The most common driver gene mutation of non-small cell lung cancer (NSCLC) in East Asians is epidermal growth factor receptor (EGFR) mutation (1). EGFR receptor tyrosine kinase inhibitors (TKIs) are preferred for EGFR-sensitive mutations in clinical setting; however, most patients experience problems with acquired resistance and disease progression after receiving EGFR-TKIs for approximately 10 to 18 months (2, 3). The use of standard platinum-based dual-drug chemotherapy as subsequent therapy had a limited effect; therefore, more effective treatment strategies should be investigated for patients who progress on EGFR-TKIs therapy.

Immune checkpoint inhibitors (ICI), including cytotoxic T lymphocyte-associated antigens-4 (CTLA-4) inhibitors, programmed cell death 1 (PD-1) inhibitors and programmed cell death ligand 1 (PD-L1) inhibitors, counteract the immunosuppressive effect of tumors and reactivate the immune response of T cells to inhibit tumor cell growth (4). Several studies have demonstrated the efficacy of ICI therapy in advanced NSCLC (5). Immune checkpoint inhibitors such as the PD-1 inhibitor Nivolumab, Pembrolizumab, and the PD-L1 inhibitor Atezolizumab, and the CTLA-4 inhibitor Ipilimumab have been approved for the treatment of advanced NSCLC. However, first-line immunotherapy has poor efficacy for NSCLC with EGFR mutations (6). Clinical studies have found that the tumor immune microenvironment is altered after receiving EGFR-TKIs targeted therapy, including an increase in CD8⁺ tumor-infiltrating lymphocytes (TILs) density, TMB, and PD-L1 expression in tumor cells (7, 8). This finding suggests the possibility of ICI monotherapy or ICI-based combination therapy for this population. Results from the IMpower150 and ORIENT31 studies further support the administration of ICI-based combination therapy in this setting (9, 10). However, the phase 3 clinical trials CheckMate-722 and KEYNOTE-789 comparing immunotherapy combined with chemotherapy with chemotherapy alone did not show significant results (11, 12). There remains controversy regarding subsequent therapeutic strategies for NSCLC patients harboring EGFR mutations previously treated with EGFR-TKIs. Therefore, our study retrospectively reviewed patients with advanced NSCLC after EGFR-TKIs failure in Shanghai Chest Hospital and analyzed clinical outcomes, safety and relevant influential factors of different treatment options.

The standard first-line treatment for patients with advanced EGFR-mutated NSCLC are EGFR-TKIs. There are limited treatment options for patients who are refractory to third-generation EGFR-TKIs and for T790M-negative patients who have received first- or second-generation EGFR-TKIs. The most commonly recommended treatment is platinum-containing doublet chemotherapy with or without bevacizumab. In vitro experiments confirmed that the co-culture system of EGFR-mutant tumor cells with immune cells was able to reduce the viability of tumor cells after treatment with PD-1 inhibitors (13). The anti-tumor effect of PD-1 inhibitors has also been demonstrated in preclinical models with EGFR mutations (14). These preclinical studies suggested the possibility of treating EGFR-mutant NSCLC with ICI. However, the initial clinical findings did not support this. Several prospective studies have shown no survival benefit from the use of immunosuppressive agents versus chemotherapy in NSCLC patients with EGFR mutations (15–18). Moreover, most clinical trials of immune checkpoint suppression have also excluded NSCLC patients with driver gene mutations. Therefore, there is an urgent need to develop immune-based combination therapies for NSCLC patients with EGFR mutations.

Our study basically included a variety of mainstream and non-mainstream treatment regimens after TKIs resistance. Analysis of these regimens revealed that the mPFS was longest in the ICA group, followed by the CA, IC, IA, and CM group, and the shortest in the IM group. The ICA group had the highest ORR and DCR, followed by the IC, CA, IA, and CM group, and the lowest was in the IM group. Further differential analysis showed that in terms of mPFS, there was no significant difference between the ICA, CA, and IC group for pairwise comparisons. The same results were obtained from the IC, IA, CM, and IM group for pairwise comparisons. There were many interesting conclusions to draw from these analyses.

