Integrating multi-omics information in deep learning architectures for joint actuarial outcome prediction in non-small-cell lung cancer patients after radiation therapy.

Researchers

Journal

Modalities

Models

Abstract

Novel actuarial deep learning neural network (ADNN) architectures are proposed for joint prediction of radiotherapy outcomes, i.e., radiation pneumonitis (RP) and local control (LC) in stage III non-small-cell lung cancer (NSCLC) patients. Unlike NTCP/TCP models that use dosimetric information solely, our proposed models consider complex interactions among multi-omics information including PET radiomics, cytokines and miRNAs. Additional time-to-event information is also utilized in the actuarial prediction.
Three architectures were investigated: ADNN-DVH considered dosimetric information only; ADNN-com integrated multi-omics information; ADNN-com-joint combined RP2 (RP grade greater or equal to 2) and LC prediction. In these architectures, differential dose volume histograms (DVHs) were fed into 1D convolutional neural networks (CNNs) for extracting reduced representations. Variational encoders (VAEs) were used to learn representations of imaging and biological data. Reduced representations were fed into Surv-Nets to predict time-to-event probabilities for RP2 and LC independently and jointly by incorporating time information into designated loss functions.
Models were evaluated on 117 retrospective patients and were independently tested on 25 newly accrued patients prospectively. A multi-institutional RTOG0617 dataset of 327 patients was used for external validation. ADNN-DVH yielded cross-validated c-indexes [95% CIs] of 0.660 [0.630 -0.690] for RP2 prediction and 0.727 [0.700-0.753] for LC prediction, outperforming a generalized Lyman model for RP2 (0.613 [0.583-0.643]) and a generalized log-logistic model for LC (0.569 [0.545-0.594]). The independent internal test and external validation yielded similar results. ADNN-com achieved an even better performance than ADNN-DVH on both cross-validation (CV) and independent internal test. Furthermore, ADNN-com-joint which yielded similar performance as ADNN-com realized joint prediction with c-indexes of 0.705 [0.676-0.734] for RP2, 0.740 [0.714-0.765] for LC and achieved an AU-FROC 0.729 [0.697-0.773] for the joint prediction of RP2 and LC.
Novel deep learning architectures that integrate multi-omics information outperformed traditional TCP/NTCP models in actuarial prediction of RP2 and LC.
Copyright © 2021. Published by Elsevier Inc.