Survival after wedge resection, segmentectomy and lobectomy for clinical stage IA non-small cell lung a systematic review and network meta-analysis

Authors: Cien Sun, Jiang Jin, Jiawen Chen, Hao Liu, Pasan Witharana, Minghui Yang, Zimin Wang, Ying Zhang, Pengfei Sheng, Yutao Chen, Chengchu Zhu, Jianfei Shen

Abstract

Background

Surgical resection remains the cornerstone of early-stage treatment for non-small cell lung cancer (NSCLC). However, the survival benefits of different surgical methods in clinical stage IA patients remain controversial. This systematic review aims to compare the efficacy of surgical methods—lobectomy, segmentectomy, and wedge resection—on survival outcomes in clinical stage IA NSCLC patients.

Methods

A systematic search was performed in PubMed, Embase, Cochrane Library, Web of Science, and ClinicalTrials.gov databases between January 2000 and November 30, 2024. Studies meeting the inclusion and exclusion criteria were identified, and hazard ratio (HR) for overall survival (OS), disease-free survival (DFS), and recurrence-free survival (RFS) were extracted from each study for pairwise and Bayesian network meta-analyses. This study protocol was registered on PROSPERO (CRD42024618659).

Results

A total of 58 retrospective studies and 3 randomized controlled trials (RCTs) were included. For patients with overall stage IA, network meta-analyses showed that both lobectomy and segmentectomy had significant advantages in OS, DFS, and RFS compared to wedge resection {HROS 0.67 [95% confidence interval (CI): 0.59–0.75], 0.75 (95% CI: 0.65–0.84); HRDFS 0.69 (95% CI: 0.54–0.89), 0.71 (95% CI: 0.55–0.92); HRRFS 0.57 (95% CI: 0.42–0.78), 0.53 (95% CI: 0.39–0.72)}. No significant differences were observed between lobectomy and segmentectomy. Based on ranking probabilities, lobectomy ranked first. In subgroup analyses, results for overall T1a/b patients were consistent with those of stage IA. When 0.5< consolidation-to-tumor ratio (CTR) <1, lobectomy and segmentectomy showed significant advantages in OS and RFS compared to wedge resection [HROS 0.57 (95% CI: 0.38–0.94), 0.52 (95% CI: 0.34–0.84); HRRFS 0.53 (95% CI: 0.32–0.83), 0.51 (95% CI: 0.31–0.84)], while no significant differences were observed otherwise. Segmentectomy ranked first in this group. When CTR =1, no significant differences were found, with lobectomy ranking first. For overall T1c patients, lobectomy demonstrated significant advantages in OS compared to segmentectomy and wedge resection [HR 0.73 (95% CI: 0.60–0.91); HR 0.60 (95% CI: 0.44–0.77)], while no significant differences were observed otherwise. Lobectomy ranked first in this group. For patients with 0.5< CTR <1, no significant differences were found, with lobectomy ranked first.

Conclusions

Lobectomy and segmentectomy provide better OS benefits compared to wedge resection in stage IA NSCLC patients, with no significant differences between lobectomy and segmentectomy. However, the optimal surgical approach should still be determined based on tumor size and CTR.

Introduction

Non-small cell lung cancer (NSCLC) accounts for approximately 85% of newly diagnosed lung cancer cases and remains the leading cause of cancer-related mortality worldwide (1). Surgical resection remains the cornerstone of early-stage NSCLC treatment, as it ensures complete tumor removal while minimizing lung function loss (2). The landmark 1995 LCSG trial historically established lobectomy as the gold standard for T1N0 NSCLC by demonstrating superior outcomes over sublobar resection (3). Subsequent advances in imaging technologies, particularly computed tomography (CT) screening, have enhanced small lesion detection and malignancy prediction (4,5), challenging lobectomy’s gold standard status and renewing consideration of sublobar resection for stage IA NSCLC.

The JCOG0201 trial established tumor size ≤2 cm and consolidation-to-tumor ratio (CTR) ≤0.25 as radiological criteria for non-invasive lung cancer (6), with subsequent 5-/10-year analyses confirming excellent survival for tumors ≤3 cm with CTR ≤0.5 (7-9). The JCOG0804 trial demonstrated that sublobar resection was a feasible and effective treatment option for peripheral ground-glass opacity (GGO)-dominant adenocarcinomas with tumor size ≤2 cm and CTR ≤0.25 (10,11). Meanwhile, the JCOG0802 trial revealed that for patients with peripheral adenocarcinomas with tumor size ≤2 cm and 0.5< CTR ≤1, segmentectomy provided superior overall survival (OS) compared to lobectomy, and recurrence-free survival (RFS) was not inferior to lobectomy (12). Additionally, The JCOG1211 trial supports segmentectomy as standard for tumors ≤3 cm with CTR ≤0.5 (13). Another randomized controlled trial (RCT), CALGB140503, confirmed that for peripheral NSCLC ≤2 cm with pathologically negative hilar and mediastinal lymph nodes, sublobar resection was an effective treatment strategy (14).

Despite these findings, concerns remain regarding the higher local recurrence rates associated with sublobar resection. In the JCOG0802 trial, the local recurrence rate was significantly higher in the segmentectomy group than in the lobectomy group (10.5%* vs. *5.4%, P=0.0018) (12). Professor Hisao Asamura, one of the study’s investigators, noted that segmentectomy is technically challenging, requires longer operative time, increases intraoperative trauma, and is associated with a higher rate of air leaks and perioperative complications. Therefore, the study did not recommend replacing lobectomy with segmentectomy based solely on its findings. Similarly, in the CALGB140503 trial, the local recurrence rate was 13.4% in the sublobar resection group compared to 10% in the lobectomy group, although this difference was not statistically significant (14). Furthermore, in the JCOG0802 trial, while segmentectomy demonstrated an OS advantage over lobectomy for pure solid nodules ≤2 cm (15), these nodules are highly invasive, warranting caution in the application of segmentectomy for such cases. Among sublobar resection techniques, particularly wedge resection, controversies remain, and further studies are required to explore its potential value.

