A National Analysis of Open vs Minimally Invasive Thymectomy for Stage I-III Thymoma

Journal Pre-proof A National Analysis of Open vs Minimally Invasive Thymectomy for Stage I-III Thymoma Chi-Fu Jeffrey Yang, Jacob Hurd, Shivani Shah, Douglas Liou, Hanghang Wang, Leah Backhus, Natalie Lui, Thomas D'Amico, Joseph Shrager, Mark Berry PII:

S0022-5223(19)37101-6

DOI:

https://doi.org/10.1016/j.jtcvs.2019.11.114

Reference:

YMTC 15485

To appear in:

The Journal of Thoracic and Cardiovascular Surgery

Received Date: 9 May 2019 Revised Date:

16 November 2019

Accepted Date: 19 November 2019

Please cite this article as: Jeffrey Yang C-F, Hurd J, Shah S, Liou D, Wang H, Backhus L, Lui N, D'Amico T, Shrager J, Berry M, A National Analysis of Open vs Minimally Invasive Thymectomy for Stage I-III Thymoma, The Journal of Thoracic and Cardiovascular Surgery (2020), doi: https:// doi.org/10.1016/j.jtcvs.2019.11.114. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Copyright © 2019 Published by Elsevier Inc. on behalf of The American Association for Thoracic Surgery

1

Title: A National Analysis of Open vs Minimally Invasive Thymectomy for Stage I-III

2

Thymoma

3 4

Running Head: MIS Thymectomy

5 6

Authors: Chi-Fu Jeffrey Yang1, Jacob Hurd2, Shivani Shah3, Douglas Liou1, Hanghang Wang2,

7

Leah Backhus1,4, Natalie Lui1, Thomas D'Amico2, Joseph Shrager1, Mark Berry1,4

8 9

Institutions and Affiliations:

10

1. Stanford University, Stanford, CA

11

2. Duke University Medical Center, Durham, NC

12

3. Harvard Medical School, Boston, MA

13

4. VA Palo Alto Health Care System, Palo Alto, Calif

14 15

Meeting Presentation: To be presented at the 99th Annual Meeting of the American Association

16

for Thoracic Surgery, 2019 Toronto, Canada

17 18

Classifications: thymoma, thymectomy, MIS surgery, VATS, robotic

19 20

Abstract Word Count: 245

21

Manuscript Word Count: 3220

22

Reference Count: 37

23

Tables & Figures Count: 7

24

Supplementary Element Count: 9

25 26

Corresponding author: Chi-Fu Jeffrey Yang. Falk Cardiovascular Research Center, 300

27

Pasteur

28

[email protected]

29

Reprints will not be available from the authors

Drive,

Stanford,

CA

94305.Ph:

650-721-6400.

Fax:

650-724-6259. Email:

30 31

Sources of Support/Conflicts of Interest: The authors have no conflicts of interest to declare. 1

32

TAD is a consultant for Scanlan (<$10,000). TD is a consultant for Medtronic (<$5,000).

2

33

Central Message

34

In this national analysis, minimally invasive thymectomy was associated with similar short-term

35

outcomes and intermediate-term survival when compared with open thymectomy for stage I-III

36

thymoma.

37

3

38

Perspective Statement

39

In this national analysis, when compared with open thymectomy, minimally invasive

40

thymectomy for stage I-III thymoma was associated with shorter length of stay and was not

41

associated with increased margin positivity, perioperative mortality, 30-day readmission rate, or

42

reduced five year survival.

4

43

Central Picture Legend

44

Open vs MIS Thymectomy for Stage I-III Thymoma: Propensity score-matched Survival

5

45

Glossary of Abbreviations:

46

MIS: Minimally Invasive

47

VATS: Video-assisted Thoracoscopic

48

NCDB: National Cancer Database

49

LOS: Length of Stay

50

RATS: Robot-assisted Thoracoscopic

51

CDCC: Charlson-Deyo Comorbidity

6

52 53 54 55

ABSTRACT

56

characterized. We compared short-term outcomes and overall survival between open and MIS

57

(video-assisted thoracoscopic [VATS] and robotic) approaches using the National Cancer Data

58

Base (NCDB).

Objective The oncologic efficacy of minimally invasive (MIS) thymectomy for thymoma is not well

59 60

Methods

61

Perioperative outcomes and survival of patients who underwent open versus MIS thymectomy

62

for clinical stage I-III thymoma from 2010 to 2014 in the NCDB were evaluated using

63

multivariable Cox proportional hazards modeling and propensity score-matched analysis.

64

Predictors of MIS use were evaluated using multivariable logistic regression. Outcomes of

65

surgical approach were evaluated using an intent-to-treat analysis.

66 67

Results

68

Of the 1,223 thymectomies that were evaluated, 317 (26%) were performed minimally invasively

69

(141 VATS and 176 robotic). The MIS group had a shorter median (IQR) length of stay (LOS)

70

when compared to the open group (3[2,4] days vs 4[3,6] days, p < 0.001). In a propensity score-

71

matched analysis of 185 open and 185 MIS (VATS + robotic) thymectomy, the MIS group

72

continued to have a shorter median LOS (3 versus 4 days, p < 0.01) but did not have significant

73

differences in margin positivity (p = 0.84), 30-day readmission (p = 0.28), 30-day mortality (p =

74

0.60) and 5-year survival (89.4% vs 81.6%, p = 0.20) when compared to the open group.

75 76

Conclusions

77

In this national analysis, MIS thymectomy was associated with shorter LOS and was not

78

associated with increased margin positivity, perioperative mortality, 30-day readmission rate, or

79

reduced overall survival when compared with open thymectomy.

80 81 82 83 7

84 85 86

INTRODUCTION

87

minimally invasive (MIS) thymectomy techniques have been developed over the past two

88

decades. Since the first case report for a video-assisted thoracoscopic (VATS) approach to

89

thymectomy for thymoma in 1992,2 studies have reported the outcomes of both VATS and

90

robotic-assisted thoracoscopic (RATS) approaches to thymectomy.3 These studies have

91

generally found that when compared to a traditional open approach, a MIS approach was

92

associated with reduced blood loss, chest tube duration and hospital length of stay and no

93

significant differences in perioperative complications, thymoma recurrence, and 5-year survival.3

94

However, concerns on the use of minimally invasive techniques for thymectomy and thymoma

95

resection have been raised due to the risk of thymoma capsule violation during minimally

96

invasive manipulation leading to pleural seeding that compromises the oncologic efficacy of the

97

procedure with ultimate thymoma recurrence. In fact, minimally invasive approaches were

98

previously considered to be appropriate only for smaller tumors less than 4 to 5 centimeters in

99

size4.

The traditional approach for a thymectomy for thymoma has been via median sternotomy,1 but

100 101

Although evidence to support those concerns or support a size limit for minimally invasive

102

resection has not been published, the majority of studies comparing outcomes of open vs MIS

103

thymectomy are single- or multi-center studies from high volume centers. To date, there have

104

only been a few national studies evaluating the role of a MIS approach in thymectomy. The

105

Japanese Association for Research on the Thymus (JART) performed a propensity score-

106

matched study of 280 patients from 32 Japanese institutions and found no difference in

107

recurrence-free and overall survival between open and VATS thymectomy.5 The reported the

8

108

results of 461 patients across four continents who underwent MIS thymectomy for thymoma6.

109

They found that minimally invasive thymectomy could achieve rates of R0 resection for

110

thymoma similar to those achieved with traditional open thymectomy. A recent U.S. National

111

Cancer Data Base (NCDB) analysis of 943 patients with stage I-II thymoma also found no

112

significant differences in R0 resection between open and minimally invasive approaches to

113

thymectomy.7 However, the ITMIG and NCDB studies did not specifically evaluate survival

114

and over 80% of patients in the ITMIG study who underwent a MIS thymectomy were from

115

Europe and Asia.

116 117

This study was therefore undertaken to evaluate the short-term outcomes and overall survival of

118

open versus MIS thymectomy for clinical stage I-III thymoma in the United States using the

119

National Cancer Data Base (NCDB), which includes data from surgeons in academic and

120

community centers across the U.S. The study objective was to test the hypothesis that the MIS

121

approach is associated with improved short-term outcomes and similar overall survival when

122

compared to thymectomy by open approaches.

123 124

METHODS

125

Data source

126

The NCDB is administered by both the American College of Surgeons Commission on Cancer

127

(CoC) and the American Cancer Society, and captures approximately 70% of all newly

128

diagnosed cases of cancer in the U.S. and Puerto Rico.8 The NCDB collects information from

129

more than 1,500 cancer centers in the U.S., and now contains more than 30 million patient

130

records.

