Postoperative recurrence defines a 3-year high-risk window in patients with EGFR-mutant lung adenocarcinoma: a multicenter cohort study

Introduction

Background

Epidermal growth factor receptor (EGFR) mutations confer marked sensitivity to tyrosine kinase inhibitors (TKIs) in patients with advanced non-small cell lung cancer (NSCLC), and EGFR-TKIs have substantially improved clinical outcomes in this population (1). However, in the resectable setting, the prognostic implications of EGFR mutation status vary according to pathological stage. While stage I EGFR-mutant (EGFR-mt) tumors are generally associated with favorable outcomes (2), patients with stage II–III disease remain at substantial risk of postoperative recurrence (3). Moreover, the benefit of conventional platinum-based adjuvant chemotherapy in patients with EGFR-mt disease appears limited, with inconsistent effects on long-term survival (4).

Adjuvant EGFR-TKIs have been investigated to address this unmet need. Although first-generation TKIs administered for fixed durations did not demonstrate clear survival benefits (5,6), more recent trials have shown improved outcomes with newer agents (7,8). These advances have reshaped the postoperative management of EGFR-mt NSCLC. However, the relationship between treatment duration and the underlying temporal distribution of recurrence risk remains unclear. This question has become increasingly important in the current era of adjuvant EGFR-TKI therapy, where treatment duration and surveillance strategies continue to evolve.

Rationale and knowledge gap

Postoperative recurrence is commonly evaluated using time-to-event endpoints, such as recurrence-free survival, which provide a summary measure of risk over the entire follow-up period. However, such approaches do not explicitly account for potential variations in recurrence hazard over time. In many malignancies, recurrence risk is known to exhibit time-dependent patterns rather than remain constant throughout the postoperative course.

Despite its clinical importance, the temporal structure of recurrence in EGFR-mt NSCLC has not been systematically characterized. It remains unclear whether recurrence hazard follows a nonlinear trajectory or whether distinct periods of heightened risk exist. A formal evaluation of such time-dependent dynamics may provide additional insights into the natural history of EGFR-mt disease and support the development of time-adapted postoperative management strategies.

Furthermore, data from treatment-naïve cohorts are essential for delineating the intrinsic recurrence patterns without the confounding effects of adjuvant targeted therapy. However, such datasets are increasingly limited in the current treatment era, highlighting the value of historical surgical cohorts.

Objective

To address this gap, we conducted a multicenter retrospective cohort study of patients with completely resected pathological stage II–III lung adenocarcinoma in the pre-adjuvant TKI era. Using competing-risk methods and Joinpoint regression analysis, we aimed to characterize the temporal dynamics of postoperative recurrence, with the aim of improving the understanding of recurrence patterns in patients with EGFR-mt disease. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1082/rc).

Methods

Study design and patient selection

This multicenter retrospective cohort study included patients from four Japanese institutions (Nagoya University Hospital, National Cancer Center Hospital East, Juntendo University Hospital, and Cancer Institute Hospital, Japanese Foundation for Cancer Research). The study database initially comprised 1,407 patients with pathological stage II or III NSCLC who underwent anatomical lung resection with systematic lymph node dissection between 2010 and 2020 and achieved pathological R0 resection status.

Patients were eligible for the present analysis if they had adenocarcinoma histology and underwent EGFR mutation testing. Finally, 787 patients were included in this study (Figure 1). During the study period, EGFR mutation testing was routinely performed according to institutional practice, primarily in patients with lung adenocarcinoma. Histological classification was based on the World Health Organization criteria (9). This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Institutional Review Board of Cancer Institute Hospital, Japanese Foundation for Cancer Research (approval No. 2020-GA-1303). All participating hospitals were informed of and agreed to this study. The requirement for written informed consent was waived because of the retrospective design of the study and use of anonymized data.

Figure 1 Study cohort selection and classification according to EGFR mutation status. This flow diagram illustrates the derivation of the study cohort from the initial population of patients with resected stage II–III non-small cell lung cancer. Among eligible patients, only those with lung adenocarcinoma who underwent EGFR mutation testing were included and subsequently classified as EGFR-mt or EGFR-wt groups. EGFR, epidermal growth factor receptor; EGFR-mt, EGFR mutant; EGFR-wt, EGFR wild-type.