Although previous studies have provided preliminary evaluations of the efficacy of sublobar resection, systematic assessment of wedge resection and detailed stratified analyses based on the CTR remain insufficient. This study systematically compares the survival outcomes of three surgical approaches based on stratified tumor diameter and CTR values, and innovatively develops a surgical decision-making fan chart. We present this article in accordance with the PRISMA-NMA reporting checklist (16) (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-816/rc).

Methods

Search strategy

The review protocol was prospectively registered in PROSPERO with the registration number CRD42024618659. A comprehensive literature search was performed in the Web of Science, PubMed, Cochrane Library, ClinicalTrials.gov, and Embase databases for studies published between January 2000 and November 2024. The search utilized the following “non-small cell lung cancer”, “clinical stage IA”, “segmentectomy”, “lobectomy”, “wedge resection”, “survival”, “overall survival (OS)”, “disease-free survival (DFS)”, and “recurrence-free survival (RFS)”. The detailed search strategy is outlined in Table S1. Additionally, we reviewed the reference lists of recent relevant reviews and meta-analyses to ensure a comprehensive literature search.

Selection criteria

The inclusion criteria for the included studies were as (I) study populations with clinical stage IA NSCLC harmonized to International Association for the Study of Lung Cancer (IASLC) 8th Edition tumor-node-metastasis (TNM) staging; (II) the surgical approaches included at least two or three of the segmentectomy, lobectomy, and wedge resection; (III) the outcomes reported included at least one of the OS, DFS or RFS; (IV) the study design was one of the RCTs, cohort studies, or case-control studies (retrospective or prospective); (V) the study included a sample size of more than 20 patients; and (VI) the study was published in English and was publicly accessible.

The exclusion criteria were as (I) studies not written in English; (II) studies available only as abstracts in databases; (III) studies that did not focus on patients with clinical stage IA NSCLC, or in which participants did not undergo at least two or three of the following surgical lobectomy, segmentectomy, or wedge resection, or where sublobar resections were not further categorized into segmentectomy and wedge resection, or outcomes that lacked extractable data on OS, DFS, or RFS; (IV) studies with low readability or credibility.

Literature screening and data extraction

The first step in screening studies involved using Endnote21 to remove duplicates. Two researchers (C.S. and J.C.) independently screened the abstracts and full texts to determine whether the studies provided relevant data based on the inclusion and exclusion criteria. Any disagreements were resolved through discussion with the involvement of a third researcher (J.S.). Finally, data were extracted from each study, including the author, publication year, patient age, gender, surgical approach, number of patients in each group, tumor size and CTR, pathological type, stage (TNM version compatibility was ensured through systematic reclassification to the 8th Edition), as well as the hazard ratio (HR) and 95% confidence interval (CI) for OS, DFS, and RFS. If HR values were not directly available in the articles, they were calculated using the number of patients in each group, OS, DFS, or RFS rates, and P values with the Microsoft Excel spreadsheet (Version 16.49) provided by Tierney* et al. *(17). CTR stratification thresholds (≤0.25, >0.25 to ≤0.5, >0.5 to <1, 1) were defined per JCOG0201 and Fleischner Society standards, with pure solid nodules categorically designated as CTR =1.

Study quality evaluation process and risk of bias

For RCTs, two researchers (C.S. and J.C.) independently evaluated the risk of bias in the included studies using the Cochrane Risk of Bias Tool. This tool addresses seven random sequence generation, allocation hiding, blindness of participants and personnel, blindness of result evaluations, incomplete result data, selective result reporting, and other sources of bias. Each domain was categorized as having a low, high, or unclear risk of bias (18). For retrospective studies, the Newcastle-Ottawa Scale (NOS) was employed to assess the risk of bias. The NOS evaluates three key the selection method of the case group and the control group, the comparability of the case group and the control group, and the method of assessing exposure (19). The assessment was conducted independently by the same two researchers (C.S. and J.C.). Additionally, publication bias was analyzed using the Egger test and the Begg test. If significant publication bias was detected (P<0.05), the trim-and-fill method was applied to adjust the results and perform further data analysis, ensuring more accurate and reliable findings.

Statistical analysis

We conducted pairwise meta-analyses using RevMan software (version 5.3) to calculate the HR and its 95% CI for survival outcomes, including OS, DFS, and RFS. Heterogeneity was assessed using the I^2^ statistic and the Q-test.

