Abstract
Objective:
We investigated whether total psoas muscle area (TPMA) was representative of the total psoas muscle volume (TPMV). Secondly, we assessed whether there was a relationship between the two commonly used single slice measurements of sarcopenia, TPMA and total abdominal muscle area (TAMA).
Methods:
Pre-operative CT imaging of 110 patients undergoing elective endovascular aneurysm repair were analyzed by two trained independent observers. TPMA was measured at individual vertebral levels between the second lumbar vertebrae and sacrum. TPMV was also estimated between the second lumbar vertebrae and sacrum. TAMA was measured at the third lumbar vertebrae (L3). Observer differences were assessed using Bland–Altman plots. Associations between the different measures were assessed using linear regression and Pearson’s correlation.
Results:
We found single slice measurements of the TPMA to be representative of the TPMV at individual levels between L2 to the sacrum. The strongest association was seen at L3 [adjusted regression coefficient 16.7, 95% confidence interval (12.1 to 21.4), p < 0.001]. There was no association between TPMA and TAMA [adjusted regression coefficient −0.7, 95% confidence interval (−4.1 to 2.8), p = 0.710].
Conclusion:
We demonstrate that measurements of the TPMA between L2 to the sacrum are representative of the TPMV, with the greatest association at the third lumbar vertebrae. There was no association between the TPMA and TAMA.
Advances in knowledge:
We demonstrate that a single slice measurement of TPMA at L3 is representative of the muscle volume, contrary to previous criticism. Future sarcopenia studies can continue to measure TPMA which is representative of the TPMV.
Introduction
Sarcopenia is a condition that is categorized by low skeletal muscle mass and declining function. It is associated with worse patient outcomes and is an increasing problem due to a rapidly ageing population.1,2 The European Working Group on Sarcopenia in Older People have recognised this and have recommended the routine assessment for sarcopenia in all patients aged above 65 years.2 Therefore, the quantification of core abdominal muscles as a surrogate marker for sarcopenia has become an increasingly popular in clinical research. The total psoas muscle area (TPMA) or the total abdominal muscle area (TAMA) from single slice computed tomography (CT) imaging at the endplates of the third or fourth lumbar vertebrae are the two commonest methods of quantifying sarcopenia.1 However, at present the optimum vertebral level to measure the TPMA is yet to be validated.
As muscles are complex three-dimensional structures, the use of single slice two-dimensional measurements of these may be criticized as being a poor representation to assess sarcopenia. The association between single slice TAMA and volumetric measurements of total abdominal muscle on CT imaging has been previously described and validated.3–5 Furthermore gender-specific cut-off values for TAMA have also been proposed which have been demonstrated to correlate with mortality.6 The assessment of the patients muscle volume using an automated quantification method with specialist software has been shown to be a better prognostic marker than area alone.7 However, similar correlations are yet to be validated between single slice measurement of TPMA and three dimensional volumetric assessment of the psoas muscle.6
The heterogeneity of the different methods of quantifying sarcopenia, TPMA and TAMA, currently make it difficult to compare and derive cut-off values for sarcopenia from single slice imaging. Measuring TAMA is more time consuming and often requires the use of specialist software. However, we have previously demonstrated that the TPMA can be easily and consistently measured on any picture archiving communications system viewer.8,9 The identification of an interchangeable relationship between TPMA and TAMA may facilitate comparative analysis of previously reported outcomes and derive clinically uniform cut-off values defining sarcopenia applicable to the general patient population.1
In this study, we investigated whether there was a relationship between total psoas muscle volume (TPMV) and TPMA, and identify the vertebral level at which the TPMA is most representative of the muscle volume. Secondly, we investigated the relationship between single slice measurements at the third lumbar vertebrae of TPMA and TAMA.
Methods
We analysed prospectively collected data from patients who have had an elective endovascular aneurysm repair (EVAR) for an abdominal aortic aneurysm (AAA).
