Pectoral muscle area derived from chest computed tomography to identify quadriceps muscle weakness in COPD

Graphical abstract

AUC: area under the curve; CT: computed tomography; (e)QS: (estimated) quadriceps strength; FEV₁ % pred: % predicted forced expiratory volume in 1 s; PMA: pectoral muscle area.

Abstract

Background

In patients with COPD, reduced quadriceps strength is associated with poor prognosis. Measuring quadriceps strength requires specific equipment and expertise, which limits its dissemination. Chest computed tomography (CT) scans are routinely performed in COPD patients. It was hypothesised that pectoral muscle area (PMA) derived from chest CT scans was related to quadriceps strength, and that low quadriceps strength could be derived from PMA in stable COPD patients.

Methods

In a retrospective cross-sectional study, thoracic CT scans and quadriceps strength were obtained simultaneously. The highest quadriceps strength value of the dominant leg was recorded (in kg). The PMA (sum of right and left measurements, in cm²) was obtained from a single axial inspiratory CT slice at the aortic arch.

Results

A total of 82 outpatients with stable COPD of varying levels of severity were analysed. The unadjusted r² between PMA and quadriceps strength was 0.32. A formula for estimating quadriceps strength from PMA was established: estimated quadriceps strength=25.2+0.41×PMA (cm²)−0.29×age (years)+6.35 (if male) (r²=0.43). The area under the receiver operating characteristic curve for identifying low quadriceps strength (lowest tercile of the studied population) using PMA was 0.76 (95% CI 0.65–0.87), with a PMA threshold of 28.5 cm².

Conclusions

PMA is a reliable surrogate for quadriceps strength assessment in routine COPD management. This easily accessible imaging biomarker could be used to identify COPD patients with low quadriceps strength, not only for prognostic purposes but also to enable them to benefit from interventions aimed at improving muscle weakness.

Shareable abstract

Pectoral muscle area derived from a chest computed tomography performed as part of routine care is reliable for identifying quadriceps weakness in COPD patients https://bit.ly/3SrBska

Introduction

COPD is a leading cause of death worldwide [1]. Low physical activity is one of the main factors associated with mortality and disabilities in COPD [2]. Lower limb muscle dysfunction largely contributes to poor exercise performance [3] and a low level of physical activity [4], and is associated with higher mortality [5] in patients with COPD. Lower limb muscle dysfunction involves mass loss and weakness of the quadriceps in both strength and endurance [6], resulting from several intermediary mechanisms, including systemic inflammation, oxidative stress [7], an imbalance between energy intake and expenditure [8] and deconditioning [9]. Recognising lower limb muscle dysfunction is important as it is a treatable trait that is amenable to improvement with appropriate management [10].

Measuring quadriceps strength (isometric or otherwise) and/or endurance requires specific equipment and expertise, which limits this assessment to clinical research settings or a few highly specialised departments. To address this issue, quantitative data obtained from a range of lower limb imaging techniques, including computed tomography (CT) [11], magnetic resonance imaging (MRI) [12] and ultrasonography [13, 14], have been used to provide an indirect approach to assess quadriceps function. However, as with physiological measurements, these methods require additional examinations beyond routine management and are rarely conducted in current practice.

Chest CT scans are systematically used in routine management of patients with COPD, particularly for screening for lung cancer, quantifying emphysema and detecting bronchiectasis [15, 16]. The cross-sectional pectoral muscle area (PMA) measured on a chest CT scan is a parameter that is increasingly studied and is highly correlated with fat-free mass in COPD [17, 18]. Associations have also been reported between PMA and several meaningful end-points in COPD [19–23], but PMA has never been used as a surrogate for quadriceps strength.

Therefore, in a prospective cohort of outpatients with varying severity grades of COPD, our primary objective was to evaluate the relationship between PMA measured on chest CT scans and quadriceps strength. In addition, we sought to assess whether PMA could be used to identify patients with the lowest quadriceps strength (lowest tercile).

Methods

Study design and population

The patients included in this retrospective study were from prospective cohorts of COPD patients followed at Grenoble University Hospital (Grenoble, France). Adult patients with mild to severe COPD, defined according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria, who presented to our outpatient department, were offered the opportunity to be included in Project COPD (NCT00404430), with inclusion between January 2007 and January 2017, and ECO-COPD (NCT03014609), with inclusion from January 2017, as described in the supplementary methods. Written informed consent was obtained from all participants. The studies were conducted in accordance with applicable good clinical practice requirements in Europe, French law, the International Council for Harmonisation E6 recommendations, and the ethical principles of the Declaration of Helsinki, and were approved by an independent ethics committee (Comité de Protection des Personnes, Grenoble IRB0006705).