9

131 132

Study Design:

133

This retrospective study was approved by the Stanford and Duke University Institutional Review

134

Board. Patients diagnosed with stage I-III thymoma (Masaoka–Koga staging) from 2010-2014 in

135

the 2015 NCDB Participant Use Data File were identified for inclusion using International

136

Classification of Diseases for Oncology, 3rd edition (ICD-O-3) histology and topography codes.

137 138

Only patients treated with thymectomy (identified using Surgical Procedure of the Primary Site

139

codes 30, 40, 50, and 60 as defined by the Facility Oncology Registry Data Standards criteria),

140

who had available data on surgical approach were included. Exclusion criteria included non-

141

malignant pathology, history of previous unrelated malignancy and age less than 18 years. The

142

primary outcome was overall survival, assessed from the time of diagnosis to the time of death or

143

last follow-up.

144

hospital, 90-day mortality, hospital length of stay (LOS), surgical margin positivity, and rates of

145

conversion to open. The years 2010-2014 were selected for analysis because data on surgical

146

approach was not available before 2010.

Secondary outcomes were 30-day mortality and readmission to the same

147 148

Statistical Analysis:

149

Patients diagnosed with stage I-III thymoma were grouped based on whether they received open

150

or minimally invasive thymectomy. Baseline characteristics and unadjusted outcomes were

151

analyzed using Pearson’s chi-square test for categorical variables and Wilcoxon Rank Sum test

152

for continuous variables.

153

10

154

Outcomes of surgical approach were assessed using an intent-to-treat analysis; patients who

155

underwent MIS or MIS converted to open thymectomy were both included in the MIS group.

156

Predictors of a MIS approach were assessed using a multivariable logistic regression model that

157

included variables felt to be relevant to treatment selection. The variables included in this

158

multivariable logistic regression model were age, sex, race, Charlson/Deyo comorbidity (CDCC)

159

score, Masaoka stage, tumor size, regional education levels, insurance type, histology, facility

160

type, and distance from facility. A multivariable Cox proportional hazards model was used to

161

assess differences in overall survival between the open and MIS thymectomy groups, adjusting

162

for the aforementioned variables.

163 164

Differences in perioperative and postoperative outcomes and overall survival between surgical

165

approaches were also assessed using a propensity score-matched analysis of open vs MIS

166

thymectomy, using methods similar to as previously described.9 Propensity scores were

167

developed, defined as the probability of treatment assignment with the MIS approach versus

168

open approach, conditional on age, sex, race, CDCC score, regional education levels, tumor size,

169

insurance type, histology, stage, year of diagnosis, distance from facility, and facility type. All

170

of the covariates selected for the model were determined a priori to be clinically relevant.

171

Applying a greedy nearest neighbor matching algorithm without replacement with a caliper of

172

0.01, the most appropriately-matched pairs were identified. After matching, balance among the

173

pairs was evaluated using standardized differences. Following propensity-score matching,

174

Kaplan-Meier analysis was used to assess the overall survival of the two groups. Secondary

175

outcomes were assessed using the Mann-Whitney U test for continuous measures and Pearson’s

176

chi-square test for discrete variables.

11

177 178

Additional subgroup analyses were performed as well to try to limit the impact of unmeasurable

179

selection biases that would impact the ability to assess differences between the different surgical

180

approaches. The aforementioned propensity score-matched analysis, using the same matching

181

algorithm and covariates detailed above, was performed for patients who underwent open vs MIS

182

thymectomy with no co-morbidities to try to better control for the possibility that either approach

183

might have been used in “sicker” patients whose co-morbidities could impact outcomes more

184

than surgical approach. In addition, a propensity score-matched analysis of patients who

185

underwent open vs MIS thymectomy for only stage I-II thymoma was also assessed using the

186

same methodology as described above to better control for the possibility that a particular

187

surgical approach might have been preferentially used for more complex tumors that already had

188

a higher risk of incomplete resection or worse overall survival.

189 190

Two additional propensity score-matched analyses were performed using the same methodology

191

as described above. The first analysis was limited to patients with tumors < 4 cm and the second

192

analysis was limited to patients with tumors > 4 cm. We also performed both a unadjusted and

193

propensity score-matched comparison of VATS vs RATS using the same methodology as noted

194

above.

195 196

Diagnostics and model balance were evaluated without any violation of major assumptions being

197

observed. A p-value of 0.05 was used to define statistical significance for all comparisons.

198

Statistical analyses were performed using Stata/MP software, version 13.1 for Mac (StataCorp,

199

College Station, TX).

12

200 201

RESULTS

202

Use and Predictors of Minimally Invasive Surgery

203

The MIS approach was used in 317 (25.9%) of 1,223 patients in the NCDB who underwent

204

thymectomy for stage I-III thymoma from 2010-2014 and met the study inclusion criteria (Figure

205

1). Figure 2 shows the percentage of open and MIS thymectomies performed per year of study.

206

MIS use increased with each year except from 2010-2011. In 2010, 45 (18.7%) of 241

207

thymectomies were performed via a MIS approach. By 2014, 84 (33.2%) of 253 thymectomies

208

were performed via a MIS. This translated to a 14.5% increase in MIS use over the 5-year study

209

period.

210 211

Baseline patient and tumor characteristics are displayed in Table 1. In univariable analysis,

212

patients undergoing MIS thymectomy were more likely to be of older age, to have smaller

213

tumors and to have earlier stage thymomas when compared to patients who received open

214

thymectomy. No significant differences were found between the MIS and open groups with

215

regard to sex, race, co-morbidities or histology. The results of multivariable analysis that

216

evaluated predictors of the MIS approach are detailed in Table 2. In this multivariable analysis,

217

patients with stage III (compared to stage I) thymoma and patients with larger tumors were less

218

likely to receive a thymectomy via a MIS approach.

219 220

Perioperative and Survival Outcomes in the Entire Cohort

221

Table 1 also details the perioperative and postoperative data of the cohorts. The MIS approach

222

was associated with a shorter median length of hospital stay than the open approach, but the two

13

223

groups did not differ significantly with regard to margin positivity, 30-day mortality, 90-day

224

mortality or 30-day readmission rate.

225 226

The median follow-up for the open group was 40.7 months (IQR 27.3, 56.8) while the median

227

follow-up for the MIS group was 35.9 months (IQR 24.9, 52.2). Kaplan–Meier analysis

228

demonstrated a 5-year survival of 86.9% (95% CI: 83.6-89.7) for the open group and 90.7%

229

(95% CI: 82.0-95.3) for MIS group (log-rank, p=0.04) (Figure 3A). In multivariable Cox

230

proportional hazards analysis, a MIS approach was not associated with worse overall survival

231

(HR: 0.57; 95% CI: 0.31-1.07; p=0.08) (Table 3).

232 233

Data regarding margin status

234

Data on margin status is detailed in Table 1. Of the patients with known data regarding margin

235

status, 26.9% (n = 226) patients in the open group had positive margins while 22.1% (n = 65)

236

patients in the MIS group had positive margins. There were no significant differences between

237

the two groups regarding margin status (p = 0.39).

238 239

Propensity Score-matched Analysis

240

Propensity-score matching was performed to create two groups of 185 patients who had an open

241

or MIS approach that were well-matched with regard to baseline patient and tumor

242

characteristics (Table 4 & Figure 4). All standardized mean differences were less than or equal

243

to 10.2%. Table 4 also shows perioperative and postoperative data for the two matched groups.

244

The MIS group did not differ significantly from the open group with regard to margin positivity,

245

30-day mortality, 90-day mortality, or 30-day readmission rate but did have a shorter median

14

246

length of hospital stay. The median follow-up for the open group was 35.9 months (IQR 25.4,

247

50.5) while the median follow-up for the MIS group was 36.4 months (IQR 25.8, 55.4). There

248

were no significant differences in five-year overall survival between the open (81.6%, 95% CI:

249

68.1-89.8) and MIS (89.4%, 95% CI: 78.6-95.0) groups (log-rank, p=0.20) (Figure 3B & Figure

250

4).

251 252

Propensity Score-matched Analysis: Data regarding margin status

253

Data on margin status is detailed in Table 4. Of the patients with known data regarding margin

254

status, 24.6% (n = 42) patients in the open group had positive margins while 19.0% (n = 33)

255

patients in the MIS group had positive margins. There were no significant differences between

256

the two groups regarding margin status (p = 0.65).

