Accepted Manuscript

Thoracic Skeletal Muscle is Associated with Adverse Outcomes After Lobectomy for
Lung Cancer

Florian J. Fintelmann, MD, Fabian M. Troschel, cand.med., Julia Mario, MD, Yves R.
Chretien, MD, PhD, Sheila J. Knoll, RN, Ashok Muniappan, MD, Henning A. Gaissert,
MD
PII: S0003-4975(18)30077-8
DOI: 10.1016/j.athoracsur.2018.01.013
Reference: ATS 31305

To appear in: The Annals of Thoracic Surgery

Received Date: 5 June 2017

Revised Date: 6 December 2017
Accepted Date: 2 January 2018

Please cite this article as: Fintelmann FJ, Troschel FM, Mario J, Chretien YR, Knoll SJ, Muniappan A,
Gaissert HA, Thoracic Skeletal Muscle is Associated with Adverse Outcomes After Lobectomy for Lung
Cancer, The Annals of Thoracic Surgery (2018), doi: 10.1016/j.athoracsur.2018.01.013.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Thoracic Skeletal Muscle is Associated with Adverse Outcomes After Lobectomy for Lung

Cancer

Running Head: Thoracic muscle and lobectomy outcomes

Florian J. Fintelmann, MD* 1, Fabian M. Troschel, cand.med.* 1, Julia Mario, MD1, Yves R. ,

PT
Chretien, MD, PhD1, Sheila J. Knoll, RN2, Ashok Muniappan, MD§2, Henning A. Gaissert, MD§ 2

RI
1) Department of Radiology, Division of Thoracic Imaging and Intervention, Massachusetts

SC
General Hospital, Boston, MA

2) Department of Surgery, Division of Thoracic Surgery, Massachusetts General Hospital,

Boston, MA

U
AN
*Dr Fintelmann and Fabian M. Troschel are co-first authors; §Drs Muniappan and Gaissert are co-
M

senior authors.
D
TE

Corresponding author:

Florian J. Fintelmann, MD
EP

Massachusetts General Hospital

C

55 Fruit Street FND-202

Boston, MA 02114
AC

Email: fintelmann@mgh.harvard.edu

Total word count: 4402/4500

Keywords: muscle, computed tomography, lung cancer, lobectomy, risk stratification

1
 ACCEPTED MANUSCRIPT
ABSTRACT
Background: Assessment of risk associated with lung cancer resection is primarily based on

evaluation of cardiopulmonary function and remains imprecise. We investigated the relationship

between thoracic muscle and early outcomes after lobectomy.

Methods: Cross-sectional area of skeletal muscle was measured at the level of the fifth thoracic

PT
vertebra on computed tomography in 135 consecutive patients prior to lobectomy for lung cancer.

Patients were stratified into low and high muscle groups using the gender-specific muscle median.

RI
Primary outcome was a composite of any postoperative complication as per Society of Thoracic

Surgeons General Thoracic Surgical Database. Secondary outcomes included postoperative

SC
respiratory complications, postoperative intensive care unit (ICU) admission, hospital length of stay

(LOS), and hospital readmission within 30 days of hospital discharge. Chi-square testing, adjusted

U
multivariable regression and likelihood ratio test were performed.
AN
Results: Patients with low muscle were significantly more likely to have any postoperative

complication and respiratory postoperative complications. While postoperative ICU admission was
M

similar for low and high muscle groups, low muscle patients experienced longer hospital LOS and a
D

higher rate of hospital readmission. Adjusted multivariable regression revealed the independent

association of thoracic muscle with all outcomes. The likelihood ratio test suggested that thoracic
TE

muscle adds predictive capability to information captured by preoperative pulmonary function testing.
EP

Conclusions: Low thoracic muscle is independently associated with increased postoperative

complications and healthcare utilization among patients undergoing lobectomy for lung cancer.
C

Evaluation of thoracic muscle may enhance risk prediction models.

AC

2
 ACCEPTED MANUSCRIPT
Patients with localized lung cancer have the greatest chance of cure with lung resection1.

Despite remarkable improvements in outcomes after lung resection in recent decades, predicting

which patient is at greatest risk for operative complications remains imprecise2,3. Current American

and European specialty society guidelines 4–6 point to pulmonary function tests (PFTs) as key

objective data for preoperative risk stratification. Lack of patient cooperation may affect these

PT
measurements, however. Indeed, several studies suggest that not all patients with marginal lung

function are at greater risk of pulmonary complications7,8, and that performance status measures

RI
should be considered to refine preoperative risk assessment8. Clearly, additional objective data are

needed to inform preoperative risk.

