Journal Pre-Proof: The Journal of Thoracic and Cardiovascular Surgery

on July 26, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc.
All rights reserved.

Downloaded for Anonymous User (n/a) at Meiyu International Health Management (Beijing) Co., Ltd. from ClinicalKey.com by Elsevier
Journal Pre-proof
Lobectomy, segmentectomy or wedge resection for peripheral clinical T1aN0 non-

small cell lung cancer: a post-hoc analysis of CALGB 140503 (Alliance)
Nasser Altorki, MD, Xiaofei Wang, MD, Bryce Damman, MD, Jennifer Mentlick, MD,
Rodney Landreneau, MD, Dennis Wigle, MD, David R. Jones, MD, Massimo Conti,
MD, Ahmad S. Ashrafi, MD, Moishe Liberman, MD, Marc de Perrot, MD, John D.
Mitchell, MD, Robert Keenan, Thomas Bauer, MD, Daniel Miller, MD, Thomas E.
Stinchcombe, MD
PII: S0022-5223(23)00612-8
DOI: https://doi.org/10.1016/j.jtcvs.2023.07.008
Reference: YMTC 19128
To appear in: The Journal of Thoracic and Cardiovascular Surgery
Received Date: 28 April 2023

Revised Date: 13 June 2023
Accepted Date: 4 July 2023
Please cite this article as: Altorki N, Wang X, Damman B, Mentlick J, Landreneau R, Wigle D, Jones DR,
Conti M, Ashrafi AS, Liberman M, de Perrot M, Mitchell JD, Keenan R, Bauer T, Miller D, Stinchcombe
TE, Lobectomy, segmentectomy or wedge resection for peripheral clinical T1aN0 non-small cell lung
cancer: a post-hoc analysis of CALGB 140503 (Alliance), The Journal of Thoracic and Cardiovascular
Surgery (2023), doi: https://doi.org/10.1016/j.jtcvs.2023.07.008.
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 © 2023 Published by Elsevier Inc. on behalf of The American Association for Thoracic Surgery
of
ro
-p
re
lP
na
ur
Jo
on July 26, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
1 Lobectomy, segmentectomy or wedge resection for peripheral clinical T1aN0 non-small cell
2 lung cancer: a post-hoc analysis of CALGB 140503 (Alliance)
3
4
5 Nasser Altorki, MD1, Xiaofei Wang, MD2 , Bryce Damman, MD3 , Jennifer Mentlick, MD3,
6 Rodney Landreneau, MD4 , Dennis Wigle, MD5, David R. Jones, MD6, Massimo Conti, MD7,
7 Ahmad S. Ashrafi, MD8, Moishe Liberman, MD9, Marc de Perrot, MD10 , John D. Mitchell,
8 MD11, Robert Keenan12, Thomas Bauer, MD13, Daniel Miller, MD14, Thomas E. Stinchcombe,
9 MD15.
10
1
11 Weill Cornell Medicine - New York-Presbyterian Hospital, New York, NY, USA.
2
12 Alliance Statistics and Data Management Center, and Biostatistics and Bioinformatics, Duke
13 University, Durham, NC, USA.
of
3
14 Alliance Statistics and Data Management Center, Mayo Clinic , Rochester, MN.
4
15 University of Pittsburgh Medical Center, Pittsburgh, PA, USA.
ro
5
16 Mayo Clinic, Rochester, MN, USA.
6
17 Memorial Sloan Kettering Cancer Center, New York, NY, USA.
-p
7
18 Institut Universitaire de Cardiologie et Pneumologie de Québec, Québec, QC, Canada.
8
19 Surrey Memorial Hospital Thoracic Group Fraser Valley Health Authority, BC, Canada.
20 9
re
Centre Hospitalier de l'Université de Montréal, Montreal, QC, Canada.
10
21 University of Toronto, Toronto, ON, Canada.
11
lP
22 University of Colorado Cancer Center, University of Colorado School of Medicine, Aurora,
23 CO, USA.
12
24 Moffitt Cancer Center, Tampa, FL, USA.
na
13
25 Hackensack Meridian Health Center, Hackensack, NJ, USA.
14
26 Emory University School of Medicine, Atlanta, GA, USA.
15
27 Duke Cancer Institute, Duke University Medical Center, Durham, NC, USA.
ur
28
29 Correspondence:
Jo
30 Nasser Altorki M.D.

