Healthspan Imaging & Monitoring

Scope and source discipline

Operational definition used in this note: “Healthspan imaging and monitoring” means imaging-derived quantitative biomarkers—body composition, visceral adipose tissue, hepatic fat, coronary artery calcium, coronary flow physiology, brain/liver/muscle structure—plus longitudinal physiological monitoring via wearables, continuous glucose monitors, insertable cardiac monitors, and connected home/clinical monitoring devices.

This note applies the requested source hierarchy:

• Tier 1: granted patents, peer-reviewed papers with verified DOI, standards-body documents, and regulator documents.
• Tier 2: company technical docs, SEC-filed disclosures, and named company/self-reported material.
• Prohibited: SEO content farms, anonymous benchmark sites, AI-generated summaries, and any source that merely quotes a patent or paper that cannot be retrieved.

All URLs, document identifiers, publication/grant dates, and access dates are listed in the final Sources section. Access date for all sources is 2026-04-26 unless otherwise stated.

A core evidentiary rule followed here: a patent is evidence of recorded claims, not proof of performance or commercial viability. Where a technical claim is load-bearing and cannot be independently cross-verified from a different institution, it is tagged [SINGLE-SOURCE].

Tier-1 technical evidence

2.1 Peer-reviewed imaging biomarkers: CT body composition

The strongest peer-reviewed evidence for automated healthspan imaging comes from a Mayo Clinic group:

• Weston et al., “Automated Abdominal Segmentation of CT Scans for Body Composition Analysis Using Deep Learning,” Radiology, 2019, DOI 10.1148/radiol.2018181432

URL: https://doi.org/10.1148/radiol.2018181432

This study is a Tier-1 peer-reviewed paper. It reports that deep learning can automatically segment skeletal muscle, subcutaneous adipose tissue, and visceral adipose tissue from routine abdominal CT data. The reported segmentation performance was high, with Dice similarity coefficients above 0.9 for the tested abdominal compartments. Exact compartment-level values should be read from the original paper.

Source-bound claim: automated CT body-composition segmentation is technically feasible and was internally validated with high agreement against manual segmentation. This is a measured-performance claim from a single institution; the specific CT metric is [SINGLE-SOURCE].

2.2 Independent corroboration from MRI body-composition profiling

A second Tier-1 peer-reviewed paper from a different research lineage supports the broader concept that imaging can move health assessment beyond BMI:

• Borga et al., “Advanced body composition assessment: from body mass index to body composition profiling,” Journal of Investigative Medicine, 2018, DOI 10.1136/jim-2018-000722

URL: https://doi.org/10.1136/jim-2018-000722

That paper reviews MRI-based body-composition quantification and makes the argument that body-composition profiling can be more clinically informative than BMI alone. It provides independent, different-institution support for the concept that quantitative imaging biomarkers are feasible and clinically relevant. This cross-validates the idea behind the CT work, though it is a different modality and not a direct replication of the same segmentation model.

Load-bearing claim: automated or quantitative imaging body-composition assessment is technically feasible.

Cross-check: two independent sources from different institutions and different modalities: Weston et al. 2019 for CT and Borga et al. 2018 for MRI. Therefore this claim is cross-verified, while the exact CT Dice metric remains [SINGLE-SOURCE].

2.3 Peer-reviewed evidence on monitoring: wearables require more than sensor accuracy

For the monitoring side, the technical risk is not only sensor performance but clinical integration:

• Smuck et al., “The emerging clinical role of wearables: factors for successful implementation in healthcare,” npj Digital Medicine, 2021, DOI 10.1038/s41746-021-00418-3

URL: https://doi.org/10.1038/s41746-021-00418-3

This Tier-1 peer-reviewed paper reports that successful wearable-health-monitoring deployment depends on workflow integration, data quality, clinician trust, and clinical evidence—not merely on sensor accuracy. The source does not itself prove that a specific wearable improves healthspan; it supports a narrower but important technical claim: implementation is a distinct feasibility barrier from sensor measurement.

Granted patent evidence

The required granted patent is:

• US Patent 8,315,812 B2 — “Method and system for patient-specific modeling of blood flow”
• Assignee: HeartFlow, Inc.
• Grant date: 2012-11-27
• Document identifier: US8315812B2
• URL: https://patents.google.com/patent/US8315812B2/en

Why this patent matters for healthspan imaging: It covers patient-specific modeling of blood flow using CT image data, the technical foundation for CT-derived fractional flow reserve, or FFR-CT. Coronary flow physiology is a cardiovascular aging/healthspan biomarker beyond gross anatomy. This is a granted patent, not an application, and therefore it can support moat/IP analysis, but it is not proof of clinical performance or commercial viability.

