The following is a summary of “Transcriptomic analysis of cutaneous squamous cell carcinoma reveals a multigene prognostic signature associated with metastasis,” published in the December 2023 issue of Dermatology by Wang, et al.

While cutaneous squamous cell carcinoma (cSCC) metastasis is infrequent, current staging methods exhibit sub-optimal predictive performance. For a study, researchers sought to address it by developing a robust gene expression profile signature for predicting the metastatic risk of primary cSCC, utilizing an unbiased whole transcriptome discovery-driven approach.

Archival samples from 237 immunocompetent patients, including primary cSCC with perilesional normal tissue, were retrospectively collected from four centers. The entire transcriptome was probed using TempO-seq, and machine learning algorithms were employed for deriving predictive signatures. The dataset was split into training and testing sets at a ratio of 3:1.

A 20-gene prognostic model was successfully developed and validated. The testing set demonstrated an accuracy of 86.0%, sensitivity of 85.7%, specificity of 86.1%, and a positive predictive value of 78.3%. Importantly, this outperformed traditional pathological staging systems, providing a more stable and accurate prediction of metastatic risk. A linear predictor was also established, showing a significant correlation with metastatic risk. The study was limited by its retrospective nature and included data from four centers. Larger prospective multicenter studies were deemed necessary to validate the findings further.

The developed 20-gene signature presented an accurate prediction model for cSCC metastatic risk. Its potential incorporation into clinical workflows could enhance the precision of metastasis prediction in primary cSCC.

Source: jaad.org/article/S0190-9622(23)02504-5/fulltext