31 Vice Chairman Cardiothoracic Surgery
32 Chief Thoracic surgery
33 Weill Cornell Medicine
34 Email: nkaltork@med.cornell.edu
35
36 Word count: 3901
37
38 Disclosure statement
39 Nasser Altorki reports grant funding from Astra Zeneca LLC and Janssen Pharmaceuticals,
40 steering committee membership for Roche/genetech and Adisory Board for Regeneron
41 Pharmaceuticals. Daniel Miller reports consulting for Ethicon Inc. and speakers bureau for
42 Articure Inc. John Mitchel reports consulting for Intuitive Surgical. David Jones reports steering
43 committee membership for Merck Pharmaceuticals and consulting for Astra Zeneca LLC. Marc
44 de Perrot reports speaker and consultant fees AstraZeneca and Bristol Myer Squib, consultant
45 fees for Merck and Roche speaker fees for Bayer. All other authors report no conflicts.
46
47
48 Support: Research reported in this publication was supported by the National Cancer Institute of
49 the National Institutes of Health under Award Numbers U10CA180821 and U10CA180882 (to
50 the Alliance for Clinical Trials in Oncology), UG1CA232760, UG1CA233184, UG1CA233253,
51 UG1CA233290, UG1CA233324, , U10CA180863, CCS 707213 (CCTG); and U10CA180888
52 (SWOG). The content is solely the responsibility of the authors and does not necessarily
53 represent the official views of the National Institutes of Health.
54 https://acknowledgments.alliancefound.org
55 ClinicalTrials.gov Identifier: NCT00499330
of
56
57
ro
58
59 Abstract
-p
60 Objective re
61 We have recently reported the primary results of CALGB 140503 (Alliance), a randomized trial
62 in patients with peripheral cT1aN0 NSCLC (AJCC 7th) treated with either lobar (LR) or sublobar
lP
63 resection (SLR). Here we report differences in disease-free survival (DFS), overall survival (OS)
64 and lung cancer specific survival (LCSS) between LR, segmental (SR) and wedge resections
na
65 (WR). We also report differences between WR and SR in surgical margins, rates of locoregional
66 recurrence (LRR) and expiratory flow rates at 6 months postoperatively.
ur
67 Methods
Jo
68 Between 6/2007 and 3/2017, 697 patients were randomized to LR (357) or SLR (340) stratified
69 by clinical tumor size, histology and smoking history. Ten patients were converted from SLR to
70 LR and 5 from LR to SLR. Survival end points were estimated by the Kaplan–Meier estimator
71 and tested by the stratified Log rank test. Kruskal-Wallis testing was used to compare margins
72 and FEV1 changes between groups; and a Chi-square test was used to test the association
73 between recurrence and groups.
74 Results
75 A total of 362 patients had LR, 131 had SR and 204 had WR. Basic demographic and clinical
76 and pathological characteristics were similar between all three groups. Five-year DFS was 64.7%
77 after LR [95% CI; 59.6-70.1%], 63.8% after SR [ 95% C; 55.6 − 73.2%] and 62.5% after WR
78 [95% CI; 55.8 − 69.9%] (Log rank, p = 0.888). Five-year OS was 78.7% after LR, 81.9% after
79 SR and 79.7% after WR (Log rank, p = 0.873). Five-year LCSS was 86.8% after LR, 89.2% after
80 SR and 89.7% after WR (Log rank, p = 0.903). LRR occurred in 12% after SR and 14% after
81 WR (p=0.295). At 6 months postoperatively, the median reduction in % FEV1 was 5% after WR
82 and 3% after SR (p=0.930)
83 Conclusions
84 In this large, randomized trial, LR, SR and WR were associated with similar survival outcomes.
85 Although LRR was numerically higher after WR compared to SR, the difference was not
86 clinically meaningful statistically significant. There was no significant difference in the reduction
87 of FEV1 between the SR and WR groups.
of
88
ro
89
90
91
-p
re
92
lP
93
94
na
95
96
ur
97
98
Jo
99
100
101
102
103
104
105
106
107
108
109
110 Key words: lung cancer, randomized trial, sublobar resection
111 Graphical abstract
of
ro
-p
re
lP
112
na
113 Central Picture

ur
Jo
114
115
116
117
118 Central message
119 In this post-hoc unplanned analysis of CALGB 140503, patients treated by lobectomy,
120 segmentectomy or wedge resections had similar survival outcomes.
121 Perspective statement
122 Recent trials showed that survival after sub-lobar resection (SLR) was non-inferior to that after
123 lobar resection for patients with clinical T1aN0 lung cancer. In CALGB 140503 both wedge and
of
124 segmental resections were acceptable options of SLR. In this unplanned analysis, we show that
ro
125 survival was similar regardless of the extent of parenchymal resection.
126
-p
re
127
lP
128
na
129
130
ur
131
Jo
132
133
134
135
136
137
138
139
140
141 Glossary of abbreviations
142 AJCC: American Joint Committee on Cancer
143 LR: lobar resection
144 SLR: Sublobar resection
145 DFS: Disease-free survival
146 OS: Overall survival
of
147 LCSS: Lung cancer specific survival
ro
148 SR: Segmental resection
149 WR: Wedge resection

-p
re
150 LRR: locoregional recurrence
lP
151 FEV1: Forced expiratory volume in 1 second

na
152 FVC; Forced vital capacity
153 DSMB: Data safety and monitoring board

ur
154 SDMC: Statistical and data management center

Jo
155 CALGB: Cancer and Leukemia Group B
156 NCI: National cancer institute
157 HR: Hazard ratio
158 CI: Confidence interval
159 CT: computerized tomography
160
161
162
163
164 Introduction
165 In 1982, the North American Lung Cancer Study Group (LCSG) initiated a prospective,
166 randomized trial comparing lobar resection (LR) to sub-lobar resection (SLR) in patients with
167 peripheral stage I non-small cell lung cancer (NSCLC)1. Two-hundred and seventy-six patients
168 who had clinical T1N0 peripheral tumors measuring 3 centimeters (cm) or less in all dimensions
169 on posteroanterior and lateral chest roentgenogram were randomly assigned to lobectomy or
of
170 limited resection by either wedge or anatomical segmentectomy. Routine computerized
ro
171 tomography was not required except if prompted by an abnormal history or physical examination
172
-p
or abnormal blood chemistries. After a minimum follow-up of 4.5 years in 247 patients eligible
re
173 for analysis, the investigators reported that in comparison to LR, SLR was associated with a
lP
174 significantly higher incidence of local recurrence, a 30% increase in overall mortality and a 50%
na
175 higher probability of lung cancer-related deaths. Lobectomy has since become the standard of
176 care for patients with clinical stage I NSCLC who had adequate cardiopulmonary function.
ur
177 Prompted by significant advances in imaging technology and refinements in TNM staging, in
Jo
178 2007 the Cancer and Leukemia Group B co-operative group (now part of the Alliance for
179 Clinical Trials in Oncology) and later the Japanese Clinical Oncology Group (JCOG) initiated a
180 non-inferiority randomized trial comparing LR to SLR in patients with clinical T1aN0 2 cm or
181 less in size2,3. The JCOG trial randomized 1100 patients to either lobar resection or anatomical
182 segmentectomy2. After a median follow-up of seven years JCOG investigators reported that
183 anatomical segmentectomy was non-inferior to lobectomy for the primary endpoint of overall
184 survival2. In CALGB140503, 697 patients were randomly assigned to either lobectomy or sub-
185 lobar resection. In the North American trial both anatomical segmentectomy and wedge resection
186 were considered acceptable modalities of sub-lobar resection. The use of wedge resection,
187 though controversial, was allowed since it had been and remains the most common type of sub-
188 lobar resection reported in National and Society registries4,5,6,7.
189 We recently reported that after a median follow-up of 7 years, the primary endpoint of disease-
190 free survival (DFS) was not inferior after SLR resection to that after lobectomy (HR 1.01,
191 90%CI: 0.83, 1.24)3. Five-year DFS after SLR was 63.6% (95%CI: 57.9%, 68.8%) and 64.1%
192 after LR (95%CI: 58.5%, 69.0%)3. The key secondary endpoint of overall survival (OS) was
of
193 similar in both arms of the trial (HR 0.95; 95%CI; 0.72, 1.26). Five-year OS was 80.3% (95%CI;
ro
194 75.5%, 84.3%) after SLR and 78.9% (95%CI: 74.1%, 82.9%) after LR. In this unplanned post-
195
-p
hoc analysis, we report differences in DFS, OS and lung cancer specific survival (LCSS)
re
196 between LR, segmentectomy (SR) and wedge resections (WR). We also report differences
lP
197 between WR and SR in surgical margins, rates of locoregional recurrence (LRR) and expiratory
na
198 flow rates at 6 months postoperatively.
199 Methods
ur
200 Study design and participants