Important evidentiary distinction: The patent is recorded IP. It supports the proposition that non-invasive CT-derived cardiovascular physiology is protectable and technically specified. It does not itself demonstrate that any product achieves a given sensitivity, specificity, or outcome benefit. No Tier-1 performance claim should be inferred from this patent.

Standards-body and regulator sources

4.1 ISO 13485:2016

• ISO 13485:2016 — Medical devices — Quality management systems — Requirements for regulatory purposes
• Publication date: 2016-03
• URL: https://www.iso.org/standard/59752.html

This is a Tier-1 standards-body document. It governs quality management for medical-device developers, including imaging-software and monitoring-device manufacturers. Its relevance to healthspan imaging/monitoring is regulatory and operational, not clinical evidence. It supports the claim that this sector is governed by formal design-control and risk-management requirements.

4.2 FDA AI/ML-enabled medical-device list

• FDA, “Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices”
• URL: https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-aiml-enabled-medical-devices
• Accessed: 2026-04-26

This is a Tier-1 regulator source. It maintains a public list of AI/ML-enabled devices authorized in the United States. It supports the claim that FDA authorization is a real market-access pathway for imaging-software and monitoring algorithms. The page itself is a registry of regulatory status; it is not an independent technical-performance benchmark.

4.3 FDA final guidance on predetermined change control plans

• FDA, “Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions”
• Publication date: 2024-12-04
• URL: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/marketing-submission-recommendations-predetermined-change-control-plan-artificial-intelligence-enabled
• Accessed: 2026-04-26

This Tier-1 regulator guidance is directly relevant to imaging-AI and monitoring-AI business models because it addresses how AI/ML-enabled devices can be updated after initial authorization. It affects the regulatory moat around adaptive imaging algorithms.

Adoption/deployment context from company filings and company documentation

5.1 Public imaging OEMs

GE HealthCare Technologies Inc. (GEHC) files 10-K/10-Q reports with the SEC. Its SEC-filed disclosures describe a business that sells imaging equipment, software, and services into highly regulated healthcare markets. The risk factors in GEHC’s 10-K filings include regulatory-authorization risk, clinical-adoption risk, and reimbursement risk. The filings do not claim that full-body MRI screening or AI body-composition analysis reduces mortality. This is an important primary-source control against marketing claims.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001932393&type=10-K&dateb=&owner=include&count=40
• CIK: 0001932393

Hologic, Inc. (HOLX) sells DEXA systems used for bone-density and body-composition analysis. Its SEC filings describe hospital, imaging-center, and clinician customers. DEXA is one of the most deployed body-composition imaging methods, but it is lower spatial resolution than CT/MRI.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000859737&type=10-K&dateb=&owner=include&count=40
• CIK: 0000859737

Koninklijke Philips N.V. (PHG) operates in imaging and patient monitoring. Its disclosed product lines include MRI/CT and hospital patient monitoring, which is relevant across the imaging and longitudinal-monitoring segments.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000313216&type=10-K&dateb=&owner=include&count=40
• CIK: 0000313216

5.2 Monitoring companies

DexCom, Inc. (DXCM) sells continuous glucose monitoring systems. In SEC filings, DexCom discloses that its revenue depends on patient adoption, payer coverage, and sensor accuracy. Continuous glucose monitoring is an important real-world metabolic-monitoring stream because glucose variability and hyperglycemia are modifiable healthspan-related signals.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001093557&type=10-K&dateb=&owner=include&count=40
• CIK: 0001093557

Abbott Laboratories (ABT) sells the FreeStyle Libre CGM platform. Its SEC filings disclose diagnostics and medical-device revenue, including diabetes care. This matters because CGM is not only a diabetes-management tool but also a metabolic-health monitoring input.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000001800&type=10-K&dateb=&owner=include&count=40
• CIK: 0000001800

ResMed Inc. (RMD) sells connected CPAP and sleep/respiratory-monitoring equipment. Sleep-disordered breathing is a major modifiable healthspan factor, and ResMed’s cloud-connected devices generate longitudinal monitoring data.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000943819&type=10-K&dateb=&owner=include&count=40
• CIK: 0000943819

Medtronic plc (MDT) sells insertable cardiac monitors and diabetes/cardiac devices. Insertable cardiac monitors generate long-duration heart-rhythm data and are a distinct monitoring form factor from wearables.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001613103&type=10-K&dateb=&owner=include&count=40
• CIK: 0001613103

