Trends and Outcomes of Lung Cancer Surgery in South Korea

1. Study design and data sources

We conducted a nationwide retrospective cohort study using data from the health insurance claims database, healthcare institution status reports from the Health Insurance Review and Assessment Service (HIRA), and death registration records from the Ministry of the Interior and Safety. Patient demographics and treatment data were obtained from the claims database; healthcare institution characteristics were derived from status reports; and mortality data, including dates of death, were collected from death registration records. Under the universal coverage system, the NHIS covers 97% of the population, while the remaining 3% receive benefits through the Medical Aid Program. HIRA reviews all claims from both NHIS and the Medical Aid Program, ensuring nationwide coverage. As a result, HIRA provides extensive and detailed data on healthcare services—including treatments, medications, procedures, and diagnoses—for nearly 50 million beneficiaries [16]. Since robotic surgery cannot be identified in the claims data, annual numbers were obtained in collaboration with Intuitive Surgical Korea Limited and used solely to analyze trends in lung cancer surgery volume and approaches.

2. Study participants

We identified patients diagnosed with or suspected of having lung cancer who subsequently underwent surgery between January 1, 2010, and December 31, 2023. Patients were initially selected if they had at least one claim for a primary or secondary diagnosis of lung cancer (C34: malignant neoplasm of the lung), abnormal lung imaging findings (R91), or suspected neoplasm (Z031). Among these, patients were classified as lung cancer surgery cases if they met all of the following criteria: (1) at least one lung cancer surgery procedure code (O1401, O1403, O1404, O1405, O1410, O1421, O1422, O1423, O1424, O1431, O1432); (2) at least one general anesthesia code (L1211, L1212); and (3) at least one malignancy pathologic testing code (C5500, C5504, C5505, C5508, C5918, C5919, C5606-C5607).

3. Variable definitions

Demographic variables included sex, age, economic status, and comorbidities. Economic status was categorized by health insurance type—either standard health insurance or Medical Aid, a tax-funded program designed to ensure healthcare access for low-income citizens. Comorbidities during the year preceding the index lung cancer surgery were defined using claims data and International Classification of Diseases, 10th revision (ICD-10) codes, and summarized using the Charlson Comorbidity Index (CCI).

The variable for preoperative treatment was defined by whether patients received radiation or chemotherapy within 180 days prior to surgery-related hospitalization. Radiation therapy was identified based on claims for 25 procedure codes across seven categories, including intensity-modulated radiotherapy, the most commonly used modality for lung cancer treatment. Chemotherapy administration was determined using 32 lung cancer–specific drugs covered under the National Health Insurance (S1 Table). Patients who received both radiation therapy and chemotherapy were classified as having undergone neoadjuvant chemoradiotherapy (NeoCCRTx), those who received chemotherapy alone as having undergone neoadjuvant chemotherapy (NeoCTx), and all others as having undergone upfront surgery.

The surgical approach was categorized as either video-assisted thoracoscopic surgery (VATS) or open surgery. Procedures were classified as VATS if claims for endoscopic procedures (N0031001, N0031002) were present; otherwise, they were categorized as open surgery. Surgical extents were classified as wedge resection (O1401, O1403, O1404, O1405), segmentectomy (O1410), lobectomy (O1421, O1422, O1423, O1424), or pneumonectomy (O1431, O1432) based on 11 procedure codes. If multiple procedures were performed during a single surgery, classification was prioritized in the following order: pneumonectomy > lobectomy > segmentectomy > wedge resection.

Healthcare institution characteristics were categorized by hospital type (tertiary or general hospital), geographic region (metropolitan: Seoul, Incheon, or Gyeonggi), and administrative division (district, city, or county). Finally, surgical outcomes included in-hospital mortality, 30-day mortality, and average length of stay.

4. Statistical analysis

Descriptive statistics were used to assess the distribution of annual surgical cases based on demographic, surgical treatment, and healthcare institution characteristics. Annual trends in surgical volume, patient characteristics, and outcomes were analyzed using annual percentage change (APC), estimated via a log-linear regression model under the assumption of a consistent trend. APC values and their statistical significance are presented in S2 Table. Logistic regression was used to analyze associations between each characteristic and either in-hospital or 30-day mortality, while linear regression identified factors influencing the length of hospital stay. A p-value of < 0.05 was considered statistically significant for all analyses. All analyses were conducted using SAS Enterprise Guide 7.1 (SAS Institute Inc.) and R 4.4.0 (R Foundation for Statistical Computing).