Jo
201 Cancer and Leukemia Group B (CALGB) 140503 was a multi-center, international, randomized,
202 non-inferiority phase III trial in patients with NSCLC clinically staged as T1aN0. Clinical
203 staging was based on the 7th edition of the TNM staging system. Patients were recruited from 83
204 academic and community-based institutions in the United States, Canada and Australia. Patients
205 were registered to the trial if they met the preoperative eligibility criteria and were randomized
206 after meeting intraoperative eligibility criteria. Preoperative eligibility criteria included the
207 presence of a peripheral lung nodule with a solid component 2 cm or less in size on preoperative
208 computerized tomography (CT) that is presumed or confirmed to be NSCLC; center of the
209 tumor, as seen on CT scanning, located in the outer third of the lung, tumor location suitable for
210 either LR or SLR (wedge or segment); Eastern Cooperative Oncology Group performance status
211 score of 0–2; no malignant disease within the previous 3 years other than non-melanoma skin
212 cancer, superficial bladder cancer, or carcinoma in situ of the cervix; no previous chemotherapy
213 or radiation therapy for the index lung cancer; no evidence of locally advanced or metastatic
214 disease; and age 18 years or older. Patients with pure ground-glass opacities or pathologically
215 confirmed N1 or N2 disease were not eligible.
of
216 Intraoperative eligibility criteria included histological confirmation of non-small-cell lung cancer
ro
217 (if not already obtained) and confirmation of N0 status by frozen section examination (for right-
218
-p
sided tumors, node levels 4, 7, and 10; for left-sided tumors, node levels 5 or 6, 7, and 10); nodes
re
219 previously sampled by mediastinoscopy, endobronchial ultrasound, or endoscopic ultrasound
lP
220 within 6 weeks prior to the definitive surgical procedure did not need to be resampled.
na
221 The trial was conducted according to the principles of the Declaration of Helsinki and the
222 International Council for Harmonisation for Good Clinical Practice guidelines. The protocol was
ur
223 approved by the institutional review board at each participating institution (IRB number:
Jo
224 0709009395 and date: June 15, 2007). All patients provided written informed consent before trial
225 enrolment. Since activation of CALGB 140503, the trial was monitored by the Alliance Data and
226 Safety Monitoring Board (DSMB) twice yearly.
227 Randomization
228 Eligible patients were registered to the trial using the OPEN registration system, a web-based
229 system for patients’ enrolment into National Cancer Institute (NCI)-sponsored cooperative
230 groups clinical trials. Once intra-operative eligibility, as described above, was confirmed,
231 patients underwent randomization (1:1) to either LR or SLR, based on a permuted-block
232 randomization scheme with stratification for radiographic tumour size (<1 cm, 1–1.5 cm, and
233 >1.5–2.0 cm), histology (squamous cell carcinoma, adenocarcinoma, and other), and smoking
234 status (never, former, and current). Assignment was not concealed from patients, surgeons,
235 nurses, data managers, or statisticians.
236 Procedures
237 The type of SLR (wedge resection or segmentectomy) was not randomly assigned but left at the
238 discretion of the surgeon. When performing a WR, a 2 cm margin around the tumor was
of
239 recommended or a margin equivalent to the greatest diameter of the primary tumor. This margin
ro
240 recommendation was arbitrarily determined by the trial investigators. The length of the margin
241
-p
was estimated by the surgeon in the deflated lung and noted in the operative record. Absence of
re
242 disease at the wedge margin as assessed by frozen section examination (histologically or
lP
243 cytologically) was strongly encouraged. If a frozen section examination of the wedge margin was
na
244 not performed, the surgeon had to ensure a 1.5 to 2 cm margin around the tumor. Patients with a
245 positive wedge margin on frozen section examination could either be have an additional WR,
ur
246 have an anatomic segmentectomy if technically possible or a LR if necessary. Mediastinal nodal

