- Email: [email protected]
The Impact of Postoperative Radiotherapy for Thymoma and Thymic Carcinoma
ORIGINAL ARTICLE
The Impact of Postoperative Radiotherapy for Thymoma and Thymic Carcinoma Matthew W. Jackson, MD,a,* David A. Palma, MD, MsC, PhD,b D. Ross Camidge, MD, PhD,c Bernard L. Jones, PhD,a Tyler P. Robin, MD, PhD,a David J. Sher, MD, MPH,d Matthew Koshy, MD,e Brian D. Kavanagh, MD, MPH,a Laurie E. Gaspar, MD, MBA,a Chad G. Rusthoven, MDa a
Department of Radiation Oncology, University of Colorado Cancer Center, Aurora, Colorado London Health Sciences Centre, London, Ontario, Canada c Division of Medical Oncology, University of Colorado Cancer Center, Aurora, Colorado d Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, Texas e Department of Radiation Oncology and Cellular Oncology, Department of Radiation Oncology, University of Illinois at Chicago, Chicago, Illinois b
Received 7 April 2016; revised 4 December 2016; accepted 1 January 2017 Available online - 25 January 2017
ABSTRACT Introduction: The optimal role for postoperative radiotherapy (PORT) for thymoma and thymic carcinoma remains controversial. We used the National Cancer Data Base to investigate the impact of PORT on overall survival (OS). Methods: Patients who underwent an operation for thymoma or thymic carcinoma were categorized into Masaoka-Koga stage groups I to IIA, IIB, III, and IV. Patients who did not undergo an operation or those who received preoperative radiation were excluded. Kaplan-Meier estimates of OS and univariate and multivariate Cox proportional hazards regression analyses were performed. Propensity score–matched analyses were performed to further control for baseline confounders. Results: From 2004 to 2012, 4056 patients were eligible for inclusion, 2001 of whom (49%) received PORT. On multivariate analysis of OS in the thymoma cohort adjusted for age, WHO histologic subtype, Masaoka-Koga stage group, surgical margins, and chemotherapy administration, PORT was associated with superior OS (hazard ratio [HR] ¼ 0.72, p ¼ 0.001). Propensity score–matched analyses confirmed the survival advantage associated with PORT. Subset analysis indicated longer OS in association with PORT for patients with stage IIB thymoma (HR ¼ 0.61, p ¼ 0.035), stage III (HR ¼ 0.69, p ¼ 0.020), and positive margins (HR ¼ 0.53, p < 0.001). The impact of PORT for stage I to IIA disease did not reach significance (HR ¼ 0.76, p ¼ 0.156). Conclusions: In this large database analysis of PORT for thymic tumors, PORT was associated with longer OS, with the greatest relative benefits observed for stage IIB to III disease and positive margins. In the absence of randomized
Journal of Thoracic Oncology
Vol. 12 No. 4: 734-744
studies assessing the value of PORT, these data may inform clinical practice. 2017 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. Keywords: Thymoma; Thymic carcinoma; Adjuvant radiation; Postoperative radiotherapy; National Cancer Data Base
Introduction The thymic epithelial tumors, thymoma and thymic carcinoma, represent the most frequent tumors of the anterior mediastinum, although they remain relatively rare overall.1 Surgery stands as the primary therapeutic modality in the curative setting, as extent of resection has consistently been shown to be independently prognostic of improved survival.2,3 The intimate association of these tumors with critical mediastinal structures can make complete resection difficult, particularly for advanced-stage disease.4,5 As a result, there has been a longstanding interest in utilizing
*Corresponding author. Disclosure: The authors declare no conflict of interest. Address for correspondence: Matthew W. Jackson, MD, Department of Radiation Oncology, University of Colorado Cancer Center, 1665 Aurora Court, Suite 1032 MS F706, Aurora, CO 80045. E-mail: matthew. [email protected] ª 2017 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. ISSN: 1556-0864 http://dx.doi.org/10.1016/j.jtho.2017.01.002
April 2017
postoperative radiotherapy (PORT) to improve outcomes. However, the precise indications for PORT and its magnitude of benefit have been difficult to ascertain.6–8 The low annual incidence of thymic tumors renders the execution of randomized trials challenging, leaving only limited, conflicting data regarding the utility of PORT. Large registry data sets, such as the Surveillance Epidemiology and End Results (SEER) database, have proved useful for investigating clinical questions in thymic tumors and other rare diseases for which there are inadequate data to address relevant questions. Given a lack of prospective evidence, several investigators have previously utilized the SEER database and other national data sets to investigate the role of PORT for thymic epithelial tumors.9–17 Although these studies have been fairly heterogeneous in their design, consistent themes have emerged. PORT appears beneficial for patients with advanced disease (Masaoka-Koga stage III–IV)9,10 but of minimal utility for completely encapsulated tumors (Masaoka-Koga stage I),13–15 and it remains controversial for Masaoka-Koga stage II disease.15,16 Improved outcomes with the addition of PORT have been more readily demonstrated for patients with aggressive thymic carcinoma, as opposed to for those with thymoma histologic subtypes.18–21 Many retrospective, large-database studies have been unable to control for important confounders, including surgical margin status, histologic subtype, and use of chemotherapy; these studies have called for future studies to incorporate the aforesaid variables to provide more robust analyses. The National Cancer Data Base (NCDB), which was queried for this study, includes additional information on these important variables. As a large hospital-based registry representing more than two-thirds of all cancer cases in the United States, the NCDB draws from a substantial number of institutions nationwide, thereby increasing the number of cases available for analysis. In this largest reported series on PORT for thymic tumors, we utilize the NCDB, with its unprecedented body of critical patient- and disease-specific factors, to describe outcomes and prognostic factors that identify patients who may benefit from adjuvant radiotherapy.
Methods Data Source and Patient Selection The NCDB is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. It is a hospital-based registry that represents 70% of all cancer cases in the United States, drawing data from more than 1500
PORT for Thymoma and Thymic Carcinoma
735
commission-accredited cancer programs. The data used in the study are derived from a deidentified NCDB 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 for the conclusions drawn from these data by the investigators.22 Before the present analysis was begun, appropriate exemption was obtained from our institutional review board. The analysis was restricted to adult patients (age 18 years) in whom thymoma or thymic carcinoma, defined as International Classification of Disease for Oncology, Third Edition (ICD-O-3), histologic codes for thymoma (8580–8585) and thymic carcinoma (8002, 8010, 8012, 8013, 8020, 8021, 8070-8072, 8074, 8140, 8240, 8243, 8246, 8586, 8588, and 8589) with corresponding topographic codes C37.9 (thymus) and C38.1 (anterior mediastinum), had been diagnosed.23 All patients included in the analysis were recorded as having undergone an operation and had their disease diagnosed from 2004 to 2012. All included cases had known treatment-related variables with respect to surgery, chemotherapy, and radiotherapy delivery, with unknowns excluded from this analysis. Treatment data coded in the NCDB are limited to the first course of treatment, defined as all methods of treatment recorded in the treatment plan and administered to the patient before disease progression or recurrence. Cases for which radiation was coded as being delivered before surgery or intraoperatively were excluded. Cases in which death occurred within 1 month of diagnosis were omitted from analysis.
Patient Demographics and Treatment Variables Potentially relevant patient and treatment characteristics were selected a priori. Retrievable data included age, race, sex, insurance status, median income, treatment facility type, distance to facility, Charlson-Deyo comorbidity score, year of diagnosis, surgical margin status, WHO histologic subtype, use of radiation, use of chemotherapy, and derived MasaokaKoga stage group. To facilitate clinical applicability of this report, cases within the database were categorized according to the Masaoka-Koga staging system for thymic epithelial tumors: stage I to IIA (invasive tumor confined to gland of origin or localized, NOS), stage IIB (adjacent connective tissue), stage III (adjacent organs/structures in mediastinum), and stage IV (further contiguous extension, lymph nodes, or metastasis). As neither the SEER database nor the NCDB codes for microscopic transcapsular invasion, stage I disease could not be distinguished from stage IIA disease.
736 Jackson et al
Journal of Thoracic Oncology
Vol. 12 No. 4
Table 1. Patient- and Disease-Specific Factors by Treatment Group Thymoma
Characteristic Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown Masaoka stage I–IIA IIB III IV Unknown
Thymic Carcinoma
Surgery Alone
Surgery Plus PORT
No.
%
No.
%
1100 73 259 134 21
(69.3) (4.6) (16.3) (8.4) (1.3)
1014 61 241 113 15
(70.2) (4.2) (16.7) (7.8) (1.0)
729 858
(45.9) (54.1)
706 738
(48.9) (51.1)
785 65 93 589 9 46
(49.5) (4.1) (5.9) (37.1) (0.6) (2.9)
831 39 89 438 21 26
(57.5) (2.7) (6.2) (30.3) (1.5) (1.8)
236 314 416 604
(15.9) (19.8) (26.2) (38.1)
222 307 403 512
(15.4) (21.3) (27.9) (35.5)
89 570 707 68 153
(5.6) (35.9) (44.5) (4.3) (9.6)
98 552 551 89 154
(6.8) (38.2) (38.2) (6.2) (10.7)
1346 241
(84.8) (15.2)
1322 122
(91.6) (8.4)
1202 306 79
(75.7) (19.3) (5.0)
1132 255 57
(78.4) (17.7) (3.9)
423 500 664
(26.7) (31.5) (41.8)
475 502 467
(32.9) (34.8) (32.3)
1147 103 145 32 160
(72.3) (6.5) (9.1) (2.0) (10.1)
733 188 331 60 132
(50.8) (13.0) (22.9) (4.2) (9.1)
813 274 317 136 47
(51.2) (17.3) (20.0) (8.6) (3.0)
431 359 451 165 38
(29.8) (24.9) (31.2) (11.4) (2.6)
p Value
Surgery Alone
Surgery Plus PORT
No.
%
No.
