Extracting medical decisions from discharge summaries is essential for downstream clinical analytics, yet the task remains challenging due to the heterogeneous structure of electronic health records. For the MedExACT track at ACL 2026, we proposed a system that achieved the 4th position. Our approach first applies dynamic section conditioning to capture the contextual dependencies inherent in each document. A transformer backbone is then augmented with category- and section-aware layer mixing, enabling us to fuse global document structure with fine-grained semantic cues. To further improve robustness, we employ an ensemble of instruction-tuned large language models for automatic section extraction, while a fairness-oriented model selection criterion ensures that performance does not degrade on minority demographic subgroups. The resulting system attains a final score of 0.5806 on the held-out test set and demonstrates significant gains over the baseline across all evaluated subpopulations.

This paper presents our system at the Radiology Report Summarization Shared Task-1B of the 22nd BioNLP Workshop 2023. Inspired by the work of the BioBART model, we continuously pre-trained a general domain BART model with biomedical data to adapt it to this specific domain. In the pre-training phase, several pre-training tasks are aggregated to inject linguistic knowledge and increase the abstractivity of the generated summaries. We present the results of our models, and also, we have carried out an additional study on the lengths of the generated summaries, which has provided us with interesting information.