Prognostic recurrence analysis method for non-small cell lung cancer based on CT imaging

In order to assist doctors in planning postoperative treatment and re-examination of patients with non-small cell lung cancer, this study proposed a prognostic recurrence analysis method for non-small cell lung cancer based on CT imaging features, aiming to use multiple CT image features to predict the prognosis recurrence of non-small cell lung cancer. Firstly, the lung tumor area was segmented and features were extracted. Secondly, the extracted feature data was optimized for removing redundant features. Then, the optimized feature data and the patient's prognosis were taken as input, the data was trained using a machine learning method, and a predictive analysis model was constructed to predict the prognosis of the non-small cell patient. Finally, experiments were designed to verify the performance of the prognostic recurrence analysis model. A total of 157 patients with non-small cell lung cancer were enrolled in the study. The experimental results showed that the predictive accuracy of the prognostic recurrence model of random forest classifier based on CT imagery grayscale, shape and texture is as high as 84.7%, which can effectively assist doctors to make more accurate prognosis for patients with non-small cell lung cancer. This model can help doctors choose treatment and review methods to prolong the patient's survival.