257 258

Propensity score-matched Analysis: Patients with No Comorbidities

259

A comparison of baseline characteristics after propensity matching between patients with no-

260

comorbidities who underwent open and MIS thymectomy for stage I-III thymoma, is detailed in

261

Supplemental Table 1. After propensity-score matching, both groups were well matched with all

262

standardized mean differences less than or equal to 11.8%. Supplemental Table 1 also shows

263

perioperative and postoperative data for both matched groups. The MIS group did not differ

264

significantly from the open group with regard to margin positivity, and 30-day and 90-day

265

mortality rates but did have a shorter length of hospital stay. The MIS group was associated with

266

a higher 30-day readmission rate than the open group. There were no significant differences in

267

five-year overall survival between the open (79.0%, 95% CI: 64.1-88.3) and MIS (89.7%, 95%

268

CI: 75.2-95.9) groups (log-rank, p = 0.07) (Supplemental Figure 1A).

15

269 270

Propensity score-matched Analysis: Patients with Stage I-II Thymoma

271

An additional propensity score-matched analysis was performed for patients with stage I-II

272

thymoma. After propensity-score matching, both groups were well matched (Supplemental

273

Table 2). Supplemental Table 2 also shows perioperative and postoperative data for the two

274

matched groups. The MIS group did not differ significantly from the open group with regard to

275

margin positivity, 30-day mortality, 90-day mortality, or 30-day readmission rate but did have a

276

shorter median length of hospital stay. There were no significant differences in five-year overall

277

survival between the open (88.5%, 95% CI: 78.0-94.2) and MIS (90.6%, 95% CI: 77.3-96.3)

278

groups (log-rank, p=0.73) (Supplemental Figure 1B).

279 280

Propensity score-matched Analysis: Tumors < 4 cm and Tumors > 4 cm

281

Two additional propensity score-matched analyses were performed. The first analysis was

282

limited to patients with tumors < 4 cm (results detailed in Supplemental Table 3). In this

283

subgroup analysis, there were no significant differences in margin positivity or overall survival

284

(Supplemental Table 3 and supplemental figure 2A). The second analysis was limited to patients

285

with tumors > 4 cm (results detailed in Supplemental Table 4). In this subgroup analysis, there

286

were no significant differences in margin positivity or overall survival (Supplemental Table 4

287

and supplemental figure 2B).

288 289

VATS vs RATS

290

A comparison of differences in baseline characteristics, perioperative outcomes and overall

291

survival between VATS vs RATS in unadjusted and propensity score-matched analysis is

16

292

detailed in Supplemental Table 5 and 6 and Supplemental Figures 3A and 3B. The RATS group

293

had a shorter length of stay than the VATS group in unadjusted analyses although in the

294

propensity score-matched analysis there were no significant differences in length of stay between

295

the groups. There were no other significant differences in perioperative outcomes and overall

296

survival.

297 298

DISCUSSION

299

In this study of stage I-III thymoma in the National Cancer Data Base, MIS thymectomy was

300

found to be used for only a minority of patients with thymoma. When compared to patients who

301

underwent open thymectomy, the MIS group did not have significantly different rates of margin

302

positivity, 30-day mortality, 90-day mortality, or 30-day readmission but did have a shorter

303

median length of hospital stay, in both unadjusted and propensity score-matched analyses. The

304

MIS group as compared to the open group had significantly better five-year overall survival in

305

unadjusted analysis and did not have worse survival in multivariable-adjusted and propensity

306

score-matched analysis. Overall, these results suggest that minimally invasive techniques can be

307

used when resecting thymomas without compromising oncologic efficacy.

308 309

Our study findings are consistent with those reported by previous studies that found that MIS

310

thymectomy was associated with similar or lower 30-day mortality,3, 7, 10-12 shorter length of

311

hospital stay,3, 11-32 and similar or better five-year overall survival rates.3, 5, 10-13, 15, 16, 22, 24-28, 33

312

The main difference between the present study findings and those previously reported is with

313

regard to the rate of positive margins. Previous studies have reported R0 resection rates ranging

314

from 47% to 100% and 44% to 100% for MIS and open thymectomy patient groups,

17

315

respectively.5-7, 11, 16, 18, 20, 22, 23, 28, 30, 31, 34-36 In the present study, while there were no significant

316

differences between open and MIS thymectomy with regard to margin status, both groups had

317

higher than expected number of positive margins. A R0 resection was achieved in 67.7% of

318

patients with open thymectomy and 72.2% of patients with MIS thymectomy. In contrast, in the

319

ITMIG study of 2053 open thymectomies and 461 minimally invasive thymectomies, 88% of all

320

thymectomies achieved a R0 resection and that a R0 resection was achieved in 94% of MIS

321

cases.6 This discrepancy is in part because ~7% of patients in each group in our present study

322

had unknown data regarding margins. However, the difference could also be that ITMIG data is

323

provided directly from participating surgeons with their own assessment of the margin status

324

based on both intraoperative findings as well as pathologic data. NCDB data is coded by

325

registrars based on pathologic reports. We speculate that there were probably cases where the

326

pathology report noted that the tumor extended to the edge of the specimen and was coded by the

327

registrars or the pathologist as being a positive margin simply because the tumor was not next to

328

normal tissue but in actuality would not have been considered a positive margin by the surgeon

329

(e.g. in the case where the tumor “hangs” into the “air” of the pleural space when the lung is

330

down). In such cases, the clinical judgement of the surgeon who performed the operation is

331

often needed to clarify whether the margin is positive.

332 333

Our study has the strength of having assembled a large cohort of patients across a wide variety of

334

academic and community center, allowing both subgroup analyses as well as generalizability

335

beyond high volume centers. However this study does have several limitations. First, it is a

336

retrospective nature and the potential presence of unobserved confounding and selection bias

337

exists. We did try to reduce bias and account for confounding variables by performing

18

338

multivariable modeling and propensity-score matching. In addition, we used propensity-

339

matching analyses of patients with no comorbidities and patients only with lower stage I-II

340

thymoma to further consider that certain approaches may have been more preferentially used for

341

“sicker” patients or less complex tumors. Second, the NCDB does not contain any details

342

regarding the exact operative approach for the open thymectomy group (e.g. sternotomy vs

343

thoracotomy) nor does it have data regarding surgeon experience. Third, as noted above, there

344

may have been some inaccuracies or inconsistencies in the margin status data. Fourth, the

345

NCDB does not have patient data regarding whether a patient had myasthenia gravis. Fifth, the

346

NCDB does not contain data on postoperative complications, disease-free and disease-specific

347

survival. Sixth, the NCDB does not have data regarding the use of enhanced recovery after

348

surgery (ERAS) or fast-track protocols that have been shown to reduce length of stay.37 The

349

findings from the study suggest that the MIS approach is associated with faster recovery, with a

350

shorter median length of stay by 1 day. However, it should be noted that with the use of ERAS

351

protocols, patients who undergo open approaches to thymectomy can also often go home within

352

3 days after an operation. Future investigation that includes data on ERAS protocol

353

implementation for thymectomies will better clarify the impact of an MIS approach on post-

354

operative recovery. Seventh, the most important limitation of the study is with regard to the

355

relatively short follow-up of the study. The indolent nature of thymoma is such that finding no

356

difference in 5-year survival does not ensure that there is no difference in tumor recurrence or

357

longer-term survival. Additional studies with longer follow-up will have to be done to ensure

358

that these results found for up to 5-year outcomes are maintained over a longer period.

359

19

360

In this national analysis of patients with stage I-III thymoma, MIS thymectomy was observed to

361

be associated with shorter length of hospital stay, and similar margin positivity, 30-day mortality,

362

90-day mortality, 30-day readmission, and five-year survival thymectomy performed via an open

363

approach. While the use of the MIS approach has been increasing, it is still only used in the

364

minority of patients undergoing thymectomy. This study supports surgeons using minimally

365

invasive techniques to resect stage I-III thymomas when they consider that those techniques will

366

allow complete resection without tumor spillage.

367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396

REFERENCES 1. 2. 3.

4. 5.

6.

7.

8.

9.

10.

