In a groundbreaking study, MedMamba emerges as a potent classifier for medical images, solving the shortcomings of CNNs and Transformers in long-range capability and high computational complexity, respectively.

• The novel Conv-SSM module is at the core of MedMamba, harnessing the strengths of CNN for local feature extraction and SSM for long-range dependencies.
• Extensive experiments on public medical datasets demonstrate MedMamba’s capacity in lesion detection.

MedMamba stands out as the first innovative application of Vision Mamba in the realm of medical classification. Its potential signals a shift towards more efficient SSM-based AI algorithms, influencing future endeavors in medical AI systems. Researchers are encouraged to explore the source code and contribute to this novel baseline in medical image classification.