Treatment response evaluation using texture analysis on 18-fluorodeoxyglucose positron- emission tomography/computed tomography in patients with locally advanced breast cancer

The objectives of this study were to (1) utilize the ability of texture analysis to evaluate the treatment response in patients with locally advanced breast cancer (LABC) and overall survival (OS) and (2) to compare the result with PET/ CT parameters. We retrospectively analyzed 57 patients with histopathologically proven LABC. 18-fluorodeoxyglucose positron-emission tomography/computed tomography (¹⁸F FDG PET/CT) scan was performed at the baseline and after 6-12 cycles of chemotherapy. Thirteen haralick texture features were derived from the baseline scan. An independent nuclear medicine physician evaluated the ¹⁸F FDG PET/CT parameters in the baseline scan. Mann Whitney U test & independent T-test were performed to determine the significance of the difference in the mean value of texture parameters & PET/ CT parameters. Cox- regression & Kaplan- Meier tests were applied for survival analysis. Mean age of patients was (46.07 ± 1.46 years). Based upon the clinical follow-up, out of 57 patients, 39 patients were categorized as responders & 18 patients as non - responders. Out of the thirteen texture parameters, two of the texture parameters: sum average (V6) with p = 0.071 & sum variance (V7) with p= 0.069 were found to approach a statistically significant level. Responders had higher values of TLG AND MTV than non-responders. Kaplan-Meier test showed the trend for the OS of the patients with the higher value of MTV & TLG observed with patients who died than in the patients who survived. No texture parameters & PET parameters were significant to differentiate between responders and non-responders and also didn't significant in differentiating between the two survival groups (died and alive). We concluded that two of the texture parameters: sum average (V6) and sum variance (V7) that were approaching significant may have role in predicting treatment response if studied in a larger sample size with a more homogenous patient population. (author)