Multimodal fusion of imaging and genomics for lung cancer recurrence prediction
This work addresses recurrence prediction for early-stage lung cancer patients, but it is incremental as it builds on existing multimodal fusion approaches.
The paper tackled predicting lung cancer recurrence by fusing CT images and genomics, achieving up to a 10% improvement in concordance-index using linear models.
Lung cancer has a high rate of recurrence in early-stage patients. Predicting the post-surgical recurrence in lung cancer patients has traditionally been approached using single modality information of genomics or radiology images. We investigate the potential of multimodal fusion for this task. By combining computed tomography (CT) images and genomics, we demonstrate improved prediction of recurrence using linear Cox proportional hazards models with elastic net regularization. We work on a recent non-small cell lung cancer (NSCLC) radiogenomics dataset of 130 patients and observe an increase in concordance-index values of up to 10%. Employing non-linear methods from the neural network literature, such as multi-layer perceptrons and visual-question answering fusion modules, did not improve performance consistently. This indicates the need for larger multimodal datasets and fusion techniques better adapted to this biological setting.