Background: Atezolizumab is superior to docetaxel for patients with advanced non-small-cell lung cancer (NSCLC) who are pretreated with platinum-based chemotherapy based on the POPLAR and OAK trials. However, patients who received prior immunotherapy (IO) were excluded from these trials. The preferred subsequent therapy for these patients remains to be docetaxel +/- ramucirumab. The efficacy and safety of atezolizumab as a subsequent therapy in IO-pretreated patients are unknown. Methods: We conducted a retrospective study of all patients with locally advanced/metastatic NSCLC who were pretreated with IO at Mayo Clinic Florida/Rochester in 2016 - 2022. Patients who received subsequent therapy of atezolizumab alone (Atezo), docetaxel (Doce), or docetaxel + ramucirumab (Doce+Ram) were included. Kaplan-Meier method and log-rank test were used for survival analysis. Results: Among 169 patients included, 51.5% were female, 91.7% were Caucasian. The median age was 67 (35 – 92). The histology was predominantly adenocarcinoma (76.3%). The most common metastatic sites were bone (51.5%), followed by CNS (44.4%). The median number of prior therapies was 2 (1 – 8). Pembrolizumab was the most used prior IO (87%), followed by durvalumab (10.7%) and nivolumab (8.3%). 13.6% (n = 23), 49.1% (n = 83), and 37.2% (n = 63) patients received subsequent Atezo, Doce, and Doce+Ram, respectively. PFS was improved in the Atezo group compared with Doce group (median PFS 3.9 vs. 3.7 months, HR 0.57, 95% CI 0.37 – 0.87, p = 0.015*). Notably, a prominent percentage of patients in the Atezo group appear to have long-term PFS benefits, demonstrated by 1-year PFS rates of 26.1%, 3.6%, and 4.8% (p = 0.003**), and 2-year PFS rates of 17.4%, 0%, and 0% (p = 0.002**), in the Atezo, Doce, and Doce+Ram groups, respectively. The median OS (months) in the Atezo, Doce, and Doce+Ram groups were 17.5, 7.6, and 11.5, respectively; and were 29.5, 7.2, and 6.9, respectively in the PD-L1 > 50% subgroups. The Atezo group showed significantly greater OS improvement compared with the Doce group (HR 0.42, 95% CI 0.26 – 0.66, p = 0.001**). This benefit appears greater and remains statistically significant in the PD-L1 > 1% (HR 0.39, 95% CI 0.22 – 0.71, p = 0.005**) and PD-L1 > 50% (HR 0.37, 95% CI 0.15 – 0.91, p = 0.022*) subgroups. 5 of 23 (21.7%) patients in the Atezo group developed immune-related adverse events. 3 patients were grade 3 or 4. No grade 5 event. Conclusions: We observed statistically significant and clinically meaningful PFS and OS benefits of atezolizumab monotherapy compared with docetaxel in patients with advanced NSCLC who were pretreated with IO. The OS benefits of atezolizumab over docetaxel +/- ramucirumab was greater in PD-L1 > 1% and PD-L1 > 50% subgroups. Atezolizumab appears to be tolerable in IO-pretreated patients. These findings need to be verified in a prospective study.