%
339 30 61 34 4
(72.4) (6.4) (13.0) (7.3) (0.9)
398 26 83 47 3
(71.5) (4.7) (14.9) (8.4) (0.5)
296 172
(63.2) (36.8)
346 211
(62.1) (37.9)
216 15 22 200 5 10
(46.2) (3.2) (4.7) (42.7) (1.1) (2.1)
305 17 35 187 5 8
(54.8) (3.1) (6.3) (33.6) (0.9) (1.4)
74 110 123 161
(15.8) (23.5) (26.3) (34.4)
86 127 129 215
(15.4) (22.8) (23.2) (38.6)
20 172 190 34 52
(4.3) (36.8) (40.6) (7.3) (11.1)
38 229 208 39 43
(6.8) (41.1) (37.3) (7.0) (7.7)
390 78
(83.3) (16.7)
504 53
(90.5) (9.5)
334 98 36
(71.4) (20.9) (7.7)
424 105 28
(76.1) (18.9) (5.0)
144 155 169
(30.8) (33.1) (36.1)
159 180 218
(28.5) (32.3) (39.1)
287 48 60 25 48
(61.3) (10.3) (12.8) (5.3) (10.3)
263 87 121 32 54
(47.2) (15.6) (21.7) (5.7) (9.7)
174 48 140 85 21
(37.2) (10.3) (29.9) (18.2) (4.5)
154 85 211 95 12
(27.6) (15.3) (37.9) (17.1) (2.2)
0.868
p Value 0.576
0.104
0.710
<0.001
0.046
0.369
0.511
0.003
0.090
<0.001
0.001
0.169
0.121
<0.001
0.579
<0.001
<0.001
<0.001
<0.001
(continued)
April 2017
PORT for Thymoma and Thymic Carcinoma
737
Table 1. Continued Thymoma
Thymic Carcinoma
Surgery Alone
Surgery Plus PORT
Characteristic
No.
%
No.
%
Chemotherapy No Yes
1361 226
(85.8) (14.2)
1131 313
(78.3) (21.7)
p Value
Surgery Alone
Surgery Plus PORT
No.
%
No.
%
337 131
(72.0) (28.0)
247 310
(44.3) (55.7)
<0.001
p Value <0.001
Note: Of the patients with thymoma, 1587 (median age 60 years, age range 18–90 years) were treated with surgery alone and 1444 (median age 59 years, age range 18–90 years) were treated with surgery plus PORT. Of the patients with thymic carcinoma, 468 (median age 63 years, age range 19–89 years) were treated with surgery alone and 557 (median age 60 years, age range 20–90 years) were treated with surgery plus PORT. PORT, postoperative radiotherapy; NOS, not otherwise specified.
Median household income in the patient’s home zip code was recorded in quartiles relative to the U.S. population. Distance to treatment facility was recorded as the distance in miles between the hospital that reported the case and the patient’s zip code centroid or city if the zip code was not available. Patient comorbidities were categorized by comorbidity score as described by Charlson and Deyo (1992), and cases were coded as 0, 1, or 2 or higher.24
replacement was performed by using the caliper match algorithm described by Coca-Perraillon,26 with the caliper width set to 0.05 times the standard deviation of the logit of the propensity score.27 These outcomes were assessed with a log-rank test, and the hazard ratio (HR) was determined by univariate Cox regression. The score calculation was blinded with respect to patient outcomes.
Results Statistical Analyses
Patient Characteristics
All statistical analyses were performed using SPSS, version 23.0 (SPSS, Inc., Chicago, IL). Median follow-up was calculated with the reverse Kaplan-Meier method described by Schemper and Smith.25 Pearson chi-square tests were used to evaluate patient- and treatmentrelated factors between patients who did or did not receive PORT. The primary end point was overall survival (OS) which is defined in the NCDB as date of diagnosis to the date of death. OS estimates were generated by using the Kaplan-Meier method and comparisons were made by using the log-rank test. Univariate hazard estimates were generated with unadjusted Cox proportional hazards models. Multivariate analysis for OS were performed by using Cox models adjusted for all variables selected a priori: age, race, sex, insurance status, median income, treatment facility type, distance to facility, Charlson-Deyo comorbidity score, year of diagnosis, surgical margin status, WHO histologic subtype, use of radiation, use of chemotherapy, and derived Masaoka-Koga stage group. To further account for confounding variables, we performed propensity score matching for patients treated with or without PORT, accounting for the same variables described in the preceding paragraph. The propensity score was calculated by using logistic regression to estimate the probability of receiving radiation. One-to-one propensity matching without
A total of 4056 patients with thymoma (n ¼ 3031) or thymic carcinoma (n ¼ 1025) who met predefined selection criteria were identified. Median follow-up was 57.2 months (range 1.08–129.15 months) for patients with thymoma and 59.5 months (range 1.15–130.23) for patients with thymic carcinoma. The median age for patients with thymoma was 60 years (range 18–90 and older) and that for patients with thymic carcinoma was 61 years (range 19–90 and older). Use of adjuvant radiotherapy was evenly split among the cohort, as 1444 patients with thymoma (47.6%) and 557 patients with thymic carcinoma (54.3%) received PORT. The interquartile range for time from surgery to the start of radiation for all patients was 41 to 77 days. The overwhelming majority of patients presented with localized disease, with a higher proportion of thymic carcinomas (17.5%) being Masaoka-Koga stage IV at diagnosis. Patient- and disease-specific characteristics for both patients with thymoma and thymic carcinoma as well as for the surgery-alone and PORT treatment groups are listed in Table 1. Differences were identified between the PORT and surgery-alone groups, with a general preponderance of adverse risk factors among those receiving PORT. Disease factors associated with the use of PORT included positive margins and advanced stage. Demographic factors associated with receipt of PORT included private medical
738 Jackson et al
Journal of Thoracic Oncology
Table 2. Thymoma: Univariate Analysis of Predictors of Overall Survival
Vol. 12 No. 4
Table 2. Continued Univariate
Univariate Variable Local Therapy Surgery alone Surgery þ PORT Age Continuous by year Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/ managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown
HR
95% CI
p Value
Variable
HR
95% CI
p Value
Histologic subtype Thymoma, NOS Type A Type AB Type B1 Type B2 Type B3 Thymic carcinoma
1 1.15 0.79 0.89 1.04 1.12
0.83–1.59 0.60–1.04 0.65–1.21 0.78–1.38 0.86–1.45
0.397 0.090 0.462 0.788 0.396
1 0.80
0.67–0.96
0.015
1.05
1.04–1.05
<0.001
1 1.16 1.14 0.62 1.16
0.76–1.76 0.90–1.44 0.41–0.94 0.55–2.46
0.500 0.283 0.025 0.695
1 1.01
Masaoka-Koga stage I–IIA IIB III IV Unknown
1 1.18 2.09 2.15 1.15
0.90–1.54 1.67–2.60 1.61–2.86 0.64–2.08
0.224 <0.001 <0.001 0.638
0.85–1.21
0.880
Chemotherapy No Yes
1 1.32
1.07–1.62
0.010
1 1.11 1.33 2.70 1.06
0.62–1.99 0.87–2.03 2.24–3.26 0.34–3.31
0.731 0.193 <0.001 0.924
1.10
0.56–2.15
0.783
1 0.89 0.79 0.65
0.68–1.17 0.61–1.03 0.50–0.84
0.416 0.079 0.001
1 0.90
0.64–1.26
0.533
0.75 0.87
0.54–1.06 0.54–1.38
0.100 0.547
0.38
0.23–0.62
<0.001
1 0.99
0.75–1.31
0.927
1 1.62 2.08
1.31–2.00 1.47–2.95
<0.001 <0.001
1 0.89 0.78
0.72–1.09 0.59–1.03
0.248 0.076
1 1.74 1.28 3.58 1.44
1.32–2.30 1.00–1.64 2.58–4.97 1.08–1.93
<0.001 0.049 <0.001 0.012 (continued)
HR, hazard ratio; CI, confidence interval; PORT, postoperative radiotherapy; NOS, not otherwise specified; R1, microscopic residual tumor; R2, macroscopic residual tumor.
insurance and residence within 50 miles of the treating facility.
OS for Patients with Thymoma On univariate analysis of the entire thymoma cohort, despite the predominance of adverse disease-specific factors in the PORT group, PORT was associated with improved OS (HR ¼ 0.80, 95% confidence interval [CI]: 0.67–0.96, p ¼ 0.015). This includes all patients with thymoma irrespective of margin status. Univariate survival analyses for all variables are displayed in Table 2. Kaplan-Meier survival curves for stage-specific subgroups are displayed in the top row of Figure (Fig. 1A–C). On multivariate analysis, PORT remained independently associated with improved OS (HR ¼ 0.72, 95% CI: 0.59–0.87, p ¼ 0.001). A high median income was also associated with improved OS. The analysis showed an overall trend toward better outcomes in the latter years of the study, which reached significance for the years 2010–2012. Variables independently associated with inferior OS on multivariate analysis include older age, increased comorbidity score, positive margins, advanced Masaoka-Koga stage (III–IV), and use of chemotherapy (Table 3). Propensity score–matched treatment groups (n ¼ 1015 per group) were well balanced for all patient- and treatment-related factors between the surgery-alone and PORT groups. For the propensity-matched thymoma cohort, OS in patients receiving PORT was superior to that
April 2017
PORT for Thymoma and Thymic Carcinoma
739
Figure 1. (Top row) Kaplan-Meier curves depicting overall survival for patients with thymoma: stage I to IIA (A), stage IIB (B), and stage III (C). (Bottom row) Kaplan-Meier curves depicting overall survival for patients with thymic carcinoma: stage I to IIA (D), stage IIB (E), and stage III (F). Abbreviation: PORT, postoperative radiotherapy.
in patients receiving an operation alone (HR ¼ 0.64, CI: 0.52–0.79, p < 0.001).
OS for Patients with Thymic Carcinoma On univariate analysis of the entire thymic carcinoma cohort, PORT was associated with improved OS (HR ¼ 0.79, 95% CI: 0.64–0.97, p ¼ 0.025). Univariate survival analyses for all variables are displayed in Table 4. Kaplan-Meier survival curves for stage-specific subgroups are displayed in the bottom row of Figure 1 (Fig. 1D–F). This includes all patients with thymic carcinoma irrespective of margin status. On multivariate analysis, PORT remained independently associated with improved OS (HR ¼ 0.73, 95% CI: 0.58–0.91, p ¼ 0.006). Variables independently associated with inferior OS on multivariate analysis include older age, uninsured status, positive margins, and advanced Masaoka-Koga stage (Table 5). Propensity score–matched treatment groups (n ¼ 345 per group) were well balanced for all patient- and treatment-related factors between the surgery-alone and PORT groups. For the propensity-matched thymoma cohort, OS in patients receiving PORT was superior to that in patients receiving an operation alone (HR ¼ 0.73, 95% CI: 0.57–0.94, p ¼ 0.016).