Network meta-analyses were performed using the gemtc package in R software (version 4.3.2). Analyses were conducted using Markov Chain Monte Carlo (MCMC) simulation techniques, enabling the integration of both direct and indirect evidence to compare treatment strategies between any two surgical approaches (20). Among the two primary frameworks for network meta-analysis—frequentist and Bayesian—we adopted the Bayesian framework due to its ability to incorporate prior knowledge, manage uncertainty, and address challenges associated with sparse data, such as estimation biases and overconfidence in frequentist methods (21,22). A random-effects consistency model was employed, running four independent chains with 50,000 iterations each, discarding the first 20,000 as burn-in. Convergence was first evaluated qualitatively by inspecting trace plots to assess the overlap among MCMC chains, where satisfactory convergence was indicated by indistinguishable trajectories across chains. Good convergence was further confirmed through density plots, where a smooth, normal distribution curve aligned with the model’s prior assumptions, and the Bandwidth value approached stability near zero. A quantitative convergence assessment was subsequently performed using the Brooks-Gelman-Rubin diagnostic. Adequate convergence was indicated by the shrink factor’s median and 97.5% values approaching 1 and stabilizing (22). Results were reported as HRs with corresponding 95% credible intervals (CrIs). Treatment rankings were based on the surface under the cumulative ranking curve (SUCRA) values, ranging from 0 (the least effective treatment) to 1 (the most effective treatment), along with the probability of superiority rankings presented in league tables and bar charts. The assumptions of transitivity and consistency, critical to network meta-analysis, were evaluated. Transitivity was assessed using Bayesian meta-regression to explore potential modifiers such as sample size, gender, age, and smoking status. Local inconsistencies between direct and indirect evidence were evaluated using results from pairwise meta-analyses and node-splitting methods. Global inconsistency was examined by comparing the model fit between consistent and inconsistent models (23).

Finally, STATA software (version SE15) was used to generate network plots, and funnel plots for publication bias, and to perform Egger’s and Begg’s tests.

Results

Screening results of the literature

From January 2000 to November 2024, a total of 4,672 publications were retrieved from major databases, including 1,244 from PubMed, 1,359 from Web of Science, 1,828 from Embase, 194 from the Cochrane Library, and 47 from ClinicalTrials.gov. After screening, 61 studies were included in the final analysis, comprising 3 RCTs and 58 retrospective studies. Among these, 58 studies reported HR for OS, 16 for DFS, and 22 for RFS (Figure 1).

Study characteristics and risk of bias assessment

The basic characteristics of the included studies are summarized in Table 1. We assessed the risk of bias for RCTs using the Cochrane Risk of Bias Tool and for retrospective studies using the NOS (Tables S2,S3). Funnel plots, along with quantitative Egger’s and Begg’s tests, were employed to evaluate potential publication bias (Figure S1). For endpoints such as OS, DFS, and RFS, we consider the likelihood of bias to be minimal, as outcomes like mortality are less susceptible to the influence of patients, physicians, or outcome assessors.

Overall analysis results for clinical stage IA

Pairwise meta-analysis

Comparing lobectomy and segmentectomy, we found significant OS advantage for lobectomy [HR 1.12 (95% CI: 1.03–1.21), I^2^=51%], across 48 studies, but no significant differences in DFS [HR 1.04 (95% CI: 0.92–1.17), I^2^=0%, 12 studies] or RFS [HR 0.89 (95% CI: 0.74–1.07), I^2^=45%, 18 studies] (Figure S2). In the comparison between segmentectomy and wedge resection, 18 studies provided OS data, 5 studies reported DFS data and 6 studies provided RFS data. Segmentectomy showed significant advantages across OS, DFS, and RFS [HROS 0.76 (95% CI: 0.69–0.85), I^2^=51%; HRDFS 0.60 (95% CI: 0.50–0.71), I^2^=0%; HRRFS 0.71 (95% CI: 0.53–0.96), I^2^=17%] (Figure S3). In the comparison between wedge resection and lobectomy, 22 studies reported OS data, while 6 studies each provided DFS and RFS data. Lobectomy demonstrated a significant advantage in OS and RFS [HROS 1.57 (95% CI: 1.40–1.76), I^2^=69%; HRRFS 2.02 (95% CI: 1.53–2.67), I^2^=39%], with no significant difference observed for DFS [HR 1.30 (95% CI: 0.98–1.74), I^2^=67%] (Figure S4).

Network meta-analysis

A total of 58 studies reported OS, 16 studies reported DFS and 22 studies reported RFS. The network relationships of the three surgical approaches are illustrated in Figure 2. Using a random-effects consistency model, diagnostic plots, including trace plots, density plots, and Gelman-Rubin-Brooks diagnostic plots, indicated satisfactory convergence, as evidenced by the Bandwidth values approaching 0 and the median and 97.5% quantiles of the shrinkage factor approaching 1 (Figure S5). The comprehensive analysis yielded the following no significant differences were observed in OS, DFS, or RFS between lobectomy and segmentectomy [HROS 0.90 (95% CI: 0.82–1.00); HRDFS 0.98 (95% CI: 0.79–1.21); HRRFS 1.07 (95% CI: 0.88–1.32)]. Compared to wedge resection, both lobectomy and segmentectomy showed significant advantages in OS, DFS, and RFS [HROS 0.67 (95% CI: 0.59–0.75), 0.75 (95% CI: 0.65–0.84); HRDFS 0.69 (95% CI: 0.54–0.89), 0.71 (95% CI: 0.55–0.92); HRRFS 0.57 (95% CI: 0.42–0.78), 0.53 (95% CI: 0.39–0.72)]. Bar plots of the ranking probabilities and SUCRA values indicated that lobectomy ranked first in terms of OS and DFS (SUCRAOS =0.99; SUCRADFS =0.79), while segmentectomy ranked first in RFS (SUCRARFS =0.88) (Figure 3).

The transitivity assumption was accepted due to the absence of significant differences (Tables S4). Local consistency was ensured through the alignment of results between pairwise meta-analysis and network meta-analysis, as well as node-splitting analyses with P values ≥0.05 (Tables S5). Global consistency was verified by the similar model fit between consistency and inconsistency models (Tables S6). Heterogeneity testing revealed some degree of heterogeneity among the included studies. To minimize the potential impact of this heterogeneity on the overall results, subgroup analyses were further conducted to explore the sources of heterogeneity and make appropriate adjustments (Tables S5).