Patients were categorized as EGFR-mt or EGFR-wild-type (EGFR-wt). EGFR-mt was defined as the presence of exon 18 point mutation (G719X), exon 19 deletions, exon 20 insertions or other exon 20 alterations (including T790M), or exon 21 point mutations (L858R and L861Q). Exon 19 deletions and L858R were classified as common mutations, whereas all others were classified as uncommon mutations. Mutation subtype information was unavailable for a subset of patients; these cases were categorized as “unknown” and were retained in all analyses without imputation. Patients who received induction therapy or had insufficient clinical data were excluded. Analyses were performed using a complete-case approach without imputation (Figure 1).

The clinicopathological variables included age, sex, performance status (PS), Charlson comorbidity index (CCI) (10), smoking history, operative procedure, postoperative complications (Clavien-Dindo classification) (11), adjuvant therapy, and cause of death. Pathological staging was based on the 8th edition of the TNM classification (12). Adjuvant therapy comprised platinum-based chemotherapy, and EGFR-TKIs were not administered in the adjuvant setting.

Postoperative surveillance

Patients were followed up every 6 months for the first 5 years after resection and annually thereafter. Follow-up evaluations included physical examination, chest radiography, blood testing (for tumor markers), and chest and abdominal computed tomography at approximately 6-month intervals for the first 5 postoperative years. Additional assessments were performed when a recurrence was suspected.

Recurrence was primarily assessed based on imaging findings in routine clinical practice, with pathological confirmation obtained when clinically indicated. Time-to-event was calculated from the date of surgery to recurrence, lung cancer-related death, non-cancer-related death, or last follow-up. Recurrent disease and second primary lung cancers were distinguished according to the Martini and Melamed’s criteria (13) and evaluated by a multidisciplinary team of thoracic oncologists.

Statistical analysis

Baseline categorical variables were compared using Fisher’s exact test.

To address potential confounding factors, 1:1 propensity score matching (PSM) was performed using preoperative variables (age, sex, PS, CCI, and smoking history) (2). Matching was restricted to preoperative variables because pathological findings such as pT, pN, and pStage are determined after resection and may lie on the causal pathway between EGFR mutation status and recurrence; therefore, their inclusion in the propensity score model would risk overadjustment. Nearest-neighbor matching was performed using a caliper width of 0.2 of the standard deviation of the logit of the propensity score. Covariate balance was assessed using standardized mean differences, with values of <0.10 considered acceptable.

The cumulative incidence of recurrence and lung cancer-related death was estimated using competing-risk methods, treating non-cancer-related death as a competing event. Differences between groups were assessed using Gray’s test (14), and subdistribution hazard ratios were estimated using the Fine-Gray model (15).

Prespecified subgroup analyses were conducted to evaluate the consistency of temporal recurrence patterns according to adjuvant platinum-based chemotherapy and clinical eligibility for cisplatin based on age (≥75 years) and/or renal function (estimated glomerular filtration rate <60 mL/min/1.73 m²) (16). Within the EGFR-mt cohort, subgroup analyses were also performed according to pathological stage (stage II vs. III) and mutation subtype (common vs. uncommon). These analyses were exploratory and not powered for formal comparison.

The annualized recurrence rates were calculated by dividing the number of recurrence events by the number of at-risk patients in each postoperative year. Joinpoint regression was used to model temporal changes in recurrence hazard and identify inflection points. Segment-specific annual percentage changes were estimated. A maximum of one Joinpoint was allowed a priori to reduce the risk of overfitting, given the number of recurrence events and the exploratory nature of the temporal hazard analysis.

An exploratory interval-specific subdistribution hazard analysis using the Fine-Gray model was performed to assess the temporal variation in the relative recurrence risk (15). Separate models were fitted for each postoperative yearly interval, including only patients at risk at the beginning of each interval.

All patients were censored on July 31, 2021. Statistical analyses were performed using EZR (version 1.70) (17), and Joinpoint regression was conducted using the Joinpoint Regression Program (version 5.4.0) (18). Two-sided P values of <0.05 were considered statistically significant.

Results

Patient characteristics

In total, 787 patients (56% of the initial 1,407-patient cohort) with completely resected pathological stage II–III lung adenocarcinoma who underwent EGFR mutation testing were included after applying eligibility criteria. Of these 787 patients, 492 had EGFR-wt tumors and 295 had EGFR-mt tumors. The median follow-up duration was 62 months.

The baseline characteristics of the groups differed (Table 1). Patients with EGFR-mt tumors were more likely to be female and light smokers or nonsmokers, to have fewer comorbidities, and to have smaller tumors; however, pathological nodal involvement was more frequent in the EGFR-mt group than in the EGFR-wt group. Other clinicopathological characteristics, including pathological stage distribution, were comparable. Among patients with EGFR-mt tumors, common mutations accounted for 72.5% of cases, whereas mutation subtype information was unavailable for 18.7% of cases.

