Abstract
Background
The significance of indeterminate margins following surgery for non-small cell lung cancer (NSCLC) is unknown. We evaluated the impact of adjuvant therapy on survival in patients with indeterminate margins.
Methods
Patients with indeterminate margins following surgery for NSCLC were identified in the National Cancer Database (NCDB) between 2004–2015, and stratified by receipt of adjuvant treatment. The primary outcome was overall survival, which was evaluated with multivariable Cox Proportional Hazards.
Results
Indeterminate margins occurred in 0.31% of 232,986 patients receiving surgery for NSCLC and was associated with worse survival compared to R0 resection (adjusted hazard ratio [HR] 1.53; 95% confidence interval [CI] 1.40–1.67). Anatomic resection was protective against the finding of indeterminate margins in logistic regression. Amongst 553 patients with indeterminate margins, 343 (62%) received no adjuvant therapy, 96 (17%) received adjuvant chemotherapy, 33 (6%) received adjuvant radiation, and 81 (15%) received adjuvant chemoradiation. Any mode of adjuvant therapy was not associated with improved survival compared to no further treatment.
Conclusion
The finding of indeterminate margins is reported in 0.31% of patients receiving curative-intent surgery for NSCLC. This was associated with worse overall survival compared to complete resection and not mitigated by adjuvant therapy. The risks and benefits of adjuvant therapy should, therefore, be carefully considered for patients with indeterminate margins after surgery for NSCLC.
Graphical Abstract
Indeterminate resection margins were associated with worse survival following surgery for NSCLC, and adjuvant therapy was not associated with improved survival.
Introduction
The finding of positive margins occurs in about 5% of patients following resection of non-small cell lung cancer (NSCLC) and has been associated with worse survival in observational studies1–3. The bronchial stump is most likely to harbor residual cancer, which can manifest as in situ disease, mucosal involvement, or peribronchial infiltration4–10. The National Comprehensive Cancer Network (NCCN) guidelines recommend either re-resection or adjuvant therapy for patients with margin-positive resection for NSCLC, though these guidelines are based on observational studies since patients with incomplete resection have been excluded from prospective, randomized controlled trials examining the utility of adjuvant therapy in lung cancer11,12.
There are no guidelines, however, to direct management of patients who have indeterminate margins on pathological review following surgery for NSCLC. Patients can have indeterminate margins due to both technical and biological factors: sectioning and preparation of the tissue specimen may result in poorly defined margins that cannot be adequately evaluated, or morphology of the cells at the margins can be equivocal for carcinoma. Indeterminate margins have been reported at an incidence of up to 8% in other cancers13–15, but there are no studies examining the incidence or significance of indeterminate margins in lung cancer.
We performed a retrospective cohort study using a large, national database to study the incidence of indeterminate margins following surgery for NSCLC, with two specific aims: (1) We evaluated whether the finding of indeterminate margins impacts survival following surgery for lung cancer, and (2) we examined the impact of adjuvant therapy on survival in patients with indeterminate margins. We hypothesized that patients with indeterminate margins would experience worse survival compared to those with margin-negative resection, and that adjuvant chemotherapy would improve survival in this group of patients.
Methods
Data Source
The National Cancer Database (NCDB) is a collaborative initiative of the American Cancer Society and the American College of Surgeons. It is a prospectively maintained registry containing information about approximately 80% of cancers diagnosed in 1500 treatment centers in the United States every year16. Data are collected by certified, independent tumor registrars.
