Abstract
Objective:
Segmentectomy has become an accepted procedure for the treatment of non–small cell lung cancer. Adequate lymph node sampling, sufficient margins, and proper tumor size selection are factors vital for achieving outcomes comparable to lobectomy. Previous studies have demonstrated poor adherence to lymph node sampling guidelines. However, national trends in the quality of segmentectomy and implications on survival are unknown.
Methods:
The National Cancer Database was used to identify patients with clinical stage I to IIA non–small cell lung cancer surgically treated between 2004 and 2018. Facility-level trends in extent of resection and segmentectomy odds of adherence to (1) 2014 Commission on Cancer guidelines of sampling 10 or more lymph nodes, (2) negative (R0) resection margins, and (3) tumor size 2 cm or less were determined. Propensity score matching was based on segmentectomy adherence to (4) a composite of all measures, and survival was evaluated with Cox models and Kaplan–Meier survival estimates.
Results:
The study included 249,391 patients with 4.4% (n = 11,006) treated with segmentectomy. The proportion of segmentectomies performed annually increased from 3.3% in 2004 to 6.1% in 2018 (P < .001). Overall, 12.6% (n = 1385) of patients who underwent segmentectomy between 2004 and 2018 were adherent to all measures, and adherence was more likely at academic programs (odds ratio, 1.56; 95% confidence interval, 1.14–2.15) than nonacademic programs (P < .001, reference). Adherence to all measures was associated with improved survival (hazard ratio, 0.67; 95% confidence interval, 0.56–0.79).
Conclusions:
As segmentectomy is increasingly established as a valid oncological option for the treatment of non–small cell lung cancer, it is important that quality remains high. This study demonstrates that continued improvement is needed.
Keywords: lung cancer, National Cancer Database, non–small cell lung cancer, outcomes, quality, segmentectomy
Graphical Abstract
Lung cancer is the second most common cancer and is the leading cause of cancer death in the United States.1 Anatomic lobectomy with mediastinal lymph node sampling has been the standard of care for management of early-stage non–small cell lung cancer (NSCLC) since the randomized controlled trial conducted by the Lung Cancer Study Group.2 Segmentectomy is commonly considered to be more technically challenging than lobectomy and was once considered a less optimal option primarily appropriate for patients with poorer preoperative pulmonary function.3–6 Over time, segmentectomy emerged as a potentially acceptable alternative to lobectomy for select patients, with several studies showing no difference in overall survival (OS), disease-free survival, or perioperative morbidity and mortality.6–18 More recently, JCOG0802/WJOG4607L reported the superiority of segmentectomy over lobectomy in OS.6 To achieve comparable outcomes to lobectomy, segmentectomy must be performed with adequate lymph node sampling and obtain sufficient resection margins on patients appropriately selected by tumor size.6,14,19–24
Segmentectomy is well described at experienced programs. However, little is known about segmentectomy used outside of the academic or research setting. Specifically, the trends in prevalence and disparities in quality of segmentectomy are unknown. Our hypothesis is that segmentectomy is becoming more common, but with facility-level disparities in quality. The objectives of this study were to describe national trends in the prevalence of segmentectomy for NSCLC in the United States, determine adherence to oncologic quality measures (QMs) for patients receiving segmentectomy, evaluate facility-level disparities in both segmentectomy prevalence and quality, and evaluate the association between segmentectomy quality and survival.
MATERIALS AND METHODS
Data Source and Study Cohort
The National Cancer Database (NCDB) is a multicenter hospital-based cancer registry that is estimated to include approximately 82% of annual cancers of the lung and bronchus in the United States.25,26 More than 1500 hospitals are approved to contribute to the NCDB, with patient data extracted by trained Certified Tumor Registrars. The NCDB was used to identify patients who had been diagnosed between 2004 and 2018 with NSCLC and had also undergone surgical treatment at the reporting facility. Inclusion criteria were detailed in a flow diagram (Figure 1). Patient information reported in the NCDB is deidentified and publicly available, and this research was determined to be exempt from review by our institution’s Institutional Review Board.
FIGURE 1.
Inclusion criteria flow diagram. The NCDB was used to identify patients with NSCLC diagnosed between 2004 and 2018. Population excluded at each stage of analysis is included on the right. The cohorts that underwent analysis are detailed in the middle. The type of analysis performed on each cohort is detailed on the left. Cox PH, Cox proportional hazard; KM SE, Kaplan–Meier survival estimate; MIS, minimally invasive surgical approach.
Quality Measures
To facilitate assessment of segmentectomy quality, we selected 3 QMs based on a combination of published guidelines and the findings from existing literature on segmentectomy. These QMs include the following.
QM1: 10 or more regional lymph nodes sampled.
Guideline adherence was evaluated for the Commission on Cancer (CoC)-published lymph node sampling QM (QM1). Based on CoC guidelines published in 2014, successful adherence was defined as at least 10 regional lymph nodes being sampled.27 Patients who had less than 10 lymph nodes sampled were considered to have failed to adhere to this QM. Notably, patients were not limited to pathological stage IA to IIB per the guideline definitions because this would exclude patients who were upstaged to a stage greater than IIB. Updated CoC guidelines published in 2020 are not available in the NCDB for evaluation.
QM2: Negative (R0) resection margins.
Successful adherence to QM2 was defined as a pathological report stating negative microscopic margins (R0). Failure was defined as positive microscopic or macroscopic margin status (R1, R2) or unspecified residual tumor. The NCDB does not provide data on margin distance.
QM3: Tumor size 2 cm or less.
Adherence to recommended tumor size of 2 cm or less for patients treated with segmentectomy was determined.6,10,16,20,28–30 Tumor size (diameter) is recorded in the NCDB as the largest dimension listed in a patient’s pathological report entered to the nearest millimeter. Patients with tumor size more than 2 cm or missing tumor size (<1%) were considered to fail to adhere to this QM. Those with missing tumor size were excluded from survival analysis.
