Extracting structured information from unstructured histopathology reports using generative pre-trained transformer 4 (GPT-4).

Deep learning applied to whole-slide histopathology images (WSIs) has the potential to enhance precision oncology and alleviate the workload of experts. However, developing these models necessitates large amounts of data with ground truth labels, which can be both time-consuming and expensive to obtain. Pathology reports are typically unstructured or poorly structured texts, and efforts to implement structured reporting templates have been unsuccessful, as these efforts lead to perceived extra workload. In this study, we hypothesised that large language models (LLMs), such as the generative pre-trained transformer 4 (GPT-4), can extract structured data from unstructured plain language reports using a zero-shot approach without requiring any re-training. We tested this hypothesis by utilising GPT-4 to extract information from histopathological reports, focusing on two extensive sets of pathology reports for colorectal cancer and glioblastoma. We found a high concordance between LLM-generated structured data and human-generated structured data. Consequently, LLMs could potentially be employed routinely to extract ground truth data for machine learning from unstructured pathology reports in the future. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

PITX2 as a Sensitive and Specific Marker of Midgut Neuroendocrine Tumors: Results from a Cohort of 1157 Primary Neuroendocrine Neoplasms.

As neuroendocrine tumors (NETs) often present as metastatic lesions, immunohistochemical assignment to a site of origin is one of the most important tasks in their pathologic assessment. Because a fraction of NETs eludes the typical expression profiles of their primary localization, additional sensitive and specific markers are required to improve diagnostic certainty. We investigated the expression of the transcription factor Pituitary Homeobox 2 (PITX2) in a large-scale cohort of 909 NET and 248 neuroendocrine carcinomas (NEC) according to the immunoreactive score (IRS) and correlated PITX2 expression groups with general tumor groups and primary localization. PITX2 expression (all expression groups) was highly sensitive (98.1%) for midgut-derived NET, but not perfectly specific, as non-midgut NET (especially pulmonary/duodenal) were quite frequently weak or moderately positive. The specificity rose to 99.5% for a midgut origin of NET if only a strong PITX2 expression was considered, which was found in only 0.5% (one pancreatic/one pulmonary) of non-midgut NET. In metastases of midgut-derived NET, PITX2 was expressed in all cases (87.5% strong, 12.5% moderate), whereas CDX2 was negative or only weakly expressed in 31.3% of the metastases. In NEC, a fraction of cases (14%) showed a weak or moderate PITX2 expression, which was not associated with a specific tumor localization. Our study independently validates PITX2 as a very sensitive and specific immunohistochemical marker of midgut-derived NET in a very large collective of neuroendocrine neoplasms. Therefore, our data argue toward implementation into diagnostic panels applied for NET as a firstline midgut marker.

Automatic structuring of radiology reports with on-premise open-source large language models.

OBJECTIVES: Structured reporting enhances comparability, readability, and content detail. Large language models (LLMs) could convert free text into structured data without disrupting radiologists' reporting workflow. This study evaluated an on-premise, privacy-preserving LLM for automatically structuring free-text radiology reports. MATERIALS AND METHODS: We developed an approach to controlling the LLM output, ensuring the validity and completeness of structured reports produced by a locally hosted Llama-2-70B-chat model. A dataset with de-identified narrative chest radiograph (CXR) reports was compiled retrospectively. It included 202 English reports from a publicly available MIMIC-CXR dataset and 197 German reports from our university hospital. Senior radiologist prepared a detailed, fully structured reporting template with 48 question-answer pairs. All reports were independently structured by the LLM and two human readers. Bayesian inference (Markov chain Monte Carlo sampling) was used to estimate the distributions of Matthews correlation coefficient (MCC), with [-0.05, 0.05] as the region of practical equivalence (ROPE). RESULTS: The LLM generated valid structured reports in all cases, achieving an average MCC of 0.75 (94% HDI: 0.70-0.80) and F1 score of 0.70 (0.70-0.80) for English, and 0.66 (0.62-0.70) and 0.68 (0.64-0.72) for German reports, respectively. The MCC differences between LLM and humans were within ROPE for both languages: 0.01 (-0.05 to 0.07), 0.01 (-0.05 to 0.07) for English, and -0.01 (-0.07 to 0.05), 0.00 (-0.06 to 0.06) for German, indicating approximately comparable performance. CONCLUSION: Locally hosted, open-source LLMs can automatically structure free-text radiology reports with approximately human accuracy. However, the understanding of semantics varied across languages and imaging findings. KEY POINTS: Question Why has structured reporting not been widely adopted in radiology despite clear benefits and how can we improve this? Findings A locally hosted large language model successfully structured narrative reports, showing variation between languages and findings. Critical relevance Structured reporting provides many benefits, but its integration into the clinical routine is limited. Automating the extraction of structured information from radiology reports enables the capture of structured data while allowing the radiologist to maintain their reporting workflow.