SC
Low muscle mass has been shown to predict poor surgical outcome9. For example, psoas

muscle area measured on preoperative computed tomography (CT) at the level of the third lumbar

U
vertebral body (L3) predicts mortality after pneumonectomy for lung cancer 10 and muscle cross-
AN
sectional area (CSA) of all muscles at L3 is associated with adverse outcomes after abdominal11–13,

vascular 14,15, cardiac16 and esophageal surgery17. Existing risk prediction models for postoperative
M

morbidity have not accounted for thoracic skeletal muscle18–21. Yet the data are readily available,
D

thanks to near-universal staging with chest CT. Given the relationship of thoracic muscles to
TE

respiratory function, we postulate that the state of thoracic musculature impacts recovery after lung

resection in lung cancer patients. We therefore investigated the relationship between the cross-
EP

sectional area of thoracic skeletal muscle and early outcome after lobectomy for lung cancer.
C

Patients and Methods

AC

The Institutional Review Board approved this retrospective analysis of prospectively collected

data. The study was compliant with the Health Insurance Portability and Accountability Act. No

outside or industry funding was provided.

Inclusion Criteria

Consecutive patients who underwent lobectomy or bilobectomy for primary lung cancer at

Massachusetts General Hospital between 7/1/2015 and 6/30/2016 were identified in the institutional

3
 ACCEPTED MANUSCRIPT
data collected for the Society of Thoracic Surgeons (STS) General Thoracic Surgery Database

(GTSD). A complete set of clinical data and a chest CT obtained within 90 days prior to surgery

showing the relevant thoracic muscles were required for inclusion. Of 184 eligible patients, 135

(73.4%) met inclusion criteria; reasons for exclusion are detailed in Figure 1.

PT
Outcomes

Primary outcome was a composite of any postoperative complication. Secondary outcomes

RI
included the composite of postoperative respiratory complications, postoperative intensive care unit

(ICU) admission, hospital length of stay (LOS), and hospital readmission within 30 days of hospital

SC
discharge. Outcomes were defined according to version 2.3 of the STS GTSD22 and had been

recorded by our Thoracic Surgical Data Manager prior to this study. Hospital LOS was defined as

U
number of days from date of surgery to date of hospital discharge. An external audit performed at the
AN
direction of the STS National Database in September 2015 found that our site’s overall data

agreement rate was 98.5%.

M
D

Image Analysis
TE

CT studies were loaded onto a workstation using OsiriX Lite software (version 7.0.2, Pixmeo,

Geneva, Switzerland) and evaluated for factors precluding muscle analysis, such as prior anatomy-
EP

altering operations or incomplete visualization due to beam hardening artifact, patient positioning or

field of view. Patients with such inadequate images were excluded (Figure 1). A single axial image
C

showing both transverse processes at the level of the fifth thoracic vertebral body (T5) was selected in
AC

each examination. We chose to measure muscle at this level because this bony landmark corresponds

closely to the aortic arch landmark used by others 23–25. A primary analyst blinded to clinical outcomes

quantified cross-sectional area (CSA) of skeletal muscles at T5 in cm2 for each patient using semi-

automated segmentation with Hounsfield unit thresholds of -29 to +150. Muscles measured included

the pectoralis major and minor, serratus anterior, external and internal intercostals, teres major,

infraspinatus, rhomboid major, trapezius, and subscapularis (Figure 2).

4
 ACCEPTED MANUSCRIPT
A board-certified radiologist verified a random sample (20%) of all cases. Inter- and intra-

analyst agreement was assessed by having the random sample re-analyzed independently three

months later by a second analyst and by the primary analyst.

Statistical Methods

PT
Analyses were performed with STATA software (version 13.0, StataCorp, College Station,

Texas) with the alpha level set at 0.05. Descriptive statistics were used to estimate frequencies, means

RI
with standard deviations (SD), and medians with interquartile ranges (IQR). Patients were stratified

into low and high muscle groups using the gender-specific muscle CSA median, such that the low

SC
muscle group had CSAs less than the gender-specific median and the high muscle group had CSAs

greater than or equal to the gender-specific median (Figure 3). We chose the gender-specific median

U
as a separator because no reference values are available to define the range of normal for thoracic
AN
muscle.

