Authors: Lauren Kari Dixon, Ella Barber, Adrian Cook, Joanne Boudour, John Conibear, Douglas West, Neal Navani, David A Cromwell
Categories: Lung Cancer, Non-Small Cell Lung Cancer, Thoracic Surgery
Source: BMJ Open Respiratory Research
Sublobar resection or lobectomy for stage Ia non-small cell lung a systematic review and meta-analysis
Authors: Lauren Kari Dixon, Ella Barber, Adrian Cook, Joanne Boudour, John Conibear, Douglas West, Neal Navani, David A Cromwell
Abstract
Background
This systematic review and meta-analysis synthesises evidence from both randomised trials and observational studies to determine whether lobectomy or sublobar resection offers improved outcomes for patients with stage Ia non-small cell lung cancer (NSCLC).
Methods
Studies (up to June 2025) comparing lobectomy and sublobar resection (segmentectomy or wedge) for clinical stage Ia NSCLC (<2 cm) were included in the random-effects meta-analyses. Risk of bias was assessed using Risk of Bias 2 for randomised trials or Risk of Bias in Non-randomised Studies of Interventions-I for observational studies.
Results
19 studies, including four randomised trials, were included. Overall survival at 5 years was comparable between lobectomy and sublobar resection (HR=1.00; 95% CI 0.84 to 1.19; I²=26%), as was disease-free survival (HR=1.05; 95% CI 0.90 to 1.23; I²=0%). Sublobar resection was associated with significantly higher local recurrence (OR=1.86; 95% CI 1.07 to 3.25; I²=73%). No differences were observed in 10-year survival (OR=0.99; 95% CI 0.27 to 3.59; I²=86%) or postoperative change in forced expiratory volume in 1 s (mean difference=–4.70; 95% CI –11.15 to 1.76; I²=99%). In 10 studies that mandated systematic hilar and mediastinal lymph node sampling, sublobar resection was associated with improved overall survival compared with lobectomy (HR=0.81; 95% CI 0.69 to 0.965; I²=0%).
Conclusion
Lobectomy and sublobar resection offer comparable long-term survival for patients with stage Ia NSCLC. While sublobar resection is associated with higher local recurrence rates, subgroup analysis suggests that when intraoperative systematic hilar and mediastinal lymph node sampling is performed, sublobar resection may offer a survival advantage.
Introduction
Lung resection surgery is the standard management approach for early-stage non-small cell lung cancer (NSCLC); traditionally, this has been in the form of a lobectomy.^1 2^ Survival following lobectomy for stage I NSCLC can be as high as 80% at 10 years.^3^ In recent years, there has been a growing trend towards the use of sublobar resections which limit the resection of normal lung tissue, especially in small, early stage tumours. Segmentectomy offers an anatomical lung cancer resection while sparing more tissue than a lobectomy. A wedge resection is not an anatomical resection but removes the tumour and a margin of normal lung tissue. The preservation of normal lung tissue is thought to result in preserved lung function and improved postoperative physical function.^4^ However, there are concerns about oncological outcomes following a more limited resection, particularly regarding the incidence of local disease recurrence.^5^
The choice between lobectomy and sublobar resection remains a key consideration in the surgical management of early-stage NSCLC. Previous meta-analyses have examined this comparison but reached conflicting conclusions about whether one approach offers superior outcomes.513 Two major randomised controlled trials (RCTs) published in 2022 significantly added to the evidence base. The JCOG0802 trial^14^ randomised patients with clinical stage Ia NSCLC to lobectomy or segmentectomy and reported a modest overall survival benefit with segmentectomy, alongside higher loco-regional recurrence but better preservation of postoperative forced expiratory volume in 1 s (FEV1). The CALGB/Alliance trial^15^ compared lobectomy with either segmentectomy or wedge resection and concluded that sublobar resection was non-inferior to lobectomy in terms of both overall and disease-free survival (DFS). While some recent meta-analyses have incorporated these two RCTs,^16 17^ they were restricted to randomised evidence alone and did not consider lymph node sampling strategies.
