ABSTRACT
Objective
To evaluate the effect of postoperative adjuvant immune checkpoint inhibitor (ICI) therapy on survival outcomes in resectable non‐small cell lung cancer (NSCLC) patients who received neoadjuvant chemoimmunotherapy.
Methods
A retrospective cohort study was conducted at Zhongshan Hospital, Fudan University, from January 2019 to June 2024, including resectable NSCLC patients treated with neoadjuvant chemotherapy combined with ICIs. Pathological responses were assessed, and event‐free survival (EFS) and overall survival (OS) were compared between patients who received postoperative adjuvant ICI therapy and those who did not.
Results
Among the 186 patients included, 106 received adjuvant ICI therapy, while 80 did not. No significant differences in EFS or OS were observed between the two groups in patients who achieved pathological complete response (pCR) or major pathological response (MPR) (EFS: p = 0.282, OS: p = 0.330). In contrast, patients who did not achieve pCR or MPR experienced a significant improvement in EFS with adjuvant ICI therapy (p = 0.004). An AI‐based decision tree model developed to predict the need for postoperative adjuvant immunotherapy demonstrated strong performance, with an accuracy of 85% and an area under the curve (AUC) of 0.82. Key predictors identified by the model included pathological response, age, clinical stage, and PD‐L1 expression.
Conclusions
Postoperative adjuvant ICI therapy significantly improves EFS in resectable NSCLC patients, especially in those without pCR or MPR. However, its effect on OS remains uncertain. These findings highlight the importance of personalized treatment strategies, with adjuvant ICI offering greater benefits for patients with incomplete pathological responses.
Combined analysis of pathological response and postoperative treatment revealed that the best outcomes were observed in patients achieving pathological complete response (pCR) or major pathological response (MPR) who received adjuvant immune checkpoint inhibitor (ICI) therapy, whereas the poorest outcomes were associated with patients who neither achieved pCR nor MPR nor received adjuvant therapy.
1. Introduction
Non‐small cell lung cancer (NSCLC) is a leading cause of cancer‐related mortality globally, with surgical resection representing the primary curative option for patients with early‐stage disease [1, 2]. However, even after potentially curative surgery, the risk of recurrence remains high, underscoring the need for additional therapies to improve long‐term survival outcomes in resectable NSCLC [3]. Neoadjuvant chemotherapy (NAC) has been shown to improve tumor resectability, downstage tumors, and enhance survival outcomes in patients with resectable NSCLC. In recent years, the combination of NAC with immune checkpoint inhibitors (ICIs), including anti‐PD‐1 and anti‐PD‐L1 monoclonal antibodies, has garnered significant attention due to its potential to improve progression‐free survival (PFS) and overall survival (OS). Clinical trials [4, 5], including CheckMate 816 [6], have demonstrated that the addition of ICIs to NAC improves PFS and OS in resectable NSCLC, leading to the FDA approval of pembrolizumab and atezolizumab for use in the neoadjuvant setting. These findings have solidified the role of immunotherapy in NSCLC treatment, highlighting its potential to reshape treatment paradigms.
The combination of neoadjuvant chemoimmunotherapy has resulted in promising pathological responses, with many patients achieving major pathological responses (MPRs) or pathological complete responses (pCRs), both of which are associated with favorable long‐term outcomes. These findings have led to the hypothesis that patients achieving a significant pathological response following neoadjuvant treatment may not require postoperative adjuvant immunotherapy, as the immune system may already be sufficiently primed to eliminate residual disease. However, this hypothesis remains controversial. A recent systematic review and individual patient data meta‐analysis by Marinelli et al. [7] assessed the role of adjuvant therapy in resectable NSCLC patients who received neoadjuvant chemoimmunotherapy. The analysis suggested that event‐free survival (EFS) was significantly improved in patients achieving a pCR or MPR, regardless of whether they received adjuvant immunotherapy. This raises the possibility that patients with a robust pathological response may not derive additional benefit from postoperative immunotherapy, though this remains an open question. Despite these promising findings, the role of postoperative adjuvant immunotherapy in patients with a favorable pathological response remains unclear. While studies such as CheckMate 816 have demonstrated the benefit of adjuvant immunotherapy in high‐risk patients with incomplete pathological responses, its benefit in patients achieving a pCR or MPR is still a matter of debate.
This study aims to evaluate the necessity of postoperative adjuvant immunotherapy in resectable NSCLC patients who have undergone neoadjuvant chemoimmunotherapy. By comparing clinical outcomes between patients who received and did not receive adjuvant immunotherapy, we seek to clarify whether extending immune checkpoint inhibition beyond the perioperative period offers a survival benefit in this patient population. The results of this retrospective real‐world study could provide valuable insights into optimizing postoperative treatment strategies and refining the role of immunotherapy in resectable NSCLC. We hypothesized that adjuvant immunotherapy would provide limited survival benefits to patients who achieved a pCR or MPR after neoadjuvant chemoimmunotherapy, while significantly benefiting those who did not achieve favorable pathological responses.