Differences between characteristics of patients in the low and high muscle groups were
M

assessed with the use of two-tailed chi-square tests for categorical variables, Student t-tests for
D

continuous variables demonstrating approximate normality, and Mann-Whitney-U-Test for

TE

continuous variables not normally distributed. Inter- and intra-analyst agreement was quantified with

intraclass correlation coefficients (ICC). The distribution of forced expiratory volume in 1 second
EP

(FEV1) % predicted was found to be normal based on the Shapiro-Wilks test and the relationship with

CSA was analyzed using linear regression.

C

Chi-square testing was used to determine the relationship between low and high muscle
AC

groups and outcomes. To apply chi-square testing to the outcome LOS, the continuous variable was

transformed into a binary variable for this analysis by stratifying patients according to the median.

Multivariable logistic regression analyses were performed to determine whether T5 muscle

CSA as the continuous explanatory variable (expressed per 10 cm2) was independently associated

with binary outcomes. The primary model was adjusted for gender, age, body mass index (BMI),

FEV1 % predicted and surgical approach. These co-variables were determined a priori based on

available literature and clinical experience21. Given that its distribution is right-skewed, hospital LOS

5
 ACCEPTED MANUSCRIPT
was stratified into quartiles and analyzed with multivariable ordinal logistic regression. The model

was limited to the co-variable gender to test for association with postoperative respiratory

complications and 30-day hospital readmission to prevent over fitting a small number of events.

The likelihood ratio test was used to compare a model consisting of age, gender, BMI,

FEV1% predicted and surgical approach (model A) to model B which included T5 muscle CSA in

PT
addition to the variables included in model A.

RI
Results

Patient Characteristics and Outcomes

SC
The study included 135 patients who underwent either lobectomy or bilobectomy (n=2),

U
predominantly via thoracoscopy (n=90). Mean age was 69 years, and women were in the majority.

Patient characteristics are detailed in Table 1.

AN
Early post-operative outcomes are summarized in Table 2. Postoperative complications
M

occurred in 37% of patients with postoperative respiratory complications in 12.6% of patients.

Postoperative ICU admission was required for 18.5% of patients and the median hospital LOS was 4
D

days. The hospital readmission rate within 30 days of hospital discharge was 5.2%. There were no
TE

operative or 30-day deaths.

EP

Muscle Measurements and Relation to PFTs

The mean T5 muscle CSA measured 190.7 ± 45.6 cm2 in men and 131.3 ± 19.5 cm2 in
C

women, with median values of 181.25 cm² and 129.38 cm², respectively. The distribution of CSA
AC

was unimodal and symmetric among both men and women (Figure 3). By definition, the low muscle

group included 67 patients and the high muscle group included 68 patients in both men and women.

Inter- and intra-class muscle measurement agreement was excellent with ICCs of 0.999, (95% CI

0.996-0.999) and 0.998 (95% CI 0.996-0.999), respectively.

FEV1 % predicted was normally distributed and compared to T5 muscle CSA using linear

regression. While this analysis uncovered corresponding increases of T5 muscle CSA and FEV1

6
 ACCEPTED MANUSCRIPT
(Figure 4), the correlation between T5 muscle CSA and FEV1 was weak, indicating considerable

unexplained variation between thoracic muscle and FEV1.

Simple Analyses

When comparing low and high muscle groups using the chi-square test, the low muscle group

PT
had higher rates of both any complication and respiratory postoperative complications, with a relative

risk (RR) of 2.36 (0.52 vs. 0.22, p<0.001) and 4.75 (0.21 vs. 0.04, p=0.004), respectively (Figure 5).

RI
The rate of postoperative ICU admission was not significantly different between low and high muscle

groups (RR 1.47; 0.22 vs. 0.15, p=0.25). Low muscle patients experienced higher rates of hospital

SC
LOS greater than the 4-day median (RR 1.52; 0.58 vs. 0.38, p=0.02) and 30-day hospital readmission

(RR 9.0; 0.09 vs. 0.01, p=0.05).