CD15 is a risk predictor and a novel target in clear cell renal cell carcinoma.

INTRODUCTION: Tumor cells use adhesion molecules like CD15 or sialylCD15 (sCD15) for metastatic spreading. We analyzed the expression of CD15 and sCD15 in clear cell renal cell carcinoma (ccRCC) regarding prognosis. METHODS: A tissue microarray containing tissue specimens of 763 patients with ccRCC was immunohistochemically stained for CD15 and sCD15, their expression quantified using digital image analysis and the impact on patients' survival analyzed. The cell lines 769p and 786o were stimulated with CD15 or control antibody in vitro and the effects on pathways activating AP-1 and tumor cell migration examined. RESULTS: ccRCC showed a broad range of CD15 and sCD15 expression. A high CD15 expression was significantly associated with favorable outcome (p<0.01) and low-grade tumor differentiation (p<0.001), whereas sCD15 had no significant prognostic value. Tumors with synchronous distant metastasis had a significantly lower CD15 expression compared to tumors without any (p<0.001) or with metachronous metastasis (p<0.01). Tumor cell migration was significantly reduced after CD15 stimulation in vitro, but there were no major effects on activating pathways of AP-1. CONCLUSION: CD15, but not sCD15, qualifies as a biomarker for risk stratification and as an interesting novel target in ccRCC. Moreover, the data indicates a contribution of CD15 to metachronous metastasis. Further research is warranted to decipher the intracellular pathways of CD15 signaling in ccRCC in order to characterize the CD15 effects on ccRCC more precisely.

Direct prediction of genetic aberrations from pathology images in gastric cancer with swarm learning.

BACKGROUND: Computational pathology uses deep learning (DL) to extract biomarkers from routine pathology slides. Large multicentric datasets improve performance, but such datasets are scarce for gastric cancer. This limitation could be overcome by Swarm Learning (SL). METHODS: Here, we report the results of a multicentric retrospective study of SL for prediction of molecular biomarkers in gastric cancer. We collected tissue samples with known microsatellite instability (MSI) and Epstein-Barr Virus (EBV) status from four patient cohorts from Switzerland, Germany, the UK and the USA, storing each dataset on a physically separate computer. RESULTS: On an external validation cohort, the SL-based classifier reached an area under the receiver operating curve (AUROC) of 0.8092 (± 0.0132) for MSI prediction and 0.8372 (± 0.0179) for EBV prediction. The centralized model, which was trained on all datasets on a single computer, reached a similar performance. CONCLUSIONS: Our findings demonstrate the feasibility of SL-based molecular biomarkers in gastric cancer. In the future, SL could be used for collaborative training and, thus, improve the performance of these biomarkers. This may ultimately result in clinical-grade performance and generalizability.

Artificial intelligence to identify genetic alterations in conventional histopathology.

Precision oncology relies on the identification of targetable molecular alterations in tumor tissues. In many tumor types, a limited set of molecular tests is currently part of standard diagnostic workflows. However, universal testing for all targetable alterations, especially rare ones, is limited by the cost and availability of molecular assays. From 2017 to 2021, multiple studies have shown that artificial intelligence (AI) methods can predict the probability of specific genetic alterations directly from conventional hematoxylin and eosin (H&E) tissue slides. Although these methods are currently less accurate than gold standard testing (e.g. immunohistochemistry, polymerase chain reaction or next-generation sequencing), they could be used as pre-screening tools to reduce the workload of genetic analyses. In this systematic literature review, we summarize the state of the art in predicting molecular alterations from H&E using AI. We found that AI methods perform reasonably well across multiple tumor types, although few algorithms have been broadly validated. In addition, we found that genetic alterations in FGFR, IDH, PIK3CA, BRAF, TP53, and DNA repair pathways are predictable from H&E in multiple tumor types, while many other genetic alterations have rarely been investigated or were only poorly predictable. Finally, we discuss the next steps for the implementation of AI-based surrogate tests in diagnostic workflows. © 2022 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Neutrophils prevent rectal bleeding in ulcerative colitis by peptidyl-arginine deiminase-4-dependent immunothrombosis.

