Journal of Pathology Clinical Research最新文献

The efficacy of pembrolizumab monotherapy versus chemotherapy increased with increasing programmed death ligand 1 (PD-L1) expression, as quantified by combined positive score (CPS; PD-L1 expression on both tumour cells and immune cells) in patients with previously treated metastatic triple-negative breast cancer (mTNBC) in the phase 3 KEYNOTE-119 study. This exploratory analysis was conducted to determine whether the expression of PD-L1 on tumour cells contributes to the predictive value of PD-L1 CPS in mTNBC. PD-L1 expression in tumour samples was assessed using PD-L1 IHC 22C3 pharmDx and quantified using both CPS and tumour proportion score (TPS; PD-L1 expression on tumour cells alone). Calculated immune cell density (CID) was defined as CPS minus TPS. The ability of each scoring method (CPS, TPS, and CID) to predict clinical outcomes with pembrolizumab was evaluated. With pembrolizumab, the area under the receiver operating characteristic curve was 0.69 (95% CI = 0.58–0.80) for CPS, 0.55 (95% CI = 0.46–0.64) for TPS, and 0.67 (95% CI = 0.56–0.77) for CID. After correction for cutoff prevalence, CPS performed as well as, if not better than, CID with respect to predicting objective response rate, progression-free survival, and overall survival. Data from this exploratory analysis suggest that, although PD-L1 expression on immune cells alone is predictive of response to programmed death 1 blockade in mTNBC, adding tumour PD-L1 expression assessment (i.e. CPS, which combines immune cell and tumour cell PD-L1 expression) may improve prediction. PD-L1 CPS thus remains an effective and broadly applicable uniform scoring system for enriching response to programmed death 1 blockade with pembrolizumab in mTNBC as well as other tumour types.

Diffuse large B-cell lymphoma (DLBCL) is a heterogeneous and prevalent subtype of aggressive non-Hodgkin lymphoma that poses diagnostic and prognostic challenges, particularly in predicting drug responsiveness. In this study, we used digital pathology and deep learning to predict responses to immunochemotherapy in patients with DLBCL. We retrospectively collected 251 slide images from 216 DLBCL patients treated with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP), with their immunochemotherapy response labels. The digital pathology images were processed using contrastive learning for feature extraction. A multi-modal prediction model was developed by integrating clinical data and pathology image features. Knowledge distillation was employed to mitigate overfitting on gigapixel histopathology images to create a model that predicts responses based solely on pathology images. Based on the importance derived from the attention mechanism of the model, we extracted histological features that were considered key textures associated with drug responsiveness. The multi-modal prediction model achieved an impressive area under the ROC curve of 0.856, demonstrating significant associations with clinical variables such as Ann Arbor stage, International Prognostic Index, and bulky disease. Survival analyses indicated their effectiveness in predicting relapse-free survival. External validation using TCGA datasets supported the model's ability to predict survival differences. Additionally, pathology-based predictions show promise as independent prognostic indicators. Histopathological analysis identified centroblastic and immunoblastic features to be associated with treatment response, aligning with previous morphological classifications and highlighting the objectivity and reproducibility of artificial intelligence-based diagnosis. This study introduces a novel approach that combines digital pathology and clinical data to predict the response to immunochemotherapy in patients with DLBCL. This model shows great promise as a diagnostic and prognostic tool for clinical management of DLBCL. Further research and genomic data integration hold the potential to enhance its impact on clinical practice, ultimately improving patient outcomes.

