Diagnostic performance of dermoscopy in cutaneous tumors: A retrospective analysis

Background Misdiagnosis of cutaneous tumors leads to inappropriate management, morbidity or mortality. Highly sensitive and specific diagnostic tools are needed. Dermoscopy evaluates surface of tumors rapidly and noninvasively but is limited to depth of dermal papillae. Histopathology is gold standard diagnostic of cutaneous tumors, but is invasive and time-consuming. This study evaluated diagnostic performance of clinical examination and dermoscopy compared to histopathology in determining diagnosis and nature of cutaneous tumors. Methods This retrospective study included cutaneous tumor patients examined clinically, dermoscopically and histopathologically at Dermatooncology and Dermatosurgery Division, Dermatology and Venereology Outpatient Clinic, Dr. Soetomo General Academic Hospital, Surabaya in 2019-2020. Clinical, dermoscopic and histopathological diagnosis were obtained from medical record. Clinical diagnosis was established by board-certified dermatologists. Dermoscopic diagnoses were established from revised two-step algorithm. Histopathological diagnoses were established by board-certified pathologists. Clinical, dermoscopic and histopathologic nature were determined from respective diagnoses. Concordance, sensitivity and specificity of clinical examination and dermoscopy were calculated with histopathology as gold standard examination. Results There were 27 subjects. Ten subjects had malignant tumors including basal cell carcinoma (7 subjects), squamous cell carcinoma, Bowen’s disease, and Kaposi’s sarcoma (1 subject each). Seventeen subjects had benign tumors including seborrheic keratosis (4 subjects), verruca vulgaris and lymphangioma (2 subjects each), solar lentigo, melanocytic nevi, pyogenic granuloma, hemangioma, pilomatrixoma, sebaceous gland hyperplasia, steatocystoma, neurofibroma and fibroepithelial polyp (1 subject each). Clinical and histopathological diagnosis showed moderate concordance (Cohen’s Kappa (K)=0.447). Dermoscopic and histopathological diagnosis showed fair concordance (K=0.346). Clinical examination showed sensitivity 70.0%, specificity 70.6%, and fair concordance (K=0.390) while dermoscopy showed sensitivity 100.0%, specificity 82.4%, and substantial concordance (K=0.776) compared with histopathology in determining malignant nature of tumors. Conclusion Dermoscopy is a valuable tool to support clinical examination, but cannot replace clinical or histopathologic examination in determining nature and diagnosis of cutaneous tumors.