Abstract
Background
National organizations have recommended quality measures for surgery in early-stage non-small cell lung cancer (NSCLC). The outcomes of adherence to these guidelines are unknown.
Methods
Clinical stage I NSCLC surgery patients were abstracted from the National Cancer Data Base (NCDB). After reviewing current guidelines, the following quality measures were selected: anatomic resection, surgery within 8 weeks of diagnosis, R0 resection, and sampling ≥10 lymph nodes. Multivariate models identified variables independently associated with receiving quality measures and a Cox model created to evaluate overall survival.
Results
Between 2004 and 2013, 133,026/133,366 (99.7%), 126,598/133,366 (94.9%), 91,472/133,366 (68.6%), and 30,041/133,366 (22.5%) patients met 1, 2, 3, or 4 measures. Income ≥ $38,000/year (OR 1.20, 1.15–1.24), insurance type (private insurance OR 1.22, 1.09 – 1.36, Medicare OR 1.16, 1.04 – 1.30), centers with ≥38 cases/year (OR 1.18, 1.14–1.22), academic institutions (OR 1.31, 1.27–1.35), and clinical Stage IB patients (OR 1.50, 1.40 – 1.60) were more likely to meet all 4 measures; while increasing age (OR 0.99, 0.99–0.99), females (0.93, 0.91 – 0.96), non-Caucasian race (OR 0.83, 0.79 – 0.87) and increasing Charlson/Deyo comorbidity score (1: OR 0.90, 0.87–0.93, ≥2: OR 0.82, 0.79–0.86) were associated with lower likelihood. Pathologic upstaging (HR 1.84, 1.78–1.89) and meeting all 4 measures (HR 0.39, 0.31–0.48) were most powerfully associated with overall survival.
Conclusions
National adherence to quality measures in stage I NSCLC resection is suboptimal. Guideline compliance is strongly associated with survival and vigorous efforts should be instituted by national societies to improve adherence.
Clinical stage I non-small cell lung cancer (NSCLC) represents approximately 15% of lung cancer diagnoses. [1] Standard therapy for Stage I NSCLC is lobectomy, with 5-year survival rates exceeding 80%. [2] Surgery offers definitive local control and mediastinal lymph node sampling may reveal upstaging in approximately 25% of patients. [3–4] For patients receiving surgery, it is important to study practice patterns and associated outcomes to ensure that optimal care is being delivered.
Multiple lung cancer quality measures (QM) have been endorsed by organizations including the American College of Chest Physicians (ACCP), the National Comprehensive Cancer Network (NCCN), and the American College of Surgeons (ACS) Commission on Cancer (Table 1). [5–7] Other measures influencing patient outcomes in Stage I NSCLC have been reported in the literature. [8] While each of these recommendations have been developed from clinical trials, database studies, and professional consensus, to date there has not been a study of outcomes when these measures are consolidated.
Table 1.
Current guidelines for cancer care in clinical Stage I NSCLC.
| Recommending Organization | Clinical Care Measure in Clinical Stage I NSCLC |
|---|---|
| National Comprehensive Cancer Network (NCCN) |
|
| American College of Surgeons Commission on Cancer (ACS CoC) |
|
| Society of Thoracic Surgeons (STS) |
|
| American College of Chest Physicians (ACCP) |
|
Asteric* indicates measures that were identifiable in the National Cancer Data Base (NCDB) for this analysis.
We selected four QM that would be identifiable in the National Cancer Data Base (NCDB): anatomic resection (including segmentectomy for tumors ≤2cm), surgery <8 weeks from diagnosis, R0 resection, and sampling of ≥10 lymph nodes. We hypothesized that these metrics would be individually and collectively associated with improved overall survival (OS). [5–8] Secondary aims included identifying patient and tumor characteristics associated with meeting these QM.
Patients and Methods
The NCDB Participant User File (PUF) for NSCLC was reviewed to identify clinical Stage I patients receiving surgery. The NCDB is a collaboration between ACS and the American Cancer Society, and captures approximately 70% of all malignancies in the United States. Patients and facilities are deidentified, and this study was exempt from our IRB. Patient, tumor, and treatment variable definitions are detailed on the NCDB PUF data dictionary website (http://ncdbpuf.facs.org/node/259). The upper quartile of facility volume (≥38 clinical Stage I NSCLC surgical cases/year) defined a ‘high volume’ facility. Time from diagnosis (radiologic or histologic) to surgery was dichotomotized into < or ≥8 weeks. [8] Surgical approach was unavailable for 47.2% of patients, and was not analyzed.
Patients with unknown resection type (n=2,664), laser excision (n=156), bronchial sleeve resection (n=224), or extended resection were excluded (n=4,400). Resections were classified as wedge, segmentectomy, lobectomy, or pneumonectomy. Bilobectomy cases are included, but not specified, in the lobectomy category. Patients that had surgery >1 year from diagnosis (n=216), received preoperative chemotherapy (n=1406) or radiotherapy (n=253) were excluded. Patients were considered as needing adjuvant therapy if their pathologic stage was reported as ≥T3 or ≥N1.