Keynes G. The surgery of the thymus gland. Br J Surg. 1946;33:201-214. Landreneau RJ, Dowling RD, Castillo WM, Ferson PF. Thoracoscopic resection of an anterior mediastinal tumor. The Annals of thoracic surgery. 1992;54:142-144. Hess NR, Sarkaria IS, Pennathur A, Levy RM, Christie NA, Luketich JD. Minimally invasive versus open thymectomy: a systematic review of surgical techniques, patient demographics, and perioperative outcomes. Annals of Cardiothoracic Surgery. 2016;5:19. Eds Selke FW dNP, Swanson SJ. . Anterior Mediastinal Masses. Sabiston & Spencer Surgery of the Chest 7th Edition. Agatsuma H, Yoshida K, Yoshino I, et al. Video-Assisted Thoracic Surgery Thymectomy Versus Sternotomy Thymectomy in Patients With Thymoma. The Annals of thoracic surgery. 2017;104:1047-1053. Burt BM, Yao X, Shrager J, et al. Determinants of Complete Resection of Thymoma by Minimally Invasive and Open Thymectomy: Analysis of an International Registry. Journal of Thoracic Oncology. 2017;12:129-136. Burt BM, Nguyen D, Groth SS, et al. Utilization of Minimally Invasive Thymectomy and Margin Negative Resection for Early Stage Thymoma. The Annals of thoracic surgery. 2019. Bilimoria KY, Stewart AK, Winchester DP, Ko CY. The National Cancer Data Base: a powerful initiative to improve cancer care in the United States. Annals of surgical oncology. 2008;15:683-690. Yang CJ, Kumar A, Gulack BC, et al. Long-term outcomes after lobectomy for non-small cell lung cancer when unsuspected pN2 disease is found: A National Cancer Data Base analysis. The Journal of thoracic and cardiovascular surgery. 2015. Hao W, Zhitao G, Jianyong D, et al. Perioperative outcomes and long-term survival in clinically early-stage thymic malignancies: video-assisted thoracoscopic thymectomy versus open approaches. Journal of Thoracic Disease. 2016;8:673-679.

20

397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

Liu TJ, Lin M-W, Hsieh M-S, et al. Video-Assisted Thoracoscopic Surgical Thymectomy to Treat Early Thymoma: A Comparison with the Conventional Transsternal Approach. Annals of surgical oncology. 2014;21:322-328. Xie A, Tjahjono R, Phan K, Yan TD. Video-assisted thoracoscopic surgery versus open thymectomy for thymoma: a systematic review. Annals of Cardiothoracic Surgery. 2015;4:495-508. Chao Y-K, Liu Y-H, Hsieh M-J, et al. Long-Term Outcomes After Thoracoscopic Resection of Stage I and II Thymoma: A Propensity-Matched Study. Annals of surgical oncology. 2015;22:1371-1376. Cheng Y-J, Kao E-L, Chou S-H. Videothoracoscopic Resection of Stage II Thymoma: Prospective Comparison of the Results Between Thoracoscopy and Open Methods. Chest. 2005;128:3010-3012. Chung JW, Kim HR, Kim DK, et al. Long-term Results of Thoracoscopic Thymectomy for Thymoma without Myasthenia Gravis. Journal of International Medical Research. 2012;40:1973-1981. Fadayomi AB, Iniguez CEB, Chowdhury R, et al. Propensity Score Adjusted Comparison of Minimally Invasive versus Open Thymectomy in the Management of Early Stage Thymoma. The Thoracic and cardiovascular surgeon. 2018;66:352-358. He Z, Zhu Q, Wen W, Chen L, Xu H, Li H. Surgical approaches for stage I and II thymoma-associated myasthenia gravis: feasibility of complete video-assisted thoracoscopic surgery (VATS) thymectomy in comparison with trans-sternal resection. Journal of Biomedical Research. 2013;27:62-70. Jurado J, Javidfar J, Newmark A, et al. Minimally Invasive Thymectomy and Open Thymectomy: Outcome Analysis of 263 Patients. The Annals of thoracic surgery. 2012;94:974-982. Kimura T, Inoue M, Kadota Y, et al. The oncological feasibility and limitations of videoassisted thoracoscopic thymectomy for early-stage thymomas. European Journal of Cardio-Thoracic Surgery. 2013;44:e214-e218. Kneuertz PJ, Kamel MK, Stiles BM, et al. Robotic Thymectomy Is Feasible for Large Thymomas: A Propensity-Matched Comparison. The Annals of thoracic surgery. 2017;104:1673-1678. Liqiang Q, Xiaoke C, Jia H, et al. A comparison of three approaches for the treatment of early-stage thymomas: robot-assisted thoracic surgery, video-assisted thoracic surgery, and median sternotomy. Journal of Thoracic Disease. 2017;9:1997-2005. Maniscalco P, Tamburini N, Quarantotto F, Grossi W, Garelli E, Cavallesco G. Longterm outcome for early stage thymoma: comparison between thoracoscopic and open approaches. The Thoracic and cardiovascular surgeon. 2015;63:201-205. Manoly I, Whistance RN, Sreekumar R, et al. Early and mid-term outcomes of transsternal and video-assisted thoracoscopic surgery for thymoma†. European Journal of Cardio-Thoracic Surgery. 2014;45:e187-e193. Marulli G, Comacchio GM, Schiavon M, et al. Comparing robotic and trans-sternal thymectomy for early-stage thymoma: a propensity score-matching study. European Journal of Cardio-Thoracic Surgery. 2018:ezy075-ezy075. Odaka M, Akiba T, Mori S, et al. Oncological outcomes of thoracoscopic thymectomy for the treatment of stages I–III thymomas. Interactive CardioVascular and Thoracic Surgery. 2013;17:285-290. 21

443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479

26.

Odaka M, Shibasaki T, Asano H, Marushima H, Yamashita M, Morikawa T. Feasibility of thoracoscopic thymectomy for treatment of early stage thymoma. Asian journal of endoscopic surgery. 2015;8:439-444. 27. Odaka M, Shibasaki T, Kato D, et al. Comparison of oncological results for early- and advanced-stage thymomas: thoracoscopic thymectomy versus open thymectomy. Surgical endoscopy. 2017;31:734-742. 28. Pennathur A, Qureshi I, Schuchert MJ, et al. Comparison of surgical techniques for earlystage thymoma: Feasibility of minimally invasive thymectomy and comparison with open resection. The Journal of Thoracic and Cardiovascular Surgery. 2011;141:694-701. 29. Wilshire CL, Vallières E, Shultz D, Aye RW, Farivar AS, Louie BE. Robotic Resection of 3 cm and Larger Thymomas Is Associated With Low Perioperative Morbidity and Mortality. Innovations:Technology and Techniques in Cardiothoracic and Vascular Surgery. 2016;11:321-326. 30. Ye B, Li W, Ge X-X, et al. Surgical treatment of early-stage thymomas: robot-assisted thoracoscopic surgery versus transsternal thymectomy. Surgical endoscopy. 2014;28:122-126. 31. Ye B, Tantai J-C, Ge X-X, et al. Surgical techniques for early-stage thymoma: Videoassisted thoracoscopic thymectomy versus transsternal thymectomy. The Journal of Thoracic and Cardiovascular Surgery. 2014;147:1599-1603. 32. Yuan Z-Y, Cheng G-Y, Sun K-L, et al. Comparative study of video-assisted thoracic surgery versus open thymectomy for thymoma in one single center. Journal of Thoracic Disease. 2014;6:726-733. 33. Sakamaki Y, Oda T, Kanazawa G, Shimokawa T, Kido T, Shiono H. Intermediate-term oncologic outcomes after video-assisted thoracoscopic thymectomy for early-stage thymoma. The Journal of Thoracic and Cardiovascular Surgery. 2014;148:12301237.e1231. 34. Keijzers M, Dingemans A-mC, Blaauwgeers H, et al. 8 Years' experience with robotic thymectomy for thymomas. Surgical endoscopy. 2014;28:1202-1208. 35. Rowse PG, Roden AC, Corl FM, et al. Minimally invasive thymectomy: the Mayo Clinic experience. 2015. 2015;4:519-526. 36. Ye B, Tantai J-C, Li W, et al. Video-assisted thoracoscopic surgery versus roboticassisted thoracoscopic surgery in the surgical treatment of Masaoka stage I thymoma. World journal of surgical oncology. 2013;11:157. 37. Martin L, Sarosiek B, Harrison M, et al. Implementing a Thoracic Enhanced Recovery Program: Lessons Learned In the First Year. The Annals of thoracic surgery. 2018.

