Sarcopenia on imaging: what DXA, CT and MRI can show

Why it matters: Sarcopenia, the loss of muscle mass and strength with age or disease, can often be measured on scans patients already have. A 2023 review in RöFo explains how DXA, CT, MRI and ultrasound detect it, and why a muscle measurement on a routine staging CT can change how cancer and surgical patients are managed.

Not only a problem of old age

Sarcopenia affects roughly one in five 70-year-olds and about half of 75-year-olds, but it also develops in younger people with cancer, chronic liver disease, muscular dystrophy or long-term steroid treatment. It should be told apart from cachexia, which is defined by weight loss: more than 5% in six months, or more than 2% in people who are already underweight (BMI below 20 kg/m²) or sarcopenic. Muscle loss combined with excess fat, known as sarcopenic obesity, is linked to cardiometabolic risk and fatty liver disease and is seen even in children and adolescents.

Which scan shows what

DXA is quick, inexpensive and low-dose: a whole-body scan delivers about 1–10 µSv. It measures appendicular lean mass, and the EWGSOP thresholds for low muscle mass are an appendicular lean mass index below 6 kg/m² in women and below 7 kg/m² in men. Its limits are two-dimensional images, lower accuracy in obesity or oedema, and results that differ between scanner manufacturers.

CT measures muscle area and density, usually at the L3 or L4 vertebral level, and gives the skeletal muscle index (muscle area divided by height squared). A density window of −29 to 150 HU has become the standard. The dose is higher, around 11 mSv for the abdomen and pelvis, but that does not matter when the measurement comes from a scan done for another reason, such as cancer staging.

MRI adds a radiation-free view of muscle quality: Dixon sequences quantify fat infiltration, and spectroscopy can follow muscle metabolism. It is expensive and slow and remains mainly a research tool. Ultrasound is cheap and reproducible but not yet standardised, and too much probe pressure can make muscles look smaller than they are.

One gap remains: EWGSOP cut-off values exist only for DXA and bioelectrical impedance, not for CT or MRI.

Where it changes care

In oncology, a low CT-based muscle index is found in roughly 35–50% of patients across tumour types and is linked to shorter survival; in pancreatic cancer, patients with low muscle values also had more side effects from chemotherapy. Before pancreatic surgery, low muscle mass identified patients at higher risk of postoperative fistulas. Muscle loss on CT was associated with mortality in necrotising pancreatitis, and pectoralis muscle area predicted hospital stay, intubation and death in COVID-19.

Artificial intelligence is making these measurements routine. Automated segmentation of CT scans has matched manual measurement, and a deep-learning analysis of 9,223 CT datasets predicted hip fractures and mortality as well as established clinical risk scores.

In numbers

• ~20% and ~50%: share of 70- and 75-year-olds with sarcopenia
• <6 and <7 kg/m²: EWGSOP DXA thresholds for low muscle mass in women and men
• 1–10 µSv vs ~11 mSv: radiation dose of whole-body DXA vs abdominal and pelvic CT
• 38 studies, 7,843 patients: meta-analysis linking a low muscle index at diagnosis to worse survival in solid tumours

For practice

Sarcopenia is usually reversible. Progressive, individualised resistance training is the core treatment, balance training lowers the risk of falls, and a protein regimen rich in leucine, for example whey protein, supports muscle protein synthesis. Reporting muscle measurements from routine CT, increasingly with AI support, lets clinicians start these measures earlier.

Source: Vogele D, Otto S, Sollmann N, Haggenmüller B, Wolf D, Beer M, Schmidt SA. Sarcopenia – Definition, Radiological Diagnosis, Clinical Significance. RöFo. 2023;195(5):393–405. doi:10.1055/a-1990-0201