Diagnosing extrapulmonary tuberculosis (EPTB) is challenging. Immunohistochemistry or immunocytochemistry has been used to diagnose tuberculosis (TB) by detection of MPT64 antigen from various extrapulmonary specimens and has shown good diagnostic performance in our previous studies. The test can distinguish between disease caused by Mycobacterium tuberculosis (Mtb) complex and nontuberculous mycobacteria and can be applied on formalin-fixed paraffin-embedded tissue. As the antibodies previously used were in limited supply, a new batch of polyclonal antibodies was developed for scale-up and evaluated for the first time in this study. Our aim was to assess the diagnostic accuracy of the MPT64 test with reproduced antibodies in the high burden settings of Pakistan and India. Patients were enrolled prospectively. Samples from suspected sites of infection were collected and subjected to histopathologic and/or cytologic evaluation, routine TB diagnostics, GeneXpert MTB/RIF (Xpert), and the MPT64 antigen detection test. Patients were followed until the end of treatment. Based on a composite reference standard (CRS), 556 patients were categorized as TB cases and 175 as non-TB cases. The MPT64 test performed well on biopsies with a sensitivity and specificity of 94% and 75%, respectively, against a CRS. For cytology samples, the sensitivity was low (36%), whereas the specificity was 81%. Overall, the MPT64 test showed higher sensitivity (73%) than Xpert (38%) and Mtb culture (33%). The test performed equally well in adults and children. We found an additive diagnostic value of the MPT64 test in conjunction with histology and molecular tests, increasing the yield for EPTB. In conclusion, immunochemical staining with MPT64 antibodies improves the diagnosis of EPTB in high burden settings and could be a valuable addition to routine diagnostics.

Everolimus is widely used in patients with advanced ER-positive, HER2-negative breast cancer. We looked at alterations in the PIK3CA/AKT/mTOR pathway in a multicenter cohort as potential biomarkers of efficacy. Patients with advanced ER-positive, HER2-negative breast cancer treated with everolimus and endocrine therapy between 2012 and 2014 in two cancer centers were included. Targeted sequencing examined mutations in PIK3CA, ESR1, and AKT1 genes. An immunochemical analysis was conducted to evaluate expression of PTEN, INPP4B, STK11, p4EBP1, and pS6. We analyzed 71 patients (44 primary tumors; 27 metastatic tissues). Median age was 63 years [58–69]. All patients had heavily pretreated advanced disease. A mutation in the PIK3CA pathway was observed in 32 samples (PIK3CA exons 10 and 21 and AKT1 exon 4 in 15.5%, 24.0%, and 5.6% of samples), and in ESR1 in 5 samples (7.0%), respectively. Most samples showed cytoplasmic expression of the PIK3CA pathway proteins. Progression-free survival was longer in patients with a pS6 or p4EBP1 histoscore ≥ median value (6.6 versus 3.7 months, p = 0.037), and in patients with a PTEN histoscore ≤ median value (7.1 versus 5.3 months, p = 0.02). Overall survival was longer in patients with pS6 ≥ 3rd quartile (27.6 versus 19.3 months, p = 0.038) and in patients with any mutation in the PIK3CA/AKT/mTOR pathway (27.6 versus 19.3 months, p = 0.011). The prognosis of patients treated with everolimus for advanced ER-positive, HER2-negative breast cancer appears primarily driven by molecular features associated with the activation of the PIK3CA/AKT/mTOR pathway.

Upper tract urothelial carcinoma (UTUC) is a rare and aggressive, yet understudied, urothelial carcinoma (UC). The more frequent UC of the bladder comprises several molecular subtypes, associated with different targeted therapies and overlapping with protein-based subtypes. However, if and how these findings extend to UTUC remains unclear. Artificial intelligence-based approaches could help elucidate UTUC's biology and extend access to targeted treatments to a wider patient audience. Here, UTUC protein-based subtypes were identified, and a deep-learning (DL) workflow was developed to predict them directly from routine histopathological H&E slides. Protein-based subtypes in a retrospective cohort of 163 invasive tumors were assigned by hierarchical clustering of the immunohistochemical expression of three luminal (FOXA1, GATA3, and CK20) and three basal (CD44, CK5, and CK14) markers. Cluster analysis identified distinctive luminal (N = 80) and basal (N = 42) subtypes. The luminal subtype mostly included pushing, papillary tumors, whereas the basal subtype diffusely infiltrating, non-papillary tumors. DL model building relied on a transfer-learning approach by fine-tuning a pre-trained ResNet50. Classification performance was measured via three-fold repeated cross-validation. A mean area under the receiver operating characteristic curve of 0.83 (95% CI: 0.67–0.99), 0.8 (95% CI: 0.62–0.99), and 0.81 (95% CI: 0.65–0.96) was reached in the three repetitions. High-confidence DL-based predicted subtypes showed significant associations (p < 0.001) with morphological features, i.e. tumor type, histological subtypes, and infiltration type. Furthermore, a significant association was found with programmed cell death ligand 1 (PD-L1) combined positive score (p < 0.001) and FGFR3 mutational status (p = 0.002), with high-confidence basal predictions containing a higher proportion of PD-L1 positive samples and high-confidence luminal predictions a higher proportion of FGFR3-mutated samples. Testing of the DL model on an independent cohort highlighted the importance to accommodate histological subtypes. Taken together, our DL workflow can predict protein-based UTUC subtypes, associated with the presence of targetable alterations, directly from H&E slides.