22

480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525

Table and Figure Legend Figure 1. Flow diagram showing schema of study subject selection Figure 2. Changes in Surgical Approaches for Thymectomy Over Time Table 1. Analysis of Open vs. Minimally Invasive (MIS [Video-Assisted Thoracoscopic or Robotic]) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data Table 2. Multivariable Logistic Regression Evaluating Predictors of the Use of Minimally Invasive (MIS) Thymectomy for Patients with Stage 1-3 Thymoma Figure 3. Survival after Open vs. Minimally Invasive (MIS) Thymectomy for Stage I-III Thymoma: A. Entire Cohort B. Propensity Score-matched Analysis Table 3. Independent Predictors of Survival After Cox Proportional Hazards Adjustment for Patients with Stage 1-3 Thymoma Table 4. Propensity-matched Analysis of Open vs. Minimally Invasive (MIS [Video-Assisted Thoracoscopic or Robotic]) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data Central Figure. Open vs MIS Thymectomy for Stage I-III Thymoma: Propensity score-matched Survival Figure 4. Graphical Abstract Demonstrating the Comparison of 5-Year Survival Outcomes between Minimally Invasive (MIS) and Open Thymectomy for Patients with Stage I-III Thymoma in the National Cancer Database through a Propensity Score-Matched Analysis Video. A National Analysis of Open vs Minimally Invasive Thymectomy for Stage I-III Thymoma Supplemental Table 1. Propensity Score-Matched Analysis of Patients without Comorbidities of Open vs Minimally Invasive (MIS [Video-Assisted Thoracoscopic or Robotic]) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data Supplemental Figure 1. Survival after Open vs. Minimally Invasive (MIS) Thymectomy for Stage I-III Thymoma. A. Propensity Score-matched Patients without Comorbidities B. Propensity Score-Matched Patients with Stage I & II Thymoma Supplemental Table 2. Propensity-matched Analysis of Open vs. Minimally Invasive (MIS) Thymectomy for Stage I & II: Baseline Characteristics, Perioperative and Postoperative Data Supplemental Table 3. Propensity Score-Matched Analysis of Patients with Tumors Less than 4cm of Open vs. Minimally Invasive (MIS) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data

23

526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544

Supplemental Table 4. Propensity Score-Matched Analysis of Patients with Tumors Greater than 4cm of Open vs. Minimally Invasive (MIS) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data Supplemental Figure 2. Survival after Open vs. Minimally Invasive (MIS) Thymectomy for Stage I-III Thymoma. A. Propensity Score-matched Patients with Tumors less than 4cm B. Propensity Score-matched Patients with Tumors Greater than 4cm Supplemental Table 5. Analysis of Patients of Video-Assisted Thoracoscopic (VATS) vs. Robotic (RATS) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data Supplemental Table 6. Propensity Score-Matched Analysis of Patients of Video-Assisted Thoracoscopic (VATS) vs. Robotic (RATS) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data Supplemental Figure 3. Survival after Video-Assisted Thoracoscopic (VATS) vs. Robotic (RATS) Thymectomy for Stage I-III Thymoma. A. Entire Cohort B. Propensity Score-matched Analysis

24

545 546

Table 1. Analysis of Open vs. MIS (VATS or Robotic) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data

Patient characteristics

Open (N=906)

MIS (N=317)

P Value

57.4 (14.1) 486 (53.6%)

59.6 (12.7) 170 (53.6%)

0.02 1.00 0.48

632 (69.8%) 157 (17.3%) <10 94 (10.4%)

236 (74.4%) 40 (12.6%) 0 37 (11.7%)

682 (75.3%) 181 (20.0%) 36 (4.0%) <10

247 (77.9%) 55 (17.4%) 14 (4.4%) <10

141 (15.6%) 249 (27.5%) 283 (31.2%) 233 (25.7%)

49 (15.5%) 59 (18.6%) 113 (35.6%) 94 (29.7%)

400 (44.2%) 30 (3.3%) 315 (34.8%) 59 (6.5%)

161 (50.8%) <10 (<2%) 99 (31.2%) 27 (8.5%)

506 (55.8%) 66 (7.3%) 266 (29.4%) <10 (<1%) 43 (4.7%)

169 (53.3%) 11 (3.5%) 120 (37.9%) <10 (<2%) <10 (<3%)

432 (47.7%) 196 (21.6%) 278 (30.7%) 65 (45, 90)

203 (64.0%) 77 (24.3%) 37 (11.7%) 49.5 (35, 70)

95 (10.5%) 201 (32.7%) 129 (14.2%) 168 (18.5%) 149 (16.4%) 164 (18.1%)

34 (10.7%) 77 (24.3%) 51 (16.1%) 65 (20.5%) 35 (11.0%) 55 (17.4%)

Baseline Characteristics Age, (years, SD) Female, n (%) Race, n (%) White Black Asian Other Charlson Deyo Comorbidity Score, n (%) 0 1 2 3+ Education (% Without HS Diploma) 21% or more 13% - 20.9% 7% - 12.9% Less than 7% Facility, n (%) Academic/Research Program Community Cancer Program Comprehensive Community Cancer Program Integrated Network Cancer Program Insurance, n (%) Private Medicaid Medicare Other Government Program Uninsured Masaoka Stage, n (%) 1-2a 2b 3 Tumor Size, mm, median (IQR) Histology, n (%) Thymoma, type A, malignant Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant Thymoma, malignant, NOS Induction therapy, n (%) Induction Chemotherapy Induction Chemoradiation Induction Radiation

0.59

0.02

0.13

0.02

<0.001

<0.001 0.29

0.09 104 (11.5%) 12 (1.3%) 13 (1.4%)

13 (4.1%) <10 (<2%) <10 (<2%)

25

Distance to Facility (IQR) MIS Approach, n (%) VATS RATS

10.4 (5.4,28)

12.3 (4.7,30.4)

N/A N/A

141 (44.5%) 176 (55.5%)

0.44

Perioperative Outcomes Conversion from MIS to open, n (%) Surgical margins, n (%) Negative Microscopic Residual Tumor Macroscopic Residual Tumor Residual Tumor, NOS Unknown Nodes Removed, n (IQR) 30-day mortality, n (%) 30-day readmission, n (%) 90-day mortality, n (%) Hospital Length of Stay, n (IQR)

34 (10.7%)

N/A 0.62

613 (67.7%) 132 (14.6%) 11 (1.2%) 83 (9.2%) 67 (7.4%) 18.2 (36.1) <10 51 (3.7%) <10 4 (3, 6)

229 (72.2%) 39 (12.3%) <10 24 (7.6%) 23 (7.3%) 17.4 (36.2) <10 15 (4.2%) <10 3 (2, 4)

0.05 0.75 0.38 0.99 <0.001

337 (32.7%) 65 (3.7%)

99 (30.8%) <10

0.06 <0.001

Postoperative Therapy Adjuvant Radiotherapy Adjuvant Chemotherapy

547 548 549

26

550 551

Table 2. Multivariable Logistic Regression Evaluating Predictors of the Use of MIS Thymectomy for Patients With Stage 1-3 Thymoma Age (per year) Female v male Race (ref = White) Black Other CDCC score (ref = 0) 1 2 3+ Masaoka Stage (ref Stage 1-2a) Stage 2b Stage 3 Tumor Size (per mm) Education: % without HS diploma (ref > 21%) 13% to 20.9% 7% to 12.9% Less than 7% Insurance type (ref = Uninsured) Private Medicaid Medicare Other Government Histology, (ref=Thymoma, type A, malignant) Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant Facility type (ref = community) Comprehensive Academic/research Integrated Network Distance from facility (per mile)

Odds Ratio 0.99 0.86

95% CI 0.97, 1.01 0.62, 1.21

P 0.53 0.40

1.09 1.36

0.65, 1.84 0.78, 2.38

0.74 0.28

0.66 0.86 0.33

0.43, 1.02 0.40, 1.84 0.04, 2.96

0.06 0.70 0.32

0.90 0.34 0.98

0.61, 1.34 0.20, 0.56 0.98, 0.99

0.61 <0.001 <0.001

0.54 0.74 0.77

0.31, 0.96 0.44, 1.24 0.45, 1.32

0.04 0.26 0.35

2.67 1.54 4.30 5.87

0.75, 9.46 0.34, 6.90 1.16, 15.97 0.94, 36.53

0.13 0.57 0.03 0.06

1.07 1.45 1.32 0.80

0.63, 1.82 0.80, 2.64 0.74, 2.34 0.42, 1.55

0.81 0.22 0.35 0.51

1.94 2.30 2.63 1.00

0.66, 5.65 0.79, 6.69 0.81, 8.57 1.00, 1.00

0.23 0.13 0.11 0.50

552 553

27

554 555

Table 3. Independent Predictors of Survival After Cox Proportional Hazards Adjustment for Patients with Stage 1-3 Thymoma

MIS (ref = Open) Age (per year) Female v male Race (ref = White) Black Other CDCC score (ref = 0) 1 2 3+ Masaoka Stage (ref Stage 1-2a) Stage 2b Stage 3 Tumor Size (per mm) Education: % without HS diploma (ref > 21%) 13% to 20.9% 7% to 12.9% Less than 7% Insurance type (ref = Uninsured) Private Medicaid Medicare Other Government Histology, (ref= Thymoma, type A, malignant) Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant Facility type (ref = community) Comprehensive Academic/research Unknown Distance from facility (per mile)