OBJECTIVE: Bleeding ulcers and erosions are hallmarks of active ulcerative colitis (UC). However, the mechanisms controlling bleeding and mucosal haemostasis remain elusive. DESIGN: We used high-resolution endoscopy and colon tissue samples of active UC (n = 36) as well as experimental models of physical and chemical mucosal damage in mice deficient for peptidyl-arginine deiminase-4 (PAD4), gnotobiotic mice and controls. We employed endoscopy, histochemistry, live-cell microscopy and flow cytometry to study eroded mucosal surfaces during mucosal haemostasis. RESULTS: Erosions and ulcerations in UC were covered by fresh blood, haematin or fibrin visible by endoscopy. Fibrin layers rather than fresh blood or haematin on erosions were inversely correlated with rectal bleeding in UC. Fibrin layers contained ample amounts of neutrophils coaggregated with neutrophil extracellular traps (NETs) with detectable activity of PAD. Transcriptome analyses showed significantly elevated PAD4 expression in active UC. In experimentally inflicted wounds, we found that neutrophils underwent NET formation in a PAD4-dependent manner hours after formation of primary blood clots, and remodelled clots to immunothrombi containing citrullinated histones, even in the absence of microbiota. PAD4-deficient mice experienced an exacerbated course of dextrane sodium sulfate-induced colitis with markedly increased rectal bleeding (96 % vs 10 %) as compared with controls. PAD4-deficient mice failed to remodel blood clots on mucosal wounds eliciting impaired healing. Thus, NET-associated immunothrombi are protective in acute colitis, while insufficient immunothrombosis is associated with rectal bleeding. CONCLUSION: Our findings uncover that neutrophils induce secondary immunothrombosis by PAD4-dependent mechanisms. Insufficient immunothrombosis may favour rectal bleeding in UC.

pT3 colorectal cancer revisited: a multicentric study on the histological depth of invasion in more than 1000 pT3 carcinomas-proposal for a new pT3a/pT3b subclassification.

BACKGROUND: Pathological TNM staging (pTNM) is the strongest prognosticator in colorectal carcinoma (CRC) and the foundation of its post-operative clinical management. Tumours that invade pericolic/perirectal adipose tissue generally fall into the pT3 category without further subdivision. METHODS: The histological depth of invasion into the pericolic/perirectal fat was digitally and conventionally measured in a training cohort of 950 CRCs (Munich). We biostatistically calculated the optimal cut-off to stratify pT3 CRCs into novel pT3a (≤3 mm)/pT3b (>3 mm) subgroups, which were then validated in two independent cohorts (447 CRCs, Bayreuth/542 CRCs, Mainz). RESULTS: Compared to pT3a tumours, pT3b CRCs showed significantly worse disease-specific survival, including in pN0 vs pN+ and colonic vs. rectal cancers (DSS: P < 0.001, respectively, pooled analysis of all cohorts). Furthermore, the pT3a/pT3b subclassification remained an independent predictor of survival in multivariate analyses (e.g. DSS: P < 0.001, hazard ratio: 4.41 for pT3b, pooled analysis of all cohorts). While pT2/pT3a CRCs showed similar survival characteristics, pT3b cancers remained a distinct subgroup with dismal survival. DISCUSSION: The delineation of pT3a/pT3b subcategories of CRC based on the histological depth of adipose tissue invasion adds valuable prognostic information to the current pT3 classification and implementation into current staging practices of CRC should be considered.

The future of artificial intelligence in digital pathology - results of a survey across stakeholder groups.