QM4: Composite adherence to all measures.
Successful adherence to all measures (QM1–3) for patients undergoing segmentectomy was determined. Any failure of QM1–3 was considered a QM4 failure.
QM5: Composite adherence to all measures with minimally invasive surgical approach.
Because the minimally invasive surgical (MIS) approach has been associated with improved perioperative outcomes, an additional QM incorporating operative approach was investigated.31 Segmentectomy successfully adherent to QM4 performed with the MIS approach (including robotic and video-assisted thoracoscopic surgery) versus open thoracotomy was defined as QM5. Operative approach data were available from 2010 to 2018, and patients who underwent the MIS approach and were converted to open were included in the MIS approach category.
Patient and Clinical Characteristics
NCDB data included patient age, race and ethnicity, sex, year of diagnosis, sociodemographic characteristics, geographic characteristics, comorbidities, and clinical characteristics. Patient age was treated as a continuous variable. Race and ethnicity were categorized as non-Hispanic White, non-Hispanic Black, Hispanic, Asian, and other or unknown. Income and educational attainment level were determined by ZIP code tabulation area estimates categorized as quartiles. Patients with missing data for income or educational attainment were categorized as missing. Insurance payor status was categorized as private or Medicare, Medicaid or no insurance, and other, unknown, or missing insurance. Rurality in the NCDB is categorized using Rural-Urban Continuum Codes, which include 9 levels that are differentiated by population density, adjacency to metropolitan counties, and commuter activity. Rural-Urban Continuum Code levels 1 to 4 were categorized as large urban and metropolitan (urban), levels 5 to 9 were categorized as small-town and completely rural (rural), and missing data were categorized as missing. Comorbidities in the NCDB are reported with the Charlson-Deyo index, which is grouped as 0, 1, 2, and greater than or equal to 3. Categories of tumor size were generated for analysis and included incomplete/missing, 1 or less, more than 1 to 2, more than 2 to 3, more than 3 to 5, more than 5 to 7, or more than 7 cm. Clinical stage was used in multivariable analysis because it reflected the information available for operative planning. Pathological stage was also included in bivariate analysis to compare staging drift. Both clinical and pathological stage were categorized as American Joint Committee on Cancer stage 0, IA, IB, IIA, IIB, IIIA, IIIB, IIIC, and IV. Histology was categorized as adenocarcinoma, squamous, large cell, carcinoid, and other (Table E1). The extent of surgical resection was categorized as wedge resection, segmentectomy (including lingulectomy), lobectomy (including bilobectomy and bronchial sleeve resection), and pneumonectomy. Patients with unknown or other extent of surgical resection were excluded.
Hospital Characteristics
Travel distance from the treatment facility to the patient’s address was calculated as the straight-line distance between the center of the patient’s ZIP code to the address of the treating facility. Travel distance was analyzed as a continuous variable. Patients with missing travel distance were excluded. Facility type was determined by the CoC accreditation program and categorized as academic research programs versus nonacademic programs. Academic research programs include Academic Comprehensive Cancer Programs as well as National Cancer Institute-designated Comprehensive Community Programs. Nonacademic programs include Comprehensive Community, Integrated Network, and Community programs. CoC cancer program designations reflect organization, capabilities, and the volume of new cancers treated. Annual hospital volume was defined as the facility-level surgical volume divided by the number of years the reporting facility was accredited to contribute to the CoC during the study period. Hospital surgical volume was categorized using LeapFrog criteria with a cutoff of 40 or more annual pulmonary resections to designate high-volume hospitals.32
Statistical Analysis
Descriptive statistics and bivariate analysis.
Chi-square tests were used to determine the significance of differences between categorical variables, including facility subcategories, by extent of resection and by adherence to each QM. Mean and standard deviation were reported for normally distributed continuous variables. Median and interquartile range were reported for non-normally distributed continuous variables. Student t tests and Mann–Whitney tests were used as appropriate to evaluate differences between continuous variables by adherence to QMs and facility type. Overall trends in the extent of surgical resection were determined, and academic programs were compared with nonacademic programs. Additionally, trends in successful adherence to QM1–4 were determined. Cochran-Armitage tests were used to assess statistical changes of nonparametric variables over time.
Multivariable logistic regression.
Model 1: Odds of receiving segmentectomy.
All patients who underwent surgical treatment were evaluated with multivariable logistic regression to determine associations between receipt of segmentectomy (Model 1) with facility type and annual surgical volume while adjusting for year of diagnosis, age, sex, race and ethnicity, rurality, travel distance, income, education, insurance status, Charlson-Deyo score, clinical stage, tumor size, and histology.
Models 2–4: Odds of segmentectomy quality measure adherence.
Patients who underwent segmentectomy were evaluated with separate multivariable logistic regression models for each QM (Models 2–4: QM1–3) to determine associations between segmentectomy QM adherence with facility type and annual surgical volume while adjusting for year of diagnosis, age, sex, race and ethnicity, rurality, travel distance, income, education, insurance status, Charlson-Deyo score, clinical stage, tumor size (excluded from Model 4: QM3 due to collinearity), and histology.
Model 5: Odds of successful adherence to all segmentectomy quality measures.
Patients who underwent segmentectomy were evaluated with multivariable logistic regression to determine associations between successful adherence to all QMs (Model 5: QM4) with facility type and annual surgical volume while adjusting for year of diagnosis, age, sex, race and ethnicity, rurality, travel distance, income, education, insurance status, Charlson-Deyo score, clinical stage, and histology.
All multivariable logistic regression model standard errors were adjusted for clustering of observations within hospitals and were reported with odds ratios (ORs) and 95% confidence intervals (CIs). Unless specified as excluded, missing data categories were included in multivariable logistic regression models to decrease bias introduced by nonrandom patient exclusion.26,33 Sensitivity analyses were performed with exclusion of missing data (Table E2).26,34
Propensity score matching.
