Between January 2019 and January 2022, 319 patients received immunotherapy in our institution. 94 patients were excluded for without progression, and 69 patients were excluded for lost to follow up in first-line. 10 patients were excluded for with other primary tumors, and 5 patients were excluded for with EGFR mutations in post-test. 141 aNSCLC patients was finally included due to treated with ICIs in second-line after immunotherapy PD in first-line. Then 20 patients were excluded for with <2 cycles IO+ChT treatment in first-line. As a result, a total of 121 patients who received PD-1/PD-L1 inhibitor as the second-line or later therapy were enrolled in the study (Fig. 1). According to the therapeutic modality, there were 53 patients in the IBP group and 68 patients in the non-IBP group. The last follow-up time was October 1, 2022. 48 patients had died by the end of the follow-up, and 73 patients were still alive. The median follow-up time was 5.6 months (range, 0.2– 29.0 months) for surviving patients and 13.1 months (range, 0.2–29.0 months) for all patients. The baseline characteristics of all patients in both groups are shown in Table 1. There were no differences between the two groups in the distribution of most variables except for treatment features. The median age of the patients in the two groups was 62 years, and the age range was 27 to 76 years. Fifty- three (43.8%) patients received IBP treatment, and 68 (56.2%) patients received non-IBP treatment. Eigthy-seven (71.9%) patients have adenocarcinoma pathology, 34 (28.1%) patients have squamous pathology. Thirty-three (28.1%) patients had brain metastases, and 20 (16.5%) patients had liver metastases. Sixty-four (52.9%) patients had never smoked. Fifty-four patients underwent PD- L1 detection: 17 (14.0%) patients had a level <1%, 19 (15.7%) patients had a level from 1 to 49%, and 18 (14.9%) patients had a level ≥ 50% (Table 1). The loss of follow-up rate was 7.4%.

The median PFS1 in the IBP group and non-IBP group in first-line therapy was 7.6 and 9.4 months, respectively. PFS1 was not significantly different between the two groups (p=0.103 Fig. 2). The median OS in the IBP group and non-IBP group was 14.1 and 10.8 months, and not statistically difference (p=0.063 Fig. 3A). The OS rates at 1 year in the IBP group and non-IBP group were 37.7% and 13.2%, respectively. The median PFS was 8.7 months for the patients receiving IBP treatment and 4.1 months for those receiving non-IBP treatment in second-line therapy. The PFS was longer in the IBP group than in the non-IBP group (p < 0.001, Fig. 3B). Partial subgroup analysis of OS was in Fig. 3C.

IBP versus non-IBP in second-line treatment of PFS1 in advanced NSCLC patients. Abbreviations: PFS, progression-free survival

IBP versus non-IBP in second-line treatment of OS (A) and PFS (B) in advanced NSCLC patients. Subgroup analysis of OS (C). Abbreviations: IBP, immunotherapy beyond progression; NSCLC, non-small-cell lung cancer; OS, overall survival

In the subgroup of 109 patients with ≥4 cycles in first-line immunotherapy, the OS and PFS were statistically different between the IBP group and non-IBP group (median OS: 15.4 vs. 10.8 months, respectively, p=0.047 Fig. 4A; median PFS: 8.7 vs. 4.4 months, respectively, p<0.001, Fig. 4B). And in the subgroup of the 90 patients with ≥6 cycles in first-line immunotherapy, the OS and PFS were statistically different between the IBP group and non-IBP group (median OS: 16.1 vs. 10.8 months, respectively, p=0.039, Fig. 4C; median PFS: 8.7 vs. 4.7 months, respectively, p=0.01, Fig. 4D). Then in the subgroup of the 67 patients with ≥8 cycles in first-line immunotherapy, the OS and PFS were statistically different between the IBP group and non-IBP group (median OS: 16.3 vs. 10.9 months, respectively, p=0.029, Fig. 4E; median PFS: 11.9 vs. 5.7 months, respectively, p=0.006, Fig. 4F).

OS (A) and PFS (B) in patients with ≥4 cycles in first-line immunotherapy subgroup of aNSCLC between IBP and non-IBP. OS (C) and PFS (D) in patients with ≥6 cycles in first-line immunotherapy subgroup of aNSCLC between IBP and non-IBP. OS (C) and PFS (D) in patients with ≥8 cycles in first-line immunotherapy subgroup of aNSCLC between IBP and non-IBP

In the subgroup of the 48 patients with PR as best response in first-line immunotherapy, the OS was different between the IBP group and non-IBP group (median OS: 18.9 vs. 10.2 months, respectively, p=0.041, Fig. 5A). The PFS in the IBP group was longer than that in the non-IBP group (median PFS: 11.6 vs. 4.4 months, respectively, p=0.023, Fig. 5B). The OS of the 63 patients with SD as best response in first-line immunotherapy was similar between the IBP group and non-IBP group (median OS: 14.1 vs. 13.5 months, p=0.389, Fig. 5C). The PFS in the IBP group was longer than that in the non-IBP group (median PFS: 7.9 vs. 4.2 months, p=0.004, Fig. 5D).

OS (A) and PFS (B) in patients with PR as best response in first-line immunotherapy subgroup of aNSCLC between IBP and non-IBP. OS (C) and PFS (D) in patients with SD as best response in first-line immunotherapy subgroup of aNSCLC between IBP and non-IBP

The IBP group showed a longer OS (median OS: 16.3 vs. 10.8 months, respectively, p=0.035, Fig. 6A) and PFS (median PFS: 9.4 vs. 4.0 months, respectively, p=0.002, Fig. 6B) than the non-IBP group for oligo progression subgroups. However, the OS (median OS: 13.1 vs. 13.5 months, respectively, p= 0.626, Fig. 6C) were not different between the two groups in the extensive progression subgroups. The PFS was longer in the IBP group than in the non-IBP group (median PFS: 6.7 vs. 4.4 months, respectively, p= 0.014, Fig. 6D).

OS (A) and PFS (B) in patients with oligo progression in first-line immunotherapy subgroup of aNSCLC between IBP and non-IBP. OS (C) and PFS (D) in patients with extensive progrssion in first-line immunotherapy subgroup of aNSCLC between IBP and non-IBP

The clinical characteristics of the patients were evaluated to determine their prognostic value for OS (Table 2). Univariate analysis indicated that age and cycles of first-line immunotherapy were associated with survival. Patients ≤65 (p=0.045) and first-line immunotherapy ≥8 cycles (p=0.046) had a better OS. For PFS, univariate analysis revealed that histology, liver meta and cycles of first-line immunotherapy were significant favorable prognostic factors (Table 3). Multivariate analysis revealed that the liver meta (p= 0.007) was favorable prognostic factors for PFS.

The rate of adverse events(AEs) in two groups was listed in Table 4. The incidence of any grade AEs in IBP group and non-IBP group was 32.1% and 29.4%, respectively. The rate of grade≥3 adverse events was 7.5% in the IBP group, compared to 5.9% in the non-IBP group. No patient died from AEs.