Breast cancers involving mutations in homologous recombination (HR) genes, most commonly BRCA1 and BRCA2 (BRCA1/2), respond well to PARP inhibitors and platinum-based chemotherapy. However, except for these specific HR genes, it is not clear which other mutations contribute to homologous recombination defects (HRD). Here, we performed next-generation sequencing of tumor tissues and matched blood samples from 119 breast cancer patients using the OncoScreen Plus panel. Genomic mutation characteristics and HRD scores were analyzed. In the HR genes, we found that BRCA1/2 and PLAB2 mutations were related to HRD. HRD was also detected in a subset of patients without germline or somatic mutations in BRCA1/2, PLAB2, or other HR-related genes. Notably, LRP1B, NOTCH3, GATA2, and CARD11 (abbreviated as LNGC) mutations were associated with high HRD scores in breast cancer patients. Furthermore, functional experiments demonstrated that silencing CARD11 and GATA2 impairs HR repair efficiency and enhances the sensitivity of tumor cells to olaparib treatment. In summary, in the absence of mutations in the HR genes, the sensitivity of tumor cells to PARP inhibitors and platinum-based chemotherapy may be enhanced in a subset of breast cancer patients with LNGC somatic mutations.

An increasing number of manuscripts related to digital and computational pathology are being submitted to The Journal of Pathology: Clinical Research as part of the continuous evolution from digital imaging and algorithm-based digital pathology to computational pathology and artificial intelligence. However, despite these technological advances, tissue analysis still relies heavily on pathologists' annotations. There are three crucial elements to the pathologist's role during annotation tasks: granularity, time constraints, and responsibility for the interpretation of computational results. Granularity involves detailed annotations, including case level, regional, and cellular features; and integration of attributions from different sources. Time constraints due to pathologist shortages have led to the development of techniques to expedite annotation tasks from cell-level attributions up to so-called unsupervised learning. The impact of pathologists may seem diminished, but their role is crucial in providing ground truth and connecting pathological knowledge generation with computational advancements. Measures to display results back to pathologists and reflections about correctly applied diagnostic criteria are mandatory to maintain fidelity during human–machine interactions. Collaboration and iterative processes, such as human-in-the-loop machine learning are key for continuous improvement, ensuring the pathologist's involvement in evaluating computational results and closing the loop for clinical applicability. The journal is interested particularly in the clinical diagnostic application of computational pathology and invites submissions that address the issues raised in this editorial.

We performed comprehensive analyses of somatic copy number alterations (SCNAs) and gene expression profiles of gastric intramucosal neoplasia (IMN) using array-based methods in 97 intestinal-type IMNs, including 39 low-grade dysplasias (LGDs), 37 high-grade dysplasias (HGDs), and 26 intramucosal carcinomas (IMCs) with stromal invasion of the lamina propria to identify the molecular mechanism of IMN. In addition, we examined gene mutations using gene panel analyses. We used cluster analyses for exclusion of arbitrariness to identify SCNA patterns and expression profiles. IMNs were classified into two distinct subgroups (subgroups 1 and 2) based on SCNA patterns. Subgroup 1 showed a genomic stable pattern due to the low frequency of SCNAs, whereas subgroup 2 exhibited a chromosomal instability pattern due to the high frequencies of SCNAs and TP53 mutations. Interestingly, although the frequencies of LGD and HGD were significantly higher in subgroup 1 than in subgroup 2, IMC was commonly found in both types. Although the expression profiles of specific mRNAs could be used to categorise subgroups 1 and 2, no clinicopathological findings correlated with either subgroup. We examined signalling pathways specific to subgroups 1 and 2 to identify the association of each subgroup with signalling pathways based on gene ontology tree visualisation: subgroups 1 and 2 were associated with haem metabolism and chromosomal instability, respectively. These findings reveal a comprehensive genomic landscape that highlights the molecular complexity of IMNs and provide a road map to facilitate our understanding of gastric IMNs.