The current analysis included COPD patients ≥40 years old, current or former smokers with a smoking history of ≥10 pack-years, for whom the database contained at least a chest CT scan allowing measurement of the PMA and the result of a quadriceps isometric force measurement performed within ±10 days of the chest CT scan.

Quadriceps strength assessment

To measure quadriceps strength, the subjects were sitting on a chair, with the knee joint of the dominant leg positioned at a 90° angle, as described elsewhere [24]. The subject was positioned in the chair in such a way as to avoid supporting the back and pelvis against the back-rest to minimise the contribution of agonist muscles other than the quadriceps. An inextensible, stretch-sensitive strain gauge (Dempo Technologies, model ABA, series 206683, maximum capacity 300 kg) was attached to the chair. The muscle length was maintained during contraction, enabling isometric force to be measured using the electrical signal generated by the strain gauge [24, 25]. This signal was sent to a dedicated unit (Globus Italia, Ergo Meter no. 10085, DC, 6V1W). Quadriceps strength was recorded in kilograms as the highest value of maximal voluntary strength recorded over three trials, with each trial separated by a recovery period of ≥2 minutes [6, 26].

CT scan measurements of PMA

For CT scan imaging, patients were positioned supine, with their arms raised above their heads. Acquisitions were performed during breath-hold at the end of inspiration, without injection of contrast. The PMA was the aggregate area (in cm²) of the right and left pectoralis major and minor muscles [20]. Two separate, blinded radiologists assessed PMA from a single axial inspiratory CT slice at the aortic arch. Predefined attenuation ranges from −50 to 90 HU were quantitatively assessed using a thoracic imaging platform (www.chestimagingplatform.org) on 3D Slicer 4.11 software. These assessors were blinded to the results of all other measurements.

Statistical analysis

The normality of the distribution of the variables was tested using the Shapiro–Wilk test after a visual inspection of distribution plots. The intraclass correlation coefficient was used to assess the reliability of the PMA measurements performed by the two radiologists. The relationship between measured quadriceps strength and PMA was assessed using simple linear regression. To derive estimated quadriceps strength from PMA, we used a multivariate analysis with a multiple linear model. The selection of covariates was based on a pragmatic approach for the primary outcome. As our aim was to use only parameters that can be obtained from a CT scan (and not by any other clinical measure) to estimate quadriceps strength, the only covariates tested were age and sex. A correlation matrix verified the absence of strong collinearity between the explanatory variables. The proportion of variance explained by the model was measured by the adjusted r². A concordance analysis between estimated and measured quadriceps strength was performed using a Bland–Altman plot. A receiver operating characteristic (ROC) curve analyses of PMA was used for the prediction of low quadriceps strength (lowest tercile of the whole population). p<0.05 was considered significant. The Bland–Altman plot was obtained using Microsoft Excel; all other statistical analyses were performed using R version 2024.09.0.

Results

Subjects

A total of 82 patients fulfilling the inclusion criteria were studied (supplementary figure 1), with no missing data. The characteristics of the subjects are presented in table 1. The median values for PMA and quadriceps strength for men and women are shown in table 2. The sample included the full spectrum of GOLD severity stages and five levels of dyspnoea, classified according to the modified Medical Research Council scale. Men and women were similar in age; nevertheless, men had significantly higher values for PMA and quadriceps strength than women (table 2). The agreement between the two radiologists for PMA measurement was excellent, with an intraclass correlation coefficient of 0.96 (95% CI 0.95–0.97).

TABLE 1.

Characteristics of the 82 patients with COPD included in the study

Men/women	50 (61%)/32 (39%)
Age, years	65±9
Height, cm	166±8
Weight, kg	76±18
Body mass index, kg·m−2	27.2±5.8
mMRC dyspnoea scale
Grade 0	26 (32%)
Grade 1	27 (33%)
Grade 2	15 (18%)
Grade 3	5 (6%)
Grade 4	9 (11%)
GOLD stage
GOLD 1	21 (26%)
GOLD 2	42 (51%)
GOLD 3	17 (20%)
GOLD 4	2 (3%)
FEV1, % predicted	64±19
FVC, % predicted	93±16
Total lung capacity, % predicted	117±20
Quadriceps strength, kg	22.3±9.8
Pectoral muscle area, cm2	29.7±9.3

Data are expressed as n (%) of patients or mean±sd, unless otherwise stated. mMRC: Modified Medical Research Council; GOLD: Global Initiative for Chronic Obstructive Lung Disease; FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity.

TABLE 2.