Secondary and Subgroup Analyses Subgroup analysis for patients with thymoma was performed to evaluate the impact of PORT by histologic
subtype, Masaoka-Koga stage group, and surgical margin status, with the corresponding forest plots depicted in Figure 2. Utilization of PORT was associated with significantly longer survival in patients with the following characteristics: histologic subtypes A and B, stage IIB or III disease, and positive margins. To help exclude the possibility that positive margins were influencing the benefit of PORT among patients with stage IIB disease, an additional subset analysis was performed. This multivariate analysis evaluated the impact of PORT in a subgroup of patients with marginnegative stage IIB thymoma. PORT remained associated with improved OS in this subgroup of 406 patients (HR ¼ 0.52, 95% CI: 0.28–0.96, p ¼ 0.037).
Discussion The optimal use of PORT for patients with thymic epithelial neoplasms is unclear, particularly for patients with Masaoka-Koga stage II disease. The 2016 National Comprehensive Cancer Network guidelines acknowledge this controversy and recommend that clinicians consider PORT for completely resected Masaoka-Koga stage II disease.28 Understandably, there is little current evidence on which to base any strong recommendation for PORT in this setting, and several smaller series have indicated no benefit for PORT in patients with stage II disease.29–31 Retrospective analyses of the Chinese Alliance for Research of Thymoma database have generally found PORT to
740 Jackson et al
Journal of Thoracic Oncology
Table 3. Thymoma: Multivariate Analysis of Predictors of Overall Survival
Vol. 12 No. 4
Table 3. Continued Multivariate
Multivariate Variable Local therapy Surgery alone Surgery plus PORT Age Continuous by year Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/ managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown
HR
95% CI
p Value
Variable
HR
95% CI
p Value
Histologic subtype Thymoma, NOS Type A Type AB Type B1 Type B2 Type B3
1 0.98 0.85 1.14 1.25 1.13
0.70–1.36 0.64–1.13 0.83–1.56 0.93–1.67 0.87–1.47
0.880 0.263 0.428 0.137 0.374
1 0.72
0.59–0.87
0.001
1.05
1.04–1.06
<0.001
1 1.09 1.14 0.87 1.44
0.70–1.69 0.87–1.48 0.56–1.34 0.67–3.08
0.711 0.345 0.525 0.349
Masaoka-Koga stage I–IIA IIB III IV Unknown
1 1.11 1.78 1.67 1.27
0.84–1.45 1.40–2.26 1.21–2.31 0.69–2.33
0.463 <0.001 0.002 0.445
1 0.99
0.83–1.18
0.894
Chemotherapy No Yes
1 1.35
1.06–1.71
0.014
1 1.05 1.21 1.31 1.70
0.58–1.90 0.77–1.88 1.02–1.69 0.22–2.24
0.881 0.411 0.034 0.552
0.94
0.47–1.90
0.867
1 0.92 0.79 0.70
0.69–1.22 0.60–1.05 0.53–0.93
0.558 0.104 0.014
1 0.81
0.57–1.14
0.223
0.75 0.64
0.53–1.07 0.40–1.04
0.114 0.071
1.19
0.67–2.11
0.551
1 0.87
0.64–1.17
0.359
1 1.37 1.47
1.11–1.70 1.02–2.11
0.004 0.038
1 0.82 0.75
0.66–1.00 0.56–0.99
0.059 0.043
1 1.67 1.22 2.68 1.29
1.25–2.23 0.94–1.58 1.86–3.86 0.96–1.75
0.001 0.136 <0.001 0.092 (continued)
HR, hazard ratio; CI, confidence interval; PORT, postoperative radiotherapy; NOS, not otherwise specified; R1, microscopic residual tumor; R2, macroscopic residual tumor.
improve OS for margin-negative patients with thymic carcinoma but not thymoma.32,33 Additionally, two previous SEER database studies lacking data on many important baseline covariates identified longer OS associated with PORT in stage III to IV disease but no corresponding improvement for patients with stage IIB disease.10,15 By contrast, our NCDB study presented here demonstrates a clear OS advantage for PORT in both stage IIB and stage III thymoma. This effect persisted on additional subgroup analysis excluding positive margins, indicating that stage IIB patients with completely resected thymomas benefit as well. A recent smaller NCDB analysis by Boothe et al. that was limited to stage II and III patients also demonstrated improved OS with PORT for patients with thymic epithelial neoplasms.34 Although there were differences in these two modern NCDB analyses owing to cohort definition and study design, both studies indicate a particular benefit of PORT for patients with positive margins and, to a lesser degree, for patients with WHO type A thymoma. Another recent, relatively well-powered study of 1263 patients from the International Thymic Malignancies Interest Group database also reported improved OS with PORT for patients with stage II and III thymoma.35 Although overall their conclusions align well with data from the NCDB, they report an increased magnitude of benefit for PORT among those patients with WHO type B1, B2, and B3 disease, which conflicts with our data and that of other investigators using the NCDB. Future analyses
April 2017
PORT for Thymoma and Thymic Carcinoma
Table 4. Thymic Carcinoma: Univariate Analysis of Predictors of Overall Survival
741
Table 4. Continued Univariate
Univariate Variable Local Therapy Surgery alone Surgery plus PORT Age Continuous by year Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/ managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown
HR
95% CI
p Value
1 0.79
0.64–0.97
0.025
1.01
1.01–1.02
0.001
1 1.42 1.14 0.75 1.14
0.92–2.19 0.85–1.54 0.48–1.16 0.37–3.56
0.119 0.391 0.195 0.822
1 0.88
0.71–1.09
0.240
1.72 1.44 1.44 0.76
1.01–2.92 0.91–2.30 1.16–1.80 0.19–3.06
0.045 0.123 0.001 0.696
0.90
0.40–2.04
0.803
1 0.97 0.89 0.94
0.69–1.35 0.64–1.25 0.69–1.28
0.838 0.502 0.715
1 1.27
0.79–2.04
0.332
1.10 0.91
0.68–1.78 0.48–1.71
0.696 0.760
1.06
0.60–1.87
0.840
1 0.93
0.68–1.27
0.660
1 1.26 1.76
0.98–1.63 1.20–2.59
0.070 0.004
1 1.05 1.10
0.82–1.35 0.82–1.48
0.698 0.524
1 1.79 1.70 2.65 1.20
1.32–2.44 1.29–2.23 1.83–3.85 0.83–1.75
<0.001 <0.001 <0.001 0.332 (continued)
1
Variable
HR
95% CI
p Value
Masaoka-Koga stage I–IIA IIB III IV Unknown
1 0.97 1.66 2.87 1.57
0.64–1.46 1.26–2.19 2.13–3.86 0.86–2.88
0.866 <0.001 <0.001 0.145
Chemotherapy No Yes
1 1.20
0.98–1.48
0.086
HR, hazard ratio; CI, confidence interval; PORT, postoperative radiotherapy; NOS, not otherwise specified; R1, microscopic residual tumor; R2, macroscopic residual tumor.
from data sets with robust local control and causespecific death information are warranted to further clarify these conflicting observations. There are a few possible explanations as to why the present NCDB analysis indicates longer OS with PORT for stage IIB disease despite the fact that many previous series have been unable to demonstrate this finding. The present study analyzed 3031 thymoma cases (633 stage IIB), making it a very large series of a rare disease, which may provide sufficient power to see a survival advantage with improved local therapy. Previous studies have at times demonstrated favorable HRs for PORT but perhaps lacked the statistical power to detect a significant difference. Similar in size to the present analysis, a recent, well-powered meta-analysis by Zhou et al. also demonstrated improved OS for patients with stage II and III thymoma.36 The NCDB also contains additional variables not available in the SEER database, such as margin status, consistent distinction of thymic carcinoma as a unique subtype, Charlson-Deyo comorbidity score, and chemotherapy use, which are all independently prognostic in the current study. Additionally, the time period analyzed may also help explain why longer OS with PORT was demonstrated in our modern NCDB analysis. Although previous studies have captured patients dating back as early as the 1970s,9,10,15 our analysis is limited to 2004–2012, which is well within the era of advanced three-dimensional conformal radiotherapy and intensity-modulated radiation therapy. Conformal dose distributions made possible by high-quality modern radiotherapy techniques have the potential to decrease dose to critical nearby thoracic organs and may improve the therapeutic ratio of PORT.37 Our inability to more specifically delineate Masaoka-Koga stage must be acknowledged as an important limitation of this analysis. Masaoka-Koga
742 Jackson et al
Journal of Thoracic Oncology
Table 5. Thymic Carcinoma: Multivariate Analysis of Predictors of Overall Survival
Vol. 12 No. 4
Table 5. Continued Multivariate
Multivariate Variable Local therapy Surgery alone Surgery plus PORT Age Continuous by year Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/ managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown
HR
95% CI
p Value
1 0.73
0.58–0.91
0.006
1.02
1.01–1.04
0.002
1 1.33 1.28 0.76 1.15
0.83–2.12 0.93–1.77 0.48–1.19 0.37–3.64
0.239 0.131 0.227 0.807
1 0.92
0.73–1.15
0.456
1.93 1.40 1.02 0.85
1.11–3.35 0.85–2.29 0.75–1.37 0.21–3.51
0.019 0.187 0.925 0.825
0.71
0.30–1.64
0.417
1 1.07 0.98 1.08
0.76–1.41 0.69–1.39 0.77–1.51
0.701 0.911 0.653
1 1.02
0.63–1.67
0.927
0.92 0.74
0.56–1.52 0.39–1.41
0.747 0.359
1.35
0.68–2.68
0.391
1 0.95
0.68–1.33
0.758
1 1.22 1.67
0.94–1.58 1.11–2.49
0.142 0.013
1 1.11 1.10
0.86–1.43 0.81–1.48
0.414 0.537
1 1.62 1.81 1.95 1.22
1.17–2.24 1.36–2.41 1.31–2.90 0.83–1.80
0.004 <0.001 0.001 0.321 (continued)
1
Variable
HR
95% CI
p Value
Masaoka-Koga stage I–IIA IIB III IV Unknown
1 0.99 1.50 2.70 1.16
0.65–1.51 1.12–2.00 1.95–3.74 0.62–2.19
0.957 0.006 <0.001 0.642
Chemotherapy No Yes
1 1.06
0.83–1.36
0.620
HR, hazard ratio; CI, confidence interval; PORT, postoperative radiotherapy; NOS, not otherwise specified; R1, microscopic residual tumor; R2, macroscopic residual tumor.
stage is not specifically recorded in either the NCDB or the SEER database, but was instead assigned on the basis of recorded information. Although the present investigation suggests an OS benefit for PORT in the patient with stage IIB thymoma, little can be gleaned about the relative benefit of PORT unique to patients with only microscopic transcapsular invasion (stage IIA). Of additional concern, chemotherapy appears negatively prognostic for patients with thymoma on multivariate analysis (HR ¼ 1.35, p ¼ 0.014), when in actuality its use may be a surrogate for advanced disease and residual confounding for which we cannot reliably control. This association was not observed for patients with thymic carcinoma, a cohort in which chemotherapy is more typically utilized, and other investigators have demonstrated improved outcomes when chemotherapy and radiation are delivered together as adjuvant therapy for locally advanced disease.38 The present analysis has several other important limitations. This study is retrospective, nonrandomized, and therefore subject to selection bias or the influence of unknown variables that cannot be controlled for in the NCDB. In addition, data regarding specific chemotherapy drugs and dosing are lacking. Finally, OS is the only outcome measure reported; information regarding other important oncologic outcomes such as local control and progression-free survival are unavailable. Efforts to account for potential selection biases included the use of multivariate models and propensity score–matched analysis with well-balanced patient characteristics.
