ABSTRACT
BACKGROUND: Current staging systems for cutaneous squamous cell carcinoma (cSCC) have limited positive predictive value for identifying patients who will experience metastasis. OBJECTIVE: To develop and validate a gene expression profile (GEP) test for predicting risk for metastasis in localized, high-risk cSCC with the goal of improving risk-directed patient management. METHODS: Archival formalin-fixed paraffin-embedded primary cSCC tissue and clinicopathologic data (n = 586) were collected from 23 independent centers in a prospectively designed study. A GEP signature was developed using a discovery cohort (n = 202) and validated in a separate, nonoverlapping, independent cohort (n = 324). RESULTS: A prognostic 40-GEP test was developed and validated, stratifying patients with high-risk cSCC into classes based on metastasis risk: class 1 (low risk), class 2A (high risk), and class 2B (highest risk). For the validation cohort, 3-year metastasis-free survival rates were 91.4%, 80.6%, and 44.0%, respectively. A positive predictive value of 60% was achieved for the highest-risk group (class 2B), an improvement over staging systems, and negative predictive value, sensitivity, and specificity were comparable to staging systems. LIMITATIONS: Potential understaging of cases could affect metastasis rate accuracy. CONCLUSION: The 40-GEP test is an independent predictor of metastatic risk that can complement current staging systems for patients with high-risk cSCC.
Subject(s)
Biomarkers, Tumor/genetics , Carcinoma, Squamous Cell/secondary , Gene Expression Profiling/methods , Skin Neoplasms/pathology , Adult , Aged , Aged, 80 and over , Carcinoma, Squamous Cell/diagnosis , Carcinoma, Squamous Cell/epidemiology , Carcinoma, Squamous Cell/genetics , Female , Gene Expression Regulation, Neoplastic , Humans , Kaplan-Meier Estimate , Male , Middle Aged , Neoplasm Staging/methods , Predictive Value of Tests , Prognosis , Prospective Studies , Risk Assessment/methods , Skin/pathology , Skin Neoplasms/diagnosis , Skin Neoplasms/epidemiology , Skin Neoplasms/genetics , Survival RateABSTRACT
BACKGROUND: The heterogeneous behavior of patients with melanoma makes prognostication challenging. To address this, a gene expression profile (GEP) test to predict metastatic risk was previously developed. This study evaluates the GEP's prognostic accuracy in an independent cohort of cutaneous melanoma patients. METHODS: This multi-center study analyzed primary melanoma tumors from 523 patients, using the GEP to classify patients as Class 1 (low risk) and Class 2 (high risk). Molecular classification was correlated to clinical outcome and assessed along with AJCC v7 staging criteria. Primary endpoints were recurrence-free (RFS) and distant metastasis-free (DMFS) survival. RESULTS: The 5-year RFS rates for Class 1 and Class 2 were 88% and 52%, respectively, and DMFS rates were 93% versus 60%, respectively (P < 0.001). The GEP was a significant predictor of RFS and DMFS in univariate analysis (hazard ratio [HR] = 5.4 and 6.6, respectively, P < 0.001 for each), along with Breslow thickness, ulceration, mitotic rate, and sentinel lymph node (SLN) status (P < 0.001 for each). GEP, tumor thickness and SLN status were significant predictors of RFS and DMFS in a multivariate model that also included ulceration and mitotic rate (RFS HR = 2.1, 1.2, and 2.5, respectively, P < 0.001 for each; and DMFS HR = 2.7, 1.3 and 3.0, respectively, P < 0.01 for each). CONCLUSIONS: The GEP test is an objective predictor of metastatic risk and provides additional independent prognostic information to traditional staging to help estimate an individual's risk for recurrence. The assay identified 70% of stage I and II patients who ultimately developed distant metastasis. Its role in consideration of patients for adjuvant therapy should be examined prospectively.