Descriptive statistics of continuous variables were reported as mean ± standard deviation. Independent sample t tests were used to analyze normally distributed continuous data. χ2 tests were used to compare categorical data. Backwards stepwise multivariate logistic regression identified variables independently associated with each QM. Variables with a significant difference (p<0.05) on univariate analysis were entered into the regression model. Cramer’s V test was performed on categorical variables considered clinically related to check for association prior to model entry. A Cox proportional hazards model identified variables independently associated with mortality. P values < 0.05 were considered statistically significant. Log-rank p values were used to compare median survival. All statistical analyses were performed in SPSS for Windows (Version 23.0. Armonk, NY: IBM Corporation, 2015).
Results
From 2004 to 2013, 146,908/241,146 (60.9%) clinical Stage I NSCLC patients received surgical resection (Figure 1). 29,649/146,908 (20.2%) received a wedge resection, 6,212/146,908 (4.2%) segmentectomy, 107,687/146,908 (73.3%) lobectomy, and 3,360/146,908 (2.3%) pneumonectomy (Table 2). 3,675/6,212 (59.2%) segmentectomies had a tumor size ≤2cm. Variables independently associated with anatomic resection included non-Caucasian race (Odds Ratio 1.12, 95% Confidence Interval 1.07 – 1.17) and clinical IB stage (OR 2.38, 2.24 – 2.54), both p<0.001. Variables associated with a decreased likelihood of anatomic resection included increasing age (per year, OR 0.97, 0.97 –0.98), female gender (OR 0.91, 0.89 – 0.94), increasing Charlson/Deyo comorbidity score (CDS 1 OR 0.83, 0.80 – 0.85, CDS ≥2 OR 0.70, 0.67 – 0.72), high volume facility (OR 0.94, 0.91 – 0.97), and academic center (OR 0.88, 0.86 – 0.91), all p<0.001.
Figure 1.
CONSORT diagram for clinical Stage I NSCLC patients receiving surgical resection in the National Cancer Data Base (NCDB) from 2004–2013.
Table 2.
Univariate analysis of patients receiving ‘quality’ surgery.
| Variable | Non-anatomical Resection (n=32,186) |
Anatomical Resection (n=114,722) |
P value |
|---|---|---|---|
| Age | 70.0 ± 9.5 | 67.5 ± 10.0 | <0.001 |
| Male | 14173 (44.0%) | 53280 (46.4%) | <0.001 |
| Caucasian | 29014 (90.9%) | 101683 (89.3%) | <0.001 |
| Private Insurance | 7638 (24.0%) | 35791 (31.6%) | <0.001 |
| Income ≥$38,000 | 26421 (83.3%) | 93184 (82.4%) | <0.001 |
| Metro county | 25891 (83.2%) | 92166 (83.0%) | 0.50 |
| <21% in community no high school degree |
26987 (85.1%) | 95398 (84.3%) | <0.001 |
| Academic Center | 12361 (38.6%) | 40764 (35.8%) | <0.001 |
| ≥38 cases per year | 8667 (26.9%) | 28213 (24.6%) | <0.001 |
| Charlson/Deyo Score ≥2 | 5473 (17.0%) | 15017 (13.1%) | <0.001 |
| Clinical Stage IB | 5255 (16.3%) | 36291 (31.6%) | <0.001 |
| Received surgery ≥8 weeks | 6016 (18.9%) | 22660 (20.0%) | <0.001 |
| ≥10 Lymph Nodes Sampled | 2282 (7.5%) | 39804 (37.5%) | <0.001 |
| ≥ R1 Resection | 1791 (5.7%) | 3116 (2.7%) | <0.001 |
| Tumor size (mm) | 19.7 ± 11.0 | 27.3 ± 15.8 | <0.001 |
| Length of inpatient stay (days) | 5.5 ± 6.2 | 7.1 ± 6.8 | <0.001 |
| 30-day readmission rate | 1711 (5.4%) | 7097 (6.3%) | <0.001 |
| 30-day mortality | 491 (1.8%) | 2158 (2.2%) | <0.001 |
| 90-day mortality | 975 (3.6%) | 3866 (4.0%) | 0.003 |
| Upstaged ≥T3 | 1106 (3.7%) | 4586 (4.2%) | 0.001 |
| Upstaged ≥N1 | 983 (6.0%) | 13451 (12.6%) | <0.001 |
| Proportion of upstaged patients receiving adjuvant chemotherapy |
791 (44.9%) | 8189 (55.3%) | <0.001 |
Compared to wedge resections, segmentectomy patients were more likely to have ≥10 lymph nodes sampled (15.9% versus 6.7%, p<0.001), surgery at a high volume center (5.9% versus 3.7%, p<0.001), positive pathologic lymph nodes (6.6% versus 5.5%, p=0.004) and were less likely to have positive margins (3.8% versus 5.7%, p<0.001). Alternatively, 3,360 patients received pneumonectomy. Compared to lobectomies, pneumonectomy patients had a higher ≥R1 rate (7.1% versus 2.6%, p<0.001) and had larger pathologic tumor sizes (43.4 ± 23.5mm versus 27.2 ± 15.4mm, p<0.001), suggesting tumor anatomy was involved this surgical approach.