U
AN
Adjusted Analyses

Adjusted multivariable logistic regression showed that higher T5 muscle CSA was
M

independently associated with lower rates of any postoperative complication and postoperative ICU
D

admission (Table 3). A regression model adjusted only for gender showed that higher T5 muscle CSA
TE

was independently associated with lower rates of postoperative respiratory complication and 30-day

hospital readmission. Multivariable ordinal logistic regression indicated that the odds of being in a
EP

higher quartile LOS decreased by 13% for each 10cm2 increase in muscle area.

The likelihood ratio test demonstrated an association of thoracic muscle with the composite of
C

all postoperative complications, postoperative ICU admission and hospital LOS since model B fit
AC

significantly better than model A (Table 4). This finding suggests that T5 muscle CSA adds

predictive capability to information captured by preoperative PFTs.

Comment

Thoracic muscle area derived from preoperative chest CT in patients undergoing lobectomy

for lung cancer is independently associated with any postoperative complication and respiratory

7
 ACCEPTED MANUSCRIPT
postoperative complications, postoperative ICU admission, prolonged hospital LOS, and hospital

readmission within 30 days of hospital discharge.

Our model considers thoracic muscle CSA in addition to 5 variables known to be associated

with outcomes after lung resection for lung cancer21. BMI serves as an indicator of nutritional status,

and skeletal muscle area corrected for gender as an indicator of frailty 26. These preliminary data

PT
suggest that characterizing thoracic muscle on routine preoperative chest CT presents an opportunity

to refine risk stratification models for lung cancer patients undergoing lobectomy.

RI
Thoracic surgeons currently do not have a simple, objective and robust preoperative risk

stratification model to predict whether a patient can tolerate a lobectomy without significant

SC
complications. Quantitative data derived from routine CT could bolster future efforts to refine

preoperative risk stratification currently largely based on lung function and surgeon judgment. The

U
absence of a strong correlation between thoracic muscle CSA and FEV1% predicted, in this study,
AN
suggests that each capture distinct patient characteristics. We do not submit that thoracic muscle area

should replace pulmonary function studies as a predictor of postoperative events. However, the results
M

of the likelihood ratio test suggest that the addition of muscle assessment may improve existing
D

preoperative risk prediction models.

TE

The identification of patients at high risk for postoperative complications from lobectomy

may lead to consideration of sublobar resection or non-operative approaches. However, if lobectomy

EP

is the most appropriate management for high risk patients because of superior oncologic outcome,

preoperative pulmonary rehabilitation and nutritional or exercise interventions may be considered27.

C

What is more, preoperative identification of patients at risk for hospital readmission allows for
AC

redirection of resources and services to mitigate that risk28.

Numerous investigations have reported the association of low lumbar muscle area with

increased postoperative complications, hospital LOS and mortality 9,11–17. However, the relation of

postoperative outcomes to thoracic muscle area has heretofore received insufficient attention, despite

an intuitive relevance of chest wall muscle strength to respiratory function. The muscles measured in

this study include primary and accessory muscles of respiration and those supporting the force of

cough. As such, the observation that low thoracic muscle area is associated with increased incidence

8
 ACCEPTED MANUSCRIPT
of post-operative pulmonary complications and 30-day hospital readmission is in line with the finding

that the most common reasons for hospital readmission after lobectomy are of respiratory origin29.

Not all studies investigating the relation of muscle to postoperative outcomes have found

significant associations. Lee et al. found that low muscle area, measured at the level of the carina, was

not significantly associated with mortality after lung transplantation30. While there was a trend

PT
towards delayed recovery in patients with low thoracic muscle, multivariable analysis was not

performed to eliminate confounding variables such as the etiology of end-stage lung disease. We

RI
suspect that reduced thoracic muscle is likely an independent risk factor for postoperative

complications not only for lobectomy but also for other thoracic operations such as lung

SC
transplantation.

Clearly, additional studies of large patient populations are required to further define the role

U
of thoracic muscle area in preoperative risk stratification. Specifically, reference values of expected
AN
thoracic muscle in men and women of different ages need to be established. Fortunately, muscle

measurements can be performed in less than 10 minutes per patient with free and universally-
M

available software, and may be entirely automated with machine learning in the future31. Future work
D

could assess the correlation of thoracic muscle measured on CT with evaluation of muscle strength
TE

and frailty, and whether the associations identified in this study are reproducible at other centers.