Two propensity score–matched (PSM) models were created to evaluate QM4 (2004–2017) and QM5 (2010–2017). Propensity scores were derived from the probability of adherence to the QM adjusted with Model 5 covariates and for facility-level clustering. Patients who received neoadjuvant therapies or had missing data were excluded (Figure 1). An interaction term between facility type and annual surgical volume was included. The caliper was defined as standard deviation of the propensity score multiplied by 0.2.35 Nearest-neighbor matching without replacement was performed using the logit of the propensity score. Finally, matching was assessed by evaluating standardized mean differences of covariates using a cutoff of 0.1.E1 Sensitivity analyses were conducted with inclusion of missing data as covariates with unchanged inference in survival analysis (Table E2).26,33,34,E2
Survival analysis.
OS was analyzed using unadjusted (before PSM) and adjusted (after PSM) Cox proportional hazards models and Kaplan–Meier survival estimates with log-rank tests. All analyses were done with Stata Version 17.
RESULTS
The study included 249,391 patients with clinical stage I to IIA NSCLC who underwent surgical treatment at 1208 reporting facilities between 2004 and 2018 (Figure 1). Segmentectomy was performed in 4.4% (n = 11,006) of cases (Figure 2). Inclusion criteria at each stage of analysis are shown in Figure 1.
FIGURE 2.
Segmentectomy is becoming more popular. Overall QM adherence is poor. Academic programs are more likely to successfully adhere to segmentectomy QMs compared with nonacademic programs. Better quality is associated with improved survival. Improvement is needed with lymph node sampling. OR, Odds ratio; HR, hazard ratio; CI, confidence interval;.
Trends in Extent of Resection
There was a trend toward increased use of parenchymal-sparing surgical techniques between 2004 and 2018 (Figure E1, Table E3, P<.001). Compared with other types of lung resection, the proportion of segmentectomies performed increased from an average of 3.3% per year in 2004 to 6.1% in 2018 (P<.001, Figure 3). Further, the proportion of segmentectomies performed in relation to other types of resection (wedge resection, lobectomy, or pneumonectomy) increased more rapidly in academic programs versus nonacademic programs (Figure E1). Overall, academic programs performed 45.5% (n = 5007) of segmentectomies, whereas 54.5% (n = 5999) were performed at nonacademic programs (P <.001, Table 1). Additionally, 63.3% (n = 6972) of segmentectomies were performed at facilities performing 40 or more annual lung resections, whereas 36.7% (n = 4034) were performed at facilities performing less than 40 annual lung resections (P<.001, Table 1). The median number of lymph nodes sampled with segmentectomy was 5 (IQR, 2–10), with a median of 6 (IQR, 3–12) for those treated at academic programs and 4 (IQR, 1–9) for those treated at nonacademic programs (P <.001). For comparison, the median number of lymph nodes sampled by lobectomy was 9 (IQR, 5–15), with 10 (IQR, 6–17) for those treated at academic programs and 9 (IQR, 5–14) for those treated at nonacademic programs (P <.001).
FIGURE 3.
National trends in segmentectomy for NSCLC: 2004–2018. Overall trends in the proportion of segmentectomies performed annually from 2004 to 2018 as a function of all therapeutic lung resections performed nationally. Academic program versus nonacademic program trends in the proportion of segmentectomies performed annually from 2004 to 2018 as a function of all therapeutic lung resections performed nationally.
TABLE 1.
Characteristics of patients who underwent segmentectomy adherent to all measures with multivariable logistic regression model 5 evaluating associations with adherence
| All segmentectomy quality measures |
|||||
|---|---|---|---|---|---|
| QM4 |
Failed to achieve all measures |
Achieved all measures |
Total |
||
| N |
9621 |
1385 |
11,006 |
Model 5 |
|
| Parameter | % | % | N | P value | OR (95% CI) |
|
| |||||
| Program type | <.001 | 1.56 (1.14–2.15) | |||
| Nonacademic | 90.9 | 9.1 | 5999 | Reference | |
| Academic | 83.2 | 16.8 | 5007 | ||
| Annual surgical volume | <.001 | ||||
| <40 | 92.2 | 7.8 | 4034 | Reference | |
| ≥40 | 84.6 | 15.4 | 6972 | 1.56 (1.18–2.07) | |
| Year of diagnosis | <.001 | ||||
| 2004* | - | - | - | ||
| 2005 | 94.0 | 6.0 | 283 | ||
| 2006 | 94.0 | 6.0 | 285 | ||
| 2007 | 91.6 | 8.4 | 381 | ||
| 2008 | 93.0 | 7.0 | 532 | ||
| 2009 | 92.9 | 7.1 | 630 | ||
| 2010 | 93.1 | 6.9 | 773 | ||
| 2011 | 93.2 | 6.8 | 750 | ||
| 2012 | 90.0 | 10.0 | 823 | ||
| 2013 | 88.6 | 11.4 | 930 | ||
| 2014 | 87.3 | 12.7 | 942 | ||
| 2015 | 83.8 | 16.2 | 1002 | ||
| 2016 | 85.5 | 14.5 | 1089 | ||
| 2017 | 80.2 | 19.8 | 1188 | ||
| 2018 | 77.4 | 22.6 | 1170 | ||
| Sex | .02 | ||||
| Female | 86.8 | 13.2 | 6510 | ||
| Male | 88.3 | 11.7 | 4496 | ||
| Age, mean (SD) | 68.3 (9.2) | 69.7 (9.3) | <.001 | ||
| Race and Ethnicity | .02 | ||||
| Non-Hispanic White | 87.3 | 12.7 | 9221 | ||
| Non-Hispanic Black | 88.0 | 12.0 | 709 | ||
| Hispanic | 90.4 | 9.6 | 270 | ||
