MesoGraph: Automatic profiling of mesothelioma subtypes from histological images

Mark Eastwood; Heba Sailem; Silviu Tudor Marc; Xiaohong Gao; Judith Offman; Emmanouil Karteris; Angeles Montero Fernandez; Danny Jonigk; William Cookson; Miriam Moffatt; Sanjay Popat; Fayyaz Minhas; Jan Lukas Robertus

doi:10.1016/j.xcrm.2023.101226

MesoGraph: Automatic profiling of mesothelioma subtypes from histological images

Mark Eastwood^*, Heba Sailem, Silviu Tudor Marc, Xiaohong Gao, Judith Offman, Emmanouil Karteris, Angeles Montero Fernandez, Danny Jonigk, William Cookson, Miriam Moffatt, Sanjay Popat, Fayyaz Minhas, Jan Lukas Robertus

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

3 Citations (Scopus)

Abstract

Mesothelioma is classified into three histological subtypes, epithelioid, sarcomatoid, and biphasic, according to the relative proportions of epithelioid and sarcomatoid tumor cells present. Current guidelines recommend that the sarcomatoid component of each mesothelioma is quantified, as a higher percentage of sarcomatoid pattern in biphasic mesothelioma shows poorer prognosis. In this work, we develop a dual-task graph neural network (GNN) architecture with ranking loss to learn a model capable of scoring regions of tissue down to cellular resolution. This allows quantitative profiling of a tumor sample according to the aggregate sarcomatoid association score. Tissue is represented by a cell graph with both cell-level morphological and regional features. We use an external multicentric test set from Mesobank, on which we demonstrate the predictive performance of our model. We additionally validate our model predictions through an analysis of the typical morphological features of cells according to their predicted score.

Original language	English
Article number	101226
Journal	Cell Reports Medicine
Volume	4
Issue number	10
DOIs	https://doi.org/10.1016/j.xcrm.2023.101226
Publication status	Published - 17 Oct 2023

Keywords

cancer subtyping
digital pathology
graph neural networks
mesothelioma
multiple instance learning

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1016/j.xcrm.2023.101226

Cite this

@article{246a31c2dbf74978aca614db4080eadf,

title = "MesoGraph: Automatic profiling of mesothelioma subtypes from histological images",

abstract = "Mesothelioma is classified into three histological subtypes, epithelioid, sarcomatoid, and biphasic, according to the relative proportions of epithelioid and sarcomatoid tumor cells present. Current guidelines recommend that the sarcomatoid component of each mesothelioma is quantified, as a higher percentage of sarcomatoid pattern in biphasic mesothelioma shows poorer prognosis. In this work, we develop a dual-task graph neural network (GNN) architecture with ranking loss to learn a model capable of scoring regions of tissue down to cellular resolution. This allows quantitative profiling of a tumor sample according to the aggregate sarcomatoid association score. Tissue is represented by a cell graph with both cell-level morphological and regional features. We use an external multicentric test set from Mesobank, on which we demonstrate the predictive performance of our model. We additionally validate our model predictions through an analysis of the typical morphological features of cells according to their predicted score.",

keywords = "cancer subtyping, digital pathology, graph neural networks, mesothelioma, multiple instance learning",

author = "Mark Eastwood and Heba Sailem and Marc, {Silviu Tudor} and Xiaohong Gao and Judith Offman and Emmanouil Karteris and Fernandez, {Angeles Montero} and Danny Jonigk and William Cookson and Miriam Moffatt and Sanjay Popat and Fayyaz Minhas and Robertus, {Jan Lukas}",

note = "Publisher Copyright: {\textcopyright} 2023 The Authors",

year = "2023",

month = oct,

day = "17",

doi = "10.1016/j.xcrm.2023.101226",

language = "English",

volume = "4",

journal = "Cell Reports Medicine",

issn = "2666-3791",

publisher = "Cell Press",

number = "10",

}

TY - JOUR

T1 - MesoGraph

T2 - Automatic profiling of mesothelioma subtypes from histological images

AU - Eastwood, Mark

AU - Sailem, Heba

AU - Marc, Silviu Tudor

AU - Gao, Xiaohong

AU - Offman, Judith

AU - Karteris, Emmanouil

AU - Fernandez, Angeles Montero

AU - Jonigk, Danny

AU - Cookson, William

AU - Moffatt, Miriam

AU - Popat, Sanjay

AU - Minhas, Fayyaz

AU - Robertus, Jan Lukas

PY - 2023/10/17

Y1 - 2023/10/17

N2 - Mesothelioma is classified into three histological subtypes, epithelioid, sarcomatoid, and biphasic, according to the relative proportions of epithelioid and sarcomatoid tumor cells present. Current guidelines recommend that the sarcomatoid component of each mesothelioma is quantified, as a higher percentage of sarcomatoid pattern in biphasic mesothelioma shows poorer prognosis. In this work, we develop a dual-task graph neural network (GNN) architecture with ranking loss to learn a model capable of scoring regions of tissue down to cellular resolution. This allows quantitative profiling of a tumor sample according to the aggregate sarcomatoid association score. Tissue is represented by a cell graph with both cell-level morphological and regional features. We use an external multicentric test set from Mesobank, on which we demonstrate the predictive performance of our model. We additionally validate our model predictions through an analysis of the typical morphological features of cells according to their predicted score.

AB - Mesothelioma is classified into three histological subtypes, epithelioid, sarcomatoid, and biphasic, according to the relative proportions of epithelioid and sarcomatoid tumor cells present. Current guidelines recommend that the sarcomatoid component of each mesothelioma is quantified, as a higher percentage of sarcomatoid pattern in biphasic mesothelioma shows poorer prognosis. In this work, we develop a dual-task graph neural network (GNN) architecture with ranking loss to learn a model capable of scoring regions of tissue down to cellular resolution. This allows quantitative profiling of a tumor sample according to the aggregate sarcomatoid association score. Tissue is represented by a cell graph with both cell-level morphological and regional features. We use an external multicentric test set from Mesobank, on which we demonstrate the predictive performance of our model. We additionally validate our model predictions through an analysis of the typical morphological features of cells according to their predicted score.

KW - cancer subtyping

KW - digital pathology

KW - graph neural networks

KW - mesothelioma

KW - multiple instance learning

UR - http://www.scopus.com/inward/record.url?scp=85174075475&partnerID=8YFLogxK

U2 - 10.1016/j.xcrm.2023.101226

DO - 10.1016/j.xcrm.2023.101226

M3 - Article

C2 - 37816348

AN - SCOPUS:85174075475

SN - 2666-3791

VL - 4

JO - Cell Reports Medicine

JF - Cell Reports Medicine

IS - 10

M1 - 101226

ER -

MesoGraph: Automatic profiling of mesothelioma subtypes from histological images

Abstract

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this