Abstract
Background/Aim: Recurrence rates in early and locally advanced non-small-cell lung cancer (NSCLC) remain high despite curative treatment. Recently, the survival benefit of immune checkpoint inhibitors (ICI) in the (neo)adjuvant setting in patients with stage II-III NSCLC has been demonstrated. This study aimed to identify predisposing factors for disease recurrence to select patients who would benefit from multimodality treatment. Patients and Methods: This retrospective observational study included patients with stage I-IIIA NSCLC discussed at the Thoracic Multidisciplinary Tumour Board of the University Hospital, Brussels, between 2017 and 2021. Results: Of the 167 patients, 34% had a recurrence, with a median time to recurrence of 9.1 months [272 (interquartile range=175-621.5) days]. The highest recurrence rate (56.5%) was observed in cTNM stage IIIA. Of surgical patients who were not eligible for (neo)adjuvant ICI according to current European reimbursement criteria, 21.7% developed disease recurrence. Twelve out of 20 patients eligible for ICI had no recurrence at a median follow-up of 34.1 months and would have been overtreated if they had received ICI therapy. Treatment modality and TNM stage were significantly associated with recurrence and worse progression-free survival (p<0.05). Stereotactic body radiotherapy, higher TNM stage and the presence of serine/threonine kinase 11 (STK11) mutation were significantly associated with worse overall survival. Conclusion: European reimbursement criteria for (neo)adjuvant ICI in surgical patients are based on TNM stage (T≥4 cm or N1/N2 disease). However, TNM stage alone does not give the full picture. In patients undergoing surgery, the presence of the STK11 mutation was significantly associated with worse overall survival. We suggest the integration of analysis of circulating tumour DNA into perioperative strategies to reduce over- and undertreatment.
- Non-small cell lung cancer
- immune checkpoint inhibitor
- ICI
- relapse risk factors
- treatment strategy
- retrospective data
Patients with early or locally advanced NSCLC (stage I-IIIA) are treated with curative intent. Adjuvant chemotherapy after radical surgery has been shown to reduce the risk of local recurrence with a modest improvement in 5-year survival of approximately 5% (1). Recurrence remains high due to subclinical spread of tumour cells. Lou et al. found that one out of five patients developed cancer recurrence after surgery (2). Research has shown that initial TNM stage, treatment modality and pathological diagnosis have an impact on the risk of recurrence (1, 2). In the last few years, several trials using immune checkpoint inhibitors (ICI) in the adjuvant, neo-adjuvant and the perioperative setting demonstrated a beneficial effect in certain patients with early-stage or locally advanced resectable NSCLC. ICIs increase the rate of pathological complete response and major pathological response compared to chemotherapy with increased disease-free survival/event-free survival. The KEYNOTE-671 trial was the first phase III study in the perioperative setting to demonstrate a statistically significant improvement in overall survival (OS) [hazard ratio (HR)=0.72, p=0.005] (3). Different regimens in the adjuvant and neo-adjuvant setting are now reimbursed for patients with stage II-IIIA NSCLC. Although ICIs are generally well tolerated by patients, serious immune-related side-effects can occur and can be life-threatening. Exposing a patient who may have been cured of their NSCLC to these risks must be a well-considered decision. Moreover, immunotherapy has a high price tag. Therefore, appropriate patient selection is important. The aim of this study was to provide a global overview of our group of patients with early or locally advanced NSCLC prior to the use of ICI in the perioperative setting and to identify factors predisposing for disease recurrence that may help to select the patient population that would benefit from multimodality treatment.
Patients and Methods
Study design and population. This was a retrospective observational study. Data were collected from the medical records of patients treated at the University Hospital of Brussels between the 1st January 2017 and 31st December 2021. Adult patients with stage I to IIIA NSCLC (8th TNM classification) were retained for further analysis (4). Exclusion criteria were: Other malignancy, histological small-cell carcinoma or neuroendocrine malignancy, death before treatment, or no treatment started, concurrent tumours where it was unclear which tumour caused the recurrence, and loss to follow-up. The following patient and tumour characteristics were collected: age, sex, smoking status, Eastern Cooperative Oncology Group (ECOG) performance status, cardiovascular history, history of previous malignancy, forced expiratory volume (FEV1), forced vital capacity, histology, clinical and (if available) pathological TNM stage, programmed death-ligand 1 (PD-L1) expression and driver mutational status. Treatment modality and any adjuvant therapy were also recorded. If surgery was performed, the following variables were collected: Type of surgery, R resection, lymphatic, vascular or perineural invasion. Time to recurrence was defined as the time from the start of treatment or the date of surgery to the first radiological sign of recurrence.
