Treatment patterns and outcomes of later-line therapy in advanced non-small cell lung cancer with unknown or no actionable genomic alterations: a multicenter real-world study in China (RECAP study)

Introduction

Lung cancer remains the leading cause of cancer-related mortality worldwide, representing a significant public health burden (1,2). Non-small cell lung cancer (NSCLC), the most common histological subtype, accounts for approximately 85% of all lung cancer diagnoses (3-5). Due to its insidious onset and limited early symptoms, a large proportion of patients are diagnosed at an advanced stage, with over 40% presenting with stage IV disease at initial diagnosis (3). The therapeutic landscape for advanced NSCLC is largely determined by the presence or absence of actionable genomic alterations (AGAs) (6,7). Targeted therapies have shown substantial efficacy in patients harboring driver mutations such as epidermal growth factor receptor (EGFR) mutations or ALK receptor tyrosine kinase (ALK) rearrangements (8,9), leading to improved survival in both clinical trials and real-world settings (10,11). However, approximately 40–60% of patients with advanced NSCLC lack AGAs, rendering them ineligible for targeted therapies (12). Optimal treatment strategies for this molecularly unselected population remain under investigation.

The introduction of immune checkpoint inhibitors (ICIs) targeting programmed death 1 (PD-1) and its ligand PD-L1 has transformed the treatment landscape of advanced NSCLC without known actionable genomic alterations and established immunotherapy-based regimens as the standard first-line (1L) treatment for many patients (13,14). Despite these advances, important clinical challenges remain. ICIs can cause serious immune-related adverse events (AEs), including pneumonitis and colitis (15,16). Moreover, up to 60% of initial responders eventually develop acquired resistance and experience disease progression (17,18). Following progression on 1L therapy, treatment selection in the second-line (2L) and third-line (3L) settings remains challenging, as evidence-based treatment recommendations are limited (19). Clinical decision-making is further complicated by substantial heterogeneity in patient and disease characteristics. Factors such as histological subtype, the presence of brain metastases, and the use of different systemic treatment regimens may significantly influence treatment selection and clinical outcomes (20-22). However, real-world evidence remains limited regarding how post–1L treatment patterns and these heterogeneous factors affect clinical outcomes in advanced NSCLC with unknown or no AGAs.

To address these knowledge gaps, we conducted the RECAP study, a multicenter, retrospective, real-world cohort study, to characterize treatment patterns and evaluate clinical outcomes associated with 2L and 3L treatment strategies in advanced NSCLC with unknown or no AGAs. We further examined treatment utilization and clinical outcomes across key sources of clinical heterogeneity, including histological subtype. We present this article in accordance with the STROBE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0397/rc).

Methods

Study design and patient characteristics

The RECAP study was a retrospective, multicenter, real-world study conducted across six tertiary cancer centers in China, including Peking University Cancer Hospital and Institute, Shandong Cancer Hospital and Institute, Shandong First Medical University and Shandong Academy of Medical Sciences, Sun Yat-sen University Cancer Center, The First Affiliated Hospital of Soochow University, Liaoning Cancer Hospital and Institute, and The First Affiliated Hospital of Fujian Medical University. It included patients with advanced NSCLC who received 2L or 3L systemic therapy between September 1, 2019, and December 31, 2022. The present analysis represents a pre-specified subset of the RECAP study, focusing exclusively on patients with unknown or no AGAs. The study flowchart is illustrated in Figure 1A. The study is registered with ClinicalTrials.gov (NCT06617390). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committees of Peking University Cancer Hospital and Institute (No. 2024YW114), Shandong Cancer Hospital and Institute, Shandong First Medical University and Shandong Academy of Medical Sciences (No. SDZLEC2024-225-01), Sun Yat-sen University Cancer Center (No. B2024-741-01), The First Affiliated Hospital of Soochow University [No. (2024) 215], Liaoning Cancer Hospital and Institute (No. 2024985), and The First Affiliated Hospital of Fujian Medical University (No. LY2024-065-01). Given the retrospective nature of the study, informed consent was waived by the ethics committees.

Figure 1 Treatment patterns. (A) Among patients enrolled in 2L therapy (2L-enrolled). (B) Among patients enrolled in 3L therapy (3L-enrolled). 1L, first-line therapy; 2L, second-line therapy; 3L, third-line therapy; A, anti-angiogenic therapy; C, chemotherapy; I, immunotherapy; Non-AGA, driver gene testing yielded negative results or was not reported; NSCLC, non-small cell lung cancer; NSQ, non-squamous NSCLC; pts, patients; SQ, squamous NSCLC; T, targeted therapy.

Patients were eligible for inclusion if they were aged 18 years or older at enrollment with stage IV NSCLC confirmed by histological or cytological examination and had received prior immunotherapy in the 1L or 2L setting. Eligible patients were classified as unknown or no AGAs if no AGAs were detected in available molecular testing results, or if AGAs status was unavailable due to missing documentation in the medical records. AGAs were defined as alterations in epidermal growth factor receptor (EGFR), ALK receptor tyrosine kinase (ALK), ROS proto-oncogene 1, receptor tyrosine kinase (ROS1), neurotrophic tyrosine receptor kinase (NTRK), B-Raf proto-oncogene, serine/threonine kinase (BRAF) V600 mutation, MET proto-oncogene, receptor tyrosine kinase (MET) exon 14 skipping mutations, and ret proto-oncogene (RET). This classification reflects real-world variability in molecular testing practices and documentation completeness across participating centers.

