Radiation Therapy for Nasopharyngeal Carcinoma: Predictability of Treatment Outcomes at Interim Re-assessment

Abstract

As most prognostic factors of nasopharyngeal carcinoma (NPC) are based on pretreatment characteristics, individual radiosensitivity differences are not considered. We perform interim target volume re-assessments with CT and/or MRI at the end of elective nodal irradiation (ENI). We retrospectively investigated whether interim lymph node (LN) re-assessment is predictive of the treatment outcome in NPC patients. Between February 1996 and June 2007 we performed radiation therapy in 43 NCP patients. At pretreatment assessment, 9 and 34 patients were recorded as N0 and N+, respectively. Thirty-four patients received combination chemotherapy. ENI was with a median dose of 39.6 Gy. Thereafter, boost irradiation was delivered based on the interim re-assessment results; the primary tumor was irradiated with a median dose of 24 Gy, the cervical LN of 27 N+ patients with a median dose of 24 Gy. Planned neck dissection (ND) was performed in 7 patients. The patients with regional failure or residual carcinoma in the LN specimen were categorized as uncontrolled, and the remaining patients were categorized as controlled. The short axis of the largest cervical LN was measured by CT or MRI at pretreatment assessment and interim re-assessment, thereafter, receiver operating characteristic (ROC) analysis was used to compare the predictability of regional control using the LN size at both assessments. Median follow-up was 61 months (range 6-176 months); 60% of all patients had an overall survival (OS) of 5 years. Residual carcinoma in the LN specimens was confirmed in 4 patients with ND. Among 15 patients with treatment failure it was local in 6, regional in 1, and distant in 8. Of the 43 NCP, 5 were categorized as uncontrolled and 38 as controlled. The area under the ROC curve (AUC) at pretreatment assessment and interim re-assessment was 0.825 and 1.000, respectively. The optimal cut-off was 16 mm (Youden index = 0.6071) and 15 mm (Youden index = 1.0000) for the two assessments, respectively. None of the 16 patients without boost irradiation to the LN manifested regional failure; their LN measured 5-10 mm at interim re-assessment. All LN in controlled NPC were ≤ 15 mm at interim re-assessment, in uncontrolled NPC they measured 17-26 mm. An LN size ≤ 15 mm at interim re-assessment was a significant factor for 5-year OS (65% vs 20%, p = 0.03) while a pretreatment size of ≤ 16 mm was not (p = 0.78). Interim re-assessment yields better predictability of treatment outcomes in NPC patients than pretreatment assessment. LN measuring ≤10 mm at interim re-assessment need not undergo boost irradiation. Patients whose LN is > 15 mm are at high risk for regional failure and their prognosis is poor.