| Asian | 82.3 | 17.7 | 254 | ||
| Other or unknown | 89.9 | 10.1 | 552 | ||
| Rurality | .02 | ||||
| Urban | 87.4 | 12.6 | 9752 | ||
| Rural | 89.1 | 10.9 | 879 | ||
| Missing | 83.5 | 16.5 | 375 | ||
| Travel distance, median (IQR) | 11.2 (4.9–28.2) | 12.0 (5.1–30.9) | .006 | ||
| Income quartiles | .11 | ||||
| Lowest | 86.8 | 13.2 | 1551 | ||
| 2 | 88.4 | 11.6 | 1987 | ||
| 3 | 88.4 | 11.6 | 2495 | ||
| Highest | 86.7 | 13.3 | 4879 | ||
| Missing* | - | - | - | ||
| Education quartiles | .08 | ||||
| Lowest | 88.0 | 12.0 | 1698 | ||
| 2 | 87.6 | 12.4 | 2572 | ||
| 3 | 88.2 | 11.8 | 3289 | ||
| Highest | 86.1 | 13.9 | 3363 | ||
| Missing* | - | - | - | ||
| Insurance status | .05 | ||||
| Medicare or private | 87.3 | 12.7 | 10,340 | ||
| Medicaid or uninsured | 88.3 | 11.7 | 478 | ||
| Unknown or other | 93.1 | 6.9 | 188 | ||
| Charlson-Deyo score | <.001 | ||||
| 0 | 85.8 | 14.2 | 5308 | ||
| 1 | 88.1 | 11.9 | 3650 | ||
| 2 | 90.3 | 9.7 | 1423 | ||
| ≥3 | 90.2 | 9.8 | 625 | ||
| Histology | .009 | ||||
| Adenocarcinoma | 86.8 | 13.2 | 7172 | ||
| Squamous | 89.5 | 10.5 | 2597 | ||
| Large Cell | 87.6 | 12.4 | 217 | ||
| Carcinoid | 86.3 | 13.7 | 702 | ||
| Other | 86.5 | 13.5 | 318 | ||
| Clinical stage | <.001 | ||||
| IA | 85.5 | 14.5 | 9107 | ||
| IB | 97.4 | 2.6 | 1539 | ||
| IIA | 93.9 | 6.1 | 360 | ||
QM, Quality measure; OR, odds ratio; CI, confidence interval; SD, standard deviation; IQR, interquartile range.
Year of diagnosis 2004, missing income, and missing education censored due to small cell size per NCDB requirements.
Segmentectomy Quality Measure Adherence
QM1: 10 or more regional lymph nodes sampled.
Overall, adherence to (QM1) CoC lymph node sampling guidelines was met in 21.2% (n = 2339) of patients (P <.001, Table E4). Patients treated at academic programs were more likely to meet guidelines (OR, 1.53; 95% CI, 1.12–2.08, vs nonacademic) (Figure E2). Patients treated at hospitals that performed 40 or more annual lung resections also were more likely to adhere to lymph node sampling guidelines (OR, 1.69; 95% CI, 1.30–2.21, vs<40 annual lung resections; Table E4). Patients with care adherent to QM1 had higher rates of nodal upstaging (7.6% cN0 to pN1–3 vs 4.1% cN0 to pN1-N3, QM1 nonadherent) and lower rates of incomplete formal TNM staging on final pathology (3.8% vs 13.0% QM1 nonadherent, all P <.001).
QM2: Negative (R0) resection.
(QM2) An R0 resection was achieved in 96.2% (n = 10,586) of patients treated with segmentectomy (P <.001, Table E5). Patients treated at academic programs were more likely to receive an R0 resection with segmentectomy (OR, 1.53; 95% CI, 1.09–2.15, vs nonacademic) (Table E5, Figure 4, and Figure E2). Increasing tumor size was associated with decreased odds of R0 resection (OR, 0.48; 95% CI, 0.31–0.75 for >3–5 cm, vs >1–2 cm) (Table E5).
FIGURE 4.
Association between academic programs and successful adherence to QMs for patients surgically treated with segmentectomy. Separate multivariable logistic regression models were used to evaluate each QM. All models were adjusted for patient, sociodemographic, clinical, geographic, and facility-level characteristics. Each model is graphically displayed on the plot with ORs and 95% CIs. ORs greater than 1.0 favor academic programs. OR, Odds ratio; CI, confidence interval; QM1, quality measure; QM2, quality measure 2; QM3, quality measure 3; QM4, quality measure 4. †Nonacademic program is the reference for each reported OR.
QM3: Tumor size 2 Cm or less.
(QM3) Patient tumor size was 2 cm or less in 60.8% (n = 6691) of all cases treated with segmentectomy (P <.001, Table E6 Of segmentectomies that were adherent to QM3, 20.9% had tumor size 1 cm less and 79.1% had tumor size 1 to 2 cm (P <.001). Patients treated at academic programs were more likely to undergo segmentectomy with tumor size 2 cm or less (OR, 1.24; 95% CI, 1.08–1.43, vs nonacademic; Figures E2 and 4). Patients who underwent segmentectomy at a hospital with an annual surgical volume 40 or more had no statistically significant difference in QM3 adherence compared with hospitals performing less than 40 annual lung resections in multivariable analysis (Table E6).
QM4: Composite adherence to all measures.
Overall, 12.6% (n = 1385) of patients who underwent segmentectomy between 2004 and 2018 met (QM1) CoC lymph node sampling guidelines and had (QM2) an R0 resection with (QM3) tumor 2 cm or less (Table 1). Adherence improved from 6.0% in 2005 (OR, 0.14; 95% CI, 0.17–0.78) to 22.6% in 2018 (OR, 1.41; 95% CI, 1.02–1.95; reference 2015, all P <.001) (Table 1, Figure 5). In the entire study period, there are notable facility-level disparities in overall quality of segmentectomy between academic programs (16.8% adherent, OR, 1.56; 95% CI, 1.14–2.15) and nonacademic programs (9.1% adherent, reference, P <.001) (Table E7, Figures 4 and 5). Although overall QM adherence improved at both academic and nonacademic programs between 2004 and 2018, adherence at academic programs remained significantly higher than in nonacademic programs (Figures 4 and 5).
