Integration of flow cytometry and computational analysis to dissect the epidermal cellular subsets in keloids that correlate with recurrence

Abstract Keloid disease is a benign skin disease without good therapy. More and more researches show epidermal abnormalities participate in keloid pathogenesis. Little is known about the relationship between the abnormal epidermal immunophenotype and clinical outcome. Nine-color flow cytometry with computational analysis was performed to detect the altered cellular subpopulation distribution in lesions. Receiver operating characteristic (ROC) curves were drawn to compare predictive ability between alteration of cell subgroup frequency and the Vancouver scar scale (VSS). The frequency of CD49fhi/CD29+/TLR7+ cellular subsets increased in the keloid epidermis compared with the healthy control. The CD49fmid-hi/CD29+/TLR7+/CD24+ cellular subpopulation level was increased significantly in keloids, while the CD49flo-mid/CD29-/TLR7-/CD24- cellular subpopulation frequency was decreased. The CD49flo/CD29-/TLR7-/CD24+/CD117+ cellular subpopulation showed an increased frequency during recurrence with a sensitivity of 66.7% and specificity of 91.7%. The area under the curve (AUC) was 0.806 for cellular subpopulation analysis, which was higher than the AUC for the VSS (0.583). The alteration of keloid epidermal subpopulation frequency is related with recurrence, which will provide an optional predictive marker for keloid recurrence and a potential target subset for investigating keloid generation.