Self-Supervised Learning and Multimodal Foundation Models for AI-Driven Medical Imaging

Dear Colleagues,

In recent years, the unprecedented growth of data and the rapid progress of deep learning have revolutionized diverse domains, from natural language processing to computer vision and bioinformatics.

In the medical field, imaging plays a central role in diagnosis, prognosis, and treatment planning. Deep learning has already demonstrated its potential to enhance diagnostic accuracy, improve efficiency, and uncover novel biomarkers. Yet, significant challenges remain: the scarcity of labeled data, the cost of expert annotations, and the inherent variability across patients, institutions, and imaging modalities.

Self-supervised learning and transfer learning have emerged as powerful strategies to overcome these barriers by harnessing large-scale unlabeled datasets. The advent of foundation models has shown how pretraining at scale, exemplified by CLIP, EVA, or DINOv3, can produce generalizable representations that accelerate progress across tasks. Translating this paradigm to medicine has the potential to fundamentally reshape imaging research and clinical practice.

Crucially, medical imaging is inherently multimodal and vision-centric: radiology, pathology, ophthalmology, endoscopy, dermatology, and microscopy each provide complementary perspectives on human health. Moreover, integrating imaging with non-imaging data such as genomics, clinical notes, or laboratory values can unlock richer and more holistic representations. The emergence of multimodal foundation models for healthcare represents a transformative opportunity to move beyond siloed approaches toward unified models of disease.

This Special Issue seeks to catalyze this transformation by gathering cutting-edge research on self-supervised learning, transfer learning, and foundation models for medical imaging, with a strong emphasis on multimodal and vision-centric innovations. We invite contributions that push the boundaries of methodology, benchmark performance, and demonstrate real-world clinical impact.

Topics of interest include, but are not limited to, the following:

• Novel self-supervised learning algorithms for medical imaging;
• Transfer learning strategies for foundation models across imaging modalities and tasks;
• Foundation models for unimodal and multimodal medical imaging;
• Vision-centric multimodal fusion of imaging and non-imaging data by leveraging foundation models;
• Large-scale pretraining pipelines and benchmarks for medical imaging;
• Robustness, interpretability, fairness, and regulatory considerations of foundation models in medical AI;
• Future challenges and opportunities for foundation models in healthcare.

We welcome contributions from researchers, clinicians, and industry experts working at the intersection of medical imaging, machine learning, and artificial intelligence. Together, this collection will chart the path toward the next generation of scalable, multimodal foundation models for the integration of artificial intelligence techniques in medicine and healthcare.

Dr. Andrea Loddo
Dr. Lorenzo Putzu
Dr. Cecilia Di Ruberto
Dr. Carsten Marr
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Journal of Imaging is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1800 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• medical imaging
• deep learning
• foundation models
• self-supervised learning
• multimodal fusion
• transfer learning
• clinical AI
• medical image segmentation
• multimodal LLM
• medical image detection