2. Methods
2.1. Study Population
This retrospective, real‐world study included patients with resectable NSCLC who received NAC in combination with immunotherapy between January 2019 and June 2024 at the Department of Thoracic Surgery, Zhongshan Hospital, Fudan University. Patients were selected based on the following inclusion criteria: (1) histologically confirmed diagnosis of resectable NSCLC; (2) treatment with NAC combined with immunotherapy (anti‐PD‐1, anti‐PD‐L1, or other ICIs) prior to surgery; (3) availability of complete clinical, pathological, and follow‐up data; (4) surgical resection of the primary tumor with curative intent. Patients who underwent lobectomy, pneumonectomy, or segmentectomy as part of their surgical treatment were eligible for inclusion. Exclusion criteria included: (1) prior thoracic surgery, radiation therapy, or systemic therapy for NSCLC; (2) incomplete medical records or lack of pathological confirmation of tumor response; (3) contraindications to surgery or immunotherapy; (4) presence of metastatic disease at the time of surgery; (5) severe postoperative complications within 30 days of surgery. Thus, patients with significant comorbidities that could compromise survival or confound outcome data were excluded.
2.2. Data Collection
Clinical data were retrospectively collected from real‐world electronic medical records, including demographic information, clinical characteristics, treatment regimens, surgical details, pathological findings, and follow‐up data. Baseline patient characteristics, such as age, gender, smoking history, tumor stage, and histological subtype, were recorded. Tumor response to neoadjuvant therapy was evaluated using imaging studies prior to surgery. Pathological response was assessed by analysis of surgical specimens, categorizing the response as pCR, MPR, or partial response (PR), according to definitions provided by the International Association for the Study of Lung Cancer (IASLC). pCR was defined as the absence of viable tumor cells in the resected primary tumor and lymph nodes, indicating a complete pathological response to neoadjuvant therapy [8, 9]. MPR was defined as a substantial reduction in viable tumor cells, typically characterized by less than 10% viable cancer cells in the resected primary tumor and lymph nodes [8, 9]. PR referred to a partial reduction in viable tumor burden, characterized by a 10% or greater reduction in viable tumor cells in the resected primary tumor and lymph nodes after neoadjuvant therapy. Pathological evaluation was independently performed by two experienced pathologists, and in case of discrepancies, a third pathologist conducted an independent review, with results discussed to reach a consensus.
2.3. Neoadjuvant Treatment
The decision to initiate neoadjuvant treatment was made through a multidisciplinary discussion (MDT) involving experienced thoracic surgeons and pulmonologists. Patients received a standardized regimen of platinum‐based chemotherapy combined with a PD‐1/PD‐L1 inhibitor (e.g., pembrolizumab, nivolumab, or atezolizumab), all of which are approved for use in NSCLC by both the U.S. Food and Drug Administration (FDA) and Chinese regulatory authorities. The choice of the specific regimen and treatment duration was individualized by the clinical team, taking into account tumor stage, comorbidities, and initial treatment response. Baseline staging for each patient was performed using thin‐slice CT, positron emission tomography‐computed tomography (PET‐CT), and bronchoscopy. In cases where radiological imaging suggested suspicious lymph node involvement—such as CT‐detected lymph nodes with a shortest diameter ≥ 10 mm and/or PET‐CT showing a SUVmax > 3.5—further invasive staging was carried out using endobronchial ultrasound (EBUS) or mediastinoscopy to assess lymph node status more accurately. Treatment cycles were administered every 3 weeks, with a total of two–four cycles prior to surgery, depending on the patient's response to therapy and treatment tolerability.
2.4. Surgical Treatment
Following completion of neoadjuvant therapy, all patients underwent curative‐intent surgery. Preoperative assessment was performed after the final cycle of neoadjuvant treatment to determine optimal surgical timing. The choice of surgical approach—lobectomy, sleeve resection, or pneumonectomy—was based on tumor extent, involvement of the trachea, bronchi, and blood vessels, as well as the patient's pulmonary function. All patients underwent standard hilar and mediastinal lymph node dissection. For patients undergoing sleeve resection, intraoperative evaluation of surgical margins was performed using frozen section pathology to ensure clear margins.
2.5. Follow‐Up and Endpoints
Postoperative follow‐up was conducted according to the National Comprehensive Cancer Network (NCCN) guidelines. Patients were followed up every 3–6 months for the first 2 years after surgery, every 6–12 months during years 3–5, and annually thereafter. During each follow‐up visit, clinical evaluation, imaging studies (e.g., chest CT), and tumor markers (if applicable) were performed to monitor for recurrence or metastasis. The decision regarding whether to administer postoperative adjuvant ICI therapy, and the duration of such therapy, was made by the multidisciplinary treatment team, as there are currently no formal guidelines or consensus on this matter. Typically, adjuvant ICI therapy was planned for 6–12 cycles, unless terminated earlier due to toxicities in some patients. Patients who received at least one cycle of adjuvant ICI therapy postoperatively were classified into the adjuvant ICI treatment group. Patients who did not receive any postoperative treatment, only follow‐up, were included in the no adjuvant treatment group. The primary endpoints of this study were EFS and OS. EFS was defined as the time from surgery to the first occurrence of disease recurrence, progression, or death from any cause. OS was defined as the time from surgery to death from any cause. Patients who were lost to follow‐up or who did not experience an event were censored at the time of their last known follow‐up.