Patient Selection and Study Design
The study was deemed exempt by our Institutional Review Board. The aim of the first part of the study was to examine the factors associated with indeterminate margins on pathology following surgery for NSCLC, and to identify the effect of indeterminate margins on survival. The 2004–2015 NCDB was used to identify patients with pathologic stage T1–3N0–2M0 NSCLC who underwent surgery. In the first part of the study, patients with missing data about margin status or who did not receive definitive surgery were excluded. These patients were then stratified by margin status: R0, R+ (positive but not otherwise specified as R1 or R2), R1 (microscopically positive margins), R2 (macroscopically positive margins), or RID (indeterminate on pathology). The NCDB describes an indeterminate margin as one that is described in the pathology report as unable to be adequately evaluated for involvement of cancer. A pathology report that does not mention the margin status or one that does not mention if the positive margin is R1 or R2 is coded separately as either missing or positive, not otherwise specified, respectively. The tumor registrars use these standardized coding guidelines in completing entries for the NCDB. However, the NCDB does not specify the anatomic location of or reason for the indeterminate margin. Background characteristics of patients with different margin status were compared using the Wilcoxon rank sum and Pearson’s chi-squared tests for continuous and categorical variables, respectively. . A multivariable logistic regression was then performed to identify factors associated with indeterminate margins compared to negative margins on pathology. In the next part of the study, the effect of margin status on survival was examined. In addition to the exclusion criteria above, patients with missing overall survival information were excluded (Supplemental Figure 1a). The primary outcome was overall survival, which was estimated using Kaplan-Meier and multivariable Cox Proportional Hazards models. Variables for the multivariable Cox models were selected a priori based on availability in the NCDB and perceived clinical importance: age, sex, race, Charlson-Deyo comorbidity index (CDCC) score, rural or urban treatment location, treatment at an academic center, insurance status, type of surgery, histology, and pathologic stage (recoded to AJCC 8th Edition). An additional survival analysis was performed after stratifying patients by pathologic stage. In the second part of the study, the objective was to understand if adjuvant therapy was associated with survival in patients with indeterminate margins. Patients with indeterminate margins following surgery for NSCLC were identified and stratified by adjuvant therapy: none, chemotherapy alone, radiation alone, or chemoradiation. Patients were included regardless of receipt of neoadjuvant chemotherapy or radiation. Patients with 90-day post-operative mortality and missing survival or adjuvant treatment data were excluded (Supplemental Figure 1b). Patients who received adjuvant therapy beyond 180 days from surgery were also excluded because they might have been treated for recurrence, which is not catalogued by the NCDB. The primary outcome was overall survival. The multivariable Cox model included the variables listed above as well as receipt of neoadjuvant therapy and radiation treatment volume. Several additional multivariable survival analyses were performed. (1) Because the survival of patients with indeterminate margins was similar to that of patients with R1 resection, the impact of adjuvant therapy on patients with R1 resection was studied. Analyses were also performed (2) including all patients regardless of 90-day postoperative survival, (3) excluding patients who underwent pneumonectomy, (4) excluding patients who underwent neoadjuvant chemotherapy or radiation, (5) including only patients with pathologic stage I disease, where the importance of adjuvant therapy may be paramount, and (6) comparing the survival of patients with RID status, stratified by receipt of adjuvant therapy, with that of patients with R0 resection.
Missing data were handled with complete case analysis. A comparison of background characteristics between patients who met inclusion criteria and who were excluded due to missing information on survival and adjuvant therapy is presented in Supplemental Table 1, while a comparison between patients with missing and nonmissing covariate data is available in Supplemental Table 2. All statistical analyses were performed using R version 3.5.1 for Mac (Vienna, Austria).
Results
Impact of Indeterminate Margins on Survival
A total of 232,986 patients underwent surgery for pT1–3N0–2M0 NSCLC during the study period: 95.8% had an R0 resection, 1.6% R+, 2.1% R1, 0.16% R2, and 0.31% RID. Patients who had a non-R0 resection were more likely to be female, undergo a non-anatomic resection, have a more advanced pathologic stage, and be less likely to have adenocarcinoma (Table 1). Five-year survival in 209,830 patients with complete survival information was 56% (95% confidence interval [CI] 55–56), 32% (95%CI 30–34), 31% (95%CI 30–33), 24% (95%CI 20–30), and 34% (95%CI 30–38) for patients with R0, R+, R1, R2, and RID status, respectively (Figure 1). In multivariable analysis, the finding of indeterminate margins was associated with worse survival compared to R0 status (Table 2). When examined by pathologic stage, RID status was associated with worse survival in patients with pathologic stage I and III disease (Supplemental Table 3). A multivariable logistic regression revealed that increasing age, a lack of insurance, and advanced pathologic stage were associated with RID status, while female sex, later year of diagnosis, and anatomic resection were protective against indeterminate margin status (Table 3).
Table 1.