Cumulative incidence of lung cancer recurrence and death

The 5-year cumulative incidence of lung cancer recurrence did not differ significantly between the EGFR-mt and EGFR-wt groups (53% vs. 44%; P=0.20), although the numerically higher rate in the EGFR-mt group was consistent with the baseline differences in nodal involvement. Similarly, the 5-year lung cancer-related death rate was comparable between the groups (29% vs. 32%; P=0.40; Figure 2A,2B).

Figure 2 Cumulative incidence of recurrence and lung cancer-related death according to EGFR mutation status. Although the overall cumulative incidence of recurrence and lung cancer-related death was comparable between EGFR-mt and -wt tumors, the shape of the recurrence curves suggests a difference in temporal dynamics. (A,B) The cumulative incidence in the overall cohort. (C,D) The propensity score-matched cohort. Non-cancer-related death was treated as a competing event. CI, confidence interval; EGFR, epidermal growth factor receptor; EGFR-mt, EGFR mutant; EGFR-wt, EGFR wild-type; HR, hazard ratio.

In the EGFR-mt group, the cumulative incidence of recurrence increased over time, with a slower rate of increase observed beyond approximately 3 years postoperatively. The lung cancer-related death curves initially differed between the groups but converged after approximately 5 years.

PSM using age, sex, PS, CCI, and smoking history yielded 240 well-balanced pairs (Table 1). After matching, the cumulative incidence of recurrence (P=0.50) and lung cancer-related death (P=0.56) remained similar between the groups. Temporal patterns were unchanged, with a decline in recurrence hazard in the EGFR-mt group at approximately year 3 and convergence of lung cancer-related death curves near year 5 (Figure 2C,2D).

Subgroup analyses

Subgroup analyses revealed consistent temporal recurrence patterns in EGFR-mt disease across clinically relevant groups (Figures 3 and 4). An early concentration of recurrence events followed by attenuation beyond 3 years was observed in patients who received adjuvant platinum chemotherapy, those ineligible for cisplatin, those with common or uncommon EGFR mutations, and those with pathological stage II or III disease. Similar patterns were observed in patients with unknown EGFR mutation subtypes, although these analyses were exploratory (Figure S1, Table S1).

Figure 3 Consistency of temporal recurrence patterns across clinically relevant subgroups. The cumulative incidence curves demonstrate a similar temporal pattern of recurrence across subgroups defined by adjuvant platinum chemotherapy (A) and cisplatin eligibility (B). In each subgroup, recurrence events appear to be concentrated in the early postoperative period, followed by attenuation over time. CI, confidence interval; EGFR, epidermal growth factor receptor; EGFR-mt, EGFR mutant; EGFR-wt, EGFR wild-type; HR, hazard ratio.

Figure 4 Temporal recurrence patterns within the EGFR-mt cohort. Within the EGFR-mt group, similar temporal patterns of recurrence were observed across pathological stage (A) and mutation subtype (B). The concentration of recurrence events in the early postoperative period followed by a decline over time was consistently observed across subgroups. CI, confidence interval; EGFR, epidermal growth factor receptor; EGFR-mt, EGFR mutant; HR, hazard ratio.

Joinpoint regression analysis and time-dependent comparison of recurrence risk

Joinpoint regression and interval-specific competing-risk analyses were performed to evaluate temporal changes in recurrence hazards (Figure 5A,5B).

Figure 5 Time-structured recurrence hazard in EGFR-mt lung adenocarcinoma. (A) Joinpoint regression analysis revealed a nonlinear temporal pattern, with an inflection point at approximately 3 years after surgery, followed by a marked decline in hazard. In contrast, no clear inflection point was observed in patients with EGFR-wt tumors. (B) Interval-specific competing-risk analysis shows that the relative recurrence risk between EGFR-mt and -wt tumors does not significantly differ across postoperative time intervals, indicating that the observed pattern reflects differences in temporal structure rather than magnitude of risk. Error bars indicate 95% confidence intervals. APC, annual percent change; EGFR, epidermal growth factor receptor; EGFR-mt, EGFR mutant; EGFR-wt, EGFR wild-type.

In patients with EGFR-mt tumors, a significant inflection point was identified approximately 3 years after surgery, after which the annualized recurrence hazard declined. Before the Joinpoint, the annual percent change in recurrence hazard was −7.45%, whereas after the Joinpoint, it was −34.64% (P=0.002 for the change in slope).