The second QM was surgical resection <8 weeks from lung cancer diagnosis. By this definition, 28,676/145,090 (19.8%) of Stage I NSCLC patients received delayed surgery (Table 3). Variables associated with an increased likelihood of delayed surgery included increasing age (per year, OR 1.01, 1.0 – 1.01), non-Caucasian race (OR 1.45, CI 1.39 – 1.52), academic center (OR 1.27, 1.24 – 1.31), increasing Charlson/Deyo score (CDS 1: OR 1.13, 1.10 – 1.17, CDS 2: OR 1.30, 1.25 – 1.36), and biopsy prior to surgery (OR 2.53, 2.46 – 2.60), all p<0.001. Variables independently associated with a decreased likelihood of delayed surgery included income ≥$38,000 (OR 0.87, 0.83 – 0.90), rural county (OR 0.84, 0.76 – 0.93), and higher education zip code (OR 0.91, 0.88 – 0.95), all p≤0.001.
Table 3.
Univariate analysis of patients receiving early versus delayed pulmonary resection.
| Variable | Surgery < 8 weeks (n=116,414) |
Surgery ≥ 8 weeks (n=28,676) |
P value |
|---|---|---|---|
| Age | 67.8 ± 10.0 | 68.8 ± 9.7 | <0.001 |
| Male | 53182 (45.7%) | 13422 (46.8%) | 0.001 |
| Caucasian | 104596 (90.5%) | 24535 (86.2%) | <0.001 |
| Private insurance | 35896 (31.2%) | 7015 (24.7%) | <0.001 |
| Income ≥$38,000 | 95523 (83.2%) | 22473 (79.5%) | <0.001 |
| Metro county | 93466 (83.0%) | 23107 (83.3%) | 0.03 |
| <21% in community no high school degree |
97710 (85.1%) | 23249 (82.2%) | <0.001 |
| Academic Center | 41040 (35.6%) | 11328 (39.7%) | <0.001 |
| ≥38 cases/year | 29209 (25.1%) | 7160 (25.0%) | 0.67 |
| Charlson/Deyo Score ≥2 | 15653 (13.4%) | 4649 (16.2%) | <0.001 |
| Clinical Stage IB | 32208 (27.7%) | 8766 (30.6%) | <0.001 |
| Biopsy prior to surgery | 46260 (40.7%) | 17612 (63.1%) | <0.001 |
| Wedge Resection | 23955 (20.6%) | 5412 (18.9%) | <0.001 |
| Segmentectomy | 4836 (4.2%) | 1289 (4.5%) | |
| Lobectomy | 85049 (73.1%) | 21242 (74.1%) | |
| Pneumonectomy | 2574 (2.2%) | 733 (2.6%) | |
| ≥10 lymph nodes removed | 32791 (30.3%) | 8788 (33.0%) | <0.001 |
| ≥R1 resection | 3759 (3.3%) | 1091 (3.9%) | <0.001 |
| Tumor size (mm) | 25.2 ± 15.2 | 27.3 ± 15.4 | <0.001 |
| Length of inpatient stay (days) | 6.6 ± 6.6 | 7.2 ± 7.1 | <0.001 |
| 30-day readmission rate | 7025 (6.1%) | 1690 (6.0%) | 0.46 |
| 30-day mortality | 2000 (2.0%) | 619 (2.6%) | <0.001 |
| 90-day mortality | 3633 (3.7%) | 1148 (4.8%) | <0.001 |
| Upstaged ≥ T3 | 4263 (3.9%) | 1366 (5.1%) | <0.001 |
| Upstaged ≥N1 | 10972 (11.3%) | 3271 (13.5%) | <0.001 |
| Proportion of upstaged patients receiving adjuvant chemotherapy |
7077 (56.1%) | 1830 (48.3%) | <0.001 |
The third QM selected was R0 resection, with 140,014/144,921 (96.6%) reporting negative margins (Table 4). Variables associated with an increased likelihood of ≥R1 margins included a Charlson/Deyo score ≥2 (OR 1.14, 1.04 – 1.24, p=0.003), clinical Stage IB (OR 1.42, 1.22 – 1.66, p<0.001), delayed surgery (OR 1.14, 1.06 – 1.22, p<0.001), and increasing tumor size (per millimeter OR 1.02, 1.01 – 1.02, p<0.001). Variables associated with a decreased likelihood of ≥R1 margins included academic centers (OR 0.85, 0.79 – 0.91), high-volume centers (OR 0.89, 0.82 – 0.96), clinical Stage IA (OR 0.78, 0.67 – 0.90), and resection type (ref: wedge; segmentectomy OR 0.61, 0.52 – 0.70 and lobectomy OR 0.34, 0.32 – 0.36), all p≤0.001.