Several limitations of our study warrant consideration. First, this is a retrospective study, and
EP

approximately 15% of patients had to be excluded due to either unavailable PFTs or CT scans.

Second, this study was performed at a single, academic tertiary care center and the findings may not
C

be applicable to other settings. Third, patient disposition issues that delay discharge may affect
AC

hospital LOS independent of thoracic muscle.

In conclusion, low thoracic muscle area is independently associated with increased

postoperative complications, postoperative ICU admission, prolonged hospital LOS, and hospital

readmission within 30 days after hospital discharge among patients undergoing lobectomy for lung

cancer. Evaluation of thoracic muscle on preoperative CT may enhance existing risk prediction

models.

9
 ACCEPTED MANUSCRIPT
References

1. Mazzone, P. Preoperative evaluation of the lung resection candidate. Cleve. Clin. J. Med. 79

Electro, eS17-22 (2012).

2. Ferguson, M. K. & Vigneswaran, W. T. Changes in patient presentation and outcomes for

major lung resection over three decades. Eur. J. Cardiothorac. Surg. 33, 497–501 (2008).

PT
3. Strand, T.-E., Bartnes, K. & Rostad, H. National trends in lung cancer surgery. Eur. J.

Cardiothorac. Surg. 42, 355–8 (2012).

RI
4. Brunelli, A., Kim, A. W., Berger, K. I. & Addrizzo-Harris, D. J. Physiologic evaluation of the

patient with lung cancer being considered for resectional surgery: Diagnosis and management

SC
of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice

guidelines. Chest 143, e166S–90S (2013).

5.
U
Brunelli, A. et al. ERS/ESTS clinical guidelines on fitness for radical therapy in lung cancer
AN
patients (surgery and chemo-radiotherapy). Eur. Respir. J. 34, 17–41 (2009).

6. Lim, E. et al. Guidelines on the radical management of patients with lung cancer. Thorax 65
M

Suppl 3, iii1-i27 (2010).

D

7. Berry, M. F. et al. Pulmonary function tests do not predict pulmonary complications after
TE

thoracoscopic lobectomy. Ann. Thorac. Surg. 89, 1044-51–2 (2010).

8. Paul, S. et al. Outcomes of Lobectomy in Patients with Severely Compromised Lung Function
EP

(Predicted Postoperative Diffusing Capacity of the Lung for Carbon Monoxide % ≤ 40%).

Ann. Am. Thorac. Soc. 10, 616–621 (2013).

C

9. Englesbe, M. J. et al. Analytic Morphomics, Core Muscle Size, and Surgical Outcomes. Ann.
AC

Surg. 256, 255–261 (2012).

10. Hervochon, R. et al. Body Mass Index and Total Psoas Area Affect Outcomes in Patients

Undergoing Pneumonectomy for Cancer. Ann. Thorac. Surg. 103, 287–295 (2017).

11. Englesbe, M., Patel, S. & He, K. Sarcopenia and Post-Liver Transplant Mortality. J Am Coll

Surg 211, 271–278 (2010).

12. Peng, P. et al. Impact of Sarcopenia on Outcomes Following Resection of Pancreatic

Adenocarcinoma. J. Gastrointest. Surg. 16, 1478–1486 (2012).

10
 ACCEPTED MANUSCRIPT
13. Miyamoto, Y. et al. Sarcopenia is a Negative Prognostic Factor After Curative Resection of

Colorectal Cancer. Ann. Surg. Oncol. 22, 2663–2668 (2015).

14. Matsubara, Y. et al. Sarcopenia is a prognostic factor for overall survival in patients with

critical limb ischemia. J. Vasc. Surg. 61, 945–950 (2015).

15. Lee, J. S.-J. et al. Frailty, core muscle size, and mortality in patients undergoing open

PT
abdominal aortic aneurysm repair. J. Vasc. Surg. 53, 912–917 (2011).

16. Zuckerman, J. et al. Psoas Muscle Area and Length of Stay in Older Adults Undergoing

RI
Cardiac Operations. Ann. Thorac. Surg. (2016). doi:10.1016/j.athoracsur.2016.09.005

17. Nishigori, T. et al. Sarcopenia as a predictor of pulmonary complications after esophagectomy

SC
for thoracic esophageal cancer. J. Surg. Oncol. 113, 678–84 (2016).