Overall analysis results for clinical stage T1a/b

Pairwise meta-analysis

In the comparison between lobectomy and segmentectomy, no significant differences were observed in OS, DFS, or RFS [HROS 1.06 (95% CI: 0.99–1.14), I^2^=27%; HRDFS 1.01 (95% CI: 0.88–1.16), I^2^=0%; HRRFS 0.99 (95% CI: 0.86–1.15), I^2^=12%] (Figure S6). Comparing segmentectomy and wedge resection, segmentectomy demonstrated a significant advantage in OS, DFS, and RFS [HROS 0.75 (95% CI: 0.70–0.82), I^2^=23%; HRDFS 0.69 (95% CI: 0.56–0.84), I^2^=0%; HRRFS 0.66 (95% CI: 0.51–0.84), I^2^=4%] (Figure S7). In the comparison between wedge resection and lobectomy, lobectomy showed significant advantages in OS and RFS [HROS 1.62 (95% CI: 1.38–1.91), I^2^=65%; HRRFS 1.92 (95% CI: 1.57–2.34), I^2^=0%], while no significant difference was observed in DFS [HRDFS 1.25 (95% CI: 0.98–1.58), I^2^=47%] (Figure S8).

Network meta-analysis

A total of 39 studies reported OS, 14 studies reported DFS and 14 studies reported RFS. Using a random-effects consistency model, convergence was satisfactory. The comprehensive analysis yielded the following compared to segmentectomy, lobectomy may have an advantage in OS, DFS, and RFS [HROS 0.95 (95% CI: 0.84–1.09); HRDFS 0.98 (95% CI: 0.81–1.17); HRRFS 0.97 (95% CI: 0.81–1.19)], but these differences did not reach statistical significance. Compared to wedge resection, both lobectomy and segmentectomy demonstrated significant advantages in OS, DFS, and RFS [HROS 0.65 (95% CI: 0.56–0.75), 0.69 (95% CI: 0.58–0.80); HRDFS 0.77 (95% CI: 0.63–0.94), 0.78 (95% CI: 0.64–0.98); HRRFS 0.53 (95% CI: 0.41–0.68), 0.54 (95% CI: 0.42–0.70)]. Bar plots of ranking probabilities and SUCRA values demonstrated that lobectomy ranked first in OS, DFS, and RFS (SUCRAOS =0.88; SUCRADFS =0.78; SUCRARFS =0.81) (Figure 4). Local consistency was ensured by the agreement between pairwise meta-analysis and network meta-analysis results or node-splitting analyses with P values ≥0.05. Heterogeneity testing indicated minimal heterogeneity among the included studies (Table S7).

Clinical stage T1a/b with CTR <0.5

Only two studies have reported OS, and three studies have addressed RFS. Given the limited number of included studies, both pairwise meta-analysis based on the frequentist approach and Bayesian network meta-analysis, which failed to form closed loops, demonstrated considerable heterogeneity. Therefore, this section qualitatively synthesizes the existing research findings.

Among the included studies, Kamigaichi* et al. reported that both wedge resection and lobectomy achieved a 5-year OS of 100%. Compared to segmentectomy, the 5-year OS and RFS showed a relative advantage for wedge resection, but this did not reach statistical significance [HROS 0.81 (95% CI: 0.46–1.45); HRRFS 0.87 (95% CI: 0.50–1.56)] (43). Motono et al. revealed that wedge resection exhibited a potential advantage in 5-year RFS compared to lobectomy, though statistical significance was not attained [HR 0.77 (95% CI: 0.13–4.45)] (52). Liu et al. *demonstrated that segmentectomy showed an advantage in 5-year OS and RFS compared to wedge resection, but this advantage was not statistically significant [HROS 0.50 (95% CI: 0.13–1.90); HRRFS 0.53 (95% CI: 0.15–1.88)] (49).

For patients with T1a/b-stage (0–2 cm) tumors and CTR <0.25, the pathological findings are predominantly adenocarcinoma* in situ *or minimally invasive adenocarcinoma, which exhibit extremely low invasiveness and minimal pleural, vascular, and lymph node involvement. The JCOG0201 study demonstrated that the specificity for the absence of lymph node metastasis and vascular invasion in this population was as high as 98.7% (6). The JCOG0804 study, a single-arm validation clinical trial, investigated 333 cases of peripheral GGO-dominant adenocarcinomas (tumor long diameter ≤2 cm, CTR ≤0.25). Among them, 258 patients underwent wedge resection with a required surgical margin of 5 mm, while 56 underwent segmentectomy. The 5-year RFS for the study population reached 99.7%, with no recurrences observed during follow-up. These findings suggest that a 5-mm margin is sufficient, and sublobar resection (82% wedge resection) is a feasible and effective therapeutic approach (10,11). For patients with T1a/b-stage (0–2 cm) tumors and 0.25< CTR <0.5, the JCOG1211 study excluded patients from the JCOG0804 criteria and included those with a tumor long diameter ≤3 cm and CTR ≤0.5. Patients underwent segmentectomy, achieving a 5-year OS and RFS of 98.2%. Even for patients with postoperative invasive adenocarcinoma pathology, the 5-year RFS reached 97%, with only two recurrences and one case of bone metastasis reported during follow-up. These findings support segmentectomy as a standard surgical procedure for this population (13).