Table 4.
Univariate analysis of patients receiving R0 versus ≥R1 resection.
| Variable | R0 resection (n=140,014) |
≥R1 resection (n=4,907) |
P value |
|---|---|---|---|
| Age | 68.0 ± 9.9 | 69.0 ± 9.8 | <0.001 |
| Male | 64049 (45.7%) | 2437 (49.7%) | <0.001 |
| Caucasian | 124587 (89.7%) | 4342 (88.9%) | 0.10 |
| Private Insurance | 41578 (30.0%) | 1286 (26.5%) | <0.001 |
| Income ≥$38,000 | 114053 (82.6%) | 3937 (81.5%) | 0.04 |
| Metro county | 112624 (83.1%) | 3873 (81.6%) | 0.006 |
| <21% in community no high school degree |
116708 (84.5%) | 4064 (84.1%) | 0.43 |
| Academic Center | 51050 (36.7%) | 1541 (31.6%) | <0.001 |
| ≥38 cases/year | 35558 (25.4%) | 1044 (21.3%) | <0.001 |
| Charlson/Deyo Score ≥2 | 19441 (13.9%) | 804 (16.4%) | <0.001 |
| Clinical Stage IB | 38723 (27.7%) | 2171 (44.2%) | <0.001 |
| Wedge Resection | 27260 (19.5%) | 1648 (33.6%) | <0.001 |
| Segmentectomy | 5901 (4.2%) | 231 (4.7%) | |
| Lobectomy | 103814 (74.1%) | 2796 (57.0%) | |
| Pneumonectomy | 3039 (2.2%) | 232 (4.7%) | |
| Received surgery <8 weeks | 111222 (80.4%) | 3759 (77.5%) | <0.001 |
| ≥10 lymph nodes removed | 40693 (31.2%) | 1150 (25.1%) | <0.001 |
| Tumor size (mm) | 25.4 ± 15.0 | 31.3 ± 19.4 | <0.001 |
| Length of inpatient stay (days) | 6.7 ± 6.7 | 7.2 ± 7.6 | <0.001 |
| 30-day readmission rate | 8362 (6.1%) | 339 (7.1%) | 0.004 |
| 30-day mortality | 2433 (2.0%) | 163 (3.8%) | <0.001 |
| 90-day mortality | 4434 (3.8%) | 312 (7.4%) | <0.001 |
| Upstaged to ≥ T3 | 4735 (3.6%) | 843 (18.4%) | <0.001 |
| Upstaged to ≥ N1 | 13180 (11.1%) | 1064 (29.4%) | <0.001 |
| Proportion of upstaged patients receiving adjuvant chemotherapy |
8076 (54.1%) | 771 (54.4%) | 0.84 |
The fourth QM analyzed was ≥10 lymph nodes sampled. Only 42,086/136,612 (30.8%) patients obtained this QM (Table 5). The proportion of patients having ≥10 lymph nodes sampled increased from 2004 (26.5%) to 2013 (34.2%), p<0.001. Variables associated with obtaining ≥10 lymph nodes included income ≥$38,000 (OR 1.16, 1.12 – 1.20), academic center (OR 1.61, 1.57 – 1.66) or a high-volume center (OR 1.27, 1.24 – 1.32), clinical Stage IB (OR 1.22, 1.14 – 1.30), delayed resection (OR 1.09, 1.05 – 1.12), and resection type (ref: wedge; segmentectomy OR 2.49, 2.28 – 2.72, lobectomy OR 8.56, 8.14 – 9.01, pneumonectomy OR 22.4, 20.4 – 24.5), all p<0.001.
Table 5.
Univariate analysis of patients receiving < versus ≥10 lymph nodes sampled.