Median (interquartile range) values for pectoral muscle area (PMA) and quadriceps strength in men and women

Characteristics	Men (n=50)	Women (n=32)	p-value
Age, years	65.6 (63.1–68.1)	62.9 (59.8–66.02)	0.14
PMA, cm2	34.1 (31.7–36.5)	22.7 (20.9–24.5)	<0.001
Quadriceps strength, kg	26.3 (23.6–29.0)	16.1 (13.8–18.3)	<0.001
Height, cm	170.6 (168.9–172.3)	160.6 (158.6–162.5)	<0.001
Weight, kg	80.6 (75.7–85.5)	67.1 (62.1–72.2)	0.002

Relationship between PMA and measured quadriceps strength

The linear relationship between PMA (in cm²) and quadriceps strength (in kg) is illustrated in figure 1. The results demonstrate a significant positive correlation between these two variables, with an estimated regression coefficient of 0.60 (95% CI 0.40–0.79; p<0.001) and r²=0.32.

FIGURE 1.

Unadjusted correlation between pectoral muscle area and quadriceps strength in patients with stable COPD. The standardised coefficient (β) with associated 95% confidence interval and p-value is shown.

Multiple linear regression analysis revealed significant relationships between quadriceps strength and the prespecified predictors, i.e. sex, age and PMA (table 3). Men had a mean quadriceps strength value 6.35 kg (95% CI 2.01–11.00 kg) higher than that of women (p=0.004). Regarding age, each additional year was associated with an average decrease in quadriceps strength of 0.29 kg (95% CI −0.48– −0.10 kg; p=0.003). Finally, each 1 cm² increase in PMA was associated with an average increase in quadriceps strength of 0.41 kg (95% CI 0.18–0.63 kg; p<0.001). The model estimating quadriceps strength based on age, sex and PMA yielded an r² value of 0.43 and could be expressed as follows:

$$\mathrm{eQS}=25.2+0.41\times \mathrm{PMA}(\mathrm{c}{\mathrm{m}}^2)-0.29\times \mathrm{age}(\mathrm{years})+6.35(\mathrm{if}\mathrm{male})$$

TABLE 3.

Multiple linear regression for estimating quadriceps strength (adjusted r²=0.43)

Characteristics	β (95% CI)	p-value
Intercept	25 (12–39)	<0.001
PMA, cm2	0.41 (0.18–0.63)	<0.001
Age, years	−0.29 (−0.48– −0.10)	0.003
Sex	6.4 (2.1–11.0)	0.004

β: regression coefficient; PMA: pectoral muscle area.

As illustrated in figure 2, there was strong agreement between measured and estimated quadriceps strength, with a bias close to zero. Although a trend towards increasing bias for the highest quadriceps forces was observed, >95% of the data points were within the 95% agreement.

FIGURE 2.

Bland–Altman plot illustrating the concordance analysis between quadriceps strength (QS) and estimated quadriceps strength (eQS) from sex, age and pectoral muscle area. The solid line represents the bias, and the dashed lines represent the limits of agreement.

Using PMA to identify COPD patients with the lowest quadriceps strength tercile

The area under the ROC curve for identifying the lowest tercile of quadriceps strength using PMA was 0.76 (95% CI 0.65–0.87), with a PMA threshold of 28.5 cm² (figure 3).

FIGURE 3.

Receiver operating characteristic curve analysis of pectoral muscle area (PMA) for predicting low quadriceps strength (lowest tercile of the whole population). The red dot represents the optimal PMA threshold of 28.5 cm², providing a sensitivity of 0.68 (horizontal dashed line) and a specificity of 0.79 (vertical dashed line). AUC: area under the curve.

Discussion

In this study, we demonstrate the reliability of PMA measured on a chest CT scan performed as part of routine care for identifying quadriceps weakness. We also derived an equation to estimate quadriceps strength from PMA.

It is well established that quadriceps strength is reduced in patients with COPD, even those who are noncachectic, compared with healthy individuals [27]. In addition, quadriceps strength independently predicts increased healthcare utilisation and mortality in COPD patients [5, 28]. However, despite its potential usefulness for patient management, quadriceps strength assessment requires specific and sometimes cumbersome equipment, and it involves the active participation of patients to achieve a maximum voluntary contraction. To overcome these difficulties, several strategies have been proposed to estimate quadriceps strength from other parameters, most often measured using imaging techniques. Some of these strategies allow direct characterisation of the quadriceps, either by CT scans [29], MRI [30] or ultrasound [31, 32]. Other strategies involve estimating quadriceps strength from measurements of total body composition, either by dual-energy X-ray absorptiometry (DEXA) [33, 34] or bioelectrical impedance [35]. As discussed below, each of these techniques has demonstrated value for predicting quadriceps strength. However, unlike our method of assessing PMA from a chest CT scan performed as part of routine care, these other methods incur additional costs and require patients to undergo additional examinations. In addition, thigh CT scans and DEXA expose patients to ionising radiation, making repeated longitudinal measurements undesirable. MRI avoids this problem, but limited accessibility and long examination times are relevant constraints. Although less demanding than MRI and not involving radiation exposure, ultrasound and bioelectrical impedance require time from both the patient and examiner, specific expertise and dedicated facilities, making them potentially difficult to use in routine practice.

