Assessing Response to Chemoradiotherapy on 18F-FDG PET Imag- es in Non-Small Cell Lung Cancer Using New Approaches of Histo- gram and Gray Level Co-Occurrence Matrix

Purpose: To investigate inherent information provided by 18F-FDG PET to ameliorate shortcomings of relying on visual inspection or sole SUV measurement in treatment assessing. Patients and methods: Twelve patients with newly diagnosed NSCLC and treated with combined Chemoradiotherapy (CRT) were involved in this study. We analyzed the percentage variation of gray value in every gray level or on the whole using histogram analysis algorithm which represents global intensity distribution. We also investigated texture parameters which describe local intensity-spatial distribution and were calculated by Gray Level Co-occurrence Matrix (GLCM). The parameters’ variation analysis involved comparison between patient’s PET scans of pretreatment and 1 month after treatment completion and correlation between characteristics variation and response degree were analyzed. The texture comparison between the same patient’s lesions on one side lung and the corresponding tissue with same size, location on the other side normal lung was involved. Results: The uniformity degree of gray level distribution on the whole and the maximum ratio decrease were well associated with tumor shrinkage and response degree. In that case, they were capable to differentiate tumor response to CRT. Texture parameters’ variation characterizing local tumor metabolism was able to differentiate the response if these parameters were taken as indices because they showed great correlation with regional response to CRT. Conclusion: We demonstrated that histogram and texture analysis methods on baseline 18F-FDG PET scans have showed more robust, discriminative in assessing response to combined CRT and may have a good application prospect in clinical practice.