| Variable | <10 Lymph Nodes Obtained (n=94,526) |
≥10 Lymph Nodes Obtained (n=42,086) |
P value |
|---|---|---|---|
| Age | 68.2 ± 10.0 | 67.6 ± 9.9 | <0.001 |
| Male | 42669 (45.1%) | 19909 (47.3%) | <0.001 |
| Caucasian | 83977 (89.5%) | 37436 (89.6%) | 0.44 |
| Private Insurance | 27340 (29.3%) | 12921 (31.1%) | <0.001 |
| Income ≥$38,000 | 76459 (82.1%) | 34709 (83.6%) | <0.001 |
| Metro county | 75652 (82.8%) | 34256 (84.0%) | <0.001 |
| <21% in community no high school degree |
78586 (84.3%) | 35158 (84.7%) | 0.09 |
| Academic Center | 31058 (33.1%) | 18434 (44.1%) | <0.001 |
| ≥38 cases/year | 21179 (22.4%) | 12212 (29.0%) | <0.001 |
| Charlson/Deyo Score ≥2 | 13696 (14.5%) | 5471 (13.0%) | <0.001 |
| Clinical Stage IB | 23761 (25.1%) | 14612 (34.7%) | <0.001 |
| Wedge Resection | 26275 (27.8%) | 1886 (4.5%) | |
| Segmentectomy | 4837 (5.1%) | 913 (2.2%) | <0.001 |
| Lobectomy | 62236 (65.8%) | 37386 (88.8%) | |
| Pneumonectomy | 1178 (1.2%) | 1901 (4.5%) | |
| Received surgery <8 weeks | 75513 (80.9%) | 32791 (78.9%) | <0.001 |
| ≥R1 Resection | 3438 (3.7%) | 1150 (2.7%) | <0.001 |
| Tumor size (mm) | 24.0 ± 14.2 | 28.8 ± 16.7 | <0.001 |
| Length of inpatient stay (days) | 6.6 ± 6.6 | 6.9 ± 6.9 | <0.001 |
| 30-day readmission rate | 5550 (6.0%) | 2597 (6.3%) | 0.03 |
| 30-day mortality | 1674 (2.1%) | 790 (2.3%) | 0.06 |
| 90-day mortality | 3090 (3.9%) | 1396 (4.0%) | 0.23 |
| Upstaged to ≥ T3 | 3426 (3.7%) | 1986 (4.9%) | <0.001 |
| Upstaged to ≥ N1 | 6551 (9.0%) | 6804 (16.7%) | <0.001 |
| Proportion of upstaged patients receiving adjuvant chemotherapy |
4150 (50.9%) | 4144 (57.3%) | <0.001 |
Of the 133,366/146,908 (90.8%) clinical Stage I NSCLC patients with complete data on all QM, 340/133,366 (0.2%) received zero QM, 133,026/133,366 (99.7%) received one QM, 126,598/133,366 (94.9%) received two QM, 91,472/133,366 (68.6%) received 3 QM, and 30,041/133,366 (22.5%) received 4 QM (Appendix Table 1). The proportion obtaining all four QM increased from 2004 (18.9%) to 2013 (24.7%), p<0.001, Figure 2. Variables associated with meeting all QM are listed in Table 6. Despite patients receiving all four QM having a higher rate of upstaging than patients meeting <4 QM, they experienced improved median OS (89.6 ± 1.33 months versus 72.7 ± 0.42 months, p<0.001). Improved median OS was seen for each individual QM obtained: 0 QM 31.4 ± 2.78 months, 1 QM 48.0 ± 0.92, 2 QM 63.0 ± 0.6, 3 QM 82.7 ± 0.72, and 4 QM 89.6 ± 1.33, p<0.001), Figure 3. Factors associated with overall mortality in Cox proportional hazards modeling are shown in Table 7. While pathologic upstaging was associated with the highest increase for mortality, an increasing number of QM met was associated with the largest decrease in mortality.
Figure 2.
Proportion of clinical stage I NSCLC patients meeting all four QM over time (p<0.001, chi-square analysis).
Table 6.
Variables independently associated with likelihood of meeting all 4 quality measures in clinical Stage I NSCLC.
| Variable | Odds Ratio (95% CI) | P value |
|---|---|---|
| Age (per year increase) | 0.99 (0.99 – 0.99) | <0.001 |
| Female | 0.93 (0.91 – 0.96) | <0.001 |
| Non-Caucasian | 0.83 (0.79 – 0.87) | <0.001 |
| Income ≥$38,000 | 1.20 (1.15 – 1.24) | <0.001 |
| Insurance (ref: uninsured) | ||
| Private | 1.22 (1.09 – 1.36) | <0.001 |
| Medicare | 1.16 (1.04 – 1.30) | 0.008 |
| Charlson/Deyo Score (ref: 0) | ||
| 1 | 0.90 (0.87 – 0.93) | <0.001 |
| ≥2 | 0.82 (0.79 – 0.86) | <0.001 |
| ≥38 cases/year | 1.18 (1.14 – 1.22) | <0.001 |
| Academic Cancer Center | 1.31 (1.27 – 1.35) | <0.001 |
| Clinical Stage IB | 1.50 (1.40 – 1.60) | <0.001 |
Figure 3.
Kaplan-Meier analysis for clinical stage I NSCLC patients, by number of QM met.
Table 7.
