PURPOSE: To evaluate the toxicities and survival outcomes associated with carbon ion radiation therapy (CIRT) using pencil beam scanning (PBS) technique and to assess the prognostic factors for patients with stage III non-small cell lung cancer (NSCLC) using a local effect model (LEM)-based biological dose calculation. METHODS AND MATERIALS: This analysis included patients with stage III NSCLC (n = 181) who received CIRT between December 2016 and June 2023. CIRT was administered at a relative biological effectiveness-weighted dose of 77 Gy (range, 69-83.6 Gy) in 22 fractions (Fx) (range, 19-24 Fx). Most patients (96.1%) underwent systemic therapy before and/or after CIRT. Toxicities and survival outcomes were recorded, and statistical analyses conducted. RESULTS: The median follow-up period was 18.2 months. Grade 1, 2, 3, and 4 acute toxicities were observed in 62.4%, 30.4%, 2.8%, and 0.6% of patients, respectively, with hematological toxicities accounting for all grade ≥ 3 acute toxicities. Grade 1, 2, 3, and 4 late toxicities occurred in 40.3%, 30.9%, 4.4%, and 1.7% of patients, respectively, with most grade ≥ 3 CIRT-induced late toxicities (72.7%) observed in patients receiving a CIRT dose ≥ 79.2 Gy. The median overall survival (OS) and progression-free survival (PFS) were 37.1 and 16.7 months, respectively. The 2-year locoregional control, OS, PFS, and distant metastasis-free survival rates were 66.1%, 64.2%, 40.3%, and 49.5%, respectively. Patients who received a CIRT dose ≤ 77 Gy had better OS (P = .047), but worse locoregional control compared with those who received higher doses (P = .026). Post-CIRT immunotherapy was an independent prognostic factor for improved OS, distant metastasis-free survival, and PFS (P = .002, P <
.002, and P <
.002, respectively). CONCLUSIONS: CIRT with pencil beam scanning was effective for locally advanced NSCLC, resulting in acceptable toxicities and promising OS outcomes, particularly with doses of 69 to 77 Gy and post-CIRT immunotherapy.