AIMS: Artificial intelligence (AI) provides a powerful tool to extract information from digitised histopathology whole slide images. During the last 5 years, academic and commercial actors have developed new technical solutions for a diverse set of tasks, including tissue segmentation, cell detection, mutation prediction, prognostication and prediction of treatment response. In the light of limited overall resources, it is presently unclear for researchers, practitioners and policymakers which of these topics are stable enough for clinical use in the near future and which topics are still experimental, but worth investing time and effort into. METHODS AND RESULTS: To identify potentially promising applications of AI in pathology, we performed an anonymous online survey of 75 computational pathology domain experts from academia and industry. Participants enrolled in 2021 were queried about their subjective opinion on promising and appealing subfields of computational pathology with a focus upon solid tumours. The results of this survey indicate that the prediction of treatment response directly from routine pathology slides is regarded as the most promising future application. This item was ranked highest in the overall analysis and in subgroups by age and professional background. Furthermore, prediction of genetic alterations, gene expression and survival directly from routine pathology images scored consistently high throughout subgroups. CONCLUSIONS: Together, these data demonstrate a possible direction for the development of computational pathology systems in clinical, academic and industrial research in the near future.

Loss of CDX2 in colorectal cancer is associated with histopathologic subtypes and microsatellite instability but is prognostically inferior to hematoxylin-eosin-based morphologic parameters from the WHO classification.

BACKGROUND: Immunohistochemical loss of CDX2 has been proposed as a biomarker of dismal survival in colorectal carcinoma (CRC), especially in UICC Stage II/III. However, it remains unclear, how CDX2 expression is related to central hematoxylin-eosin (HE)-based morphologic parameters defined by 2019 WHO classification and how its prognostic relevance is compared to these parameters. METHODS: We evaluated CDX2 expression in 1003 CRCs and explored its prognostic relevance compared to CRC subtypes, tumour budding and WHO grade in the overall cohort and in specific subgroups. RESULTS: CDX2-low/absent CRCs were enriched in specific morphologic subtypes, right-sided and microsatellite-instable (MSI-H) CRCs (P < 0.001) and showed worse survival characteristics in the overall cohort/UICC Stage II/III (e.g. DFS: P = 0.005) and in microsatellite stable and left-sided CRCs, but not in MSI-H or right-sided CRCs. Compared with CDX2, all HE-based markers showed a significantly better prognostic discrimination in all scenarios. In multivariate analyses including all morphologic parameters, CDX2 was not an independent prognostic factor. CONCLUSION: CDX2 loss has some prognostic impact in univariate analyses, but its prognostic relevance is considerably lower compared to central HE-based morphologic parameters defined by the WHO classification and vanishes in multivariate analyses incorporating these factors.

Chronic intestinal inflammation drives colorectal tumor formation triggered by dietary heme iron in vivo.

The consumption of red meat is associated with an increased risk for colorectal cancer (CRC). Multiple lines of evidence suggest that heme iron as abundant constituent of red meat is responsible for its carcinogenic potential. However, the underlying mechanisms are not fully understood and particularly the role of intestinal inflammation has not been investigated. To address this important issue, we analyzed the impact of heme iron (0.25 µmol/g diet) on the intestinal microbiota, gut inflammation and colorectal tumor formation in mice. An iron-balanced diet with ferric citrate (0.25 µmol/g diet) was used as reference. 16S rRNA sequencing revealed that dietary heme reduced α-diversity and caused a persistent intestinal dysbiosis, with a continuous increase in gram-negative Proteobacteria. This was linked to chronic gut inflammation and hyperproliferation of the intestinal epithelium as attested by mini-endoscopy, histopathology and immunohistochemistry. Dietary heme triggered the infiltration of myeloid cells into colorectal mucosa with an increased level of COX-2 positive cells. Furthermore, flow cytometry-based phenotyping demonstrated an increased number of T cells and B cells in the lamina propria following heme intake, while γδ-T cells were reduced in the intraepithelial compartment. Dietary heme iron catalyzed formation of fecal N-nitroso compounds and was genotoxic in intestinal epithelial cells, yet suppressed intestinal apoptosis as evidenced by confocal microscopy and western blot analysis. Finally, a chemically induced CRC mouse model showed persistent intestinal dysbiosis, chronic gut inflammation and increased colorectal tumorigenesis following heme iron intake. Altogether, this study unveiled intestinal inflammation as important driver in heme iron-associated colorectal carcinogenesis.