Cox Proportional Hazards Model for overall survival in clinical Stage I NSCLC
| Variable | Hazard Ratio (95% Confidence Interval) |
P value |
|---|---|---|
| Age (per year increase) | 1.03 (1.02 –1.03) | <0.001 |
| Female | 0.71 (0.70 – 0.73) | <0.001 |
| Non-Caucasian | 0.91 (0.87 – 0.95) | <0.001 |
| Income ≥$38,000 | 0.89 (0.86 – 0.92) | <0.001 |
| Population (ref: metro) Urban |
1.11 (1.08 – 1.15) | <0.001 |
| Insurance type (ref: uninsured) | ||
| Private | 0.83 (0.75 – 0.92) | <0.001 |
| Medicaid | 1.14 (1.02 – 1.29) | 0.03 |
| Charlson/Deyo Score (ref: 0) | ||
| 1 | 1.19 (1.16 – 1.22) | <0.001 |
| ≥2 | 1.47 (1.42 – 1.52) | <0.001 |
| ≥38 cases/year | 0.96 (0.94 – 0.99) | 0.007 |
| Academic Cancer Center | 0.95 (0.92 – 0.97) | <0.001 |
| Tumor size (per mm increase) | 1.01 (1.0 – 1.01) | <0.001 |
| Inpatient Length of Stay (per day) | 1.03 (1.02 – 1.03) | <0.001 |
| 30-Day Readmission | 1.25 (1.20 – 1.30) | <0.001 |
| Upstaged to ≥T3 or ≥N1 | 1.84 (1.78 – 1.89) | <0.001 |
| Number of QM Met (ref: 0) | ||
| 1 | 0.70 (0.56 – 0.88) | 0.003 |
| 2 | 0.56 (0.45 – 0.70) | <0.001 |
| 3 | 0.44 (0.36 – 0.55) | <0.001 |
| 4 | 0.39 (0.31 – 0.48) | <0.001 |
Comment
We selected four QM from the literature and organizations including the NCCN, ACS Commission on Cancer, and ACCP to evaluate their current use and relationship to survival in clinical Stage I NSCLC. These measures included anatomic resection (including segmentectomy if tumor size ≤2cm), receiving surgery <8 weeks from diagnosis, R0 resection, and sampling ≥10 lymph nodes. Meeting these QM was associated with socioeconomic factors, institution type, and clinical substage. It also found that patients receiving all four QM have a higher rate of pathologic upstaging, which may be missed by non-anatomic resections and/or lower lymph node counts. Finally, each additional QM met was associated with improved OS.
A review of clinical stage IA patients from the NCDB (2003 – 2011) found that lobectomy was associated with improved 5-year OS compared to sublobar resection. [9] An earlier population-based analysis in the SEER database from 1998 – 2007, found similar overall and cancer-specific survival benefits for lobectomy in Stage I NSCLC, even by tumor size subgroups. [10] A propensity-matched Stage IA analysis of resection type found that wedge resection or segmentectomy patients continued to have an increased mortality hazard compared to lobectomy patients. [11] Similar to our analysis, the propensity matched analysis found an increased positive margin rate, and decreased lymph node sampling and subsequent upstaging rates. [11] However, our dichotomization of quality surgery was based on studies that have suggested no survival difference for segmentectomies with tumors ≤2cm. [12–14] Recommendations by the ACCP state that for medically compromised patients “sublobar resection should involve an anatomical segmentectomy whenever possible”. [5] Certainly, an important driver for survival in sublobar resection includes adequate lymph node sampling, most likely due to improved staging accuracy. [9,14]
Previous work has shown that an increasing time interval from diagnosis to surgery is associated with increased mortality hazard. [8, 15] Work by our group found that propensity matched patients receiving surgery past eight weeks from diagnosis were more likely to be upstaged and have decreased median OS. [8] In our current analysis, drivers of delayed resection included socioeconomic, comorbidity, and institutional factors. These findings suggest multiple intervention points for improving access and coordination of care after diagnosis. With expanding lung cancer screening efforts, efficient evaluation of early stage lung cancer patients will be crucial in optimizing outcomes.
ACCP guidelines currently recommend that tumors <2cm have a margin greater than its maximal diameter, and that tumors ≥2cm should have ≥ 2cm margins. [5] We were unable to study tumor margin distance, therefore R0 status was selected as a surrogate measure. A previous study examining microscopic residual disease (R1) in Stage I NSCLC documented a decrease in 5-year OS (37% versus 62% for R0 patients, p<0.001). [16] Despite our finding that sublobar resections were performed at a slightly higher rate at high-volume and academic centers, patients at these facilities were less likely to have positive margins. While negative margin status was the QM most frequently met, it is strongly influenced by resection type. This remains an area for intervention, especially when only one-third of ≥R1 patients were receiving postoperative radiotherapy.
Our final QM, sampling of ≥10 lymph nodes, has been recommended by both the NCCN and the ACS. [6,7] However, these recommendations do not specify lymph node location. Alternatively, the ACCP recommends Stage I patients receive systematic mediastinal lymph node sampling or dissection. [5] Since 2005, decreased mortality for patients receiving more extensive lymph node sampling has been noted, likely due to more accurate staging. [17–19] This was the QM least frequently met, and is modifiable. Surgeons may feel that the final lymph node number may be partially out of their control, however interventions such as using a pre-labeled collection kit have increased lymph node yield. [20–21]
This study found that for every QM met, there was an association with improved OS. While this may be related to unmeasured patient comorbidities, it still brings into focus modifiable factors that we know improve survival: choosing a segmentectomy with lymph node sampling when possible, minimizing wait times to surgery, selecting a surgical approach to optimize margins, and sampling ≥10 lymph nodes. Of note, both academic centers and high volume centers were independently seen to have improved odds of receiving all four quality measures and independently associated with a decreased mortality hazard. Referral of clinical Stage I NSCLC to these center types may also improve rates of national adherence. An interplay of structural measures (center type, surgeon specialty, center resources) and process measures (defined as health care activities performed on patients, like those measures examined in this analysis) are likely influencing crucial short- and long-term outcome measures (30-day readmission, 30- and 90-day mortality, and overall survival) and need to be considered in future quality studies for lung cancer patients.
