2016 until November 2022. Propensity scoring and an adjusted logistic regression model were used for analysis.

Results

We included 68 patients with an average follow-up of 189 days. The most common etiology of pancreatitis was gallstones (35.3%), most PFCs were walled-off necrosis (61.8%), and the mean size was 14.7 cm (SD ± 5.9 cm). Overall clinical success was 88.2%, without significant differences between LAMSs and DPPSs vs LAMSs alone (95.7% vs 84.4%; P = 0.18; aOR = 4.6; 95% CI, 0.5-41.4). We found no statistically significant differences in rates of LAMS occlusion (aOR = 0.47; 95% CI, 0.09-2.5), infection (aOR = 1.03; 95% CI, 0.17-6.2), bleeding (aOR = 0.4; 95% CI, 0.03-5), or stent migration (aOR = 0.42; 95% CI, 0.04-4.1) between the 2 groups.

Conclusion

This retrospective cohort study found no statistically significant differences in the safety or efficacy of cystgastrostomy with LAMSs and DPPSs vs LAMSs alone. Larger, prospective trials comparing these strategies are needed.

Background and Aims

Gastrointestinal bleeding (GIB) is a common morbidity among continuous-flow left ventricular assist device (CF-LVAD) recipients. The study objectives were to investigate the incidence, recurrence, risk factors, and medical and endoscopic management of GIB among CF-LVAD patients.

Methods

This is a single-center retrospective study analyzing 563 CF-LVAD recipients implanted at a large tertiary referral center from 2007 to 2018 with 619 bleeding events characterized. Patients with device exchanges or history of right ventricular assist devices were excluded. Chart review data pertaining to demographics, LVAD characteristics, endoscopic reports, and medications were collected.

Results

The incidence of GIB was 44%, with a median time to first bleed of 133 days and a recurrence rate of 53%. Approximately 13% of patients had ≥5 GIB events. There is an association of reduced GIB among patients taking angiotensin-converting enzyme inhibitors (ACEis). There was no association with GIB and acid-reducing medications, antiplatelet medications, and history of GIB before device implantation. Arteriovenous malformations were the most common bleeding etiology. CF-LVAD patients that bled more frequently were older, had their CF-LVAD longer, and bled earlier after LVAD implantation.

Conclusion

This is the largest comprehensive single-center GIB study among CF-LVAD recipients. The incidence and recurrence of GIB among CF-LVAD recipients is underreported. Push enteroscopy improves diagnostic yield for upper GIB. Special attention should be given to CF-LVAD patients that bleed within the first 18 days of implantation. There is an associative, dose-dependent benefit of ACEi medications among CF-LVAD recipients. Further studies on this effect are warranted.

Previously, the gold standard surgery for early gastric cancer was gastrectomy. This practice has changed radically with the introduction of advanced endoscopic resection techniques, including endoscopic submucosal dissection (ESD). Compared with traditional gastrectomies, the minimally invasive nature of ESDs allows patients to recover faster and experience less postoperative pain. Nevertheless, procedurists performing gastric ESDs are hampered by the limitations of current endoscopic instruments. The lack of degrees of freedom of existing endoscopic instruments renders it challenging for procedurists to achieve triangulation of instruments, optimal retraction of tissues, and adequate exposure of the operating field. Robotic endoscopic platforms solve these challenges by providing robotic end-effectors with high degrees of freedom. This review analyzes the benefits of robotic endoscopic platforms in reducing the learning curve, procedure times, and complication rates in performing gastric ESD. The review also explores future development of robotic endoscopic platforms that can enhance the efficacy and efficiency of gastric ESD, such as the development of adjunct endoscopic instruments to create and close full-thickness incisions, incorporating haptic feedback into robotic endoscopic platforms, and establishing a structured training program for procedurists.

