A predictive model for the recurrence of intracranial aneurysms following coil embolization

OBJECTIVE: This study aimed to identify risk factors for intracranial aneurysms (IAs) recurrence and establish a predictive model to aid evaluation. METHODS: A total of 302 patients with 312 IAs undergoing coil embolization between September 2017 and October 2022 were divided into two groups based on digital subtraction angiography follow-up. Clinical characteristics, operation-related factors, and morphologies were measured. Cox proportional hazard regression was used to identify the risk factors. Hazard ratios (HRs) were used to score points, and a predictive model was established. The test cohorts consisted of 51 IAs. Receiver operating characteristic curves were generated to determine the cutoff values and area under the curves (AUCs). A Delong test was performed to compare the AUCs. RESULTS: Diameter maximum (D max) (p < 0.001, HR = 1.221), Raymond-Roy occlusion classification (RROC) II or III (p = 0.004, HR = 2.852), and ruptured status (p < 0.001, HR = 7.782) were independent risk factors for the recurrence of IAs. A predictive model was established: D max + 2 (*) RROC (II or III; yes = 1, no = 0) + 6 (*) ruptured status (yes = 1; no = 0). The AUC of the predictive model (0.818) was significantly higher than those of D max (0.704), RROC (II or III) (0.645), and rupture status (0.683), respectively (Delong test, p < 0.05). The cutoff values of the predictive model and D max were 9.75 points and 6.65 mm, respectively. CONCLUSION: The D max, RROC (II or III), and ruptured status could independently predict the recurrence of IAs after coil embolization. Our model could aid in practical evaluations.