3D total body photography, a promising innovation for early skin cancer detection: narrative review (Preprint)

BACKGROUND Skin cancer is a global health concern due to its high and still increasing incidence and associated healthcare cost. Belgium is no exception as one in five people are diagnosed with skin cancer before the age of 75. A promising innovation, the VECTRA WB360, a three-dimensional total body photography system allows clinicians to objectively compare the totality of the skin on a macroscopic level on further appointments. The integrated lesion visualisation software allows automated detection, counts and assessment of skin lesions. Detailed comparison of individual lesions is possible through the attached digital dermatoscope. OBJECTIVE This study aims to review available literature on the use of the Vectra in research and clinical settings, and to summarise the clinical utility, advantages and limitations reported for this system. METHODS An electronic literature search was conducted on PubMed using a combination of following search terms: 3D imaging, VECTRA WB360, melanoma, non-melanoma skin cancer, their synonyms and associated entry terms. RESULTS Our literature search yielded 11 relevant papers. According to multiple studies, the VECTRA WB360 images were of a high enough quality to allow on-screen diagnosis of melanoma and nonmelanoma skin cancers by dermatologists. The integrated lesion visualisation software is capable of detecting and counting naevi and distinguishing melanoma from other skin lesions with high accuracy. Integrating a convolutional neural network enhances both the sensitivity and specificity of the software. However, dermatologists achieved greater specificity and thus remained superior to machine and artificial intelligence. CONCLUSIONS At this time the Vectra 3D TBP provided high enough image quality to detect NMSC by a clinician remotely examining the images without the addition of dermoscopy. The longitudinal comparation of 3D TBP images has aided the detection of melanomas in research studies. The CNN for naevi classification has high accuracy and the CNN for lesion malignancy risk (DEXI) reported higher accuracy than the one used for the 2D TBP system. Despite these promising results, expert overview is still recommended, and AI should be used as a support tool.