Conclusions In this contemporary NCDB analysis, we present a large, contemporary evaluation of PORT for patients
April 2017
PORT for Thymoma and Thymic Carcinoma
743
Figure 2. Forest plot for patients with thymoma: impact of PORT on survival by histologic subtype, Masaoka-Koga stage, and surgical margin. PORT, postoperative radiotherapy; HR, hazard ratio.
with thymoma or thymic carcinoma. A significant OS advantage was associated with PORT for patients with derived Masaoka-Koga stage IIB or stage III disease; no significant OS benefit was observed for patients with early-stage disease (I-IIA). Substantial benefits from PORT were also observed among patients with positive surgical margins, although patients with intermediateand advanced-stage thymoma with negative margins also appeared to benefit from PORT. In the absence of randomized data assessing the value of PORT, these data may inform clinical practice.
References 1. Engels EA. Epidemiology of thymoma and associated malignancies. J Thorac Oncol. 2010;5:S260–S265. 2. Maggi G, Casadio C, Cavallo A, Cianci R, Molinatti M, Ruffini E. Thymoma: results of 241 operated cases. Ann Thorac Surg. 1991;51:152–156. 3. Blumberg D, Port JL, Weksler B, et al. Thymoma: a multivariate analysis of factors predicting survival. Ann Thorac Surg. 1995;60:908–914. 4. Curran W, Kornstein M, Brooks J, Turrisi A. Invasive thymoma: the role of mediastinal irradiation following complete or incomplete surgical resection. J Clin Oncol. 1988;6:1722–1727. 5. Häfner MF, Roeder F, Sterzing F, et al. Postoperative Strahlentherapie bei Patienten mit epithelialem Thymuskarzinom (TET). Strahlenther Onkol. 2015;191:133– 140 [in German]. 6. Fuller CD, Housman DM, Thomas CR. Radiotherapy for thymoma and thymic carcinoma. Hematol Oncol Clin North Am. 2008;22:489–507. 7. Lombe DC, Jeremic B. A Review of the place and role of radiotherapy in thymoma. Clin Lung Cancer. 2015;16:406–412. 8. Lim YJ, Kim E, Kim HJ, et al. Survival impact of adjuvant radiotherapy in Masaoka stage II-IV thymomas–a systematic review and meta-analysis. Int J Radiat Oncol Biol Phys. 2016;94:1129–1136.
9. Patel S, Macdonald OK, Nagda S, Bittner N, Suntharalingam M. Evaluation of the role of radiation therapy in the management of malignant thymoma. Int J Radiat Oncol Biol Phys. 2012;82:1797–1801. 10. Lim YJ, Kim HJ, Wu H-G. Role of postoperative radiotherapy in nonlocalized thymoma: propensity-matched analysis of Surveillance, Epidemiology, and End Results Database. J Thorac Oncol. 2015;10:1357–1363. 11. Weksler B, Shende M, Nason KS, Gallagher A, Ferson PF, Pennathur A. The role of adjuvant radiation therapy for resected stage III thymoma: a population-based study. Ann Thorac Surg. 2012;93:1822–1829. 12. Weksler B, Dhupar R, Parikh V, Nason KS, Pennathur A, Ferson PF. Thymic carcinoma: a multivariate analysis of factors predictive of survival in 290 patients. Ann Thorac Surg. 2013;95:299–303. 13. Forquer JA, Rong N, Fakiris AJ, Loehrer PJ, Johnstone PA. Postoperative radiotherapy after surgical resection of thymoma: differing roles in localized and regional disease. Int J Radiat Oncol Biol Phys. 2010;76:440–445. 14. Nakagawa K, Yokoi K, Nakajima J, et al. Is thymomectomy alone appropriate for stage I (T1N0M0) thymoma? Results of a propensity-score analysis. Ann Thorac Surgy. 2016;101:520–526. 15. Fernandes AT, Shinohara ET, Guo M, et al. The role of radiation therapy in malignant thymoma: a Surveillance, Epidemiology, and End Results database analysis. J Thorac Oncol. 2010;5:1454–1460. 16. Omasa M, Sozu T, Sato T, et al. Postoperative radiotherapy is effective for thymic carcinoma but not for thymoma in stage II and III thymic epithelial tumors: the Japanese Association for Research on the Thymus Database study. Cancer. 2015;121:1008–1016. 17. Mariano C, Ionescu DN, Cheung WY, et al. Thymoma: a population-based study of the management and outcomes for the province of British Columbia. J Thorac Oncol. 2013;8:109–117. 18. Ruffini E, Detterbeck F, Van Raemdonck D, et al. Thymic carcinoma: a cohort study of patients from the European Society of Thoracic Surgeons database. J Thorac Oncol. 2014;9:541–548.
744 Jackson et al 19. Tseng Y-H, Lin Y-H, Tseng Y-C, et al. Adjuvant therapy for thymic carcinoma–a decade of experience in a Taiwan National Teaching Hospital. PloS One. 2016;11: e0146609. 20. Ahmad U, Yao X, Detterbeck F, et al. Thymic carcinoma outcomes and prognosis: results of an international analysis. J Thorac Cardiovasc Surg. 2015;149:95–101. 21. Hishida T, Nomura S, Yano M, et al. Long-term outcome and prognostic factors of surgically treated thymic carcinoma: results of 306 cases from a Japanese Nationwide Database study. Eur J Cardiothorac Surg. 2016;49:835–841. 22. National Cancer Data Base. https://www.facs.org/quality %20programs/cancer/ncdb. Accessed September 15, 2015. 23. International Agency for Research on Cancer. International Classification of Diseases for Oncology including updates as at Sep 01 2011, approved by the IARC/WHO Committee for ICD-O-3. ICD-O-3 Online 2011; 3.1. http://codes.iarc.fr/ codegroup/2. Accessed September 15, 2015. 24. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619. 25. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials. 1996;17:343–346. 26. Coca-Perraillon M. Local and global optimal propensity score matching. SAS Global Forum 2007. http://www2.sas.com/proceedings/forum2007/1852007.pdf. Accessed October 15, 2015. 27. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10:150–161. 28. Ettinger D. NCCN guidelines version 2.2016. Thymomas and thymic carcinomas. http://www.nccn.org/professionals/ physician_gls/pdf/thymic.pdf. Accessed March 15, 2016. 29. Utsumi T, Shiono H, Kadota Y, et al. Postoperative radiation therapy after complete resection of thymoma has little impact on survival. Cancer. 2009;115: 5413–5420.
Journal of Thoracic Oncology
Vol. 12 No. 4
30. Korst RJ, Kansler AL, Christos PJ, Mandal S. Adjuvant radiotherapy for thymic epithelial tumors: a systematic review and meta-analysis. Ann Thorac Surg. 2009;87:1641–1647. 31. Rena O, Papalia E, Oliaro A, et al. Does adjuvant radiation therapy improve disease-free survival in completely resected Masaoka stage II thymoma? Eur J Cardiothorac Surg. 2007;31:109–113. 32. Fu H, Gu Z-T, Fang W-T, et al. Long-term survival after surgical treatment of thymic carcinoma: a retrospective analysis from the Chinese Alliance for Research of Thymoma database. Ann Surg Oncol. 2016;23:619–625. 33. Liu Q, Gu Z, Yang F, et al. The role of postoperative radiotherapy for stage I/II/III thymic tumor—results of the ChART retrospective database. J Thorac Dis. 2016;8:687. 34. Boothe D, Orton A, Thorpe C, Kokeny K, Hitchcock YJ. Postoperative radiotherapy in locally invasive malignancies of the thymus: patterns of care and survival. J Thorac Oncol. 2016;11:2218–2226. 35. Rimner A, Yao X, Huang J, et al. Postoperative radiation therapy is associated with longer overall survival in completely resected stage II and III thymoma—an analysis of the International Thymic Malignancies Interest Group retrospective database. J Thorac Oncol. 2016;11:1785–1792. 36. Zhou D, Deng X-F, Liu Q-X, Zheng H, Min J-X, Dai J-G. The effectiveness of postoperative radiotherapy in patients with completely resected thymoma: a meta-analysis. Ann Thorac Surg. 2016;101:305–310. 37. Bucci MK, Bevan A, Roach M. Advances in radiation therapy: conventional to 3D, to IMRT, to 4D, and beyond. CA: Cancer J Clin. 2005;55:117–134. 38. Leuzzi G, Rocco G, Ruffini E, et al. Multimodality therapy for locally advanced thymomas: a propensity score– matched cohort study from the European Society of Thoracic Surgeons Database. J Thorac Cardiovasc Surg. 2016;151:47–57.
