Probabilistic approach to longitudinal response prediction: application to radiomics from brain cancer imaging

TL;DR

This paper introduces a probabilistic model for predicting disease progression over time using radiomic features from brain cancer imaging, effectively handling uncertainty and reducing reliance on intermediate follow-up data.

Contribution

The study presents a novel probabilistic approach that integrates baseline and follow-up data for longitudinal prediction, outperforming existing models in uncertainty management and dimensionality control.

Findings

Model is competitive with state-of-the-art methods.

Effectively accounts for uncertainty in predictions.

Reduces need for intermediate follow-up data.

Abstract

Longitudinal imaging analysis tracks disease progression and treatment response over time, providing dynamic insights into treatment efficacy and disease evolution. Radiomic features extracted from medical imaging can support the study of disease progression and facilitate longitudinal prediction of clinical outcomes. This study presents a probabilistic model for longitudinal response prediction, integrating baseline features with intermediate follow-ups. The probabilistic nature of the model naturally allows to handle the instrinsic uncertainty of the longitudinal prediction of disease progression. We evaluate the proposed model against state-of-the-art disease progression models in both a synthetic scenario and using a brain cancer dataset. Results demonstrate that the approach is competitive against existing methods while uniquely accounting for uncertainty and controlling the growth of problem dimensionality, eliminating the need for data from intermediate follow-ups.