Abstract: Objective To compare the performance differences between the multidisciplinary team (MDT) model and the conventional diagnostic and treatment model for lung cancer, and to explore a high-quality development pathway for optimizing lung cancer diagnostic and treatment resources. Methods A retrospective analysis was conducted on electronic medical record data of lung cancer patients at Shanghai Chest Hospital from March 2025 to December 2025. Patients were divided into an MDT group and a conventional care group based on whether they were admitted to the integrated oncology ward. Statistical analyses were performed using the Mann-Whitney U test, trend chi-square test, and chi-square test. Subgroup analyses were conducted according to tumor stage (Ⅰ-Ⅳ) and relative weight (RW) risk stratification. Gamma regression and negative binomial regression models with robust standard errors clustered by patient ID were used to analyze the effects of MDT on cost per hospitalization and hospitalization frequency during the observation period. To adjust for differences in observation time, the logarithm of observation days was used as an offset in the negative binomial regression. Heterogeneity across tumor stage and RW subgroups was assessed using interaction models. To enhance the robustness of causal inference, propensity score matching (PSM) and inverse probability of treatment weighting (IPTW) analyses were performed as sensitivity validations. Results A total of 4, 758 patients with primary lung cancer were included, comprising 365 (7.7%) in the MDT group and 4, 393 (92.3%) in the conventional care group. After adjusting for confounding factors, the MDT model significantly reduced hospitalization frequency during the observation period by 48.8% (IRR=0.512, 95% CI:0.463-0.567, P<0.001). After further adjusting for observation days, MDT still significantly reduced hospitalization frequency (IRR=0.834, 95% CI:0.698-0.997, P=0.046), with the effect direction consistent with the main analysis. No statistically significant difference was observed in cost per hospitalization (IRR=0.942, 95% CI:0.865-1.027, P=0.178). Heterogeneity analysis revealed that the cost effect varied by tumor stage:cost increased in stage Ⅰ patients (IRR=2.002), decreased in stage Ⅱ (IRR=0.705) and stage Ⅳ (IRR=0.743) patients, and showed no change in stage III patients. No significant cost effects were observed across RW subgroups. The reduction in hospitalization frequency was significant across all tumor stage and RW subgroups (IRR range:0.450-0.680), with no significant interactions (all P>0.05). Sensitivity analyses confirmed the robustness of the hospitalization frequency effect (PSM:IRR=0.581; IPTW:IRR=0.520). The PSM analysis showed a significant reduction in cost per hospitalization in the MDT group (IRR=0.878, 95% CI:0.778-0.992, P=0.036), while the IPTW analysis was consistent with the main analysis and showed no statistical significance (IRR=0.943, 95% CI:0.862-1.032, P=0.218). Conclusions The MDT model for lung cancer significantly reduces hospitalization frequency; however, its effect on cost per hospitalization is population-selective, with increased costs in early-stage (stage Ⅰ) patients and decreased costs in late-stage (stages Ⅱ and Ⅳ) patients. The implementation of the MDT model should adopt precise patient stratification management, prioritizing the optimal patient population to achieve the optimal allocation of medical resources.