The PASS (polychromy, asymmetry, specific structures) algorithm for dermoscopy had good diagnostic accuracy and was easier to perform than other dermoscopic algorithms, according to the results of a multicenter retrospective analysis published in the International Journal of Dermatology.
Dermoscopy has been shown to be a useful and non-invasive diagnostic test for cutaneous melanoma, but its accuracy depends on observer experience.
Investigators from centers in Spain sought to develop a novel algorithm that was easy to use while maintaining diagnostic accuracy. To that end, 1120 dermoscopic images collected at 3 centers were pooled and used to develop the PASS algorithm, which awarded 1 point each for the presence of polychromy (3 or more colors), asymmetry of color or structure in 1 or 2 axes, and 1 melanoma-specific dermoscopic structure; and 2 points for 2 or more melanoma-specific structures.
Continue Reading
The images comprised melanocytic nevi without histological atypia (51%), melanomas (28.6%), melanocytic nevi with histological atypia (10.8%), Spitz nevi (5.3%), and blue nevi (4.3%).
Overall, 800 images were of benign melanocytic tumors and 320 were of melanomas. More of the benign tumors were scored as PASS 0 (46.6%), 1 (20.4%), or 2 (20%), whereas more of the melanomas were scored as PASS 3 (23.4%) or 4 (68.4%).
For diagnosing melanomas, a PASS 4 score had a sensitivity of 68.4%, specificity of 96.4%, positive predictive value of 88.3%, negative predictive value of 88.4%, and accuracy of 88.4%.
The interobserver agreement between 3 independent investigators showed moderate agreement in PASS outcomes (k, 0.621).
The major limitation of this technique is that dermoscopy itself depends on user experience.
These data indicated that the PASS algorithm had a good diagnostic accuracy and seemed to be easier to perform than other algorithms. These findings should be confirmed in additional studies.
Reference
Avilés-Izquierdo JA, García-Piqueras P, Ciudad-Blanco C, et al. Do not PASS any melanoma without diagnosis: a new simplified dermoscopic algorithm. Int J Dermatol. Published online January 20, 2023. doi:10.1111/ijd.16573