FIGURE 5.
National trends in rates of adherence to segmentectomy QMs with facility-level comparison between academic and nonacademic programs: 2004–2018. A, Overall QM adherence is displayed for each QM, including QM 1 (≥10 lymph nodes sampled), QM 2 (R0 resection), QM 3 (tumor size ≤2 cm), and QM 4 (adherence to all measures). B, Academic versus nonacademic program trends in adherence to segmentectomy QM 4 (adherence to all measures). QM1, Quality measure; QM2, quality measure 2; QM3, quality measure 3; QM4, quality measure 4.
Survival Analysis
Median survival for the PSM 2004–2017 cohort (n = 2102) was 120.8 (95% CI, 100.6–134.9) months for patients adherent to QM4 and 95.7 (95% CI, 78.5–108.6) months for those nonadherent to QM4 (Table E8, Figure 6). Adherence to QM4 was associated with improved OS in unadjusted (HR, 0.70; 95% CI, 0.61–0.80) and adjusted (HR, 0.67; 95% CI, 0.56–0.79) Cox proportional hazards models (Table E9). Kaplan–Meier survival estimates demonstrated an association with improved OS for patients adherent to QM4 (Table E9, Figure 6, P <.001).
FIGURE 6.
Kaplan–Meier survival curve examining the association between adherence to all QMs and OS among PSM patients who underwent segmentectomy: 2004–2018. CI, Confidence interval; SE, survival estimate; QM4, quality measure 4 (achieved all measures).
Adherence to QM5 was associated with improved OS in unadjusted (HR, 0.52; 95% CI, 0.41–0.67) and adjusted (HR, 0.67; 95% CI, 0.51–0.88, PSM cohort) Cox models (Tables E10 and E11).
DISCUSSION
Our analysis demonstrates that although lobectomy remains the most popular surgical option, segmentectomy is becoming a more common approach to treat NSCLC in the United States. There are distinct disparities in facility-level and overall QM adherence, especially with respect to the extent of lymph node sampling at the time of operation. Adherence to QMs is associated with improved OS.
Although lobectomy has remained the standard of care for resectable NSCLC, a growing body of literature suggesting noninferiority of segmentectomy to lobectomy for outcomes including OS, disease-free survival, and locoregional recurrence was published during our study period.8–10,12 Randomized controlled trials have demonstrated noninferiority in perioperative morbidity and mortality.15,16 These findings have led to increased interest in segmentectomy, and it is likely that interest will intensify with the recent publication of JCOG0802/WJOG4607L results demonstrating superiority of segmentectomy versus lobectomy in OS for early-stage NSCLC.6 Our study found increased uptake of segmentectomy as a proportion of all lung resections performed at academic research programs relative to nonacademic programs. However, due to their larger numbers, more segmentectomies are performed at nonacademic programs overall. Our study found patients treated with segmentectomy at nonacademic programs are less likely to receive key elements of surgical decision-making and technique described in the literature that define both a high-quality operation and appropriate patient selection.
Only 60.8% of segmentectomies in this study met the proposed tumor size criteria of 2 cm or less.6,10,16,28–30 Because segmentectomy has been shown to be a highly effective alternative to lobectomy for high-risk patients with poor pulmonary function, it is possible that some patients with primary tumor size greater than 2 cm were treated with segmentectomy for this reason.29 Also, select patients with part-solid/ground-glass tumors between 2.1 and 3 cm who undergo segmentectomy may have noninferior outcomes in OS and disease-free survival for early-stage NSCLC.E3 Accordingly, tumor size greater than 2 cm should not be viewed as an absolute contraindication to segmentectomy, and it is not expected that adherence to this QM would approach 100%. However, although there are select patients with tumors greater than 2 cm who would benefit from segmentectomy, tumor size 2 cm or less remains the contemporary cutoff for segmentectomy indications to achieve outcomes comparable to lobectomy.6,10,16,28–30
Segmentectomy is characterized by low rates of adherence to adequate lymph node sampling QMs in our data, which is defined as sampling of at least 10 lymph nodes in the mediastinal and hilar stations with anatomic resection of the primary tumor.27 This finding remains consistent with adherence rates to lymph node sampling analyzed in NSCLC previously. The low adherence rate is, in part, inherent to the smaller pathological specimen removed with a parenchymal-sparing technique. This is evident in our data that show only 21.2% of segmentectomy cases had guideline-adherent nodal assessment. Although this is more favorable than the 8.3% rate of nodal sampling seen with wedge resections, it highlights an opportunity to improve surgical quality. By contrast the rate of nodal sampling adherence with lobectomy (43.2%) is double and with pneumonectomy (64.5%) 3 times that seen with segmentectomy (P <.001). Despite low rates of guideline adherence, only 14% of patients underwent segmentectomy without some type of nodal assessment.
Two main outcomes have been examined regarding the number of lymph nodes sampled for NSCLC: patient survival and accurate cancer staging. Previous research suggests that median survival was best in patients with greater than 10 lymph nodes evaluated.E4–E6 Prior publications also suggest even higher lymph node sampling thresholds are needed for accurate cancer staging.E7,E8 However, the American College of Surgeons Oncology Group Z0030 trial demonstrated that the specific nodal stations sampled are more important than the absolute number of lymph nodes sampled.E9 More recent studies suggest that patient survival after segmentectomy may not be influenced by lymph node sampling, at least for patients with tumors 2 cm or less.E10 Furthermore, in 2020, the CoC updated the lymph node guidelines for NSCLC to focus on lymph node stations.E11,E12 However, these changes are not yet reflected in the NCDB, and this study of patients diagnosed between 2004 and 2018 is the most up-to-date nationwide analysis of segmentectomy lymph node QMs of which we are aware.