Study population
We randomly selected pre-operative abdominal CT angiogram (CTA) scans routinely performed as part of the assessment for intervention. All scans were performed in the supine position with a breath-hold to minimize motion artefact. Patients were all identified from the Health Quality Improvement Partnership National Vascular Registry, a prospectively maintained database, from January 2008 and December 2014.10 Inclusion into the study required the patient to have an abdominal CT with the psoas muscle clearly identifiable from the second lumbar vertebrae to the sacrum. Patients were excluded if they had incomplete imaging with missing portions. Ethical approval was granted by the local radiology research authorization group and Health Research Authority (IRAS project ID, 228484).
Covariate assessment
Data were reviewed from the National Vascular Registry for baseline age, gender, height and weight. These are all parameters that are routinely collected, however, medical records were also reviewed to ensure all data collected were accurate.
Image analysis
Imaging was performed using a Siemens Somatom Definition AS CT scanner with the patient in the supine position with a breath-hold to minimize motion artefact. Slice thicknesses were between 1 and 2.5mm. Scans were assessed for inclusion by a single investigator, who did not participate in any images analysis, using the picture archiving and communications system viewer IMPAX (AGFA-Gevaert Group, Mortsel, Belgium) and ImageJ (National Institute of Health, Bethesda). Two independent observers (Rater 1 [R1] and Rater 2 [R2]) were trained by an investigator with prior expertise. R1 was a clinician with 2 years post-graduate clinical experience and R2 was a post-graduate research fellow with 3 years post-graduate clinical experience. The trainer was a surgical fellow with 5 years postgraduate clinical experience.
TPMA was measured by manually tracing around the area of the left and right psoas muscle at each vertebral level of the transverse processes from L2 to the sacrum (Figure 1). TPMV was calculated by multiplying each individual TPMA by the distance between the corresponding vertebral levels. TAMA was measured using a fully automated technique. In brief, single slice images at the third lumbar vertebrae were downloaded in the digital imaging and communications in medicine format with preservation of actual dimensions to avoid magnification indices and scales. Analysis was subsequently performed using ImageJ (National Institute of Health, Bethesda) by setting the Hounsfield unit range between −30 and 130.8 TAMA was calculated by measuring all the abdominal muscles, namely: psoas, erector spinae, quadratus lumborum, transversus abdominis, external and internal oblique’s, and rectus abdominis muscles based on the total pixel densities (Figure 2). We acknowledge that there are a variety of other software packages and methods that may also be used.
Figure 1.
Illustration of TPMV calculated from L2 to sacrum. Red line shows manual tracing of TPMA. Example shows TPMA at L2 (4.2 cm2), L3 (10.0 cm2), L4 (16.8 cm2), L5 (22.4 cm2), Sacrum (23.1 cm2). Distances: 2.9 cm between L1 and L2, 3.1 cm between L2 and L3, 4.4 cm between L3 and L4, 5.0 cm between L4 and L5, 5.3 cm between L5 and sacrum. Therefore, the TPMV calculated is 351.5 cm3. TPMA, total psoas muscle area; TPMV, totalpsoas muscle volume.
Figure 2.
TAMA measurement using the automated technique. Red highlights tissue matching pixel density with HU between −30 to 130. TAMA, total abdominal muscle area; HU, Hounsfield unit.
Statistical analysis
Measurements were made in centimetres (cm). TPMA and TAMA are reported as an area (cm2) and TPMV as a volume (cm3). Statistical analyses were performed using Minitab 17 (Minitab Inc., Pennsylvania).11 Continuous variables were checked for normality and reported as a mean and standard deviation or 95% confidence interval (CI). Non-parametric data was reported as a median and interquartile range (IQR). Categorical variable are reported as absolute numbers (n) and percentages (%), and were compared using the χ2 test. Statistical significance was defined as a two-tailed p-value < 0.05.