Background characteristics of patients who underwent surgery for NSCLC, stratified by margin status. R0= margin negative; R+ = margin positive, not otherwise specified; R1=microscopically margin positive; R2=macroscopically margin positive; RID=margin indeterminate for positivity
| R0 (n=223,250) (%) | R+ (n=3642) (%) | R1 (n=4993) (%) | R2 (n=375) (%) | RID (n=726) (%) | p value | |
|---|---|---|---|---|---|---|
| Age (years, median) | 68 | 68 | 68 | 68 | 70 | <0.001 |
| Sex (female) | 116772(52) | 1671(46) | 2356(47) | 167(45) | 344(47) | <0.001 |
| Race | <0.001 | |||||
| White | 198165(89) | 3150(87) | 4421(89) | 331(88) | 633(88) | |
| Black | 17478(8) | 377(10) | 415(8) | 32(9) | 70(10) | |
| Other | 6075(3) | 85(3) | 138(3) | 11(3) | 13(2) | |
| Year of diagnosis, median (inter-quartile range) | 2010(2007–2013) | 2010(2007–2013) | 2010(2006–2013) | 2009(2006–2012) | 2008(2006–2011) | <0.001 |
| CDCC Score | 0.004 | |||||
| 0 | 116416(52) | 1902(52) | 2500(50) | 164(44) | 380(52) | |
| 1 | 76850(34) | 1252(34) | 1790(36) | 142(38) | 242(33) | |
| 2+ | 29984(13) | 488(14) | 703(14) | 69(18) | 104(15) | |
| Insurance status | <0.001 | |||||
| Private | 68209(31) | 1104(31) | 1455(29) | 119(32) | 192(27) | |
| Government | 148623(67) | 2390(67) | 3407(69) | 242(65) | 497(70) | |
| None | 3527(2) | 91(2) | 83(2) | 10(3) | 20(3) | |
| Facility location | 0.028 | |||||
| Metro | 177752(80) | 2824(78) | 3930(79) | 297(79) | 580(80) | |
| Urban | 33354(15) | 622(17) | 776(15) | 61(16) | 102(14) | |
| Rural | 12144(5) | 196(5) | 287(6) | 17(5) | 44(6) | |
| Academic center | 83223(37) | 1170(32) | 1613(32) | 126(34) | 292(40) | <0.001 |
| Surgery | ||||||
| Wedge resection | 27254(12) | 611(17) | 714(14) | 84(33) | 228(32) | <0.001 |
| Segmentectomy | 7706(4) | 91(3) | 148(3) | 10(2) | 31(4) | |
| Lobectomy | 169905(76) | 2209(61) | 3235(65) | 195(52) | 277(38) | |
| Pneumonectomy | 9130(4) | 322(9) | 476(10) | 26(7) | 28(4) | |
| Other | 9255(4) | 409(10) | 420(8) | 60(16) | 162(22) | |
| Pathologic stage | <0.001 | |||||
| IA | 117958(53) | 850(23) | 1010(20) | 64(17) | 294(40) | |
| IB | 26481(12) | 308(9) | 468(9) | 28(8) | 60(8) | |
| IIA | 13009(6) | 182(5) | 284(6) | 17(4) | 26(4) | |
| IIB | 36154(16) | 918(25) | 1350(27) | 60(16) | 103(14) | |
| IIIA | 27708(12) | 1306(36) | 1745(35) | 200(53) | 231(32) | |
| IIIB | 1940(1) | 78(2) | 136(3) | 6(2) | 12(2) | |
| Histology | <0.001 | |||||
| Adenocarcinoma | 103143(46) | 1409(39) | 1866(37) | 143(38) | 293(40) | |
| Squamous cell cancer | 61166(27) | 1339(37) | 1837(37) | 133(35) | 196(27) | |
| Large cell cancer | 6004(3) | 94(3) | 154(3) | 14(4) | 23(3) | |
| Other | 52937(24) | 800(21) | 1136(23) | 85(23) | 214(30) | |
| Postoperative 90-day mortality | 8659(4) | 303(9) | 373(8) | 46(14) | 54(9) | <0.001 |
Figure 1.
Kaplan-Meier survival curves for patients undergoing surgery for non-small cell lung cancer, stratified by margin status: R0 = margin negative; R+ = margin positive, not otherwise specified; R1= microscopically margin positive; R2= macroscopically margin positive; RID = margin indeterminate. Shaded regions represent the 95% confidence interval for each survival curve. P-value represents the result of the log-rank test
Table 2.
Multivariable Cox Proportional Hazards model for factors associated with survival for patients with NSCLC. R0= margin negative; R+ = positive margin, not otherwise specified; R1 = microscopically positive; R2= macroscopically positive; RID = indeterminate
| 95% Confidence Interval | ||||
|---|---|---|---|---|
| Variable | Hazard Ratio | Lower | Upper | p-value |
| Age (per year) | 1.03 | 1.03 | 1.03 | <0.001 |
| Female sex (reference: male) | 0.75 | 0.74 | 0.76 | <0.001 |
| Race (reference: White) | ||||
| Black | 1.03 | 1.00 | 1.05 | 0.02 |
| Charlson-Deyo Comorbidity Index (reference: 0) | ||||
| 1 | 1.17 | 1.16 | 1.19 | <0.001 |
| 2+ | 1.45 | 1.42 | 1.48 | <0.001 |
| Insurance status (reference: private) | ||||
| Government | 1.18 | 1.16 | 1.20 | <0.001 |
| None | 1.28 | 1.21 | 1.35 | <0.001 |
| Academic center | 0.88 | 0.87 | 0.90 | <0.001 |
| Pathologic stage (reference: IA) | ||||
| IB | 1.29 | 1.26 | 1.31 | <0.001 |
| IIA | 1.41 | 1.37 | 1.45 | <0.001 |
| IIB | 1.77 | 1.74 | 1.80 | <0.001 |
| IIIA | 2.41 | 2.37 | 2.46 | <0.001 |
| IIIB | 2.88 | 2.73 | 3.04 | <0.001 |
| Histology (reference: adenocarcinoma) | ||||
| Squamous cell carcinoma | 1.08 | 1.07 | 1.10 | <0.001 |
| Large cell lung cancer | 1.23 | 1.19 | 1.28 | <0.001 |
| Other | 0.86 | 0.85 | 0.88 | <0.001 |
| Surgery (reference: wedge resection) | ||||
| Segmentectomy | 0.85 | 0.82 | 0.89 | <0.001 |
| Lobectomy | 0.72 | 0.71 | 0.74 | <0.001 |
| Pneumonectomy | 0.91 | 0.88 | 0.94 | <0.001 |
| Margin status (reference: R0) | ||||
| R+ | 1.55 | 1.49 | 1.62 | <0.001 |
| R1 | 1.52 | 1.47 | 1.58 | <0.001 |
| R2 | 1.64 | 1.45 | 1.85 | <0.001 |
| RID | 1.53 | 1.40 | 1.67 | <0.001 |
Table 3.
