Aim:
To determine the diagnostic accuracy of total body photography (TBP) for the detection of melanoma in adults.
Methods:
We searched MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials from May 2020 to June 2026 to update the systematic review by Ji Xu et al. 2021. Searches were imported into Covidence and two reviewers independently screened titles, abstracts, and full texts to establish eligibility. Summary estimates for the proportion of biopsies positive for melanoma were calculated with a random effects model meta-analysis. Risk of bias was assessed using a modified version of QUADAS-2.
Results:
Eighteen studies (20 publications) were included, comprising eight new studies. Most studies were observational, including nine retrospective studies and eight prospective studies, and one randomised clinical trial. Fourteen studies reported data on 2D-TBP, while four studies reported data on 3D-TBP. No studies used artificial intelligence to diagnose lesions. Risk of bias was generally judged as low for the index test domain but high or unclear for participant selection, reference standard, and flow/timing domains. Applicability concerns were low for most studies. The pooled proportion of biopsies positive for melanoma for 2D-TBP was 13% (95% confidence interval (CI) 9%–20%) and 10% (95% CI 3%–31%) for 3D-TBP. The number needed to biopsy to diagnose a melanoma ranged between 2.3 – 19.6 for 2D-TBP and 2.7 – 37.7 for 3D-TBP, and the positive predictive value ranged between 5 – 43% for 2D-TBP and 3 – 37% for 3D-TBP. One study reported sufficient information to allow the calculation of sensitivity and specificity for TBP, with 3D-TBP achieving 100% sensitivity and 93% specificity compared with histopathology for diagnosis of melanoma (including invasive and in situ).
Conclusion:
Estimation of diagnostic accuracy for TBP is limited by insufficient reporting of sensitivity and specificity data, and variability in TBP workflows.