RESUMEN
BACKGROUND: We aimed to develop a new tumor response assessment method for lung SBRT. METHODS: In total, 132 lung cancer patients with 134 tumors who received SBRT treatment with daily CBCT were included in this study. The information about tumor size (area), contrast (contrast-to-noise ratio (CNR)), and density/attenuation (µ) was derived from the CBCT images for the first and the last fractions. The ratios of tumor area, CNR, and µ (RA, RCNR, Rµ) between the last and first fractions were calculated for comparison. The product of the three rations was defined as a new parameter (R) for assessment. Tumor response was independently assessed by a radiologist based on a comprehensive analysis of the CBCT images. RESULTS: R ranged from 0.27 to 1.67 with a mean value of 0.95. Based on the radiologic assessment results, a receiver operation characteristic (ROC) curve with the area under the curve (AUC) of 95% was obtained and the optimal cutoff value (RC) was determined as 1.1. The results based on RC achieved a 94% accuracy, 94% specificity, and 90% sensitivity. CONCLUSION: The results show that R was correlated with early tumor response to lung SBRT and that using R for evaluating tumor response to SBRT would be viable and efficient.