Multivariable logistic regression model for variables independently associated with indeterminate margins following surgery for NSCLC
| 95% Confidence Interval | ||||
|---|---|---|---|---|
| Predictor | Odds Ratio | Lower | Upper | p-value |
| Age (per year) | 1.01 | 1.00 | 1.02 | 0.02 |
| Female sex (reference: male) | 0.85 | 0.73 | 0.99 | 0.04 |
| Race (reference: White) | ||||
| Black | 1.23 | 0.95 | 1.59 | 0.11 |
| Year of diagnosis | 0.92 | 0.90 | 0.94 | <0.001 |
| Charlson-Deyo Comorbidity Index (reference: 0) | ||||
| 1 | 1.02 | 0.86 | 1.20 | 0.82 |
| 2+ | 1.13 | 0.91 | 1.42 | 0.27 |
| Insurance status (reference: private) | ||||
| Government | 1.09 | 0.89 | 1.32 | 0.41 |
| None | 2.01 | 1.23 | 3.24 | 0.004 |
| Academic center | 1.07 | 0.92 | 1.25 | 0.37 |
| Pathologic stage (reference: IA) | ||||
| IB | 1.15 | 0.86 | 1.53 | 0.34 |
| IIA | 1.02 | 0.67 | 1.56 | 0.93 |
| IIB | 1.54 | 1.22 | 1.96 | <0.001 |
| IIIA | 3.83 | 3.17 | 4.62 | <0.001 |
| IIIB | 3.34 | 1.81 | 6.17 | <0.001 |
| Histology (reference: adenocarcinoma) | ||||
| Squamous cell carcinoma | 0.98 | 0.81 | 1.19 | 0.85 |
| Large cell lung cancer | 1.00 | 0.65 | 1.55 | 0.99 |
| Surgery (reference: wedge resection) | ||||
| Segmentectomy | 0.48 | 0.33 | 0.71 | <0.001 |
| Lobectomy | 0.17 | 0.14 | 0.20 | <0.001 |
| Pneumonectomy | 0.19 | 0.12 | 0.29 | <0.001 |
Impact of Adjuvant Therapy on Survival for Patients with Indeterminate Margins
A total of 553 patients with indeterminate margins met criteria for the second part of the study: 343 (62%), 96 (17%), 33 (6%), and 81 (15%) received no adjuvant therapy, adjuvant chemotherapy, adjuvant radiation, and adjuvant chemoradiation, respectively. Compared to patients not receiving adjuvant therapy, those receiving adjuvant therapy were more likely to have private insurance and an advanced pathologic stage (Table 3). Unadjusted five-year survival, conditional on 90-day postoperative survival, was 44%(95% 39–50) for patients receiving no adjuvant therapy, 38%(95%CI 29–50) for those receiving chemotherapy, 15% (95%CI 6–36) for those receiving radiation, and 26%(95%CI 17–38) for patients undergoing chemoradiation (Figure 2). Adjuvant chemotherapy and chemoradiation were not associated with survival in multivariable regression, but adjuvant radiation was associated with worse survival (Table 4).
Figure 2.
Kaplan-Meier survival curves for patients with indeterminate margins following surgery for non-small cell lung cancer, stratified by type of adjuvant therapy: chemo = chemotherapy alone; RT = radiation alone; CRT = chemotherapy and radiation. Shaded regions represent the 95% confidence interval for each survival curve. P-value represents the result of the log-rank test. Survival is conditional on 90-day postoperative survival
Table 4.
