Artificial intelligence for detection of microsatellite instability in colorectal cancer-a multicentric analysis of a pre-screening tool for clinical application

Publikation: Beitrag in FachzeitschriftForschungsartikelBeigetragenBegutachtung

Beitragende

  • A Echle - , Universitätsklinikum Aachen (Autor:in)
  • N Ghaffari Laleh - , Else Kröner Fresenius Zentrum für Digitale Gesundheit, Universitätsklinikum Aachen (Autor:in)
  • P Quirke - , University of Leeds (Autor:in)
  • H I Grabsch - , University of Leeds (Autor:in)
  • H S Muti - , Zentrum für Chirurgie - ITS, Else Kröner Fresenius Zentrum für Digitale Gesundheit, Universitätsklinikum Aachen (Autor:in)
  • O L Saldanha - , Universitätsklinikum Aachen (Autor:in)
  • S F Brockmoeller - , University of Leeds (Autor:in)
  • P A van den Brandt - , Maastricht University (Autor:in)
  • G G A Hutchins - , University of Leeds (Autor:in)
  • S D Richman - , University of Leeds (Autor:in)
  • K Horisberger - , Universitätsspital Zürich (Autor:in)
  • C Galata - , Universität Heidelberg (Autor:in)
  • M P Ebert - , Universitätsmedizin Mannheim, Universitätsklinikum Heidelberg, Universität Heidelberg (Autor:in)
  • M Eckardt - , Universitätsmedizin Mannheim, Universitätsklinikum Heidelberg, Universität Heidelberg (Autor:in)
  • M Boutros - , Deutsches Krebsforschungszentrum (DKFZ) (Autor:in)
  • D Horst - , Charité – Universitätsmedizin Berlin (Autor:in)
  • C Reissfelder - , Universität Heidelberg (Autor:in)
  • E Alwers - , Deutsches Krebsforschungszentrum (DKFZ) (Autor:in)
  • T J Brinker - , Deutsches Krebsforschungszentrum (DKFZ) (Autor:in)
  • R Langer - , Inselspital - Universitätsspital Bern, Universität Bern (Autor:in)
  • J C A Jenniskens - , Maastricht University (Autor:in)
  • K Offermans - , Maastricht University (Autor:in)
  • W Mueller - , Gemeinschaftspraxis Pathologie (Autor:in)
  • R Gray - , University of Oxford (Autor:in)
  • S B Gruber - , Center for Precision Medicine and Department of Medical Oncology, City of Hope National Medical Center (Autor:in)
  • J K Greenson - , City of Hope Comprehensive Cancer Center (Autor:in)
  • G Rennert - , Technion-Israel Institute of Technology (Autor:in)
  • J D Bonner - , City of Hope National Medical Center (Autor:in)
  • D Schmolze - , City of Hope Comprehensive Cancer Center (Autor:in)
  • J Chang-Claude - , Universitätsklinikum Hamburg-Eppendorf (UKE) (Autor:in)
  • H Brenner - , Deutsches Krebsforschungszentrum (DKFZ) (Autor:in)
  • C Trautwein - , Universitätsklinikum Aachen (Autor:in)
  • P Boor - , Universitätsklinikum Aachen (Autor:in)
  • D Jaeger - , Klinik und Poliklinik für Kinderchirurgie, Nationales Zentrum für Tumorerkrankungen (NCT) Heidelberg, Universitätsklinikum Heidelberg (Autor:in)
  • N T Gaisa - , Universitätsklinikum Aachen (Autor:in)
  • M Hoffmeister - , Deutsches Krebsforschungszentrum (DKFZ) (Autor:in)
  • N P West - , University of Leeds (Autor:in)
  • J N Kather - , Else Kröner Fresenius Zentrum für Digitale Gesundheit, Universitätsklinikum Aachen, Leeds Teaching Hospitals NHS Trust, University of Leeds (Autor:in)

Abstract

BACKGROUND: Microsatellite instability (MSI)/mismatch repair deficiency (dMMR) is a key genetic feature which should be tested in every patient with colorectal cancer (CRC) according to medical guidelines. Artificial intelligence (AI) methods can detect MSI/dMMR directly in routine pathology slides, but the test performance has not been systematically investigated with predefined test thresholds.

METHOD: We trained and validated AI-based MSI/dMMR detectors and evaluated predefined performance metrics using nine patient cohorts of 8343 patients across different countries and ethnicities.

RESULTS: Classifiers achieved clinical-grade performance, yielding an area under the receiver operating curve (AUROC) of up to 0.96 without using any manual annotations. Subsequently, we show that the AI system can be applied as a rule-out test: by using cohort-specific thresholds, on average 52.73% of tumors in each surgical cohort [total number of MSI/dMMR = 1020, microsatellite stable (MSS)/ proficient mismatch repair (pMMR) = 7323 patients] could be identified as MSS/pMMR with a fixed sensitivity at 95%. In an additional cohort of N = 1530 (MSI/dMMR = 211, MSS/pMMR = 1319) endoscopy biopsy samples, the system achieved an AUROC of 0.89, and the cohort-specific threshold ruled out 44.12% of tumors with a fixed sensitivity at 95%. As a more robust alternative to cohort-specific thresholds, we showed that with a fixed threshold of 0.25 for all the cohorts, we can rule-out 25.51% in surgical specimens and 6.10% in biopsies.

INTERPRETATION: When applied in a clinical setting, this means that the AI system can rule out MSI/dMMR in a quarter (with global thresholds) or half of all CRC patients (with local fine-tuning), thereby reducing cost and turnaround time for molecular profiling.

Details

OriginalspracheEnglisch
Aufsatznummer100400
FachzeitschriftESMO open
Jahrgang7
Ausgabenummer2
PublikationsstatusVeröffentlicht - Apr. 2022
Peer-Review-StatusJa

Externe IDs

PubMedCentral PMC9058894
Scopus 85125539306

Schlagworte

Ziele für nachhaltige Entwicklung

Schlagwörter

  • Artificial Intelligence, Colorectal Neoplasms/diagnosis, DNA Mismatch Repair/genetics, Early Detection of Cancer, Humans, Microsatellite Instability