There are limitations to this analysis, some of which have been mentioned. Additionally, we cannot analyze subgroup populations for whom a wedge resection was the only likely surgical option, or was a diagnostic procedure, versus patients that could have tolerated an anatomical resection. There are also other QM that would have appropriate for analysis, but are frequently missing (i.e. surgical approach) or not available (pre-operative staging or mediastinal lymph node sampling rates). Other quality gaps have been documented in lung cancer care: a review of patients receiving lung cancer surgery from 2007–2013 in a national database found that approximately 40% were not receiving recommended pre-operative workup (PFTs, staging CT, or PET). [22] Surgical approach would be particularly interesting, as VATS series have shown decreased complication rates compared to open thoracotomy, but debates regarding lymph node yield remain. [4, 23–24] Additionally, recurrence data is not available, therefore the relationship that these QM have with cancer-free survival is unknown.
Improving national adherence to these quality measures will require evaluation by individual institutions and surgeons to identify gaps in recommended care. Even single-surgeon auditing has demonstrated improved rates of anatomical resections and number of mediastinal lymph nodes obtained. [25] It is likely that there is high variability of adherence to quality measures both between and within cancer centers, and identifying this variability will be crucial in developing tailored plans to facilitate adherence. With the anticipated rise in earlier stage lung cancers due to expanded lung cancer screening efforts, this is an exciting time to optimize quality care, and ultimately, survival.
Supplementary Material
Acknowledgments
Pamela Samson, MD, MPHS has grant support through NIH Cardiothoracic Surgery T32 HL07776. Varun Puri, MD, MSCI has grant funding through NIH K07CA178120 and K12CA167540-02. The NCDB is not responsible for the analytic methodology used in this study, and the conclusions drawn are solely those of the authors.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Presented at the Fifty-second Annual Meeting of The Society of Thoracic Surgeons, Phoenix, AZ, Jan 23–27, 2016. Winner of the J. Maxwell Chamberlain Memorial Award for General Thoracic Surgery.
The Appendix Table can be viewed in the online version of this article [insert doi number] on http://www.annalsthoracicsurgery.org
References
- 1.SEER Fact Sheets: Lung and Bronchus Cancer. [Accessed 1/2016];Surveillance, Epidemiology, and End Results Program. http://seer.cancer.gov/statfacts/html/lungb.html.
- 2.Whitson BA, Groth SS, Duval SJ, et al. Surgery for Early-Stage NSCLC: A Systematic Review of VATS Versus Thoracotomy Approaches to Lobectomy. Ann Thorac Surg. 2008;86:2008–2018. doi: 10.1016/j.athoracsur.2008.07.009. [DOI] [PubMed] [Google Scholar]
- 3.Licht PB, Jorgensen OD, Ladegaard L, et al. A National Study of Nodal Upstaging after Thoracoscopic Versus Open Lobectomy for Clinical Stage I Lung Cancer. Ann Thorac Surg. 2013;96:943–950. doi: 10.1016/j.athoracsur.2013.04.011. [DOI] [PubMed] [Google Scholar]
- 4.Merritt RE, Hoang CD, Shrager JB. Lymph node evaluation achieved by open lobectomy compared with thoracoscopic lobectomy for N0 lung cancer. Ann Thorac Surg. 2013;96(4):1171–1177. doi: 10.1016/j.athoracsur.2013.05.044. [DOI] [PubMed] [Google Scholar]
- 5.Howington JA, Blum MG, Chang AC, et al. Treatment of Stage I and II NSCLC. Chest. 2013;143((5)Suppl):e278S–e313S. doi: 10.1378/chest.12-2359. [DOI] [PubMed] [Google Scholar]
- 6.National Comprehensive Cancer Network. [Acessed 11/24/15];NCCN Clinical Practice Guidelines in Oncology: NSCLC. Version 2.2016. http://www.nccn.org/professionals/physician_gls/pdf/nscl.pdf. [Google Scholar]
- 7.American College of Surgeons Commission on Cancer. Cancer Programs Practice Profile Reports (CP3R): Lung Measure Specifications. [Accessed 11/24/15]; https://www.facs.org/~/media/files/quality%20programs/cancer/lungmeasuredocumentation_05272015.ashx. [Google Scholar]