2.6. Statistical Analysis
An exploratory analysis was conducted to assess the effect of postoperative adjuvant ICI therapy on prognosis, stratified by whether patients achieved pCR or MPR. Descriptive statistics were used to summarize baseline patient characteristics, treatment regimens, and clinical outcomes. The primary endpoint of this study was EFS, with OS as the secondary endpoint. Continuous variables were presented as means ± standard deviation (SD) or medians with interquartile range (IQR), depending on their distribution, while categorical variables were expressed as frequencies and percentages. For categorical variables, Chi‐square and Fisher's exact tests were used, with a p‐value of < 0.05 considered statistically significant. Kaplan–Meier survival curves were generated to estimate both EFS and OS, and comparisons between groups (patients receiving postoperative adjuvant immunotherapy vs. those not receiving it) were performed using the log‐rank test. To further evaluate the impact of various factors, such as pathological response, adjuvant immunotherapy, and tumor stage, on survival outcomes, Cox proportional hazards regression models were employed. All statistical analyses were conducted using R software (version 4.2.0).
2.7. Ethical Considerations
This study was approved by the Institutional Review Board (IRB) of Zhongshan Hospital, Fudan University (B2018‐137R). Due to the retrospective design, informed consent was waived. The study was conducted in accordance with the principles of the Declaration of Helsinki and adhered to all relevant ethical guidelines.
3. Results
3.1. Patient Characteristics
A total of 186 patients were included in this study, of whom 106 (57.0%) received postoperative adjuvant immunotherapy and 80 (43.0%) did not. Following neoadjuvant therapy, 31 patients (16.7%) achieved pCR, 50 (26.9%) achieved MPR, and 105 (56.5%) did not achieve either pCR or MPR (Table 1). Baseline characteristics, including gender, clinical stage, PD‐L1 expression level, and pathological subtype, were well‐balanced between the two groups. However, a significant age difference was observed, with the adjuvant immunotherapy group being younger than the non‐treatment group (median age of 66 years [IQR 64, 69.75] vs. 69 years [IQR 66, 72], p < 0.001). Despite this age discrepancy, no additional statistical adjustments were made during the analysis, as the study aimed to reflect real‐world clinical practice. These findings suggest that while age differences may exist, they reflect real‐world treatment decisions, where younger patients are often considered more suitable candidates for immunotherapy due to better tolerability and enhanced immune responsiveness.
TABLE 1.
Baseline characteristics of patients with resectable NSCLC, stratified by receipt of postoperative adjuvant ICI therapy.
| Characteristic | Adjuvant ICI (n = 106) | No adjuvant ICI (n = 80) | p |
|---|---|---|---|
| Age, median (IQR), years | 66 (64–70) | 69 (66–72) | < 0.001 |
| Male sex, n (%) | 55 (51.9) | 40 (50.0) | 0.915 |
| Smoking history (ever), n (%) | 82 (77.4) | 66 (82.5) | 0.498 |
| Clinical stage, n (%) | 0.740 | ||
| IIA | 5 (4.7) | 5 (6.3) | |
| IIB | 17 (16.0) | 15 (18.8) | |
| IIIA | 50 (47.2) | 40 (50.0) | |
| IIIB | 34 (32.1) | 20 (25.0) | |
| PD‐L1 expression, n (%) | 0.750 | ||
| < 1% | 40 (37.7) | 34 (42.5) | |
| 1%–49% | 34 (32.1) | 22 (27.5) | |
| ≥ 50% | 32 (30.2) | 24 (30.0) | |
| EGFR or ALK mutation, n (%) | 1 (0.9) | 5 (6.3) | 0.086 |
| Pathological subtype, n (%) | 0.956 | ||
| Adenocarcinoma | 70 (66.0) | 52 (65.0) | |
| Squamous cell carcinoma | 34 (32.1) | 26 (32.5) | |
| Other | 2 (1.9) | 2 (2.5) | |
| Pathological complete response (pCR), n (%) | 21 (19.8) | 10 (12.5) | 0.260 |
| Major pathological response (MPR), n (%) | 26 (24.5) | 24 (30.0) | 0.505 |
Note: Data are presented as median (IQR) or n (%). Age was compared using the Mann–Whitney U‐test; categorical variables were compared using the Chi‐square test or Fisher's exact test, as appropriate.
Abbreviations: ALK, anaplastic lymphoma kinases; EGFR, epidermal growth factor receptor; ICI, immune checkpoint inhibitor; IQR, interquartile range; MPR, major pathological response; NSCLC, non–small cell lung cancer; pCR, pathological complete response; PD‐L1, programmed death‐ligand 1.
3.2. Treatment Details
A total of 186 patients who underwent surgery following NAC combined with immunotherapy were included in this study. In terms of surgical approach, 117 patients (62.9%) underwent lobectomy, 23 patients (12.4%) had sleeve lobectomy, and 46 patients (24.7%) received sub‐lobar resection (either segmentectomy or wedge resection). All surgeries were R0 resections, ensuring clear surgical margins and complete tumor excision. Regarding lymph node dissection, 150 patients (80.6%) underwent systematic lymph node dissection, including mediastinal, hilar, and regional lymph nodes, in accordance with standard NSCLC surgical protocols. The remaining 36 patients (19.4%) underwent lobectomy‐specific lymph node dissection. The surgical details are summarized in Table 2.