50 images were analysed by two trained independent investigators (R1 and R2) to measure the TPMA, TAMA and TPMV. Repeat measurements of all images were made by Rater 1 (e.g. R1a and R1b). Intra observer and inter observer differences were evaluated using Bland–Altman plots and differences in measurements evaluated using Student’s t-test. The limits of agreement were calculated as two standard deviation from the mean difference calculated between observer measurements.
We calculated that we would require at least 85 patients in this study in order to assess for a correlation of greater than 0.7 with 95% significance at 80% power. All images were analyzed by R1. Comparisons between the individual methods of assessing core muscle were evaluated using measurements recorded by Rater one as single observer measurements are likely to represent clinical practice. The relationship between TPMV and TPMA: at the second lumbar vertebrae (L2), third lumbar vertebrae (L3), fourth lumbar vertebrae (L4), fifth lumbar vertebrae (L5) and sacrum were assessed using Pearson’s correlation and linear regression analysis. Similarly, the relationship between TPMA vs TAMA were assessed using Pearson’s correlation and linear regression analysis. All regression analyses were adjusted for age and sex.
Results
Patient characteristics
In total, CT scans from 110 consecutive patients were analysed in this study. Scans were performed between October 2008 and July 2014. The median age of patients was 77.5 years (IQR 71.3–81.0) and there were 96 (87.3%) males. The median height was 174.0 cm (IQR 165.5–179.0) and the medium weight was 77.0 kg (IQR 69.0–95.0).
Intra- and interobserver differences
Intra- and interobserver difference are shown in Figure 3. No significant differences in measurements were identified between observers (Table 1). Single observer variation (R1) was lowest with TAMA (mean 309.1 cm2 ±0.1 [0.03%]), followed by TPMA (mean 8.8 cm2 ±0.3 [3.7%]) and then TPMV (mean 337.4 cm3 ±13.6 [4.0%]).
Figure 3.
Bland–Altman plots showing intra- and inter- observer differences in measurements of TPMA, TAMA and TPMV. TAMA, total abdominal muscle area; TPMA, total psoas musclearea; TPMV, total psoas muscle volume.
Table 1.
Observer measurements of TPMA, TAMA and TPMV
| Measurement | R1a Mean (SD) |
R1b Mean (SD) |
R2 Mean (SD) |
Intraobserver differences | Interobserver differences | ||
| Mean (SD) | p-value | Mean (SD) | p-value | ||||
| TPMV [cm3] |
337.6 (105.3) |
337.1 (101.0) |
338.8 (106.6) |
0.5 (13.6) |
0.981 | −1.1 (11.8) |
0.957 |
| TPMA [cm2] |
8.8 (3.5) |
8.9 (3.6) |
8.8 (3.4) |
−0.03 (0.3) |
0.963 | 0.03 (0.3) |
0.924 |
| TAMA [cm2] |
309.1 (66.8) |
309.0 (66.7) |
309.1 (66.8) |
0.1 (0.1) |
0.997 | −0.003 (0.01) |
1.000 |
TAMA, total abdominal muscle area; TPMA, total psoas muscle area; TPMV, total psoas muscle volume.
Comparative analysis
Measurements of TPMV, TPMA and TAMA were all normally distributed. Mean measurements of the TPMA at each vertebral level, L2 to the sacrum, are highlighted in Table 2. As anticipated, males had larger psoas muscle size than females (Table 2). The mean TPMV was 334.0 ± 101.6 cm3 and the mean TAMA was 303.7 ± 62.2 cm2. The mean distance between measurements at the: sacrum and L5 was 4.9 ± 1.0 cm, L5 to L4 was 4.5 ± 0.9 cm, L4 to L3 was 4.3 ± 0.8 cm, L3 to L2 was 3.3 ± 0.9 cm and L2 to first lumber vertebrae was 3.2 ± 0.9 cm.
Table 2.
