This research project aims to develop computerized imaging features of FDG PET-CT in these patients which receive surgical intervention for suspected lung cancer, determine the likelihood of malignancy corresponding to individual imaging features and expect to distinguish benign lesions (tuberculosis infection and inflammation) from malignant tumors. The PET/CT scanning needs several minutes and dozens of the respiratory cycles to obtain images. The physiological effects of respiratory movement and heart motion can cause imaging blurring especially for lung images. The standard chest CT scans for diagnosis are performed after the subjects hold their breath at maximum inspiration. The CT images of PET/CT are usually obtained when the subjects breathe freely and are used for attenuation correction and anatomic positioning. We compare the image features of lung lesions, such as attenuation, contour and margin and morphology, between the two sets of CT images and evaluate if the CT image of PET/CT are adequate for interpretation and diagnosis of these lung lesions or not. Fused PET/CT images offer functional and anatomic information. We computerize the CT image features and standard uptake value (SUV) of PET/CT for these lung lesions to evaluate the likelihood of malignancy of each parameter according to the pathological diagnosis. The computerized image features can provide supplementary information for clinical physicians to make accurate diagnosis.