In contrast, no significant inflection point was identified in patients with EGFR-wt tumors, and the recurrence hazard curve did not show a distinct temporal shift. Moreover, the subdistribution hazard ratios for recurrence did not differ significantly between the EGFR-mt and EGFR-wt tumors across the postoperative time intervals, with all 95% confidence intervals crossing unity.

Discussion

Key findings

The principal finding of this study was that postoperative recurrence in patients with resected stage II–III EGFR-mt lung adenocarcinoma followed a time-dependent pattern, with recurrence hazard being concentrated within the first 3 postoperative years and attenuating thereafter. This transition, objectively identified using Joinpoint regression, was not observed in patients with EGFR-wt tumors, suggesting that this temporal pattern reflects an EGFR-mt-specific temporal pattern rather than a general postoperative recurrence phenomenon.

The temporal shift in EGFR-mt disease was consistently observed across clinically relevant subgroups, including patients who received adjuvant platinum chemotherapy, those ineligible for cisplatin, and those with various EGFR mutation subtypes. Patients with unknown mutation subtypes showed similar recurrence curves, suggesting that missing subtype data did not materially influence the observed dynamics. The reproducibility of this pattern suggests that recurrence timing in patients with EGFR-mt disease may reflect underlying disease characteristics rather than treatment-related factors alone.

Interval-specific competing-risk analyses revealed no significant difference in recurrence risk between patients with EGFR-mt and those with EGFR-wt tumors at any postoperative interval, supporting the interpretation that differences in recurrence curves are primarily driven by temporal dynamics within the EGFR-mt group. Overall, these findings provide a descriptive temporal framework for understanding postoperative recurrence risk, with potential implications for postoperative management.

Strengths and limitations

This study had several strengths. The multicenter design across four Japanese institutions yielded a substantial sample of 787 patients with pathological stage II–III lung adenocarcinoma, all treated in the pre-adjuvant TKI era, providing a dataset that is increasingly difficult to obtain prospectively. The use of competing-risk methods appropriately addressed the competing nature of lung cancer-related death and recurrence, and Joinpoint regression enabled objective identification of temporal inflection points without imposing predetermined thresholds. PSM confirmed that the observed temporal patterns were robust to baseline differences between the EGFR-mt and -wt groups.

This study also had limitations. First, the multicenter design may have introduced variability in postoperative management among institutions. Differences in surveillance and imaging protocols may have influenced recurrence detection and cause-of-death classification. In addition, PSM was restricted to preoperative variables; therefore, residual imbalance in pathological tumor burden, aggressiveness, and microscopic invasive features cannot be excluded. Furthermore, invasive histological features associated with aggressive recurrence behavior, including lymphovascular invasion, visceral pleural invasion, and spread through air spaces (STAS), were not uniformly available across participating institutions and therefore could not be incorporated into the present analyses. Imbalances in these pathological factors may have influenced recurrence timing patterns and represent additional sources of residual confounding. Postoperative treatment-related variables, including heterogeneity in adjuvant chemotherapy regimens, treatment completion rates, dose intensity, and salvage therapies after recurrence, were also not uniformly available because of the retrospective multicenter design and long study period. These factors may have influenced recurrence dynamics and survival outcomes and therefore represent further potential sources of residual confounding. However, because the primary aim was to characterize temporal recurrence dynamics rather than absolute recurrence rates, these sources of variability are unlikely to have materially affected the overall temporal patterns observed in this study. Second, site-specific recurrence patterns could not be systematically evaluated. Third, EGFR mutation subtype information was unavailable in a subset of patients, which may have introduced additional molecular heterogeneity and limited more detailed subtype-specific interpretation. In addition, some subgroup analyses were limited by sample size. Finally, decreasing numbers at risk in later time periods may have affected the stability of hazard estimates. In addition, Joinpoint regression was intentionally restricted to a maximum of one inflection point to reduce the risk of overfitting given the number of recurrence events and the exploratory nature of the temporal hazard analysis. Therefore, smaller secondary fluctuations in recurrence hazard may not have been captured, and more flexible continuous hazard modeling approaches may further refine characterization of postoperative recurrence dynamics in future larger-scale studies. Generalizability beyond Japanese populations should be interpreted with caution.

