Changes in Apparent Diffusion Coefficient (ADC) in Serial Weekly MRI during Radiotherapy in Patients with Head and Neck Cancer: Results from the PREDICT-HN Study
Sweet Ping Ng, Carlos Cárdenas, Houda Bahig, Baher Elgohari, Jihong Wang, Jason M. Johnson, Amy C. Moreno, Shalin Shah, Adam S. Garden, Jack Phan, G. Brandon Gunn, Steven J. Frank, Yao Ding, Lumine Na, Ying Yuan, Diana L. Urbauer, Abdallah Mohamed, David I. Rosenthal, William H. Morrison, Michael MacManus, Clifton D. Fuller
Abstract
Background: The PREDICT-HN study aimed to systematically assess the kinetics of imaging MR biomarkers during head and neck radiotherapy. Methods: Patients with intact squamous cell carcinoma of the head and neck were enrolled. Pre-, during, and post-treatment MRI were obtained. Serial GTV and ADC measurements were recorded. The correlation between each feature and the GTV was calculated using Spearman’s correlation coefficient. The linear mixed model was used to evaluate the change in GTV over time. Results: A total of 41 patients completed the study. The majority (76%) had oropharyngeal cancer. A total of 36 patients had intact primary tumours that can be assessed on MRI, and 31 patients had nodal disease with 46 nodes assessed. Median primary GTV (GTVp) size was 14.1cc. The rate of GTVp shrinkage was highest between pre-treatment and week 4. Patients with T3-T4 tumours had a 3.8-fold decrease in GTVp compared to T1-T2 tumours. The ADC values correlated with residual GTVp. The median nodal volume (GTVn) was 12.4cc. No clinical features were found to correlate with GTVn reduction. The overall change in ADC for GTVn from pre-treatment was significant for 35th−95th percentiles in weeks 1−4 (p < 0.001). Conclusion: A discrepancy in the trajectory of ADC between primary and nodal sites suggested that they exhibit different treatment responses and should be analysed separately in future studies.