Background characteristics of patients with indeterminate margins, stratified by adjuvant treatment. None=no adjuvant therapy; Chemo=chemotherapy alone; RT=radiation therapy alone; CRT=chemotherapy with radiation
| None (n=343)(%) | Chemo (n=96)(%) | RT (n=33)(%) | CRT (n=81)(%) | p value | |
|---|---|---|---|---|---|
| Age (years, median) | 70 | 66 | 72 | 67 | <0.001 |
| Sex (female) | 182(53) | 41(43) | 10(30) | 31(38) | 0.008 |
| Race | 0.30 | ||||
| White | 304(90) | 84(87) | 25(78) | 73(90) | |
| Black | 27(8) | 11(12) | 6(19) | 8(10) | |
| Other | 8(2) | 1(1) | 1(3) | 0(0) | |
| Year of diagnosis, median (inter-quartile range) | 2008(2006–2011) | 2008(2006–2011) | 2009(2006–2010) | 2008(2006–2011) | 0.56 |
| CDCC Score | 0.80 | ||||
| 0 | 182(53) | 48(50) | 17(52) | 45(56) | |
| 1 | 114(33) | 37(39) | 11(33) | 22(27) | |
| 2+ | 47(14) | 11(11) | 5(15) | 14(17) | |
| Insurance status | 0.009 | ||||
| Private | 81(24) | 39(41) | 6(20) | 31(38) | |
| Government | 243(73) | 52(55) | 24(80) | 48(59) | |
| None | 10(3) | 4(4) | 0(0) | 2(3) | |
| Facility location | 0.33 | ||||
| Metro | 272(79) | 83(87) | 26(79) | 64(79) | |
| Urban | 51(15) | 8(8) | 7(21) | 14(17) | |
| Rural | 20(6) | 5(5) | 0(0) | 3(4) | |
| Academic center | 148(43) | 33(34) | 11(33) | 27(33) | 0.19 |
| Surgery | 0.001 | ||||
| Wedge resection | 133(39) | 25(26) | 11(33) | 24(30) | |
| Segmentectomy | 17(5) | 1(1) | 5(15) | 2(3) | |
| Lobectomy | 142(41) | 47(49) | 8(24) | 31(38) | |
| Pneumonectomy | 13(4) | 6(6) | 1(3) | 2(3) | |
| Other | 38(11) | 17(18) | 8(24) | 21(26) | |
| Pathologic stage | <0.001 | ||||
| IA | 187(54) | 11(12) | 12(36) | 9(11) | |
| IB | 33(10) | 4(4) | 4(12) | 1(1) | |
| IIA | 17(5) | 3(3) | 0(0) | 3(4) | |
| IIB | 46(13) | 23(24) | 4(12) | 13(16) | |
| IIIA | 58(17) | 52(54) | 12(36) | 51(63) | |
| IIIB | 2(1) | 3(3) | 1(3) | 4(5) | |
| Histology | 0.005 | ||||
| Adenocarcinoma | 132(38) | 47(49) | 8(24) | 17(21) | |
| Squamous cell cancer | 85(25) | 25(26) | 10(30) | 39(48) | |
| Large cell cancer | 7(2) | 2(2) | 11(33) | 20(25) | |
| Other | 119(35) | 22(23) | 4(12) | 5(6) | |
| Neoadjuvant chemotherapy | 19(6) | 4(4) | 2(6) | 5(6) | 0.41 |
| Neoadjuvant radiation | 21(6) | 5(5) | 0(0) | (0) | 0.07 |
| Radiation volume (median in Gy with IQR) | 0(0–0) | 0(0–0) | 50.4(40.0–62.5) | 50.4(45.0–63.0) | N/A |
In an analysis of patients regardless of 90-day postoperative survival, adjuvant therapy was not associated with a survival benefit (Supplemental Table 4). Adjuvant therapy was also not associated with a survival benefit in subgroup analyses excluding patients who underwent pneumonectomy, excluding those who received neoadjuvant therapy, or in a cohort of patients with exclusively pathologic stage I disease. In a separate analysis of patients with R1 resection, adjuvant chemotherapy and chemoradiation were associated with improved survival while adjuvant radiation was not. In an analysis comparing patients with R0 and R1 resection with those with RID margin status stratified by receipt of adjuvant therapy (Supplemental Figure 2), RID patients receiving adjuvant therapy (HR 1.51; 95%CI 1.31–1.75; p<0.001) and not receiving adjuvant therapy (HR 1.54; 95%CI 1.37–1.73; p<0.001) had worse survival compared to R0 patients on multivariable analysis.
Discussion
In this NCDB analysis, we found that patients with indeterminate margins following surgery for NSCLC experienced worse survival compared to those with an R0 resection. Anatomic resection was independently associated with a low probability of indeterminate margins. Further, the receipt of adjuvant therapy was not associated with a survival benefit even after adjustment for pathologic stage. Our study suggests that the risks and benefits of adjuvant therapy should be weighed carefully before offering it to patients with indeterminate margins.