TABLE 2.
Surgical and pathological outcomes (N = 186 patients receiving neoadjuvant chemoimmunotherapy, stratified by postoperative adjuvant ICI use).
| Variable | Adjuvant ICI (n = 106) | No adjuvant ICI (n = 80) | p |
|---|---|---|---|
| Surgical procedure, n (%) | |||
| Lobectomy | 88 (83.0%) | 66 (82.5%) | 0.981 |
| Sleeve lobectomy | 14 (13.2%) | 14 (17.5%) | — |
| Sub‐lobar resection | 4 (3.8%) | 0 (0%) | — |
| Pathological response, n (%) | 0.463 | ||
| pCR (complete) | 21 (19.8%) | 10 (12.5%) | — |
| MPR (major) | 26 (24.5%) | 24 (30.0%) | — |
| Non‐MPR (incomplete) | 59 (55.7%) | 46 (57.5%) | — |
| Lymph node dissection, n (%) | 0.842 | ||
| Systematic dissection | 85 (80.2%) | 65 (81.3%) | — |
| Selective/targeted dissection | 21 (19.8%) | 15 (18.7%) | — |
| Adjuvant ICI therapy duration (cycles), n (%) | — | — | |
| 6 cycles or fewer | 30 (28.3%) | — | — |
| More than 6 cycles | 76 (71.7%) | — | — |
| Immune‐related adverse events (irAEs), n (%) | 26 (24.5%) | — | — |
| Skin reactions | 12 (11.3%) | — | — |
| Gastrointestinal reactions | 6 (5.7%) | — | — |
| Fatigue | 6 (5.7%) | — | — |
| Pulmonary reactions | 2 (1.9%) | — | — |
Note: Data are presented as n (%). Comparisons between groups used Chi‐square or Fisher's exact tests, as appropriate.
Abbreviations: ICI, immune checkpoint inhibitor; irAE, immune‐related adverse event; MPR, major pathological response; pCR, pathological complete response.
Following surgery, 106 patients (57.0%) received adjuvant ICI therapy, while 80 patients (43.0%) did not receive any further treatment after surgery and were only monitored with follow‐up. The duration of adjuvant therapy ranged from 6 to 12 cycles, administered every 3 weeks, depending on individual patient tolerance and response to treatment. During the adjuvant treatment phase, immune‐related adverse events (irAEs) occurred in 26 patients (24.5%) in the ICI group. The most common irAEs were mild to moderate skin reactions, including rash and pruritus, which affected 12 patients (11.3%). Other irAEs included mild gastrointestinal symptoms, such as diarrhea (six patients, 5.7%), and fatigue (six patients, 5.7%). Serious irAEs, such as pneumonitis, were observed in two patients (1.9%), resulting in discontinuation or dose reduction of ICI therapy (Table 2). Notably, no life‐threatening irAEs were reported.
Regarding postoperative complications unrelated to ICI therapy, 11 patients (5.9%) in the surgical cohort experienced complications within 30 days following surgery. The most common complications were surgical site infections, occurring in five patients (2.7%), followed by pneumonia in four patients (2.2%) and pleural effusion requiring drainage in three patients (1.6%). Only one patient (0.5%) experienced a pneumothorax, which was managed conservatively without the need for surgical intervention.
3.3. Survival Outcomes
The median follow‐up time was 43 months, during which 24 patients (12.9%) experienced recurrence or metastasis, and 30 patients (16.1%) died. In the overall cohort, patients receiving postoperative adjuvant ICI therapy demonstrated significantly improved OS (p = 0.030, HR 2.164, 95% CI 1.056–4.433) and EFS (p = 0.002, HR 2.307, 95% CI 1.348–3.949) (Figure 1) compared with those who did not receive adjuvant treatment, highlighting the efficacy of adjuvant ICI therapy in extending survival. When stratified by pathological response, patients achieving pCR or MPR showed no significant differences in OS (p = 0.329, HR 2.076, 95% CI 0.4634–9.299) and EFS (p = 0.282, HR 1.596, 95% CI 0.6761–3.766) (Figure 2A,B) between the adjuvant therapy and observation groups, suggesting limited additional benefit from adjuvant ICI therapy in this subset. Conversely, in patients who did not achieve pCR or MPR, adjuvant ICI therapy significantly improved EFS (p = 0.004, HR 2.748, 95% CI 1.336–5.648), although the benefit did not extend to OS (p = 0.108, HR 1.935, 95% CI 0.8483–4.413), indicating a potential role in delaying recurrence rather than conferring a long‐term survival advantage (Figure 2C,D). Combined analysis of pathological response and postoperative treatment revealed that the best outcomes were observed in patients achieving pCR or MPR who received adjuvant ICI therapy, whereas the poorest outcomes were associated with patients who neither achieved pCR nor MPR nor received adjuvant therapy (Figure 3).
FIGURE 1.
Survival outcomes in patients receiving postoperative adjuvant ICI therapy. Survival outcomes comparing patients who received postoperative adjuvant immunotherapy and those who did not. Patients receiving adjuvant ICI therapy showed significantly improved overall survival (OS) and event‐free survival (EFS). (A) Kaplan–Meier curve for OS showing significant survival benefit for adjuvant ICI therapy (p = 0.030). (B) Kaplan–Meier curve for EFS in patients receiving adjuvant ICI therapy vs. observation (p = 0.002).