Several international guidelines support the use of sublobar resection in selected patients with early-stage NSCLC. The National Comprehensive Cancer Network (NCCN) guideline recommends sublobar resection for peripheral tumours less than 2 cm in size with a consolidation-to-tumour ratio below 0.5 on CT imaging and no evidence of lymph node involvement.^18^ The European Society for Medical Oncology (ESMO) adopted a more conservative stance, considering anatomical segmentectomy acceptable primarily for pure ground-glass opacities or minimally invasive adenocarcinomas.^19^ In contrast, the National Institute for Health and Care Excellence continues to recommend lobectomy as the standard approach for operable disease.^1^ The varied recommendations highlight ongoing uncertainty regarding oncological equivalence.
The aim of this study was to perform a meta-analysis that combined the extended evidence from RCTs and observational studies to determine whether lobectomy or sublobar resection results in improved outcomes for patients with stage Ia NSCLC.
Methods
This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)^20^ statement and was registered on PROSPERO (registration CRD42023443965). The data, analytical methods and study materials will be made available to other researchers on request to the corresponding author.
Patient and public involvement
Patients and the public were not involved in this research.
Eligibility criteria
The review included RCTs and observational studies that reported overall survival (OS) for adult patients (aged 18 years or over) with a preoperative diagnosis of NSCLC stage Ia (with tumour diameter <2 cm) undergoing surgical resection. Staging was determined according to the seventh edition of the tumour, node, metastases classification of malignant tumours or as stated within the publication. The review was restricted to studies which compared a sublobar procedure (either wedge resection, segmentectomy or combination of the two) against standard lobectomy. Any surgical approach (minimally invasive, open, robotic-assisted) could have been used. Studies were excluded if they allocated patients to lobectomy or sublobar resection based on explicit criteria (eg, poor lung function tests). We also included studies that reported at least one of the other outcomes mentioned below, even if the primary outcome of OS was not reported. There was no restriction on the sample size or date of publication.
Search strategy
A literature search was conducted using PubMed, the Cochrane Central Register of Controlled Trials (CENTRAL) and the Web of Science Core Collection. Studies were eligible if published from database inception to June 2025. The search combined terms related to surgical procedure, study design, cancer type and staging. An example of the full search string used in PubMed is provided ‘((lobectomy) OR (lobar resection) OR (lobe resection)) AND ((segmentectomy) OR (sublobar resection) OR (wedge resection) OR (segmental resection)) AND ((NSCLC) OR (non-small cell lung cancer) OR (non small cell lung cancer)) AND ((stage 1) OR (stage one) OR (stage 1A) OR (T1a) OR (early stage))’.
No language restrictions were applied. The literature search was initially performed in 2023 and last updated in June 2025. The complete search strategies for all databases, including search dates, terms and number of records retrieved, are provided in online supplemental table S1.
Study selection
Search results were imported into Rayyan (https://www.rayyan.ai/)^21^ and the publication titles and abstracts were screened against the eligibility criteria by two reviewers (LKD and EB), independently. Full-text articles were obtained for all potentially eligible studies and assessed for inclusion against predefined eligibility criteria. Discrepancies in study selection were resolved through discussion and consensus between the two reviewers. A PRISMA flow diagram summarises the study selection process (see online supplemental figure S1).
Data extraction
Information was extracted from each study by two independent reviewers using a predefined proforma in accordance with guidance of the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies (CHARMS).^22^ Summary tables described the details of the study design, preoperative patient characteristics, the surgical operations and patient outcomes. The quality of each included study was quality assessed using either the Cochrane Risk of Bias 2 (ROB2) tool^23^ for randomised trials or the Risk of Bias in Non-randomised Studies of Interventions (ROBINS-I) tool^24^ for observational studies. The ROB2 tool uses five domains on which the risk of bias is rated as ‘low’, ‘some concerns’ or ‘high’. The ROBINS-I tool is based on how well the observational study mimics a target trial, and uses seven bias domains (see online supplemental material); the risk of bias is rated as ‘low’, ‘moderate’, ‘serious’ or ‘critical’, with a rating of ‘low risk’ meaning the study is judged to be equivalent to a high-quality RCT. For both tools, the ratings on each domain are used to form an overall risk of bias assessment. Each study was quality assessed by two independent reviewers (LKD and EB) and results were compared; any differences were discussed and a consensus was reached.