Patients were excluded if they had participated in clinical trials for lung cancer during the study period; had concurrent malignancies at diagnosis (excluding non-metastatic non-melanoma skin cancers, carcinoma in situ, or benign tumors); or had been diagnosed with another malignancy within five years of their NSCLC diagnosis. Additional exclusion criteria included histological diagnosis of small cell lung cancer, neuroendocrine tumors, or mixed small cell and non-small cell histology; incomplete medical records; or clinical ineligibility as judged by the investigators.

Data collection and patient grouping

Clinical data were retrospectively extracted from the electronic medical records and hospital information systems of all participating centers. Patients were followed until October 30, 2023. Variables collected included demographic characteristics, disease features, treatment history, and clinical outcomes. Patients were stratified by the line of systemic therapy at enrollment (2L group or 3L group). Further subgroup classification was based on the presence or absence of new brain metastases at enrollment, use of docetaxel-containing versus non-docetaxel regimens in the 2L setting, and histological subtype (squamous vs. non-squamous NSCLC). Treatment regimens were classified into the following six categories: anti-angiogenic therapy (A), chemotherapy (C), chemotherapy-based combinations (C+), immunotherapy (I), immunotherapy-based combinations (I+), and other regimens (O). Cases with no available treatment data were recorded as not applicable (N/A).

Follow-up and observation outcomes

Patients were followed until October 30, 2023. In the cohort with unknown or no AGAs, observation outcomes included treatment patterns, effectiveness, and the incidence of AEs across different subgroups.

Effectiveness was assessed using real-world progression-free survival (rwPFS), time to treatment discontinuation (TTD), time to next treatment or death (TTNT), and real-world overall survival (rwOS). rwPFS was defined as the interval from the initiation of each treatment line to either documented disease progression or death from any cause, whichever occurred first. TTD was defined as the time from treatment initiation to the earliest occurrence of disease progression, treatment discontinuation, or death. For patients who continued treatment without documented discontinuation or death, TTD was censored at the date of the last lung cancer–related medication for the corresponding treatment line. TTNT was defined as the interval from the initiation of each line of therapy to the start date of the subsequent line of treatment. For patients with no subsequent treatment or no recorded start date for the next line, TTNT was censored at the date of the last medical record entry. rwOS was calculated from the date of advanced NSCLC diagnosis to the date of death from any cause. For patients not observed to have died during the follow-up period, rwOS was censored at the date of the last medical record entry or the last known survival time.

AEs were recorded irrespective of causality and encompassed the date of onset, type, severity (graded per standard criteria), outcome, treatment impact (e.g., discontinuation, interruption, or dose modification), and whether the event was fatal. All AEs were graded according to the Common Terminology Criteria for Adverse Events (CTCAE). AE data were derived solely from medical records obtained during hospitalization and outpatient clinic visits.

Statistical analysis

The full analysis set (FAS) included patients who met all inclusion criteria and none of the exclusion criteria. The efficacy analysis set (EAS) comprised FAS patients with evaluable effectiveness outcomes (rwPFS, TTD, TTNT, rwOS). The safety set (SS) included FAS patients with at least one safety evaluation. Missing data were not imputed. Continuous variables were summarized as mean (standard deviation) or median [interquartile range (IQR)], and categorical variables as counts and percentages. Time-to-event outcomes were estimated using the Kaplan–Meier method, with median values and 95% confidence intervals (CIs) reported. All analyses were performed using SAS (version 9.2) and R (version 4.2.1).

Results

Baseline characteristics

The FAS of the unknown or no AGAs cohort included 383 patients with NSCLC (Figure S1), all of whom were also included in the EAS and SS. The mean age was 61.11±9.94 years. Female patients accounted for 62 (16.2%) cases. Comorbidities were present in 256 (66.8%) patients. Surgical resection of the primary tumor had been performed in 40 (10.4%) patients, and 141 (36.8%) had received radiotherapy. Intracranial metastases at baseline were reported in 124 (32.4%) patients, while 6 (1.6%) had no distant metastases. At enrollment, 302 (78.9%) were receiving 2L therapy (2L-enrolled cohort) and 81 (21.1%) 3L therapy (3L-enrolled cohort) at study entry. Histologically, among the 376 patients with available subtype data, 153 (40.7%) had squamous NSCLC, 217 (57.7%) had adenocarcinoma, 4 (1.1%) had adenosquamous carcinoma, 2 (0.5%) had other subtypes (Table 1).

Biomarker testing

Among the cohort, 264 (68.9%) patients underwent biomarker testing. Of these, 171 (64.8%) had PD-L1 expression assessed (32.6% had PD-L1 <1%, 34.9% had PD-L1 1-50%, and 32.6% had PD-L1 >50%), and 211 (79.9%) underwent genetic testing. Reported testing methods included next-generation sequencing (NGS) in 17 (8.1%) patients and polymerase chain reaction (PCR) in 7 (3.3%); the method was unspecified in the remaining 88.6%. Among patients with available results, all tested negative for EGFR, ALK, ROS1, NTRK, BRAF, MET, and RET mutations (Table S1).

Treatment patterns

Among 2L-enrolled patients (Figure 1A; Table 2), the most common 1L regimen was immunotherapy plus chemotherapy (I + C), administered to 179 patients (59.3%), followed by chemotherapy alone (C) in 40 patients (13.2%) and immunotherapy + anti-angiogenic therapy + chemotherapy (I + A + C) in 31 patients (10.3%). In the 2L setting, I + C remained the predominant regimen (92 patients, 30.5%), followed by C (57 patients, 18.9%) and I + A + C (45 patients, 14.9%). 3L treatments were more heterogeneous, with A used in 25 patients (25.3%), I + C in 19 (19.2%), C in 18 (18.2%), and I + A + C in 13 (13.1%).

