Effectiveness Of Esophageal Gastric Junction Tumor Motion With And Without A Pneumatic Abdominal Compression In Precise IGRT Era

Abstract

In the study 140 patients with II-III stage were included. After long course of chemoradiation and further surgery the pathology examination showed pCR in 38 (27.1%) of the patients. From every patients' CTsimulation image, the tumor volumes extracted 850 features (119,000 in total number). KS test showed statistically difference between pCR-groups and non-pCR. Logistic regression showed high value of wavelet contribution to model. Multilayer perceptron model showed high predictive importance of wavelet in the model. Applying TwoStep cluster analysis for classifying the texture and wavelet parameters three cluster were defined with predominant wavelet features importance. Conclusion: Current study reports, for the first time, the feasibility of using non-diagnostic CT images as data source for texture analysis in combination with wavelets features analysis for predicting pCR in locally advanced rectal cancer patients. The model performance shows the importance of including wavelets features in radiomics analysis.