Outcomes
The primary outcome was OS from the time of surgery. Data were extracted on 5-year survival for each surgical intervention, as well as the HR (both crude and adjusted values) for the sublobar procedure. Secondary outcomes were cancer-specific survival at 5 years, DFS at 5 years, local and distant disease recurrence, mean reduction in FEV1 and OS at 10 years. Cancer-specific survival refers to the time from surgery during which a patient survives without dying from cancer. DFS was defined as the time from surgery to first diagnosis of local or distant recurrence. The same outcome statistics were extracted for these outcomes from each study. We extracted the mean difference in FEV1 at 6 months, when reported.
Statistical analysis
Study results are presented as percentages for categorical values and as means and 95% CIs or SD for continuous numerical variables. When available, HR and SEs were directly extracted from papers. Otherwise, HRs were calculated from reported Kaplan-Meier graphs for OS or DFS.^25 26^ Where studies reported both unadjusted and adjusted effect estimates, adjusted HRs or ORs were used for the meta-analysis. The meta-analyses of the survival and recurrence outcomes were conducted using random-effects models to combine the logarithm of the HR (with its variance). The random-effects models were fitted using the DerSimonian and Laird method for pooling estimates. Other binary outcomes were pooled using the Mantel-Haenszel method with random-effects modelling. The continuous outcome (mean change in FEV1) was analysed using inverse variance weighting under a random-effects model. Subgroup analyses were one restricted to studies that mandated systematic hilar and mediastinal lymph node assessment, and another limited to studies comparing segmentectomy only with lobectomy.
Forest plots were used to describe the results reporting HR, OR and 95% CIs. Heterogeneity was assessed using the χ²-based Q test and the I² statistic, which quantifies the percentage of total variation across studies due to heterogeneity rather than chance. Funnel plots were used to assess for possible publication bias. Throughout this study, statistical tests were two-sided and significance was defined as p<0.05. Statistical analysis was performed using Review Manager software (Review Manager (RevMan) V.5.4.1,^27^ and Stata.^28^
Results
Quantity and quality of the evidence
The search strategy returned 882 results from PubMed, 134 results from Web of Science Core Collection and 105 results from the Cochrane library. After removing duplicate results, 794 results were reviewed by titles and abstracts against the inclusion criteria. Subsequently, 58 studies remained and were considered in the full-text review stage. An additional 20 studies were found by hand-searches and cross-search of the references; 6 of those were included in the final analysis. Finally, a total of 19 studies1415 2945 were included in the meta-analysis (see tables1 2).
Of the included studies, 4 were randomised trials^14 15 29 30^ and 15 were observational studies.3145 Of the observational studies, all were retrospective cohort designs; five studies used propensity score matching for their analysis.353843 The studies were published between 2001 and 2023.
In total, 8503 patients were reviewed for this meta-analysis in either the sublobar (n=3075) or lobectomy (n=5428) group. 12 studies limited the sublobar group to include only segmentectomies (2 RCT),1430 3339 41 42 45 6 studies included both segmentectomies and wedge resections (2 RCTs)^15 29 31 40 43 44^ and 1 observational study limited the sublobar group to wedge resections.^32^
The approach of surgery used was described in 14 studies with 4 video-assisted thoracoscopic surgery (VATS) or open,^14 15 30 44^ 3 VATS,^35 37 43^ 3 open,^38 42 45^ 2 VATS, open or robotic assisted thoracoscopic surgery (RATS),^32 33^ 1 RATS^34^ and 1 database analysis study that included all approaches.^36^
Of the included studies, 15 (including all 4 RCTs) reported the outcome of 5-year OS1415 29 30 32 3543 45 and 11 studies reported 5-year DFS.^14 15 29 30 33 35 38 40 42 43 45^ The rates of OS and DFS within each intervention group can be found in the supplementary material (online supplemental table S2). Follow-up varied between the studies; the median follow-up for studies reporting OS ranged from 37 months to 190 months.
Quality assessment summary diagrams can be found in the supplementary material (online supplemental table S3 and figures S2 and S3). All RCTs included had at least some concerns when assessed using the ROB2 tool. There was a lack of blinding of patients and assessors and the conversion rate of patients assessed for eligibility to patients randomised was seldom reported. RCTs tended to have good follow-up rates and reported appropriate outcomes. With regard to the observational studies, assessment with the ROBINS-1 tool, the majority of included studies had moderate or serious concerns and only one study had a low risk of bias.^32^ A domain of particular concern for bias of the evidence base was the selection of patients. Overall, bias from completeness of follow-up and the reporting of outcomes was low.
