The field of decision-focused learning is moving towards more scalable and efficient methods, with a focus on reducing the dependence on expensive solver calls and improving decision alignment. Researchers are exploring new frameworks and techniques, such as dual-guided surrogates and bidirectional feedback, to enhance the performance of predictive models in downstream decision-making problems. Additionally, the concept of digital biomarkers is expanding to include multimodal data, which raises important ethical implications for knowledge, responsibility, and governance in data-driven medicine. Noteworthy papers include: A Dual Perspective on Decision-Focused Learning, which introduces a scalable objective that preserves decision alignment while reducing solver dependence. From Sequential to Recursive: Enhancing Decision-Focused Learning with Bidirectional Feedback, which proposes a novel framework that introduces bidirectional feedback between downstream optimization and upstream prediction.