Measurements of TPMA from L2 to the sacrum by gender
| Spinal level | TPMA | |||
| Male (N = 96) |
Female (N = 14) |
p-value | Overall | |
| L2 [cm2] | 2.6 ± 1.8 | 1.7 ± 1.4 | 0.058 | 2.5 ± 1.8 |
| L3 [cm2] | 9.4 ± 3.1 | 6.3 ± 2.4 | <0.001 | 9.0 ± 3.2 |
| L4 [cm2] | 16.9 ± 4.2 | 12.2 ± 3.7 | <0.001 | 16.3 ± 4.4 |
| L5 [cm2] | 23.8 ± 5.2 | 17.0 ± 4.5 | <0.001 | 22.9 ± 5.6 |
| Sacrum [cm2] | 25.6 ± 5.7 | 18.5 ± 5.3 | <0.001 | 24.7 ± 6.1 |
TPMA, total psoas muscle area.
Significant positive correlations were observed between measurements of TPMV and TPMA at all levels, L2 to sacrum (Figure 4). Regression analysis highlighted significant associations between measurements of the TPMA, from L2 to the sacrum, with TPMV (Table 3). The greatest association between single slice measurements and volumetric analysis was seen at L3 [unadjusted regression coefficient 20.1, 95% CI (15.4 to 24.9), p < 0.001, adjusted regression coefficient 16.7, 95% CI 12.1 to 21.4, p < 0.001].
Figure 4.
Scatter graph illustrating relationship between TPMV and single slice TPMA measurements at: (a) L2, (b) L3, (c) L4, (d) L5 and (e) sacrum. TPMA, total psoas muscle area; TPMV, totalpsoas muscle volume.
Table 3.
Linear regression analysis comparing measurements of TPMA from L2 to sacrum in relation to TPMV
| Spinal level | Unadjusted | Adjusted | ||||
| Regression coefficient | 95% CI | p-value | Regression coefficient | 95% CI | p-value | |
| L2 | 16.7 | 6.2–27.2 | 0.002 | 12.1 | 2.6–21.5 | 0.013 |
| L3 | 20.1 | 15.4–24.9 | <0.001 | 16.7 | 12.1–21.4 | <0.001 |
| L4 | 17.7 | 14.9–20.6 | <0.001 | 15.6 | 12.5–18.7 | <0.001 |
| L5 | 14.1 | 12.0–16.3 | <0.001 | 12.9 | 10.3–15.5 | <0.001 |
| Sacrum | 11.8 | 9.5–14.0 | <0.001 | 10.3 | 7.6–12.9 | <0.001 |
CI, confidence interval; TPMA, total psoas muscle area; TPMV, total psoas muscle volume.
Adjusted for age and sex.
There was no significant correlation between TPMA and TAMA (Figure 5). No association was evident when comparing measurements of TPMA and TAMA [unadjusted regression coefficient 2.9, 95% CI (−0.8 to 6.6), p = 0.126 and adjusted regression coefficient −0.7, 95% CI (−4.1 to 2.8), p = 0.710].
Figure 5.
Scatter graph demonstrating the relationship between TPMA and TAMA. TAMA, total abdominal muscle area; TPMA, total psoas musclearea.
Discussion
In this study, we demonstrate single slice measurements of the TPMA to be representative of the TPMV on CT imaging. Measurements of the TPMA at any vertebral level between L2 to the sacrum were found to be significantly representative of the TPMV. Therefore, it may be plausible to use measurements of TPMA at any of these levels when limited by the images available from routine imaging. This is important as it may not always be possible to measure the TPMA at the L3 vertebrae if the CT sequence has not captured this section. However, it is important to appreciate that the observed regression coefficients and confidence interval were almost identical for TPMA measurements at L3 and L4 therefore utilization of measurements at either level is acceptable.
