The impact of target dosimetry on patients' locoregional recurrence in nasopharyngeal carcinoma: A propensity score-matched analysis.

PURPOSE:To analyze the impact of target dosimetry using propensity score matching (PSM) on patients' locoregional recurrence for nasopharyngeal carcinoma (NPC) and to find significant dose-volume factors of recurrence. METHODS:Sixty-eight nasopharyngeal carcinoma (NPC) patients with recorded locoregional recurrence were enrolled in this study. These patients were treated with IMRT in 2009-2010 in our department. Another 198 NPC patients without recurrence were randomly selected from the same treatment time period. The median follow-up time for all patients was 49.0 months. PSM was performed to match the recurrence and nonrecurrence cohorts. Dose-volume histograms (DVHs) of treatment planning were extracted for statistical analysis. Cox hazard model and Kaplan-Meier log-rank analysis were performed to evaluate correlations between PTV dose coverage and local/region recurrence. RESULTS:Propensity score matching balanced the clinical factors in two matches. Univariate cox survival model showed D90 and D95 were significantly correlated to the recurrence, and the D90 was the most significant (p = 0.036) one. The results of multivariate analysis show that only D90 is required for recurrence prediction when collinear dosimetric factors are considered. KM log-rank analysis showed that patients have significant local/regional control differences (p-value = 0.036, log-rank) in the D90 >101% and D90 <101% groups. CONCLUSION:D90 corresponds to significant dose-volume factors. PTV dose coverage has a significant impact on locoregional recurrence in NPC clinical routine patients.