Hazard Ratio 0.57 1.05 0.64

95% CI 0.31, 1.07 1.03, 1.08 0.40, 1.02

P 0.08 <0.001 0.062

0.92 0.37

0.46, 1.82 0.13, 1.06

0.80 0.06

1.17 0.85 2.01

0.68, 2.01 0.26, 2.77 0.46, 8.87

0.57 0.78 0.36

1.88 2.05 1.00

1.06, 3.34 1.17, 3.60 1.00, 1.01

0.03 0.01 0.27

0.93 1.08 0.63

0.45, 1.90 0.54, 2.19 0.29, 1.38

0.83 0.82 0.25

0.74 1.17 1.03 0.91

0.17, 3.24 0.21, 6.56 0.22, 4.74 0.07, 11.12

0.69 0.86 0.97 0.94

0.93 0.89 1.04 1.38

0.40, 2.19 0.41, 1.94 0.44, 2.47 0.62, 3.05

0.87 0.76 0.94 0.43

1.98 1.67 3.02 1.00

0.46, 8.54 0.38, 7.23 0.65, 14.16 1.00, 1.00

0.36 0.50 0.16 0.97

556 557

28

558 559

Table 4. Propensity-matched Analysis of Open vs. MIS (VATS or Robotic) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data

Patient characteristics

Open (N=185)

MIS (N=185)

Standardized Difference

Baseline Characteristics Age, (years, SD) Female, n (%) Race, n (%) White Black Asian Other Charlson Deyo Comorbidity Score, n (%) 0 1 2 3+ Education (% Without HS Diploma), n (%) 21% or more 13% - 20.9% 7% - 12.9% Less than 7% Facility, n (%) Academic/Research Program Community Cancer Program Comprehensive Community Cancer Program Integrated Network Cancer Program Insurance, n (%) Private Medicaid Medicare Other Government Program Uninsured Masaoka Stage, n (%) 1-2a 2b 3 Tumor Size, mm (IQR) Histology, n (%) Thymoma, type A, malignant Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant Induction Therapy, n (%) Induction Chemotherapy Induction Chemoradiation Induction Radiation Year of Diagnosis, (IQR) Distance to Facility (IQR)

62.6 (11.1) 100 (54.1%)

61.6 (10.4) 97 (52.4%)

6.5 <0.1

129 (69.7%) 38 (20.5%)

145 (78.4%) 22 (11.9%)

4.9 1.5

0 (0%)

0 (0%)

<0.1

<10

18 (9.7%)

5.3

143 (77.3%) 33 (17.8%) <10 0 (0%)

138 (74.6%) 37 (20.0%) <10 <10 (<1%)

7.7 5.6 2.6 6.7

33 (17.8%) 34 (18.4%) 70 (37.8%) 47 (25.4%)

25 (13.5%) 37 (20.0%) 64 (34.6%) 59 (31.9%)

<0.1 5.2 2.3 7.2

87 (47.0%) <10 75 (40.5%) 16 (8.6%)

96 (51.9%) <10 68 (36.8%) 16 (9.8%)

8.7 6.8 5.7 2.0

95 (51.4%) <10 77 (41.6%) <10 <10

96 (51.9%) <10 75 (40.5%) <10 <10

<0.1 4.7 1.2 4.7 5.7

116 (62.7%) 40 (21.6%) 29 (15.7%) 53 (35,75)

110 (59.5%) 49 (26.5%) 26 (14.1%) 50 (40,70)

4.5 9.0 4.2 1.0

32 (17.3%) 64 (34.6%) 32 (17.3%) 38 (20.5%) 19 (10.3%)

32 (17.3%) 58 (31.9%) 29 (15.7%) 40 (21.6%) 26 (14.1%)

<0.1 7.2 4.2 2.5 10.2

11 (5.9%) <10 0 2012 (2011,2013) 11.3 (5.6,25.4)

<10 <10 <10 2013 (2011,2013) 14 (5.2,29.3)

5.8 5.3 9.5 1.6 7.7

29

Perioperative Outcomes Conversion from open to MIS, n (%) Surgical margins, n (%) Negative Microscopic Residual Tumor Macroscopic Residual Tumor Residual Tumor, NOS Unknown Nodes Removed, n (SD) 30-day mortality, n (%) 30-day readmission, n (%) 90-day mortality, n (%) Hospital Length of Stay, n (IQR)

19 (10.3%)

p-value N/A 0.84

129 (69.7%) 26 (14.1%) <10 15 (8.1%) 14 (7.6%) 19 (36.9) <10 <10 <10 4 (3, 5)

141 (76.2%) 21 (11.4%) <10 11 (5.9%) 11 (5.9%) 18.7 (37.6) <10 <10 <10 3 (2, 4)

0.56 1.00 0.28 0.60 <0.001

62 (33.5%) <10

55 (29.7%) <10

0.43 0.78

Postoperative Therapy Adjuvant Radiotherapy Adjuvant chemotherapy

560

30

561 562 563

Supplemental Table 1. Propensity Score-Matched Analysis of Patients without Comorbidities of Open vs. MIS (VATS or Robotic) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data

Patient characteristics

Baseline Characteristics Age, (years, SD) Female, n (%) Race, n (%) White Black Asian Other Education (% Without HS Diploma), n (%) 21% or more 13% - 20.9% 7% - 12.9% Less than 7% Facility, n (%) Academic/Research Program Community Cancer Program Comprehensive Community Cancer Program Integrated Network Cancer Program Insurance, n (%) Private Medicaid Medicare Other Government Program Uninsured Masaoka Stage, n (%) 1-2a 2b 3 Tumor Size, mm, median (IQR) Histology, n (%) Thymoma, type A, malignant Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant Induction therapy, n (%) Induction Chemotherapy Induction Chemoradiation Induction Radiation Year of Diagnosis, (IQR)

Open (N=141)

MIS (N=141)

Standardized Differences (%)

61.4 (12.0) 77 (49.5%)

61.4 (10.3) 78 (54.2%)

0.1 1.4

117 (83.0%) 15 (10.6%)

113 (80.1%) 14 (9.9%)

6.4 2.0

0 (0%) <10

0 (0%) 14 (9.9%)

<0.1 11.5

20 (14.1%) 26 (18.4%) 46 (32.6%) 49 (34.8%)

22 (15.6%) 28 (19.9%) 47 (33.3%) 44 (31.2%)

3.9 3.4 1.5 7.8

82 (58.1%) <10 41 (29.1%) 14 (9.9%)

76 (53.9%) <10 49 (34.8%) 12 (8.5%)

8.5 10.6 11.8 5.2

81 (57.4%) <10 52 (36.9%) <10 0 (0%)

78 (55.3%) <10 54 (38.3%) <10 <10

4.3 3.0 3.0 6.1 3.7

83 (58.9%) 36 (25.5%) 22 (15.6%) 52 (35,70)

83 (58.9%) 38 (30.0%) 20 (14.2%) 50 (35,70)

<0.1 3.3 3.6 0.6

21 (14.9%) 44 (31.2%) 24 (17.0%) 30 (21.3%) 22 (15.6%)

21 (14.9%) 46 (32.6%) 26 (18.4%) 29 (20.6%) 19 (13.5%)

<0.1 3.2 3.6 1.6 5.8

13 (9.2%) 0 (0%) 0 (0%) 2013 (2011,2013)

<10 0 (0%) <10 2013 (2011,2013)

10.4 <0.1 7.4 1.0

31

Distance to Facility (IQR) Perioperative Outcomes Surgical margins Negative Microscopic Residual Tumor Macroscopic Residual Tumor Residual Tumor, NOS Unknown Nodes Removed, n (SD) 30-day mortality, n (%) 30-day readmission, n (%) 90-day mortality, n (%) Hospital Length of Stay, n (IQR)

12.4 (6.2,25.4)

11 (4.9, 24.7)

96 (68.1%) 14 (9.9%) <10 15 (10.6%) 14 (9.9%) 19.1 (37.4) 0 (0%) 0 (0%) 0 (0%) 4 (3, 5)

102 (72.3%) 18 (12.8%) 0 (0%) 10 (7.1%) 11 (7.8%) 19.9 (38.6) <10 <10 <10 3 (2, 4)

6.2 p-value 0.54

0.91 0.37 0.04 0.22 <0.001

Postoperative Therapy Adjuvant Radiotherapy Adjuvant chemotherapy

51 (36.2%) <10

46 (32.6%) <10

0.53 0.16

564 565

32

566 567 568

Supplemental Table 2. Propensity-matched Analysis of Open vs. MIS Thymectomy for Stage I & II: Baseline Characteristics, Perioperative and Postoperative Data