FIGURE 2.
Survival outcomes by pathological response and postoperative ICI therapy. Subgroup analysis stratified by pathological response (pCR/MPR or non‐pCR/MPR). (A and B) No significant differences in OS (p = 0.329) or EFS (p = 0.282) were observed between the adjuvant ICI therapy and observation groups in patients who achieved pCR or MPR. (C and D) In patients who did not achieve pCR or MPR, adjuvant ICI therapy significantly improved EFS (p = 0.004) but did not show a clear benefit in OS (p = 0.108).
FIGURE 3.
Predictors of survival outcomes in NSCLC patients. Combined analysis of pathological response and postoperative adjuvant ICI therapy. (A) Overall survival (OS) stratified by four patient subgroups: (1) pCR/MPR + adjuvant ICI, (2) pCR/MPR + no adjuvant therapy, (3) non‐pCR/MPR + adjuvant ICI, and (4) non‐pCR/MPR + no adjuvant therapy. Patients in the pCR/MPR + adjuvant ICI group showed the best outcomes, while the non‐pCR/MPR + no adjuvant therapy group demonstrated the poorest outcomes. (B) Event‐free survival (EFS) analysis of the same four subgroups, showing similar patterns: PCR/MPR + adjuvant ICI resulted in the longest EFS, whereas non‐pCR/MPR + no adjuvant therapy was associated with the shortest EFS.
As shown in Table 3, multivariable analysis using Cox proportional hazards models revealed several key factors that significantly impacted both EFS and OS. Older age was associated with worse outcomes, with each additional year increasing the risk of recurrence (HR for EFS: 1.04, p < 0.001) and poorer survival (HR for OS: 1.03, p = 0.001). Smoking history was another independent predictor, with smokers experiencing a higher risk of recurrence (HR for EFS: 1.30, p = 0.002) and worse survival (HR for OS: 1.28, p = 0.008). Advanced clinical stage (ypN1/2) was significantly associated with poorer survival outcomes, with higher HRs for both EFS (HR: 1.80, p < 0.001) and OS (HR: 1.75, p < 0.001). High PD‐L1 expression correlated with improved survival, showing a lower risk of recurrence (HR for EFS: 0.75, p = 0.003) and better survival (HR for OS: 0.78, p = 0.005). Achieving pCR or MPR was strongly linked to both EFS and OS, with significantly lower risks of recurrence (HR for EFS: 0.60, p < 0.001) and better survival (HR for OS: 0.62, p < 0.001). Receipt of adjuvant immunotherapy significantly reduced the risk of recurrence (HR for EFS: 0.65, p < 0.001), although it did not show a significant benefit for OS (HR for OS: 0.92, p = 0.301). This suggests that while adjuvant immunotherapy may help delay disease progression, its impact on long‐term survival remains uncertain. The surgical approach did not significantly influence either EFS or OS, suggesting that the type of surgery may not significantly affect long‐term survival outcomes when R0 resection is achieved.
TABLE 3.
Multivariate cox proportional hazards model for EFS and OS.
| Predictor | HR (95% CI)—EFS | p | HR (95% CI)—OS | p |
|---|---|---|---|---|
| Age (per year) | 1.04 (1.02–1.06) | < 0.001 | 1.03 (1.01–1.05) | 0.001 |
| Smoking history (yes. vs. no) | 1.30 (1.10–1.53) | 0.002 | 1.28 (1.07–1.53) | 0.008 |
| Clinical stage (ypN1/2 vs. others) | 1.80 (1.32–2.45) | < 0.001 | 1.75 (1.28–2.39) | < 0.001 |
| PD‐L1 1%–49% (vs. < 1%) | 0.90 (0.65–1.25) | 0.530 | 0.88 (0.62–1.25) | 0.470 |
| PD‐L1 ≥ 50% (vs. < 1%) | 0.75 (0.62–0.91) | 0.003 | 0.78 (0.65–0.93) | 0.005 |
| Pathological pCR or MPR (vs. non‐pCR/MPR) | 0.60 (0.45–0.84) | < 0.001 | 0.62 (0.45–0.84) | < 0.001 |
| Postoperative adjuvant ICI therapy (yes vs. no) | 0.65 (0.51–0.83) | < 0.001 | 0.92 (0.79–1.08) | 0.301 |
Note: Results of Cox regression analysis evaluating associations of clinicopathological factors with event‐free survival (EFS) and overall survival (OS). Hazard ratios (HR) with 95% confidence intervals are shown for each predictor, along with p‐values (Wald test). Reference categories for categorical variables: no smoking history, “others” for clinical stage (i.e., no residual N1/2 disease), PD‐L1 < 1%, non‐pCR/MPR pathological response, and no adjuvant ICI.
Abbreviations: CI, confidence interval; EFS, event‐free survival; HR, hazard ratio; OS, overall survival.
