As healthcare moves toward more personalized and preventive care, quantitative imaging is emerging as a powerful tool for uncovering health insights beyond conventional clinical measurements. AIRS Medical is advancing this approach with SwiftSight Body Composition, an AI-powered MRI solution designed to quantify fat distribution, muscle quality, muscle asymmetry, and liver fat in an MRI scan lasting less than five minutes, followed by automated analysis.

In this interview with MedTech Spectrum, Jina Park, Chief Strategy Officer at AIRS Medical, discusses the clinical need behind SwiftSight Body Composition, the role of AI in automating tissue segmentation and quantitative assessment, and the challenges of integrating the technology into existing MRI workflows. She also explores the importance of clinical validation, regulatory compliance, and physician oversight, alongside the solution’s potential applications in preventive health, metabolic health, weight management, and healthy aging. With its first customer deployment announced at SimonMed Imaging, Park shares AIRS Medical’s vision for expanding AI-powered quantitative MRI and enabling more consistent, longitudinal insights from existing imaging infrastructure.

What clinical need did AIRS Medical aim to address with SwiftSight Body Composition, and how does AI-powered MRI analysis improve the assessment of muscle, adipose tissue, and liver fat compared with conventional approaches?

Body weight and body mass index (BMI) are useful indicators, but they do not reveal how fat and muscle are distributed or provide a meaningful assessment of muscle quality. Two individuals of the same age and BMI can have significantly different levels of visceral fat, muscle mass, intramuscular fat infiltration and liver fat, with important implications for metabolic and musculoskeletal health.

SwiftSight Body Composition was developed to make these measurements accessible through a short MRI examination, combining AI-powered tissue analysis with quantitative reporting and age- and sex-matched benchmarking.

Conventional approaches each have strengths and limitations:

• Dual-energy X-ray absorptiometry (DXA) is a fast, established method for assessing bone density and body composition. Some systems estimate visceral fat, but DXA provides less detailed characterization of individual muscle groups, intramuscular fat infiltration and liver fat. It also involves a very low dose of ionizing radiation.
• CT provides detailed anatomical information and can quantify fat and muscle, but radiation exposure is a consideration when repeated examinations are required.
• Conventional MRI offers excellent soft-tissue contrast without ionizing radiation. However, manual tissue segmentation and quantification can be time-consuming, operator-dependent and difficult to standardize across a large imaging organization.

SwiftSight uses AI-powered segmentation to analyze defined MRI regions of the abdomen and thighs, distinguishing visceral from subcutaneous fat and measuring muscle proportion, intramuscular fat infiltration (myosteatosis) and bilateral muscle asymmetry. It also quantifies liver fat using proton density fat fraction (PDFF).

Unlike weight or BMI, these measurements provide tissue-specific information about body composition. By automating analysis and bringing the results together in a structured report, SwiftSight gives clinicians a quantitative baseline that can support patient assessment and follow-up.

SwiftSight Body Composition quantifies visceral and subcutaneous fat, muscle quality, bilateral muscle balance, and liver fat. How did AIRS Medical develop and validate the AI segmentation and quantification algorithms to ensure accuracy and consistency across these measurements?

SwiftSight Body Composition was developed using MRI datasets incorporating different scanner vendors, field strengths and acquisition conditions. Expert-annotated images provided reference information for identifying relevant fat and muscle compartments.

The segmentation models analyze visceral and subcutaneous adipose tissue, abdominal wall, psoas and paraspinal muscles, and major thigh muscle groups including the quadriceps, hamstrings and adductors.

The validation program addressed three principal areas:

• Segmentation and measurement accuracy: Comparing automated segmentation with expert-defined reference annotations to evaluate anatomical agreement and the resulting quantitative measurements.
• Measurement consistency: Assessing repeatability and performance under supported imaging conditions, including differences in acquisition and scanner configuration.
• Reference-based interpretation: Comparing quantitative measurements against age- and sex-matched reference populations to provide context for individual patient results.