Jo
247 staging was required either by simple sampling (minimal requirement for randomization as
248 described above), systematic sampling (sampling of 3 or more mediastinal nodal stations) or
249 complete dissection of the mediastinal lymph nodes as per standard institutional practices .
250 Outcomes
251 The primary objective of this post-hoc analysis was to compare DFS, OS and LCSS after
252 lobectomy, anatomical segmentectomy and WR. Disease-free survival was defined as any
253 disease recurrence or death from any cause, whichever occurred first. Overall survival was
254 defined as the time between randomization and death from all causes. Lung cancer specific
10
255 survival was defined as the time between randomization and death from lung cancer; all other
256 events, including recurrence and death from any cause, were censored at the occurrence of these
257 events. Secondary objectives included comparisons of resection margins, incidence of loco-
258 regional and systemic recurrence, and expiratory flow rates 6 months postoperatively between
259 the segmentectomy and wedge resection groups. For the purpose of this trial, locoregional
260 recurrence was defined as recurrent disease in the lung parenchyma and/or the hilar nodes of the
261 index lobe. Regional recurrence was defined as isolated mediastinal nodal recurrence. All other
of
262 recurrence was deemed systemic.
ro
263 Statistical analysis
264
-p
The trial was designed to have approximately 80% power with 351 events to reject the null
re
265 hypothesis that the hazard ratio of sub-lobar resection over lobectomy is less than Δ=1.306 by
lP
266 stratified Log rank test for non-inferiority at one-sided significance level of 5% when the true
na
267 hazard ratio is 1. Δ denotes the pre-specified non-inferiority margin, for which there is 5%
268 chance that the null hypothesis is rejected when the hazard rate of sub-lobar resection is 30.6%
ur
269 higher than lobectomy. The trial accrued 697 patients between June 2007 and March 2017.
Jo
270 Based on interim analyses conducted up to November 2021 and a validation analysis in March
271 2022, the Alliance DSMB recommended unanimously to release the data, and terminate further
272 DSMB monitoring of the trial.
273 While the primary analysis of efficacy endpoints was based on the intent-to-treat (ITT)
274 population2, the current post-hoc analysis is based on surgical procedures actually performed
275 regardless of treatment assignment. Survival endpoints were characterized using the Kaplan-
276 Meier estimator. The p-values for testing the survival function difference between surgical
277 procedures for DFS, OS and LCSS were obtained from the stratified Log rank test with tumor
11
278 size, histology and smoking status as stratification factors. Hazard ratios (HRs) and their
279 confidence intervals (CIs) were estimated using stratified Cox Proportional Hazards models.
280 Violation of the proportional hazards assumption was evaluated by the method of Schoenfeld
281 residuals(8).
282 Cumulative incidence functions for lung cancer related deaths and competing causes of deaths
283 after LR, SR and WR were compared using the Gray method and the associated hazard ratios
284 (HRs) and confidence intervals (CIs) estimated by the Fine-Gray sub-distribution hazard
of
285 model9,10. Disease recurrence rates were summarized by surgical procedure, and an association
ro
286 between recurrence and procedure was tested through the Chi-square test. Similariyl, Chi-square
287
-p
test was used to test the association between positive surgrical margin with surgical procedures.
re
288 The comparison of resection margins and changes in pulmonary functions (forced expiratory
lP
289 volume in 1 second [FEV1] and forced vital capacity [FVC]), between WR and SR were tested
na
290 with the Kruskal–Wallis test. Differences in the rate of positive surgical margins between WR
291 and SR was tested with Fisher’s Exact test.

ur
292 Data quality was ensured by review of data by the Alliance Statistical and Data Management
Jo
293 Center (SDMC) and the study chairperson (NA), following Alliance policies. The efficacy
294 endpoints, including disease-free and overall survival, have been independently validated by an
295 Alliance SDMC statistician, who is not associated with the study. All statistical analyses were
296 conducted by the study statisticians and statistical programmers with the data lock down on June
297 21, 2022. Data management and statistical analysis were done in SAS 9.4 and graphs were
298 generated in R 3.6.3. The corresponding author and CALGB/Alliance statistician developed the
299 trial design, had full access to the raw data, and analyzed the data. The first author wrote the first
300 draft of the manuscript. All authors had the opportunity to revise the paper and vouch for the
12
301 data, the analysis, and adherence to the protocol. This trial is registered with ClinicalTrials.gov,
302 number NCT00499330.
303 Role of the funding source
304 The funder approved the trial design but had no role in data collection, data interpretation, data
305 analysis, or writing of the report. There were no agreements concerning confidentiality of the
306 data between the funding source and the authors or the participating institutions.
307
of
308 Results
ro
309 Between June 15, 2007, and March 13, 2017, 1080 patients with suspected or confirmed T1aN0
310
-p
non-small-cell lung cancer were preregistered to the trial by 125 surgeons at 83 participating
re
311 institutions. A total of 697 (65%) patients met preoperative and intraoperative eligibility criteria
lP
312 and were randomly assigned to either LR (n=357) or SLR (n=340)). Among 340 patients
na
313 assigned to SLR, 201 (59%) underwent wedge resection and 129 (38%) had an anatomical
314 segmental resection. Ten patients (3%) were converted from SLR to LR and five (1.4%) from
ur
315 LR to SLR. In aggregate, based on procedures actually performed, 362 patients had a lobectomy,
Jo
316 131 had anatomical segmentectomy and 204 patients had a wedge resection.
317 Baseline demographic and clinical and pathological characteristics were generally similar
318 between all three groups particularly regarding the three prespecified stratification variables
319 (Table 1). Male patients were slightly over-represented among patients receiving wedge
320 resections. Compared to patients who had lobar or segmental resections, patients receiving a WR
321 were more likely to have had either simple nodal sampling, the minimal nodal assessment
322 required for randomization or mediastinal node dissection. Overall, mediastinal nodal assessment
13
323 by either dissection, systematic sampling or simple sampling was performed in nearly all
324 patients.
325
326 Survival
327 Five-year DFS was 64.7% after LR [95% CI; 59.6-70.1%], 63.8% after SR [ 95% C; 55.6 −
328 73.2%] and 62.5% after WR [95% CI; 55.8 − 69.9%] (Log rank, p = 0.888) (Figure IA, Figure
329 4). Five-year OS was 78.7% [95%CI;74.5-83.2%] after LR, 81.9% [95%CI;75.3-89.1%] after SR
of
330 and 79.7% after WR [95%CI;74.2-85.6%] (Log rank, p = 0.873) (Figure IB, Figure4). LCSS was
ro
331 86.8%% after LR, 89.2% after SR and 89.7% after WR (Log rank, p = 0.903) (Figure IC). We
332
-p
had previously reported that the cumulative incidence of deaths from lung cancer and other
re
333 causes of death was similar between lobar and sub-lobar resection3. In the current analysis, there
lP
334 were 19 lung cancer related deaths in the SR group and 28 among patients who had a wedge
na
335 resection. There were no differences in the cumulative incidence of lung cancer (p=0.694) or
336 competing causes of death (p=0.448) between the two groups.