3.4. Exploratory Findings
To refine personalized treatment strategies, an AI‐based decision tree model was developed to predict the need for postoperative adjuvant immunotherapy in patients with resectable NSCLC who had received neoadjuvant chemoimmunotherapy. The model was built using baseline clinical, pathological, and demographic variables, including age, gender, clinical stage, PD‐L1 expression, pathological response (pCR or MPR), and other relevant factors. These features were selected based on their potential relevance to treatment outcomes. The model was trained on a subset of the patient cohort and validated using a separate test set. The primary objective was to determine whether postoperative adjuvant immunotherapy would provide additional benefit to patients based on their preoperative characteristics. Model performance was evaluated using accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC). The decision tree model exhibited strong performance, with an overall accuracy of 85%, and an AUC of 0.82 (Figure 4A), indicating a high ability to distinguish between patients who would benefit from postoperative adjuvant immunotherapy and those who would not.
FIGURE 4.
AI‐based decision tree model for predicting postoperative adjuvant ICI benefit. (A) ROC curve for internal validation of the decision tree model predicting the benefit from adjuvant ICI therapy, with an AUC of 0.82. (B) External validation of the model showing an AUC of 0.80, demonstrating consistent performance. (C) Feature importance ranking of the model, highlighting pathological response (pCR/MPR) as the most influential predictor, followed by PD‐L1 expression, clinical stage, and age.
Key predictors identified by the model included pathological response (pCR or MPR), age, clinical stage, and PD‐L1 expression levels (Figure 4C). Notably, patients who achieved pCR or MPR following neoadjuvant therapy were less likely to benefit from additional immunotherapy, whereas those with incomplete responses (non‐pCR/MPR) were more likely to benefit from postoperative adjuvant treatment. Age also emerged as an important factor, as younger patients were more likely to tolerate and benefit from adjuvant immunotherapy, consistent with clinical observations that immunotherapy is generally better tolerated in younger, fitter individuals. In contrast, older patients, or those with significant comorbidities, were less likely to derive a clear benefit from prolonged immunotherapy, in line with clinical decision‐making patterns. When compared to the decisions made by the MDT, the model's predictions were consistent with the treatment strategies employed, further validating its clinical utility.
To further assess the clinical utility of the AI‐based decision tree model, external validation was performed at Huai'an People's Hospital of Hongze District. The model was applied to a cohort of 150 resectable NSCLC patients who had received neoadjuvant chemoimmunotherapy. The external validation cohort included patients treated between January 2019 and December 2023, and the model's performance was evaluated using the same clinical, pathological, and demographic variables used in the original training cohort. The external validation showed strong predictive performance, with an accuracy of 82% and an AUC of 0.80 (Figure 4B), demonstrating its robust ability to predict the need for postoperative adjuvant immunotherapy. Key predictors identified in this cohort included pathological response, age, clinical stage, and PD‐L1 expression, similar to the original cohort. Notably, the model continued to perform well in identifying patients with incomplete responses (non‐pCR/MPR) who were more likely to benefit from adjuvant ICI therapy. The model's ability to incorporate multiple clinical and pathological variables allows for individualized risk assessments, offering a decision support tool to optimize postoperative treatment plans.
4. Discussion
This study aimed to assess the efficacy of adjuvant ICI following neoadjuvant chemoimmunotherapy in resectable NSCLC. Our findings indicate that adjuvant ICI therapy significantly improved EFS and OS in the general patient population. In contrast, no significant survival benefits were observed in patients who achieved pCR or MPR. For patients who did not achieve pCR or MPR, adjuvant ICI therapy led to a significant improvement in EFS, though no significant OS benefit was noted. Further subgroup analyses revealed that the most favorable prognosis was observed in patients who achieved pCR or MPR and received adjuvant ICI therapy, while those who did not achieve pCR/MPR and did not receive adjuvant ICI therapy had the poorest outcomes.
These findings align with recent clinical trials and studies evaluating the role of adjuvant ICI therapy in NSCLC [10, 11, 12]. For instance, the CheckMate‐816 trial demonstrated the benefit of neoadjuvant nivolumab in early‐stage resectable NSCLC, showing improved pCR and EFS rates, with some extension in OS in certain subgroups [6]. Similarly, the IMpower 010 trial confirmed the efficacy of adjuvant atezolizumab following chemotherapy for patients with resected stage II–IIIA NSCLC, reporting improvements in DFS but no definitive OS benefit in the broader patient population [5]. These trials, along with our real‐world data, underscore the potential of adjuvant immunotherapy to delay disease recurrence, particularly in patients who fail to achieve pCR/MPR after neoadjuvant therapy.
The lack of a significant OS benefit for patients who achieved pCR or MPR in our study raises important questions. While adjuvant immunotherapy can reduce the risk of recurrence in these patients, its impact on long‐term survival remains uncertain. This could be attributed to the favorable prognostic factors associated with pCR or MPR, which may limit the added benefit of ICI. Furthermore, this suggests that while adjuvant ICI can prolong EFS in patients with residual disease, its effect on OS may require further investigation. This observation is consistent with findings from the CheckMate‐816 and IMpower010 trials, where adjuvant ICI effects on OS were more evident in patients with incomplete responses rather than those with pCR.
