The field of medical image segmentation and diagnosis is rapidly advancing with the development of innovative methods that leverage foundation models and test-time adaptation. Recent research has focused on improving the accuracy and efficiency of medical image segmentation, particularly in scenarios where annotated data is scarce or unavailable. The use of vision-language models and multimodal adaptation frameworks has shown promise in bridging the gap between general-purpose models and medical image diagnosis. Noteworthy papers include AutoMiSeg, which introduced a zero-shot and automatic segmentation pipeline, and MedBridge, which proposed a lightweight multimodal adaptation framework for accurate medical image diagnosis. Additionally, the development of incremental relationship-guided segmentation and robust annotation-free wound segmentation methods has demonstrated significant advancements in digital pathology and automated wound assessment.