Search

ObjectiveMismatch Repair Deficiency (dMMR) / Microsatellite Instability (MSI) is a key biomarker in colorectal cancer (CRC). Universal screening of CRC patients for dMMR/MSI status is now recommended, but contributes to increased workload for pathologists and delayed therapeutic decisions. Deep learning has the potential to ease dMMR/MSI testing in clinical practice, yet no comprehensive validation of a clinically approved tool has been conducted.DesignWe developed an MSI pre-screening tool, MSIntuit, that uses deep learning to identify MSI status from H&E slides. For training, we used 859 slides from the TCGA database. A blind validation was subsequently performed on an independent dataset of 600 consecutive CRC patients. Each slide was digitised using Phillips-UFS and Ventana-DP200 scanners. Thirty dMMR/MSI slides were used for calibration on each scanner. Prediction was then performed on the remaining 570 patients following an automated quality check step. The inter and intra-scanner reliability was studied to assess MSIntuit’s robustness.ResultsMSIntuit reached a sensitivity and specificity of 97% (95% CI: 93-100%) / 46% (42-50%) on DP200 and of 95% (90-98%) / 47% (43-51%) on UFS scanner. MSIntuit reached excellent agreement on the two scanners (Cohen’s κ: 0.82) and was repeatable across multiple rescanning of the same slide (Fleiss’ κ: 0.82).ConclusionWe performed a successful blind validation of the first clinically approved AI-based tool for MSI detection from H&E slides. MSIntuit reaches sensitivity comparable to gold standard methods (92-95%) while ruling out almost half of the non-MSI population, paving the way for its use in clinical practice.

Our manuscript «PACpAInt: a deep learning approach to identify molecular subtypes of pancreatic adenocarcinoma on histology slides » being currently under review, a sample of code and data developed for the paper under consideration is made available to the journal referees. This sample is shared to the manuscript reviewers only for reproducibility purposes and should not be used for any other purpose. Inside code / pacpaint, one can find README.md which contains the information to launch the experiments.

Introduction: Pancreatic adenocarcinoma (PAC) is highly heterogeneous, resulting in overall ineffectiveness of most anti-tumor treatments. Two tumor subtypes (Classical and Basal) and two stromal subtypes (“active” and “inactive”) have been described. The Basal and the active stroma have worse prognosis. These subtypes could also be predictive of the response to different chemotherapies. To date, molecular subtype classification or phenotype quantification can only be defined by RNAseq, a complex technique sensitive to the quantity and quality of samples, requiring a timescale limiting its routine application. We propose a deep learning approach (PACpAInt) to predict molecular subtypes in PAC on routine histological slides. Patients and Methods: 424 digitalized HES slides of 202 resected PAC from 3 centers with clinical and transcriptomic data were used as training cohort. 3 validation cohorts were used (i) 250 resected PAC from a 4th center including 97 cases with an exact HES/RNAseq spatial match and all tumor slides digitalized (n = 891); (ii) 126 resected PAC from the TCGA (HES + RNAseq); (iii) 25 liver biopsies from metastatic PAC (HES + RNAseq). A multi-step deep learning model was developed to recognize tumor tissue, tumor from stroma cells, and then predicts their transcriptomic molecular subtypes, either at the level of an entire slide, or at the tile level (squares of 112 μm) allowing to study intratumor heterogeneity. Results: PACpAInt correctly predicted the tumor subtype at the whole slide level (AUC = 0.86 and 0.81 in 2 validation cohorts) and improved for samples with unambiguous molecular subtype (AUC = 0.91 and 0.88) confirming the limit of a binary approach. Similar results were obtained on liver biopsies (AUC = 0.85 and 0.92 on unambiguous cases). PACpAInt independently predicted progression-free and overall survival (PFS HR=1.37 [1.16 - 1.62] and OS HR=1.27 [1.08 - 1.49]). Analysis of all tumor slides from 77 Classical cases showed that 39% were heterogeneous with a Basal contingent. These cases had shorter PFS (15 vs. 47 months, p= 0.001) and OS (31 vs. 64 months, p= 5e-5).The analysis of intratumoral heterogeneity using PACpAInt predicted the molecular subtype of tumor and stroma cells of each tile within a slide (> 6 million tiles analyzed). 61% of cases had a main subtype either Classical (42%) or Basal (19%). 39% of the cases were ambiguous could be considered either hybrid (10%) with coexistence of Basal and Classical cells, or intermediate (29%) corresponding to homogeneous tumors but of intermediate differentiation. This classification had a strong prognostic impact (OS: 45.1 vs 33.0 vs 23.4 vs 13.6 months, resp. Classical, intermediate, hybrid, Basal; p Conclusion: This study provides the first PAC subtyping tool widely usable in clinical practice, opening the possibility of molecular classification useful for routine care and clinical trials. Citation Format: Charlie Saillard, Flore Delecourt, Benoit Schmauch, Olivier Moindrot, Magali Svrcek, Armelle Bardier-Dupas, Jean Francois Emile, Mira Ayadi, Vinciane Rebours, Louis De Mestier, Pascal Hammel, Cindy Neuzillet, Jean Baptiste Bachet, Juan Iovanna, Nelson Dusetti, Yuna Blum, Magali Richard, Yasmina Kermezli, Valerie Paradis, Mikhail Zaslavskiy, Pierre Courtiol, Aurelie Kamoun, Remy Nicolle, Jerome Cros. PACpAInt: A deep learning approach to identify molecular subtypes of pancreatic adenocarcinoma on histology slides [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 472.

