Purpose/Objective(s): Optimal treatment for nonesmall cell lung cancer (NSCLC), depending on patients' clinical stage, performance status, and preferences, may include surgical resection, radiation therapy, chemotherapy (in various combinations), or no treatment. We evaluated variation in stage-stratified use of treatment modalities in and out of a multidisciplinary (MultiD) program in a large community-based healthcare system. We hypothesized that MultiD care will optimize modality use. Materials/Methods: 2014-2015 NSCLC Tumor Registry data from 4 hospitals within 1 tri-state healthcare system retrospectively reviewed; MultiD data prospectively collected from a weekly MultiD tumor conference. MultiD data were analyzed separately from the tumor registry data. Any MultiD patients present in the tumor registry data were removed. Staging was based on information available before commencement of any treatment modality (defined as the actual treatment received). Patients were clustered into 3 stage groupings: I-II, III, and IV. Stage-stratified frequencies of treatment modalities across the 4 hospitals and the MultiD program were compared using Chi-squared tests. Results: 1,614 NSCLC patients were seen outside and 248 within the MultiD program. Demographics were similar except MultiD had a higher proportion of African American (26% v 33%, PZ.01) and Medicaid/ uninsured (8.3% v 22%) and fewer commercially insured (55% v 40%) patients (P<.001). Stage distributions for non-MultiD versus MultiD were: 29% versus 46% (stage I/II); 20% versus 25% (stage III); 52% versus 29% (stage IV). There were striking differences in use of modalities for non-MultiD versus MultiD: surgery for stage I/II disease 48% versus 63% and no treatment, 16% versus 6% (PZ.01); multimodality therapy for stage III, 62% versus 75%, and no treatment 19% versus 10% (PZ.02); chemotherapy for stage IV, 57% versus 70% stage IV, radiation only, 12% versus 13%, and no treatment 29% versus 9% (P<.001). Patient demographics were similar across the 4 hospitals, but modality use varied widely. For stage I/II, surgical resection rates ranged from 24%-62% and no treatment from 7%-42%; for stage III, multimodality treatment ranged from 40%-81%, and no treatment from 13%-36%; for stage IV, chemotherapy rates ranged from 47%-59%, and no treatment from 25%-43% (P<.001 for all stages). Conclusion: Despite less favorable patient demographics, stage-stratified care within a community-based MultiD Program was more directed towards active and multimodality care. Great heterogeneity in pattern of care exists across hospitals within the same healthcare system. Further research into the causes of stage-stratified treatment variation, and the potential use of structured MultiD Programs to improve access to care is warranted.