OBJECTIVE: This study aimed to analyze the risk of lymph node metastases (LNM) in three different VSCC subgroups (HPV-positive (HPVpos), HPV-negative p53 wildtype (HPVneg/p53wt) and HPV-negative p53 abnormal (HPVneg/p53abn)), and develop a predictive model for clinical use. METHODS: A retrospective cohort study was performed, collecting data from all surgically treated VSCC patients between 2000 and 2022 from two oncology clinics. The primary outcome was the risk of groin LNM at diagnosis. Prognostic variables for LNM were identified using uni- and multivariate analyses. A model was created to estimate the probability of LNM at diagnosis. RESULTS: A total of 516 patients were included, of which 94 (18.2 %) were HPVpos, 117 (22.7 %) HPVneg/p53wt, and 305 (59.1 %) HPVneg/p53abn. LNM rates were 17.0 %, 26.5 %, and 35.1 %, respectively (p = .002). Molecular subgroup remained a significant predictor of LNM after adjusting for age, tumor size, and depth of invasion (p = .028). A model using these variables was developed to predict LNM at diagnosis. CONCLUSION: HPVneg/p53abn VSCCs have a higher risk of LNM compared to HPVpos VSCCs. HPVneg/p53wt VSCC are considered an intermediate risk group. Molecular subgroups contribute to LNM risk assessment at diagnosis. We developed a well-performing, clinically feasible model to predict the risk of LNM at diagnosis.