To our knowledge this is the first study to describe the incidence and importance of the finding of indeterminate margins in lung cancer, though our study corroborates the equivocal observational literature about the utility of adjuvant therapy for margin-positive resection for lung cancer. While the NCCN guidelines recommend consideration of either re-resection or adjuvant therapy for positive margins, retrospective cohort studies have had conflicting findings, with some demonstrating no benefit with adjuvant therapy for positive margins4,6,7, and others demonstrating a survival benefit with adjuvant chemotherapy for earlier stage disease and chemoradiation for locally advanced cancer1–3. These studies reflect the heterogeneity of patients with incomplete resection: they may have been unable to tolerate an extensive operation, received an inadequate resection or lymphadenectomy, or have remnant in situ disease rather than invasion17.
Our study revealed that patients with indeterminate margins who underwent adjuvant therapy did not experience a survival benefit compared to those who did not have adjuvant therapy. One explanation is that adjuvant therapy truly does not confer a significant survival benefit in this patient population. Another is that there were too few patients in our study to establish a meaningful survival benefit, though the point estimates and confidence intervals do not suggest this. A third possibility is that patients with incomplete resection, as observed above, are a heterogeneous population, and numerous technical and biologic factors can result in indeterminate margins. As a result, our finding that patients with indeterminate margins do not benefit from adjuvant therapy must be understood as the aggregate effect seen in a disparate population. The decision to offer adjuvant therapy should therefore be tailored to the patient in a multi-disciplinary setting like tumor board. There are few studies on the significance of indeterminate margins in other cancers, with no studies examining the role of adjuvant therapy.
Our study has several limitations. The most important limitation in a retrospective cohort analysis is selection bias, because we do not understand the reasons that each patient was assigned to a treatment, nor do we have access to the medical records that might shed more light on this issue. For instance, patients with possible in situ disease remaining in the bronchial stump may have been less likely to receive adjuvant therapy, but may also have had a better prognosis than someone who received adjuvant therapy for the possibility of invasion. Some patients may have had a preoperative plan for adjuvant therapy as well. For example, many patients in our cohort had more advanced disease, which was likely the major driver for adjuvant therapy. The decision to offer adjuvant therapy may also have been influenced by patients’ postoperative clinical course, and the NCDB does not have detailed information about this; we used 90-day postoperative survival as a proxy. The decision to offer adjuvant therapy is multidisciplinary and is predicated on numerous factors that were not available to us in this study. We attempted to adjust for these differences with multivariable regression and supplemental analyses. Our study was also limited by a small cohort size, though the rarity of the finding of indeterminate margins makes the NCDB a better source of data than single-institution studies. The study was also limited by the accuracy of coding in the NCDB. A small cohort like ours is susceptible to significant shifts in analysis if even a small number of patients had been miscategorized. Similarly, missing data may also have affected our findings, since only 76% of patients with indeterminate margins were included in the final cohort. Since most of the excluded patients had 90-day postoperative mortality or unknown 90-day postoperative survival, we performed an additional analysis (Supplemental Table 4) in patients regardless of 90-day postoperative survival, with similar findings. In addition, while we used complete case analysis for our multivariable regression, only 18 patients (3%) were excluded due to missing covariate data (Supplemental Table 2). We were limited by the variables available in the NCDB as well, which does not catalogue the reasons a margin was deemed indeterminate, the anatomic location of the indeterminate margin, disease recurrence, repeat resection, type of chemotherapy, preoperative pulmonary function, or disease-free survival.
In this NCDB analysis, the finding of indeterminate margins was reported in about 0.31% of patients undergoing surgery for NSCLC, and was associated with worse survival compared to patients who had R0 resection (Figure 3). Anatomic resection was found to be protective against the finding of indeterminate margins. The addition of adjuvant therapy was not associated with improved survival for patients with indeterminate margins even after adjustment for pathologic stage. The risks and benefits should be weighed carefully before offering adjuvant therapy to patients with indeterminate margins.
Figure 3.
In this National Cancer Database analysis, the finding of indeterminate margins was reported in 0.31% of patients undergoing resection for non-small cell lung cancer. Indeterminate margins were associated with worse survival compared to margin-negative resection. Adjuvant therapy in any form was not associated with a survival benefit compared to no adjuvant therapy in patients with indeterminate margins. RT= radiation; ChemoRT = chemotherapy and radiation
Supplementary Material
Supplemental Figure 2. Kaplan-Meier survival curves for patients undergoing surgery for non-small cell lung cancer, stratified by margin status and receipt of adjuvant therapy: R0 = margin negative; R1= microscopically margin positive; RID (no adjuvant) = margin indeterminate without any adjuvant therapy; RID (with adjuvant)= margin indeterminate with any type of adjuvant therapy. Shaded regions represent the 95% confidence interval for each survival curve. P-value represents the result of the log-rank test
Supplemental Figure 1. Patient selection scheme for (a) first part of study examining the comparative survival of patients by margin status and (b) the second part of the study examining the impact of adjuvant therapy on survival in patients with indeterminate margins following resection for lung cancer
Table 5.