Among 3L-enrolled patients (Figure 1B; Table 2), 1L treatment most commonly consisted of C (43 patients, 53.1%) and A + C (32 patients, 39.5%). In the 2L setting, I + C (92 patients, 30.5%) and C (57 patients, 18.9%) remained the most utilized regimens. By 3L, A + C (17 patients, 21.0%), A (15 patients, 18.5%), and I + A (14 patients, 17.3%) were the most frequently used approaches, while the use of targeted therapy drugs (T) remained infrequent (3 patients, 3.7%).

Across all treatment lines in the 2L-enrolled group, patients with non-squamous NSCLC were less likely to receive I + C compared to those with squamous NSCLC, while squamous NSCLC was less frequently treated with A than non-squamous NSCLC (Figure 1A,1B). Among non-squamous NSCLC patients enrolled in the 2L-enrolled group, 128 (76.2%) received I-based combination therapy in the 1L setting. Among these, 87 (51.8%) continued I-based combination in 2L, and 71 (42.2%) switched to C-based treatment. 21 (12.5%) received I-based combination again in 3L, while others transitioned to A (14 patients, 8.3%), C (11 patients, 6.5%), or other regimens (Figure 2A). Notably, all patients who received C or C- based treatment without ICIs (29 patients, 17.2%) in 1L subsequently transitioned to I or I-based combination therapy in 2L. Among patients with squamous NSCLC enrolled in the 2L-enrolled group, 87 (70.2%) received I-based combination as their 1L regimen, with 81 (65.3%) receiving I-based combination in 2L and 19 (15.3%) of them received I-based combination again in 3L, 9 (7.3%) transitioned to A, 7 (5.6%) to C (Figure 2B). Others switched to C (21 patients, 16.9%), C-based combination (10 patients, 8.1%), or other regimens in 2L. Similarly, all patients who initially received C-based regimen (32 patients, 25.8%) in 1L subsequently transitioned to I-based combination regimens in 2L. The detailed chemotherapy agents administered to the 2L-enrolled patients and the 3L-enrolled patients are shown in Table S2.

Figure 2 Sankey diagrams of treatment sequences among patients enrolled in 2L therapy (2L-enrolled). (A) Treatment sequences in patients with non-squamous NSCLC; (B) Treatment sequences in patients with squamous NSCLC. 1L, first-line; 2L, second-line; 3L, third-line therapy; A, anti-angiogenic therapy; C, chemotherapy; C+, chemotherapy-based combinations including A + C, I + C, and I + A + C; I, immunotherapy; I+, immunotherapy-based combinations including I + C, I + A, I + C + A, and I + T + C; N/A, no available treatment data; NSCLC, non-small cell lung cancer; O, regimens including T, T + A, T + C, and T + A + C; T, targeted therapy.

Clinical outcomes

The median follow-up duration for the overall cohort was 5.7 months (IQR: 2.4–10.6). Among 2L-enrolled patients receiving 2L treatment, the disease progression rate was 42.3%, with a median rwPFS of 7.0 months (95% CI: 6.0–9.7), median TTD of 6.8 months (95% CI: 5.0–8.7), and median TTNT of 18.7 months [95% CI: 13.8–not available (NA); Table 3; Figure 3A], without survival differences observed among the pathological types. For 2L regimens, I + C (n=92) showed a lower progression rate of 32.6% and a longer median rwPFS of 10.1 months (95% CI: 5.9–23.2). C alone (n=57) had a progression rate of 43.9% and median rwPFS of 4.6 months (95% CI: 3.8–7.4). Anti-vascular endothelial growth factor (VEGF)-based therapies (n=143) yielded the following: A + C (n=47): progression rate 48.94%, median rwPFS 4.8 months (95% CI: 3.5–NA); I + A + C (n=45): progression rate 48.9%, median rwPFS 8.0 months (95% CI: 6.8–NA); I + A (n=34): progression rate 50.0%, median rwPFS 13.7 months (95% CI: 4.1–NA); I alone (n=14) was associated with a 35.7% progression rate and median rwPFS of 8.0 months (95% CI: 2.4–NA), while A alone (n=9) had a 33.33% progression rate and median rwPFS of 6.3 months (95% CI: 1.6–NA). Outcomes including disease progression, death, censoring, rwPFS, mTTD, and mTTNT across treatment groups are summarized in Table 3.

Figure 3 Median rwPFS by treatment regimen in 2L therapy for 2L-enrolled patients. (A) Patients with NSCLC. (B) Patients with NSQ. (C) Patients with SQ. 2L, second-line; CI, confidence interval; mPFS, median progression-free survival; NSCLC, non-small cell lung cancer; NSQ, non-squamous NSCLC; rwPFS, real-world progression-free survival; SQ, squamous NSCLC.

Among 3L-enrolled patients, the disease progression rate during 3L treatment was 37.0%, with a median rwPFS of 7.2 months (95% CI: 5.0–12.8), median TTD of 6.2 months (95% CI: 4.8–8.9), and median TTNT of 12.8 months (95% CI: 8.3–NA) (Table 3). Among specific regimens: A + C (n=17) showed a progression rate of 35.3%, median rwPFS of 5.8 months (95% CI: 4.4–NA). A monotherapy (n=15) had a progression rate of 26.7%, median rwPFS of 7.2 months (95% CI: 2.3–NA). I + A (n=14) showed a progression rate of 42.9%, median rwPFS at 8.3 months (95% CI: 4.7–NA). C alone (n=11) had a 36.4% progression rate, median rwPFS of 5.0 months (95% CI: 4.8–NA). I + A + C (n=8) showed a 37.5% progression rate, with median rwPFS of 13.6 months (95% CI: 10.6–NA). Small cohorts receiving T (n=3) or T + C (n=2) showed high progression rates (33.3% and 100%, respectively), with median rwPFS of 3.7 months (95% CI: 0.9–NA). Outcomes such as disease progression, death, censoring, rwPFS, mTTD, and mTTNT across treatment groups are summarized in Table 3.

