Policy-Driven, Multimodal Deep Learning for Predicting Visual Fields from the Optic Disc and OCT Imaging.

ABSTRACT

PURPOSE: To develop and validate a deep learning (DL) system for predicting each point on visual fields (VFs) from disc and OCT imaging and derive a structure-function mapping. DESIGN: Retrospective, cross-sectional database study. PARTICIPANTS: A total of 6437 patients undergoing routine care for glaucoma in 3 clinical sites in the United Kingdom. METHODS: OCT and infrared reflectance (IR) optic disc imaging were paired with the closest VF within 7 days. EfficientNet B2 was used to train 2 single-modality DL models to predict each of the 52 sensitivity points on the 24-2 VF pattern. A policy DL model was designed and trained to fuse the 2 model predictions. MAIN OUTCOME MEASURES: Pointwise mean absolute error (PMAE). RESULTS: A total of 5078 imaging scans to VF pairs were used as a held-out test set to measure the final performance. The improvement in PMAE with the policy model was 0.485 (0.438, 0.533) decibels (dB) compared with the IR image of the disc alone and 0.060 (0.047, 0.073) dB with to the OCT alone. The improvement with the policy fusion model was statistically significant (P < 0.0001). Occlusion masking shows that the DL models learned the correct structure-function mapping in a data-driven, feature agnostic fashion. CONCLUSIONS: The multimodal, policy DL model performed the best; it provided explainable maps of its confidence in fusing data from single modalities and provides a pathway for probing the structure-function relationship in glaucoma.