1. Total number of lung cancer operations and surgical approaches

Between 2010 and 2023, a total of 124,334 lung cancer operations were identified using the health insurance claims database, and an additional 1,740 robotic lung cancer operations were estimated by Intuitive Surgical Korea Limited. During this period, the annual number of operations in South Korea increased from 4,557 in 2010 to 14,184 in 2023, representing more than a threefold growth (APC, 8.86%; p < 0.001). Fig. 1A illustrates the trend in the total number of lung cancer operations, including robotic operations.

Fig. 1.

Trends in pulmonary resection of lung cancer. Number of lung cancer surgery cases from 2010 to 2023 in Korea. Robotic surgery numbers were estimated in tens. (A) Total cases. (B) By various surgical approaches. VATS, video-assisted thoracoscopic surgery.

The number of operations by approach is shown in Fig. 1B. VATS became the predominant surgical approach, increasing from 52.9% of cases in 2010 to 94.8% in 2023 (APC, 4.11%; p < 0.001). Conversely, the proportion of open thoracotomy procedures decreased from 45.6% to 2.1% (APC, –20.75%; p < 0.001). Although robotic surgery represented a relatively small fraction, it accounted for 3.17% of all procedures, with 450 cases in 2023—surpassing the number of open operations for the first time (APC, 11.49%; p < 0.001). However, as detailed patient information on robotic surgery was unavailable, these cases were excluded from further analysis.

2. Patient demographics

The proportion of female patients undergoing surgery increased from 32.0% in 2010 to 44.7% in 2023 (APC, 2.39%; p < 0.001). The percentage of patients aged 70-79 years rose from 26.3% to 32.3% (APC, 1.60%; p < 0.001), and those aged 80 years or older increased from 2.0% to 6.2% (APC, 9.63%; p < 0.001). In contrast, there was no significant change in the proportion of patients aged 60-69 years (APC, 0.30%; p=0.178), whereas the proportions of younger age groups declined (APC for < 40, 40-49, and 50-59 years: –1.94%, –3.51%, and –3.37%, respectively; all p < 0.05). The proportion of patients with CCI scores of 7 or higher increased from 9.0% in 2010 to 17.4% in 2023, indicating a trend toward performing surgery on patients with greater comorbidities (APC, 6.07%; p < 0.001). Health insurance remained the predominant payer, covering 95.7% of operations in 2010 and 95.2% in 2023 (APC, –0.04%; p=0.013), while the proportion of Medical Aid cases increased slightly from 4.3% to 4.8% (APC, 0.91%; p=0.017) (Table 1).

Table 1.

Basic characteristics of patients who underwent pulmonary resection for lung cancer

3. Surgical treatment

Lobectomy remained the most common procedure; however, its proportion continuously declined from 82.7% in 2010 to 63.1% in 2023 (APC, –1.91%; p < 0.001). In contrast, sub-lobar resections increased, with wedge resection rising from 9.0% to 18.5% (APC, 5.72%; p < 0.001) and segmentectomy increasing from 3.0% to 17.9% (APC, 14.63%; p < 0.001). Pneumonectomy declined sharply from 5.3% to 0.5% (APC, –15.43%; p < 0.001) (Table 1, Fig. 2A).

Fig. 2.

Trends of pulmonary resection of lung cancer. Number of lung cancer surgery cases from 2010 to 2023 in Korea. (A) Extent of surgical resection excluding robotic surgery. (B) Distribution by preoperative treatment, excluding robotic surgery. CCRTx, concurrent chemoradiotherapy; CTx, chemotherapy.

Upfront surgery, performed without neoadjuvant treatment, accounted for 94.2% of cases in 2010 and increased to 98.2% in 2023 (APC, 0.30%; p < 0.001). Although the absolute number of operations following neoadjuvant therapy remained stable—averaging 162 procedures annually for NeoCCRTx and 155 for NeoCTx—the proportions of patients receiving NeoCCRTx (declining from 2.2% to 0.8%; APC, –6.10%; p < 0.001) and NeoCTx (declining from 3.5% to 1.1%; APC, –8.64%; p < 0.001) both decreased (Table 1, Fig. 2B).

