Background: The Draf 3 procedure has been demonstrated to be effective in the treatment of refractory frontal sinus disease. A variety of clinical factors may contribute to the change of frontal neo-ostium (FNO) area after this procedure. Imaging plays a vital role in the evaluation and follow-up after surgery and provides useful prognostic information. Objectives: (1) To investigate the influence of local anatomic factors on FNO after a Draf 3 procedure by radiological measurements on imaging software and (2) to explore other predictive factors of FNO restenosis. Methods: Twenty-four patients with chronic rhinosinusitis who underwent a Draf 3 procedure and were followed up for more than 12 months (2012-2014) were enrolled in this study. Data on patient demographics, medical history, and computed tomography scans were collected. Anatomic dimensions were measured with OsiriX (R) (Pixmeo, Geneva, Switzerland). Stenotic ostium was defined as a loss of more than 50% of the original intraoperative area. Multivariate linear regression was used to assess independent factors linked to frontal neo-osteogenesis 1 year after the surgery. A receiver operating characteristic curve was built for the cutoff value of preoperative dimension to predict restenosis of FNO area. Results: A significant association was demonstrated between the minimum anteroposterior diameter (FOAP(MIN)) of the frontal ostium preoperatively and FNO area 1 year postoperatively. Multivariate linear regression showed that FOAP(MIN) of the frontal ostium preoperatively and the percentage of serum eosinophils (EOSs) correlated with the contour area of the FNO 1 year postoperatively. FOAP(MIN) <= 3.592 mm yielded a sensitivity of 95% and a specificity of 75% for the diagnosis of FNO restenosis. Conclusions: Both FOAP(MIN) and EOS are independent predictors of the contour area of the FNO 1 year postoperatively. Patients with a higher risk of developing frontal ostium restenosis can be identified preoperatively by measurements of the frontal sinus anatomic dimensions.