The prediction of pulmonary nodules being canceration is of vital importance to the diagnosis and treatment of early lungadenocarcinoma. However, most traditional research paradigms only focus on image data at a single point in time, whichmakes it easy to ignore the correlation of lung computer tomography (CT) at multiple points in time. Based on the the correlationbetween longitudinal images and medical characteristics of patients with pulmonary nodules, a multimodal feature analysisframework is proposed in this paper.The multimodal analysis framework systematically evaluates the risk of cancer progressionin patients by incorporating diverse sources of information, including radiological features, depth features, and relevant riskfactors within lung cancer imaging. The experiment result show that the accuracy of nodule canceration prediction reached89.15%, which proved the effectiveness of the method proposed in this study. This comprehensive framework holds greatpromise in contributing to more precise diagnostics and personalized treatment strategies in the context of lung cancer.