Clone Selection Artificial Intelligence Algorithm-Based Positron Emission Tomography-Computed Tomography Image Information Data Analysis for the Qualitative Diagnosis of Serous Cavity Effusion in Patients with Malignant Tumors

Abstract

This study aimed to investigate the application of positron emission tomography- (PET-) computed tomography (CT) image information data combined with serous cavity effusion based on clone selection artificial intelligence algorithm in the diagnosis of patients with malignant tumors. A total of 97 patients with PET-CT scanning and empirically confirmed as serous cavity effusion were retrospectively analyzed in this study. The clone selection artificial intelligence algorithm was applied to register the PET-CT images, and the patients were rolled into a benign effusion group and a malignant effusion group according to the benign and malignant conditions of the serous cavity effusion. Besides, the causes of patients from the two groups were analyzed, and there was a comparison of their physiological conditions. Subsequently, CT values of different KeV, lipid/water, water/iodine, and water/calcium concentrations were measured, and the differences of the above quantitative parameters between benign and malignant serous cavity effusion were compared, as well as the registration results of the clone algorithm. The results showed that the registration time and misalignment times of clonal selection algorithm (13.88, 0) were lower than those of genetic algorithm (18.72, 8). There were marked differences in CT values of 40–60 keV and 130–140 keV between the two groups. The concentrations of lipid/water, water/iodine, and water/calcium in basal substances of the malignant effusion group were obviously higher than the concentrations of the benign effusion group ( $P<0.05$ ). Benign and malignant effusions presented different manifestations in PET-CT, which was conducive to the further diagnosis of malignant tumors. Based on clone selection artificial intelligence algorithm, PET-CT could provide a new multiparameter method for the identification of benign and malignant serous cavity effusions and benign and malignant tumors.