Study objectives. The primary endpoint was disease recurrence (locoregional or distant) and its predisposing factors. The secondary endpoints were progression-free survival (PFS), defined as the time from diagnosis to disease progression, recurrence or death, and OS.
Statistical analysis. Analyses were performed using MedCalc (version 22, MedCalc Software Ltd., Ostend, Belgium). Baseline characteristics were analysed using Mann–Whitney U-test and chi-squared test. Logistic regression analysis and stratified multivariable Cox proportional hazards modelling were used to identify predictors of defined outcomes. PFS and OS were analysed using the Kaplan–Meier method. Statistical significance was set at a two-tailed value of p<0.05.
Results
Clinical characteristics. A total of 188 patients met the inclusion criteria, an additional 21 patients were excluded due to loss to follow-up or death from other causes. Patient characteristics are shown in Table I. The study population was predominantly male (62.9%) with a mean age of 69.0 (range=42-89) years. The majority were former smokers (56.3%) with a mean of 31.2 pack-years. The mean forced expiratory volume was 2.25 l and the mean forced vital capacity was 3.29 l. The majority had an ECOG 0 or 1 performance score (92.2%). There was an almost even distribution with regard to cardiovascular history, with 87 (52.1%) individuals without and 80 (47.9%) with such history. History of previous malignancy was assessed and almost a third (31.1%) had a history of cancer. The majority had an ECOG performance score of 0 or 1 (92.2%). The most common clinical TNM stage at diagnosis was IA2 (22.8%), followed by IB (14.4%), IA3 (13.8%) and IIIA (13.8%). Most patients underwent surgery (61.7%). The majority underwent lobectomy (78.7%), while less than a fifth underwent sublobar resection (19.4%) and only two patients underwent pneumonectomy (1.9%). R0 resection was achieved in 92% of cases. Lymphovascular invasion was present in half of the cases (51%), while perineural invasion was present in only 11%. The second most common treatment was stereotactic body radiotherapy (SBRT, 28.7%). Adenocarcinoma was the most common histopathological finding (64.7%). PD-L1 status was assessed in 137 patients (82.0%), of whom 43.1% were PD-L1 negative, 29.9% and 27.0% of patients had a PD-L1 expression of 1-50% and >50%, respectively. Next-generation sequencing was used to identify the most common mutations in 129 patients (77.2%). Approximately 30.2% had a Kirsten rat sarcoma viral oncogene homolog (KRAS) mutation and 41.1% were tumor protein p53 (TP53)-positive, while 14.7% had an epidermal growth factor receptor (EGFR) mutation. Other notable mutations were serine/threonine kinase 11 (STK11) (8/129), phosphatidylinositol-4,5-biphosphate 3-kinase, catalytic subunit alpha (PIK3CA) (9/129), V-raf murine sarcoma viral oncogene homolog B1 (BRAF) (9/129), anaplastic lymphoma kinase (ALK) (5/129) and MET proto-oncogene, receptor tyrosine kinase (MET) exon 14 (5/129). Three cases with non-receptor tyrosine kinases (NRTK) mutations and one with ROS proto-oncogene (ROS1) mutation were detected.
Baseline clinical characteristics of study patients overall and by no recurrence versus recurrence.
Time, pattern and risk of recurrence. Median (interquartile range) follow-up time was 34.1 months [1,023 (696-1,505) days]. A total of 57 patients (34%) had a recurrence event, with a median time to recurrence of 9.1 months [272 (175-621.5) days]. In 36 out of these 57 patients (63%), the recurrence occurred within 1 year of follow-up and was defined as early recurrence. Half of the patients had locoregional recurrence (49%) and half had distal recurrence (51%). The most common sites of distal recurrence were the contralateral lung and adrenal gland.