Overall survival
15 studies reported on survival at 5 years, 4 RCTs^14 15 29 30^ and 11 observational studies,323543 45 comprising 6269 patients having sublobar (n=2685) or lobectomy (n=3584). There was no significant difference in OS at 5 years between sublobar resection and lobectomy (HR=1.00, 95% CI (0.84 to 1.19), p=1.00, I^2^=26%, figure 1). This finding is consistent when considering both only randomised (HR=0.92, 95% CI 0.66 to 1.29, p=0.65, I^2^=67%) and observational (HR=1.16, 95% CI 0.97 to 1.40, p=0.11, I^2^=0%) studies.
When limited to studies with an experimental arm of segmentectomy specifically,1430 3539 41 42 45 the findings were consistent (HR=1.07, 95% CI 0.78 to 1.46), p=0.67, I^2^=51%, (online supplemental figure S7). Repeating the analysis and excluding studies with serious or critical risk of bias on quality assessment leaving seven studies,^14 15 29 30 32 36 38^ no difference in survival was found between sublobar resection and lobectomy (HR=1.02, 95% CI 0.79 to 1.33, p=0.86, I^2^=65%).
Three studies reported OS at 10 years^31 38 41^ for 720 patients who had undergone sublobar resection (n=222) or lobectomy (n=498). There was no significant difference in 10-year survival between the procedures (OR=0.99, 95% CI 0.27 to 3.59, p=0.99, I^2^=86%, online supplemental figure S4).
Cancer-specific survival at 5 years
Cancer-specific survival was reported by three observational studies.^33 40 42^ Each study gave an estimate of cancer-specific survival at 5 years, and reported an HR. In total, 712 patients were included in the analysis, with the size of the sublobar group (n=293) being smaller than the lobectomy (n=419). Sublobar resection was associated with significantly worse cancer-specific survival (HR=1.71, 95% CI 1.07 to 2.73, p=0.03, I^2^=53%, online supplemental figure S5).
Disease-free survival at 5 years
In total, 11 studies reported the 5-year DFS following surgery,^14 15 29 30 33 35 38 40 42 43 45^ including 3724 patients in the sublobar (n=1788) and lobectomy (n=1936) groups. There was no significant difference in the DFS at 5 years between sublobar resection and lobectomy (HR=1.05, 95% CI 0.90 to 1.23, p=0.52, I^2^=0%, figure 2). The results are similar for randomised trials^14 15 29 30^ (HR=1.00, 95% CI 0.83 to 1.21, p=1.00, I^2^=0%) and observational studies^33 35 38 40 42 43 45^ (HR=1.16, 95% CI 0.89 to 1.52, p=0.26, I^2^=0%) when these types of study were analysed separately.
When limited to studies with an experimental arm of segmentectomy specifically,^14 25 30 33 35 38 42^ no difference in DFS at 5 years was found (HR=1.05, 95% CI 0.87 to 1.29, p=0.61, I^2^=0%, online supplemental figure S8).
After excluding studies with serious or critical risk of bias on quality assessment, leaving five studies,^14 15 29 30 38^ no difference in DFS was found between sublobar resection and lobectomy (HR=1.04, 95% CI 0.87 to 1.25, p=0.65, I^2^=0%).
Local recurrence was reported by 11 studies1415 30 33 37 38 40 4245 of 4419 patients who underwent sublobar resection (n=1994) or lobectomy (n=2425) surgical management. Sublobar resection was associated with significantly higher local recurrence (OR=1.86, 95% CI 1.07 to 3.25, p=0.03, I^2^=73%, figure 3).
Distant recurrence was reported by 10 studies1415 30 33 37 40 4245 of 4182 patients who underwent sublobar resection (n=1830) or lobectomy (n=2352). Between sublobar resection and lobectomy, there was no difference in terms of distant recurrence (OR=1.04, 95% CI 0.74 to 1.45, p=0.83, I^2^=43%, online supplemental figure S6).