We thank the authors for their interest in our study, that aimed at building deep learning-based models for the prediction of survival after resection of hepatocellular carcinoma (HCC).(1) They raise the potential issue of adjuvant treatment as a source of bias.

Deep learning methods for digital pathology analysis are an effective way to address multiple clinical questions, from diagnosis to prediction of treatment outcomes. These methods have also been used to predict gene mutations from pathology images, but no comprehensive evaluation of their potential for extracting molecular features from histology slides has yet been performed. We show that HE2RNA, a model based on the integration of multiple data modes, can be trained to systematically predict RNA-Seq profiles from whole-slide images alone, without expert annotation. Through its interpretable design, HE2RNA provides virtual spatialization of gene expression, as validated by CD3- and CD20-staining on an independent dataset. The transcriptomic representation learned by HE2RNA can also be transferred on other datasets, even of small size, to increase prediction performance for specific molecular phenotypes. We illustrate the use of this approach in clinical diagnosis purposes such as the identification of tumors with microsatellite instability., RNA-sequencing of tumour tissue can provide important diagnostic and prognostic information but this is costly and not routinely performed in all clinical settings. Here, the authors show that whole slide histology slides—part of routine care—can be used to predict RNA-sequencing data and thus reduce the need for additional analyses.

Deep learning methods for digital pathology analysis have proved an effective way to address multiple clinical questions, from diagnosis to prognosis and even to prediction of treatment outcomes. They have also recently been used to predict gene mutations from pathology images, but no comprehensive evaluation of their potential for extracting molecular features from histology slides has yet been performed. We propose a novel approach based on the integration of multiple data modes, and show that our deep learning model, HE2RNA, can be trained to systematically predict RNA-Seq profiles from whole-slide images alone, without the need for expert annotation. HE2RNA is interpretable by design, opening up new opportunities for virtual staining. In fact, it provides virtual spatialization of gene expression, as validated by double-staining on an independent dataset. Moreover, the transcriptomic representation learned by HE2RNA can be transferred to improve predictive performance for other tasks, particularly for small datasets. As an example of a task with direct clinical impact, we studied the prediction of microsatellite instability from hematoxylin & eosin stained images and our results show that better performance can be achieved in this setting.

4141 Background: Pancreatic adenocarcinoma (PAC) is predicted to be the second cause of death by cancer in 2030 and its prognosis has seen little improvement in the last decades. PAC is a very heterogeneous tumor with preeminent stroma and multiple histological aspects. Omic studies confirmed its molecular heterogeneity, possibly one of the main factors explaining the failure of most clinical trials. Two and three transcriptomic subtypes of tumor cells and stroma respectively, were described with major prognostic and predictive implications. The tumor subtypes, Basal-like and Classical, have been shown by several groups to be predictive of the response to first line chemotherapy. As of today, these subtypes can only be defined by RNA profiling which is limited by the quantity and quality of the samples (formalin fixation and low cellularity) as well as by the analytical delay that may restrict its application in routine care. In addition, tumors may harbor a mixture of several subtypes limiting their interpretation using bulk transcriptomic approaches and thereby their clinical use. Here, we propose a multistep approach using deep learning models to predict tumor components and their molecular subtypes on routine histological preparations. Methods: 728 whole-slide digitized histological slides corresponding to 350 consecutive resected PAC from four centers with clinical and transcriptomic data were assembled and used as a discovery set. PAC from TCGA (n = 134) was used as a validation set. Tumor regions from slides of the discovery set were annotated to train a multistep deep learning model that first recognizes tumor tissue and then predicts tumor and stroma cells molecular subtypes assessed by the published PurIST algorithm. Results: The tumor detection model was very efficient (AUC = 0.98 in the TCGA validation cohort). In the discovery set, the Basal-like/Classical classification performance of the model by cross validation was 0.79 (AUC) and reached 0.86 when restricted to samples with a high-confidence RNA-defined molecular subtype.Subtypes defined by the model were independently associated with overall survival in multivariate analysis (HR = 2.56 [1.87 - 3.49], pval < 0.001), and association was higher relatively to PurIST RNA subtypes (HR = 1.60 [1.17 - 2.19] pval < 0.001). In the validation cohort, the model had an overall AUC of 0.82, and 0.89 in the subset of “subtype-pure” tumors. In addition to demonstrating the value of histology-based deep learning models for tumor subtyping in PAC, these results also show the limit of molecular-based subtyping in highly heterogeneous samples. Conclusions: This study provides the first PAC subtyping tool usable worldwide in clinical practice, finally opening the possibility of patient molecular stratification in routine care and clinical trials.

