BT27 How reliable is teledermatology in diagnosing skin cancer? A retrospective analysis of skin cancer outcomes from a single centre

Abstract Teledermatology has become a pivotal tool to meet the increasing demands on skin cancer services. Factors such as referral quality and clinical judgement need to be considered when evaluating its effectiveness and reliability. A single-centre retrospective study was conducted to evaluate the diagnostic accuracy of a pilot teledermatology service used to triage 2-week-wait skin cancer referrals from primary care. Between May and July 2023 (total 158 referrals), 45% of 2-week-wait teledermatology referrals were accepted into a face-to-face (F2F) clinic, of which 80% were seen in a 2-week-wait clinic, either by dermatology or plastic surgery. The remaining accepted cases were redirected into a general dermatology clinic. Of the 87 referrals returned, 31% were due to poor-quality photographs or insufficient information provided, and 69% were thought to be appropriate for primary care management. A detailed analysis of 100 referrals was conducted to assess the referral quality and clinical outcomes. Of the 51 patients with a suspected skin cancer diagnosed on teledermatology, 21% (5 of 24) of melanomas, 17% (4 of 23) of squamous cell carcinomas and 75% (3 of 4) of basal cell carcinomas were confirmed on histological analysis. Of the 21 patients who were suspected to have a precancerous or benign lesion on teledermatology, two were diagnosed with a malignancy on histology. When comparing the teledermatology diagnosis vs. histological diagnosis as the gold standard, diagnostic accuracy was 0.66 95% confidence interval (CI) 0.51–0.78, with 92% sensitivity, 42% specificity and κ = 0.33 (95% CI 0.11–0.56). This was comparable with the diagnostic accuracy of a F2F clinic review, which was 0.66 (95% CI 0.51–0.79) with 96% sensitivity, 38% specificity and κ = 0.33 (95% CI 0.13–0.54). When only the F2F diagnosis was used as the gold standard, the diagnostic accuracy of teledermatology was higher at 0.78 (95% CI 0.66–0.87), with 91% sensitivity, 56% specificity and κ = 0.5 (95% CI 0.28–0.71). Our study demonstrated that teledermatology can be a useful tool to reduce unnecessary F2F clinic reviews, if adequate information and photographs are provided. It has demonstrated high skin cancer diagnostic sensitivity, but with compromised specificity. There was a moderate level of concordance between diagnoses made in teledermatology and the F2F clinic. Although teledermatology and F2F assessment had very comparable diagnostic accuracies, a higher sensitivity and lower specificity were noted on F2F assessment. This could be attributed to the initial diagnosis made on teledermatology influencing the subsequent clinical decision in a F2F clinic. Ongoing caution, critical reflection and more studies are warranted to elucidate teledermatology’s full potential in the diagnosis and management of skin cancer.