Purpose: investigate the metabolic changes on liver caused by chemotherapy by utilizing F18-FDG PET/CT scan by using Radiomics.
Methods: 83 Patients with lymphoma underwent PET/CT scan before, after treatment and in follow up stage. Hounsfield units (HU) of liver and spleen ,maximum and mean standard uptake value (SUVmax,SUVmean), mean correlation for background, metabolic tumor volume (MTV), total lesion glycolysis (TLG) were elevated. The image undergo Radiomics's analytical system by using (LIFEx v.4) get 4 histogram -intensity features and textural features find by six gray levels co-occurrence matrix (GLCM) features, all were elevated.
Results: 45 patients (54.2%) during treatment had liver morphological changes in HU of liver, HU of spleen from non contrast CT, HU of liver in CT scan with contrast and liver volume. There are functional changes in SUVmax, MTV, and TLG. Changes in MTV and TLG were related to changes in SUVmax and SUVmean. There are also morphological changes proved by first order statics [histogram features (skewness, entropy, and kurtosis)]. By increasing number of cycles of chemotherapy lead to textural changes proved by second order statics [GLCM features (homogeneity, energy, contrast, correlation, entropy and dissimilarity)] were elevated.
Conclusion: most of features from F18-FDG PET/CT provide strong independent predictors for fatty liver changes in patients with lymphoma under chemotherapy. The utility of such features should be confirmed by larger clinical studies before considering their potential integration into decisional algorithms aimed at personalized medicine.