Implementation of Decision Tree Models in Differentiating Hard-to-Diagnose Lung Nodules

We have studied the importance of various lung characteristics obtained through computed tomography (CT), in conjunction with other factors, for distinguishing between lung pathologies such as hamartoma (HM) and tuber-culoma (TB). We have developed a model based on the decision tree algorithm trained on clinical data. This enabled us to identify key characteristics and their interactions, as well as to define the main clinical factors that precisely characterize HM and TB states. The model demonstrated that the radiology factor characterizing changes in the surrounding tissue of the pathological nodule has the most contribution.