A CROSS-MODAL skin imaging system may soon transform how dermatologists diagnose skin diseases, offering near-histologic accuracy without the need for a biopsy. In this study, researchers evaluated whether this non-invasive imaging method could identify microscopic skin features that traditionally require tissue sampling and histopathologic analysis.
Integrating Skin Imaging with Histologic Analysis
The diagnostic study enrolled 65 adults scheduled for routine skin biopsies at two U.S. dermatology clinics between October 2022 and August 2023. Each participant underwent cross-modal skin imaging, a technique combining multiple imaging modalities, prior to biopsy. The resulting images were then compared to conventional haematoxylin and eosin (H&E)–stained slides to determine whether the technology could replicate key histologic details.
Expert comparative readers used the training portion of the data set to validate specific tissue features on cross-modal images and train a group of blinded physician readers. In subsequent testing, the blinded readers demonstrated remarkable accuracy, correctly identifying primary histologic features 96.4% of the time and secondary features 98.5% of the time. Interrater agreement was also extremely high, with Fleiss κ values exceeding 0.9 for both regional and feature-based assessments. Importantly, no adverse events were reported, underscoring the procedure’s safety.
The results indicate that cross-modal imaging can reliably visualise skin structures in vivo that correspond closely with traditional histopathologic findings. According to the authors, the findings support the technology’s recently cleared role by the U.S. Food and Drug Administration as an aid to clinical judgment, potentially allowing dermatologists to assess lesions noninvasively while preserving diagnostic accuracy.
If adopted widely, the technology could streamline dermatologic diagnostics, reduce the need for invasive biopsies, and enhance patient comfort, particularly for those requiring serial monitoring of skin lesions. Future research will likely explore its role in differentiating benign from malignant conditions and integrating image-based analysis into routine clinical workflows.
Reference
Arron ST et al. Cross-modal imaging in non-invasive identification of histologic features of skin. JAMA Dermatol. 2025; doi:10.1001/jamadermatol.2025.4318.