ur
337 Recurrence
Jo
338 Overall recurrence developed in 31.5% after SR and 31.2% after WR. Distant relapse occurred in
339 16.9% and 14.9% after SR and WR resection respectively. Locoregional recurrence occurred in
340 12.3% after SR and 14.4% after WR (p=0.295). We had previously reported that locoregional
341 recurrence occurred in 10% of patients after LR (3). Freedom from locoregional recurrence at five
342 years was 86.6% [95%CI;82.6-90.8%] after LR, 84.6% [95%CI;77.9-91.9%] after SR and 82.7%
343 [95%CI;77.0-88.9%] after WR (Log rank p=0.232) (Figure 3).
344 Resection Margins
14
345 Information on surgical margins was available for 136 patients after WR (66%) and 76 patients
346 after SR (58%). Median margin length was 1.6 cm after WR and 2.0 cm after SR (p=0.030). The
347 proportion of patients within each category of margin length in the WR and SR group is shown
348 in Table 2. The majority of patients in either group had margins exceeding 1 cm in length. The
349 median margin to clinical tumor ratio was 1.2 after WR and 1.3 after SR (p=0.070). The
350 proportions of patients within each category of margin to clinical tumor ratio in the WR and SR
351 groups are shown in Table 2. Patients who had a wedge resection were more likely to have a
of
352 margin to tumor ratio less than one compared to those who had a SR.
ro
353 The median margin length and margin to clinical tumor ratio in each tumor size category (<1 cm,
354
-p
1-1.5 cm,>1.5 cm) for patients who had WR and SR are shown in Table 3. In patients who had a
re
355 wedge resection, the median length of the margin was 1.7 cm for tumor < 1 cm in size, 1.5 cm
lP
356 for tumors 1.0-1.5 cm in size and 2 cm for tumors exceeding 1.5 centimeters in diameter.
na
357 Intraoperative histological or cytological assessment of the surgical margin by frozen section
358 examination (FS) was performed in 281 of 335 patients (84%) after SLR. Frozen section
ur
359 assessment of the margin was done in 178 patients after WR (87%) and 103 patients after
Jo
360 segmentectomy (76%). Eight patients who had a positive margin detected by FS after wedge
361 resection (4.4%) were treated by an immediate additional wedge resection (n=5) or an immediate
362 lobectomy (n=3). Two patients had a positive margin on FS examination after segmentectomy
363 (1.5%); one had an extended segmentectomy and the other had a lobectomy, both performed
364 during the index operation. None of these 10 patients developed subsequent disease recurrence.
365 Overall, a positive surgical margin by either frozen or permanent section examination was
366 present in 3 patients after LR (0.8%), 2 patients after SR (1.5%) and 10 patients after WR
367 (Fisher’s exact, p=0.007).
15
368 Changes in expiratory flow rates
369 At 6 months postoperatively, the median reduction from baseline in percent predicted FEV1 was
370 5% after WR and 3% after SR (p=0.930) and the median reduction from baseline in percent
371 predicted FVC was 3% after segmentectomy and 4% after wedge resection (Table 4).
372 Discussion
373 In this post-hoc analysis of CALGB 140503, we found that disease-free and overall survival, the
374 primary and secondary endpoints, were similar regardless of the extent of pulmonary resection.
of
375 Specifically, we found that there was no clinically meaningful difference in DFS, OS and LCSS
ro
376 between patients treated by lobectomy, and those treated by either wedge resection or anatomical
377
-p
segmentectomy. Although these results are reassuring and consistent with the reported primary
re
378 results of the trial3, they should be interpreted with caution since the trial was not powered to
lP
379 determine non-inferiority in survival outcomes between the two modalities of sub-lobar
na
380 resection.
381 We had also previously reported that there was no difference in the cumulative incidence of lung
ur
382 cancer related and competing causes of death between patients treated by lobectomy and those
Jo
383 treated by sublobar resection3. Here we found that although the cumulative incidence of lung
384 cancer related deaths appeared numerically higher after SR compared to WR, the difference was
385 not statistically significant. Additionally, we observed that after both wedge and segmental
386 resection, the risk of lung cancer related deaths appeared to sharply increase between years 2 and
387 4 postoperatively, followed by a small residual risk that persisted beyond five years of follow-up.
388 We had previously reported that recurrence rates were similar between patients treated by lobar
389 resection and those in whom a sub-lobar resection was performed3. Here we report that there
390 were no significant intergroup differences in systemic or locoregional recurrence rates between
16
391 the two modalities of sub-lobar resection and that at least 50% of recurrences were systemic in
392 nature.
393 Although, the resection margin after segmentectomy was predictably longer compared to that
394 after wedge resection, the difference in locoregional recurrence between WR and SR was neither
395 statistically significant nor clinically meaningful. Freedom from loco-regional recurrence was
396 84.6% after SR and 82.7% after WR. For patients having a wedge resection, the trial protocol
397 had recommended a resection margin of two centimeters or one equivalent to the diameter of the
of
398 tumor. It is reassuring that the median resection margin after wedge resection was 1.6
ro
399 centimeters which compares favorably with the median clinical tumor size of 1.5 centimeters in
400
-p
the WR group. Furthermore, more than 48% of patients in the WR group had a resection margin
re
401 at least 2 centimeters in length. Overall, a positive margin was present in 4.9% patients who had
lP
402 a WR compared to 1.5% of patients who had a SR. Ten of twelve patients who underwent SLR
na
403 and had a positive margin, had their positive margins detected on frozen section examination and
404 remedied by additional parenchymal resection.