The Impact of Postoperative Radiotherapy for Thymoma and Thymic Carcinoma Matthew W. Jackson, MD,a,* David A. Palma, MD, MsC, PhD,b D. Ross Camidge, MD, PhD,c Bernard L. Jones, PhD,a Tyler P. Robin, MD, PhD,a David J. Sher, MD, MPH,d Matthew Koshy, MD,e Brian D. Kavanagh, MD, MPH,a Laurie E. Gaspar, MD, MBA,a Chad G. Rusthoven, MDa a
Department of Radiation Oncology, University of Colorado Cancer Center, Aurora, Colorado London Health Sciences Centre, London, Ontario, Canada c Division of Medical Oncology, University of Colorado Cancer Center, Aurora, Colorado d Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, Texas e Department of Radiation Oncology and Cellular Oncology, Department of Radiation Oncology, University of Illinois at Chicago, Chicago, Illinois b
Received 7 April 2016; revised 4 December 2016; accepted 1 January 2017 Available online - 25 January 2017
ABSTRACT Introduction: The optimal role for postoperative radiotherapy (PORT) for thymoma and thymic carcinoma remains controversial. We used the National Cancer Data Base to investigate the impact of PORT on overall survival (OS). Methods: Patients who underwent an operation for thymoma or thymic carcinoma were categorized into Masaoka-Koga stage groups I to IIA, IIB, III, and IV. Patients who did not undergo an operation or those who received preoperative radiation were excluded. Kaplan-Meier estimates of OS and univariate and multivariate Cox proportional hazards regression analyses were performed. Propensity score–matched analyses were performed to further control for baseline confounders. Results: From 2004 to 2012, 4056 patients were eligible for inclusion, 2001 of whom (49%) received PORT. On multivariate analysis of OS in the thymoma cohort adjusted for age, WHO histologic subtype, Masaoka-Koga stage group, surgical margins, and chemotherapy administration, PORT was associated with superior OS (hazard ratio [HR] ¼ 0.72, p ¼ 0.001). Propensity score–matched analyses confirmed the survival advantage associated with PORT. Subset analysis indicated longer OS in association with PORT for patients with stage IIB thymoma (HR ¼ 0.61, p ¼ 0.035), stage III (HR ¼ 0.69, p ¼ 0.020), and positive margins (HR ¼ 0.53, p < 0.001). The impact of PORT for stage I to IIA disease did not reach significance (HR ¼ 0.76, p ¼ 0.156). Conclusions: In this large database analysis of PORT for thymic tumors, PORT was associated with longer OS, with the greatest relative benefits observed for stage IIB to III disease and positive margins. In the absence of randomized
Journal of Thoracic Oncology
Vol. 12 No. 4: 734-744
studies assessing the value of PORT, these data may inform clinical practice. 2017 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. Keywords: Thymoma; Thymic carcinoma; Adjuvant radiation; Postoperative radiotherapy; National Cancer Data Base
Introduction The thymic epithelial tumors, thymoma and thymic carcinoma, represent the most frequent tumors of the anterior mediastinum, although they remain relatively rare overall.1 Surgery stands as the primary therapeutic modality in the curative setting, as extent of resection has consistently been shown to be independently prognostic of improved survival.2,3 The intimate association of these tumors with critical mediastinal structures can make complete resection difficult, particularly for advanced-stage disease.4,5 As a result, there has been a longstanding interest in utilizing
*Corresponding author. Disclosure: The authors declare no conflict of interest. Address for correspondence: Matthew W. Jackson, MD, Department of Radiation Oncology, University of Colorado Cancer Center, 1665 Aurora Court, Suite 1032 MS F706, Aurora, CO 80045. E-mail: matthew. [email protected] ª 2017 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. ISSN: 1556-0864 http://dx.doi.org/10.1016/j.jtho.2017.01.002
April 2017
postoperative radiotherapy (PORT) to improve outcomes. However, the precise indications for PORT and its magnitude of benefit have been difficult to ascertain.6–8 The low annual incidence of thymic tumors renders the execution of randomized trials challenging, leaving only limited, conflicting data regarding the utility of PORT. Large registry data sets, such as the Surveillance Epidemiology and End Results (SEER) database, have proved useful for investigating clinical questions in thymic tumors and other rare diseases for which there are inadequate data to address relevant questions. Given a lack of prospective evidence, several investigators have previously utilized the SEER database and other national data sets to investigate the role of PORT for thymic epithelial tumors.9–17 Although these studies have been fairly heterogeneous in their design, consistent themes have emerged. PORT appears beneficial for patients with advanced disease (Masaoka-Koga stage III–IV)9,10 but of minimal utility for completely encapsulated tumors (Masaoka-Koga stage I),13–15 and it remains controversial for Masaoka-Koga stage II disease.15,16 Improved outcomes with the addition of PORT have been more readily demonstrated for patients with aggressive thymic carcinoma, as opposed to for those with thymoma histologic subtypes.18–21 Many retrospective, large-database studies have been unable to control for important confounders, including surgical margin status, histologic subtype, and use of chemotherapy; these studies have called for future studies to incorporate the aforesaid variables to provide more robust analyses. The National Cancer Data Base (NCDB), which was queried for this study, includes additional information on these important variables. As a large hospital-based registry representing more than two-thirds of all cancer cases in the United States, the NCDB draws from a substantial number of institutions nationwide, thereby increasing the number of cases available for analysis. In this largest reported series on PORT for thymic tumors, we utilize the NCDB, with its unprecedented body of critical patient- and disease-specific factors, to describe outcomes and prognostic factors that identify patients who may benefit from adjuvant radiotherapy.
Methods Data Source and Patient Selection The NCDB is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. It is a hospital-based registry that represents 70% of all cancer cases in the United States, drawing data from more than 1500
PORT for Thymoma and Thymic Carcinoma
735
commission-accredited cancer programs. The data used in the study are derived from a deidentified NCDB 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 for the conclusions drawn from these data by the investigators.22 Before the present analysis was begun, appropriate exemption was obtained from our institutional review board. The analysis was restricted to adult patients (age 18 years) in whom thymoma or thymic carcinoma, defined as International Classification of Disease for Oncology, Third Edition (ICD-O-3), histologic codes for thymoma (8580–8585) and thymic carcinoma (8002, 8010, 8012, 8013, 8020, 8021, 8070-8072, 8074, 8140, 8240, 8243, 8246, 8586, 8588, and 8589) with corresponding topographic codes C37.9 (thymus) and C38.1 (anterior mediastinum), had been diagnosed.23 All patients included in the analysis were recorded as having undergone an operation and had their disease diagnosed from 2004 to 2012. All included cases had known treatment-related variables with respect to surgery, chemotherapy, and radiotherapy delivery, with unknowns excluded from this analysis. Treatment data coded in the NCDB are limited to the first course of treatment, defined as all methods of treatment recorded in the treatment plan and administered to the patient before disease progression or recurrence. Cases for which radiation was coded as being delivered before surgery or intraoperatively were excluded. Cases in which death occurred within 1 month of diagnosis were omitted from analysis.
Patient Demographics and Treatment Variables Potentially relevant patient and treatment characteristics were selected a priori. Retrievable data included age, race, sex, insurance status, median income, treatment facility type, distance to facility, Charlson-Deyo comorbidity score, year of diagnosis, surgical margin status, WHO histologic subtype, use of radiation, use of chemotherapy, and derived MasaokaKoga stage group. To facilitate clinical applicability of this report, cases within the database were categorized according to the Masaoka-Koga staging system for thymic epithelial tumors: stage I to IIA (invasive tumor confined to gland of origin or localized, NOS), stage IIB (adjacent connective tissue), stage III (adjacent organs/structures in mediastinum), and stage IV (further contiguous extension, lymph nodes, or metastasis). As neither the SEER database nor the NCDB codes for microscopic transcapsular invasion, stage I disease could not be distinguished from stage IIA disease.
736 Jackson et al
Journal of Thoracic Oncology
Vol. 12 No. 4
Table 1. Patient- and Disease-Specific Factors by Treatment Group Thymoma
Characteristic Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown Masaoka stage I–IIA IIB III IV Unknown
Thymic Carcinoma
Surgery Alone
Surgery Plus PORT
No.
%
No.
%
1100 73 259 134 21
(69.3) (4.6) (16.3) (8.4) (1.3)
1014 61 241 113 15
(70.2) (4.2) (16.7) (7.8) (1.0)
729 858
(45.9) (54.1)
706 738
(48.9) (51.1)
785 65 93 589 9 46
(49.5) (4.1) (5.9) (37.1) (0.6) (2.9)
831 39 89 438 21 26
(57.5) (2.7) (6.2) (30.3) (1.5) (1.8)
236 314 416 604
(15.9) (19.8) (26.2) (38.1)
222 307 403 512
(15.4) (21.3) (27.9) (35.5)
89 570 707 68 153
(5.6) (35.9) (44.5) (4.3) (9.6)
98 552 551 89 154
(6.8) (38.2) (38.2) (6.2) (10.7)
1346 241
(84.8) (15.2)
1322 122
(91.6) (8.4)
1202 306 79
(75.7) (19.3) (5.0)
1132 255 57
(78.4) (17.7) (3.9)
423 500 664
(26.7) (31.5) (41.8)
475 502 467
(32.9) (34.8) (32.3)
1147 103 145 32 160
(72.3) (6.5) (9.1) (2.0) (10.1)
733 188 331 60 132
(50.8) (13.0) (22.9) (4.2) (9.1)
813 274 317 136 47
(51.2) (17.3) (20.0) (8.6) (3.0)
431 359 451 165 38
(29.8) (24.9) (31.2) (11.4) (2.6)
p Value
Surgery Alone
Surgery Plus PORT
No.
%
No.
