Purpose: To evaluate the value of skull-base abnormality on MRI for predicting local recurrence in nasopharyngeal carcinoma.
Materials and methods: Between November 1988 and February 1997, 48 patients with NPC were examined with both MRI (1.5 T) and CT prior to radiation therapy. T classification (1987 UICC) based on physical examination and CT findings were T1 in 3 cases, T2 in 22, T3 in 9, and T4 in 14. On MRI, low-intensity tissue with Gd enhancement in the marrow of the skull was considered to be a suspicious finding of skull-base invasion. CT simulation was performed in all patients. The total dose to the primary tumor was 60-75 Gy (mean, 67 Gy). The mean follow-up period was 42 months.
Results: All 14 T4 patients had abnormal tissue in the marrow of the skull base on MRI. Thirty-eight percent (13 of 34) of T1-3 patients were suspected to have skull-base invasion based on MRI (0% for T1, 27% [6 of 22] for T2, and 78% [7 of 9] for T3). The 5-year local control rate was significantly different between T1-3 and T4 tumors (97% vs. 69%, p < 0.025) but was not different by the presence of the MRI abnormality in the skull base.
Conclusion: Skull-base invasion suspected solely by MRI does not relate to local recurrence provided that careful treatment planning is performed with the aid of MRI and CT simulator.