Factors associated with recurrence. Patient-related factors: Sex, age, smoking status, cardiovascular history, and history of malignancy were not significantly associated with disease recurrence. An association was found between recurrence and performance status (p=0.01), with an odds ratio (OR) of 4.73 [95% confidence interval (CI)=1.28-17.45] for ECOG 2 (p=0.02) compared to ECOG 0.
Tumour-related factors: No significant association was found between recurrence and histology, PD-L1 expression and the (driver) mutational status. TNM stage was significantly associated with recurrence (p=0.005); in particular, patients with stage IIB and IIIA had a high risk of recurrence with ORs of 3.75 (95% CI=1.16-12.12, p=0.02) and 4.87 (95% CI=1.56-15.16, p=0.006), respectively. For T stage alone, there was a trend towards an association with recurrence (p=0.08). N2 disease was associated with a significant risk of recurrence with an OR of 13.43 (95% CI=1.57-114.80, p=0.01) compared to N0. Patients without vascular invasion tended to have a reduced risk of recurrence with OR 0.42 (95% CI=0.15-1.17, p=0.09).
Treatment-related factors: Compared to surgery, SBRT was significantly associated with recurrence, with an OR of 2.54 (95% CI=1.21-5.27, p=0.0127). Surgical patients were subdivided into those who underwent sublobar resection, lobectomy and pneumonectomy, which did not correlate with disease recurrence (p=0.62).
Patients at high risk of disease recurrence after surgery, defined as those with T ≥4 cm or N1/N2 disease, are now eligible for adjuvant or neo-adjuvant immunotherapy. Under the current European Medicines Agency reimbursement criteria, 20 of our patients (19.4%) would be eligible for such therapy based on TNM stage, of whom eight had recurrence (40%). This also means that 12 out of 20 patients (60%) without a recurrence event at a median follow-up of 34.1 months in this trial would have been treated with immunotherapy under this criterion, and potentially have been overtreated. Of the patients who did not meet reimbursement criteria, 18 patients (21.7%) had a recurrence.
Survival. Mean PFS was shorter in patients who received SBRT compared to those who underwent surgery (44 versus 59.6 months; HR=2.8, 95% CI=1.48-5.28, p=0.0014; Figure 1A). In addition, patients with a TNM stage of IIA, IIB or IIIA had an increased hazard of recurrence compared to stages below IIA (HR=2.52, 95% CI=1.39-4.56, p=0.0022; Figure 1B). The mean PFS for patients with TNM IIA-IIIA was 42.8 months, compared to 57.7 months for those with TNM 0-IB.
Kaplan–Meier curves of the progression-free survival per significantly associated factors. (A) Progression-free survival by treatment modality. Stereotactic body radiotherapy (SBRT) vs. surgery: Hazard ratio (HR)=2.8, 95% confidence interval (CI)=95% CI=1.48-5.28, p=0.0014. (B) Progression-free survival by TNM stage. IIA-IIIA vs. 0-IB: HR=2.52, 95% CI=1.39-4.56, p=0.0022.
Patients undergoing SBRT had worse OS compared to those who underwent surgery (HR=6.62, 95% CI=3.69-11.87, p<0.0001; Figure 2A). TNM stage of IIA or higher was associated with worse OS (HR=1.91, 95% CI=1.13-3.21, p=0.0146; Figure 2B). For the patients undergoing surgery, a higher TNM stage was associated with worse OS (HR=2.85, 95% CI=1.00-8.12, p=0.0491), but the presence of a STK11 mutation was also associated with worse OS in this selected group (HR=50.8, 95% CI=10.10-255.36, p<0.0001; Figure 3).
Kaplan–Meier curves of the overall survival by significantly associated factors. (A) Overall survival by treatment modality. Stereotactic body radiotherapy (SBRT) vs. surgery: Hazard ratio (HR)=6.62, 95% confidence interval (CI)=3.69-11.87, p<0.0001. (B) Overall survival by TNM stage. IIA-IIIA vs. 0-IB: HR=1.91, 95% CI=1.13-3.21, p=0.0146.
Kaplan–Meier curve of overall survival in surgical patients by serine/threonine kinase 11 (STK11) mutation. STK11 mutation vs. no mutation: hazard ratio=50.8, 95% confidence interval=10.10-255.36, p<0.0001.
