Can diffusion-weighted magnetic resonance imaging predict tumor recurrence of uterine cervical cancer after concurrent chemoradiotherapy?

Purpose: To retrospectively investigate the utility of diffusion-weighted imaging (DWI) for predicting clinical outcome after concurrent chemoradiotherapy (CCRT) in uterine cervical cancer. Materials and methods: Seventy-four consecutive patients with biopsy-proven cervical cancer who received CCRT underwent DWI at 3T. All patients had MR examinations before therapy (preTx) and at 4 weeks of initiating therapy (midTx). At each point, ADC (apparent diffusion coefficient) was measured in the tumors and ADC change between preTx and midTx were also calculated. For predicting tumor recurrence, MR variables and clinical variables were evaluated and the results were compared. Results: During a mean follow-up of 32.1 months, tumor recurrence developed in 15 (20%) patients: local recurrence (n = 7), distant metastasis (n = 5), and both (n = 3). MidTx tumor ADCs and tumor ADC changes between preTx and midTx were significantly different between the recurrence and non-recurrence groups (P < 0.05), while preTx tumor ADCs were not significantly different between the groups (P = 0.892). Univariate analysis revealed that histologic type, stage, preTx tumor size and volume, and tumor ADC change were significantly related to tumor recurrence (all P < 0.05). However, on multivariate analysis, tumor ADC changes [hazard ratio (HR) 0.886; 95% confidence interval (CI) 0.836–0.940; P = 0.001] and histological type (HR 6.063; 95% CI 1.404–26.187; P = 0.016) were the significant independent predictors of tumor recurrence. Conclusion: Tumor ADC changes between preTx and midTx might be a useful biomarker for the prediction of cervical cancer recurrence after CCRT.