Survival Analysis for Idiopathic Pulmonary Fibrosis using CT Images and Incomplete Clinical Data

Ahmed Shahin, Joseph Jacob, Daniel Alexander, David Barber
Proceedings of The 5th International Conference on Medical Imaging with Deep Learning, PMLR 172:1057-1074, 2022.

Abstract

Idiopathic Pulmonary Fibrosis (IPF) is an inexorably progressive fibrotic lung disease with a variable and unpredictable rate of progression. CT scans of the lungs inform clinical assessment of IPF patients and contain pertinent information related to disease progression. In this work, we propose a multi-modal method that uses neural networks and memory banks to predict the survival of IPF patients using clinical and imaging data. The majority of clinical IPF patient records have missing data (e.g. missing lung function tests). To this end, we propose a probabilistic model that captures the dependencies between the observed clinical variables and imputes missing ones. This principled approach to missing data imputation can be naturally combined with a deep survival analysis model. We show that the proposed framework yields significantly better survival analysis results than baselines in terms of concordance index and integrated Brier score. Our work also provides insights into novel image-based biomarkers that are linked to mortality.

Cite this Paper

BibTeX
@InProceedings{pmlr-v172-shahin22a, title = {Survival Analysis for Idiopathic Pulmonary Fibrosis using CT Images and Incomplete Clinical Data}, author = {Shahin, Ahmed and Jacob, Joseph and Alexander, Daniel and Barber, David}, booktitle = {Proceedings of The 5th International Conference on Medical Imaging with Deep Learning}, pages = {1057--1074}, 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/shahin22a/shahin22a.pdf}, url = {https://proceedings.mlr.press/v172/shahin22a.html}, abstract = {Idiopathic Pulmonary Fibrosis (IPF) is an inexorably progressive fibrotic lung disease with a variable and unpredictable rate of progression. CT scans of the lungs inform clinical assessment of IPF patients and contain pertinent information related to disease progression. In this work, we propose a multi-modal method that uses neural networks and memory banks to predict the survival of IPF patients using clinical and imaging data. The majority of clinical IPF patient records have missing data (e.g. missing lung function tests). To this end, we propose a probabilistic model that captures the dependencies between the observed clinical variables and imputes missing ones. This principled approach to missing data imputation can be naturally combined with a deep survival analysis model. We show that the proposed framework yields significantly better survival analysis results than baselines in terms of concordance index and integrated Brier score. Our work also provides insights into novel image-based biomarkers that are linked to mortality.} }
Endnote
%0 Conference Paper %T Survival Analysis for Idiopathic Pulmonary Fibrosis using CT Images and Incomplete Clinical Data %A Ahmed Shahin %A Joseph Jacob %A Daniel Alexander %A David Barber %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-shahin22a %I PMLR %P 1057--1074 %U https://proceedings.mlr.press/v172/shahin22a.html %V 172 %X Idiopathic Pulmonary Fibrosis (IPF) is an inexorably progressive fibrotic lung disease with a variable and unpredictable rate of progression. CT scans of the lungs inform clinical assessment of IPF patients and contain pertinent information related to disease progression. In this work, we propose a multi-modal method that uses neural networks and memory banks to predict the survival of IPF patients using clinical and imaging data. The majority of clinical IPF patient records have missing data (e.g. missing lung function tests). To this end, we propose a probabilistic model that captures the dependencies between the observed clinical variables and imputes missing ones. This principled approach to missing data imputation can be naturally combined with a deep survival analysis model. We show that the proposed framework yields significantly better survival analysis results than baselines in terms of concordance index and integrated Brier score. Our work also provides insights into novel image-based biomarkers that are linked to mortality.
APA
Shahin, A., Jacob, J., Alexander, D. & Barber, D.. (2022). Survival Analysis for Idiopathic Pulmonary Fibrosis using CT Images and Incomplete Clinical Data. Proceedings of The 5th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 172:1057-1074 Available from https://proceedings.mlr.press/v172/shahin22a.html.

Related Material