For liver fat assessment, SwiftSight uses MRI-derived PDFF, an established quantitative imaging biomarker. The software analyzes three regions of interest within the liver, calculates an average fat fraction and presents the result alongside a visual fat-distribution map.

SwiftSight Body Composition received FDA 510(k) clearance on October 1, 2026, under K261612. The FDA determined that the device was substantially equivalent to a legally marketed predicate device for its intended use.

As with any quantitative imaging technology, the specific performance characteristics—including segmentation accuracy, measurement repeatability and reproducibility across acquisition conditions—must be interpreted within the scope of the product's validation data and cleared labeling.

The solution features an MRI scan time of under five minutes followed by automated analysis. What were the key technical challenges in achieving this, and how does the solution integrate with existing MRI infrastructure and imaging-center workflows?

A central development objective was to make quantitative body composition practical for high-volume MRI environments, without requiring lengthy examinations or extensive manual post-processing.

SwiftSight Body Composition is designed around an MRI acquisition of less than five minutes, followed by automated AI analysis. The acquisition time refers to the required imaging sequences; overall examination and reporting times depend on the site's workflow.

The primary technical challenges included:

• Acquisition efficiency: Obtaining the necessary abdominal, thigh and liver images within a short imaging protocol while preserving the information required for quantitative analysis.
• Scanner interoperability: Supporting image analysis across different MRI vendors, field strengths and supported acquisition protocols.
• Workflow automation: Automating image routing, segmentation, quantification, reference-population comparison and report generation to reduce manual measurement and processing.

The system analyzes selected anatomical regions, including images near the third lumbar vertebra (L3) and at the mid-thigh, together with liver imaging. These regional measurements characterize the tissues examined and can provide estimates relevant to broader body composition assessment; they are not direct measurements of total whole-body tissue volumes.

Integration is designed around existing MRI infrastructure. The required images are acquired on the organization's scanners, transferred through standard DICOM workflows and processed by SwiftSight. The resulting Body Composition Report combines quantitative measurements, reference information and annotated images.

Deployment requires appropriate imaging protocols, system configuration and DICOM connectivity, but does not require purchasing a separate imaging modality. SwiftMR can also support accelerated image acquisition where appropriate.

For multi-site imaging organizations, this approach enables a consistent analysis and reporting process across supported scanners and locations, while reducing reliance on time-consuming manual measurements.

The Body Composition Report provides quantitative measurements alongside age- and sex-matched percentile rankings. How can clinicians use this information in areas such as preventive health, weight management, metabolic health, and longevity?

The clinical value of body composition analysis comes from understanding not only how much fat or muscle a patient has, but also where those tissues are distributed, how their composition compares with relevant reference populations and whether measurements change over time.

SwiftSight brings quantitative measurements, percentile rankings, reference ranges and annotated MRI images into one report. This gives clinicians a structured way to interpret results and discuss them with patients.

Age- and sex-matched percentiles show where an individual's measurements fall relative to a reference population. The report also presents expected and optimal ranges to help contextualize the findings. These comparisons are interpretive aids rather than standalone diagnostic thresholds, and results must be considered in the context of the patient's overall clinical assessment.

Potential applications include:

• Preventive health and whole-body MRI: Establish quantitative baseline measurements of fat distribution, muscle composition and liver fat as part of preventive imaging services.
• Weight management: Assess changes in visceral and subcutaneous fat while monitoring regional muscle composition during lifestyle interventions, medical weight management or GLP-1-based treatment.
• Metabolic health: Provide additional information about visceral adiposity, liver fat and muscle quality that may not be apparent from weight or BMI alone.
• Healthy aging and longevity: Follow changes in muscle proportion and fat infiltration that may be relevant to age-related changes in body composition.
• Rehabilitation and performance: Quantify regional muscle characteristics and left-right asymmetry to support clinical assessment and follow-up.

Longitudinal assessment is particularly valuable because a single measurement provides only a snapshot. Repeated quantitative assessments allow clinicians to examine changes in tissue composition over time, while accounting for measurement variability and differences in imaging conditions.

