Sarcopenia, characterized by reduced muscle mass and quality, is common in chronic respiratory diseases like interstitial lung diseases (ILD), particularly idiopathic pulmonary fibrosis (IPF) [1]. It is linked to poorer clinical outcomes and may benefit of early nutritional or rehabilitative interventions [2–7]. The gold standard for diagnosing sarcopenia involves specialized assessments by a nutritionist, but these may be delayed due to resource limitations [1]. Recent studies suggest that high-resolution computed tomography (HRCT)-derived thoracic muscle measurements could be an effective tool for diagnosing sarcopenia [2]. Specifically, single-slice cross-sectional area of pectoral muscles (PMA) at the 4th thoracic vertebra is a well-established method for assessing muscle mass and quality. HRCT provides advantages in ILD patients, as it is routinely used for diagnostic and follow-up purposes, allowing for body composition analysis without additional radiation or cost. This approach has been demonstrated to correlate with frailty and survival, particularly in oncologic and liver disease populations [8,9]. Our study aims to identify a cut-off value of PM index (PMI) in Caucasian patients with fibrotic ILDs to diagnose sarcopenia using chest HRCT.

We evaluated data from two cohort of patients: (i) an Italian cohort of consecutive adults with IPF from 4 Italian referral centers (Fondazione IRCCS San Gerardo dei Tintori, Monza; San Giuseppe Hospital, MultiMedica IRCCS, Milan; G. Salvini Hospital, Garbagnate Milanese and ASST Santi Paolo e Carlo, Milan) enrolled in the nutrIPF study (clinicaltrial.gov NCT03770845), whose clinical characteristics have been described before [10]; (ii) a Canadian cohort of consecutive adult patients with fibrotic ILDs (namely IPF, Hypersensitivity pneumonitis (HP) and idiopathic Non-specific Interstitial Pneumonia (NSIP)) from a single referral center in London, Canada, whose clinical characteristics have also been previously published [11].

For the purpose of this analysis we only included patients of Caucasian ethnicity and none of the patients was receiving systemic steroids at the time of the evaluation of sarcopenia.

The presence of sarcopenia was evaluated for both cohorts using the criteria developed by Baumgartner et al. [12], while for the Italian cohort the definition recommended by European Working Group on Sarcopenia in Older People 2 (EWGSOP2) 2019 (1) was also applied. We used the following references to define reduced Skeletal Muscle Index (SMI), measured through bioelectrical impedance analysis: SMI<8.87kg/m2 for males and <6.42kg/m2 for females [13].

In contrast, the EWGSOP2 2019 criteria, are based on both muscle mass, strength and physical performance, and allow for the identification of patients with probable sarcopenia, in addition to those with overt sarcopenia (confirmed or severe). Muscle strength was evaluated through hand grip strength for both dominant and non-dominant limbs: measurements were performed by hand-held dynamometer (KERN MAP 1.2 version), repeated three times for each side and the best value was recorded [14]. BIA was performed using a standard tetra-polar technique with patients studied in the supine position with electrodes connected to the hands and feet (seca mBCA 525) [15].

The measurement of the PMA at HRCT scan was performed by two radiologists with 9 (C.M.) and 7 (D.G.G.) years of experience that did not have access to patients’ information. The most recent unenhanced CT scan acquired within 3 months from the evaluation of sarcopenia was used for analysis. All CT examinations were performed on multi-detector scanners ranging from 64 to 256 slices. The chest CT acquisition parameters are reported in Supplementary Table 1.

Axial images at the level of the fourth thoracic vertebra (T4) were used for analysis. The total PMA was calculated as the sum of the bilateral pectoralis major and minor muscle areas. Muscle attenuation (density) within the same ROIs was measured in Hounsfield units. PMI was calculated as the ratio between PMA and height squared (cm2/m2).

This study received Ethics Committee approval (#1867, ASST Monza, October 2018 and Protocol No. 104028 and 103186, Western University Research Ethics Board).