Systematic nodal assessment
To explore whether rigorous lymph node staging influences survival, we performed a subgroup analysis limited to studies that mandated systematic sampling or dissection of hilar and mediastinal lymph nodes.^14 15 33 35 37 39 40 42 43 45^ Of these studies, nine reported OS at 5 years^14 15 35 37 39 40 42 43 45^; sublobar resection was associated with better OS (HR=0.81, 95% CI 0.69 to 0.96, p=0.01; I²=0%), as shown in figure 4. However, DFS, reported in eight studies that mandated systematic sampling or dissection of lymph nodes^14 15 33 35 40 42 43 45^ showed no statistically significant difference between lobectomy and sublobar resection (HR=1.03, 95% CI 0.87 to 1.22), p=0.73; I²=0%, online supplemental figure S9).
Mean reduction in FEV1
Mean reduction in FEV1 from pre-surgery to post-surgery was reported by four studies^14 15 34 35^ of 2242 patients who underwent sublobar (n=1029) or lobectomy (n=1213). No significant difference was found between the surgical procedures (mean difference=−4.70, 95% CI −11.15 to 1.76, p=0.15, I^2^=99%, online supplemental figure S10).
Discussion
In this meta-analysis, comprising 8503 patients undergoing either sublobar resection or lobectomy as the surgical management of NSCLC stage Ia (<2 cm), we found similar 5-year rates of overall and DFS. Sublobar resection was associated with significantly higher local recurrence and worse cancer-specific survival. We did not find a difference from the few studies that reported 10-year OS. There was no difference found in mean reduction of FEV1 preoperatively and postoperatively. Importantly, in studies that mandated systematic lymph node sampling, the sublobar cohort showed improved OS, highlighting the importance of nodal staging quality.
At 5 years, OS was not different following sublobar resection or lobectomy. This finding has been consistently reported in RCTs^15 29 30^ and other meta-analyses.^12 13 16 17 46 47^ While we did not find a difference in disease recurrence within 5 years overall, we did find that local recurrence was more common in the sublobar group during the studies follow-up intervals. This seems to have resulted in worse cancer-specific survival at 5 years in the sublobar group. Although only three studies reported this outcome^33 40 42^ and were therefore included in our analysis, further studies are needed to support this as a conclusion.
This study adds novel insight to the existing literature by including a subgroup analysis of studies that mandated systematic hilar and mediastinal lymph node assessment. Interestingly, in this analysis, sublobar resection was associated with better OS than lobectomy. This challenges long-held assumptions about the superiority of lobectomy and suggests that when adequate nodal staging is ensured, sublobar approaches may offer not only comparable but potentially superior long-term survival. This finding warrants further investigation and may have direct implications for surgical decision-making in early-stage NSCLC. The lack of a difference in DFS within the same subgroup further suggests that the observed OS benefit may relate to non-cancer factors, such as reduced perioperative morbidity or preserved pulmonary function in the sublobar group.^4 48^
It is an interesting and recurring finding that the disease recurrence does not result in worse survival. One RCT found better survival in the sublobar resection cohort despite an increase in disease recurrence.^14^ A limitation of the current evidence base is understanding the treatment and outcomes of sublobar resection patients following disease recurrence. Due to the small number of studies reporting 10-year OS,^31 38 41^ it cannot be reliably concluded that the increased local recurrence rate following sublobar resection is not affecting survival after 5 years. It is hypothesised that the effects of limited pulmonary resections compared with lobectomy result in preserved pulmonary function of patients which confers a survival benefit,^48^ even if our study does not support the argument for preserved respiratory function following sublobar resection.
Our meta-analysis of four studies comparing mean reduction in FEV1 following sublobar resection and lobectomy did not show a significant difference.^14 15 34 35^ There remains a hypothesis that preserving lung tissue would result in preserved respiratory function. Preserving lung function may also allow treatment with curative intent of synchronous, metachronous or recurrent lesions. An RCT in 2022 found a sublobar resection resulted in a 3.5% lesser reduction in FEV1 postoperatively when compared with lobectomy.^14^ However, it is difficult to translate a 3% preservation of FEV1 into a patient’s functionality. Furthermore, given FEV1 is reliant on patient effort, the lack of blinding to the surgical procedure may have led to performance bias.^49 50^ So far, no RCT has reported any patient-reported outcome measures and given the inherent importance of respiratory function to quality of life and individual functionality, this should be a research priority in the future.