- 8.Samson P, Patel A, Garrett T, et al. Effects of Delayed Surgical Resection on Short-Term and Long-Term Outcomes in Clinical Stage I NSCLC. Ann Thorac Surg. 2015;99(6):1906–1912. doi: 10.1016/j.athoracsur.2015.02.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Speicher PJ, Gu L, Gulack BC, et al. Sublobar Resection for Clinical Stage IA NSCLC in the United States. Clin Lung Cancer. 2015 doi: 10.1016/j.cllc.2015.07.005. S1525-7304(15)00185-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Whitson BA, Groth SS, Andrade RS, et al. Survival After Lobectomy Versus Segmentectomy for Stage I NSCLC: A Population-Based Analysis. Ann Thorac Surg. 2011;92:1943–1950. doi: 10.1016/j.athoracsur.2011.05.091. [DOI] [PubMed] [Google Scholar]
- 11.Khullar OV, Liu Y, Gillespie T, et al. Survival After Sublobar Resection versus Lobectomy for Clinical Stage IA Lung Cancer. J Thorac Oncol. 2015;10:1625–1633. doi: 10.1097/JTO.0000000000000664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Tsutani Y, Miyata Y, Nakayama H, et al. Oncologic outcomes of segmentectomy compared to lobectomy for clinical stage IA lung adenocarcinoma: Propensity score-matched analysis in a multicenter study. J Thorac Cardiovasc Surg. 2013;146:358–364. doi: 10.1016/j.jtcvs.2013.02.008. [DOI] [PubMed] [Google Scholar]
- 13.Altorki NK, Yip R, Hanaoka T, et al. Sublobar resection is equivalent to lobectomy for clinical stage IA lung cancer in solid nodules. J Thorac Cardiovasc Surg. 2014;147:754–764. doi: 10.1016/j.jtcvs.2013.09.065. [DOI] [PubMed] [Google Scholar]
- 14.Wolf AS, Richards WG, Jaklitsch MT, et al. Lobectomy versus sublobar resection for small (2cm or less) NSCLC. Ann Thorac Surg. 2011;92(5):1819–1823. doi: 10.1016/j.athoracsur.2011.06.099. [DOI] [PubMed] [Google Scholar]
- 15.Kanarek N, Hooker C, Matheiu L, et al. Survival after community diagnosis of early-stage NSCLC. Am J Med. 2014;127:443–449. doi: 10.1016/j.amjmed.2013.12.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hancock JG, Rosen JE, Antonicelli A, et al. Impact of adjuvant treatment for microscopic residual disease after NSCLC surgery. Ann Thorac Surg. 2015;99(2):406–413. doi: 10.1016/j.athoracsur.2014.09.033. [DOI] [PubMed] [Google Scholar]
- 17.Ludwig MA, Goodman M, Miller DL, et al. Postoperative survival and the number of lymph nodes sampled during resection of node-negative NSCLC. Chest. 2005;128(3):1545–1550. doi: 10.1378/chest.128.3.1545. [DOI] [PubMed] [Google Scholar]
- 18.Osarogiagbon RU, Ogbata O, Yu X. Number of lymph nodes associated with maximal reduction of long-term mortality risk in pathologic node-negative NSCLC. Ann Thorac Surg. 2014;97(2):385–393. doi: 10.1016/j.athoracsur.2013.09.058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ou SH, Zell JA. Prognostic Significance of the number for lymph nodes removed at lobectomy in Stage IA NSCLC. J Thorac Oncol. 2008;3(8):880–886. doi: 10.1097/JTO.0b013e31817dfced. [DOI] [PubMed] [Google Scholar]
- 20.Osarogiagbon RU, Miller LE, Ramirez RA, et al. Use of a surgical specimen-collection kit to improve mediastinal lymph node examination of resectable lung cancer. J Thorac Oncol. 2012;7(8):1276–1282. doi: 10.1097/JTO.0b013e318257fbe5. [DOI] [PubMed] [Google Scholar]
- 21.Osarogiagbon RU, Ramirez RA, Wang CG, et al. Dual intervention to improve pathologic staging of resectable lung cancer. Ann Thorac Surg. 2013;96(6):1975–1981. doi: 10.1016/j.athoracsur.2013.07.009. [DOI] [PubMed] [Google Scholar]
- 22.Flanagan MR, Varghese TK, Backhus LM, et al. Gaps in Guideline-Concordant Use of Diagnostic Tests Among Lung Cancer Patients. Ann Thorac Surg. 2015;100:2006–2012. doi: 10.1016/j.athoracsur.2015.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Paul S, Altorki NK, Sheng S, et al. Thoracoscopic lobectomy is associated with lower morbidity than open lobectomy: a propensity-matched analysis from the STS database. J Thorac Cardiovasc Surg. 2010;139(2):366–378. doi: 10.1016/j.jtcvs.2009.08.026. [DOI] [PubMed] [Google Scholar]
- 24.Martin JT, Durbin EB, Chen L, et al. Nodal Upstaging During Lung Cancer Resection is Associated with Surgical Approach. Ann Thorac Surg. 2016;101(1):238–245. doi: 10.1016/j.athoracsur.2015.05.136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hagan ME, Williams ST, Socci L, et al. Completing the Audit Cycle Improves Surgical Standards in Lung Cancer: Why Do Some Patients Still Not Receive the Best Care? J Thorac Oncol. 2013;8:779–782. doi: 10.1097/JTO.0b013e31828c61a5. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.