A descriptive analysis was performed on the two cohorts using Fisher's exact test or the Mann–Whitney test for comparisons. In the Italian sample, the agreement between PMA measurements done by the two radiologists was evaluated using the Intraclass Correlation Coefficient (ICC).

PMI cut-offs for the prediction of overt sarcopenia were estimated separately for the two cohorts and on the whole pooled sample, by maximizing the Youden index for the area under the receiver operating characteristic (ROC) curve (AUC). Sensitivity and specificity of the identified cut-off was then assessed overall, and stratified by sex. 95% Confidence Intervals (95% CI) were computed through bootstrap method.

In the Italian cohort, we also conducted exploratory univariate multinomial logistic regression analyses using the three categories of sarcopenia as the outcome (no sarcopenia, probable sarcopenia, overt sarcopenia) and PMI as the predictor, in different forms.

In this study, we included two cohorts of Caucasian patients: the first cohort consisted of 59 Italian patients with IPF (median [IQR] age 72 [67–79] years, 20% female), and the second cohort included 28 Canadian patients with fibrotic interstitial lung disease (ILD), comprising 22 with IPF, 5 with NSIP, and 2 with HP. The median [IQR] age in the Canadian cohort was 71 [63–79] years, with 36% female. The clinical characteristics of the study cohorts are summarized in Table 1, while characteristics stratified by sarcopenic status are reported in Supplementary Table 2.

In the Canadian cohort, 12 patients (42.9%) met the criteria for overt sarcopenia according to the definition by Baumgartner et al. In the Italian cohort, 6 patients (10.2%) were classified as having overt sarcopenia according to the criteria by Baumgartner et al., while using the 2019 EWGSOP2 criteria, which allow for staging of sarcopenia, 8 patients were identified with probable sarcopenia, and 6 had overt sarcopenia (3 patients each were classified as having confirmed and severe sarcopenia).

The ICC was 0.99 (95% CI: 0.99–1.00), demonstrating excellent concordance between observers in regards to PMA measurements.

In the Italian cohort the optimal cut-off PMI value (95%CI) to identify patients with overt sarcopenia was 11.01 (10.01–14.81)cm2/m2, with an AUC of 0.76 (95%CI: 0.57–0.93), p-value 0.04, Fig. 1a. The sensitivity and specificity (95%CI) of the cut-off were 0.83 (0.50–1.00) and 0.79 (0.34–0.98), respectively. In the Canadian cohort, the optimal cut-off PMI value (95%CI) to identify patients with overt sarcopenia was 12.65 (8.81–13.18)cm2/m2, with an AUC of 0.76 (95%CI: 0.56–0.92), p-value 0.02, Fig. 1b. The sensitivity and specificity (95%CI) of the cut-off were 1.00 (0.58–1.00) and 0.63 (0.38–0.94), respectively. By merging the two datasets into a single cohort of 87 patients, the optimal cut-off PMI value (95%CI) to identify patients with overt sarcopenia was 12.90 (10.00–12.90)cm2/m2, with an AUC of 0.81 (95%CI: 0.70–0.90), p-value<0.001, Fig. 1c. The sensitivity and specificity (95%CI) of the cut-off were 0.94 (0.67–1.00) and 0.67 (0.54–0.93), respectively. When stratifying the sample by sex and applying the 12.90 cut-off, sensitivity and specificity were respectively 1.00 (1.00–1.00) and 0.40 (0.13–0.67) among females, and 0.91 (0.73–1.00) and 0.68 (0.56–0.80) among males.

Fig. 1.

Receiver operating characteristic (ROC) curve for PMI, reporting the Area Under the Curve (AUC) and the optimal cut-off point based on the Youden Index for the (a) Italian cohort, (b) Canadian cohort, (c) Italian+Canadian cohort.

