Generalisation of automatic tumour segmentation in histopathological whole-slide images across multiple cancer types

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Skrede, Ole-Johan, Pradhan, Manohar, Isaksen, Maria Xepapadakis, Hveem, Tarjei Sveinsgjerd, Vlatkovic, Ljiljana, Nesbakken, Arild, Lindemann, Kristina, Kristensen, Gunnar B, Kasius, Jenneke, Zeimet, Alain G, Brustugun, Odd Terje, Busund, Lill-Tove Rasmussen, Richardsen, Elin H, Haug, Erik Skaaheim, Brennhovd, Bjørn, Rewcastle, Emma, Lillesand, Melinda, Kvikstad, Vebjørn, Janssen, Emiel, Kerr, David J, Liestøl, Knut, Albregtsen, Fritz, Kleppe, Andreas
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918159377760256
author Skrede, Ole-Johan
Pradhan, Manohar
Isaksen, Maria Xepapadakis
Hveem, Tarjei Sveinsgjerd
Vlatkovic, Ljiljana
Nesbakken, Arild
Lindemann, Kristina
Kristensen, Gunnar B
Kasius, Jenneke
Zeimet, Alain G
Brustugun, Odd Terje
Busund, Lill-Tove Rasmussen
Richardsen, Elin H
Haug, Erik Skaaheim
Brennhovd, Bjørn
Rewcastle, Emma
Lillesand, Melinda
Kvikstad, Vebjørn
Janssen, Emiel
Kerr, David J
Liestøl, Knut
Albregtsen, Fritz
Kleppe, Andreas
author_facet Skrede, Ole-Johan
Pradhan, Manohar
Isaksen, Maria Xepapadakis
Hveem, Tarjei Sveinsgjerd
Vlatkovic, Ljiljana
Nesbakken, Arild
Lindemann, Kristina
Kristensen, Gunnar B
Kasius, Jenneke
Zeimet, Alain G
Brustugun, Odd Terje
Busund, Lill-Tove Rasmussen
Richardsen, Elin H
Haug, Erik Skaaheim
Brennhovd, Bjørn
Rewcastle, Emma
Lillesand, Melinda
Kvikstad, Vebjørn
Janssen, Emiel
Kerr, David J
Liestøl, Knut
Albregtsen, Fritz
Kleppe, Andreas
contents Deep learning is expected to aid pathologists by automating tasks such as tumour segmentation. We aimed to develop one universal tumour segmentation model for histopathological images and examine its performance in different cancer types. The model was developed using over 20 000 whole-slide images from over 4 000 patients with colorectal, endometrial, lung, or prostate carcinoma. Performance was validated in pre-planned analyses on external cohorts with over 3 000 patients across six cancer types. Exploratory analyses included over 1 500 additional patients from The Cancer Genome Atlas. Average Dice coefficient was over 80% in all validation cohorts with en bloc resection specimens and in The Cancer Genome Atlas cohorts. No loss of performance was observed when comparing the universal model with models specialised on single cancer types. In conclusion, extensive and rigorous evaluations demonstrate that generic tumour segmentation by a single model is possible across cancer types, patient populations, sample preparations, and slide scanners.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11182
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Generalisation of automatic tumour segmentation in histopathological whole-slide images across multiple cancer types
Skrede, Ole-Johan
Pradhan, Manohar
Isaksen, Maria Xepapadakis
Hveem, Tarjei Sveinsgjerd
Vlatkovic, Ljiljana
Nesbakken, Arild
Lindemann, Kristina
Kristensen, Gunnar B
Kasius, Jenneke
Zeimet, Alain G
Brustugun, Odd Terje
Busund, Lill-Tove Rasmussen
Richardsen, Elin H
Haug, Erik Skaaheim
Brennhovd, Bjørn
Rewcastle, Emma
Lillesand, Melinda
Kvikstad, Vebjørn
Janssen, Emiel
Kerr, David J
Liestøl, Knut
Albregtsen, Fritz
Kleppe, Andreas
Image and Video Processing
Artificial Intelligence
Computer Vision and Pattern Recognition
Deep learning is expected to aid pathologists by automating tasks such as tumour segmentation. We aimed to develop one universal tumour segmentation model for histopathological images and examine its performance in different cancer types. The model was developed using over 20 000 whole-slide images from over 4 000 patients with colorectal, endometrial, lung, or prostate carcinoma. Performance was validated in pre-planned analyses on external cohorts with over 3 000 patients across six cancer types. Exploratory analyses included over 1 500 additional patients from The Cancer Genome Atlas. Average Dice coefficient was over 80% in all validation cohorts with en bloc resection specimens and in The Cancer Genome Atlas cohorts. No loss of performance was observed when comparing the universal model with models specialised on single cancer types. In conclusion, extensive and rigorous evaluations demonstrate that generic tumour segmentation by a single model is possible across cancer types, patient populations, sample preparations, and slide scanners.
title Generalisation of automatic tumour segmentation in histopathological whole-slide images across multiple cancer types
topic Image and Video Processing
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.11182