ur
405 Counterintuitively, the reduction in expiratory flow rates at 6 months postoperatively was similar
Jo
406 between the WR and the SR group. These results appear consistent with the previously reported
407 trial results and suggest the potential limitations of expiratory flow rates as an appropriate metric
408 for assessment of pulmonary function following lung resection3.
409 Overall, the findings from this post-hoc analysis appear to challenge the long-held concept that
410 wedge resections are inherently suboptimal oncological procedures. Our data showed no
411 clinically meaningful intergroup differences in either survival outcomes or in the incidence of
412 disease recurrence. However, these results are only applicable to a highly selected group of
413 patients who present with peripheral tumors two centimeters or less in size and in whom the
17
414 absence of nodal disease had been confirmed by pathological examination of lymph nodes from
415 at least one major hilar and two mediastinal nodal stations. Additionally, the trial protocol
416 emphasized the need for obtaining an “adequate” resection margin when preforming a wedge
417 resection and strongly encouraged intraoperative pathological assessment of the resection
418 margin. To that end, an “adequate” margin was arbitrarily defined by the investigators as a two-
419 centimeter margin around the tumor or a margin equivalent to the clinical tumor size.
420 Our findings show that these recommendations are appropriate as they were associated with an
of
421 acceptably low incidence of positive surgical margin following wedge resection and similar
ro
422 survival outcomes to those following anatomic resections.
423
-p
The findings of this study, particularly those related to wedge resection, will inevitably raise
re
424 questions regarding the role of competing ablative strategies particularly stereotactic whole body
lP
425 radiotherapy (SBRT). To our knowledge, there are no data from randomized trials supporting
na
426 the use of SBRT in this patient population or indeed in patients with NSCLC who are otherwise
427 good surgical candidates. Two randomized trials compared lobar resection with SBRT in a
ur
428 surgically fit population, both of which were terminated prematurely after randomizing less than
Jo
429 3% of their target accrual (STARS trial [NCT00840749];ROSEL trial [NCT00687986]). Chang
430 et al pooled all 58 patients randomized in both trials and reported their outcomes(11). Recurrence-
431 free survival at 3 years was 86% for 31 patients treated by SBRT and 80% in 27 patients treated
432 by lobar resection(11). Recurrent disease developed in 3 patients in the surgery group and 6
433 patients in the SBRT group. Interestingly, 4 of 6 patients who developed recurrence after SBRT
434 had nodal recurrence. In our opinion, the role of SBRT in this cohort of patients remains
435 controversial.
18
436 There are three important limitations to the results of this post-hoc analysis. First, the trial was
437 not powered to test for non-inferiority between the two modalities of sub-lobar resection and
438 therefore the results should be construed as hypothesis generating rather than the result of
439 hypothesis testing. Additionally, patients were not randomly assigned to the type of sublobar
440 resection performed. Although we found no difference between the two modalities of sublobar
441 resection in any of the examined survival endpoints, it is possible that smaller differences in
442 survival might emerge should a randomized trial with a larger sample size be performed.
of
443 Accordingly, we make no recommendations for or against a specific type of sub-lobar resection
ro
444 but suggest that either modality is acceptable within the constraints defined by the trial. Second,
445
-p
although our margin data are unique since they are collected in a controlled prospective trial
re
446 setting, information on the surgical margin was available from only 63% of patients treated by
lP
447 sub-lobar resection. However, given the overall trial results and the high rate of locoregional
na
448 control after WR and SR, we are confident that these margin results are representative of all
449 patients who had sub-lobar resection. Finally, we are unable to evaluate the impact of
ur
450 pathological variables such as adenocarcinoma subtypes, lympho-vascular invasion or spread

Jo
451 through the airspaces on outcome. Realistically, however, logistical challenges posed by the
452 large number of participating centers and the extended duration of the trial, precluded central
453 pathology review. Additionally, some of these variables were either not identified (STAS and
454 adenocarcinoma subtypes) or were not thought to be of prognostic value at the initiation of the
455 trial.
456 In summary, this post-hoc analysis shows that there are no meaningful differences in survival
457 based on the extent of parenchymal resection and that both wedge resection and anatomical
458 segmentectomy are acceptable treatment options for patients with peripheral T1a non-small cell
19
459 lung without metastases to the hilar and mediastinal nodes.. However, it is important to once
460 again reiterate that the trial was adequately powered to test for noninferiority of sublobar
461 resection compared to lobar resection for the primary end point of disease-free survival. With a
462 hazard ratio (sublobar resection vs. lobar resection) of 1.01 and a p-value for noninferiority of
463 0.02, noninferiority for the primary endpoint was established. The results of all comparisons
464 beyond the primary end point, including the results of the current study, should be regarded as
465 hypothesis-generating rather than hypothesis-testing. Neither p-values nor the widths of the
of
466 confidence intervals were adjusted for multiplicity and may not be used in place of hypothesis
ro
467 testing. That said, in the absence of further randomized trials, our results should be considered
468
-p
the best available evidence suggesting that in this specific cohort of patients, there is no
re
469 difference in oncological otcomes based on the extent of parenchymal resection.
lP
470
na
471
472 References
ur
473 1- Ginsberg RJ, Rubinstein LV; Lung Cancer Study Group. Randomized trial of lobectomy
Jo
474 versus limited resection for T1 N0 non-small cell lung cancer. Ann Thorac Surg 1995; 60:
475 615-23.
476 2- Saji H, Okada M, Tsuboi M, Nakajima R, Suzuki K, Aokage K, Aoki T, Okami J, Yoshino I,
477 Ito H, Okumura N, Yamaguchi M, Ikeda N, Wakabayashi M, Nakamura K, Fukuda H,
478 Nakamura S, Mitsudomi T, Watanabe SI, Asamura H; West Japan Oncology Group and
479 Japan Clinical Oncology Group. Segmentectomy versus lobectomy in small-sized peripheral
480 non-small-cell lung cancer (JCOG0802/WJOG4607L): a multicentre, open-label, phase 3,
481 randomised, controlled, non-inferiority trial. Lancet. 2022 Apr 23;399(10335):1607-1617
20
482 3- Altorki N, Wang X, Kozono D, Watt C, Landrenau R, Wigle D, Port J, Jones DR, Conti M,
483 Ashrafi AS, Liberman M, Yasufuku K, Yang S, Mitchell JD, Pass H, Keenan R, Bauer T,
484 Miller D, Kohman LJ, Stinchcombe TE, Vokes E. Lobar or Sublobar Resection for Peripheral
485 Stage IA Non-Small-Cell Lung Cancer. N Engl J Med. 2023 Feb 9;388(6):489-498.
486 4- Mery CM, Pappas AN, Burt BM, Bueno R, Linden PA, Sugarbaker DJ, Jaklitsch MT.
487 Diameter of non-small cell lung cancer correlates with long-term survival: implications for T
488 stage. Chest. 2005 Nov;128(5):3255-60
of
489 5- Speicher PJ, Gu L, Gulack BC, Wang X, D'Amico TA, Hartwig MG, Berry MF. Sublobar
ro
490 Resection for Clinical Stage IA Non-small-cell Lung Cancer in the United States.Clin Lung
491 Cancer. 2016 Jan;17(1):47-55