Clear cell renal cell carcinoma (ccRCC) is one of the most common subtypes of renal cancer, with 30% of patients presenting with systemic disease at diagnosis. This aggressiveness is a consequence of the activation of epithelial–mesenchymal transition (EMT) caused by many different inducers or regulators, signaling cascades, epigenetic regulation, and the tumor environment. Alterations in EMT-related genes and transcription factors are associated with poor prognosis in ccRCC. EMT-related factors suppress E-cadherin expression and are associated with tumor progression, local invasion, and metastasis. The aim of this study was to investigate the expression levels and prognostic significance of macrophage migration inhibitory factor (MIF), β-catenin, and E-cadherin in ccRCC patients. We examined these proteins immunohistochemically in tumor areas and adjacent normal tissues resected from patients with ccRCC. Analysis of the cancer genome atlas (TCGA) cohort was performed to verify our results. Kaplan–Meier analysis showed that median overall survival (OS) was significantly shorter in patients with tumors exhibiting high MIFⁿ and MIF^m-c levels compared to those with low MIFⁿ and MIF^m-c levels (p = 0.03 and p = 0.007, respectively). In the TCGA cohort, there was a significant correlation between MIF expression and OS (p < 0.0001). In conclusion, this study provides further evidence for the biological and prognostic value of MIF in the context of EMT as a potential early prognostic marker for advanced-stage ccRCC.

Up to 30% of colorectal cancers (CRCs) develop from sessile serrated lesions (SSLs). Within the serrated neoplasia pathway, at least two principally distinct oncogenetic routes exist generating microsatellite-stable and microsatellite-instable CRCs, respectively. Aberrant DNA methylation (DNAm) is found early in the serrated pathway and might play a role in both oncogenetic routes. We studied a cohort of 23 SSLs with a small focus (<10 mm) of dysplasia or cancer, 10 of which were MLH1 deficient and 13 MLH1 proficient. By comparing, for each SSL, the methylation status of (1) the region of dysplasia or cancer (SSL-D), (2) the nondysplastic SSL (SSL), and (3) adjacent normal mucosa, differentially methylated probes (DMPs) and regions (DMRs) were assessed both genome-wide as well as in a tumor-suppressor gene-focused approach. By comparing DNAm of MLH1-deficient SSL-Ds with their corresponding SSLs, we identified five DMRs, including those annotating for PRDM2 and, not unexpectedly, MLH1. PRDM2 gene promotor methylation was associated with MLH1 expression status, as it was largely hypermethylated in MLH1-deficient SSL-Ds and hypomethylated in MLH1-proficient SSL-Ds. Significantly increased DNAm levels of PRDM2 and MLH1, in particular at ‘critical’ MLH1 probe sites, were to some extent already visible in SSLs as compared to normal mucosa (p = 0.02, p = 0.01, p < 0.0001, respectively). No DMRs, nor DMPs, were identified for SSLs destined to evolve into MLH1-proficient SSL-Ds. Our data indicate that, within both arms of the serrated CRC pathway, the majority of the epigenetic alterations are introduced early during SSL formation. Promoter hypermethylation of PRDM2 and MLH1 on the other hand specifically initiates in SSLs destined to transform into MLH1-deficient CRCs suggesting that the fate of SSLs may not necessarily result from a stochastic process but possibly is already imprinted and predisposed.