Among 2L-enrolled patients receiving 2L treatment, those with non-squamous NSCLC had a median rwPFS of 7.4 months (95% CI: 6.1–12.9), while those with squamous NSCLC had a median rwPFS of 6.5 months (95% CI: 4.3–10.2) (Figure 3B,3C). For I-based regimens, the median rwPFS was 11.8 months in non-squamous NSCLC and 9.7 months in squamous NSCLC (Figure 3B,3C). For C-based regimens, median rwPFS was 6.5 months in non-squamous NSCLC and 3.0 months in squamous NSCLC (Figure 3B,3C). Detailed 3L rwPFS outcomes stratified by histology and enrollment line were presented in Figure S2.

In 2L-enrolled patients, those in the non-docetaxel group (n=256) had a median rwPFS of 8.0 months (95% CI: 6.6–11.2), compared to 4.0 months (95% CI: 3.0–6.8) in the docetaxel group (n=46). Within the non-docetaxel subgroup, I + A achieved the longest median rwPFS of 13.7 months (95% CI: 4.1–NA), while in the docetaxel group, I + C yielded the highest median rwPFS of 6.7 months (95% CI: 10.6–NA) (Table S3).

Among patients presenting with new brain metastases at enrollment, all eight 2L-enrolled patients experienced disease progression, with a median rwPFS of 5.0 months (95% CI: 4.0–NA). The 3L-enrolled patient with new brain metastasis (n=1) had a median rwPFS of 12.8 months (Table S4).

Safety

During the study period, 50 patients (13.1%) experienced at least one AE, with a total of 94 AEs recorded. AEs led to treatment discontinuation in 12 patients (3.1%), treatment interruption in 4 (1.0%), and dose modification in 4 (1.0%). The most common AEs were myelosuppression (32 events in 25 patients, 6.5%) and gastrointestinal reactions (21 events in 21 patients, 5.5%) and infection (10 events in 9 patients, 2.4%). Other AEs included skin and mucosal abnormalities (7 patients, 1.8%), fatigue (4 patients, 1.0%). Severe events were rare, with grade 3 AEs in 9 patients (2.4%) and grade 4 AEs in 4 (1.0%). AE outcomes were not reported in 31 patients (8.1%), remission was documented in 26 (6.8%), and 2 patients (0.5%) experienced persistent AEs without improvement (Table 4).

Discussion

The RECAP study provides a comprehensive real-world overview of 2L and 3L treatment patterns and clinical outcomes in advanced NSCLC with unknown or no AGAs in China. I-based regimens predominate, especially in 2L therapy. Among patients enrolled at treatment initiation, those receiving I + A in 2L and I + A + C in 3L demonstrated median rwPFS of 13.7 and 13.6 months, respectively. Subgroup analyses demonstrated differential rwPFS responses to 2L regimens across histological subtypes, with non-docetaxel therapies showing numerically longer rwPFS than docetaxel-based treatments. These findings underscore the complexity of clinical decision-making in later lines and the critical need for more detailed real-world evidence to guide optimal treatment selection for advanced NSCLC with unknown or no AGAs.

The study revealed distinct treatment trajectories, highlighting evolving therapeutic strategies across lines of care. Between 2019 and 2024, the 2L treatment paradigm for patients without AGAs according to the China guidelines remained unchanged, with I and C consistently recommended for both non-squamous and squamous NSCLC patients (PS 0–2). For squamous NSCLC patients deemed ineligible for I or C, A excluding bevacizumab was proposed as an alternative (8,23). In this study, real-world 2L treatment patterns exhibited greater diversity compared to guideline-recommended regimens, with the most common approaches including I + C, C, A + C, I + A + C or I + A. Among 2L-enrolled patients, the use of I + C decreased from 59.3% in the 1L setting to 30.5% in the 2L setting, whereas the use of C increased from 13.3% to 18.9%, likely reflecting clinical adjustments after progression or suboptimal response to 1L immunotherapy (8,24-26). Between 2019 and 2024, the 3L treatment paradigm for patients without AGA included I and C as core strategies. For non-squamous NSCLC patients, anlotinib (A) was recommended following failure of two prior chemotherapy regimens, whereas peripheral-type squamous NSCLC patients could also receive anlotinib (A) as a 3L option (8,23). In this study, the use of A increased substantially from 2.98% in the 2L setting to 25.3% in the 3L setting. For 3L enrollees, C and I + C dominated 1L and 2L regimens, respectively. By 3L, treatment patterns diversified, with similar frequencies across A + C (21.0%), A (18.5%), I + A (17.3%), I + C (14.8%), and C (13.6%), indicating a shift toward individualized approaches. Overall, I + C demonstrated higher utilization rates in 1L and 2L settings among patients with unknown or no AGAs, whereas 3L treatment exhibited greater diversity, with A regimens showing the numerically highest proportion.