PARP-1 protects against colorectal tumor induction, but promotes inflammation-driven colorectal tumor progression.

Colorectal cancer (CRC) is one of the most common tumor entities, which is causally linked to DNA repair defects and inflammatory bowel disease (IBD). Here, we studied the role of the DNA repair protein poly(ADP-ribose) polymerase-1 (PARP-1) in CRC. Tissue microarray analysis revealed PARP-1 overexpression in human CRC, correlating with disease progression. To elucidate its function in CRC, PARP-1 deficient (PARP-1-/-) and wild-type animals (WT) were subjected to azoxymethane (AOM)/ dextran sodium sulfate (DSS)-induced colorectal carcinogenesis. Miniendoscopy showed significantly more tumors in WT than in PARP-1-/- mice. Although the lack of PARP-1 moderately increased DNA damage, both genotypes exhibited comparable levels of AOM-induced autophagy and cell death. Interestingly, miniendoscopy revealed a higher AOM/DSS-triggered intestinal inflammation in WT animals, which was associated with increased levels of innate immune cells and proinflammatory cytokines. Tumors in WT animals were more aggressive, showing higher levels of STAT3 activation and cyclin D1 up-regulation. PARP-1-/- animals were then crossed with O6-methylguanine-DNA methyltransferase (MGMT)-deficient animals hypersensitive to AOM. Intriguingly, PARP-1-/-/MGMT-/- double knockout (DKO) mice developed more, but much smaller tumors than MGMT-/- animals. In contrast to MGMT-deficient mice, DKO animals showed strongly reduced AOM-dependent colonic cell death despite similar O6-methylguanine levels. Studies with PARP-1-/- cells provided evidence for increased alkylation-induced DNA strand break formation when MGMT was inhibited, suggesting a role of PARP-1 in the response to O6-methylguanine adducts. Our findings reveal PARP-1 as a double-edged sword in colorectal carcinogenesis, which suppresses tumor initiation following DNA alkylation in a MGMT-dependent manner, but promotes inflammation-driven tumor progression.

Appendiceal goblet cell carcinoids and adenocarcinomas ex-goblet cell carcinoid are genetically distinct from primary colorectal-type adenocarcinoma of the appendix.

The appendix gives rise to goblet cell carcinoids, which represent special carcinomas with distinct biological and histological features. Their genetic background and molecular relationship to colorectal adenocarcinoma is largely unknown. We therefore performed a next-generation sequencing analysis of 25 appendiceal carcinomas including 11 goblet cell carcinoids, 7 adenocarcinomas ex-goblet cell carcinoid, and 7 primary colorectal-type adenocarcinomas, using a modified Colorectal Cancer specific Panel comprising 32 genes linked to colorectal and neuroendocrine tumorigenesis. The mutational profiles of these neoplasms were compared with those of conventional adenocarcinomas, mixed adenoneuroendocrine carcinomas, and neuroendocrine carcinomas of the colorectum. In addition, a large-scale pan-cancer sequencing panel covering 409 genes was applied to selected cases of goblet cell carcinoid/adenocarcinoma ex-goblet cell carcinoid (n=2, respectively). Mutations in colorectal cancer-related genes (eg, TP53, KRAS, APC) were rare to absent in both, goblet cell carcinoids and adenocarcinomas ex-goblet cell carcinoid, but frequent in primary colorectal-type adenocarcinomas of the appendix. Additional large-scale sequencing of selected goblet cell carcinoids and adenocarcinomas ex-goblet cell carcinoid revealed mutations in Wnt-signaling-associated genes (USP9X, NOTCH1, CTNNA1, CTNNB1, TRRAP). These data suggest that appendiceal goblet cell carcinoids and adenocarcinomas ex-goblet cell carcinoid constitute a morphomolecular entity, histologically and genetically distinct from appendiceal colorectal-type adenocarcinomas and its colorectal counterparts. Altered Wnt-signaling associated genes, apart from APC, may act as potential drivers of these neoplasms. The absence of KRAS/NRAS mutations might render some of these tumors eligible for anti-EGFR directed therapy regimens.