The “Expert Consensus on the Rational Surgical Management of Pulmonary Nodules ≤2 cm in Diameter in China (2024)” released by Chinese experts in August 2024 recommended wedge resection for GGO lesions located in the outer one-third of the lung with a CTR ≤0.25. For pulmonary nodules with 0.25< CTR <0.5, either wedge resection or segmentectomy may be considered. For nodules located in the inner two-thirds of the lung with CTR ≤0.5, segmentectomy was recommended as the first choice (85). In September 2024, the “Chinese Medical Association Guidelines for the Clinical Management of Lung Cancer (2024 Edition)” strongly recommended sublobar resection for peripheral T1a/b N0 patients with ground-glass components, prioritizing segmentectomy. For patients with poor pulmonary reserve or significant comorbidities, wedge resection was recommended, followed by segmentectomy as a secondary option (86). Similarly, the Japanese Lung Cancer Guidelines (2024 Edition), released in September, recommended sublobar resection for peripheral tumors ≤2 cm with CTR ≤0.25 and strongly advocated segmentectomy for tumors with 0.25< CTR <0.5. In November 2024, the National Comprehensive Cancer Network (NCCN) Guidelines for Non-Small Cell Lung Cancer (Version 11.2024) emphasized that sublobar resection should be strongly considered for peripheral T1a/b N0 tumors ≤2 cm (87).

In conclusion, for clinical T1a/b-stage (0–2 cm) patients with CTR <0.5, sublobar resection is recommended. For patients with CTR <0.25, wedge resection is suggested, whereas for those with 0.25< CTR <0.5, the choice between wedge resection and segmentectomy should be further determined based on tumor location, solid component, pulmonary function, and patient tolerance.

Clinical stage T1a/b with 0.5< CTR <1

Pairwise meta-analysis

In the comparison between lobectomy and segmentectomy, lobectomy demonstrated a potential advantage in OS, DFS, and RFS [HROS 0.91 (95% CI: 0.77–1.09), I^2^=0%; HRDFS 0.73 (95% CI: 0.46–1.16), I^2^=0%; HRRFS 0.95 (95% CI: 0.74–1.22), I^2^=0%], but none of these differences reached statistical significance (Figure S9). In the comparison between segmentectomy and wedge resection, segmentectomy showed significant advantages in OS and RFS [HROS 0.50 (95% CI: 0.35–0.70), I^2^=1%; HRRFS 0.53 (95% CI: 0.38–0.73), I^2^=0%], while no significant difference was observed in DFS [HR 0.68 (95% CI: 0.31–1.51), I^2^=69%] (Figure S10). In the comparison between wedge resection and lobectomy, lobectomy exhibited an advantage in RFS [HR 1.65 (95% CI: 1.22–2.25), I^2^=0%], but no significant differences were found in OS or DFS [HROS 1.74 (95% CI: 0.99–3.06), I^2^=67%; HRDFS 1.27 (95% CI: 0.45–3.56), I^2^=88%] (Figure S11).

Network meta-analysis

A total of 8 studies reported OS, 3 studies reported DFS and 5 studies reported RFS. The random-effects consistency model indicated satisfactory convergence. The combined results were as compared with segmentectomy, lobectomy showed a potential disadvantage in OS, DFS, and RFS [HROS 1.10 (95% CI: 0.79–1.58); HRDFS 1.23 (95% CI: 0.45–3.41); HRRFS 1.03 (95% CI: 0.68–1.48)], but none of these differences reached statistical significance. Compared with wedge resection, both lobectomy and segmentectomy showed significant advantages in OS and RFS [HROS 0.57 (95% CI: 0.38–0.94), 0.52 (95% CI: 0.34–0.84); HRRFS 0.53 (95% CI: 0.32–0.83), 0.51 (95% CI: 0.31–0.84)], while no significant differences were observed in DFS [HR 0.89 (95% CI: 0.35–2.43), 0.72 (95% CI: 0.30–1.79)]. The bar plots of ranking probabilities and the SUCRA demonstrated that segmentectomy ranked highest in OS, DFS, and RFS (SUCRAOS =0.87; SUCRADFS =0.73; SUCRARFS =0.78) (Figure 5).

Clinical stage T1a/b with CTR =1

Pairwise meta-analysis

In the comparison between lobectomy and segmentectomy, lobectomy showed a potential advantage in OS, DFS, and RFS [HROS 1.03 (95% CI: 0.84–1.25), I^2^=35%; HRDFS 1.02 (95% CI: 0.83–1.26), I^2^=0%; HRRFS 1.14 (95% CI: 0.88–1.47), I^2^=17%], but none of these differences reached statistical significance (Figure S12). In the comparison between segmentectomy and wedge resection, no significant differences were observed in OS or DFS [HROS 0.81 (95% CI: 0.40–1.63), I^2^=64%; HRDFS 1.03 (95% CI: 0.50–2.12), I^2^=82%], and RFS data were not reported (Figure S13). In the comparison between wedge resection and lobectomy, lobectomy showed a significant advantage in OS [HR 2.17 (95% CI: 1.06–4.41), I^2^=90%], while DFS and RFS data were not available (Figure S14).