Background and Aims

Current guidelines recommend that the decision to pursue surveillance colonoscopy in older adults with polyps be individualized, yet how these recommendations are communicated has not been characterized. We aimed to evaluate the effectiveness of endoscopist recommendations after colonoscopy in communicating the need for future surveillance in older adults.

Methods

We performed a single-center, retrospective chart review of adults age ≥ 75 years who underwent colonoscopy for polyp surveillance or screening during which polyps were detected. We performed content analysis of the recommendations from both colonoscopy reports and post-pathology follow-up letters. Recommendations were classified as optimal or less optimal based on whether they were clear, contained a rationale, and maintained consistency between the report and follow-up letter.

Results

Between 2012 and 2019, there were 1428 colonoscopies performed by 17 endoscopists, of which 874 (61%) were optimal and 554 (39%) were less optimal. Among the less optimal recommendations, 76 (14%) lacked a recommendation, 233 (42%) were unclear, and 409 (74%) lacked a rationale. Among the 954 post-pathology follow-up letters, 80 (8%) were inconsistent with the original colonoscopy report recommendation, of which 30 (38%) resulted in a change in management. The frequency of less optimal recommendations ranged from 0% to 50% by endoscopist.

Conclusion

Following colonoscopy in older adults, we found that roughly one-third of the reports were less than optimal, and there was sizable variation in individual endoscopist performance. Discrepancies between colonoscopy reports and patient follow-up letters could be minimized by avoiding providing recommendations on future colonoscopy before pathologic interpretation.

Endoscopic submucosal dissection (ESD) is the current standard treatment for early intramucosal gastrointestinal neoplasia, as ESD achieved a higher rate of en bloc resection and lower rate of local recurrence, especially for neoplastic lesions greater than 20 mm, compared with endoscopic mucosal resection. However, colorectal ESD remains technically challenging, with a significant risk of complications, especially perforation during dissection. The higher risk of perforation is related to the intrinsic operational difficulties of using flexible endoscopy and dissection via the coaxial ESD device, as well as the unfavorable anatomical characteristics, including thin colonic wall and haustrations. The lack of bimanual tissue manipulation, mucosal traction, and poor visualization of dissection plane led to the challenges in colorectal ESD. The development of endoluminal flexible robotic systems aims to overcome these challenges with innovative flexible robotics to refine the endoscope and instrumentations, some of which have already been used in clinical trials. Preclinical studies demonstrated promising outcomes with using flexible robotics in improving the learning curve of colorectal ESD. This review will also explore future development of endoluminal flexible robotics and a training system, which are crucial to ensure an effective, efficient, and safe training program for novice endoscopists to perform robotic colorectal ESD.

Background and Aims

The adenoma detection rate (ADR), is a key quality metric for colonoscopy, but it might fail to measure differences in adenoma detection in different parts of the colon. Because many post-colonoscopy cancers arise in the proximal colon, differences in the ADR of proximal vs distal adenomas could be clinically important. Our aim was to determine the feasibility of measuring ADR for both the proximal and distal colon and to understand differences based on age, sex, and race.

Methods

We queried a large claims database (Health Care Service Corporation) to screen colonoscopies from 2016-2018 to determine the location and pathology of detected polyps. We measured the ADR in both the proximal (P-ADR) and distal (D-ADR) colon to determine differences in the ratio of P-ADR vs D-ADR by age, sex, and race.

Results

The cohort included 93,163 women and 89,132 men. A progressive increase in both P-ADR and D-ADR occurs with advancing age, with the P-ADR/D-ADR ratio rising from 1.2 in patients under 40 years of age to 2.65 in patients age 75 and older, in both men and women. This ratio is affected by the experience of the endoscopist. A higher volume of procedures is directly related to higher polyp detection, specifically the P-ADR.

Conclusion

As age advances, there is an increasing predominance of proximal adenomas relative to distal adenomas in both men and women, in all races. With appropriate coding, it is feasible to measure the ADR for the proximal and distal colon. Measurement of P-ADR could be an important new quality metric.