In our study, we included both segmentectomy and non-anatomical wedge resection in the sublobar group. This decision was made to reflect the inclusion criteria of high-profile RCTs in this research area. However, many would argue that a segmentectomy is superior to a wedge resection in terms of oncological resection given the anatomical resection boundaries.^51^ Still, including wedge resections with segmentectomies may reflect a more realistic and pragmatic surgical approach.^52^ So far, two RCTs have randomised patients to segmentectomy specifically for comparison with lobectomy and they have reached differing conclusions.^14 30^ Meta-analyses that have included only segmentectomies and compared with lobectomy for early stage NSCLC have shown similar results to our study and subgroup analysis for segmentectomy only^13 46 47^ with similar findings for survival and disease recurrence.
Our findings also contribute to ongoing debate around international guideline recommendations. While both the ESMO and NCCN recommend sublobar resection for selected patients with early-stage NSCLC with favourable tumour characteristics,^18 19^ these guidelines are nuanced and not universally adopted. For instance, the National Institute of Clinical Excellence does not currently recommend sublobar resection as a first-line option. The comparable overall and DFS demonstrated in this meta-analysis, particularly within rigorously staged cohorts, supports a reassessment of these recommendations and suggests that sublobar resection may be a valid alternative to lobectomy in appropriately selected patients.
Limitations
Further to the limitations of secondary research, there are specific limitations of this study. Other than the stage and size limitation of the NSCLC, we did not differentiate the location of the tumour in relation to the lung field, centrally or peripherally located tumours, or the appearance of the tumour on CT scans, solid, semi-solid or ground-glass opacification. Spread through air spaces is another important phenomenon associated with recurrence that was not considered.
While surgery is the gold standard management of patients with preoperatively staged Ia NSCLC, adjuvant treatments such as systemic anti-cancer therapy may also prove important for survival should stage be more advanced after histology analysis of the surgical specimen. This meta-analysis did not consider the use of other anticancer treatments and this was inconsistently reported in the literature.
Our analysis, limited to studies mandating hilar and mediastinal lymph node sampling, showed improved OS for sublobar resections. Overall, the literature comprises a variety of lymph node sampling strategies and there were differences within individual studies for lymph node sampling between lobectomy and sublobar groups. The RCT that concluded segmentectomy was superior to lobectomy for stage Ia NSCLC adopted a strategy of systematic lymph node dissection.^14^ Systematic lymph node dissection following segmentectomy is considered a necessity as it improves staging accuracy, increases proportion for patients suitable for adjuvant therapies and helps to predict recurrence.^53^
The expected survival of early stage NSCLC is favourable and therefore, considering survival and recurrence at 5 years may not give adequate time for any differences to be realised. Only three observational studies reported 10-year OS,^31 38 41^ and no difference was found following the two procedures. Longer follow-up of the more recent randomised trials will be useful to clarify this.
A notable limitation of our study is the risk of bias present in the included studies, as evaluated using the ROB2 and ROBINS-I tools. Many of the studies displayed significant methodological concerns, such as selection bias, performance bias and confounding variables. These biases can potentially distort the true effect size and weaken the validity of our findings. This high risk of bias necessitates a cautious interpretation of the individual study conclusions and the results of this meta-analysis. Overall, this study has highlighted the need for well-designed RCTs and observational studies with longer follow-up periods and PROMs to provide more robust evidence on the comparative efficacy of lobectomy versus sublobar resection for stage Ia NSCLC.
Conclusions
This study has shown that survival at 5 and 10 years following lobectomy and sublobar resection for stage Ia NSCLC is broadly comparable, with DFS at 5 years also similar between the two surgical approaches. While sublobar resection was associated with a higher rate of local recurrence, subgroup analysis limited to studies that mandated systematic hilar and mediastinal lymph node assessment revealed a significant OS benefit for sublobar resection. This finding challenges existing assumptions about lobectomy as the standard of care and highlights the potential oncological adequacy of sublobar surgery when rigorous staging is ensured. Longer-term data are still needed to determine whether increased local recurrence associated with sublobar resection translates into excess mortality or morbidity. Furthermore, data on respiratory function and patient-reported outcome measures remain limited, and future studies should prioritise these aspects to better inform patient-centred surgical decision-making.