VEGFR2 Signaling Prevents Colorectal Cancer Cell Senescence to Promote Tumorigenesis in Mice With Colitis.

BACKGROUND & AIMS: Senescence prevents cellular transformation. We investigated whether vascular endothelial growth factor (VEGF) signaling via its receptor, VEGFR2, regulates senescence and proliferation of tumor cells in mice with colitis-associated cancer (CAC). METHODS: CAC was induced in VEGFR2(ΔIEC) mice, which do not express VEGFR2 in the intestinal epithelium, and VEGFR2(fl/fl) mice (controls) by administration of azoxymethane followed by dextran sodium sulfate. Tumor development and inflammation were determined by endoscopy. Colorectal tissues were collected for immunoblot, immunohistochemical, and quantitative polymerase chain reaction analyses. Findings from mouse tissues were confirmed in human HCT116 colorectal cancer cells. We analyzed colorectal tumor samples from patients before and after treatment with bevacizumab. RESULTS: After colitis induction, VEGFR2(ΔIEC) mice developed significantly fewer tumors than control mice. A greater number of intestinal tumor cells from VEGFR2(ΔIEC) mice were in senescence than tumor cells from control mice. We found VEGFR2 to activate phosphatidylinositol-4,5-bisphosphate-3-kinase and AKT, resulting in inactivation of p21 in HCT116 cells. Inhibitors of VEGFR2 and AKT induced senescence in HCT116 cells. Tumor cell senescence promoted an anti-tumor immune response by CD8(+) T cells in mice. Patients whose tumor samples showed an increase in the proportion of senescent cells after treatment with bevacizumab had longer progression-free survival than patients in which the proportion of senescent tumor cells did not change before and after treatment. CONCLUSIONS: Inhibition of VEGFR2 signaling leads to senescence of human and mouse colorectal cancer cells. VEGFR2 interacts with phosphatidylinositol-4,5-bisphosphate-3-kinase and AKT to inactivate p21. Colorectal tumor senescence and p21 level correlate with patient survival during treatment with bevacizumab.

Colitis-associated neoplasia: molecular basis and clinical translation.

Crohn's disease and ulcerative colitis are both associated with an increased risk of inflammation-associated colorectal carcinoma. Colitis-associated cancer (CAC) is one of the most important causes for morbidity and mortality in patients with inflammatory bowel diseases (IBD). Colitis-associated neoplasia distinctly differs from sporadic colorectal cancer in its biology and the underlying mechanisms. This review discusses the molecular mechanisms of CAC and summarizes the most important genetic alterations and signaling pathways involved in inflammatory carcinogenesis. Then, clinical translation is evaluated by discussing new endoscopic techniques and their contribution to surveillance and early detection of CAC. Last, we briefly address different types of concepts for prevention (i.e., anti-inflammatory therapeutics) and treatment (i.e., surgical intervention) of CAC and give an outlook on this important aspect of IBD.

TROP2 in colorectal carcinoma: associations with histopathology, molecular phenotype, and patient prognosis.

Antibody-drug conjugates (ADCs) directed to trophoblast cell surface antigen 2 (TROP2) have gained approval as a therapeutic option for advanced triple-negative breast cancer, and TROP2 expression has been linked to unfavourable outcomes in various malignancies. In colorectal carcinoma (CRC), there is still a lack of comprehensive studies on its expression frequency and its prognostic implications in relation to the main clinicopathological parameters. We examined the expression of TROP2 in a large cohort of 1,052 CRC cases and correlated our findings with histopathological and molecular parameters, tumour stage, and patient outcomes. TROP2 was heterogeneously expressed in 214/1,052 CRCs (20.3%), with only a fraction of strongly positive tumours. TROP2 expression significantly correlated with an invasive histological phenotype (e.g. increased tumour budding/aggressive histopathological subtypes), advanced tumour stage, microsatellite stable tumours, and p53 alterations. While TROP2 expression was prognostic in univariable analyses of the overall cohort (e.g. for disease-free survival, p < 0.001), it exhibited distinct variations among important clinicopathological subgroups (e.g. right- versus left-sided CRC, microsatellite stable versus unstable CRC, Union for International Cancer Control [UICC] stages) and lost its significance in multivariable analyses that included stage and CRC histopathology. In summary, TROP2 is quite frequently expressed in CRC and associated with an aggressive histopathological phenotype and microsatellite stable tumours. Future clinical trials investigating anti-TROP2 ADCs should acknowledge the observed intratumoural heterogeneity, given that only a subset of TROP2-expressing CRC show strong positivity. The prognostic implications of TROP2 are complex and show substantial variations across crucial clinicopathological subgroups, thus indicating that TROP2 is a suboptimal parameter to predict patient prognosis.