%
339 30 61 34 4
(72.4) (6.4) (13.0) (7.3) (0.9)
398 26 83 47 3
(71.5) (4.7) (14.9) (8.4) (0.5)
296 172
(63.2) (36.8)
346 211
(62.1) (37.9)
216 15 22 200 5 10
(46.2) (3.2) (4.7) (42.7) (1.1) (2.1)
305 17 35 187 5 8
(54.8) (3.1) (6.3) (33.6) (0.9) (1.4)
74 110 123 161
(15.8) (23.5) (26.3) (34.4)
86 127 129 215
(15.4) (22.8) (23.2) (38.6)
20 172 190 34 52
(4.3) (36.8) (40.6) (7.3) (11.1)
38 229 208 39 43
(6.8) (41.1) (37.3) (7.0) (7.7)
390 78
(83.3) (16.7)
504 53
(90.5) (9.5)
334 98 36
(71.4) (20.9) (7.7)
424 105 28
(76.1) (18.9) (5.0)
144 155 169
(30.8) (33.1) (36.1)
159 180 218
(28.5) (32.3) (39.1)
287 48 60 25 48
(61.3) (10.3) (12.8) (5.3) (10.3)
263 87 121 32 54
(47.2) (15.6) (21.7) (5.7) (9.7)
174 48 140 85 21
(37.2) (10.3) (29.9) (18.2) (4.5)
154 85 211 95 12
(27.6) (15.3) (37.9) (17.1) (2.2)
0.868
p Value 0.576
0.104
0.710
<0.001
0.046
0.369
0.511
0.003
0.090
<0.001
0.001
0.169
0.121
<0.001
0.579
<0.001
<0.001
<0.001
<0.001
(continued)
April 2017
PORT for Thymoma and Thymic Carcinoma
737
Table 1. Continued Thymoma
Thymic Carcinoma
Surgery Alone
Surgery Plus PORT
Characteristic
No.
%
No.
%
Chemotherapy No Yes
1361 226
(85.8) (14.2)
1131 313
(78.3) (21.7)
p Value
Surgery Alone
Surgery Plus PORT
No.
%
No.
%
337 131
(72.0) (28.0)
247 310
(44.3) (55.7)
<0.001
p Value <0.001
Note: Of the patients with thymoma, 1587 (median age 60 years, age range 18–90 years) were treated with surgery alone and 1444 (median age 59 years, age range 18–90 years) were treated with surgery plus PORT. Of the patients with thymic carcinoma, 468 (median age 63 years, age range 19–89 years) were treated with surgery alone and 557 (median age 60 years, age range 20–90 years) were treated with surgery plus PORT. PORT, postoperative radiotherapy; NOS, not otherwise specified.
Median household income in the patient’s home zip code was recorded in quartiles relative to the U.S. population. Distance to treatment facility was recorded as the distance in miles between the hospital that reported the case and the patient’s zip code centroid or city if the zip code was not available. Patient comorbidities were categorized by comorbidity score as described by Charlson and Deyo (1992), and cases were coded as 0, 1, or 2 or higher.24
replacement was performed by using the caliper match algorithm described by Coca-Perraillon,26 with the caliper width set to 0.05 times the standard deviation of the logit of the propensity score.27 These outcomes were assessed with a log-rank test, and the hazard ratio (HR) was determined by univariate Cox regression. The score calculation was blinded with respect to patient outcomes.
Results Statistical Analyses
Patient Characteristics
All statistical analyses were performed using SPSS, version 23.0 (SPSS, Inc., Chicago, IL). Median follow-up was calculated with the reverse Kaplan-Meier method described by Schemper and Smith.25 Pearson chi-square tests were used to evaluate patient- and treatmentrelated factors between patients who did or did not receive PORT. The primary end point was overall survival (OS) which is defined in the NCDB as date of diagnosis to the date of death. OS estimates were generated by using the Kaplan-Meier method and comparisons were made by using the log-rank test. Univariate hazard estimates were generated with unadjusted Cox proportional hazards models. Multivariate analysis for OS were performed by using Cox models adjusted for all variables selected a priori: age, race, sex, insurance status, median income, treatment facility type, distance to facility, Charlson-Deyo comorbidity score, year of diagnosis, surgical margin status, WHO histologic subtype, use of radiation, use of chemotherapy, and derived Masaoka-Koga stage group. To further account for confounding variables, we performed propensity score matching for patients treated with or without PORT, accounting for the same variables described in the preceding paragraph. The propensity score was calculated by using logistic regression to estimate the probability of receiving radiation. One-to-one propensity matching without
A total of 4056 patients with thymoma (n ¼ 3031) or thymic carcinoma (n ¼ 1025) who met predefined selection criteria were identified. Median follow-up was 57.2 months (range 1.08–129.15 months) for patients with thymoma and 59.5 months (range 1.15–130.23) for patients with thymic carcinoma. The median age for patients with thymoma was 60 years (range 18–90 and older) and that for patients with thymic carcinoma was 61 years (range 19–90 and older). Use of adjuvant radiotherapy was evenly split among the cohort, as 1444 patients with thymoma (47.6%) and 557 patients with thymic carcinoma (54.3%) received PORT. The interquartile range for time from surgery to the start of radiation for all patients was 41 to 77 days. The overwhelming majority of patients presented with localized disease, with a higher proportion of thymic carcinomas (17.5%) being Masaoka-Koga stage IV at diagnosis. Patient- and disease-specific characteristics for both patients with thymoma and thymic carcinoma as well as for the surgery-alone and PORT treatment groups are listed in Table 1. Differences were identified between the PORT and surgery-alone groups, with a general preponderance of adverse risk factors among those receiving PORT. Disease factors associated with the use of PORT included positive margins and advanced stage. Demographic factors associated with receipt of PORT included private medical
738 Jackson et al
Journal of Thoracic Oncology
Table 2. Thymoma: Univariate Analysis of Predictors of Overall Survival
Vol. 12 No. 4
Table 2. Continued Univariate
Univariate Variable Local Therapy Surgery alone Surgery þ PORT Age Continuous by year Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/ managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown
HR
95% CI
p Value
Variable
HR
95% CI
p Value
Histologic subtype Thymoma, NOS Type A Type AB Type B1 Type B2 Type B3 Thymic carcinoma
1 1.15 0.79 0.89 1.04 1.12
0.83–1.59 0.60–1.04 0.65–1.21 0.78–1.38 0.86–1.45
0.397 0.090 0.462 0.788 0.396
1 0.80
0.67–0.96
0.015
1.05
1.04–1.05
<0.001
1 1.16 1.14 0.62 1.16
0.76–1.76 0.90–1.44 0.41–0.94 0.55–2.46
0.500 0.283 0.025 0.695
1 1.01
Masaoka-Koga stage I–IIA IIB III IV Unknown
1 1.18 2.09 2.15 1.15
0.90–1.54 1.67–2.60 1.61–2.86 0.64–2.08
0.224 <0.001 <0.001 0.638
0.85–1.21
0.880
Chemotherapy No Yes
1 1.32
1.07–1.62
0.010
1 1.11 1.33 2.70 1.06
0.62–1.99 0.87–2.03 2.24–3.26 0.34–3.31
0.731 0.193 <0.001 0.924
1.10
0.56–2.15
0.783
1 0.89 0.79 0.65
0.68–1.17 0.61–1.03 0.50–0.84
0.416 0.079 0.001
1 0.90
0.64–1.26
0.533
0.75 0.87
0.54–1.06 0.54–1.38
0.100 0.547
0.38
0.23–0.62
<0.001
1 0.99
0.75–1.31
0.927
1 1.62 2.08
1.31–2.00 1.47–2.95
<0.001 <0.001
1 0.89 0.78
0.72–1.09 0.59–1.03
0.248 0.076
1 1.74 1.28 3.58 1.44
1.32–2.30 1.00–1.64 2.58–4.97 1.08–1.93
<0.001 0.049 <0.001 0.012 (continued)
HR, hazard ratio; CI, confidence interval; PORT, postoperative radiotherapy; NOS, not otherwise specified; R1, microscopic residual tumor; R2, macroscopic residual tumor.
insurance and residence within 50 miles of the treating facility.
OS for Patients with Thymoma On univariate analysis of the entire thymoma cohort, despite the predominance of adverse disease-specific factors in the PORT group, PORT was associated with improved OS (HR ¼ 0.80, 95% confidence interval [CI]: 0.67–0.96, p ¼ 0.015). This includes all patients with thymoma irrespective of margin status. Univariate survival analyses for all variables are displayed in Table 2. Kaplan-Meier survival curves for stage-specific subgroups are displayed in the top row of Figure (Fig. 1A–C). On multivariate analysis, PORT remained independently associated with improved OS (HR ¼ 0.72, 95% CI: 0.59–0.87, p ¼ 0.001). A high median income was also associated with improved OS. The analysis showed an overall trend toward better outcomes in the latter years of the study, which reached significance for the years 2010–2012. Variables independently associated with inferior OS on multivariate analysis include older age, increased comorbidity score, positive margins, advanced Masaoka-Koga stage (III–IV), and use of chemotherapy (Table 3). Propensity score–matched treatment groups (n ¼ 1015 per group) were well balanced for all patient- and treatment-related factors between the surgery-alone and PORT groups. For the propensity-matched thymoma cohort, OS in patients receiving PORT was superior to that
April 2017
PORT for Thymoma and Thymic Carcinoma
739
Figure 1. (Top row) Kaplan-Meier curves depicting overall survival for patients with thymoma: stage I to IIA (A), stage IIB (B), and stage III (C). (Bottom row) Kaplan-Meier curves depicting overall survival for patients with thymic carcinoma: stage I to IIA (D), stage IIB (E), and stage III (F). Abbreviation: PORT, postoperative radiotherapy.
in patients receiving an operation alone (HR ¼ 0.64, CI: 0.52–0.79, p < 0.001).
OS for Patients with Thymic Carcinoma On univariate analysis of the entire thymic carcinoma cohort, PORT was associated with improved OS (HR ¼ 0.79, 95% CI: 0.64–0.97, p ¼ 0.025). Univariate survival analyses for all variables are displayed in Table 4. Kaplan-Meier survival curves for stage-specific subgroups are displayed in the bottom row of Figure 1 (Fig. 1D–F). This includes all patients with thymic carcinoma irrespective of margin status. On multivariate analysis, PORT remained independently associated with improved OS (HR ¼ 0.73, 95% CI: 0.58–0.91, p ¼ 0.006). Variables independently associated with inferior OS on multivariate analysis include older age, uninsured status, positive margins, and advanced Masaoka-Koga stage (Table 5). Propensity score–matched treatment groups (n ¼ 345 per group) were well balanced for all patient- and treatment-related factors between the surgery-alone and PORT groups. For the propensity-matched thymoma cohort, OS in patients receiving PORT was superior to that in patients receiving an operation alone (HR ¼ 0.73, 95% CI: 0.57–0.94, p ¼ 0.016).
