Shining a Light on the Hidden Shadows of Sarcopenia

AI tools can help radiologists improve patient care by identifying conditions often overlooked in clinical practice

BY NICK KLENSKE September 11, 2025

Tagliafico

Boutin

Sarcopenia, the age-related loss of muscle mass and strength, is a condition that becomes more common with age, particularly after 60. Likewise, sarcopenic obesity is a condition characterized by the co-occurrence of sarcopenia and obesity.

It is estimated between 6% and 22% of older adults suffer from sarcopenia, which can lead to decreased physical function, increased risk of falls and other health problems. Sarcopenic obesity is also associated with increased mortality risk.

The two conditions are typically diagnosed using a multi-step process involving screening (BMI, waist circumference, etc.) and staging. Imaging also plays a role in diagnosing sarcopenia and sarcopenic obesity, with CT and MRI considered the gold standards for imaging muscle quantity and quality.

Unfortunately, a lack of standardized definitions, the gradual onset of conditions, and the high costs and limited availability of diagnostic tools have led to the underdiagnosis of sarcopenia.

But what if imaging could do more?

“As sarcopenia is something we regularly see on inpatient scans, radiology could play a major role in early detection and awareness,” said Alberto Tagliafico, MD, a radiologist at the University of Genova in Italy, who co-wrote a 2022 study on the use of imaging in sarcopenia.

Despite its prevalence on scans, sarcopenia tends to get ignored amid all the other diagnoses. “As radiologists, we look at muscle and fat on most imaging examinations, but we haven’t had an efficient way to measure if these tissues are normal or not,” added Robert Boutin, MD, clinical professor, radiology at Stanford University School of Medicine in California.

Dr. Boutin recently co-authored a Radiology study that looked at whether sarcopenia and sarcopenic obesity are underdiagnosed clinically in electric health records (EHRs) and CT reports. “We aimed to shine a light on the hidden shadows of sarcopenia and sarcopenic obesity within clinical records and CT reports, revealing what often goes unnoticed,” he said.

“As muscle and fat can now be measured automatically, it’s important to know if AI measurements might support radiologists and clinicians in identifying patients with undiagnosed sarcopenia and sarcopenic obesity.”

— ROBERT BOUTIN, MD

Identifying Patients with Undiagnosed Sarcopenia

To find out how often sarcopenia and sarcopenic obesity are diagnosed, the study reviewed EHRs and abdominal CT reports for 17,646 patients at an academic hospital. Researchers then audited those cases using an AI tool that can automatically measure muscle and fat on routine CT scans.

Researchers also compared the diagnostic frequency of sarcopenia and sarcopenic obesity in their hospital’s patients with the diagnostic frequency shown at 274 other institutions.

What they found was that sarcopenia was rarely diagnosed in the hospital’s EHR (0.05%) or across 274 other institutions (0.1%). However, researchers discovered CT findings of sarcopenia in 28.5% of patients, similar to the prevalence estimates in the sarcopenia literature.

In addition, coexisting sarcopenia and obesity (sarcopenic obesity) was not documented at all among the 17,646 patients yet was detected in by CT imaging in 5.7% of the patients.

“By comparing clinical documentation with objective imaging measurements for a large number of patients, we were confronted with a massive discrepancy that has potential implications for both individual patients and population health,” Dr. Boutin noted.

A Catalyst for Change in Radiology

Fortunately, the study also showed that there is a solution capable of addressing this discrepancy: automated measurement using an AI tool on routinely acquired CT scans.

“Given that AI tools are now available to quantify body composition features, such as low muscle mass and excess adiposity, this study acts as a catalyst for change in radiology,” Dr. Boutin said. “It can transform how we screen and follow patients for sarcopenia using CT scans that are already being acquired for routine indications, without additional radiation exposure or diagnostic testing.”

According to Dr. Boutin, this improved diagnosis of sarcopenia could be the key that unlocks better treatment pathways and risk prediction, allowing health care providers to tailor recommendations to their patients. For example, considering recent sarcopenia and obesity management guidelines with specific prescriptions of exercise and nutrition. There is now a heightened need to explore the real-time implementation of AI tools as a value-added feature to routinely acquired CT scans.

Addressing the Silent Epidemic of Sarcopenia

With sarcopenia and sarcopenic obesity often overlooked in clinical practice, Dr. Boutin believes that the sensible integration of AI measurement tools will help radiologists and other providers efficiently enhance patient management and clinical outcomes.

“We hope our research serves as a wake-up call, urging the entire medical community to help address the silent epidemic of sarcopenia and sarcopenic obesity,” he concluded.