Charlie Saillard1, Benoit Schmauch1, Oumeima Laifa1, Matahi Moarii1, Sylvain Toldo1, Mikhail Zaslavskiy1, Elodie Pronier1, Alexis Laurent2,3, Giuliana Amaddeo3,4,5, Hélène Regnault5, Daniele Sommacale2,3,4, Marianne Ziol6,7, Jean-Michel Pawlotsky3,4,8, Sébastien Mulé3,4,9, Alain Luciani3,4,9, Gilles Wainrib1, Thomas Clozel1, Pierre Courtiol1, Julien Calderaro3,4,10. 1. Owkin Lab, Owkin, Paris, France. 2. Assistance Publique-Hôpitaux de Paris, Department of Hepatobiliary and Digestive Surgery, Henri Mondor Hospital, Créteil, France. 3. Paris Est Créteil University, UPEC, Créteil, France. 4. INSERM U955, Team "Pathophysiology and Therapy of Chronic Viral Hepatitis and Related Cancers", Créteil, France. 5. Assistance Publique-Hôpitaux de Paris, Department of Hepatology, Henri Mondor Hospital, Créteil, France. 6. Assistance Publique-Hôpitaux de Paris, Department of Pathology, Jean Verdier Hospital, Bondy, France. 7. Functional Genomics of Solid Tumors, INSERM-1162, Paris 13 University, Paris, France. 8. National Reference Center for Viral Hepatitis B, C and Delta, Department of Virology, Henri Mondor Hospital, Créteil, France. 9. Assistance Publique-Hôpitaux de Paris, Department of Medical Imaging, Henri Mondor Hospital, Créteil, France. 10. Assistance Publique-Hôpitaux de Paris, Department of Pathology, Henri Mondor Hospital, Créteil, France.
Abstract
BACKGROUND AND AIMS: Standardized and robust risk-stratification systems for patients with hepatocellular carcinoma (HCC) are required to improve therapeutic strategies and investigate the benefits of adjuvant systemic therapies after curative resection/ablation. APPROACH AND RESULTS: In this study, we used two deep-learning algorithms based on whole-slide digitized histological slides (whole-slide imaging; WSI) to build models for predicting survival of patients with HCC treated by surgical resection. Two independent series were investigated: a discovery set (Henri Mondor Hospital, n = 194) used to develop our algorithms and an independent validation set (The Cancer Genome Atlas [TCGA], n = 328). WSIs were first divided into small squares ("tiles"), and features were extracted with a pretrained convolutional neural network (preprocessing step). The first deep-learning-based algorithm ("SCHMOWDER") uses an attention mechanism on tumoral areas annotated by a pathologist whereas the second ("CHOWDER") does not require human expertise. In the discovery set, c-indices for survival prediction of SCHMOWDER and CHOWDER reached 0.78 and 0.75, respectively. Both models outperformed a composite score incorporating all baseline variables associated with survival. Prognostic value of the models was further validated in the TCGA data set, and, as observed in the discovery series, both models had a higher discriminatory power than a score combining all baseline variables associated with survival. Pathological review showed that the tumoral areas most predictive of poor survival were characterized by vascular spaces, the macrotrabecular architectural pattern, and a lack of immune infiltration. CONCLUSIONS: This study shows that artificial intelligence can help refine the prediction of HCC prognosis. It highlights the importance of pathologist/machine interactions for the construction of deep-learning algorithms that benefit from expert knowledge and allow a biological understanding of their output.
BACKGROUND AND AIMS: Standardized and robust risk-stratification systems for patients with hepatocellular carcinoma (HCC) are required to improve therapeutic strategies and investigate the benefits of adjuvant systemic therapies after curative resection/ablation. APPROACH AND RESULTS: In this study, we used two deep-learning algorithms based on whole-slide digitized histological slides (whole-slide imaging; WSI) to build models for predicting survival of patients with HCC treated by surgical resection. Two independent series were investigated: a discovery set (Henri Mondor Hospital, n = 194) used to develop our algorithms and an independent validation set (The Cancer Genome Atlas [TCGA], n = 328). WSIs were first divided into small squares ("tiles"), and features were extracted with a pretrained convolutional neural network (preprocessing step). The first deep-learning-based algorithm ("SCHMOWDER") uses an attention mechanism on tumoral areas annotated by a pathologist whereas the second ("CHOWDER") does not require human expertise. In the discovery set, c-indices for survival prediction of SCHMOWDER and CHOWDER reached 0.78 and 0.75, respectively. Both models outperformed a composite score incorporating all baseline variables associated with survival. Prognostic value of the models was further validated in the TCGA data set, and, as observed in the discovery series, both models had a higher discriminatory power than a score combining all baseline variables associated with survival. Pathological review showed that the tumoral areas most predictive of poor survival were characterized by vascular spaces, the macrotrabecular architectural pattern, and a lack of immune infiltration. CONCLUSIONS: This study shows that artificial intelligence can help refine the prediction of HCC prognosis. It highlights the importance of pathologist/machine interactions for the construction of deep-learning algorithms that benefit from expert knowledge and allow a biological understanding of their output.
Authors: Raphael M Kronberg; Lena Haeberle; Melanie Pfaus; Haifeng C Xu; Karina S Krings; Martin Schlensog; Tilman Rau; Aleksandra A Pandyra; Karl S Lang; Irene Esposito; Philipp A Lang Journal: Cancers (Basel) Date: 2022-04-13 Impact factor: 6.575