Multivariable Cox Proportional Hazards model for factors associated with survival for patients with indeterminate margins following surgery for NSCLC, conditional on 90-day postoperative survival
| 95% Confidence Interval | ||||
|---|---|---|---|---|
| Variable | Hazard Ratio | Lower | Upper | p-value |
| Age (per year) | 1.02 | 1.01 | 1.04 | 0.002 |
| Female sex (reference: male) | 0.68 | 0.54 | 0.86 | 0.001 |
| Race (reference: White) | ||||
| Black | 0.68 | 0.45 | 1.04 | 0.07 |
| Charlson-Deyo Comorbidity Index (reference: 0) | ||||
| 1 | 1.24 | 0.97 | 1.58 | 0.08 |
| 2+ | 1.47 | 1.07 | 2.02 | 0.02 |
| Insurance status (reference: private) | ||||
| Government | 1.23 | 0.93 | 1.64 | 0.15 |
| None | 1.78 | 0.97 | 3.27 | 0.06 |
| Academic center | 1.06 | 0.84 | 1.34 | 0.64 |
| Pathologic stage (reference: IA) | ||||
| IB | 1.28 | 0.82 | 1.99 | 0.28 |
| IIA | 1.05 | 0.58 | 1.92 | 0.87 |
| IIB | 1.43 | 0.98 | 2.07 | 0.06 |
| IIIA | 2.46 | 1.79 | 3.36 | <0.001 |
| IIIB | 1.65 | 0.73 | 3.73 | 0.23 |
| Neoadjuvant therapy | 1.24 | 0.73 | 2.12 | 0.42 |
| Histology (reference: adenocarcinoma) | ||||
| Squamous cell carcinoma | 1.04 | 0.80 | 1.36 | 0.77 |
| Large cell lung cancer | 1.13 | 0.62 | 2.08 | 0.69 |
| Other | 0.83 | 0.63 | 1.09 | 0.18 |
| Surgery (reference: wedge resection) | ||||
| Segmentectomy | 0.79 | 0.46 | 1.34 | 0.38 |
| Lobectomy | 0.63 | 0.48 | 0.83 | 0.001 |
| Pneumonectomy | 0.77 | 0.43 | 1.38 | 0.38 |
| Radiation treatment volume (per cGy) | 1.00 | 1.00 | 1.00 | 0.63 |
| Adjuvant therapy (reference: none) | ||||
| Chemotherapy | 1.08 | 0.80 | 1.47 | 0.61 |
| Radiation | 1.76 | 1.13 | 2.74 | 0.01 |
| Chemoradiation | 1.31 | 0.92 | 1.87 | 0.13 |
Perspective Statement.
Our study demonstrates that indeterminate margins are associated with worse survival in NSCLC compared to R0 resection. However, adjuvant therapy is not associated with improved survival in patients with indeterminate margins following surgery for NSCLC, and the risks and benefits of adjuvant therapy should be weighed before offering it to these patients.
Acknowledgements and Funding:
The American College of Surgeons is in a Business Associate Agreement that includes a data use agreement with each of its Commission on Cancer accredited hospitals. The data used in the study are derived from a de-identified National Cancer Data Base file. The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analytic or statistical methodology used or the conclusions drawn from these data by the investigators.
We are grateful to Hanghang Wang, M.D., Ph.D., for statistical support.
Drs. Raman and Voigt were supported by a National Institutes of Health T-32 grant 5T32CA093245 in surgical oncology. Dr. Jawitz was supported by a National Institutes of Health T-32 grant 5T32HL069749 in clinical research.
Footnotes
This paper was presented at the Western Thoracic Surgical Association annual meeting in Squaw Valley in June 2019.