Because MRI does not use ionizing radiation, repeat imaging can be considered when clinically appropriate without the radiation exposure associated with CT.

SwiftSight provides quantitative information to support clinical assessment. It does not independently diagnose disease, determine treatment, or establish whether an intervention has been clinically successful.

As AI becomes more integrated into medical imaging, how does AIRS Medical approach clinical validation, regulatory compliance, and quality assurance? What safeguards ensure clinicians remain appropriately involved in interpreting results?

AIRS Medical operates under a medical device quality management system and has achieved certification through the Medical Device Single Audit Program (MDSAP).

SwiftSight Body Composition received FDA 510(k) clearance in the United States under K261612. The clearance establishes substantial equivalence to a legally marketed device for the specified intended use.

Product development follows regulated software development, verification, validation and change-control processes. AIRS Medical also maintains post-market processes for reviewing reported issues, monitoring product performance and implementing corrective actions when required.

These quality and regulatory controls are distinct from the clinical interpretation of an individual patient's results.

Clinicians remain responsible for evaluating the measurements and determining their significance.

• Quantitative decision support: SwiftSight provides tissue measurements, visualizations and reference-based comparisons. It does not independently diagnose medical conditions or prescribe treatment.
• Visual verification: Reports include annotated images of the analyzed anatomy so healthcare professionals can review the tissue regions used to generate the measurements.
• Clinical oversight: Results are interpreted alongside medical history, laboratory findings, other imaging results and physical examination, according to the provider's established clinical workflow.

AIRS Medical has also established a Clinical Advisory Board, with founding members Dr. Ann K. Jay and Dr. Gregory J. Moore, to provide external clinical and strategic perspectives on the development of quantitative MRI, longitudinal imaging and AI-enabled clinical workflows.

The objective is to combine automated quantitative analysis with appropriate medical oversight, ensuring that clinicians remain central to how the information is interpreted and used.

With SimonMed Imaging set to deploy SwiftSight Body Composition as your first customer, what are AIRS Medical’s plans for expanding across health systems and imaging networks? Do you see quantitative AI-driven MRI supporting a broader range of biomarkers and longitudinal health assessments in the future?

SimonMed Imaging has been announced as the first customer to deploy SwiftSight Body Composition. The introduction builds on AIRS Medical's established relationship with SimonMed and reflects the opportunity to extend quantitative MRI capabilities within large imaging organizations.

For health systems and imaging networks, SwiftSight Body Composition offers a way to expand existing MRI services without investing in an additional imaging modality.

Organizations operating whole-body MRI, preventive imaging, executive health, healthy-aging or weight-management programs can incorporate quantitative body composition assessment into their service offerings.

The commercial opportunity lies in extending the value of existing MRI infrastructure. Rather than adding another scanner or relying on a separate manual analysis process, providers can introduce a quantitative service built around short acquisition times, automated tissue analysis and structured reporting.

For multi-site organizations, the same approach can be implemented across supported scanners and locations, helping establish a more consistent service model while allowing deployment requirements to be managed within the organization's existing imaging and IT environment.

SwiftSight Body Composition is also part of a broader platform strategy.

AIRS Medical is extending SwiftSight beyond brain volumetry, with diffusion tensor imaging (DTI) planned for November 2026 and quantitative joint applications planned for 2027. These developments are intended to expand the range of anatomical measurements and clinical information available through MRI.

Our longer-term vision is to move MRI beyond the interpretation of individual images toward a more quantitative and longitudinal approach to patient assessment.

This requires more than automated segmentation. It requires measurement accuracy, reproducibility, meaningful reference information and clinical interpretation, supported by technology that can operate across increasingly complex MRI environments.

AIRS Medical's objective is to help radiologists and imaging organizations obtain more performance, consistency and intelligence from the MRI infrastructure they already own—while building the foundation for the next generation of AI-powered quantitative MRI.