In this study, patients with squamous NSCLC more frequently received I-based regimens than those with non-squamous NSCLC, consistent with the higher immunogenicity and clinical responsiveness observed in squamous histology (8,24). Among squamous NSCLC patients who experienced failure of first-line immunotherapy, 60.9% received subsequent I-based regimens across subsequent lines. In contrast, 44.7% of non-squamous NSCLC patients received further I-based therapy following failure of 1L immunotherapy. In contrast, a higher proportion of non-squamous NSCLC patients received A + C compared with squamous patients. This scenario deviates from guideline recommendations that advocate for A monotherapy alone (8,24), reflecting a trend toward more personalized treatment strategies in clinical practice. Supporting the use of I-based therapy, a meta-analysis encompassing multiple lines of treatment demonstrated its association with reduced risk of progression and death in both histological subtypes (27). Beyond inherent immunogenicity, squamous histology is associated with a higher risk of bleeding complications with anti-angiogenic agents. This safety concern limits treatment options for squamous patients, increasing their relative dependence on, and utilization of, immunotherapy-chemotherapy combinations compared to non-squamous patients. The substantial use of I-based regimens after failure of 1L immunotherapy observed in our study may reflect growing clinical interest in immunotherapy rechallenge strategies. The 2024 Chinese expert consensus on post-progression management of advanced NSCLC includes immunotherapy rechallenge as one of several potential treatment approaches, alongside chemotherapy alone and chemotherapy combined with anti-angiogenic therapy (28). Retrospective studies have reported that immunotherapy rechallenge across treatment lines in advanced NSCLC with non-AGA is associated with improved overall survival (29) and comparable PFS between squamous and non-squamous histologies [hazard ratio (HR): 1.170, 95% CI: 0.694–1.972] (30). However, accumulating data indicates that the efficacy of immunotherapy rechallenge is heterogeneous (31,32). In particular, ICI monotherapy rechallenges in unselected NSCLC populations who discontinued initial immunotherapy due to disease progression have generally demonstrated limited clinical benefit (33). In contrast, more favorable outcomes have been reported in selected patients, including those who achieved durable disease control with initial immunotherapy and experienced relapse after a treatment-free interval, whereas benefit appears limited in patients with primary progression during initial treatment or uncertain in those discontinuing due to AEs (31). Emerging evidence also suggests that combination strategies incorporating ICIs and anti-angiogenic agents may demonstrate encouraging activity in previously treated NSCLC (34). Collectively, these findings suggest that the benefit of immunotherapy-based retreatment strategies is unlikely to be uniform across all patients and may depend on careful patient selection. Taken together, our findings demonstrate substantial real-world use of subsequent immunotherapy-based regimens after prior immunotherapy exposure, particularly among patients with squamous NSCLC. These observations underscore the need for improved patient stratification to identify patients most likely to benefit from immunotherapy-based retreatment strategies following progression on first-line immunotherapy.

In this real-world cohort, the median follow-up duration was 5.7 months. Among 2L-enrolled patients, the disease progression rate was 42.4%, with a median rwPFS of 7.0 months, median TTD of 6.8 months, and median TTNT of 18.7 months. These findings broadly align with previous clinical trials evaluating 2L therapies in non-AGA advanced NSCLC populations (8,35). In 3L-enrolled patients, the progression rate was slightly lower (37.0%), with a median rwPFS of 7.2 months, median TTD of 6.2 months, and median TTNT of 12.8 months. Notably, the observed rwPFS in the 3L setting exceeded typical clinical trial results, which generally range from 3 to 6 months (36,37). This study also demonstrated that A + I (median rwPFS, 13.7 months) and I + C (median rwPFS, 10.1 months) showed superior effectiveness in 2L treatment compared to other regimens. This may reflect more individualized treatment selection, selective inclusion of fitter patients, and the relatively short observation window. It is critical to consider the potential confounding effect of pseudoprogression, a phenomenon unique to immunotherapy where initial tumor enlargement or new lesion formation precedes subsequent regression (38).

Treatment effectiveness varied considerably by regimen. For 2L treatment in 2L-enrolled patients, the combination of I + A achieved the longest median rwPFS (13.7 months), followed by I + C (10.1 months). For 3L treatment in 3L-enrolled patients, I + A + C tended to yield more favorable outcomes, with a median rwPFS of 13.6 months, notably higher than the 4.5 months observed with I + C. It is important to note that rwPFS data may be influenced by pseudoprogression, a well-documented phenomenon associated with immunotherapy. The median OS was not estimable within the current follow-up period. The relatively low mortality rate may be attributed to challenges in patient follow-up, such as patients not returning or seeking care elsewhere, lack of systematic long-term survival data collection, and the limited duration of follow-up, all likely contributing to incomplete mortality ascertainment. Collectively, these findings suggest that regimens integrating anti-angiogenic agents may provide additive or synergistic benefit in the context of immunotherapy resistance, warranting further prospective investigation (8,24). As treatments were administered in routine clinical practice and outcomes may be influenced by confounding, observed differences in efficacy across groups, including those involving immunotherapy which may be affected by biases such as pseudoprogression, should be interpreted cautiously and considered exploratory.

Subgroup analyses in the RECAP study showed differential effects across clinical and pathological subgroups. Our study of 2L I-based regimens demonstrated median rwPFS durations of 11.8 months in non-squamous NSCLC and 9.7 months in squamous NSCLC patients, showing similar efficacy patterns to those reported for 1L immunotherapy plus chemotherapy in real-world settings (35,39). However, C-based therapies, particularly those combined with A, were also more effective in non-squamous than in squamous NSCLC, reinforcing known histological differences in treatment response (25). In 3L therapy, rwPFS in non-squamous NSCLC patients was 3.8 months and 5.8 months among 2L and 3Lenrollees, respectively. Conversely, in 2L therapy, rwPFS in squamous NSCLC patients among 2L-enrolled and 3L-enrolled were 3.9 months and 9.0 months, respectively. From the data, 3L-enrolled patients demonstrated numerically better rwPFS compared to those in 2L-enrolled patients. Notably, squamous NSCLC patients showed superior rwPFS outcomes relative to non-squamous NSCLC subgroups. However, these findings must be interpreted with caution due to small sample size and heterogeneous patient populations, which may introduce potential biases in subgroup analyses.