In univariate multinomial logistic regression analyses PMI was considered as: (i) a continuous variable, (ii) a categorical variable using a cut-off of 12.9cm2/m2, and (iii) a categorical variable with two thresholds set at 11 and 13cm2/m2 (Table 2).

All three models confirmed an association between PMI and overt sarcopenia, with borderline statistical significance, more pronounced when using the categorical variables. No significant association was observed with probable sarcopenia.

Our study is the first to identify, in an international cohort of Caucasian patients with fibrotic ILDs, a PMI cut-off that predicts the presence of sarcopenia with good sensitivity and acceptable specificity. This finding has both strengths and limitations.

On one hand, this approach offers practical advantages: pulmonologists and radiologists can easily implement PMI assessment using chest HRCT routinely performed for IPF diagnosis and follow-up. Moreover, PMI measurement can be automated through artificial intelligence-based protocols [16]. The identified cut-offs showed high sensitivity with few false negatives but lower specificity with more false positives. As a screening rather than diagnostic tool, PMI may support early identification of ILD patients with sarcopenia and help select those requiring further evaluation according to 2019 EWGSOP2 criteria, also in light of multiple studies highlighting the importance of early identification of ILD patients with sarcopenia [17].

On the other hand, PMI alone may not detect probable sarcopenia and may be insufficient for early diagnosis. CT-based PMI assessment should therefore be complemented by physical performance measures such as the 6-minute walking test, 4-meter gait speed, or timed up and go.

The fact that the PMI cut-off identified to predict sarcopenia status differs slightly between the Italian (11.02cm2/m2) and Canadian (12.65cm2/m2) cohorts, despite both consisting of Caucasian patients with fibrotic ILDs, may be attributable to several factors. On one hand, there are currently no established parameters to standardize PMA values by weight and sex. As shown in the results, the Canadian cohort included a higher proportion of women and obese individuals compared to the Italian cohort. These differences may have contributed to the observed variation between the two populations.

Although our study is the first to propose PMI cut-off values for the prediction of sarcopenia, previous research has already demonstrated that PMA is a reliable surrogate for whole-body lean mass, and its reduction is associated with both markers of ILD severity, such as decreased forced vital capacity and diffusing capacity for carbon monoxide [18], and increased overall mortality in patients with IPF [4].

This study has several limitations. First, the small sample size prevented external validation and, although sensitivity and specificity were reported stratified by sex, the low sensitivity (0.40) observed in the female population may suggest that a lower cut-off would have been more appropriate for this group; however, the limited sample size and the low proportion of females among patients with IPF precluded further analysis.

To address this, we estimated bootstrap 95% confidence intervals for the cut-off and its performance measures to reflect result uncertainty and encourage validation in larger, multi-ethnic cohorts. Additionally, no validated formulas are available to adjust PMI for sex. Second, findings are limited to Caucasian patients with fibrotic ILD, reducing generalizability. Third, the 2019 EWGSOP2 criteria were applied only to the Italian cohort, while the Canadian cohort used a muscle mass–based definition, although patient selection substantially overlapped where both definitions were tested. This is due to the fact that the study was conducted prior to 2021; therefore, definitions of sarcopenia other than the 2019 EWGSOP2 criteria were also used. Fourth, our study included HRCT examinations acquired on different CT scanners with heterogeneous detector configurations and reconstruction algorithms. Although acquisition parameters were relatively standardized across centers, residual inter-scanner variability cannot be excluded and may partly contribute to the observed differences in PMI cut-off values between cohorts. Future multicenter studies with harmonized acquisition protocols are warranted to further improve the generalizability and reproducibility of the results.

In conclusion, a PMI cut-off of 12.90cm2/m2 showed high sensitivity (94%) and moderate specificity (67%) for predicting sarcopenia in Caucasian patients with fibrotic ILDs. Integrated into routine chest CT evaluation, PMI may serve as a screening tool for overt sarcopenia, but it remains limited in detecting early or probable disease without complementary assessments of muscle strength and physical performance.