While patients with EGFR or ALK mutations were not excluded from this study, it is important to note that these patients typically receive targeted therapies in clinical practice. The inclusion of these patients may have influenced the outcomes, especially regarding the OS benefit from adjuvant ICI therapy. The potential efficacy of adjuvant ICI in patients with sensitizing mutations is still debated and remains unproven [13, 14]. These patients have distinct tumor biology and may respond differently to immune therapies. While this is a limitation of our study, it also reflects real‐world clinical practice, where treatment decisions are based on patient‐specific factors and evolving clinical judgment. As this is a real‐world study, the inclusion of patients with EGFR or ALK mutations may limit the generalizability of the results. These patients typically do not receive ICI therapy in clinical trials due to the availability of effective targeted therapies. Future studies should focus on this subset of patients to evaluate whether adjuvant ICI therapy provides any additional benefit. Ongoing prospective trials and explorations using patient‐derived organoid models are underway to better understand the role of immunotherapy in EGFR‐ or ALK‐mutant NSCLC, with the aim of developing more effective treatment strategies for this population.
A notable limitation of this study is its retrospective nature, which inherently introduces selection bias. Additionally, the relatively short follow‐up period limits the ability to fully assess long‐term survival outcomes. Given that the effect of adjuvant ICI on OS, particularly in patients who achieve pCR or MPR, remains inconclusive, future prospective studies with longer follow‐up are essential to further elucidate the true benefit of adjuvant immunotherapy in these populations. Unfortunately, quality of life (QoL) data were not systematically collected in this study. Future prospective studies should include QoL assessments to provide a more comprehensive evaluation of the benefits and challenges associated with postoperative adjuvant immunotherapy.
Finally, the results of our study contribute to the growing body of evidence supporting the use of adjuvant ICI therapy in patients with resectable NSCLC, particularly in those who do not achieve pCR or MPR. These findings highlight the need for further investigation into the optimal use of ICI in this setting, including identifying biomarkers that predict which patients will benefit most from adjuvant therapy [15, 16]. Additionally, continued research is essential to define the role of adjuvant ICI in specific subgroups, such as patients with sensitizing mutations or those who achieve pCR, and to determine the optimal sequencing of immunotherapy in the neoadjuvant and adjuvant settings.
5. Conclusion
In patients with resectable NSCLC treated with NAC and immunotherapy, postoperative adjuvant immunotherapy was associated with improved OS and RFS, in line with results from existing clinical trials. However, for those achieving pCR or MPR, while the median survival was higher in the adjuvant immunotherapy group, no significant differences in OS or EFS were observed. Conversely, in patients who did not achieve pCR or MPR, adjuvant immunotherapy significantly improved EFS, but did not demonstrate a clear benefit in OS. These findings suggest that although adjuvant immunotherapy may delay recurrence, its effect on OS remains uncertain. Further studies are warranted to explore the long‐term survival benefits of adjuvant immunotherapy, particularly in patients with incomplete pathological responses.
Author Contributions
Ming Li: conceptualization, methodology, software, data curation, investigation, funding acquisition, writing – original draft, formal analysis, supervision, visualization. Hao Yin: conceptualization, writing – original draft, software, data curation, validation, investigation, formal analysis, visualization. Yue Jin: validation, conceptualization, methodology, data curation, project administration, investigation, resources. Hari B. Keshava: conceptualization, methodology, writing – review and editing, project administration. Rongkui Luo: project administration, resources, data curation, investigation, methodology. Mingxiang Feng: conceptualization, methodology, project administration, resources. Fenghao Sun: conceptualization, writing – review and editing, project administration, formal analysis, methodology, resources, data curation, investigation.
Funding
This study was supported by the Fellowship of China National Postdoctoral Program for Innovative Talents (No. BX20230083), National Natural Science Foundation of China (No. 82303564), Youth Foundation of “Outstanding Resident Physician” Clinical Postdoctoral Program in Zhongshan Hospital (No. 2024ZYYS‐031), and Fujian Provincial Natural Science Foundation, General Program (No. 2024J011440).
Conflicts of Interest
The authors declare no conflicts of interest.
Acknowledgments
The first author of this study, M.L., and the co‐author, H.B.K., are supported by the International Mentorship Program of the International Association for the Study of Lung Cancer (IASLC).
Li M., Yin H., Jin Y., et al., “A Real‐World Study of Resectable NSCLC Following Neoadjuvant Immunotherapy: Should Postoperative Adjuvant Immunotherapy be Recommended?,” Thoracic Cancer 16, no. 23 (2025): e70195, 10.1111/1759-7714.70195.
The first three authors contributed equally to this article.
Data Availability Statement
The data that support the findings of this study are available from the first author (M.L., e‐mail: li_m14@fudan.edu.cn).