Secondary and Subgroup Analyses Subgroup analysis for patients with thymoma was performed to evaluate the impact of PORT by histologic
subtype, Masaoka-Koga stage group, and surgical margin status, with the corresponding forest plots depicted in Figure 2. Utilization of PORT was associated with significantly longer survival in patients with the following characteristics: histologic subtypes A and B, stage IIB or III disease, and positive margins. To help exclude the possibility that positive margins were influencing the benefit of PORT among patients with stage IIB disease, an additional subset analysis was performed. This multivariate analysis evaluated the impact of PORT in a subgroup of patients with marginnegative stage IIB thymoma. PORT remained associated with improved OS in this subgroup of 406 patients (HR ¼ 0.52, 95% CI: 0.28–0.96, p ¼ 0.037).
Discussion The optimal use of PORT for patients with thymic epithelial neoplasms is unclear, particularly for patients with Masaoka-Koga stage II disease. The 2016 National Comprehensive Cancer Network guidelines acknowledge this controversy and recommend that clinicians consider PORT for completely resected Masaoka-Koga stage II disease.28 Understandably, there is little current evidence on which to base any strong recommendation for PORT in this setting, and several smaller series have indicated no benefit for PORT in patients with stage II disease.29–31 Retrospective analyses of the Chinese Alliance for Research of Thymoma database have generally found PORT to
740 Jackson et al
Journal of Thoracic Oncology
Table 3. Thymoma: Multivariate Analysis of Predictors of Overall Survival
Vol. 12 No. 4
Table 3. Continued Multivariate
Multivariate Variable Local therapy Surgery alone Surgery plus PORT Age Continuous by year Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/ managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown
HR
95% CI
p Value
Variable
HR
95% CI
p Value
Histologic subtype Thymoma, NOS Type A Type AB Type B1 Type B2 Type B3
1 0.98 0.85 1.14 1.25 1.13
0.70–1.36 0.64–1.13 0.83–1.56 0.93–1.67 0.87–1.47
0.880 0.263 0.428 0.137 0.374
1 0.72
0.59–0.87
0.001
1.05
1.04–1.06
<0.001
1 1.09 1.14 0.87 1.44
0.70–1.69 0.87–1.48 0.56–1.34 0.67–3.08
0.711 0.345 0.525 0.349
Masaoka-Koga stage I–IIA IIB III IV Unknown
1 1.11 1.78 1.67 1.27
0.84–1.45 1.40–2.26 1.21–2.31 0.69–2.33
0.463 <0.001 0.002 0.445
1 0.99
0.83–1.18
0.894
Chemotherapy No Yes
1 1.35
1.06–1.71
0.014
1 1.05 1.21 1.31 1.70
0.58–1.90 0.77–1.88 1.02–1.69 0.22–2.24
0.881 0.411 0.034 0.552
0.94
0.47–1.90
0.867
1 0.92 0.79 0.70
0.69–1.22 0.60–1.05 0.53–0.93
0.558 0.104 0.014
1 0.81
0.57–1.14
0.223
0.75 0.64
0.53–1.07 0.40–1.04
0.114 0.071
1.19
0.67–2.11
0.551
1 0.87
0.64–1.17
0.359
1 1.37 1.47
1.11–1.70 1.02–2.11
0.004 0.038
1 0.82 0.75
0.66–1.00 0.56–0.99
0.059 0.043
1 1.67 1.22 2.68 1.29
1.25–2.23 0.94–1.58 1.86–3.86 0.96–1.75
0.001 0.136 <0.001 0.092 (continued)
HR, hazard ratio; CI, confidence interval; PORT, postoperative radiotherapy; NOS, not otherwise specified; R1, microscopic residual tumor; R2, macroscopic residual tumor.
improve OS for margin-negative patients with thymic carcinoma but not thymoma.32,33 Additionally, two previous SEER database studies lacking data on many important baseline covariates identified longer OS associated with PORT in stage III to IV disease but no corresponding improvement for patients with stage IIB disease.10,15 By contrast, our NCDB study presented here demonstrates a clear OS advantage for PORT in both stage IIB and stage III thymoma. This effect persisted on additional subgroup analysis excluding positive margins, indicating that stage IIB patients with completely resected thymomas benefit as well. A recent smaller NCDB analysis by Boothe et al. that was limited to stage II and III patients also demonstrated improved OS with PORT for patients with thymic epithelial neoplasms.34 Although there were differences in these two modern NCDB analyses owing to cohort definition and study design, both studies indicate a particular benefit of PORT for patients with positive margins and, to a lesser degree, for patients with WHO type A thymoma. Another recent, relatively well-powered study of 1263 patients from the International Thymic Malignancies Interest Group database also reported improved OS with PORT for patients with stage II and III thymoma.35 Although overall their conclusions align well with data from the NCDB, they report an increased magnitude of benefit for PORT among those patients with WHO type B1, B2, and B3 disease, which conflicts with our data and that of other investigators using the NCDB. Future analyses
April 2017
PORT for Thymoma and Thymic Carcinoma
Table 4. Thymic Carcinoma: Univariate Analysis of Predictors of Overall Survival
741
Table 4. Continued Univariate
Univariate Variable Local Therapy Surgery alone Surgery plus PORT Age Continuous by year Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/ managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown
HR
95% CI
p Value
1 0.79
0.64–0.97
0.025
1.01
1.01–1.02
0.001
1 1.42 1.14 0.75 1.14
0.92–2.19 0.85–1.54 0.48–1.16 0.37–3.56
0.119 0.391 0.195 0.822
1 0.88
0.71–1.09
0.240
1.72 1.44 1.44 0.76
1.01–2.92 0.91–2.30 1.16–1.80 0.19–3.06
0.045 0.123 0.001 0.696
0.90
0.40–2.04
0.803
1 0.97 0.89 0.94
0.69–1.35 0.64–1.25 0.69–1.28
0.838 0.502 0.715
1 1.27
0.79–2.04
0.332
1.10 0.91
0.68–1.78 0.48–1.71
0.696 0.760
1.06
0.60–1.87
0.840
1 0.93
0.68–1.27
0.660
1 1.26 1.76
0.98–1.63 1.20–2.59
0.070 0.004
1 1.05 1.10
0.82–1.35 0.82–1.48
0.698 0.524
1 1.79 1.70 2.65 1.20
1.32–2.44 1.29–2.23 1.83–3.85 0.83–1.75
<0.001 <0.001 <0.001 0.332 (continued)
1
Variable
HR
95% CI
p Value
Masaoka-Koga stage I–IIA IIB III IV Unknown
1 0.97 1.66 2.87 1.57
0.64–1.46 1.26–2.19 2.13–3.86 0.86–2.88
0.866 <0.001 <0.001 0.145
Chemotherapy No Yes
1 1.20
0.98–1.48
0.086
HR, hazard ratio; CI, confidence interval; PORT, postoperative radiotherapy; NOS, not otherwise specified; R1, microscopic residual tumor; R2, macroscopic residual tumor.
from data sets with robust local control and causespecific death information are warranted to further clarify these conflicting observations. There are a few possible explanations as to why the present NCDB analysis indicates longer OS with PORT for stage IIB disease despite the fact that many previous series have been unable to demonstrate this finding. The present study analyzed 3031 thymoma cases (633 stage IIB), making it a very large series of a rare disease, which may provide sufficient power to see a survival advantage with improved local therapy. Previous studies have at times demonstrated favorable HRs for PORT but perhaps lacked the statistical power to detect a significant difference. Similar in size to the present analysis, a recent, well-powered meta-analysis by Zhou et al. also demonstrated improved OS for patients with stage II and III thymoma.36 The NCDB also contains additional variables not available in the SEER database, such as margin status, consistent distinction of thymic carcinoma as a unique subtype, Charlson-Deyo comorbidity score, and chemotherapy use, which are all independently prognostic in the current study. Additionally, the time period analyzed may also help explain why longer OS with PORT was demonstrated in our modern NCDB analysis. Although previous studies have captured patients dating back as early as the 1970s,9,10,15 our analysis is limited to 2004–2012, which is well within the era of advanced three-dimensional conformal radiotherapy and intensity-modulated radiation therapy. Conformal dose distributions made possible by high-quality modern radiotherapy techniques have the potential to decrease dose to critical nearby thoracic organs and may improve the therapeutic ratio of PORT.37 Our inability to more specifically delineate Masaoka-Koga stage must be acknowledged as an important limitation of this analysis. Masaoka-Koga
742 Jackson et al
Journal of Thoracic Oncology
Table 5. Thymic Carcinoma: Multivariate Analysis of Predictors of Overall Survival
Vol. 12 No. 4
Table 5. Continued Multivariate
Multivariate Variable Local therapy Surgery alone Surgery plus PORT Age Continuous by year Race White, non-Hispanic White, Hispanic African American Others Unknown Sex Male Female Insurance status Private insurance/ managed care Not insured Medicaid Medicare Other government insurance Unknown Median income <$30,000 $30,000–$34,999 $35,000–$45,999 $46,000 Facility type Community cancer program Comprehensive community cancer program Academic/research Integrated network cancer program NOS Distance to facility, miles <50 50 Charlson-Deyo comorbidity score 0 1 2 Year of diagnosis 2004–2006 2007–2009 2010–2012 Surgical margin status Negative, R0 Positive, NOS R1 R2 Unknown
HR
95% CI
p Value
1 0.73
0.58–0.91
0.006
1.02
1.01–1.04
0.002
1 1.33 1.28 0.76 1.15
0.83–2.12 0.93–1.77 0.48–1.19 0.37–3.64
0.239 0.131 0.227 0.807
1 0.92
0.73–1.15
0.456
1.93 1.40 1.02 0.85
1.11–3.35 0.85–2.29 0.75–1.37 0.21–3.51
0.019 0.187 0.925 0.825
0.71
0.30–1.64
0.417
1 1.07 0.98 1.08
0.76–1.41 0.69–1.39 0.77–1.51
0.701 0.911 0.653
1 1.02
0.63–1.67
0.927
0.92 0.74
0.56–1.52 0.39–1.41
0.747 0.359
1.35
0.68–2.68
0.391
1 0.95
0.68–1.33
0.758
1 1.22 1.67
0.94–1.58 1.11–2.49
0.142 0.013
1 1.11 1.10
0.86–1.43 0.81–1.48
0.414 0.537
1 1.62 1.81 1.95 1.22
1.17–2.24 1.36–2.41 1.31–2.90 0.83–1.80
0.004 <0.001 0.001 0.321 (continued)
1
Variable
HR
95% CI
p Value
Masaoka-Koga stage I–IIA IIB III IV Unknown
1 0.99 1.50 2.70 1.16
0.65–1.51 1.12–2.00 1.95–3.74 0.62–2.19
0.957 0.006 <0.001 0.642
Chemotherapy No Yes
1 1.06
0.83–1.36
0.620
HR, hazard ratio; CI, confidence interval; PORT, postoperative radiotherapy; NOS, not otherwise specified; R1, microscopic residual tumor; R2, macroscopic residual tumor.