Comparison with similar research

Prior studies have established that EGFR-mt stage II–III disease is associated with higher recurrence risk than stage I disease (2,3) and that platinum-based adjuvant chemotherapy provides limited long-term benefit in this population (4). Randomized trials of first-generation adjuvant EGFR-TKIs, including the IMPACT and CTONG1104 trials, did not demonstrate significant improvement in overall survival with 2-year treatment regimens (5,6). In contrast, subsequent trials using newer agents have reported improved outcomes with longer treatment durations (7,8).

The present study complements these findings by providing a formal characterization of the temporal dynamics of postoperative recurrence in a TKI-naïve cohort. While prior studies have primarily reported aggregate survival outcomes, our analysis suggests that recurrence in patients with EGFR-mt disease is not temporally uniform but follows a time-dependent pattern. This temporal characterization may provide an additional perspective for interpreting differences observed across adjuvant treatment strategies, although such interpretations remain hypothesis-generating.

Explanations of findings

The present findings indicate that postoperative recurrence in patients with EGFR-mt NSCLC is concentrated in the early postoperative period and attenuates thereafter, suggesting a non-uniform temporal distribution of recurrence risk. One possible explanation is that micrometastatic disease present at the time of surgery contributes to early recurrence events, whereas later recurrence may reflect differences in recurrence dynamics over time. Histological invasive features associated with postoperative dissemination, including STAS, may also contribute to aggressive early recurrence behavior in patients with EGFR-mt tumors. Recent pathological studies have suggested that STAS is associated with frequent postoperative recurrence in patients with EGFR-mt lung adenocarcinoma (19). However, these interpretations remain speculative, as invasive histological features such as STAS were not systematically available in the present multicenter cohort and therefore could not be directly evaluated in this study.

Recognition of time-dependent recurrence risk may have implications for postoperative management. Current guidelines do not stratify surveillance strategies according to EGFR mutation status, and the observed temporal pattern may provide a basis for exploring risk-adapted follow-up approaches in future studies.

The recurrence dynamics identified in this study may also support further investigation into integrated risk stratification approaches. The incorporation of molecular residual disease assessment or circulating tumor DNA monitoring into prospective studies may help refine our understanding of recurrence timing and factors associated with postoperative recurrence patterns (20,21). Overall, temporal patterns and molecular biomarkers may inform future individualized postoperative management strategies, as explored in ongoing studies (22).

Implications and actions needed

From a clinical perspective, these findings suggest that postoperative recurrence risk in patients with EGFR-mt disease is not constant over time and may be concentrated within an early postoperative period. This observation may have implications for the design of postoperative management strategies, including both surveillance and treatment approaches.

While randomized trials have established the efficacy of adjuvant EGFR-TKIs (7,8), the relationship between treatment duration and the natural temporal distribution of recurrence risk remains uncertain. Our findings raise the possibility that time-dependent recurrence dynamics could be considered in future studies evaluating postoperative treatment strategies. Whether aligning treatment duration with periods of higher recurrence risk may be relevant for hypothesis generation in future studies, and whether such alignment translates into clinical benefit remains to be determined.

Although causal inference cannot be established from this observational study, these results provide a framework for considering time-adapted approaches to postoperative management. Further validation in prospective studies will be required to determine the clinical applicability of this concept.

Conclusions

In this treatment-naïve multicenter cohort, postoperative recurrence in patients with resected stage II–III EGFR-mt lung adenocarcinoma was predominantly concentrated within the first 3 postoperative years, with a significant inflection in recurrence hazard at approximately 3 years and marked attenuation thereafter. This time-structured pattern was not observed in patients with EGFR-wt tumors, suggesting a disease-specific temporal pattern rather than a general postoperative recurrence phenomenon. These findings characterize the temporal pattern of postoperative recurrence in patients with EGFR-mt disease, support consideration of mutation-status-aware postoperative surveillance during the early postoperative period, and provide a descriptive framework for future studies investigating postoperative management strategies, including adjuvant EGFR-TKI therapy.

Acknowledgments

We would like to thank Editage for the English language editing.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1082/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1082/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1082/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1082/coif). The 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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Institutional Review Board of Cancer Institute Hospital, Japanese Foundation for Cancer Research (approval No. 2020-GA-1303). All participating hospitals were informed of and agreed to this study. The requirement for written informed consent was waived because of the retrospective design of the study and use of anonymized data.