4. Institutional and regional distribution

The proportion of operations performed at tertiary hospitals increased from 75.9% in 2010 to 81.9% in 2023 (APC, 0.64%; p < 0.001), while that of operations at general hospitals declined from 24.1% to 18.1% (APC, –2.95%; p < 0.001). The proportion of operations in metropolitan areas remained stable, changing from 76.6% to 76.4% (APC, 0.03%; p=0.671). In 2023, when classified by administrative division, district-level institutions accounted for 71.5% of cases, city-level institutions for 23.1%, and county-level institutions for 5.4%, with minimal changes over time (Table 1).

5. Length of hospital stay and operative mortality

The median hospital stay decreased from 13 days in 2010 to 7 days in 2023 (APC, –4.34%; p < 0.001), a trend confirmed by multivariable linear regression analysis. Additionally, linear regression identified several factors associated with prolonged hospitalization, including male sex, older age, Medical Aid status, higher CCI scores (≥ 7), neoadjuvant therapy, open surgery, broader surgical extent, treatment at general hospitals, and care at provincial or city district-level hospitals (Table 2, S3 Table).

Table 2.

Multivariable analysis for in-hospital mortality, 30-day mortality and hospital stay

In-hospital mortality declined from 1.74% in 2010 to 0.48% in 2023 (APC, –8.26%; p < 0.001), while 30-day mortality decreased from 2.45% to 0.76% (APC, –8.32%; p < 0.001) (Fig. 3). This declining trend was also evident in multivariable logistic regression analysis. Higher mortality rates were observed among male patients, older individuals, Medical Aid beneficiaries, and those with higher CCI scores (≥ 5). Patients undergoing open surgery, more extensive procedures (lobectomy and pneumonectomy versus wedge resection), or neoadjuvant therapy experienced significantly higher mortality, particularly in general hospitals and provincial regions (Table 2, S4 and S5 Tables).

Fig. 3.

Trends of surgical outcomes. (A) In-hospital and 30-day mortality rates. (B) Hospital stays (day), presented as median (line) and interquartile range (shaded area).

1. Increasing lung cancer incidence and surgical volume

The rise in lung cancer operations corresponds to an increasing crude incidence rate of lung cancer in Korea—from 42.8 to 61.6 per 100,000 between 2010 and 2021. However, the age-standardized incidence rate remained stable (62.9 vs. 59.3 per 100,000), indicating that lung cancer risk within individual age groups has not significantly changed. This suggests that the overall increase is driven by a growing population of older adults, who exhibit a higher incidence than younger individuals [17]. This trend is also seen in our study, with higher proportions of older and high-comorbidity patients undergoing lung cancer surgery.

The proportion of female patients undergoing surgery increased significantly from 32.0% to 43.5% between 2010 and 2021, outpacing the overall increase in female lung cancer diagnoses (from 29.5% to 33.0%) during the same period. A key factor is the higher proportion of localized-stage lung cancer among women, as shown by Surveillance, Epidemiology, and End Results (SEER) staging, which increased from 21.1% in 2010 to 37.3% in 2021 (compared to 18.2% to 24.5% in men) [17]. Since most female lung cancer patients in South Korea are never-smokers, this trend aligns with the global rise in lung cancer among never-smokers (LCINS) [5,13]. Potential risk factors for LCINS include secondhand smoke exposure, cooking fumes, radon, air pollution, occupational hazards, genetic susceptibility, and hormonal replacement therapy, although definitive causal relationships remain unclear [18]. Moreover, adenocarcinoma—the predominant histologic subtype in never-smokers—is often peripherally located and less aggressive, making these patients more likely candidates for surgery [19,20].

While lung cancer incidence in Korea increased 1.5-fold from 21,373 cases in 2010 to 31,616 in 2021, the number of lung cancer operations tripled during the same period (from 4,557 in 2010 to 12,609 in 2021). This suggests that improved early detection and expanding surgical indications have both contributed to the rise in operations.