Combination regimens were superior to monotherapy. In terms of PFS, the ICA group was longer than the CM or IM group in our study. In terms of the ORR and DCR, the ICA group was better than the IM group. The prospective study ORIENT31 included 444 patients with EGFR mutations after EGFR-TKIs failure, randomly assigned to sintilimab plus IBI305 plus pemetrexed and cisplatin (ICA), sintilimab plus pemetrexed and cisplatin (IC), or pemetrexed and cisplatin groups (CM). The mPFS of the ICA group was 6.9 months and the mPFS for the CM group was 4.3 months in ORIENT31. The difference in PFS between the two groups was statistically significant (P<0.0001) (9). The mPFS in both groups was consistent with the results of our study. Otherwise, our study analysis showed a longer mPFS in the CA group than in the CM or IM group. According to a prospective study, the mPFS of patients with advanced NSCLC received bevacizumab plus chemotherapy was 6.2 months and chemotherapy alone was 4.5 months, and the difference was statistically significant (19). The results were also consistent with the results of our analysis. However, one study showed that chemotherapy combined with anti-angiogenesis was not superior to chemotherapy alone in patients with advanced non-small cell lung cancer (20). Many studies have confirmed the poor efficacy of ICI monotherapy in patients with EGFR-mutant NSCLC. A Japanese randomized controlled study included 102 patients with EGFR-mutant NSCLC after TKIs resistance, with a mPFS of 1.7 months in the nivolumab group and 5.6 months in the carboplatin-pemetrexed group and differences were significant (21). Therefore, CA efficacy was better than IM, which was in line with our expectations.

A subsequent question has to be raised: which regimen has better efficacy comparing ICA with CA? The prospective study Impower150 included 124 chemotherapy-naive NSCLC patients with EGFR mutations, of which 91 patients with EGFR sensitive mutations were assigned to three different regimens: atezolizumab plus bevacizumab plus carboplatin plus paclitaxel (ABCP), bevacizumab plus carboplatin plus paclitaxel (BCP), and atezolizumab plus carboplatin plus paclitaxel (ACP). The results showed that in the 124 EGFR mutation subgroup, the HR for ABCP versus BCP was 0.61 (95% CI 0.36-1.03). The mPFS was 10.2 months for the ABCP group, 6.9 months for the BCP group. With respect to the mOS, the ABCP group also improved compared to the BCP group (10). In conclusion, ICA tended to be better than CA, which was consistent with our results. Given that IMpower150 is a small-scale subgroup analysis, it cannot be concluded that ICA was superior to the CA regimen, and the results need to be confirmed by more prospective studies. The mPFS from each treatment group obtained from the IMpower150 study had a gap to compare with our results, and it needed to be considered that we were based on real-world research; most patients received overline therapy, immunotherapy for these patients was a very posterior treatment, and our study showed that the line of therapy on the efficacy of immunotherapy is very obvious. In addition, IMpower150 enrolled patients before osimertinib was approved as first-line treatment.

In addition, pairwise combination regimens of anti-angiogenesis, immunotherapy, and chemotherapy included the CA, IC, and IA group. For the CA or IA group, anti-angiogenesis combination chemotherapy or immunotherapy was feasible. According to previous studies, there was a synergistic effect between chemotherapeutic agents and antiangiogenic agents. Elevated VEGF levels could lead to tumor vascular disorders, increased permeability and interstitial pressure, and affect the delivery of chemotherapeutic agents into the tumor (22). Bevacizumab could promote the delivery of chemotherapeutic agents into the tumor (23). On the other hand, there was also synergy between antiangiogenic agents and immune checkpoint inhibitors, and antiangiogenic agents showed immunomodulatory effects (24), could improve the tumor microenvironment for immunosuppression in patients with EGFR mutations. Our results indicate that the CA group had a significant longer mPFS than the IA group. However, the ORR and DCR were similar between the two groups. For the CA and IC group, combining immunotherapy or anti-angiogenesis based on chemotherapy, the former mPFS was longer than the latter, but there was no significant difference. In the IMpower150 study, in the EGFR mutation subgroup, the HR for PFS with ACP versus BCP was 1.14 (95% CI 0.73-1.18), mPFS 6.9 months in the ACP group, and mPFS 6.9 months in the BCP group. However, in the 56 patients who had previously received EGFR-TKIs, the mPFS of the BCP group was 6.1 months, which was slightly longer than that of the ACP group (mPFS=5.7 months) (10). The results of the subgroup analysis of IMpower150 were similar to our conclusions. Another real-world retrospective study from Shanghai Pulmonary Hospital achieved similar conclusions: the mPFS was 6.90 months in the CA group and 7.59 months in the IC group, and there was no significant difference in PFS (25). However, another retrospective study showed that OS was worse in the IC group comparing with the CA group (HR 2.37, 95%CI 1.09-5.65, P=0.030) (26). Similarly, comparing IC with IA regimens, based on immunotherapy combined with chemotherapy or anti-angiogenesis, the former mPFS was longer than the latter, no significant difference was observed in PFS.