According to current clinical guidelines, docetaxel remained the recommended standard for 2L therapy. However, in this study, patients initiating 2L treatment with non-docetaxel regimens (n=256) demonstrated a numerically longer median rwPFS (8.0 months) compared to those receiving docetaxel-containing regimens (n=46; median rwPFS, 4.0 months). Even among patients treated with chemotherapy alone, non-docetaxel regimens appeared to confer longer rwPFS than docetaxel-containing regimens (4.6 vs. 4.0 months). Beyond sample size limitations, these differences may reflect unaccounted baseline characteristics, comorbidities, prior therapies, and drug tolerance, which were not evaluated in this study. Within these groups, the I + A regimen yielded the longest rwPFS in the non-docetaxel cohort, while I + C showed better outcomes in the docetaxel cohort. These findings aligned with current clinical perspectives favoring I-based combinations over chemotherapy in later-line settings (8,24,25,40).

In the RECAP study, AEs were reported in 50 patients (13.1%), with a total of 94 events. The most frequent were myelosuppression (6.5%) and gastrointestinal reactions (5.5%), consistent with toxicity profiles from prior clinical studies (41,42). Less common AEs included skin and mucosal abnormalities, fatigue, and infections. Grade 3 and 4 AEs occurred in 9 (2.4%) and 4 (1.0%) patients, respectively. Treatment discontinuation due to AEs was documented in 12 patients (3.1%), while treatment interruption occurred in 4 patients (1.0%) due to 12 AE episodes. Dose modifications were required in 1.0% of patients. These toxicity rates were lower than those typically observed in clinical trials, possibly reflecting real-world clinical discretion in dose adjustments and patient selection with favorable baseline performance status (43). Encouragingly, nearly half of all AEs (46.8%) resolved with appropriate management, indicating most toxicities were controllable. Nonetheless, some patients experienced recurrent, unresolved, or worsening AEs, underscoring the importance of ongoing surveillance and tailored supportive care in later-line therapy (44).

This study had several limitations warranting consideration. First, as a retrospective study, the presence of missing data may have introduced selection and information biases. During screening, 2,030 patients were excluded due to extensive missing data, reflecting inconsistencies in medical record documentation in China. This limitation is particularly evident in biomarker profiling: among 383 patients, PD-L1 expression or driver gene results were available for 264, while the remaining patients’ driver gene status was unknown. Biomarker profiling data were retrospectively extracted from medical records, and verification of testing methods was not feasible, resulting in substantial missing information. In addition, mortality data were difficult to ascertain retrospectively, further limiting the completeness and reliability of survival outcomes. Residual confounding cannot be excluded, as multivariable analyses were not performed because of substantial treatment heterogeneity and limited sample sizes and outcome events in several subgroups; moreover, such analyses were not pre-specified in the study protocol or statistical analysis plan. Second, the median follow-up of 5.7 months may have been insufficient to reliably assess long-term outcomes, particularly OS, potentially underestimating progression events, as well as durable treatment effects. Disease progression was determined based on clinical documentation rather than standardized imaging schedules or centralized radiologic review, introducing measurement bias, particularly for patients receiving immunotherapy or immunotherapy-based combinations due to possible pseudoprogression. Early in immunotherapy use, limited consensus on defining and managing pseudoprogression further affected the reliability of progression endpoints. Third, imbalances in sample sizes across 2L and 3L subgroups, especially for less common regimens, limited statistical power and confidence in subgroup comparisons, and the small number of patients with baseline brain metastases restricted interpretability and generalizability of rwPFS, TTD, and TTNT estimates. Finally, safety data were derived from routinely documented inpatient and outpatient medical records rather than through protocol-specified prospective safety assessments. Consequently, mild or transient AEs that did not require medical attention may not have been consistently captured, potentially resulting in an underestimation of treatment-related toxicity. Accordingly, the reported AEs rates should be interpreted with caution.

Conclusions

In this multicenter real-world study, we evaluated 383 patients with advanced NSCLC and unknown or no AGAs who received 2L or 3L systemic therapy. The RECAP study characterized treatment distribution and effectiveness across therapy lines, revealing that I + C was the most frequently used regimen in the 2L setting, whereas A + C predominated in the 3L setting. I-based regimens were commonly used across treatment lines in this population. Patients receiving I-based combination therapies generally showed numerically longer rwPFS compared with other regimens in both 2L- and 3L-enrolled patients, particularly for regimens incorporating anti-angiogenic therapy. However, these findings should be interpreted cautiously given the retrospective study design, treatment heterogeneity, and exploratory nature of the analyses. The substantial use of I-based regimens after prior immunotherapy exposure reflects contemporary real-world treatment practices and should not be interpreted as evidence supporting routine immunotherapy rechallenge after disease progression. Further prospective studies are warranted to better define the role of subsequent immunotherapy-based treatment strategies, including combinations with anti-angiogenic agents, in later-line management of advanced NSCLC with unknown or no AGAs.

Acknowledgments

This study was sponsored by Daiichi Sankyo (China) Holdings Co., Ltd. And in July 2020, AstraZeneca entered into a global development and commercialization collaboration agreement with Daiichi Sankyo for Datopotamab Deruxtecan (Dato-DXd). Medical writing support was provided by Yujie Ma, MM, of Shanghai MedSci Healthcare Co. Ltd. (Shanghai, China). The abstract was presented at ESMO Asia Congress, 2025.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0397/rc

Data Sharing Statement: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0397/dss

