Medication Recommendation via Dual Molecular Modalities and Multi-Step Enhancement

TL;DR

This paper introduces BiMoRec, a bimodal molecular recommendation framework that integrates 3D and 2D molecular structures with multi-step enhancement and contrastive pretraining, achieving state-of-the-art results in medication recommendation.

Contribution

The paper proposes a novel bimodal graph contrastive pretraining approach combined with multi-step enhancement for improved medication recommendation using 3D molecular structures.

Findings

Achieves state-of-the-art performance on MIMIC datasets.

Effectively fuses 2D and 3D molecular information.

Enhances molecular representations through multi-step recalibration.

Abstract

Existing works based on molecular knowledge neglect the 3D geometric structure of molecules and fail to learn the high-dimensional information of medications, leading to structural confusion. Additionally, it does not extract key substructures from a single patient visit, resulting in the failure to identify medication molecules suitable for the current patient visit. To address the above limitations, we propose a bimodal molecular recommendation framework named BiMoRec, which introduces 3D molecular structures to obtain atomic 3D coordinates and edge indices, overcoming the inherent lack of high-dimensional molecular information in 2D molecular structures. To retain the fast training and prediction efficiency of the recommendation system, we use bimodal graph contrastive pretraining to maximize the mutual information between the two molecular modalities, achieving the fusion of 2D and 3D molecular graphs. Additionally, we designed a molecular multi-step enhancement mechanism to re-calibrate the molecular weights. Specifically, we employ a pre-training method that captures both 2D and 3D molecular structure representations, along with substructure representations, and leverages contrastive learning to extract mutual information. We then use the pre-trained encoder to generate molecular representations, enhancing them through a three-step process: intra-visit, molecular per-visit, and latest-visit. Finally, we apply temporal information aggregation to generate the final medication combinations. Our implementation on the MIMIC-III and MIMIC-IV datasets demonstrates that our method achieves state-of-the-art performance.