Application of PET/CT image under convolutional neural network model in postoperative pneumonia virus infection monitoring of patients with non-small cell lung cancer

It was to study the adoption of positron emission computed tomography (PET-CT) based on the convolutional neural networks (CNN) model in the monitoring of postoperative pneumonia virus infection in patients with non-small cell lung cancer (NSCLC). 120 patients with NSCLC were set as the research object. CNN model was constructed and applied to PET-CT images to identify lesions and screen tumor markers for detection. Then, the patients were randomly divided into group A (CT), group B (PET-CT), group C (PET-CT based on artificial neural network model), and group D (PET-CT diagnosis based on CNN model), 30 cases in each group, and infection surveillance was conducted. The result showed that the accuracy (Acc), sensitivity (Sen), and specificity (Spe) of PET-CT image recognition based on the CNN model were 99.31%, 100%, and 98.31%, respectively. The proportion of serum neutrophils, white blood cell count, and PCT content in group D three days after operation were significantly lower than those in groups B, C, and A (P < 0.05). The proportions of patients with surgical wound infection and lung infection in group D were 6.54% and 15.38% respectively, which were significantly lower than those in groups B, C, and A (P < 0.05). The complication rates of patients in groups A, B, C, and D were 32.4%, 30.2%, 28.75, and 8.7%, respectively. The complication rate of patients in group D was significantly lower than that of the other three groups (P < 0.05). In short, PET-CT images based on the CNN model had high accuracy, sensitivity, and specificity in monitoring postoperative pneumonia virus infection in NSCLC patients. Applying it to the patient’s virus infection monitoring can effectively prevent the patient’s lung and surgical wound infection and improve the patient’s postoperative recovery effect.