18. Wright, C. D. et al. Predictors of prolonged length of stay after lobectomy for lung cancer: a

U
Society of Thoracic Surgeons General Thoracic Surgery Database risk-adjustment model. Ann.
AN
Thorac. Surg. 85, 1857–65; discussion 1865 (2008).

19. DeLuzio, M. R. et al. A model for predicting prolonged length of stay in patients undergoing
M

anatomical lung resection: a National Surgical Quality Improvement Program (NSQIP)

D

database study. Interact. Cardiovasc. Thorac. Surg. 23, 208–15 (2016).

TE

20. Kozower, B. D. et al. STS database risk models: predictors of mortality and major morbidity

for lung cancer resection. Ann. Thorac. Surg. 90, 875-81–3 (2010).
EP

21. Fernandez, F. G. et al. The Society of Thoracic Surgeons Lung Cancer Resection Risk Model:

Higher Quality Data and Superior Outcomes. Ann. Thorac. Surg. 102, 370–7 (2016).
C

22. The Society of Thoracic Surgeons General Thoracic Surgery Database Major Procedure Data
AC

Collection Form Version 2.3.

23. Diaz, A. A. et al. Chest CT Measures of Muscle and Adipose Tissue in COPD. Acad. Radiol.

21, 1255–1261 (2014).

24. McDonald, M.-L. N. et al. Quantitative computed tomography measures of pectoralis muscle

area and disease severity in chronic obstructive pulmonary disease. A cross-sectional study.

Ann. Am. Thorac. Soc. 11, 326–34 (2014).

25. Kim, Y. S., Kim, E. Y., Kang, S. M., Ahn, H. K. & Kim, H. S. Single cross-sectional area of

11
 ACCEPTED MANUSCRIPT
pectoralis muscle by computed tomography - correlation with bioelectrical impedance based

skeletal muscle mass in healthy subjects. Clin. Physiol. Funct. Imaging n/a-n/a (2015).

doi:10.1111/cpf.12333

26. Mueller, N. et al. Can Sarcopenia Quantified by Ultrasound of the Rectus Femoris Muscle

Predict Adverse Outcome of Surgical Intensive Care Unit Patients as well as Frailty? A

PT
Prospective, Observational Cohort Study. Ann. Surg. 264, 1116–1124 (2016).

27. Pouwels, S. et al. Preoperative exercise therapy in lung surgery patients: A systematic review.

RI
Respir. Med. 109, 1495–1504 (2015).

28. Stiles, B. M. et al. Incidence and Factors Associated With Hospital Readmission After

SC
Pulmonary Lobectomy. Ann. Thorac. Surg. 101, 434–443 (2016).

29. Rajaram, R., Ju, M. H., Bilimoria, K. Y., Ko, C. Y. & DeCamp, M. M. National evaluation of

U
hospital readmission after pulmonary resection. J. Thorac. Cardiovasc. Surg. 150, 1508–14.e2
AN
(2015).

30. Lee, S. et al. Sarcopenia of thoracic muscle mass is not a risk factor for survival in lung
M

transplant recipients. J. Thorac. Dis. 8, 2011–2017 (2016).

D

31. Lee, H. et al. Pixel-Level Deep Segmentation : Artificial Intelligence Quantifies Muscle on
TE

Computed Tomography for Body Morphometric Analysis. J. Digit. Imaging (2017).

doi:10.1007/s10278-017-9988-z
C EP
AC

12
 ACCEPTED MANUSCRIPT
Table 1. Patient characteristics

Patient characteristics Low Muscle High Muscle p value

Group Group
Gender, no. (%)
Female 39 (58.2) 39 (57.4) 0.9*
Male 28 (41.8) 29 (42.7)
Age, median (IQR), years 72 (67, 67 (59.5, 0.0001**
80) 73.5)
Body Mass Index, median (IQR), kg/m2 25.8 (22.7, 26.7 (24, 0.1**

PT
29.1) 33.2)
T5 Muscle CSA, mean (SD), cm2
Female 116.5 (9.6) 145.94 (15.4) <0.001***

RI
Male 153.2 (18.1) 226.9 (32.9) <0.001***
Zubrod score, no. (%)
Stage 0 39 (58.2) 39 (57.4) 0.6*

SC
Stage 1 27 (40.3) 29 (42.7)
Stage 2 1 (1.5) 0 (0)
Preoperative chemoradiation therapy, no. (%)
Yes 1 (1.5) 9 (13.2) 0.009*