Patient characteristics Baseline Characteristics Age, (years, SD) Female, n (%) Race, n (%) White Black Asian Other Charlson Deyo Comorbidity Score, n (%) 0 1 2 3+ Education (% Without HS Diploma) 21% or more 13% - 20.9% 7% - 12.9% Less than 7% Facility, n (%) Academic/Research Program Community Cancer Program Comprehensive Community Cancer Program Integrated Network Cancer Program Insurance, n (%) Private Medicaid Medicare Other Government Program Uninsured Masaoka Stage, n (%) 1-2a 2b Tumor Size, mm, median (IQR) Histology, n (%) Thymoma, type A, malignant Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant Induction therapy, n (%) Induction Chemotherapy Induction Chemoradiation

Open (N=165)

MIS (N=165)

Standardized Differences

63.2 (11.1) 92 (55.8%)

62.0 (10.6) 88 (53.3%)

8.8 4.9

132 (80.0%) 17 (10.3%)

127 (77.0%) 21 (12.7%)

6.8 6.8

0 (0%)

0 (0%)

0.0

16 (9.7%)

17 (10.3%)

2.0

120 (72.7%) 33 (20.0%) 12 (7.3%) 0 (0%)

120 (72.7%) 33 (20.0%) 11 (6.7%) <10 (<1%)

0 0 2.8 7.2

22 (13.3%) 41 (24.8%) 58 (35.2%) 44 (26.7%)

23 (13.9%) 36 (21.8%) 55 (33.3%) 51 (30.9%)

1.7 7.3 3.8 9.4

83 (50.2%) <10 66 (40.0%) 13 (7.9%)

81 (49.1%) <10 64 (38.8%) 16 (9.7%)

2.4 2.7 2.5 6.5

79 (47.9%) <10 74 (44.8%) <10 <10

88 (53.3%) <10 66 (40.0%) <10 <10

10.9 5.1 10.1 0.0 3.3

120 (72.7%) 45 (27.3%) 50 (35, 69)

118 (71.5%) 47 (28.5%) 50 (35, 65)

2.6 2.6 0.6

29 (17.6%) 58 (35.2%) 31 (18.8%) 30 (18.2%) 17 (10.3%)

29 (17.6%) 53 (32.1%) 29 (17.6%) 33 (20.0%) 21 (12.7%)

0.0 6.5 3.2 4.4 6.8

<10 0 (0%)

<10 <10

11.2 10.9

33

Induction Radiation Distance to Facility (IQR) Perioperative Outcomes Surgical margins, n (%) Negative Microscopic Residual Tumor Macroscopic Residual Tumor Residual Tumor, NOS Unknown Nodes Removed, n (SD) 30-day mortality, n (%) 30-day readmission, n (%) 90-day mortality, n (%) Hospital Length of Stay, n (IQR)

0 (0%) 10.9 (4.9, 20)

<10 (<2%) 11.8 (5.2, 24)

16.7 2.5 p-value 0.81

128 (77.6%) 19 (11.5%) 0 (0%) <10 12 (7.3%) 25.1 (41.4) 0 (0%) <10 <10 4 (3,5)

130 (78.8%) 16 (9.7%) 0 (0%) <10 10 (6.1%) 21.0 (39.2) <10 <10 <10 3 (1,4)

0.23 0.51 0.29 0.90 <0.001

51 (30.9%) <10

46 (27.9%) <10

0.55 0.74

Postoperative Therapy Adjuvant Radiotherapy Adjuvant chemotherapy

569 570

34

571 572 573

Supplemental Table 3. Propensity Score-Matched Analysis of Patients with Tumors Less than 4cm of Open vs. MIS (VATS or Robotic) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data

Patient characteristics

Baseline Characteristics Age, (years, SD) Female, n (%) Race, n (%) White Black Asian Other Charlson Deyo Comorbidity Score, n (%) 0 1 2 3+

Education (% Without HS Diploma), n (%) 21% or more 13% - 20.9% 7% - 12.9% Less than 7% Facility, n (%) Academic/Research Program Community Cancer Program Comprehensive Community Cancer Program Integrated Network Cancer Program Insurance, n (%) Private Medicaid Medicare Other Government Program Uninsured Masaoka Stage, n (%) 1-2a 2b 3 Tumor Size, mm (IQR) Histology, n (%) Thymoma, type A, malignant Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant

Open (N=37)

MIS (N=37)

Standardized Differences (%)

60.2 (9.0) 22 (59.5%)

60.5 (10.5) 23 (62.2%)

1.9 5.4

29 (78.4%) <10

27 (73.0%) <10

12.3 15.5

0 <10

0 <10

<0.1 <0.1

30 (81.1%)

29 (78.4%)

<10 <10

<10 <10

0

0

6.4 <0.1 12.2 <0.1

<10 <10 11 (29.7%) 13 (35.1%)

<10 11 (29.7%) 10 (27.0%) 12 (32.4%)

7.4 19.6 6.0 5.7

20 (54.1%) <10 15 (40.5%) <10

23 (62.2%) <10 11 (29.7%) <10

16.2 19.7 22.5 10.2

21 (56.8%) 0 13 (35.1%) 0 <10

21 (56.8%) 0 14 (37.8%) 0 <10

<0.1 <0.1 5.5 <0.1 14.5

31 (83.8%) <10 <10 30 (20,34)

29 (78.4%) <10 <10 30 (25,35)

12.2 13.7 <0.1 21.8

<10 13 (35.1%) <10 10 (27.0%) <10

<10 14 (37.8%) <10 <10 <10

7.2 5.9 7.4 24.2 8.7

35

Induction therapy, n (%) Induction Chemotherapy Induction Chemoradiation Induction Radiation Year of Diagnosis, (IQR) Distance to Facility (IQR) Perioperative Outcomes Surgical margins Negative Microscopic Residual Tumor Macroscopic Residual Tumor Residual Tumor, NOS Unknown Nodes Removed, n (SD) 30-day mortality, n (%) 30-day readmission, n (%) 90-day mortality, n (%) Hospital Length of Stay, n (IQR)

<10 0 0 2012 (2012,2014) 15 (6.5,25.4)

0 0 0 2013 (2012,2014) 8 (2.5,19.8)

23.2 <0.1 <0.1 4.0 10.2 p-value

31 (83.8%) 0 0 <10 <10 14.2 (34) 0 <10 0 3.5 (2,4)

31 (83.8%) <10 0 0 <10 23 (40.6) <10 0 <10 2 (1,4)

0.11 0.36 0.31 0.36 0.09

10 (27.0%) <10

<10 <10

0.41 0.30

0.50

Postoperative Therapy Adjuvant Radiotherapy Adjuvant chemotherapy

574 575

36

576 577 578

Supplemental Table 4. Propensity Score-Matched Analysis of Patients with Tumors Greater than 4cm of Open vs. MIS (VATS or Robotic) Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data

Patient characteristics

Baseline Characteristics Age, (years, SD) Female, n (%) Race, n (%) White Black Asian Other Charlson Deyo Comorbidity Score, n (%) 0 1 2 3+

Education (% Without HS Diploma), n (%) 21% or more 13% - 20.9% 7% - 12.9% Less than 7% Facility, n (%) Academic/Research Program Community Cancer Program Comprehensive Community Cancer Program Integrated Network Cancer Program Insurance, n (%) Private Medicaid Medicare Other Government Program Uninsured Masaoka Stage, n (%) 1-2a 2b 3 Tumor Size, mm (IQR) Histology, n (%) Thymoma, type A, malignant Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant

Open (N=137)

MIS (N=137)

Standardized Differences (%)

62.0 (11.5) 62 (45.3%)

61.9 (10.4) 67 (48.9%)

0.8 7.3

97 (70.8%) 22 (16.1%)

105 (76.6%) 16 (11.7%)

13.0 12.2

0 18 (13.1%)

0 16 (11.7%)

<0.1 4.7

100 (73.0%)

99 (72.2%) 30 (21.9%)

0

<10 <10

1.7 5.5 10.5 8.0

21 (15.3%) 21 (15.3%) 51 (37.2%) 44 (32.1%)

18 (13.1%) 28 (20.4%) 48 (35.0%) 43 (31.4%)

5.9 12.1 4.6 1.6

69 (50.4%) <10 52 (38.0%) 12 (8.8%)

67 (48.9%) <10 53 (38.7%) 13 (9.5%)

2.9 <0.1 1.5 2.6

70 (51.1%) <10 55 (40.1%) <10 0

72 (52.6%) <10 56 (40.9%) <10 <10

2.9 8.9 1.6 5.9 3.8

83 (60.6%) 27 (19.7%) 27 (19.7%) 60 (50-80)

80 (58.4%) 34 (24.8%) 23 (16.8%) 60 (48-75)

