Factors associated with recurrence and nomogram of OKC treated with curettage or marsupialization to predict early recurrence: a retrospective study

Predicting early recurrence in odontogenic keratocyst (OKC) is difficult due to multiple recurrence factors and limited high-quality evidence. This study aimed to identify factors associated with recurrence and establish a nomogram for patients undergoing conservative surgery (curettage or marsupialization). We retrospectively enrolled patients with OKC who underwent conservative surgical treatment at the Stomatology Hospital, Zhejiang University School of Medicine (from September 2017 to December 2021), with follow-up until December 2022. Demographic, surgical, imaging, and pathological features were analysed. Significant variables were entered into multivariable Cox and logistic regression to identify independent predictors, and nomograms were developed. One hundred fifteen patients were included (overall recurrence rate 26.1%). Involvement of the maxilla and mandible (HR = 8.427, 95% CI: 3.031–23.434, P < 0.001) and elevated absolute neutrophil count (ANC) (HR = 4.773, 95% CI: 1.914–11.905, P = 0.001) were significantly associated with a shorter recurrence-free period (RFP). The concordance index (C‑index) of the Cox‑based nomogram was 0.660 (95% CI: 0.552–0.767), suggesting modest discriminative ability for predicting recurrence‑free probability. Involvement of both right and left jaws (OR = 8.882, 95% CI: 1.936–40.735, P = 0.005) and cyst volume > 4.27 cm 3 (OR = 4.290, 95% CI: 1.623–11.341, P = 0.003) were independent predictors for recurrence. The nomogram demonstrated moderate discrimination for OKC recurrence (AUC = 0.725; internal validation AUC = 0.772). In the primary Cox analysis, ANC and site of onset (maxilla or mandible) are independent predictors of recurrence-free probability in OKC patients undergoing conservative surgery. In the complementary logistic analysis, cyst volume, along with site of onset (left or right jaw), showed significant independent associations with recurrence status. The proposed nomogram may serve as a supportive tool for individualized risk presentation and may facilitate more targeted postoperative follow-up care. However, external validation in larger, prospective cohorts is required before clinical application.