Discussion
In this study, we investigated the risk of recurrence and its predisposing factors in early and locally advanced NSCLC. Overall, 34% of the study population had a recurrence, which is comparable to rates reported in the literature (2, 5-9). The observed median time to recurrence of 9.1 months was shorter in this study compared to other studies. This may be due to the fact that those studies only included surgical patients (5, 7, 8). In our study, 63% of the patients had early recurrence. The rates of distant and locoregional recurrence observed were almost equal, whereas in the literature the predominant pattern is distant (5-7).
We found a significant association between recurrence and poor performance status. It is known that the initial TNM stage at diagnosis is associated with the risk of recurrence, which is confirmed in this study, with nodal stage being the most important factor (5, 6, 8, 9). In our study, SBRT was significantly associated with a higher risk of recurrence compared to surgery, with an OR of 2.7, confirming surgery as the best treatment option. There was no significant correlation between recurrence and the type of surgery (10-12). Saji et al. studied segmentectomy versus lobectomy in small-sized peripheral NSCLCs (tumour ≤2 cm) and showed more locoregional recurrences in the segmentectomy group, but no significant difference in overall recurrence-free survival.
Looking at multivariate analysis for PFS and OS, the importance of TNM stage and treatment modality, especially surgery, was highlighted.
Current trials investigating the role of (neo)adjuvant immunotherapy in early-stage NSCLC, such as adjuvant (IMPOWER010), neo-adjuvant (Checkmate 816) and perioperative (AEGAN, Keynote 671) are primarily based on TNM stage and the presence of PD-L1 expression (3, 13-15). However, questions have been raised as to whether TNM stage and PD-L1 expression are the only possible (bio)markers to distinguish who will benefit from an ICI. In our study, 40% of patients eligible for adjuvant or neoadjuvant immunotherapy based on TNM stage had a disease relapse and 21.7% of patients relapsed while they did not meet TNM criteria.
In an effort to identify other predisposing factors for disease recurrence, we investigated the impact of driver mutations. In the small cohort of surgical patients, a significant correlation was found between STK11 mutation and reduced OS. The association of STK11 mutations with poor effectiveness of anti-PD-1/L1 therapies and worse clinical outcomes in NSCLC has already been described (10, 16). Tsukasa et al. reported reduced efficacy of pembrolizumab in patients with low STK11 expression in advanced NSCLC, despite high levels of PD-L1 expression. However, the correlation between expression and mutation is not known. In several studies, TP53 mutation was also associated with poor clinical outcome and increased risk of developing new distant metastases (17-20). This was not confirmed in our study, but our study population was rather small. Further research is needed to quantify the role of these mutations and the impact of co-mutations in the response to ICIs, as this may have implications for perioperative management in early and locally advanced NSCLC.
The detection of circulating tumour DNA is another way to better select patients with early NSCLC who are more likely to benefit more from adjuvant immunotherapy treatment. It has been shown that positivity for circulating tumour DNA after curative treatment may indicate an increased risk of recurrence (11, 12).
Finally, this study has several limitations. Firstly, it was a relatively small retrospective study conducted at a single academic institution, which introduces bias in the selection of patients and the information collected. Secondly, driver mutation status and PD-L1 expression were not known for the whole group.
Conclusion
Current European reimbursement criteria for (neo)adjuvant ICI in surgical patients are based on TNM stage (T≥4 cm or N1/N2 disease). However, TNM stage alone does not provide the full picture. As many as 21.7% of surgical patients who were not eligible for (neo)adjuvant ICI according to the current criteria still developed disease recurrence. Our study showed that beside TNM stage, STK11 mutation was associated with worse overall survival in surgical patients, highlighting the importance of integrating these mutations into treatment strategies. In addition, we suggest the integration of circulating tumour DNA detection into perioperative strategies to reduce over- and undertreatment.
Footnotes
Authors’ Contributions
TC: Conceptualization, formal analysis, investigation, writing - original draft, visualization. VK: conceptualization, Methodology, Validation, Writing: review and editing, Supervision.
Conflicts of Interest
The Authors declare that they have no conflicts of interest.
- Received August 31, 2024.
- Revision received October 2, 2024.
- Accepted October 16, 2024.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.