-p
re
492 6- Servais EL, Blasberg JD, Brown LM, Towe CW, Seder CW, Onaitis MW, Block MI, David
lP
493 EA. The Society of Thoracic Surgeons General Thoracic Surgery Database: 2022 Update on
na
494 Outcomes and Research. Ann Thorac Surg. 2023 Jan;115(1):43-49
495 7- Salati M, Brunelli A, Decaluwe H, Szanto Z, Dahan M, Varela G, Falcoz PE; ESTS DB
ur
496 Committee. Report from the European Society of Thoracic Surgeons Database 2017: patterns
Jo
497 of care and perioperative outcomes of surgery for malignant lung neoplasm. Eur J
498 Cardiothorac Surg. 2017 Dec 1;52(6):1041-1048
499 8- Schonfeld D. Partial residuals for the proportional hazards model. Biometrika. 1982;69:238–
500 41
501 9- Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing
502 risk. Ann Stat 1988; 16: 1141-54.
503 10- Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J
504 Am Stat Assoc 1999; 94: 496-509.
21
505 11- Chang JY, Mehran RJ, Feng L, Verma V, Liao Z, Welsh JW, Lin SH, O'Reilly MS, Jeter
506 MD, Balter PA, McRae SE, Berry D, Heymach JV, Roth JA; STARS Lung Cancer Trials
507 Group. Stereotactic ablative radiotherapy for operable stage I non-small-cell lung cancer
508 (revised STARS): long-term results of a single-arm, prospective trial with prespecified
509 comparison to surgery. Lancet Oncol. 2021 Oct;22(10):1448-1457
510
511
512
513
514
of
515
516
ro
517
518
-p
519
520
re
521
522
lP
523
524
525
na
526
527
ur
528
529
530
Jo
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
22
546
547
548
549
550
551
552 Table 1: Baseline demographic, clinical and pathological characteristics
Lobectomy Wedge Resection Segmentectomy Total
P-value
(N= 362) (N= 204) (N= 131) (N=697)
of
Age 0.83781
Median (Years) 67.7 68.5 66.9 67.9
ro
Gender, n (%) 0.03382
Male 146 (40.3%) 102 (50.0%) 49 (37.4%) 297 (42.6%)
-p
Female 216 (59.7%) 102 (50.0%) 82 (62.6%) 400 (67.4%)
ECOG Performance Status, n (%) 0.08382
0 255 (70.4%) 153 (75.0%) 105 (80.2%) 513 (73.6%)
re
1/2 107 (29.6%) 51 (25.0%) 26 (19.8%) 184 (26.4%)
Maximum diameter of tumor using
0.97441
lung window settings (cm)
lP
Median 1.5 1.5 1.5 1.5

Tumor Size (cm), n (%) 0.94582
<1.0 30 (8.3%) 19 (9.3%) 9 (6.9%) 58 (8.3%)
na
1.0-1.5 182 (50.3%) 103 (50.5%) 69 (52.7%) 354 (50.8%)

>1.5-2.0 150 (41.4%) 82 (40.2%) 53 (40.5%) 285 (40.9%)
ur
Smoking Status, n (%) 0.30112

Never 37 (10.2%) 13 (6.4%) 13 (9.9%) 63 (9.0%)
Former 180 (49.7%) 98 (48.0%) 71 (54.2%) 349 (50.1%)
Jo
Current 145 (40.1%) 93 (45.6%) 47 (35.9%) 285 (40.9%)

Histology, n (%) 0.84502
Squamous Cell Carcinoma 52 (14.4%) 30 (14.7%) 16 (12.2%) 98 (14.1%)
Adenocarcinoma 229 (63.3%) 126 (61.8%) 89 (67.9%) 444 (63.7%)
Other 81 (22.4%) 48 (23.5%) 26 (19.8%) 155 (22.2%)
Tumor location area, n (%) 0.22992
Right upper lobe 130 (35.9%) 73 (35.8%) 45 (34.4%) 248 (35.6%)
Right middle lobe 17 (4.7%) 14 (6.9%) 4 (3.1%) 35 (5.0%)
Right lower lobe 44 (12.2%) 33 (16.2%) 21 (16.0%) 98 (14.1%)
Left upper lobe 102 (28.2%) 58 (28.4%) 30 (22.9%) 190 (27.3%)
Left lower lobe 66 (18.2%) 25 (12.3%) 28 (21.4%) 119 (17.1%)
Lingula 3 (0.8%) 1 (0.5%) 3 (2.3%) 7 (1.0%)
Surgical approach, n (%) 0.66592
Thoracotomy 49 (13.5%) 24 (11.8%) 20 (15.3%) 93 (13.3%)
VATS 291 (80.4%) 169 (82.8%) 100 (76.3%) 560 (80.3%)
VATS, conversion 22 (6.1%) 11 (5.4%) 11 (8.4%) 44 (6.3%)
Mediastinal node staging 0.02201
Complete dissection 96 (26.7%) 58 (28.7%) 28 (21.4%) 182 (26.3%)
23
Systematic sampling 185 (51.4%) 84 (41.6%) 80 (61.1%) 349 (50.4%)
Simple sampling 79 (21.9%) 60 (29.7%) 23 (17.6%) 158 (22.8%)
Patients with positive nodes 0.04831
0 338 (93.4%) 199 (97.5%) 127 (96.9%) 664 (95.3%)
1+ 24 (6.6%) 5 (2.5%) 4 (3.1%) 33 (4.7%)
Mean number of nodes sampled (SD) 4.7 (1.45) 4.0 (1.30) 4.9 (1.46) 4.6 (1.46) <.00013
553
554
of
ro
-p
re
lP
na
ur
Jo
24
Figure 1: Disease-free survival (A); Overall survival (B); Lung Cancer Specific Survival (C)
555
(A)
of
(B)
ro
-p
re
lP
na
ur
Jo
(C)
25
556 Figure 2: Cumulative incidence of lung cancer and competing causes of deaths
557 between SR and WR.
558
559
560
561
562
563
564
565
566
567
of
568
569
ro
570
571
-p
572
573
574
re
575
lP
576
577
578 Figure 3: Locoregional recurrence-free survival for WR and SR
na
579
ur
Jo
580
581
582
583
26
584
585
586
587
588
589
590
591
of
ro
-p
re
lP
na
ur
Jo
27
592 Figure 4: Graphical abstract
of
ro
-p
re
593
594
lP
595
596
na
597
598
599
ur
600
Jo
601
602
603
604
605
606
607
608
609
610
611
612
613
614
28
615 Table 2. Margin length and margin to tumor ratio after wedge and segmental
616 resection
Wedge Resection Segmentectomy
P-value
(N=136) (N= 76)
Surgical margin length (cm)