stage is not specifically recorded in either the NCDB or the SEER database, but was instead assigned on the basis of recorded information. Although the present investigation suggests an OS benefit for PORT in the patient with stage IIB thymoma, little can be gleaned about the relative benefit of PORT unique to patients with only microscopic transcapsular invasion (stage IIA). Of additional concern, chemotherapy appears negatively prognostic for patients with thymoma on multivariate analysis (HR ¼ 1.35, p ¼ 0.014), when in actuality its use may be a surrogate for advanced disease and residual confounding for which we cannot reliably control. This association was not observed for patients with thymic carcinoma, a cohort in which chemotherapy is more typically utilized, and other investigators have demonstrated improved outcomes when chemotherapy and radiation are delivered together as adjuvant therapy for locally advanced disease.38 The present analysis has several other important limitations. This study is retrospective, nonrandomized, and therefore subject to selection bias or the influence of unknown variables that cannot be controlled for in the NCDB. In addition, data regarding specific chemotherapy drugs and dosing are lacking. Finally, OS is the only outcome measure reported; information regarding other important oncologic outcomes such as local control and progression-free survival are unavailable. Efforts to account for potential selection biases included the use of multivariate models and propensity score–matched analysis with well-balanced patient characteristics.
Conclusions In this contemporary NCDB analysis, we present a large, contemporary evaluation of PORT for patients
April 2017
PORT for Thymoma and Thymic Carcinoma
743
Figure 2. Forest plot for patients with thymoma: impact of PORT on survival by histologic subtype, Masaoka-Koga stage, and surgical margin. PORT, postoperative radiotherapy; HR, hazard ratio.
with thymoma or thymic carcinoma. A significant OS advantage was associated with PORT for patients with derived Masaoka-Koga stage IIB or stage III disease; no significant OS benefit was observed for patients with early-stage disease (I-IIA). Substantial benefits from PORT were also observed among patients with positive surgical margins, although patients with intermediateand advanced-stage thymoma with negative margins also appeared to benefit from PORT. In the absence of randomized data assessing the value of PORT, these data may inform clinical practice.
References 1. Engels EA. Epidemiology of thymoma and associated malignancies. J Thorac Oncol. 2010;5:S260–S265. 2. Maggi G, Casadio C, Cavallo A, Cianci R, Molinatti M, Ruffini E. Thymoma: results of 241 operated cases. Ann Thorac Surg. 1991;51:152–156. 3. Blumberg D, Port JL, Weksler B, et al. Thymoma: a multivariate analysis of factors predicting survival. Ann Thorac Surg. 1995;60:908–914. 4. Curran W, Kornstein M, Brooks J, Turrisi A. Invasive thymoma: the role of mediastinal irradiation following complete or incomplete surgical resection. J Clin Oncol. 1988;6:1722–1727. 5. Häfner MF, Roeder F, Sterzing F, et al. Postoperative Strahlentherapie bei Patienten mit epithelialem Thymuskarzinom (TET). Strahlenther Onkol. 2015;191:133– 140 [in German]. 6. Fuller CD, Housman DM, Thomas CR. Radiotherapy for thymoma and thymic carcinoma. Hematol Oncol Clin North Am. 2008;22:489–507. 7. Lombe DC, Jeremic B. A Review of the place and role of radiotherapy in thymoma. Clin Lung Cancer. 2015;16:406–412. 8. Lim YJ, Kim E, Kim HJ, et al. Survival impact of adjuvant radiotherapy in Masaoka stage II-IV thymomas–a systematic review and meta-analysis. Int J Radiat Oncol Biol Phys. 2016;94:1129–1136.
9. Patel S, Macdonald OK, Nagda S, Bittner N, Suntharalingam M. Evaluation of the role of radiation therapy in the management of malignant thymoma. Int J Radiat Oncol Biol Phys. 2012;82:1797–1801. 10. Lim YJ, Kim HJ, Wu H-G. Role of postoperative radiotherapy in nonlocalized thymoma: propensity-matched analysis of Surveillance, Epidemiology, and End Results Database. J Thorac Oncol. 2015;10:1357–1363. 11. Weksler B, Shende M, Nason KS, Gallagher A, Ferson PF, Pennathur A. The role of adjuvant radiation therapy for resected stage III thymoma: a population-based study. Ann Thorac Surg. 2012;93:1822–1829. 12. Weksler B, Dhupar R, Parikh V, Nason KS, Pennathur A, Ferson PF. Thymic carcinoma: a multivariate analysis of factors predictive of survival in 290 patients. Ann Thorac Surg. 2013;95:299–303. 13. Forquer JA, Rong N, Fakiris AJ, Loehrer PJ, Johnstone PA. Postoperative radiotherapy after surgical resection of thymoma: differing roles in localized and regional disease. Int J Radiat Oncol Biol Phys. 2010;76:440–445. 14. Nakagawa K, Yokoi K, Nakajima J, et al. Is thymomectomy alone appropriate for stage I (T1N0M0) thymoma? Results of a propensity-score analysis. Ann Thorac Surgy. 2016;101:520–526. 15. Fernandes AT, Shinohara ET, Guo M, et al. The role of radiation therapy in malignant thymoma: a Surveillance, Epidemiology, and End Results database analysis. J Thorac Oncol. 2010;5:1454–1460. 16. Omasa M, Sozu T, Sato T, et al. Postoperative radiotherapy is effective for thymic carcinoma but not for thymoma in stage II and III thymic epithelial tumors: the Japanese Association for Research on the Thymus Database study. Cancer. 2015;121:1008–1016. 17. Mariano C, Ionescu DN, Cheung WY, et al. Thymoma: a population-based study of the management and outcomes for the province of British Columbia. J Thorac Oncol. 2013;8:109–117. 18. Ruffini E, Detterbeck F, Van Raemdonck D, et al. Thymic carcinoma: a cohort study of patients from the European Society of Thoracic Surgeons database. J Thorac Oncol. 2014;9:541–548.
744 Jackson et al 19. Tseng Y-H, Lin Y-H, Tseng Y-C, et al. Adjuvant therapy for thymic carcinoma–a decade of experience in a Taiwan National Teaching Hospital. PloS One. 2016;11: e0146609. 20. Ahmad U, Yao X, Detterbeck F, et al. Thymic carcinoma outcomes and prognosis: results of an international analysis. J Thorac Cardiovasc Surg. 2015;149:95–101. 21. Hishida T, Nomura S, Yano M, et al. Long-term outcome and prognostic factors of surgically treated thymic carcinoma: results of 306 cases from a Japanese Nationwide Database study. Eur J Cardiothorac Surg. 2016;49:835–841. 22. National Cancer Data Base. https://www.facs.org/quality %20programs/cancer/ncdb. Accessed September 15, 2015. 23. International Agency for Research on Cancer. International Classification of Diseases for Oncology including updates as at Sep 01 2011, approved by the IARC/WHO Committee for ICD-O-3. ICD-O-3 Online 2011; 3.1. http://codes.iarc.fr/ codegroup/2. Accessed September 15, 2015. 24. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619. 25. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials. 1996;17:343–346. 26. Coca-Perraillon M. Local and global optimal propensity score matching. SAS Global Forum 2007. http://www2.sas.com/proceedings/forum2007/1852007.pdf. Accessed October 15, 2015. 27. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10:150–161. 28. Ettinger D. NCCN guidelines version 2.2016. Thymomas and thymic carcinomas. http://www.nccn.org/professionals/ physician_gls/pdf/thymic.pdf. Accessed March 15, 2016. 29. Utsumi T, Shiono H, Kadota Y, et al. Postoperative radiation therapy after complete resection of thymoma has little impact on survival. Cancer. 2009;115: 5413–5420.
Journal of Thoracic Oncology
Vol. 12 No. 4
30. Korst RJ, Kansler AL, Christos PJ, Mandal S. Adjuvant radiotherapy for thymic epithelial tumors: a systematic review and meta-analysis. Ann Thorac Surg. 2009;87:1641–1647. 31. Rena O, Papalia E, Oliaro A, et al. Does adjuvant radiation therapy improve disease-free survival in completely resected Masaoka stage II thymoma? Eur J Cardiothorac Surg. 2007;31:109–113. 32. Fu H, Gu Z-T, Fang W-T, et al. Long-term survival after surgical treatment of thymic carcinoma: a retrospective analysis from the Chinese Alliance for Research of Thymoma database. Ann Surg Oncol. 2016;23:619–625. 33. Liu Q, Gu Z, Yang F, et al. The role of postoperative radiotherapy for stage I/II/III thymic tumor—results of the ChART retrospective database. J Thorac Dis. 2016;8:687. 34. Boothe D, Orton A, Thorpe C, Kokeny K, Hitchcock YJ. Postoperative radiotherapy in locally invasive malignancies of the thymus: patterns of care and survival. J Thorac Oncol. 2016;11:2218–2226. 35. Rimner A, Yao X, Huang J, et al. Postoperative radiation therapy is associated with longer overall survival in completely resected stage II and III thymoma—an analysis of the International Thymic Malignancies Interest Group retrospective database. J Thorac Oncol. 2016;11:1785–1792. 36. Zhou D, Deng X-F, Liu Q-X, Zheng H, Min J-X, Dai J-G. The effectiveness of postoperative radiotherapy in patients with completely resected thymoma: a meta-analysis. Ann Thorac Surg. 2016;101:305–310. 37. Bucci MK, Bevan A, Roach M. Advances in radiation therapy: conventional to 3D, to IMRT, to 4D, and beyond. CA: Cancer J Clin. 2005;55:117–134. 38. Leuzzi G, Rocco G, Ruffini E, et al. Multimodality therapy for locally advanced thymomas: a propensity score– matched cohort study from the European Society of Thoracic Surgeons Database. J Thorac Cardiovasc Surg. 2016;151:47–57.