Network meta-analysis

A total of 9 studies reported OS, 4 studies reported DFS and 4 studies reported RFS. The random-effects consistency model showed satisfactory convergence. The combined results were as compared with segmentectomy, lobectomy demonstrated potential advantages in OS, DFS, and RFS [HROS 0.99 (95% CI: 0.66–1.60); HRDFS 0.94 (95% CI: 0.62–1.45); HRRFS 0.83 (95% CI: 0.46–1.32)]. However, none of these outcomes were statistically significant. Compared with wedge resection, lobectomy and segmentectomy showed potential advantages in OS, DFS, and RFS [HROS 0.63 (95% CI: 0.35–1.24), 0.63 (95% CI: 0.35–1.18); HRDFS 0.81 (95% CI: 0.52–1.49), 0.86 (95% CI: 0.58–1.48); HRRFS 0.44 (95% CI: 0.19–1.02), 0.54 (95% CI: 0.23–1.35)]. Similarly, the differences were not statistically significant. The bar plots of ranking probabilities and the SUCRA demonstrated that lobectomy ranked highest in OS, DFS, and RFS (SUCRAOS =0.72; SUCRADFS =0.71; SUCRARFS =0.90) (Figure 6).

Radiologically pure-solid NSCLC, lacking GGO, is a highly aggressive tumor characterized by a higher pathological invasiveness compared to part-solid NSCLC with GGO. Such invasiveness includes lymphovascular invasion, blood vessel invasion, lymph node metastasis, spread through air spaces (STAS), and nodal involvement (88-92). A supplemental analysis of the JCOG0201 study showed worse OS in patients with radiologically pure-solid NSCLC compared to those with part-solid NSCLC (89). Similarly, a supplemental analysis from the JCOG0802 study revealed that the local recurrence rate in the segmentectomy group for radiologically pure-solid nodules (16.1%) was significantly higher than in the lobectomy group (7.7%) (P=0.0021) (15). Pure-solid nodules are more likely to have STAS positivity, and studies have demonstrated that even with a sufficient surgical margin-to-tumor ratio of ≥1, recurrence cannot be significantly reduced for STAS-positive nodules (93,94). For patients with occult lymph node metastasis, residual tumors may negatively impact the outcomes of sublobar resections, and the effectiveness of segmentectomy remains controversial (95,96). The incidence of occult lymph node metastasis in IA1–2 stage pure-solid NSCLC ranges from 11.1% to 17.7%, and segmentectomy may sometimes be less effective than lobectomy in achieving thorough lymph node dissection (91,97,98). Pure-solid nodules are also more likely to exhibit high-risk pathological subtypes, such as micropapillary components. Studies have identified that micropapillary components accounting for ≥5% of the tumor are an independent risk factor for postoperative recurrence in patients undergoing sublobar resection, with both RFS and OS being significantly worse in the segmentectomy group compared to the lobectomy group (99). Furthermore, studies by Nomori* et al. and Chen et al. *suggested that larger segmentectomy resections may not necessarily preserve more lung function compared to lobectomy, as the amount of preserved lung tissue does not always translate into significant functional benefits (100,101). Therefore, based on the above evidence, lobectomy remains the recommended surgical approach for radiologically pure-solid NSCLC with a tumor size of 0–2 cm.

Overall analysis results for clinical stage T1c

Pairwise meta-analysis

In the comparison between lobectomy and segmentectomy, lobectomy demonstrated a significant advantage in OS and DFS [HROS 1.36 (95% CI: 1.27–1.46), I^2^=29%; HRDFS 1.42 (95% CI: 1.07–1.89), I^2^=0%]. However, it did not reach statistical significance in RFS [HR 0.98 (95% CI: 0.73–1.32), I^2^=0%] (Figure S15). In the comparison of segmentectomy and wedge resection, segmentectomy showed a significant advantage in OS [HR 0.73 (95% CI: 0.57–0.94), I^2^=63%] (Figure S16). In the comparison between wedge resection and lobectomy, lobectomy exhibited an advantage in OS [HR 1.83 (95% CI: 1.29–2.61), I^2^=88%] (Figure S17). Data on DFS and RFS were not reported for the latter two comparisons.

Network meta-analysis

A total of 15 studies reported on OS, while 3 studies reported on DFS, with satisfactory convergence of the random effects consistency model. The results of the comprehensive analysis are as compared to segmentectomy, lobectomy demonstrated a significant advantage in OS [HR 0.73 (95% CI: 0.60–0.91)] and may have an advantage in DFS [HR 0.68 (95% CI: 0.36–1.18)], though this too did not reach statistical significance. Compared to wedge resection, lobectomy showed a significant advantage in OS [HR 0.60 (95% CI: 0.44–0.77)] and may have an advantage in DFS [HR 0.52 (95% CI: 0.23–1.11)]. Similarly, the difference was not statistically significant. Segmentectomy may have an advantage in both OS and DFS [HROS 0.81 (95% CI: 0.60–1.04); HRDFS 0.76 (95% CI: 0.37–1.69)], but the difference was not statistically significant. The bar chart of ranking probabilities for different surgical approaches and the SUCRA values demonstrated that lobectomy ranks high in both OS and DFS (SUCRAOS =0.99; SUCRADFS =0.95) (Figure 7). An additional four studies reported on RFS, but only the comparison between lobectomy and segmentectomy was available, thus precluding the possibility of a network meta-analysis.

Clinical stage T1c with CTR <0.5

There were no clinical stage T1c patients with a CTR <0.5 included in the studies reviewed in this article, and therefore pairwise meta-analysis and network meta-analysis could not be conducted. Therefore, this section will qualitatively synthesize the findings of existing research.