Our research adds to the growing body of evidence utilising the measurements of core abdominal muscles from imaging as a surrogate marker for sarcopenia. Shen et al previously demonstrated the association between measurements of the TAMA 5 cm above the L4–L5 junction to be associated with volumetric measurements of the abdominal muscle volume on CT imaging in a healthy cohort of patients.3 These finding were confirmed by Mourtzakis et al who demonstrated the relationship between single slice imaging of fat free mass to whole body fat free mass.4 However, it must be noted that these measures were validated on either a normal cohort of people or cancer patients. Despite the correlation between single slice TAMA and abdominal muscle volume, few studies have demonstrated the correlation between single slice measurements of the TPMA and the TPMV. In this study, we have demonstrated a relationship between TPMA and TPMV, and have identified that the measurement of TPMA at L3 (the most widely used level) is most representative of the TPMV. Our group has previously shown that the measurement of TPMA is reproducible and independent of observer bias.8
Similar to the findings of Rutten et al, we demonstrate no correlation between single slice measurements of TPMA and TAMA.12 Therefore, studies evaluating outcomes in relation to sarcopenia utilising these different measures cannot be reliably compared. We have previously reported that measurements of the TPMA may be easily performed by manually tracing around the psoas muscle at the third lumbar vertebrae without the need for specialist software or clinical experience.8 Therefore, we would recommend the routine measurement of the TPMA for the quantification of sarcopenia as it may be easily utilised in clinical practice.
Similar to previous studies, we demonstrate that females have lower measurements of the psoas muscle in comparison to males. The number of females in our study were too few to allow for assessment of whether these relationships identified by gender were statistically relevant. However, it is important to acknowledge that the study was powered to detect the associations between TPMA and TPMV as well as TPMA and TAMA.
However, measurements of TPMA had the lowest variation when compared to TPMV as demonstrated by percentage variation of the standard deviation against the mean intraobserver measurements. This finding is expected as due to the compounding effect of errors accumulating at each level analyzed. The utilisation of the TPMA as an assessment tool for sarcopenia instead of TPMV may also facilitate and reduce the numbers needed for patient recruitment when powering future prospective studies. Although TAMA had the least variation, this was primarily due to a completely autonomous method of measurement which may not be routinely applicable out of the research setting.
It is important to acknowledge that our study estimates the TPMV as we did not utilize specialist software as described in previous studies.13 The TPMV was calculated as blocks based on two-dimensional assessment of the psoas muscle at each vertebral level and then multiplying the area by the distance between vertebral levels. This was intentional as we wanted to replicate the potential real-life application of risk stratification using the psoas muscle as not all clinicians in the National Health Service in the United Kingdom have access to specialist software for this type of image analysis. We acknowledge that this might be a crude measurement given the large interslice distances.
Conclusions
In conclusion, we demonstrate an association between measurements of TPMA and TPMV. Measurements of TPMA may be made at any vertebral level between the sacrum and L2 and be reflective of the TPMV, with the greatest association at L3. We also demonstrate the absence of any association between TAMA and TPMA therefore outcomes reported with either measure cannot be reliably compared and results translated.
Footnotes
Acknowledgment: There are no conflicts of interest. No external funding sources were required to conduct this work. MAW and MD are clinical research training fellows supported by the British Heart Foundation. MAB is a intermediate research fellow supported by the British Heart Foundation.
Contributor Information
Mohammed Abdul Waduud, Email: m.a.waduud@leeds.ac.uk.
Pratik Adusumilli, Email: p.adusumilli@nhs.net.
Michael Drozd, Email: m.drozd@leeds.ac.uk.
Marc A Bailey, Email: m.a.bailey@leeds.ac.uk.
Gary Cuthbert, Email: gary.cuthbert@nhs.net.
Christopher Hammond, Email: christopherhammond@nhs.net.
