Histogram analysis of apparent diffusion coefficient maps in the prognosis of patients with locally advanced head and neck squamous cell carcinoma: Comparison of different region of interest selection methods.

Objective: To investigate the influence of different region of interest (ROI) selection methods on the histogram analysis of apparent diffusion coefficient (ADC) maps and to compare their performance in predicting overall survival (OS) in patients with locally advanced head and neck squamous cell carcinoma (HNSCC).

Methods: A total of 73 patients with locally advanced HNSCC who underwent pretreatment diffusion-weighted magnetic resonance imaging were included. Based on the largest slice ROI (ROI) and whole tumor ROI (ROI), ADC histogram parameters including mean ADC (ADC); median ADC (ADC); 10th, 25th, 75th, and 90th percentiles of ADC values (ADC, ADC, ADC, and ADC); kurtosis; and skewness were obtained. Intraclass correlation coefficients (ICCs) and Bland-Altman plots were used to evaluate measurement reproducibility. The association of ADC histogram parameters and clinicopathological factors with OS was analyzed using log-rank tests and Cox regression.

Results: The measurements of ADC histogram parameters based on ROI showed better reproducibility than ROI (ICCs for ROI: 0.772-0.961; ICCs for ROI: 0.511-0.851). The higher ADC values (ADC, ADC, ADC, and ADC based on both ROIs; ADC75 based on ROI) and lower kurtosis based on ROI were significantly associated with worse OS of patients with locally advanced HNSCC (all P < 0.05). In the multivariate Cox analysis, ADC measured with ROI (P = 0.019, hazard ratio = 2.63, 95% confidence interval 1.17-5.90) was an independent prognostic factor after adjusting for clinicopathological factors.

Conclusions: ROI selection methods could influence ADC histogram analysis. ADC based on ROI had the best independent prognostic value for patients with locally advanced HNSCC.