JCOG1211 is a single-arm validation clinical trial that excluded patients from the JCOG0804 study, focusing on patients with tumors having a long diameter of ≤3 cm and a CTR of ≤0.5, who underwent segmentectomy. Follow-up results for 154 patients with tumors measuring 2–3 cm showed that the 5-year OS and RFS rates both reached 98.0%. The postoperative decline in expiratory volume in 1 second​(FEV1) at 1 year was 7.3%, which is lower than the 12.0% observed in the JCOG0804 lobectomy group, with an intergroup difference of 4.7% (P<0.0001). During follow-up, only 2 cases of recurrence and 1 case of bone metastasis were reported, indicating that segmentectomy is both safe and effective for this patient population (13). However, this study has certain limitations; the current results suggest that the incidence of local recurrence is extremely low during the 5-year follow-up period, but this does not necessarily imply that recurrences will not occur after 5 years. Therefore, whether segmentectomy can provide comparable long-term outcomes to lobectomy remains to be supported by high-quality prospective research evidence. The ECTOP-1012 prospective trial is currently underway, and we look forward to more studies addressing the efficacy gaps for this subset of patients.

Clinical stage T1c with 0.5< CTR <1

Pairwise meta-analysis

In the comparison between lobectomy and segmentectomy, there were no significant differences in OS and RFS [HROS 1.06 (95% CI: 0.30–3.72), I^2^=75%; HRRFS 0.73 (95% CI: 0.40–1.36), I^2^=65%] (Figure S18).

Network meta-analysis

A total of 4 studies reported OS, and the convergence of the random effects consistency model was satisfactory. The results of the comprehensive analysis are as compared to segmentectomy, lobectomy may have an advantage in OS [HR 0.90 (95% CI: 0.27–3.31)], but this also did not achieve statistical significance. Compared to wedge resection, both lobectomy and segmentectomy may have advantages in OS [HR 0.51 (95% CI: 0.06–4.34) and 0.57 (95% CI: 0.06–4.51)], although these did not reach statistical significance. The bar chart of ranking probabilities for different surgical approaches and the SUCRA values demonstrated that lobectomy ranks first in OS (SUCRA =0.66) (Figure 8). Only one study reported on DFS. Additionally, two studies reported on RFS, but due to only the comparison between lobectomy and segmentectomy being available, a network meta-analysis could not be conducted.

Clinical stage T1c with CTR =1

Only one study reported on both OS and DFS, hence pairwise meta-analysis and network meta-analysis could not be conducted. Therefore, this section will qualitatively synthesize the findings of existing research.

In the literature included in this article, the study by Soh* et al. *indicated that lobectomy may have advantages over segmentectomy in OS and DFS [HROS 1.48 (95% CI: 0.96–2.28); HRDFS 1.17 (95% CI: 0.82–1.66)], although neither reached statistical significance. Compared to wedge resection, lobectomy demonstrated significant advantages [HROS 1.25 (95% CI: 1.06–1.42); HRDFS 1.76 (95% CI: 1.26–1.46)], and segmentectomy also showed significant benefit [HROS 0.61 (95% CI: 0.39–0.95); HRDFS 0.79 (95% CI: 0.67–0.93)] (69).

Research indicates that among patients with radiologically confirmed pure solid NSCLC larger than 2–3 cm, 22% observed STAS (+) (98). Additionally, one study reported that 36% of lung adenocarcinomas greater than 2–3 cm histopathologically included the micropapillary or solid subtype (102). Other studies have shown that the incidence of occult lymph node metastasis in clinically staged IA3 pure solid NSCLC ranges from 17.3% to 36.0% (91,97,98), with some hilar lymph nodes being difficult to clear during segmentectomy. Such nodules exhibit high invasiveness, making it crucial to ensure adequate surgical margins and thorough lymph node dissection to prevent local regional recurrence. In summary, for patients with clinical T1c stage and a CTR of 1, lobectomy is recommended.

Based on the tumor long diameter and CTR values, combined with the results of this meta-analysis and existing large clinical studies, we have constructed a fan chart recommending surgical approaches (Figure 9).

Discussion

In this meta-analysis, we systematically summarized the survival outcomes of lobectomy, segmentectomy, and wedge resection in patients with clinical IA NSCLC. Overall results demonstrated that lobectomy and segmentectomy have significant advantages over wedge resection, with no significant differences observed between the two. The choice of the optimal surgical approach should take into account tumor size and the CTR value. The CTR, as an easily measurable clinical indicator, intuitively reflects the proportion of solid tumor components, thereby indirectly assessing the biological behavior and malignancy of the tumor. The CTR is closely related to patient survival outcomes and recurrence risk, and it plays an important guiding role in the selection of surgical methods. However, it must be acknowledged that CTR is not a flawless metric, as its accuracy in quantifying complex consolidation patterns remains contentious. Future biomarkers—such as deep learning radiomics—may surpass its prognostic value. The principal significance of CTR lies in empowering surgeons with intraoperative decision-making autonomy regarding resection extent. Furthermore, the selection of surgical approaches for solid nodules has been a subject of considerable debate in prior studies (103,104). This research offers important insights into this matter. Given the high invasiveness of solid nodules, lobectomy remains a safe and reliable surgical option that is strongly recommended to mitigate the risk of recurrence and enhance patient prognosis.