References:
- 1.Smeltzer MP, Lin CC, Kong F-M (Spring), Jemal A, Osarogiagbon RU. Survival impact of postoperative therapy modalities according to margin status in non–small cell lung cancer patients in the United States. The Journal of Thoracic and Cardiovascular Surgery. 2017;154(2):661–672.e10. doi: 10.1016/j.jtcvs.2017.03.085 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hancock JG, Rosen JE, Antonicelli A, et al. Impact of Adjuvant Treatment for Microscopic Residual Disease After Non-Small Cell Lung Cancer Surgery. The Annals of Thoracic Surgery. 2015;99(2):406–413. doi: 10.1016/j.athoracsur.2014.09.033 [DOI] [PubMed] [Google Scholar]
- 3.Osarogiagbon RU, Lin CC, Smeltzer MP, Jemal A. Prevalence, Prognostic Implications, and Survival Modulators of Incompletely Resected Non–Small Cell Lung Cancer in the U.S. National Cancer Data Base. Journal of Thoracic Oncology. 2016;11(1):e5–e16. doi: 10.1016/j.jtho.2015.08.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hofmann HS, Taege C, Lautenschläger C, Neef H, Silber RE. Microscopic (R1) and macroscopic (R2) residual disease in patients with resected non-small cell lung cancer. Eur J Cardiothorac Surg. 2002;21(4):606–610. doi: 10.1016/S1010-7940(02)00030-1 [DOI] [PubMed] [Google Scholar]
- 5.Massard G, Doddoli C, Gasser B, et al. Prognostic implications of a positive bronchial resection margin. Eur J Cardiothorac Surg. 2000;17(5):557–565. doi: 10.1016/S1010-7940(00)00384-5 [DOI] [PubMed] [Google Scholar]
- 6.Lequaglie C, Conti B, Brega Massone PP, Giudice G. Unsuspected residual disease at the resection margin after surgery for lung cancer: fate of patients after long-term follow-up. Eur J Cardiothorac Surg. 2003;23(2):229–232. doi: 10.1016/S1010-7940(02)00750-9 [DOI] [PubMed] [Google Scholar]
- 7.Ghiribelli C, Voltolini L, Paladini P, Luzzi L, Di Bisceglie M, Gotti G. Treatment and survival after lung resection for non-small cell lung cancer in patients with microscopic residual disease at the bronchial stump. Eur J Cardiothorac Surg. 1999;16(5):555–559. doi: 10.1016/S1010-7940(99)00310-3 [DOI] [PubMed] [Google Scholar]
- 8.Sawabata N, Ohta M, Matsumura A, et al. Optimal distance of malignant negative margin in excision of nonsmall cell lung cancer: a multicenter prospective study. The Annals of Thoracic Surgery. 2004;77(2):415–420. doi: 10.1016/S0003-4975(03)01511-X [DOI] [PubMed] [Google Scholar]
- 9.El-Sherif A, Fernando HC, Santos R, et al. Margin and Local Recurrence After Sublobar Resection of Non-Small Cell Lung Cancer. Ann Surg Oncol. 2007;14(8):2400–2405. doi: 10.1245/s10434-007-9421-9 [DOI] [PubMed] [Google Scholar]
- 10.Snijder RJ, Brutel de la Rivière A, Elbers HJJ, van den Bosch JMM. Survival in Resected Stage I Lung Cancer With Residual Tumor at the Bronchial Resection Margin. The Annals of Thoracic Surgery. 1998;65(1):212–216. doi: 10.1016/S0003-4975(97)01114-4 [DOI] [PubMed] [Google Scholar]
- 11.Winton T, Livingston R, Johnson D, et al. Vinorelbine plus Cisplatin vs. Observation in Resected Non–Small-Cell Lung Cancer. New England Journal of Medicine. 2005;352(25):2589–2597. doi: 10.1056/NEJMoa043623 [DOI] [PubMed] [Google Scholar]
- 12.Cisplatin-Based Adjuvant Chemotherapy in Patients with Completely Resected Non–Small-Cell Lung Cancer. New England Journal of Medicine. 2004;350(4):351–360. doi: 10.1056/NEJMoa031644 [DOI] [PubMed] [Google Scholar]
- 13.Makazu M, Sakamoto T, So E, et al. Relationship between indeterminate or positive lateral margin and local recurrence after endoscopic resection of colorectal polyps. Endosc Int Open. 2015;3(3):E252–E257. doi: 10.1055/s-0034-1391853 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mirkin KA, Hollenbeak CS, Wong J. Survival impact of neoadjuvant therapy in resected pancreatic cancer: A Prospective Cohort Study involving 18,332 patients from the National Cancer Data Base. International Journal of Surgery. 2016;34:96–102. doi: 10.1016/j.ijsu.2016.08.523 [DOI] [PubMed] [Google Scholar]
- 15.Smitt MC, Nowels KW, Zdeblick MJ, et al. The importance of the lumpectomy surgical margin status in long-term results of breast conservation. Cancer. 1995;76(2):259–267. [DOI] [PubMed] [Google Scholar]
- 16.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(3):683–690. doi: 10.1245/s10434-007-9747-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.D’Amico TA. Get it right the first time. The Journal of Thoracic and Cardiovascular Surgery. 2017;154(2):650–651. doi: 10.1016/j.jtcvs.2017.03.126 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure 2. Kaplan-Meier survival curves for patients undergoing surgery for non-small cell lung cancer, stratified by margin status and receipt of adjuvant therapy: R0 = margin negative; R1= microscopically margin positive; RID (no adjuvant) = margin indeterminate without any adjuvant therapy; RID (with adjuvant)= margin indeterminate with any type of adjuvant therapy. Shaded regions represent the 95% confidence interval for each survival curve. P-value represents the result of the log-rank test
Supplemental Figure 1. Patient selection scheme for (a) first part of study examining the comparative survival of patients by margin status and (b) the second part of the study examining the impact of adjuvant therapy on survival in patients with indeterminate margins following resection for lung cancer