From whole-slide image to biomarker prediction: end-to-end weakly supervised deep learning in computational pathology.

Hematoxylin- and eosin-stained whole-slide images (WSIs) are the foundation of diagnosis of cancer. In recent years, development of deep learning-based methods in computational pathology has enabled the prediction of biomarkers directly from WSIs. However, accurately linking tissue phenotype to biomarkers at scale remains a crucial challenge for democratizing complex biomarkers in precision oncology. This protocol describes a practical workflow for solid tumor associative modeling in pathology (STAMP), enabling prediction of biomarkers directly from WSIs by using deep learning. The STAMP workflow is biomarker agnostic and allows for genetic and clinicopathologic tabular data to be included as an additional input, together with histopathology images. The protocol consists of five main stages that have been successfully applied to various research problems: formal problem definition, data preprocessing, modeling, evaluation and clinical translation. The STAMP workflow differentiates itself through its focus on serving as a collaborative framework that can be used by clinicians and engineers alike for setting up research projects in the field of computational pathology. As an example task, we applied STAMP to the prediction of microsatellite instability (MSI) status in colorectal cancer, showing accurate performance for the identification of tumors high in MSI. Moreover, we provide an open-source code base, which has been deployed at several hospitals across the globe to set up computational pathology workflows. The STAMP workflow requires one workday of hands-on computational execution and basic command line knowledge.

A deep-learning workflow to predict upper tract urothelial carcinoma protein-based subtypes from H&E slides supporting the prioritization of patients for molecular testing.

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.

End-to-end prognostication in colorectal cancer by deep learning: a retrospective, multicentre study.

BACKGROUND: Precise prognosis prediction in patients with colorectal cancer (ie, forecasting survival) is pivotal for individualised treatment and care. Histopathological tissue slides of colorectal cancer specimens contain rich prognostically relevant information. However, existing studies do not have multicentre external validation with real-world sample processing protocols, and algorithms are not yet widely used in clinical routine. METHODS: In this retrospective, multicentre study, we collected tissue samples from four groups of patients with resected colorectal cancer from Australia, Germany, and the USA. We developed and externally validated a deep learning-based prognostic-stratification system for automatic prediction of overall and cancer-specific survival in patients with resected colorectal cancer. We used the model-predicted risk scores to stratify patients into different risk groups and compared survival outcomes between these groups. Additionally, we evaluated the prognostic value of these risk groups after adjusting for established prognostic variables. FINDINGS: We trained and validated our model on a total of 4428 patients. We found that patients could be divided into high-risk and low-risk groups on the basis of the deep learning-based risk score. On the internal test set, the group with a high-risk score had a worse prognosis than the group with a low-risk score, as reflected by a hazard ratio (HR) of 4·50 (95% CI 3·33-6·09) for overall survival and 8·35 (5·06-13·78) for disease-specific survival (DSS). We found consistent performance across three large external test sets. In a test set of 1395 patients, the high-risk group had a lower DSS than the low-risk group, with an HR of 3·08 (2·44-3·89). In two additional test sets, the HRs for DSS were 2·23 (1·23-4·04) and 3·07 (1·78-5·3). We showed that the prognostic value of the deep learning-based risk score is independent of established clinical risk factors. INTERPRETATION: Our findings indicate that attention-based self-supervised deep learning can robustly offer a prognosis on clinical outcomes in patients with colorectal cancer, generalising across different populations and serving as a potentially new prognostic tool in clinical decision making for colorectal cancer management. We release all source codes and trained models under an open-source licence, allowing other researchers to reuse and build upon our work. FUNDING: The German Federal Ministry of Health, the Max-Eder-Programme of German Cancer Aid, the German Federal Ministry of Education and Research, the German Academic Exchange Service, and the EU.