Study Limitations
There are some limitations to this study. The NCDB may not capture nodes retrieved from mediastinal staging procedures including endobronchial ultrasound or cervical mediastinoscopy, which could lead to underestimates. Tumor size, as reported by the NCDB, is generated primarily from pathological reports and clinical to pathological concordance is unknown, although estimated to be 70% to 80% in exploratory analysis. Other important limitations are that the presence of spread through air spaces and parenchymal surgical margin distance relative to tumor size are not available in the NCDB.E13–E16 We also cannot determine the presence of practice patterns that result in nonanatomic segmentectomy, although the prevalence of this is unknown. Despite 1208 cancer programs contributing patient data directly to this study, the NCDB only collects data at CoC-accredited hospitals and the findings may not be generalizable to care provided at nonaccredited hospitals.E17 However, the NCDB provides data on more than 80% of lung cancer diagnoses as well as greater than 70% of cancer diagnoses overall, making it the largest national cancer registry available.25
CONCLUSIONS
As segmentectomy is increasingly established as a valid oncological option for the treatment of NSCLC, it is important that quality remains high. Although segmentectomy quality increased between 2004 and 2018, academic programs performed significantly better than nonacademic programs. Better quality of segmentectomy was associated with improved survival. Overall, continued improvement is needed.
Webcast
You can watch a Webcast of this AATS meeting presentation by going to: https://www.aats.org/resources/1544.
Supplementary Material
PERSPECTIVE.
As segmentectomy is increasingly established as a valid oncological option for the treatment of NSCLC, it is important that quality remains high. Although segmentectomy quality increased between 2004 and 2018, academic programs performed better than nonacademic programs and better quality is associated with improved survival. Overall, continued improvement is needed in lymph node sampling.
CENTRAL MESSAGE.
As segmentectomy is increasingly established as a valid oncological option for the treatment of NSCLC, it is important that quality remains high. This study demonstrates that improvement is needed.
Acknowledgments
Research reported in this publication was supported by the National Institute on Minority Health and Health Disparities of the National Institutes of Health under Award Number T37MD014248 (C.D.L.), the National Cancer Institute under Award Number K07 CA216330 (D.D.O.), and the National Heart, Lung, and Blood Institute under Award Number R01HL145478 (A.B.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Abbreviations and Acronyms
- CI
confidence interval
- CoC
Commission on Cancer
- HR
hazard ratio
- MIS
minimally invasive surgical
- NCDB
National Cancer Database
- NSCLC
non–small cell lung cancer
- OR
odds ratio
- OS
overall survival
- PSM
propensity score match
- QM
quality measure
- SE
survival estimate
Biography
Discussion
Presenter: Charles Logan
Dr Bernard Park (New York, NY). You are congratulated for a well-executed and presented study that has used the NCDB to evaluate the trend and performance of segmentectomy for the primary surgical treatment of localized stage I and II lung cancer. Over a 14-year period, showing an increasing trend in both the performance of segmentectomy and the quality of segmentectomy resections. However, despite these positive trends, the study emphasizes the need for improvements in both the use of segmentectomy in appropriate patients overall and the adherence of quality measures, both in academic and community centers. You mentioned that 14% of patients undergoing segmentectomy had no lymph nodes in the specimen. What proportion of these patients do you think had, in essence, a wedge resection? How would you propose to more accurately ascertain the rate of true anatomic segmentectomy, or do you think this is not a relevant query?
Dr Charles Logan (Seward, Alaska). In the NCDB, segmentectomies and wedge sections are separate variables. Now, in the community, the prevalence of a stapled segmentectomy or a large wedge resection is likely higher, but we don’t know what the prevalence is. It is possible that some of those patients are, in essence, a wedge resection. However, when we looked back and at all patients who were missing lymph nodes, we saw that you have missing lymph nodes for patients who underwent lobectomies or even pneumonectomies, and so it’s possible that all these patients in our study who had no lymph nodes are actually anatomic segmentectomies, and it’s alarming that 14% did not receive any type of lymph node sampling with resection.
Dr Park. Great. The lymph node assessment quality measure was based on an older 2014 CoC guideline that has recently been changed from lymph node count to stations assessed. Will this be measurable in the NCDB in hindsight or moving forward, and how would this affect the results of the current study should you choose to repeat it with the newer guideline?
Dr Logan. I think the new guidelines are likely a better way to assess. Those guidelines being 3 or more mediastinal stations and 1 or more hilar stations being assessed. However, looking at some of the reports from local institutions, including ours, it appears that adherence to the new quality measure may be even lower than it was to the old lymph node quality measure, and so it’s possible that these adherence numbers may become lower than what we see here. To my knowledge, there’s no plan to retrospectively go back in the NCDB and add nodal station information. However, I expect that we would have access to both nodal stations and lymph node counts going into the future, and we may be able to get an idea of how on target we were by comparing those future data. That being said, I’m looking forward to looking at the nodal stations and seeing what that shows us.
Dr Park. The overall rate and the most contemporary rate of segmentectomy were only 4.4% and 6.1%, respectively. What in your opinion should be the goal for limited resection as a proportion of overall curative resection, anatomic limited resection?
Dr Logan. This is a great question.
Dr Park. It’s a lower number.
Dr Logan. The proportion could be much higher, I think, and the results of the CALGB [140503] Alliance trial are going to inform the answer to this question. But just looking at segmentectomy applied appropriately to the potential population, that number could be 30% or even 50% looking at tumor size and peripheral location. But it’s hard to say. As a future push, segmentectomies should become more common, especially as stage shifting toward early-stage disease increases with lung cancer screening. With increases in how many patients with early-stage cancer are out there, there is going to be competition on how to best treat those patients, and segmentectomy offers more potential for accurate staging.