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0397/prf

Funding: This work was supported by Daiichi Sankyo (China) Holdings Co., Ltd., the Wu Jieping Medical Foundation Clinical Research Special Fund (No. 320.6750.2023-05-50), and the Peking University Cancer Hospital Clinical Research Fund (No. QNJJ202323).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0397/coif). L.M. is employed by Daiichi Sankyo (China) Holdings Co., Ltd. J.Z. declares receiving payment or honoraria for lectures and has engaged in academic/educational collaborations with Novartis, Pfizer, AstraZeneca, and Be-one. Additionally, J.Z. has participated in a Data Safety Monitoring Board or Advisory Board for Zai-lab. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committees of Peking University Cancer Hospital and Institute (No. 2024YW114), Shandong Cancer Hospital and Institute, Shandong First Medical University and Shandong Academy of Medical Sciences (No. SDZLEC2024-225-01), Sun Yat-sen University Cancer Center (No. B2024-741-01), The First Affiliated Hospital of Soochow University [No. (2024) 215], Liaoning Cancer Hospital and Institute (No. 2024985), and The First Affiliated Hospital of Fujian Medical University (No. LY2024-065-01). Given the retrospective nature of the study, informed consent was waived by the ethics committees.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Hendriks LEL, Remon J, Faivre-Finn C, et al. Non-small-cell lung cancer. Nat Rev Dis Primers 2024;10:71. [Crossref] [PubMed]
3. Zeng H, Ran X, An L, et al. Disparities in stage at diagnosis for five common cancers in China: a multicentre, hospital-based, observational study. Lancet Public Health 2021;6:e877-87. [Crossref] [PubMed]
4. Luo YH, Chiu CH, Scott Kuo CH, et al. Lung Cancer in Republic of China. J Thorac Oncol 2021;16:519-27. [Crossref] [PubMed]
5. Howlader N, Forjaz G, Mooradian MJ, et al. The Effect of Advances in Lung-Cancer Treatment on Population Mortality. N Engl J Med 2020;383:640-9. [Crossref] [PubMed]
6. Meyer ML, Fitzgerald BG, Paz-Ares L, et al. New promises and challenges in the treatment of advanced non-small-cell lung cancer. Lancet 2024;404:803-22. [Crossref] [PubMed]
7. Jeon H, Wang S, Song J, et al. Update 2025: Management of Non Small-Cell Lung Cancer. Lung 2025;203:53. [Crossref] [PubMed]
8. Guidelines for the Diagnosis and Treatment of Non - small Cell Lung Cancer. 2024. Available online: https://www.csco.org.cn/cn/index.aspx
9. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®) Non-Small Cell Lung Cancer. 2025. Available online: https://education.nccn.org/Sep2025
10. Zhou F, Guo H, Xia Y, et al. The changing treatment landscape of EGFR-mutant non-small-cell lung cancer. Nat Rev Clin Oncol 2025;22:95-116. [Crossref] [PubMed]
11. Cooper AJ, Sequist LV, Lin JJ. Third-generation EGFR and ALK inhibitors: mechanisms of resistance and management. Nat Rev Clin Oncol 2022;19:499-514. [Crossref] [PubMed]
12. Tan AC, Tan DSW. Targeted Therapies for Lung Cancer Patients With Oncogenic Driver Molecular Alterations. J Clin Oncol 2022;40:611-25. [Crossref] [PubMed]
13. Dunne EG, Fick CN, Isbell JM, et al. The Emerging Role of Immunotherapy in Resectable Non-Small Cell Lung Cancer. Ann Thorac Surg 2024;118:119-29. [Crossref] [PubMed]
14. Reck M, Remon J, Hellmann MD. First-Line Immunotherapy for Non-Small-Cell Lung Cancer. J Clin Oncol 2022;40:586-97. [Crossref] [PubMed]
15. Martins F, Sofiya L, Sykiotis GP, et al. Adverse effects of immune-checkpoint inhibitors: epidemiology, management and surveillance. Nat Rev Clin Oncol 2019;16:563-80. [Crossref] [PubMed]
16. Lahiri A, Maji A, Potdar PD, et al. Lung cancer immunotherapy: progress, pitfalls, and promises. Mol Cancer 2023;22:40. [Crossref] [PubMed]
17. Memon D, Schoenfeld AJ, Ye D, et al. Clinical and molecular features of acquired resistance to immunotherapy in non-small cell lung cancer. Cancer Cell 2024;42:209-224.e9. [Crossref] [PubMed]
18. Zhang D, Huang J, Zhang C, et al. Progress on PD-1/PD-L1 Checkpoint Inhibitors in Lung Cancer. Zhongguo Fei Ai Za Zhi 2019;22:369-79. [Crossref] [PubMed]
19. Cheng Y, Zhang T, Xu Q. Therapeutic advances in non-small cell lung cancer: Focus on clinical development of targeted therapy and immunotherapy. MedComm (2020) 2021;2:692-729. [Crossref] [PubMed]
20. Inomata M, Hirai T, Seto Z, et al. Clinical Parameters for Predicting the Survival in Patients with Squamous and Non-squamous-cell NSCLC Receiving PD-1 Inhibitor Therapy. Pathol Oncol Res 2020;26:327-33. [Crossref] [PubMed]
21. Okuno T, Isobe T, Tsubata Y. Current pharmacologic treatment of brain metastasis in non-small cell lung cancer. Clin Exp Metastasis 2024;41:549-65. [Crossref] [PubMed]
22. Eguren-Santamaria I, Sanmamed MF, Goldberg SB, et al. PD-1/PD-L1 Blockers in NSCLC Brain Metastases: Challenging Paradigms and Clinical Practice. Clin Cancer Res 2020;26:4186-97. [Crossref] [PubMed]
23. Guidelines for the Diagnosis and Treatment of Non - small Cell Lung Cancer. 2019. Available online: https://www.csco.org.cn/cn/index.aspx
