Weakly-supervised learning for image-based classification of primary melanomas into genomic immune subgroups

Lucy Godson, Navid Alemi, Jérémie Nsengimana, Graham P. Cook, Emily L. Clarke, Darren Treanor, D. Timothy Bishop, Newton-Bishop Julia, Ali Gooya
Proceedings of The 5th International Conference on Medical Imaging with Deep Learning, PMLR 172:423-440, 2022.

Abstract

Determining early-stage prognostic markers and stratifying patients for effective treatment are two key challenges for improving outcomes for melanoma patients. Previous studies have used tumour transcriptome data to stratify patients into immune subgroups, which were associated with differential melanoma specific survival and potential treatment strategies. However, acquiring transcriptome data is a time-consuming and costly process. Moreover, it is not routinely used in the current clinical workflow. Here we attempt to overcome this by developing deep learning models to classify gigapixel H&E stained pathology slides, which are well established in clinical workflows, into these immune subgroups. Previous subtyping approaches have employed supervised learning which requires fully annotated data, or have only examined single genetic mutations in melanoma patients. We leverage a multiple-instance learning approach, which only requires slide-level labels and uses an attention mechanism to highlight regions of high importance to the classification. Moreover, we show that pathology-specific self-supervised models generate better representations compared to pathology-agnostic models for improving our model performance, achieving a mean AUC of 0.76 for classifying histopathology images as high or low immune subgroups. We anticipate that this method may allow us to find new biomarkers of high importance and could act as a tool for clinicians to infer the immune landscape of tumours and stratify patients, without needing to carry out additional expensive genetic tests.

Cite this Paper

BibTeX
@InProceedings{pmlr-v172-godson22a, title = {Weakly-supervised learning for image-based classification of primary melanomas into genomic immune subgroups}, author = {Godson, Lucy and Alemi, Navid and Nsengimana, J\'er\'emie and Cook, Graham P. and Clarke, Emily L. and Treanor, Darren and Bishop, D. Timothy and Julia, Newton-Bishop and Gooya, Ali}, booktitle = {Proceedings of The 5th International Conference on Medical Imaging with Deep Learning}, pages = {423--440}, year = {2022}, editor = {Konukoglu, Ender and Menze, Bjoern and Venkataraman, Archana and Baumgartner, Christian and Dou, Qi and Albarqouni, Shadi}, volume = {172}, series = {Proceedings of Machine Learning Research}, month = {06--08 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v172/godson22a/godson22a.pdf}, url = {https://proceedings.mlr.press/v172/godson22a.html}, abstract = {Determining early-stage prognostic markers and stratifying patients for effective treatment are two key challenges for improving outcomes for melanoma patients. Previous studies have used tumour transcriptome data to stratify patients into immune subgroups, which were associated with differential melanoma specific survival and potential treatment strategies. However, acquiring transcriptome data is a time-consuming and costly process. Moreover, it is not routinely used in the current clinical workflow. Here we attempt to overcome this by developing deep learning models to classify gigapixel H&E stained pathology slides, which are well established in clinical workflows, into these immune subgroups. Previous subtyping approaches have employed supervised learning which requires fully annotated data, or have only examined single genetic mutations in melanoma patients. We leverage a multiple-instance learning approach, which only requires slide-level labels and uses an attention mechanism to highlight regions of high importance to the classification. Moreover, we show that pathology-specific self-supervised models generate better representations compared to pathology-agnostic models for improving our model performance, achieving a mean AUC of 0.76 for classifying histopathology images as high or low immune subgroups. We anticipate that this method may allow us to find new biomarkers of high importance and could act as a tool for clinicians to infer the immune landscape of tumours and stratify patients, without needing to carry out additional expensive genetic tests.} }
Endnote
%0 Conference Paper %T Weakly-supervised learning for image-based classification of primary melanomas into genomic immune subgroups %A Lucy Godson %A Navid Alemi %A Jérémie Nsengimana %A Graham P. Cook %A Emily L. Clarke %A Darren Treanor %A D. Timothy Bishop %A Newton-Bishop Julia %A Ali Gooya %B Proceedings of The 5th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2022 %E Ender Konukoglu %E Bjoern Menze %E Archana Venkataraman %E Christian Baumgartner %E Qi Dou %E Shadi Albarqouni %F pmlr-v172-godson22a %I PMLR %P 423--440 %U https://proceedings.mlr.press/v172/godson22a.html %V 172 %X Determining early-stage prognostic markers and stratifying patients for effective treatment are two key challenges for improving outcomes for melanoma patients. Previous studies have used tumour transcriptome data to stratify patients into immune subgroups, which were associated with differential melanoma specific survival and potential treatment strategies. However, acquiring transcriptome data is a time-consuming and costly process. Moreover, it is not routinely used in the current clinical workflow. Here we attempt to overcome this by developing deep learning models to classify gigapixel H&E stained pathology slides, which are well established in clinical workflows, into these immune subgroups. Previous subtyping approaches have employed supervised learning which requires fully annotated data, or have only examined single genetic mutations in melanoma patients. We leverage a multiple-instance learning approach, which only requires slide-level labels and uses an attention mechanism to highlight regions of high importance to the classification. Moreover, we show that pathology-specific self-supervised models generate better representations compared to pathology-agnostic models for improving our model performance, achieving a mean AUC of 0.76 for classifying histopathology images as high or low immune subgroups. We anticipate that this method may allow us to find new biomarkers of high importance and could act as a tool for clinicians to infer the immune landscape of tumours and stratify patients, without needing to carry out additional expensive genetic tests.
APA
Godson, L., Alemi, N., Nsengimana, J., Cook, G.P., Clarke, E.L., Treanor, D., Bishop, D.T., Julia, N. & Gooya, A.. (2022). Weakly-supervised learning for image-based classification of primary melanomas into genomic immune subgroups. Proceedings of The 5th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 172:423-440 Available from https://proceedings.mlr.press/v172/godson22a.html.

Related Material