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

Journal

The Lancet. Digital health
ISSN: 2589-7500
Titre abrégé: Lancet Digit Health
Pays: England
ID NLM: 101751302

Informations de publication

Date de publication:
Jan 2024
Historique:
received: 05 04 2023
revised: 21 07 2023
accepted: 12 10 2023
medline: 21 12 2023
pubmed: 21 12 2023
entrez: 20 12 2023
Statut: ppublish

Résumé

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. 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. 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. 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. 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.

Sections du résumé

BACKGROUND 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 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 RESULTS
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 CONCLUSIONS
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 BACKGROUND
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.

Identifiants

DOI: 10.1016/S2589-7500(23)00208-X PMID: 38123254
pubmed: 38123254
pii: S2589-7500(23)00208-X
doi: 10.1016/S2589-7500(23)00208-X
pii:
doi:

Types de publication

Journal Article

Langues

eng

Sous-ensembles de citation

IM

Pagination

e33-e43

Informations de copyright

Copyright © 2024 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.

Déclaration de conflit d'intérêts

Declaration of interests JNK has received consulting fees from Owkin; DoMore Diagnostics; Panakeia; and Histofy; furthermore, JNK holds shares and holds a leadership role in StratifAI and has received honoraria for lectures by Bayer, Eisai, Merck Sharp & Dohme (MSD), Bristol-Myers Squibb (BMS), Roche, Pfizer, and Fresenius and has participated on a Data Safety Monitoring Board or Advisory Board for Bayer, Eisai, MSD, BMS, Roche, and Pfizer. PQ and NPW declare research funding from Roche and PQ consulting and speaker services for Roche. JJ declares consulting fees from WHO (Development of Digital Health Solution) and CARE International, Papua New Guinea (Development of Survey Instruments). SF declares grants and contracts from Bundesministerium für Bildung und Forschung, Deutsche Forschungsgemeinschaft, and German Cancer Aid, and payment or honoraria from BMS, MSD, European Society for Medical Oncology, and Deutsche Gesellschaft für Pathologie. DT holds shares in StratifAI. WM is a shareholder of Gemeinschaftspraxis Pathologie Starnberg, a private pathology practice in Germany. All other authors declare no competing interests.

Auteurs

Xiaofeng Jiang (X)

Else Kroener Fresenius Center for Digital Health, Technical University Dresden, Dresden, Germany; Department of Medicine III, University Hospital Rheinisch-Westfälische Technische Hochschule Aachen, Aachen, Germany.

Michael Hoffmeister (M)

Division of Clinical Epidemiology and Ageing Research, German Cancer Research Center, Heidelberg, Germany.

Hermann Brenner (H)

Division of Clinical Epidemiology and Ageing Research, German Cancer Research Center, Heidelberg, Germany; German Cancer Consortium, German Cancer Research Center, Heidelberg, Germany; Division of Preventive Oncology, German Cancer Research Center and National Center for Tumour Diseases, Heidelberg, Germany.

Hannah Sophie Muti (HS)

Else Kroener Fresenius Center for Digital Health, Technical University Dresden, Dresden, Germany.

Tanwei Yuan (T)

Division of Clinical Epidemiology and Ageing Research, German Cancer Research Center, Heidelberg, Germany.

Sebastian Foersch (S)

Institute of Pathology, University Medical Center Mainz, Mainz, Germany.

Nicholas P West (NP)

Pathology and Data Analytics, Leeds Institute of Medical Research at St James's, University of Leeds, Leeds, UK.

Alexander Brobeil (A)

Institute of Pathology, National Center for Tumour Diseases, University Hospital Heidelberg, Heidelberg, Germany; Tissue Bank, National Center for Tumour Diseases, University Hospital Heidelberg, Heidelberg, Germany.

Jitendra Jonnagaddala (J)

School of Population Health, Faculty of Medicine and Health, University of New South Wales Sydney, Kensington, NSW, Australia.

Nicholas Hawkins (N)

School of Medical Sciences, Faculty of Medicine and Health, University of New South Wales Sydney, Kensington, NSW, Australia.

Robyn L Ward (RL)

School of Medical Sciences, Faculty of Medicine and Health, University of New South Wales Sydney, Kensington, NSW, Australia; Faculty of Medicine and Health, The University of Sydney, Camperdown, NSW, Australia.

Titus J Brinker (TJ)

Digital Biomarkers for Oncology Group, German Cancer Research Center, Heidelberg, Germany.

Oliver Lester Saldanha (OL)

Else Kroener Fresenius Center for Digital Health, Technical University Dresden, Dresden, Germany; Department of Medicine III, University Hospital Rheinisch-Westfälische Technische Hochschule Aachen, Aachen, Germany.

Jia Ke (J)

Department of General Surgery (Colorectal Surgery), Guangdong Provincial Key Laboratory of Colorectal and Pelvic Floor Diseases, and Biomedical Innovation Center, The Sixth Affiliated Hospital, Sun Yat-sen University, Guangzhou, China.

Wolfram Müller (W)

Gemeinschaftspraxis Pathologie, Starnberg, Germany.

Heike I Grabsch (HI)

Pathology and Data Analytics, Leeds Institute of Medical Research at St James's, University of Leeds, Leeds, UK; Department of Pathology, GROW School for Oncology and Reproduction, Maastricht University Medical Center, Maastricht, Netherlands.

Philip Quirke (P)

Pathology and Data Analytics, Leeds Institute of Medical Research at St James's, University of Leeds, Leeds, UK.

Daniel Truhn (D)

Department of Diagnostic and Interventional Radiology, University Hospital Rheinisch-Westfälische Technische Hochschule Aachen, Aachen, Germany.

Jakob Nikolas Kather (JN)

Else Kroener Fresenius Center for Digital Health, Technical University Dresden, Dresden, Germany; Department of Medicine III, University Hospital Rheinisch-Westfälische Technische Hochschule Aachen, Aachen, Germany; Pathology and Data Analytics, Leeds Institute of Medical Research at St James's, University of Leeds, Leeds, UK; Medical Oncology, National Center for Tumour Diseases, University Hospital Heidelberg, Heidelberg, Germany; Department of Medicine I, University Hospital Dresden, Dresden, Germany. Electronic address: jakob-nikolas.kather@alumni.dkfz.de.

Classifications MeSH