Hand eczema (HE) is a highly prevalent chronic disease affecting 15% of the population. Clinical grading systems are time-intensive and require training, making overall coarse assessment the default approach in practice. We present an image-based deep learning method to automatically evaluate HE lesion surface and determine their anatomical repartition. This retrospective study was based on two datasets: (A) 312 HE pictures and (B) 215 hand pictures. Eleven dermatologists annotated HE lesions in (A) and a student labeled (B) pictures with 37 hand anatomical sub-regions. Separate segmentation deep learning models (DLM) were trained for A and B. Predictions from both DLMs were merged and used to generate textual reports of patient conditions. DLM performance was evaluated using precision and sensitivity with 95% confidence interval. Intra-class correlation of predicted lesion surface with experts’ annotations was also evaluated. The HE DLM achieved a precision and sensitivity of respectively 75% (64-82) and 69% (55-81), while the anatomy DLM achieved in average 83% (80-85) and 85% (82-88) for the same. The lesion surface intra-class correlation was 0.94 (0.90-0.96). An example of textual report is: “Both hands show eczema on the palmar side, namely on 51% of the fingertips, 34% of the fingers, 27% of the palms and 4% of the wrists.” Our approach can automatically generate a precise anatomical stratification of HE lesions. It is reproducible and has no intra- and inter-observer variability. The method could be performed remotely to enable frequent follow-up and teledermatology treatment.

ESDR143 - Automated Anatomical Mapping of Hand Eczema

The scientific meeting of the European Society for Dermatological Research (ESDR) is held yearly, usually in September. With an international attendance of more than 1000 skin scientists from around the globe, the four-day event includes short oral presentations from peer-reviewed abstracts, keynote lectures from international experts, sponsored symposia, satellite programs, educational events as well as an important focus on posters, networking and building collaborations amongst scientists. For more information about this year’s meeting visit our website.