Today, pathology imaging is one of the most common and inexpensive diagnostic/prognostic tools used in oncology, while more sophisticated methods such as next generation sequencing (NGS) remain relatively expensive and not routinely used in a clinical setting. Deep convolutional neural networks (CNNs) have emerged as an important image analysis technology enhancing the workflow of pathologists and improving the prediction of patient prognosis and response to treatment​​. Recently, a few attempts have been made to predict molecular features from tissue imaging using CNNs. While these preliminary results are encouraging, there have been no systematic attempts to link Whole Slide Images (WSIs) to transcriptomic profiles. In this study, we developed a cutting-edge deep learning model named HE2RNA, specifically customized for the direct prediction of gene expression from H&E-stained WSIs without need for annotation from pathologists. Our model was trained and tested on 8,725 patients from 28 different cancer types available at The Cancer Genome Atlas (TCGA). HE2RNA accurately predicted the expression of six gene signatures related to well known cancer hallmarks (angiogenesis, hypoxia, DNA repair, cell cycle and immunity) and performed particularly well for signalling pathways involved in immune cell activation. This indicates that suitably designed deep learning models can recognize subtle structures in tissue imaging and relate them to molecular portraits. Moreover, HE2RNA is designed to generate a spatial representation (virtual map) of any well-predicted gene expression overlaying the H&E slide. Such a virtual map was validated on a double-stained H&E/CD3 slide obtained from an independent hepatocellular carcinoma sample. This spatialization could be useful in augmenting the pathologists' workflow by providing a virtual multiplexed staining for each H&E slide while overcoming the technical issues associated with immunohistochemistry. Various important prognostic factors, such as microsatellite instability (MSI), are derived from molecular features. Microsatellite instability refers to the hypermutability of short repetitive genomic sequences caused by impaired DNA mismatch repair. These mutations frequently observed in gastric and colorectal cancer are associated with better response to immunotherapy. We show that the transcriptomic representation learned by our model can be used to improve the performance of MSI status prediction for small datasets of WSI. This type of setting is common since large databases of matched RNA-Seq profiles and WSI are widely available, while databases of matched MSI status and WSI are more scarce. In the future, such technologies could therefore facilitate universal screening of molecular biomarkers and improved identification of patients that could benefit from new therapeutic strategies. Citation Format: Elodie Pronier, Benoît Schmauch, Alberto Romagnoni, Charlie Saillard, Pascale Maillé, Julien Calderaro, Meriem Sefta, Sylvain Toldo, Mikhail Zaslavskiy, Thomas Clozel, Matahi Moarii, Pierre Courtiol, Gilles Wainrib. HE2RNA: A deep learning model for transcriptomic learning from digital pathology [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2105.

Purpose The purpose of this study was to create an algorithm that simultaneously detects and characterizes (benign vs. malignant) focal liver lesion (FLL) using deep learning. Materials and methods We trained our algorithm on a dataset proposed during a data challenge organized at the 2018 Journees Francophones de Radiologie. The dataset was composed of 367 two-dimensional ultrasound images from 367 individual livers, captured at various institutions. The algorithm was guided using an attention mechanism with annotations made by a radiologist. The algorithm was then tested on a new data set from 177 patients. Results The models reached mean ROC-AUC scores of 0.935 for FLL detection and 0.916 for FLL characterization over three shuffled three-fold cross-validations performed with the training data. On the new dataset of 177 patients, our models reached a weighted mean ROC-AUC scores of 0.891 for seven different tasks. Conclusion This study that uses a supervised-attention mechanism focused on FLL detection and characterization from liver ultrasound images. This method could prove to be highly relevant for medical imaging once validated on a larger independent cohort.