Citations

Abstract Background The loss-of-function (LOF) classification of most missense variants in tumor suppressor breast cancer genes BRCA1, BRCA2, PALB2 , and RAD51C remains unclassified and confounds clinical actionability. Classifying these variants is challenging due to their rarity, leading clinicians to rely on in silico predictive methods. Protein stability changes are associated with function, making stability predictors valuable. Stability predictions upon missense variant perturbations require high-resolution protein structures. However, the availability of these high-resolution structures is lacking. This study explores using generative AI to predict high-resolution protein structures, which can then be analyzed with in silico protein stability prediction methods to assess LOF activity in ordered regions of the protein. This study also determines the appropriate in silico protein stability and dedicated in silico missense prediction methods in dbNSFP v4.7 database to predict LOF activity in ordered regions of these four genes. Functional classifications from homology recombination DNA repair (HDR) assays and variant classifications from the ClinVar database provide a reliable dataset for evaluating the performance of these in silico prediction methods. Results Complex AlphaFold2 structures of the BRCA1-C terminal (BRCT) domain and the DNA-binding (DB) domain of BRCA2, analyzed using protein stability tool FoldX predicts LOF activity from missense variants significantly better than experimentally-derived structures in ordered regions. The BRCT domain achieved an Area Under the Curve (AUC)= 0.861 (95 % CI:0.858–0.863) and AUC= 0.842 (95 % CI:0.840–0.845), while the DB domain achieved an AUC= 0.836 (95 % CI:0.8322–0.841), compared to AUC= 0.847 (95 % CI:0.844–0.850) and AUC= 0.835 (95 % CI:0.832–0.837) from the BRCT domain, and AUC= 0.830 (95 % CI:0.821–0.8320) from the DB domain from experimentally-derived structures. Protein stability does not predict LOF activity from missense variants better than dedicated in silico missense predictors. Overall, we find that AlphaMissense ranks highly, with an average AUC= 0.890 (95 % CI 0.886–0.895) from ordered regions across these four cancer genes, compared to all other in silico missense predictors present in the dbNSFP database. Conclusions The study reveals that generative AI protein predicted structures can outperform experimentally-derived structures in evaluating LOF activity from predicted protein stability in ordered regions of genes BRCA1, BRCA2, PALB2 and RAD51C. The study also highlights the predictive performance of AlphaMissense as the premier in silico missense prediction method to predict LOF activity from missense variants in these four tumor suppressor breast cancer genes. The code for this study can be downloaded for free on GitHub ( https://github.com/rohandavidg/CarePred )

Context.— Generative artificial intelligence (AI) technologies are rapidly transforming numerous fields, including pathology, and hold significant potential to revolutionize educational approaches. Objective.— To explore the application of generative AI, particularly large language models and multimodal tools, for enhancing pathology education. We describe their potential to create personalized learning experiences, streamline content development, expand access to educational resources, and support both learners and educators throughout the training and practice continuum. Data Sources.— We draw on insights from existing literature on AI in education and the collective expertise of the coauthors within this rapidly evolving field. Case studies highlight practical applications of large language models, demonstrating both the potential benefits and unique challenges associated with implementing these technologies in pathology education. Conclusions.— Generative AI presents a powerful tool kit for enriching pathology education, offering opportunities for greater engagement, accessibility, and personalization. Careful consideration of ethical implications, potential risks, and appropriate mitigation strategies is essential for the responsible and effective integration of these technologies. Future success lies in fostering collaborative development between AI experts and medical educators, prioritizing ongoing human oversight and transparency to ensure that generative AI augments, rather than supplants, the vital role of educators in pathology training and practice.