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. Lynch TJ, Bell DW, Sordella R, et al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 2004;350:2129-39. [Crossref] [PubMed]
2. Suda K, Mitsudomi T, Shintani Y, et al. Clinical Impacts of EGFR Mutation Status: Analysis of 5780 Surgically Resected Lung Cancer Cases. Ann Thorac Surg 2021;111:269-76. [Crossref] [PubMed]
3. Ito M, Miyata Y, Tsutani Y, et al. Positive EGFR mutation status is a risk of recurrence in pN0-1 lung adenocarcinoma when combined with pathological stage and histological subtype: A retrospective multi-center analysis. Lung Cancer 2020;141:107-13. [Crossref] [PubMed]
4. Isaka T, Ito H, Yokose T, et al. Prognostic factors for relapse-free survival in stage IB-IIIA primary lung adenocarcinoma by epidermal growth factor receptor mutation status. BMC Cancer 2022;22:966. [Crossref] [PubMed]
5. Tada H, Mitsudomi T, Yamanaka T, et al. Adjuvant gefitinib versus cisplatin/vinorelbine in Japanese patients with completely resected, EGFR-mutated, stage II-III non-small cell lung cancer (IMPACT, WJOG6410L): A randomized phase 3 trial. J Clin Oncol 2021;39:8501.
6. Zhong WZ, Wang Q, Mao WM, et al. Gefitinib Versus Vinorelbine Plus Cisplatin as Adjuvant Treatment for Stage II-IIIA (N1-N2) EGFR-Mutant NSCLC: Final Overall Survival Analysis of CTONG1104 Phase III Trial. J Clin Oncol 2021;39:713-22. [Crossref] [PubMed]
7. Herbst RS, Wu YL, John T, et al. Adjuvant Osimertinib for Resected EGFR-Mutated Stage IB-IIIA Non-Small-Cell Lung Cancer: Updated Results From the Phase III Randomized ADAURA Trial. J Clin Oncol 2023;41:1830-40. [Crossref] [PubMed]
8. Tsuboi M, Herbst RS, John T, et al. Overall Survival with Osimertinib in Resected EGFR-Mutated NSCLC. N Engl J Med 2023;389:137-47. [Crossref] [PubMed]
9. Nicholson AG, Tsao MS, Beasley MB, et al. The 2021 WHO Classification of Lung Tumors: Impact of Advances Since 2015. J Thorac Oncol 2022;17:362-87. [Crossref] [PubMed]
10. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373-83. [Crossref] [PubMed]
11. Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg 2009;250:187-96. [Crossref] [PubMed]
12. Chansky K, Detterbeck FC, Nicholson AG, et al. The IASLC Lung Cancer Staging Project: External Validation of the Revision of the TNM Stage Groupings in the Eighth Edition of the TNM Classification of Lung Cancer. J Thorac Oncol 2017;12:1109-21.
13. Martini N, Melamed MR. Multiple primary lung cancers. J Thorac Cardiovasc Surg 1975;70:606-12.
14. Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Statist 1988;16:1141-54.
15. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496-509.
16. Noritake O, Nakamura S, Kinoshita F, et al. Prognostic impact of adjuvant therapy for cisplatin-unfit patients with non-small-cell lung cancer: A multicenter analysis. Lung Cancer 2024;188:107470. [Crossref] [PubMed]
17. Kanda Y. Investigation of the freely available easy-to-use software 'EZR' for medical statistics. Bone Marrow Transplant 2013;48:452-8. [Crossref] [PubMed]
18. Kim HJ, Fay MP, Feuer EJ, et al. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 2000;19:335-51. [Crossref] [PubMed]
19. Lee J, Han YB, Kim S, et al. Differential prevalence and prognostic significance of spread through air spaces according to oncogenic driver mutations in lung adenocarcinoma. J Thorac Cardiovasc Surg 2026;171:1320-1331.e2. [Crossref] [PubMed]
20. Pellini B, Chaudhuri AA. Circulating Tumor DNA Minimal Residual Disease Detection of Non-Small-Cell Lung Cancer Treated With Curative Intent. J Clin Oncol 2022;40:567-75. [Crossref] [PubMed]
21. Herbst RS, John T, Grohé C, et al. Molecular residual disease analysis of adjuvant osimertinib in resected EGFR-mutated stage IB-IIIA non-small-cell lung cancer. Nat Med 2025;31:1958-68. [Crossref] [PubMed]
22. Soo RA, de Marinis F, Han JY, et al. TARGET: A Phase II, Open-Label, Single-Arm Study of 5-Year Adjuvant Osimertinib in Completely Resected EGFR-Mutated Stage II to IIIB NSCLC Post Complete Surgical Resection. Clin Lung Cancer 2024;25:80-4. [Crossref] [PubMed]