24. Riely GJ, Wood DE, Ettinger DS, et al. Non-Small Cell Lung Cancer, Version 4.2024, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw 2024;22:249-74. [Crossref] [PubMed]
25. de Castro G Jr, Kudaba I, Wu YL, et al. Five-Year Outcomes With Pembrolizumab Versus Chemotherapy as First-Line Therapy in Patients With Non-Small-Cell Lung Cancer and Programmed Death Ligand-1 Tumor Proportion Score ≥ 1% in the KEYNOTE-042 Study. J Clin Oncol 2023;41:1986-91. [Crossref] [PubMed]
26. Mazieres J, Rittmeyer A, Gadgeel S, et al. Atezolizumab Versus Docetaxel in Pretreated Patients With NSCLC: Final Results From the Randomized Phase 2 POPLAR and Phase 3 OAK Clinical Trials. J Thorac Oncol 2021;16:140-50. [Crossref] [PubMed]
27. Li F, Zhai S, Lv Z, et al. Effect of histology on the efficacy of immune checkpoint inhibitors in advanced non-small cell lung cancer: A systematic review and meta-analysis. Front Oncol 2022;12:968517. [Crossref] [PubMed]
28. Chinese Expert Consensus on the Assessment of Resistance to First-Line Immunotherapy and Management Approaches in Patients with Driver Gene-Negative Advanced NSCLC. (2024 Edition). National Medical Journal of China 2024;104:411-26.
29. Feng J, Chen X, Wei J, et al. Safety and efficacy of immune checkpoint inhibitor rechallenge in advanced non-small cell lung cancer: a retrospective study. Sci Rep 2024;14:2315. [Crossref] [PubMed]
30. Xu M, Wang Y, Wang K, et al. Efficacy of immune checkpoint inhibitor rechallenge in initial immunotherapy responders with advanced non-small cell lung cancer: A single-center retrospective study. Invest New Drugs 2024;42:703-15. [Crossref] [PubMed]
31. Livanou ME, Nikolaidou V, Skouras V, et al. Efficacy of NSCLC Rechallenge with Immune Checkpoint Inhibitors following Disease Progression or Relapse. Cancers (Basel) 2024;16:1196. [Crossref] [PubMed]
32. Knetki-Wróblewska M, Chmielewska I, Wojas-Krawczyk K, et al. Rechallenge of immunotherapy in non-small cell lung cancer patients — navigating indications and evolving perspectives. Oncology in Clinical Practice 2024;20:428-37.
33. Watanabe H, Kubo T, Ninomiya K, et al. The effect and safety of immune checkpoint inhibitor rechallenge in non-small cell lung cancer. Jpn J Clin Oncol 2019;49:762-5. [Crossref] [PubMed]
34. Liu L, Yang H, Xiong Y, et al. Efficacy, safety, and exploratory biomarker analysis of rechallenge with immune checkpoint inhibitors combined with anlotinib in previously treated advanced non-small cell lung cancer. BMC Pulm Med 2026; Epub ahead of print. [Crossref]
35. Borghaei H, Gettinger S, Vokes EE, et al. Five-Year Outcomes From the Randomized, Phase III Trials CheckMate 017 and 057: Nivolumab Versus Docetaxel in Previously Treated Non-Small-Cell Lung Cancer. J Clin Oncol 2021;39:723-33. [Crossref] [PubMed]
36. Gadgeel S, Rodríguez-Abreu D, Speranza G, et al. Updated Analysis From KEYNOTE-189: Pembrolizumab or Placebo Plus Pemetrexed and Platinum for Previously Untreated Metastatic Nonsquamous Non-Small-Cell Lung Cancer. J Clin Oncol 2020;38:1505-17. [Crossref] [PubMed]
37. Socinski MA, Nishio M, Jotte RM, et al. IMpower150 Final Overall Survival Analyses for Atezolizumab Plus Bevacizumab and Chemotherapy in First-Line Metastatic Nonsquamous NSCLC. J Thorac Oncol 2021;16:1909-24. [Crossref] [PubMed]
38. Park HJ, Kim KW, Pyo J, et al. Incidence of Pseudoprogression during Immune Checkpoint Inhibitor Therapy for Solid Tumors: A Systematic Review and Meta-Analysis. Radiology 2020;297:87-96. [Crossref] [PubMed]
39. Wang Z, Wu L, Li B, et al. Toripalimab Plus Chemotherapy for Patients With Treatment-Naive Advanced Non-Small-Cell Lung Cancer: A Multicenter Randomized Phase III Trial (CHOICE-01). J Clin Oncol 2023;41:651-63. [Crossref] [PubMed]
40. Reck M, Popat S, Grohé C, et al. Anti-angiogenic agents for NSCLC following first-line immunotherapy: Rationale, recent updates, and future perspectives. Lung Cancer 2023;179:107173. [Crossref] [PubMed]
41. O'Byrne K, Popoff E, Badin F, et al. Long-term comparative efficacy and safety of nivolumab plus ipilimumab relative to other first-line therapies for advanced non-small-cell lung cancer: A systematic literature review and network meta-analysis. Lung Cancer 2023;177:11-20. [Crossref] [PubMed]
42. Leonetti A, Perrone F, Puntoni M, et al. Real-world outcomes of Italian patients with advanced non-squamous lung cancer treated with first-line pembrolizumab plus platinum-pemetrexed. Eur J Cancer 2024;202:114006. [Crossref] [PubMed]
43. Levy BP, Felip E, Reck M, et al. TROPION-Lung08: phase III study of datopotamab deruxtecan plus pembrolizumab as first-line therapy for advanced NSCLC. Future Oncol 2023;19:1461-72. [Crossref] [PubMed]
44. Hines JB, Bowar B, Levine E, et al. Targeted Toxicities: Protocols for Monitoring the Adverse Events of Targeted Therapies Used in the Treatment of Non-Small Cell Lung Cancer. Int J Mol Sci 2023;24:9429. [Crossref] [PubMed]