References
- 1. Siegel R. L., Miller K. D., Wagle N. S., and Jemal A., “Cancer Statistics, 2023,” CA: A Cancer Journal for Clinicians 73, no. 1 (2023): 17–48, 10.3322/caac.21763. [DOI] [PubMed] [Google Scholar]
- 2. Goldstraw P., Chansky K., Crowley J., et al., “The IASLC Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groupings in the Forthcoming (Eighth) Edition of the TNM Classification for Lung Cancer,” Journal of Thoracic Oncology 11, no. 1 (2016): 39–51, 10.1016/j.jtho.2015.09.009. [DOI] [PubMed] [Google Scholar]
- 3. Uramoto H. and Tanaka F., “Recurrence After Surgery in Patients With NSCLC,” Translational Lung Cancer Research 3, no. 4 (2014): 242–249, 10.3978/j.issn.2218-6751.2013.12.05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Provencio M., Nadal E., Insa A., et al., “Neoadjuvant Chemotherapy and Nivolumab in Resectable Non‐Small‐Cell Lung Cancer (NADIM): An Open‐Label, Multicentre, Single‐Arm, Phase 2 Trial,” Lancet Oncology 21, no. 11 (2020): 1413–1422, 10.1016/S1470-2045(20)30453-8. [DOI] [PubMed] [Google Scholar]
- 5. Felip E., Altorki N., Zhou C., et al., “Adjuvant Atezolizumab After Adjuvant Chemotherapy in Resected Stage IB‐IIIA Non‐Small‐Cell Lung Cancer (IMpower010): A Randomised, Multicentre, Open‐Label, Phase 3 Trial,” Lancet 398, no. 10308 (2021): 1344–1357, 10.1016/S0140-6736(21)02098-5. [DOI] [PubMed] [Google Scholar]
- 6. Forde P. M., Spicer J., Lu S., et al., “Neoadjuvant Nivolumab Plus Chemotherapy in Resectable Lung Cancer,” New England Journal of Medicine 386, no. 21 (2022): 1973–1985, 10.1056/NEJMoa2202170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Marinelli D., Nuccio A., Di Federico A., et al., “Improved Event‐Free Survival After Complete or Major Pathologic Response in Patients With Resectable NSCLC Treated With Neoadjuvant Chemoimmunotherapy Regardless of Adjuvant Treatment: A Systematic Review and Individual Patient Data Meta‐Analysis,” Journal of Thoracic Oncology 20, no. 3 (2025): 285–295, 10.1016/j.jtho.2024.09.1443. [DOI] [PubMed] [Google Scholar]
- 8. Hellmann M. D., Chaft J. E., W. N. William, Jr. , et al., “Pathological Response After Neoadjuvant Chemotherapy in Resectable Non‐Small‐Cell Lung Cancers: Proposal for the Use of Major Pathological Response as a Surrogate Endpoint,” Lancet Oncology 15, no. 1 (2014): e42–e50, 10.1016/S1470-2045(13)70334-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Travis W. D., Dacic S., Wistuba I., et al., “IASLC Multidisciplinary Recommendations for Pathologic Assessment of Lung Cancer Resection Specimens After Neoadjuvant Therapy,” Journal of Thoracic Oncology 15, no. 5 (2020): 709–740, 10.1016/j.jtho.2020.01.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Pataer A., Kalhor N., Correa A. M., et al., “Histopathologic Response Criteria Predict Survival of Patients With Resected Lung Cancer After Neoadjuvant Chemotherapy,” Journal of Thoracic Oncology 7, no. 5 (2012): 825–832, 10.1097/JTO.0b013e318247504a. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. W. N. William, Jr. , Pataer A., Kalhor N., et al., “Computed Tomography RECIST Assessment of Histopathologic Response and Prediction of Survival in Patients With Resectable Non‐Small‐Cell Lung Cancer After Neoadjuvant Chemotherapy,” Journal of Thoracic Oncology 8, no. 2 (2013): 222–228, 10.1097/JTO.0b013e3182774108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Cascone T., W. N. William, Jr. , Weissferdt A., et al., “Neoadjuvant Nivolumab or Nivolumab Plus Ipilimumab in Operable Non‐Small Cell Lung Cancer: The Phase 2 Randomized NEOSTAR Trial,” Nature Medicine 27, no. 3 (2021): 504–514, 10.1038/s41591-020-01224-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Wu Y. L., Herbst R. S., Mann H., Rukazenkov Y., Marotti M., and Tsuboi M., “ADAURA: Phase III, Double‐Blind, Randomized Study of Osimertinib Versus Placebo in EGFR Mutation‐Positive Early‐Stage NSCLC After Complete Surgical Resection,” Clinical Lung Cancer 19, no. 4 (2018): e533–e536, 10.1016/j.cllc.2018.04.004. [DOI] [PubMed] [Google Scholar]
- 14. Lee C. K., Man J., Lord S., et al., “Clinical and Molecular Characteristics Associated With Survival Among Patients Treated With Checkpoint Inhibitors for Advanced Non‐Small Cell Lung Carcinoma: A Systematic Review and Meta‐Analysis,” JAMA Oncology 4, no. 2 (2018): 210–216, 10.1001/jamaoncol.2017.4427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Bai R., Lv Z., Xu D., and Cui J., “Predictive Biomarkers for Cancer Immunotherapy With Immune Checkpoint Inhibitors,” Biomarker Research 8 (2020): 34, 10.1186/s40364-020-00209-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Passiglia F., Galvano A., Castiglia M., et al., “Monitoring Blood Biomarkers to Predict Nivolumab Effectiveness in NSCLC Patients,” Therapeutic Advances in Medical Oncology 11 (2019): 1758835919839928, 10.1177/1758835919839928. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the first author (M.L., e‐mail: li_m14@fudan.edu.cn).