Deep learning identifies inflamed fat as a risk factor for lymph node metastasis in early colorectal cancer.

The spread of early-stage (T1 and T2) adenocarcinomas to locoregional lymph nodes is a key event in disease progression of colorectal cancer (CRC). The cellular mechanisms behind this event are not completely understood and existing predictive biomarkers are imperfect. Here, we used an end-to-end deep learning algorithm to identify risk factors for lymph node metastasis (LNM) status in digitized histopathology slides of the primary CRC and its surrounding tissue. In two large population-based cohorts, we show that this system can predict the presence of more than one LNM in pT2 CRC patients with an area under the receiver operating curve (AUROC) of 0.733 (0.67-0.758) and patients with any LNM with an AUROC of 0.711 (0.597-0.797). Similarly, in pT1 CRC patients, the presence of more than one LNM or any LNM was predictable with an AUROC of 0.733 (0.644-0.778) and 0.567 (0.542-0.597), respectively. Based on these findings, we used the deep learning system to guide human pathology experts towards highly predictive regions for LNM in the whole slide images. This hybrid human observer and deep learning approach identified inflamed adipose tissue as the highest predictive feature for LNM presence. Our study is a first proof of concept that artificial intelligence (AI) systems may be able to discover potentially new biological mechanisms in cancer progression. Our deep learning algorithm is publicly available and can be used for biomarker discovery in any disease setting. © 2021 The Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Clinical-Grade Detection of Microsatellite Instability in Colorectal Tumors by Deep Learning.

BACKGROUND & AIMS: Microsatellite instability (MSI) and mismatch-repair deficiency (dMMR) in colorectal tumors are used to select treatment for patients. Deep learning can detect MSI and dMMR in tumor samples on routine histology slides faster and less expensively than molecular assays. However, clinical application of this technology requires high performance and multisite validation, which have not yet been performed. METHODS: We collected H&E-stained slides and findings from molecular analyses for MSI and dMMR from 8836 colorectal tumors (of all stages) included in the MSIDETECT consortium study, from Germany, the Netherlands, the United Kingdom, and the United States. Specimens with dMMR were identified by immunohistochemistry analyses of tissue microarrays for loss of MLH1, MSH2, MSH6, and/or PMS2. Specimens with MSI were identified by genetic analyses. We trained a deep-learning detector to identify samples with MSI from these slides; performance was assessed by cross-validation (N = 6406 specimens) and validated in an external cohort (n = 771 specimens). Prespecified endpoints were area under the receiver operating characteristic (AUROC) curve and area under the precision-recall curve (AUPRC). RESULTS: The deep-learning detector identified specimens with dMMR or MSI with a mean AUROC curve of 0.92 (lower bound, 0.91; upper bound, 0.93) and an AUPRC of 0.63 (range, 0.59-0.65), or 67% specificity and 95% sensitivity, in the cross-validation development cohort. In the validation cohort, the classifier identified samples with dMMR with an AUROC of 0.95 (range, 0.92-0.96) without image preprocessing and an AUROC of 0.96 (range, 0.93-0.98) after color normalization. CONCLUSIONS: We developed a deep-learning system that detects colorectal cancer specimens with dMMR or MSI using H&E-stained slides; it detected tissues with dMMR with an AUROC of 0.96 in a large, international validation cohort. This system might be used for high-throughput, low-cost evaluation of colorectal tissue specimens.

Endoscopic biopsies from gastrointestinal carcinomas and their suitability for molecular analysis: a review of the literature and recommendations for clinical practice and research.

Endoscopic biopsies (EBs) are the gold standard for diagnosing gastrointestinal carcinoma yet no guidelines address EB use for prognostic and predictive molecular testing. This review summarizes the reported quantity and quality of EBs, their relationship with molecular test failure rates and the resultant concordance between EB and resection specimen. Studies reporting molecular testing on gastrointestinal carcinoma EBs published between 2002 and 2014 were identified. Details regarding EB quantity, quality, tumour content, molecular test failure rates as well as causes and concordance with resection specimens were reviewed. Seventy-five studies were identified. Eighteen (24%) reported the mean EB number per patient (median: 2.1, range: 1-6.6 EBs). Sixty-one (81%) reported the frequency of test failure (median: 0%, range: 0-100%). Twenty-two (29%) investigated EB and resection specimen concordance (range: 0-100%). EB quantity and quality affected neither concordance nor failure rate. In summary, few studies currently report EB quantity, EB quality or EB and resection specimen concordance. Reliable molecular testing in EBs appears achievable, and can be representative of resection specimens. Concordance depends upon the testing methodology and biomarker heterogeneity within the tumour. To improve patient care, EB sampling, processing and reporting requires standardization and needs optimization for each biomarker individually.

Benchmarking weakly-supervised deep learning pipelines for whole slide classification in computational pathology.