Dr Park. There’s ample evidence that in sublobar resections, there are other quality measures such as distance of the margin relative to the size of the primary, presence of spread through airspaces, and macropapillary subtype. You alluded to that toward the end of your talk. Do you think there’s a simple way of incorporating these other factors to assess quality of segmentectomy, or is that not going to be possible through a large administrative database?
Dr Logan. It may be possible in the future. We were able to look at micropapillary histology, and there were no observations with that histology in this dataset. In the future, as long as this variable is created and that it’s easily abstracted from a synoptic pathology report, you could add that, as long as it was put in there very simply. The caveat there is that if you start to make these quality measures too detailed, then lower resource areas may not be able to fill in a lot of those small but important things that we want to know, and so we may develop blind spots because of that.
Dr Park. Okay. Great job.
Unidentified Speaker #1. One more quick question and a quick answer. Dr Eagen?
Dr Thomas Egan (Chapel Hill, NC). I want to point out that the lymphatic drainage of the lung goes out to the visceral pleura and then around to the hilum, and for a segmentectomy to be adequate, you’ve got to be able to be sure there isn’t disease in the visceral pleura because the lymphatic drainage doesn’t necessarily follow the segments. And one of these days, we’ll be able to look at segments and determine if there are tumor cells in the visceral pleura and then go back to that patient and say, “You need a lobe,” and I think that’s ultimately going to be the downstream message. And your comments about segmentectomies between institutions and between surgeons may not be the same operation.
Dr Logan. That’s wonderful. Thank you.
Footnotes
Conflict of Interest Statement
The authors reported no conflicts of interest.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.
The data used in the study are derived from a deidentified NCDB file. The American College of Surgeons and the CoC have not verified and are not responsible for the analytic or statistical methodology used or the conclusions drawn from these data by the investigator.
This study uses only deidentified, publicly available data and was determined to be exempt from review by our institution’s Institutional Review Board.
References
- 1.American Cancer Society. Cancer statistics center; 2022. Accessed February 7, 2022. http://cancerstatisticscenter.cancer.org [Google Scholar]
- 2.Ginsberg RJ, Rubinstein LV. Randomized trial of lobectomy versus limited resection for T1 N0 non-small cell lung cancer. Lung Cancer Study Group. Ann Thorac Surg. 1995;60:615–22; discussion 622–3. [DOI] [PubMed] [Google Scholar]
- 3.Eguchi T, Sato T, Shimizu K. Technical advances in segmentectomy for lung cancer: a minimally invasive strategy for deep, small, and impalpable tumors. Cancers (Basel). 2021;13:3137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Shimizu K, Nagashima T, Ohtaki Y, Obayashi K, Nakazawa S, Kamiyoshihara M, et al. Analysis of the variation pattern in right upper pulmonary veins and establishment of simplified vein models for anatomical segmentectomy. Gen Thorac Cardiovasc Surg. 2016;64:604–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Nagashima T, Shimizu K, Ohtaki Y, Obayashi K, Nakazawa S, Mogi A, et al. Analysis of variation in bronchovascular pattern of the right middle and lower lobes of the lung using three-dimensional CT angiography and bronchography. Gen Thorac Cardiovasc Surg. 2017;65:343–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Saji H, Okada M, Tsuboi M, Nakajima R, Suzuki K, Aokage K, et al. Segmentectomy versus lobectomy in small-sized peripheral non-small-cell lung cancer (JCOG0802/WJOG4607L): a multicentre, open-label, phase 3, randomised, controlled, non-inferiority trial. Lancet. 2022;399:1607–17. [DOI] [PubMed] [Google Scholar]
- 7.Macke RA, Schuchert MJ, Odell DD, Wilson DO, Luketich JD, Landreneau RJ. Parenchymal preserving anatomic resections result in less pulmonary function loss in patients with Stage I non-small cell lung cancer. J Cardiothorac Surg. 2015;10:49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.El-Sherif A, Gooding WE, Santos R, Pettiford B, Ferson PF, Fernando HC, et al. Outcomes of sublobar resection versus lobectomy for stage I non-small cell lung cancer: a 13-year analysis. Ann Thorac Surg. 2006;82:408–15; discussion 415–6. [DOI] [PubMed] [Google Scholar]
- 9.Tsutani Y, Miyata Y, Nakayama H, Okumura S, Adachi S, Yoshimura M, et al. Oncologic outcomes of segmentectomy compared with lobectomy for clinical stage IA lung adenocarcinoma: propensity score-matched analysis in a multicenter study. J Thorac Cardiovasc Surg. 2013;146:358–64. [DOI] [PubMed] [Google Scholar]
- 10.Cao J, Yuan P, Wang Y, Xu J, Yuan X, Wang Z, et al. Survival rates after lobectomy, segmentectomy, and wedge resection for non-small cell lung cancer. Ann Thorac Surg. 2018;105:1483–91. [DOI] [PubMed] [Google Scholar]
- 11.Yao F, Wang J, Yao J, Xu L, Qian J, Cao Y. Early experience with video-assisted thoracoscopic anatomic segmentectomy. J Laparoendosc Adv Surg Tech A. 2018; 28:819–26. [DOI] [PubMed] [Google Scholar]