For patients with clinical T1a/b stage, when CTR <0.25, the JCOG0804 study suggests that wedge resection can be a feasible and effective treatment option. However, research by Kakinuma* et al. *indicated that among nodules with a long diameter <3 cm and a solid component <5 mm, only 14% of pure GGO nodules exceeded 2 mm in diameter after 5 years of CT follow-up (105). The JCOG0804 study included over 50% pure GGO patients, raising the question of whether such patients require surgical intervention. The forthcoming JCOG1906 study is expected to provide clearer answers regarding the optimal timing for intervention and the issue of overtreatment. When 0.25< CTR <0.5, the JCOG1211 study supports segmentectomy as the standard surgical approach. However, since 83% of patients in this study had non-invasive cancer, whether these patients can opt for wedge resection still needs further exploration. Existing studies suggest that for non-invasive cancers ≤2 cm, wedge resection may be a reasonable choice, considering segmentectomy only when surgical margins are insufficient (11,106). Nevertheless, segmentectomy theoretically offers better radical resection outcomes (107). In the JCOG1211 study, 54% of patients underwent complex segmentectomy, indicating that simple segmentectomy may not meet the requirements for tumors larger than 2 cm. The optimal surgical approach—segmentectomy versus wedge resection—remains undetermined, pending results from ongoing prospective clinical trials such as ECTOP-1020. When 0.5< CTR <1, this meta-analysis shows the superiority of segmentectomy in this patient group. The JCOG0802 study indicates that segmentectomy is a reliable treatment option. However, retrospective studies point out that wedge resection may have comparable prognostic effects to segmentectomy in certain cases and is safer (108,109). As multicenter RCTs like JCOG1909 progress, more evidence will help us better evaluate the selection of these surgical approaches. When CTR =1, lobectomy ranked first in this meta-analysis. The JCOG0802 study indicated that segmentectomy has a significant advantage over lobectomy in OS, but the recurrence rate nearly doubled (15). Considering the high recurrence risk of pure solid nodules (such as STAS positivity, micropapillary, and other high-risk pathological subtypes, as well as occult lymph node metastasis), removing less lung tissue does not necessarily preserve more lung function. Therefore, the primary surgery should strive for radical tumor resection to avoid postoperative recurrence and improve patients’ quality of life.

For patients with clinical T1c stage, when CTR <0.5, the JCOG1211 study indicated that segmentectomy is safe and effective. However, considering that 83% of non-invasive cancer patients may not experience recurrence or metastasis, the appropriateness of segmentectomy still needs further discussion (13). The prospective study ECTOP-1012 is expected to provide more evidence. When 0.5< CTR <1, the results of this meta-analysis suggest that lobectomy is the preferred treatment option. The subgroup analysis of the JCOG0201 study indicated that patients with higher CTR values have a greater risk of recurrence, making sub-lobar resection inappropriate. However, this group included a considerable number of pN1–N2 patients and patients with CTR =1 with a high recurrence risk, and if pN0 patients with CTR <1 were re-grouped, the results might differ (9).

Through a systematic review of recent meta-analyses, we found that multiple studies comparing lobectomy versus segmentectomy demonstrate non-significant differences in survival outcomes for overall stage IA and T1a/b patients (2,110), whereas lobectomy shows significant advantages in T1c patients (111). Additional comparisons between lobectomy and sublobar resection (including segmentectomy and wedge resection) similarly indicate no statistically significant differences in overall stage IA cohorts (112-115). These findings align substantially with our meta-analysis conclusions. While these prior studies have established a critical foundation, two areas warranting further exploration are the independent efficacy evaluation of wedge resection and granular stratified analyses based on tumor diameter and CTR. Notably, although Shi* et al. *conducted a network meta-analysis comparing three surgical approaches—revealing significantly improved OS with lobectomy versus wedge resection—their study similarly lacked subgroup stratification and exclusively incorporated retrospective designs (116). To our knowledge, this study represents the first systematic comparison of survival outcomes across three surgical modalities stratified by tumor diameter and CTR values, and innovatively constructs a surgical decision-making fan chart. This visualization tool provides critical support for individualized surgical selection, demonstrating significant clinical utility. We look forward to future studies, including JCOG1906, JCOG1909, and ECTOP-1020, providing more high-quality evidence for the surgical treatment of early lung cancer. In addition, when selecting surgical approaches, many clinicians not only focus on survival outcomes but also consider important functional outcomes such as intraoperative trauma, postoperative lung function, perioperative complications, and quality of life. This study was unable to obtain sufficient data on these outcomes; however, we recognize that these factors also play a role in the decision-making process. Therefore, future research should focus on systematically evaluating these outcomes. By integrating these multidimensional results, we can achieve a more comprehensive understanding of the impact of different surgical methods on overall patient health and quality of life.

Limitation

There are several limitations in this meta-analysis. First, the substantial amount of retrospective study data is subject to selective reporting bias, and the lack of propensity score matching to adjust for baseline differences introduces uncertainty in the conclusions. While global bias tests indicate acceptable risk, interpretations of specific subgroups with limited primary studies require caution and await validation in large-scale prospective cohorts. Second, we utilized aggregated data from each study rather than individual patient data, leading to inherent differences in study design and heterogeneity among participants, which may act as confounding factors affecting the results. Third, factors such as tumor characteristics and patients’ preoperative conditions may contribute to biases in the choice of surgical methods across the studies. Fourth, landmark studies such as the LCSG trial could not be included in our analytical framework due to unstratified sublobar resection categories; however, they nevertheless made pivotal contributions to the evolution of surgical paradigms for lung cancer therapy. Fifth, methodological heterogeneity in CTR assessment across studies could not be systematically quantified, constituting a significant uncontrolled confounder; consequently, subgroup interpretations warrant circumspection.

Conclusions

For patients with clinical stage IA NSCLC, lobectomy and segmentectomy generally demonstrate advantages over wedge resection, with no significant differences observed between lobectomy and segmentectomy. The optimal surgical approach should take multiple factors into account, including the tumor size and CTR. This meta-analysis highlights the need for high-quality clinical research to better understand the optimal surgical treatment of early-stage lung cancer, especially for specific subgroups of patients.

Supplementary