ISIC-REPO; ISIC Skin Imaging Repository Enhancements for Promoting Interoperability and Utilization

SUMMARY The International Skin Imaging Collaboration (ISIC)'s overarching goal is to leverage technology to reduce skin cancer mortality and morbidity. The ISIC Archive, created and led by Memorial Sloan Kettering Cancer Center since 2016, is the leading publicly available skin imaging repository. With >240 000 images, >6000 registered

users, and 17-21 million image downloads per month, the ISIC Archive has made major contributions to dermatology teaching and the development of artificial intelligence (AI) for skin cancer diagnosis. ISIC has also been central to the development and adoption of standards for skin imaging to promote quality and

interoperability. ISIC has hosted 5 machine learning Grand Challenges and hosts continuous Live Challenges using ISIC Archive images. ISIC Challenges have attracted 6818 participants and 25 326 submissions. ISIC has directly published 21 scientific papers, while another 260 peer reviewed papers and 3060 technical reports

have relied on ISIC data. Here, we propose to develop the ISIC Archive into an even greater resource for research through major enhancements that focus on making the ISIC Archive fully compliant with FAIR principles (findability, accessibility, interoperability, and reusability), while expanding the Archive with additional

support for broadly used data formats (eg, Digital Communications Standards in Medicine [DICOM]) and skin imaging modalities (clinical imaging, total body photography [TBP]). We propose to leverage ISIC's leadership in the DICOM community to improve software support for established skin imaging standards, while publishing

new standards for TBP, and to improve quality and interoperability for these imaging modalities. We have organized key milestones for this proposal under 4 specific aims: (1) to bring the ISIC Archive into rigorous compliance with FAIR principles; (2) to overcome the challenges of ingesting and sharing clinical images and

representations of TBP while maintaining patient privacy and addressing sensitivity; (3) to strengthen the interoperability of ISIC data through the use of DICOM; and (4) to implement approaches to facilitate scaling, disseminating, and sustaining the ISIC data repository. The resulting ISIC Archive will be a more scientifically

robust, impactful, and sustainable resource. ISIC will continue to promote use of the Archive and engage the dermatology and computer science communities through Challenges, workshops, and publications. By promoting and supporting research to effectively leverage skin imaging and AI to diagnose skin cancer earlier

and more accurately, we will save lives and reduce the morbidity associated with unnecessary biopsies and the treatment of advanced disease.