Our model, which estimates quadriceps strength based on age, sex and PMA, achieved an r² value of 0.43. In addition, it was observed that the agreement between measured and estimated quadriceps strength was better the lower the quadriceps strength. This is advantageous and relevant for clinical practice because PMA would be used to identify patients with low quadriceps strength. The accuracy of our estimation of quadriceps strength using PMA is at least as good as those reported using other methods. In 34 patients with COPD, a significant positive correlation was found between thigh muscle cross-sectional area measured on CT scan and quadriceps strength (r²=0.42; p<0.0001), suggesting a relationship between quadriceps size and strength in COPD [36]. Using MRI in a small group of 10 patients with COPD, only a moderate correlation was found between the cumulative cross-sectional area of several muscles of the limb (quadriceps, hamstrings, triceps surae and tibialis anterior) and quadriceps strength (r²=0.18) [37]. Nevertheless, in this group of patients, a strong negative correlation was found between muscle quality estimated from the lipid/total proton ratio and quadriceps strength (r²=0.53), suggesting that poor muscle quality may correlate better with muscle weakness than muscle atrophy in COPD [37]. In a pilot study of 30 patients with COPD, the cross-sectional rectus femoris muscle area measured by ultrasound shared a strong linear relationship with quadriceps strength (r²=0.60) [31]. More recently, in a large group of 169 patients with COPD, a weaker relationship was found between cross-sectional muscle area measured by ultrasound and quadriceps strength, with r²=0.26 [32]. Regarding models estimating quadriceps strength in COPD patients using body composition measurements, a three-compartment model, including appendicular skeletal muscle index, bone mineral content and android/gynoid percentage fat mass measured with DEXA, found r²=0.47 in men and r²=0.51 in women after adjustment for age and forced expiratory volume in 1 s [38]. Finally, in a large group of 502 patients with stable COPD, several parameters measured by bioelectrical impedance including angle phase, fat-free mass and fat-free mass index displayed a positive correlation with quadriceps strength, with r² values of 0.44, 0.42 and 0.25, respectively [39].

Our results suggest that PMA could help to identify COPD patients with low quadriceps strength, as has previously been suggested for patients with interstitial lung disease [40]. This should be considered relevant because such patients could be offered interventions to improve muscle mass and function. Indeed, it has been shown that the ratio between quadriceps strength and thigh cross-sectional area is similar in patients with COPD and in healthy individuals [36], suggesting that the muscle contractile apparatus is preserved in COPD and that restoring muscle mass should normalise muscle strength. However, it should not be ignored that muscle mass is not the only determinant of strength, and that metabolic and histological alterations, which are frequent in COPD [41], may contribute to reduced strength but cannot be identified by PMA.

Limitations of the study

There are several limitations in our study. First, when patient inclusion began in the cohorts, quadriceps strength was recommended to be expressed in kilograms [42]. Data in kilograms do not consider the lever arm and do not allow the force to be measured in terms of torque in N·m, as is now recommended for expressing quadriceps strength [43]. Second, it is possible that our results cannot be generalised to patients with very severe COPD, as such patients were under-represented in our cohort. Third, our quadriceps strength prediction equation and the PMA threshold for identifying COPD patients with low quadriceps strength have not been validated in an independent cohort. Fourth, although all patients were stable at the time of assessment, recent COPD exacerbations were not systematically recorded and were not an exclusion criterion. This may have introduced variability in muscle function. Fifth, as we have shown that age and sex contribute considerably to the significant correlation between PMA and quadriceps strength, it is likely that the PMA threshold we determined from the ROC curve to identify low quadriceps strength may lead to identifying older women (who also have the lowest quadriceps strength), rather than necessarily identifying patients with “abnormal” muscle strength. An alternative approach involving two ROC curves, one for men and one for women, and using the quadriceps strength/age quotient could overcome this limitation. However, for such an analysis, the number of patients of each sex in our cohort is insufficient to provide precise thresholds.

Conclusion

Our analysis suggests that the use of PMA measured on chest CT scans performed as part of routine care for pulmonary monitoring of patients with COPD could facilitate better identification of those who are eligible for interventions targeting lower limb muscle function. This represents a promising example of augmented interpretation of chest CT scans and could form the basis for a phenotyping tool in routine clinical practice. In the near future, studies measuring pectoral muscle density in addition to PMA could enable even more accurate identification of COPD patients with muscle weakness. In a context where chest CT scans will be increasingly performed, particularly in the context of lung cancer screening, it seems essential to clearly define and critically evaluate the roles and limitations of such a tool.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material

00380-2025.SUPPLEMENT