- 12.Subramanian M, McMurry T, Meyers BF, Puri V, Kozower BD. Long-term results for clinical stage IA lung cancer: comparing lobectomy and sublobar resection. Ann Thorac Surg. 2018;106:375–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lim TY, Park S, Kang CH. A meta-analysis comparing lobectomy versus segmentectomy in stage I non-small cell lung cancer. Korean J Thorac Cardiovasc Surg. 2019;52:195–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kim DH, Na KJ, Park IK, Kang CH, Kim YT, Park S. Long-term outcomes in stage I lung cancer after segmentectomy with a close resection margin. J Chest Surg. 2021;54:361–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Altorki NK, Wang X, Wigle D, Gu L, Darling G, Ashrafi AS, et al. Perioperative mortality and morbidity after sublobar versus lobar resection for early-stage non-small-cell lung cancer: post-hoc analysis of an international, randomised, phase 3 trial (CALGB/Alliance 140503). Lancet Respir Med. 2018;6:915–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Suzuki K, Saji H, Aokage K, Watanabe SI, Okada M, Mizusawa J, et al. Comparison of pulmonary segmentectomy and lobectomy: safety results of a randomized trial. J Thorac Cardiovasc Surg. 2019;158:895–907. [DOI] [PubMed] [Google Scholar]
- 17.Ghaly G, Kamel M, Nasar A, Paul S, Lee PC, Port JL, et al. Video-assisted thoracoscopic surgery is a safe and effective alternative to thoracotomy for anatomical segmentectomy in patients with clinical stage I non-small cell lung cancer. Ann Thorac Surg. 2016;101:465–72; discussion 472. [DOI] [PubMed] [Google Scholar]
- 18.Kneuertz PJ, Zhao J, D’Souza DM, Abdel-Rasoul M, Merritt RE. National trends and outcomes of segmentectomy in the Society of Thoracic Surgery Database. Ann Thorac Surg. 2021;113:1361–9. [DOI] [PubMed] [Google Scholar]
- 19.Cao J, Xu J, He Z, Yuan P, Huang S, Lv W, et al. Prognostic impact of lymphadenectomy on outcomes of sublobar resection for stage IA non-small cell lung cancer ≤2 cm. J Thorac Cardiovasc Surg. 2018;156:796–805.e4. [DOI] [PubMed] [Google Scholar]
- 20.Stiles BM, Mao J, Harrison S, Lee B, Port JL, Sedrakyan A, et al. Extent of lymphadenectomy is associated with oncological efficacy of sublobar resection for lung cancer ≤2 cm. J Thorac Cardiovasc Surg. 2019;157:2454–65.e1. [DOI] [PubMed] [Google Scholar]
- 21.He Z, Li Z, Xu S, Wu W, Zhu Q, Wang J, et al. Prognostic significance of lymph node count removed at sublobar resection in pathologic stage IA non-small-cell lung cancer: a population-based analysis. Clin Lung Cancer. 2021;22:e563–73. [DOI] [PubMed] [Google Scholar]
- 22.Qu X, Wang K, Zhang T, Shen H, Dong W, Liu Q, et al. Long-term outcomes of stage I NSCLC (≤3 cm) patients following segmentectomy are equivalent to lobectomy under analogous extent of lymph node removal: a PSM based analysis. J Thorac Dis. 2017;9:4561–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Handa Y, Tsutani Y, Mimae T, Miyata Y, Ito H, Nakayama H, et al. Appropriate extent of lymphadenectomy in segmentectomy: a multicenter study. J Clin Oncol. 2021;51:451–8. [DOI] [PubMed] [Google Scholar]
- 24.Zhou H, Tapias LF, Gaissert HA, Muniappan A, Wright CD, Wain JC, et al. Lymph node assessment and impact on survival in video-assisted thoracoscopic lobectomy or segmentectomy. Ann Thorac Surg. 2015;100:910–6. [DOI] [PubMed] [Google Scholar]
- 25.Bilimoria KY, Stewart AK, Winchester DP, Ko CY. The National Cancer Data Base: a powerful initiative to improve cancer care in the United States. Ann Surg Oncol. 2008;15:683–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Merkow RP, Rademaker AW, Bilimoria KY. Practical guide to surgical data sets: National Cancer Database (NCDB). JAMA Surg. 2018;153:850–1. [DOI] [PubMed] [Google Scholar]
- 27.Odell DD, Feinglass J, Engelhardt K, et al. Evaluation of adherence to the Commission on Cancer lung cancer quality measures. J Thorac Cardiovasc Surg. 2019;157:1219–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Li F, Zhao Y, Yuan L, Wang S, Mao Y. Oncologic outcomes of segmentectomy vs lobectomy in pathologic stage IA (≤2 cm) invasive lung adenocarcinoma: a population-based study. J Surg Oncol. 2020;121:1132–9. [DOI] [PubMed] [Google Scholar]
- 29.Stiles BM, Mao J, Harrison S, Lee B, Port JL, Altorki NK, et al. Sublobar resection for node-negative lung cancer 2–5 cm in size. Eur J Cardiothorac Surg. 2019;56:858–66. [DOI] [PubMed] [Google Scholar]
- 30.Deng B, Cassivi SD, de Andrade M, Nichols FC, Trastek VF, Wang Y, et al. Clinical outcomes and changes in lung function after segmentectomy versus lobectomy for lung cancer cases. J Thorac Cardiovasc Surg. 2014;148: 1186–92.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kneuertz PJ, Abdel-Rasoul M, D’Souza DM, Zhao J, Merritt RE. Segmentectomy for clinical stage I non-small cell lung cancer: National benchmarks for nodal staging and outcomes by operative approach. Cancer. 2022;128:1483–92. [DOI] [PubMed] [Google Scholar]
- 32.Safety in numbers: hospital performance on Leapfrog’s surgical volume standard based on results of the 2019 Leapfrog Hospital Survey. Washington DC; 2020. Accessed February 8, 2022. https://psnet.ahrq.gov/issue/safety-numbers-hospital-performance-leapfrogs-surgical-volume-standard-based-results-2019 [Google Scholar]
- 33.Hamilton BH, Ko CY, Richards K, Hall BL. Missing data in the American College of Surgeons National Surgical Quality Improvement Program are not missing at random: implications and potential impact on quality assessments. J Am Coll Surg. 2010;210:125–39.e2. [DOI] [PubMed] [Google Scholar]
- 34.Hoskin TL, Boughey JC, Day CN, Habermann EB. Lessons learned regarding missing clinical stage in the National Cancer Database. Ann Surg Oncol. 2019; 26:739–45. [DOI] [PubMed] [Google Scholar]
- 35.Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res. 2011;46: 399–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.