Artificial intelligence (AI) can extract visual information from histopathological slides and yield biological insight and clinical biomarkers. Whole slide images are cut into thousands of tiles and classification problems are often weakly-supervised: the ground truth is only known for the slide, not for every single tile. In classical weakly-supervised analysis pipelines, all tiles inherit the slide label while in multiple-instance learning (MIL), only bags of tiles inherit the label. However, it is still unclear how these widely used but markedly different approaches perform relative to each other. We implemented and systematically compared six methods in six clinically relevant end-to-end prediction tasks using data from N=2980 patients for training with rigorous external validation. We tested three classical weakly-supervised approaches with convolutional neural networks and vision transformers (ViT) and three MIL-based approaches with and without an additional attention module. Our results empirically demonstrate that histological tumor subtyping of renal cell carcinoma is an easy task in which all approaches achieve an area under the receiver operating curve (AUROC) of above 0.9. In contrast, we report significant performance differences for clinically relevant tasks of mutation prediction in colorectal, gastric, and bladder cancer. In these mutation prediction tasks, classical weakly-supervised workflows outperformed MIL-based weakly-supervised methods for mutation prediction, which is surprising given their simplicity. This shows that new end-to-end image analysis pipelines in computational pathology should be compared to classical weakly-supervised methods. Also, these findings motivate the development of new methods which combine the elegant assumptions of MIL with the empirically observed higher performance of classical weakly-supervised approaches. We make all source codes publicly available at https://github.com/KatherLab/HIA, allowing easy application of all methods to any similar task.

Genomic and epigenomic EBF1 alterations modulate TERT expression in gastric cancer.

Transcriptional reactivation of telomerase catalytic subunit (TERT) is a frequent hallmark of cancer, occurring in 90% of human malignancies. However, specific mechanisms driving TERT reactivation remain obscure for many tumor types and in particular gastric cancer (GC), a leading cause of global cancer mortality. Here, through comprehensive genomic and epigenomic analysis of primary GCs and GC cell lines, we identified the transcription factor early B cell factor 1 (EBF1) as a TERT transcriptional repressor and inactivation of EBF1 function as a major cause of TERT upregulation. Abolishment of EBF1 function occurs through 3 distinct (epi)genomic mechanisms. First, EBF1 is epigenetically silenced via DNA methyltransferase, polycomb-repressive complex 2 (PRC2), and histone deacetylase activity in GCs. Second, recurrent, somatic, and heterozygous EBF1 DNA-binding domain mutations result in the production of dominant-negative EBF1 isoforms. Third, more rarely, genomic deletions and rearrangements proximal to the TERT promoter remobilize or abolish EBF1-binding sites, derepressing TERT and leading to high TERT expression. EBF1 is also functionally required for various malignant phenotypes in vitro and in vivo, highlighting its importance for GC development. These results indicate that multimodal genomic and epigenomic alterations underpin TERT reactivation in GC, converging on transcriptional repressors such as EBF1.

Tumour-microenvironment interactions: role of tumour stroma and proteins produced by cancer-associated fibroblasts in chemotherapy response.

BACKGROUND: Cytotoxic chemotherapy improves survival for some, but not all, cancer patients. Non-responders may experience unnecessary toxicity and cancer progression, thus creating an urgent need for biomarkers that can predict the response to chemotherapy. So far, the search for such biomarkers has primarily been focused on the cancer cells and less on their surrounding stroma. This stroma is known to act as a key regulator of tumour progression and, in addition, has been associated with drug delivery and drug efficacy. Fibroblasts represent the major cell type in cancer-associated stroma and they secrete extracellular matrix proteins as well as growth factors. This Medline-based literature review summarises the results from studies on epithelial cancers and aimed at investigating relationships between the quantity and quality of the intra-tumoral stroma, the cancer-associated fibroblasts, the proteins they produce and the concomitant response to chemotherapy. Biomarkers were selected for review that are known to affect cancer-related characteristics and patient prognosis. RESULTS: The current literature supports the hypothesis that biomarkers derived from the tumour stroma may be useful to predict response to chemotherapy. This notion appears to be related to the overall quantity and cellularity of the intra-tumoural stroma and the predominant constituents of the extracellular matrix. CONCLUSION: Increasing evidence is emerging showing that tumour-stroma interactions may not only affect tumour progression and patient prognosis, but also the response to chemotherapy. The tumour stroma-derived biomarkers that appear to be most appropriate to determine the patient's response to chemotherapy vary by tumour origin and the availability of pre-treatment tissue. For patients scheduled for adjuvant chemotherapy, the most promising biomarker appears to be the PLAU: SERPINE complex, whereas for patients scheduled for neo-adjuvant chemotherapy the tumour stroma quantity appears to be most relevant.