EPIC-Survival: End-to-end Part Inferred Clustering for Survival Analysis, with Prognostic Stratification Boosting

Hassan Muhammad, Chensu Xie, Carlie S Sigel, Michael Doukas, Lindsay Alpert, Amber Lea Simpson, Thomas J Fuchs
Proceedings of the Fourth Conference on Medical Imaging with Deep Learning, PMLR 143:520-531, 2021.

Abstract

Histopathology-based survival modelling has two major hurdles. Firstly, a well-performing survival model has minimal clinical application if it does not contribute to the stratification of a cancer patient cohort into different risk groups, preferably driven by histologic morphologies. In the clinical setting, individuals are not given specific prognostic predictions, but are rather predicted to lie within a risk group which has a general survival trend. Thus, It is imperative that a survival model produces well-stratified risk groups. Secondly, until now, survival modelling was done in a two-stage approach (encoding and aggregation). EPIC-Survival bridges encoding and aggregation into an end-to-end survival modelling approach, while introducing stratification boosting to encourage the model to not only optimize ranking, but also to discriminate between risk groups. In this study we show that EPIC-Survival performs better than other approaches in modelling intrahepatic cholangiocarcinoma (ICC), a historically difficult cancer to model. We found that stratification boosting further improves model performance and helps identify specific histologic differences, not commonly sought out in ICC.

Cite this Paper

BibTeX
@InProceedings{pmlr-v143-muhammad21a, title = {{EPIC}-Survival: End-to-end Part Inferred Clustering for Survival Analysis, with Prognostic Stratification Boosting}, author = {Muhammad, Hassan and Xie, Chensu and Sigel, Carlie S and Doukas, Michael and Alpert, Lindsay and Simpson, Amber Lea and Fuchs, Thomas J}, booktitle = {Proceedings of the Fourth Conference on Medical Imaging with Deep Learning}, pages = {520--531}, year = {2021}, editor = {Heinrich, Mattias and Dou, Qi and de Bruijne, Marleen and Lellmann, Jan and Schläfer, Alexander and Ernst, Floris}, volume = {143}, series = {Proceedings of Machine Learning Research}, month = {07--09 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v143/muhammad21a/muhammad21a.pdf}, url = {https://proceedings.mlr.press/v143/muhammad21a.html}, abstract = {Histopathology-based survival modelling has two major hurdles. Firstly, a well-performing survival model has minimal clinical application if it does not contribute to the stratification of a cancer patient cohort into different risk groups, preferably driven by histologic morphologies. In the clinical setting, individuals are not given specific prognostic predictions, but are rather predicted to lie within a risk group which has a general survival trend. Thus, It is imperative that a survival model produces well-stratified risk groups. Secondly, until now, survival modelling was done in a two-stage approach (encoding and aggregation). EPIC-Survival bridges encoding and aggregation into an end-to-end survival modelling approach, while introducing stratification boosting to encourage the model to not only optimize ranking, but also to discriminate between risk groups. In this study we show that EPIC-Survival performs better than other approaches in modelling intrahepatic cholangiocarcinoma (ICC), a historically difficult cancer to model. We found that stratification boosting further improves model performance and helps identify specific histologic differences, not commonly sought out in ICC.} }
Endnote
%0 Conference Paper %T EPIC-Survival: End-to-end Part Inferred Clustering for Survival Analysis, with Prognostic Stratification Boosting %A Hassan Muhammad %A Chensu Xie %A Carlie S Sigel %A Michael Doukas %A Lindsay Alpert %A Amber Lea Simpson %A Thomas J Fuchs %B Proceedings of the Fourth Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2021 %E Mattias Heinrich %E Qi Dou %E Marleen de Bruijne %E Jan Lellmann %E Alexander Schläfer %E Floris Ernst %F pmlr-v143-muhammad21a %I PMLR %P 520--531 %U https://proceedings.mlr.press/v143/muhammad21a.html %V 143 %X Histopathology-based survival modelling has two major hurdles. Firstly, a well-performing survival model has minimal clinical application if it does not contribute to the stratification of a cancer patient cohort into different risk groups, preferably driven by histologic morphologies. In the clinical setting, individuals are not given specific prognostic predictions, but are rather predicted to lie within a risk group which has a general survival trend. Thus, It is imperative that a survival model produces well-stratified risk groups. Secondly, until now, survival modelling was done in a two-stage approach (encoding and aggregation). EPIC-Survival bridges encoding and aggregation into an end-to-end survival modelling approach, while introducing stratification boosting to encourage the model to not only optimize ranking, but also to discriminate between risk groups. In this study we show that EPIC-Survival performs better than other approaches in modelling intrahepatic cholangiocarcinoma (ICC), a historically difficult cancer to model. We found that stratification boosting further improves model performance and helps identify specific histologic differences, not commonly sought out in ICC.
APA
Muhammad, H., Xie, C., Sigel, C.S., Doukas, M., Alpert, L., Simpson, A.L. & Fuchs, T.J.. (2021). EPIC-Survival: End-to-end Part Inferred Clustering for Survival Analysis, with Prognostic Stratification Boosting. Proceedings of the Fourth Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 143:520-531 Available from https://proceedings.mlr.press/v143/muhammad21a.html.

Related Material