4.5 11.9 7.0 2.2

17 (12.4%) 44 (32.1%) 27 (19.7%) 30 (21.9%) 19 (13.9%)

19 (13.9%) 44 (32.1%) 24 (17.5%) 29 (21.2%) 21 (15.3%)

4.5 <0.1 5.5 1.7 3.8

27 (19.7%) 10 (7.3%)

37

Induction therapy, n (%) Induction Chemotherapy Induction Chemoradiation Induction Radiation Year of Diagnosis, (IQR) Distance to Facility (IQR) Perioperative Outcomes Surgical margins Negative Microscopic Residual Tumor Macroscopic Residual Tumor Residual Tumor, NOS Unknown Nodes Removed, n (SD) 30-day mortality, n (%) 30-day readmission, n (%) 90-day mortality, n (%) Hospital Length of Stay, n (IQR)

17 (12.4%) <10 <10 2012 (2011-2013) 11.3 (5.2-25.4)

<10 <10 <10 2013 (2011-2013) 15.7 (5.8-30.7)

18.9 <0.1 5.6 3.1 16.3 p-value

98 (71.5%) 19 (13.9%) 0 11 (8.0%) <10 17.3 (35.2) 0 <10 <10 4 (3,6)

97 (70.8%) 18 (13.1%) <10 <10 12 (8.8%) 18.4 (37.3) <10 <10 <10 3 (2,4)

0.53 0.61 1.00 1.00 <0.001

45 (32.8%) <10

44 (32.1%) <10

0.90 1.00

0.50

Postoperative Therapy Adjuvant Radiotherapy Adjuvant chemotherapy

579 580

38

581 582

Supplemental Table 5. Analysis of Patients of VATS vs. Robotic Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data

Patient characteristics Baseline Characteristics Age, (years, SD) Female, n (%) Race, n (%) White Black Asian Other Charlson Deyo Comorbidity Score, n (%) 0 1 2 3+

Education (% Without HS Diploma), n (%) 21% or more 13% - 20.9% 7% - 12.9% Less than 7% Facility, n (%) Academic/Research Program Community Cancer Program Comprehensive Community Cancer Program Integrated Network Cancer Program Insurance, n (%) Private Medicaid Medicare Other Government Program Uninsured Masaoka Stage, n (%) 1-2a 2b 3 Tumor Size, mm (IQR) Histology, n (%) Thymoma, type A, malignant Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant Induction therapy, n (%) Induction Chemotherapy

VATS (N=141)

RATS (N=176)

62 75 (53.2%)

59 95 (54.0%)

104 (73.8%) 22 (15.6%)

132 (75.0%) 18 (10.2%)

0 (0%)

0 (0%)

15 (8.5%)

26 (14.2%)

116 (82.3%) 19 (13.5%) <10 0

131 (74.4%) 36 (20.5%) <10 <10

p-value 0.19 0.89 0.24

0.30

0.43 21 (14.9%) 25 (17.7%) 45 (31.9%) 48 (34.0%)

28 (15.9%) 34 (19.3%) 68 (38.6%) 46 (26.1%) 0.10

66 (46.8%) <10 51 (36.2%) <10

95 (54.0%) <10 48 (27.3%) 19 (10.8%)

65 (46.1%) <10 59 (41.8%) <10 <10

104 (59.1%) <10 61 (34.7%) <10 <10

0.20

0.27 86 (61.0%) 34 (24.1%) 21 (14.9%) 52 (36, 78)

117 (66.5%) 43 (24.4%) 16 (9.1%) 45 (35, 63)

14 (9.9%) 35 (24.8%) 24 (17.0%) 26 (18.4%) 21 (14.9%)

20 (11.4%) 43 (24.4%) 27 (15.3%) 39 (22.2%) 14 (8.0%)

0.05 0.40

0.53 <10

<10

39

Induction Chemoradiation Induction Radiation Year of Diagnosis, (IQR) Distance to Facility (IQR) Perioperative Outcomes Conversion to Open, n (%) Surgical margins Negative Microscopic Residual Tumor Macroscopic Residual Tumor Residual Tumor, NOS Unknown Nodes Removed, n (SD) 30-day mortality, n (%) 30-day readmission, n (%) 90-day mortality, n (%) Hospital Length of Stay, n (IQR)

<10 <10 2013 (2011,2014) 10.9 (4, 29.2)

<10 <10 2013 (2011.5, 2013) 12.9 (5.8, 31.3)

26 (18.4%)

<10

98 (69.5%) 17 (12.1%) <10 12 (8.5%) 12 (8.5%) 0 (0,2) <10 <10 <10 3 (2, 4)

131 (74.4%) 22 (12.5%) 0 12 (6.8%) 11 (6.3%) 0 (0,6) 0 <10 0 2 (1, 4)

0.46 0.27 0.46 0.06 0.02

42 (29.8%) <10

57 (32.4%) 0

0.62 <0.01

0.95 0.07

<0.01 0.60

Postoperative Therapy Adjuvant Radiotherapy Adjuvant chemotherapy

583

40

584 585

Supplemental Table 6. Propensity Score-Matched Analysis of Patients of VATS vs. Robotic Thymectomy: Baseline Characteristics, Perioperative and Postoperative Data

Patient characteristics

Baseline Characteristics Age, (years, SD) Female, n (%) Race, n (%) White Black Asian Other Charlson Deyo Comorbidity Score, n (%) 0 1 2 3+ Education (% Without HS Diploma), n (%) 21% or more 13% - 20.9% 7% - 12.9% Less than 7% Facility, n (%) Academic/Research Program Community Cancer Program Comprehensive Community Cancer Program Integrated Network Cancer Program Insurance, n (%) Private Medicaid Medicare Other Government Program Uninsured Masaoka Stage, n (%) 1-2a 2b 3 Tumor Size, mm (IQR) Histology, n (%) Thymoma, type A, malignant Thymoma, type AB, malignant Thymoma, type B1, malignant Thymoma, type B2, malignant Thymoma, type B3, malignant Induction therapy, n (%)

VATS (N=77)

RATS (N=77)

61.1 (12.2) 45 (58.4%)

60.9 (10.7) 46 (59.7%)

1.0 2.6

61 (79.2%) 11 (14.3%)

65 (84.4%) <10 (<12%)

11.9 7.5

0 (0%) <10

0 <10

Standardized Differences (%)

<0.1 8.5

62 (80.5%) 11 (14.3%) <10 0

63 (81.8%) 10 (13.0%) <10 0

3.1 3.5 <0.1 <0.1

12 (15.6%) 12 (15.6%) 20 (26.0%) 32 (41.6%)

14 (18.2%) 16 (20.8%) 27 (35.1%) 20 (26.0%)

7.0 12.7 18.6 34.9

38 (49.4%) <10 29 (37.7%) <10

43 (55.8%) 0 19 (24.7%) 12 (15.6%)

13.5 50.0 28.4 38.5

39 (50.6%) <10 33 (42.9%) <10 0

38 (49.4%) <10 33 (42.9%) <10 0

2.6 <0.1 <0.1 10.5 <0.1

48 (62.3%) 19 (24.7%) 10 (13.0%) 45 (35,70)

46 (59.7%) 21 (27.3%) 10 (13.0%) 45 (30,65)

5.4 6.0 <0.1 2.0

12 (15.6%) 23 (30.0%) 15 (19.5%) 16 (20.8%) 11 (14.3%)

16 (20.8%) 23 (30.0%) 11 (14.3%) 23 (30.0%) <10

15.4 <0.1 13.0 21.0 27.1

41

Induction Chemotherapy Induction Chemoradiation Induction Radiation Year of Diagnosis, (IQR) Distance to Facility (IQR) Perioperative Outcomes Surgical margins Negative Microscopic Residual Tumor Macroscopic Residual Tumor Residual Tumor, NOS Unknown Nodes Removed, n (SD) 30-day mortality, n (%) 30-day readmission, n (%) 90-day mortality, n (%) Hospital Length of Stay, n (IQR) Postoperative Therapy Adjuvant Radiotherapy Adjuvant chemotherapy

<10 <10 <10 2013 (2011,2014) 7.5 (2.8,20.0)

<10 0 <10 2013 (2012,2013) 13.5 (5.9,42.6)

18.2 14.7 <0.1 1.0 32.0 p-value 0.52

56 (72.7%) <10 <10 <10 <10 12.7 (31.6) <10 <10 <10 3 (2,4)

59 (76.6%) <10 0 <10 <10 16.7 (35.8) 0 <10 0 2 (1, 3)

0.79 0.32 0.34 0.22 0.08

22 (28.6%) <10

27 (35.1%) 0

0.39 0.02

586

42