Although combination therapy was superior to monotherapy, some exceptions existed. For example, the IC group had a longer mPFS than the CM and IM group, but there was no significant difference in PFS. CheckMate-722 and KEYNOTE-789 study also showed no survival benefit between the IC and CM group (11, 12). However, the results of the latest ORIENT31 second interim analysis showed that mPFS was 5.5 months in the IC group and 4.3 months in the CM group, which was similar with our results. But the difference between these two groups in ORIENT31 trail was statistically significant (P=0.016) (27). In addition, the mPFS of the IA group was longer than the CM and IM group, and no significant difference was observed. For IA and CM regimen, the prospective study of Runbo Zhong’s team showed that anlotinib combined with PD-1 inhibitors exhibited better mPFS than chemotherapy group (4.33 vs 3.6 months), a significant increase of nearly 1 month (P=0.005). The mOS was 14.17 months in the combination therapy group and 9.00 months in the chemotherapy group, with a significant extension of 5.17 months (P=0.029). The DCR of anlotinib combined with PD-1 inhibitor group was 92.1% (28). Compared with our analysis results, we found that the study IA regimen mPFS was longer, the DCR was higher, and the CM regimen mPFS was shorter.

In terms of safety, despite higher rates of grade ≥3 TRAEs, our results show that the ICA group was generally well tolerated with no new safety signals. This is similar to previous studies of such regimens in other patient populations (9, 10).

Multivariate analysis showed that for NSCLC patients with EGFR mutations who failed EGFR-TKIs treatment, the earlier they received immunotherapy, the superior benefit in PFS. A clinical study reported that patients who received EGFR-TKIs had alteration in their tumor microenvironment that contributed to immunotherapy. If other treatments were performed during this period, it might lead to the interference of the favorable tumor microenvironment, which could affect the subsequent treatment of ICI (29). Therefore, the timing of immunotherapy administration was equally important.

Subgroups with common EGFR mutations receiving immunotherapy were analyzed in our study. The results indicated that patients with the 21L858R mutation tended to have longer mPFS compared with the 19del-mutant population, but the difference was not statistically significant. However, a clinical study showed that patients with the L858R mutation receiving immunotherapy had a better response and an OS benefit compared with 19del population (30).

Analysis of the population detecting PD-L1 status showed that there was no significant difference in the median PFS, ORR and DCR in the PD-L1 high expression group compared with the low expression and negative groups. However, clinically relevant studies were controversial. A single-arm study of toripalimab combined with chemotherapy at the Shanghai Pulmonary Hospital confirmed that the PD-L1 expression levels were not associated with ICI efficacy (31). The study by Shunli Peng et al. found that the ORR in the high PD-L1 expression group was higher than that in the low PD-L1 expression group in EGFR mutant patients receiving ICI (32). Another study showed that TMB might be a more suitable biomarker for predicting ICI efficacy than PD-L1 (33). Although high PD-L1 expression suggested a better response to PD-1/PD-L1 inhibitors, it was not absolute, and some patients with high PD-L1 expression still could not benefit from ICI. The reasons included a range of factors, such as tumor mutation burden (TMB), coexisting other gene mutations, microsatellite instability (MSI), “cold” or “hot” tumor microenvironment can affect the efficacy of ICI (34–37).

Our study had several limitations. It was retrospective in nature, leading to the possibility of bias in clinical data. No additional patients receiving dual immunotherapy were included because only eight patients were identified when we reviewed the medical records. However, our study is the first to compare most treatment regimens after EGFR-TKIs failure. Moreover, because driver gene-positive patients did not respond to ICI, most of the current studies did not set an IM group for patients after receiving EGFR-TKIs, and our study precisely included these population. In contrast, our study might able to draw more comprehensive conclusions. To some extent, our analytical results provided a reference for subsequent immunotherapy strategies for similar patients. ICI-based combination therapy, especially immunotherapy combination chemotherapy and anti-angiogenesis might be the preferred treatment regimen for advanced NSCLC patients with EGFR mutations after EGFR- TKIs failure.

In conclusion, our results suggest that immunotherapy combination chemotherapy and anti-angiogenesis are more likely to prolong PFS and improve ORR and DCR than other strategies. However, more prospective clinical studies are still needed to further confirm the efficacy of this strategy.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The studies involving humans were approved by institutional ethical review board of Shanghai Chest Hospital (ethical approval number:KS22021). The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin because individual consent for this retrospective analysis was waived.

YY: Conceptualization, Formal Analysis, Methodology, Validation, Writing – review & editing. XL: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. HH: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. YS: Conceptualization, Validation, Writing – review & editing. QJ: Conceptualization, Data curation, Validation, Writing – review & editing.