2. Shift toward less invasive and lung-sparing surgery

Recent trends in Korea reveal a marked transition toward less invasive and lung-preserving surgical strategies for lung cancer. The proportion of patients receiving neoadjuvant therapy has steadily declined, while upfront surgery increased from 94.2% in 2010 to 98.2% in 2023. This shift is likely driven by earlier detection via widespread LDCT screening, as well as a prevailing preference among Korean thoracic surgeons for immediate resection due to concerns over disease progression and surgical complexity following neoadjuvant therapy. Furthermore, the clinical use of neoadjuvant immunotherapy remains limited, as it was only approved in Korea in October 2022 and is not yet widely reimbursed. Nevertheless, the use of neoadjuvant strategies, particularly immunotherapy, is expected to rise as evidence and clinical experience accumulate.

Minimally invasive approaches have become the dominant surgical modality, with VATS accounting for 97.8% of all cases in 2023. VATS is supported by international guidelines for its benefits in reducing postoperative complications, shortening recovery time, and preserving oncologic out-comes [21,22]. Although robotic-assisted thoracic surgery represented only 3.17% of cases in 2023, it surpassed open surgery volume for the first time. Unlike in the United States, where over 60% of lobectomies are now performed robotically according to the Society of Thoracic Surgeons General Thoracic Surgery Database [23], adoption in Korea remains modest due to limited insurance coverage and the earlier widespread implementation of VATS. Nonetheless, increasing familiarity with robotic systems may facilitate future uptake, particularly if evidence emerges supporting superior outcomes in Asian populations [24].

Parallel to the rise in minimally invasive techniques, the use of sub-lobar resections has expanded significantly. In 2023, sub-lobar resections accounted for 36.5% of all procedures, driven by growing evidence—such as from JCOG-0802/WJOG4607L and CALGB 140503—that these approaches offer oncologic equivalence to lobectomy in small peripheral tumors [6,25]. The use of pneumonectomy has declined markedly, often supplanted by lung-sparing techniques like sleeve lobectomy [26]. These collective trends reflect a broader shift toward surgical de-escalation, allowing more elderly and comorbid patients to become surgical candidates and contributing to the overall rise in lung cancer operations in Korea. This evolution in practice parallels global developments and highlights Korea’s alignment with evidence-based, patient-centered surgical care.

3. Improving surgical outcomes and addressing disparities

Surgical outcomes significantly improved from 2010 to 2023. The median hospital stay was reduced from 13 days to 7 days, in-hospital mortality decreased from 1.74% to 0.48%, and 30-day mortality dropped from 2.45% to 0.76%. Despite an increase in high-risk factors such as older age and higher CCI scores, these improvements are likely attributable to a higher proportion of favorable prognostic factors—such as more female patients, upfront operations, minimally invasive approaches, sub-lobar resections, and treatment at tertiary centers—which have collectively enhanced postoperative recovery and reduced mortality.

Disparities persist, with poorer outcomes observed among Medical Aid beneficiaries and patients treated at provincial hospitals, highlighting the need for enhanced perioperative care and social support programs. High-volume tertiary centers were associated with better outcomes, consistent with previous studies demonstrating a volume-outcome relationship in thoracic surgery [27]. Nonetheless, small-volume centers can achieve comparable results with structured surgical training, standardized perioperative protocols, comorbidity optimization, and multidisciplinary decision-making [28].

Currently, no national registry comprehensively tracks lung cancer surgery volume and outcomes in South Korea. In contrast, Japan’s National Clinical Database (NCD) is a nationwide online system that registers surgical patients as part of the Japanese Surgical Board Certification System. The NCD facilitates the collection of data on patient characteristics, operative information, and postoperative events [29]. Establishing a similar registry in South Korea would enable more precise data collection, facilitate policy development, and improve national surgical quality standards.

4. Limitations

This study has several limitations. As a claims-based analysis, it lacks detailed clinical data. First, the absence of precise staging information limits stage-specific mortality analysis. To address this, staging was approximated using neoadjuvant therapy administration, in line with SEER classification. Second, some cases may have involved biopsies or salvage surgery for advanced disease. However, non-surgical biopsy is generally preferred for advanced lung cancer, and the low rate of preoperative therapy suggests that most operations were intended to be curative. Third, robotic surgery data were unavailable in the claims database and were instead estimated separately. Given that robotic operations account for only a very small proportion, their exclusion likely had minimal impact on our findings. Despite these limitations, this study represents the first nationwide, long-term analysis of lung cancer surgery trends in Korea, providing valuable data for future policy and clinical improvements.