U
No 66 (98.5) 59 (86.8)
Cigarette smoking ever, no. (%)
AN
Yes 59 (88.1) 53 (77.9) 0.12*
No 8 (11.9) 15 (22.1)
FEV1 % predicted, mean (SD), percent 84.3 (22.3) 93.5 (19.3) 0.01***
ASA Score, no. (%)
M

I 0 (0) 2 (2.94) 0.2*

II 23 (34.3) 30 (44.1)
III 44 (65.7) 34 (50.0)
D

IV 0 (0) 1 (1.5)
V 0 (0) 1 (1.5)
Coronary Artery Disease, no. (%)
TE

Yes 9 (11.9) 8 (11.8) 0.98*

No 59 (88.1) 60 (88.2)
Peripheral Vascular Disease, no. (%)
EP

Yes 6 (9.0) 2 (2.9) 0.14*

No 61 (91.0) 66 (97.1)
Diabetes, no. (%)
Yes 11 (16.4) 9 (13.2) 0.6*
C

No 56 (83.6) 59 (86.8)
Hypertension, no. (%)
AC

Yes 38 (56.7) 35 (51.5) 0.5*

No 29 (43.4) 33 (48.5)
Prior Cardiothoracic Surgery, no. (%)
Yes 4 (6.0) 4 (5.9) 0.98*
No 63 (94.0) 64 (94.1)
Procedure type, no. (%)
Lobectomy 66 (98.5) 67 (98.5) 0.99*
Bilobectomy 1 (1.5) 1 (1.5)
Surgical approach, no. (%)
Thoracotomy 19 (28.4) 26 (38.2) 0.2*
Thoracoscopy 48 (71.6) 42 (61.8)
Lobe resected, no. (%) 0.2*
RUL 15 (22.4) 24 (35.3)

13
 ACCEPTED MANUSCRIPT
RML 3 (4.5) 7 (10.3)
RLL 13 (19.4) 6 (8.9)
LUL 17 (25.4) 15 (22.1)
LLL 18 (26.9) 15 (22.1)
RUL + RML 1 (1.5) 0 (0)
RLL + RML 0 (0) 1 (1.5)
Lung cancer stage, no. (%)
Stage IA 47 (70.2) 44 (64.7) 0.3*
Stage IB 13 (19.4) 8 (11.8)
Stage IIA 3 (4.5) 9 (13.2)

PT
Stage IIB 3 (4.5) 4 (5.9)
Stage IIIA 1 (1.5) 1 (1.5)
Stage IV 0 (0) 2 (2.9)

RI
Lung cancer type, no. (%)
Adenocarcinoma 52 (76.1) 52 (76.5) 0.3*
Squamous cell 12 (17.9) 8 (11.8)

SC
Neuroendocrine 2 (3.0) 7 (10.3)
Mucoepidermoid 1 (1.5) 1 (1.5)
ASA, American Society of Anesthesiology; CSA, cross-sectional area; FEV1, forced expiratory

U
volume in 1 second; IQR, interquartile range; LLL, left lower lobe; LUL, left upper lobe; RLL, right
AN
lower lobe; RML, right middle lobe; RUL, right upper lobe; SD, standard deviation. * Chi-square test;

** Mann-Whitney-U-Test; *** Student T-test

M
D
TE
C EP
AC

14
 ACCEPTED MANUSCRIPT
Table 2. Outcomes

Outcome
Any postoperative complication (composite), no. (%)
Yes 50 (37.0)
No 85 (63.0)
Postoperative respiratory complications (composite), no. (%)
Yes 17 (12.6)
No 118 (87.4)

PT
Postoperative intensive care unit admission, no. (%)
Yes 25 (18.5)
No 110 (81.5)
Hospital length of stay, median (IQR), days 4 (3-6)

RI
Hospital readmission within 30 days of hospital discharge, no. (%)
Yes 7 (5.2)
No 128 (94.8)

SC
IQR, interquartile range

U
AN
M
D
TE
C EP
AC

15
 ACCEPTED MANUSCRIPT

Table 3. Adjusted multivariable associations between thoracic skeletal muscle and outcomes