Background and Aims: This investigation examined the association of pancreatitis and pancreatitis-related pain with serum levels of two endocannabinoid molecules such as anandamide (AEA) and 2-arachidonoylglycerol (2-AG) and two paracannabinoid molecules such as oleoylethanolamide (OEA) and palmitoylethanolamide (PEA). Methods: A case–control study was conducted within the Prospective Evaluation of Chronic Pancreatitis for Epidemiological and Translational Studies, including participants with no pancreas disease ( N = 56), chronic abdominal pain of suspected pancreatic origin or indeterminate chronic pancreatitis (CP) ( N = 22), acute pancreatitis ( N = 33), recurrent acute pancreatitis ( N = 57), and definite CP ( N = 63). Results: Circulating AEA concentrations were higher in women than in men ( p = 0.0499), and PEA concentrations were higher in obese participants than those who were underweight/normal or overweight ( p = 0.003). Asymptomatic controls with no pancreatic disease had significantly ( p = 0.03) lower concentrations of AEA compared with all disease groups combined. The highest concentrations of AEA were observed in participants with acute pancreatitis, followed by those with recurrent acute pancreatitis, chronic abdominal pain/indeterminant CP, and definite CP. Participants with pancreatitis reporting abdominal pain in the past year had significantly ( p = 0.04) higher concentrations of AEA compared with asymptomatic controls. Levels of 2-AG were significantly lower ( p = 0.02) among participants reporting abdominal pain in the past week, and pain intensity was inversely associated with concentrations of 2-AG and OEA. Conclusions: Endocannabinoid levels may be associated with stage of pancreatitis, perhaps through activation of the CB1 receptor. Validation of our findings would support the investigation of novel therapeutics, including cannabinoid receptor-1 antagonists, in this patient population.

PURPOSE Cancer of unknown primary (CUP) is a syndrome comprising metastatic cancers without a clinically identified primary site. Although patients with CUP have an unfavorable prognosis, treatment with site-specific therapies guided by clinical features, standard pathology, and molecular assays can improve overall survival. The 92-gene assay (CancerTYPE ID) is a gene expression–based classifier that helps identify the tissue of origin for metastatic cancers with unknown or uncertain diagnoses. This study reports the frequency of selected molecular aberrations of oncogenes, including KRAS, IDH1/2, BRCA1/2, and BRAF, in patients with CUP in the MOSAIC database to highlight potential treatment options. METHODS MOSAIC is a database of patients with CUP submitted for CancerTYPE ID testing and NeoTYPE biomarker testing. Tumor biopsy samples were analyzed by CancerTYPE ID for tumor type identification and further tested for molecular aberrations of oncogenes, including KRAS, IDH1/2, BRCA1/2, and BRAF. RESULTS CancerTYPE ID identified a specific tumor type in 92.5% (2,929 of 3,168) of CUP cases in the MOSAIC database. The most commonly identified histological type was adenocarcinoma (75.4%), with pancreaticobiliary being the most common molecularly diagnosed cancer (24.9%). Aberrations in KRAS, IDH1/2, BRCA, and BRAF genes were identified in 18.8% (n = 597) of biopsies. A cancer-specific US Food and Drug Administration (FDA)–approved or investigational targeted therapy was potentially available for 24.6% (n = 147) of these patients. CONCLUSION This retrospective analysis supports incorporating CancerTYPE ID into the evaluation for patients with CUP to help determine the tissue of origin and identify actionable genetic alterations. This approach may allow more patients with CUP to benefit from site-specific FDA-approved targeted therapies or enrollment into clinical trials.

BackgroundDuctal features alone may not offer high diagnostic sensitivity or most accurate disease severity of chronic pancreatitis (CP).PurposeDiagnose CP based on multiparametric MRI and MRCP features.Study TypeProspective.PopulationBetween February 2019 and May 2021, 46 control (23 males, 49.3 ± 14.1 years), 45 suspected (20 males, 48.7 ± 12.5 years), and 46 definite (20 males, 53.7 ± 14.6 years) CP patients were enrolled at seven hospitals enrolled in the MINIMAP study. CP classification was based on imaging findings and clinical presentation.Field Strength and Sequences1.5 T. T1‐weighted (T1W) spoiled gradient echo, T1 map with variable flip angle, dual‐echo Dixon, secretin‐enhanced MRCP before and after secretin infusion.AssessmentDual‐echo fat fraction (FF), T1 relaxation time, extracellular volume (ECV), T1 signal intensity ratio of the pancreas to the spleen (T1 score), arterial‐to‐venous enhancement ratio (AVR), pancreatic tail diameter (PTD), pancreas volume, late gadolinium enhancement, pancreatic ductal elasticity (PDE), and duodenal filling grade of secretin‐enhanced MRCP were measured.Statistical TestsLogistic regression analysis generated CP‐MRI and secretin‐enhanced CP‐SMRI scores. Receiver operating characteristics analysis was used to differentiate definite CP from control. Interobserver agreement was assessed using Lin's concordance correlation coefficient.ResultsCompared to control, definite CP cohort showed significantly higher dual‐echo FF (7% vs. 11%), lower AVR (1.35 vs. 0.85), smaller PTD (2.5 cm vs. 1.95 cm), higher ECV (28% vs. 38%), and higher incidence of PDE loss (6.5% vs. 50%). With the cut‐off of >2.5 CP‐MRI score (dual‐echo FF, AVR, and PTD) and CP‐SMRI score (dual‐echo FF, AVR, PTD, and PDE) had cross‐validated area under the curves of 0.84 (sensitivity 87%, specificity 68%) and 0.86 (sensitivity 89%, specificity 67%), respectively. Interobserver agreement for both CP‐MRI and CP‐SMRI scores was 0.74.ConclusionThe CP‐MRI and CP‐SMRI scores yielded acceptable performance and interobserver agreement for the diagnosis of CP.Evidence Level1Technical EfficacyStage 2

Abstract Breast cancer is one of the most common cancers in women. With early diagnosis, some breast cancers are highly curable. Classifying normal versus tumor breast tissues from microscopy images is an ideal case to use for deep learning and could help diagnose breast cancer with higher reproducibility. Since data preprocessing and hyperparameter tuning have impacts on breast cancer classification accuracies, training a classifier with appropriate data preprocessing and optimized hyperparameters could improve breast cancer classification accuracy. Using 12 combinations of model architectures, data preprocessing, and hyperparameter configurations, the validation accuracy was calculated using the BreAst Cancer Histology (BACH) dataset. The DenseNet201, a non-specialized model architecture, with transfer learning approach achieved 98.61% validation accuracy compared to only 64.00% for the digital pathology (DP)-specialized model architecture. The combination of image data preprocessing and hyperparameters have a profound impact on the performance of deep neural networks for image classification. To identify a well-performing neural network to classify tumor versus normal breast histology, researchers should not only focus on developing new models specifically for DP, since hyperparameter tuning for existing deep neural networks in the computer vision field could also achieve a better prediction accuracy.

INTRODUCTION: Chronic Pancreatitis Prognosis Score (COPPS) was developed to discriminate disease severity and predict risk for future hospitalizations. In this cohort study, we evaluated if COPPS predicts the likelihood of hospitalization(s) in an American cohort. METHODS: The Chronic Pancreatitis, Diabetes, and Pancreatic Cancer consortium provided data and serum from subjects with chronic pancreatitis (N = 279). COPPS was calculated with baseline data and stratified by severity (low, moderate, and high). Primary endpoints included number and duration of hospitalizations during 12-month follow-up. RESULTS: The mean ± SD COPPS was 8.4 ± 1.6. COPPS correlated with all primary outcomes: hospitalizations for any reason (number: r = 0.15, P = 0.01; duration: r = 0.16, P = 0.01) and pancreas-related hospitalizations (number: r = 0.15, P = 0.02; duration: r = 0.13, P = 0.04). The severity distribution was 13.3% low, 66.0% moderate, and 20.8% high. 37.6% of subjects had ≥1 hospitalization(s) for any reason; 32.2% had ≥1 pancreas-related hospitalizations. All primary outcomes were significantly different between severity groups: hospitalizations for any reason (number, P = 0.004; duration, P = 0.007) and pancreas-related hospitalizations (number, P = 0.02; duration, P = 0.04). The prevalence of continued drinking at follow-up ( P = 0.04) was higher in the low and moderate groups. The prevalence of anxiety at enrollment ( P = 0.02) and follow-up ( P < 0.05) was higher in the moderate and high groups. DISCUSSION: Statistically, COPPS significantly correlated with hospitalization outcomes, but the correlations were weaker than in previous studies, which may be related to the outpatient nature of the PROspective Evaluation of Chronic Pancreatitis for EpidEmiologic and Translational StuDies cohort and lower prevalence of high severity disease. Studies in other prospective cohorts are needed to understand the full utility of COPPS as a potential tool for clinical risk assessment and intervention.

PURPOSE Long-term outcomes of patients with stage I human epidermal growth factor receptor 2 (HER2)–positive breast cancer receiving adjuvant trastuzumab emtansine (T-DM1) remain undefined, and prognostic predictors represent an unmet need. METHODS In the ATEMPT phase II trial, patients with stage I centrally confirmed HER2-positive breast cancer were randomly assigned 3:1 to adjuvant T-DM1 for 1 year or paclitaxel plus trastuzumab (TH). Coprimary objectives were to compare the incidence of clinically relevant toxicities between arms and to evaluate invasive disease-free survival (iDFS) with T-DM1. Correlative analyses included the HER2DX genomic tool, multiomic evaluations of HER2 heterogeneity, and predictors of thrombocytopenia. RESULTS After a median follow-up of 5.8 years, 11 iDFS events were observed in the T-DM1 arm, consistent with a 5-year iDFS of 97.0% (95% CI, 95.2 to 98.7). At 5 years, the recurrence-free interval (RFI) was 98.3% (95% CI, 97.0 to 99.7), the overall survival was 97.8% (95% CI, 96.3 to 99.3), and the breast cancer-specific survival was 99.4% (95% CI, 98.6 to 100). Comparable iDFS was observed with T-DM1 irrespective of tumor size, hormone receptor status, centrally determined HER2 immunohistochemical score, and receipt of T-DM1 for more or less than 6 months. Although ATEMPT was not powered for this end point, the 5-year iDFS in the TH arm was 91.1%. Among patients with sufficient tissue for HER2DX testing (n = 187), 5-year outcomes significantly differed according to HER2DX risk score, with better RFI (98.1% v 81.8%, hazard ratio [HR], 0.10, P = .01) and iDFS (96.3% v 81.8%, HR, 0.20, P = .047) among patients with HER2DX low-risk versus high-risk tumors, respectively. CONCLUSION Adjuvant T-DM1 for 1 year leads to outstanding long-term outcomes for patients with stage I HER2-positive breast cancer. A high HER2DX risk score predicted a higher risk of recurrence in ATEMPT.

Context.— Computational pathology combines clinical pathology with computational analysis, aiming to enhance diagnostic capabilities and improve clinical productivity. However, communication barriers between pathologists and developers often hinder the full realization of this potential. Objective.— To propose a standardized framework that improves mutual understanding of clinical objectives and computational methodologies. The goal is to enhance the development and application of computer-aided diagnostic (CAD) tools. Design.— This article suggests pivotal roles for pathologists and computer scientists in the CAD development process. It calls for increased understanding of computational terminologies, processes, and limitations among pathologists. Similarly, it argues that computer scientists should better comprehend the true use cases of the developed algorithms to avoid clinically meaningless metrics. Results.— CAD tools improve pathology practice significantly. Some tools have even received US Food and Drug Administration approval. However, improved understanding of machine learning models among pathologists is essential to prevent misuse and misinterpretation. There is also a need for a more accurate representation of the algorithms’ performance compared to that of pathologists. Conclusions.— A comprehensive understanding of computational and clinical paradigms is crucial for overcoming the translational gap in computational pathology. This mutual comprehension will improve patient care through more accurate and efficient disease diagnosis.