Median 1.6 2 0.0311
Size of surgical margin (cm), n (%) 0.2542
< 1 cm 23 (16.9%) 11 (14.5%)
1-<2 cm 47 (34.6%) 16 (21.1%)
2-<3 cm 39 (28.7%) 26 (34.2%)
of
3-<4 cm 15 (11.0%) 13 (17.1%)
4-<6 cm 11 (8.1%) 9 (11.8%)
ro
6+ cm 1 (0.7%) 1 (1.3%)
Margin/clinical tumor ratio
-p
Median re 1.2 1.3 0.0701
Distribution of margin/tumor ratio 0.1622
<1 52 (38.2%) 17 (22.7%)
lP
1-<2 54 (39.7%) 38 (50.7%)

2-<3 20 (14.7%) 15 (20.0%)
na
3-<4 4 (2.9%) 3 (4.0%)

4-<6 6 (4.4%) 2 (2.7%)
ur
617 1 2
Kruskal-Wallis p-value; Chi-Square p-value
618
Jo
29
619 Table 3: Margin length and margin/tumor ratio for different tumor size categories
Wedge resection
Tumor size (cm)
<1.0 1.0-1.5 >1.5-2.0

P-value
(N=19) (N=103) (N=82)
Median clinical tumor size 0.9 1.3 1.8 <.00011
Surgical margin 0.58731
N (Missing) 10 (9) 69 (34) 57 (25)
Median (cm) 1.7 1.5 2
Margin/clinical tumor size 0.00811

N (Missing) 10 (9) 69 (34) 57 (25)
of
Median (cm) 2.3 1.3 1
ro
Segmentectomy
Tumor size (cm)
-p
<1.0 1.0-1.5 >1.5-2.0
P-value
(N=9) (N=69) (N=53)
re
Median clinical tumor size 1.2 1.3 1.8 <.00011
lP
Surgical margin 0.89351
N (Missing) 6 (3) 37 (32) 33 (20)
Median (cm) 2.5 2 2
na
Margin/clinical tumor size 0.03401

N (Missing) 6 (3) 36 (33) 33 (20)
ur
Median (Range) 2.5 1.5 1.1
620 1
Kruskal-Wallis p-value
Jo
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
30
637
Table 4: Change in flow rates from baseline 5-7 months postoperatively
Wedge
Segmentectomy
resection P-value
N=81 N=87
Change in Predicted FEV1 (%)
Median -3% -5% 0.93041

Change in Predicted FVC (%)
Median -4% -3% 0.99111
of
1Kruskal-Wallis p-value
638
ro
-p
re
lP
na
ur
Jo
31
of
ro
-p
re
lP
na
ur
Jo
of
ro
-p
re
lP
na
ur
Jo
of
ro
-p
re
lP
na
ur
Jo
of
ro
-p
re
lP
na
ur
Jo
of
ro
-p
re
lP
na
ur
Jo
Figure S1: CONSORT Diagram for CALGB 140503
of
ro
-p
re
lP
na
ur
Jo

Journal Pre-Proof: The Journal of Thoracic and Cardiovascular Surgery

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Journal Pre-Proof: The Journal of Thoracic and Cardiovascular Surgery

Uploaded by

Copyright:

Available Formats

on July 26, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc.

All rights reserved.

Lobectomy, segmentectomy or wedge resection for peripheral clinical T1aN0 non-

To appear in: The Journal of Thoracic and Cardiovascular Surgery

Received Date: 28 April 2023

30 Nasser Altorki M.D.

111 Graphical abstract

113 Central Picture

118 Central message

120 segmentectomy or wedge resections had similar survival outcomes.

121 Perspective statement

141 Glossary of abbreviations

142 AJCC: American Joint Committee on Cancer

143 LR: lobar resection

144 SLR: Sublobar resection

145 DFS: Disease-free survival

146 OS: Overall survival

149 WR: Wedge resection

151 FEV1: Forced expiratory volume in 1 second

152 FVC; Forced vital capacity

153 DSMB: Data safety and monitoring board

154 SDMC: Statistical and data management center

155 CALGB: Cancer and Leukemia Group B

156 NCI: National cancer institute

157 HR: Hazard ratio

158 CI: Confidence interval

159 CT: computerized tomography

188 lobar resection reported in National and Society registries4,5,6,7.

198 flow rates at 6 months postoperatively.

200 Study design and participants

215 confirmed N1 or N2 disease were not eligible.

226 Safety Monitoring Board (DSMB) twice yearly.

231 patients underwent randomization (1:1) to either LR or SLR, based on a permuted-block

235 nurses, data managers, or statisticians.

246 have an anatomic segmentectomy if technically possible or a LR if necessary. Mediastinal nodal

272 DSMB monitoring of the trial.

291 and SR was tested with Fisher’s Exact test.

302 number NCT00499330.

303 Role of the funding source

336 competing causes of death (p=0.448) between the two groups.

343 [95%CI;77.0-88.9%] after WR (Log rank p=0.232) (Figure 3).

344 Resection Margins

367 (Fisher’s exact, p=0.007).

404 remedied by additional parenchymal resection.

408 for assessment of pulmonary function following lung resection3.

450 pathological variables such as adenocarcinoma subtypes, lympho-vascular invasion or spread

477 Ito H, Okumura N, Yamaguchi M, Ikeda N, Wakabayashi M, Nakamura K, Fukuda H,

480 non-small-cell lung cancer (JCOG0802/WJOG4607L): a multicentre, open-label, phase 3,

481 randomised, controlled, non-inferiority trial. Lancet. 2022 Apr 23;399(10335):1607-1617

488 stage. Chest. 2005 Nov;128(5):3255-60

491 Cancer. 2016 Jan;17(1):47-55

494 Outcomes and Research. Ann Thorac Surg. 2023 Jan;115(1):43-49

498 Cardiothorac Surg. 2017 Dec 1;52(6):1041-1048

502 risk. Ann Stat 1988; 16: 1141-54.

504 Am Stat Assoc 1999; 94: 496-509.

Median 1.5 1.5 1.5 1.5

1.0-1.5 182 (50.3%) 103 (50.5%) 69 (52.7%) 354 (50.8%)

Smoking Status, n (%) 0.30112

Current 145 (40.1%) 93 (45.6%) 47 (35.9%) 285 (40.9%)

Surgical margin length (cm)

Margin/clinical tumor ratio

1-<2 54 (39.7%) 38 (50.7%)

3-<4 4 (2.9%) 3 (4.0%)