Outcomes
Any postoperative Postoperative Postoperative Hospital length of Hospital
complication respiratory intensive care unit stay readmission
(composite) complications admission (in quartiles) within 30 days of
(composite) hospital discharge
Independent OR (95% CI) p OR (95% p OR (95% CI) p OR (95% CI) p OR (95% CI) p

PT
variables value CI) value value value value
T5 muscle 0.86 (0.75- 0.04 0.8 (0.65- 0.04 0.73 (0.56- 0.02 0.87 (0.78- 0.02 0.58 (0.37- 0.02
0.995) 0.98) 0.95) 0.98) 0.91)
CSA
Age 1.01 (0.97- 0.73 0.98 (0.93- 0.38 1.03 (0.99- 0.11

RI
1.05) 1.03) 1.06)
Gender
Male Ref 1.00 Ref 1.00 Ref 1.00 Ref 1.00 Ref 1.00
Female 0.43 (0.34- 0.15 0.27 0.05 0.62 (0.13- 0.54 0.22 (0.08- 0.003 0.07 (0.01- 0.01

SC
1.34) (0.07- 2.88) 0.61) 0.52)
1.01)
FEV1 % 0.98 (0.97- 0.11 0.97 (0.95- 0.04 0.99 (0.97- 0.09
1.00) 0.999) 1.00)
predicted
BMI 1.02 (0.97- 0.45 1.00 (0.94- 0.98 1.00 (0.96- 0.87

U
1.07) 1.07) 1.06)
Surgical
approach
AN
Ref 1.00 Ref 1.00 Ref 1.00
Thoracoscopy
1.94 (0.89- 0.09 11.01 (3.57- <0.01 9.31 (4.42- <0.01
Thoracotomy 4.22) 34.00) 19.60)
M

BMI, body mass index; CI, confidence interval; CSA, cross-sectional area; FEV1, forced expiratory

volume in 1 second; OR, odds ratio; Ref, reference.

D
TE
C EP
AC

16
 ACCEPTED MANUSCRIPT
Table 4. Results of likelihood ratio test

Any postoperative Postoperative Postoperative Hospital length Hospital

complication respiratory intensive care unit of stay readmission within
(composite) complications admission (in quartiles) 30 days of hospital
(composite) discharge
p value 0.036 Not applicable 0.008 0.016 Not applicable

PT
RI
U SC
AN
M
D
TE
C EP
AC

17
 ACCEPTED MANUSCRIPT
Figure Legends

Figure 1. Inclusion/exclusion criteria.

Figure 2. Example of thoracic muscle measurement. Axial computed tomography images of a 70

year-old male in the low muscle group (A) and of a 71 year-old male in the high muscle group (B)

demonstrate different muscle cross-sectional areas (red) at the level of the fifth thoracic vertebral

PT
body.

Figure 3. Distribution of muscle measurements. Histograms illustrate distribution of T5 muscle cross-

RI
sectional area with 30 cm2 intervals for men (A) and 10 cm2 intervals for women (B). The gender-

specific median (vertical line) separates patients into low and high muscle groups.

SC
Figure 4. Relationship between thoracic muscle and FEV1 % predicted. As muscle cross-sectional

area (CSA) increases, FEV1 increases for both men (A) (y = 0.12x + 61.61, r = 0.2884, p = 0.03) and

U
women (B) (y = 0.33x + 49.63, r = 0.2834, p = 0.012). While the correlations (p) are significant, the
AN
respective coefficients of determination (r) are weak, suggesting considerable unexplained variation

between these variables.

M

Figure 5. Incidence of outcomes in low and high muscle groups. Bars illustrate incidence of
D

outcomes among patients with low and high muscle cross-sectional area.
TE
C EP
AC

18
 ACCEPTED MANUSCRIPT
Alphabetical list of abbreviations
ASA American Society of Anesthesiology
BMI Body mass index
CI Confidence interval
CSA Cross-sectional area
CT Computed tomography

PT
FEV1 Forced expiratory volume in 1 second
GTSD General Thoracic Surgery Database

RI
ICC Intraclass correlation coefficient
ICU Intensive care unit

SC
IQR Interquartile range
L3 Third lumbar vertebral body

U
LOS Length of stay
OR Odds ratio
AN
PFTs Pulmonary function tests
RR Relative risk
M

SD Standard deviation
STS Society of Thoracic Surgeons
D

T5 Fifth thoracic vertebral body

TE
C EP
AC

19
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC