Advancing Pediatric Healthcare: The Introduction of AI-Enhanced Muscle Mass Growth Charts

A team of scientists has pioneered the use of artificial intelligence to develop growth charts specifically for monitoring children’s muscle mass, utilizing the most extensive pediatric MRI dataset available. This innovation allows for more precise health evaluations and the possibility of preemptive actions against muscle deterioration.

A Brigham-led team utilized AI in examining MRI scans, culminating in a novel growth standard and a rapid, reproducible method for gauging lean muscle mass indicators in children.

Researchers, using a combination of artificial intelligence and the most substantial collection of pediatric brain MRI data amassed thus far, have crafted a novel growth chart dedicated to the observation of muscle mass in children. A new study orchestrated by the team at Brigham and Women’s Hospital—affiliated with the Mass General Brigham healthcare system—unveils an AI-based tool, marking a first in offering a standardized, precise, and consistent method to evaluate and monitor muscle mass indications via routine MRI scans. The findings were published in the “Nature Communications” journal on November 9.

Understanding Muscle Mass Monitoring

Dr. Ben Kann, MD, a leading figure in radiation oncology at the Brigham and part of the Mass General Brigham’s Artificial Intelligence in Medicine Program, noted the absence of a conventional metric for measuring muscle mass, particularly in pediatric cancer patients. This gap inspired the use of AI to measure the temporalis muscle thickness and establish a universal standard.

Kann’s team has devised a growth chart that allows for swift and immediate tracking of muscle thickness in children, enabling healthcare providers to ascertain if their growth aligns with ideal parameters.

The Significance of Lean Muscle Mass

Human lean muscle mass is closely associated with the quality of life, daily functionality, and general health and longevity. Low levels of lean muscle mass, such as those seen in sarcopenia, heighten the risk of premature death and susceptibility to a variety of life-quality impacting illnesses.

Despite the critical nature of lean muscle mass, traditional methods, including the Body Mass Index (BMI), fail to provide a detailed analysis of muscle composition. BMI, while reflective of weight, does not differentiate between fat and muscle.

Temporalis muscle thickness has been recognized for decades as indicative of body lean muscle mass, yet measuring this muscle’s thickness in a clinical setting has proven challenging, without a benchmark for distinguishing normal from abnormal levels. Conventional methods, often manual, are not only laborious but also lack standardization.

Groundbreaking Research and Discoveries

In their quest to address this challenge, the research collective applied an advanced learning model to MRI scans from patients with pediatric brain tumors at the Boston Children’s Hospital/Dana-Farber Cancer Institute, in association with the Boston Children’s Radiology Department. Analyzing 23,852 brain MRIs from healthy individuals aged 4 to 35, they computed temporalis muscle thickness and constructed normative growth charts. The analysis by AI demonstrated accuracy across a broad patient spectrum, rivaling the assessments of trained professionals.

Clinical Implications

Kann suggests these growth charts could be employed similarly to how height and weight are used, helping to determine if a child’s muscle mass falls within normal ranges. This method could prove beneficial for patients undergoing routine brain MRIs for conditions such as pediatric cancers and neurodegenerative diseases, allowing for prompt intervention in cases of observed muscle loss to combat the adverse effects of sarcopenia and diminished muscle mass.

The technique does present limitations, notably dependent on the quality of the MRI scans, which can impact the precision of measurements and their interpretation. Furthermore, the scarcity of MRI datasets from regions outside the United States and Europe restricts the global applicability of these findings.

Looking Ahead

Dr. Kann and his team contemplate extending the use of iTMT beyond current practices, potentially justifying more frequent MRI scans for a broader patient group. Their goal is to enhance the model by training it on a wider variety of cases. They envision future iTMT applications that might allow for the tracking and prediction of morbidity and the identification of critical physiological states necessitating medical intervention.

Citation: Zapaishchykova, A. et al. presents the study “Automated Temporalis Muscle Quantification and Growth Charts for Children Through Adulthood,” published on November 9, 2023, in Nature Communications.
DOI: 10.1038/s41467-023-42501-1

Authorship includes contributors from Brigham such as Anna Zapaishchykova, Kevin X. Liu, among others, with additional collaboration from Paul Catalano and Viviana Benitez, to name a few.

The study acknowledges funding from various NIH grants and the European Research Council, with KL being supported by the NIH Loan Repayment Program.

Frequently Asked Questions (FAQs) about AI pediatric growth charts

More about AI pediatric growth charts

• AI-Driven Pediatric Muscle Mass Charts
• Brigham and Women’s Hospital Research
• Mass General Brigham Healthcare System
• National Institutes of Health Funding
• European Research Council Grants
• Nature Communications Journal