REFERENCES
- 1. Jones K , Gordon-Weeks A , Coleman C , Silva M . Radiologically determined sarcopenia predicts morbidity and mortality following abdominal surgery: a systematic review and meta-analysis . World J Surg 2017. ; 9 : 1 – 14 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Cruz-Jentoft AJ , Baeyens JP , Bauer JM , Boirie Y , Cederholm T , Landi F , et al. . Sarcopenia: European consensus on definition and diagnosis: report of the European Working Group on sarcopenia in older people . Age Ageing 2010. ; 39 : 412 – 23 10.1093/ageing/afq034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Shen W , Punyanitya M , Wang Z , Gallagher D , St-Onge M-P , Albu J , et al. . Total body skeletal muscle and adipose tissue volumes: estimation from a single abdominal cross-sectional image . J Appl Physiol 2004. ; 97 : 2333 – 8 . doi: 10.1152/japplphysiol.00744.2004 [DOI] [PubMed] [Google Scholar]
- 4. Mourtzakis M , Prado CMM , Lieffers JR , Reiman T , McCargar LJ , Baracos VE . A practical and precise approach to quantification of body composition in cancer patients using computed tomography images acquired during routine care . Appl Physiol Nutr Metab 2008. ; 33 : 997 – 1006 . doi: 10.1139/H08-075 [DOI] [PubMed] [Google Scholar]
- 5. Gomez-Perez SL , Haus JM , Sheean P , Patel B , Mar W , Chaudhry V , et al. . Measuring abdominal circumference and skeletal muscle from a single cross-sectional computed tomography image: A step-by-step guide for clinicians using National Institutes of health ImageJ . J Parenter Enteral Nutr 2016. ; 40 : 308 – 18 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Prado CMM , Lieffers JR , McCargar LJ , Reiman T , Sawyer MB , Martin L , et al. . Prevalence and clinical implications of sarcopenic obesity in patients with solid tumours of the respiratory and gastrointestinal tracts: a population-based study . Lancet Oncol 2008. ; 9 : 629 – 35 . doi: 10.1016/S1470-2045(08)70153-0 [DOI] [PubMed] [Google Scholar]
- 7. Amini N , Spolverato G , Gupta R , Margonis GA , Kim Y , Wagner D , et al. . Impact total psoas volume on short- and long-term outcomes in patients undergoing curative resection for pancreatic adenocarcinoma: a new tool to assess sarcopenia . J Gastrointest Surg 2015. ; 19 : 1593 – 602 . doi: 10.1007/s11605-015-2835-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Waduud MA , Drozd M , Linton E , Wood B , Manning J , Bailey MA , et al. . Influences of clinical experience in the quantification of morphometric sarcopenia: a cohort study . Br J Radiol 2018. ; 91 : 1088 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Waduud MA , Wood B , Keleabetswe P , Manning J , Linton E , Drozd M , et al. . Influence of psoas muscle area on mortality following elective abdominal aortic aneurysm repair . Br J Surg 2019. ; 106 : 367 – 74 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Waton S , Johal A , Heikkilä K , Cromwell D , Loftus I . National Vascular Registry: 2015 annual report . The Royal College of Surgeons of England 2015. ;. [Google Scholar]
- 11. Minitab I . MINITAB release 17: statistical software for windows . USA: : The British Institute of Radiology. ; 2014. . [Google Scholar]
- 12. Rutten IJG , Ubachs J , Kruitwagen RFPM , Beets-Tan RGH , Olde Damink SWM , Van Gorp T . Psoas muscle area is not representative of total skeletal muscle area in the assessment of sarcopenia in ovarian cancer . J Cachexia Sarcopenia Muscle 2017. ; 8 : 630 – 8 . doi: 10.1002/jcsm.12180 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Zargar H , Almassi N , Kovac E , Ercole C , Remer E , Rini B , et al. . Change in psoas muscle volume as a predictor of outcomes in patients treated with chemotherapy and radical cystectomy for muscle-invasive bladder cancer . Bladder Cancer 2017. ; 3 : 57 – 63 . doi: 10.3233/BLC-160080 [DOI] [PMC free article] [PubMed] [Google Scholar]