Boston Scientific Corp. (BSX) also sells insertable cardiac monitors. This widens the cardiac-monitoring exposure beyond Medtronic and supports the claim that cardiac monitoring is a multi-vendor public-market category.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000885725&type=10-K&dateb=&owner=include&count=40
• CIK: 0000885725

5.3 Consumer platforms and enabling hardware

Apple Inc. (AAPL) sells Apple Watch with heart-rate, rhythm, and activity monitoring. Apple’s SEC filings describe a consumer wearable ecosystem; the company files regulatory submissions for specific features such as ECG and irregular-rhythm notifications. Apple’s disclosures are therefore primary-source evidence of consumer health-monitoring deployment, not independent proof of healthspan outcomes.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000320193&type=10-K&dateb=&owner=include&count=40
• CIK: 0000320193

Alphabet Inc. (GOOGL) is the parent of Verily/health-data initiatives. Its exposure is to health data platforms, research partnerships, and AI infrastructure rather than a single imaging device. SEC filings provide primary-source disclosure of Alphabet’s broader strategy and risk factors.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001652044&type=10-K&dateb=&owner=include&count=40
• CIK: 0001652044

NVIDIA Corp. (NVDA) supplies GPUs used in training and inference for deep-learning imaging models. It is an enabling-hardware supplier, not a clinical-imaging company. Its SEC filings disclose data-center/accelerator revenue and risk factors.

• SEC EDGAR page: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001045810&type=10-K&dateb=&owner=include&count=40
• CIK: 0001045810

5.4 Direct-to-consumer full-body MRI operators

Prenuvo markets proactive full-body MRI scans directly to consumers and employers. This is company self-reported marketing material, so it is Tier 2, not independent clinical validation.

• URL: https://www.prenuvo.com/
• Pricing page: https://www.prenuvo.com/pricing

Ezra also markets cash-pay full-body MRI to consumers. This is Tier 2 self-reported material.

• URL: https://ezra.com/

These companies monetize healthspan imaging via cash-pay scans or employer health programs. They are private and do not file 10-K reports; therefore their marketing claims cannot be cross-referenced against SEC risk-factor filings in the same way as GE HealthCare, Hologic, or Dexcom.

Contradictions and source-bound claims

Contradiction 1: DTC marketing versus filed disclosure language

DTC full-body MRI operators such as Prenuvo market scans as proactive early-detection tools for often hundreds of conditions. Those claims are self-reported and appear on company websites (https://www.prenuvo.com/, accessed 2026-04-26). By contrast, a public imaging OEM, GE HealthCare, does not claim in its SEC-filed 10-K risk factors that full-body screening MRI or AI-derived body-composition measurements reduce mortality. GEHC instead discloses regulatory, clinical-adoption, and reimbursement risks via its SEC EDGAR 10-K filings (https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001932393&type=10-K&dateb=&owner=include&count=40, accessed 2026-04-26).

This is a discrepancy between marketing claims and filed disclosure caution. It does not prove the DTC companies are wrong, but it does mean the strongest publicly verifiable language from a large public imaging OEM is materially more conservative than direct-to-consumer marketing.

Contradiction 2: Patent claims are not performance claims

HeartFlow’s US 8,315,812 B2 is a granted patent for patient-specific blood-flow modeling. It is often described as foundational to CT-derived FFR. That is true as an IP claim. But the patent is not a clinical-outcome study and does not establish sensitivity, specificity, or mortality benefit. This note does not treat the patent as proof that FFR-CT improves healthspan.

Contradiction 3: Wearable marketing versus implementation evidence

Consumer wearables are marketed as health-improvement tools, but the peer-reviewed literature emphasizes implementation barriers: data governance, clinician workflow, and outcome evidence. The Smuck et al. 2021 paper (DOI 10.1038/s41746-021-00418-3) supports the narrower implementation-risk claim. It is not a validation that any specific wearable improves healthspan.

Quantitative worked calculation

Illustrative capacity model for a cash-pay full-body MRI scanner

This calculation is intentionally simple and transparent. It is a capacity calculation, not a market forecast or company-specific income statement.

Formula:

\[

R = P \times V \times D

\]

Where:

• \(R\) = annual realized revenue per scanner
• \(P\) = net realized price per scan
• \(V\) = scans per day
• \(D\) = operating days per year

Inputs used:

Calculation:

\[

R = 2,500 \times 4 \times 240 = \$2,400,000

\]

Sensitivity:

• If \(P = \$2,000\), then \(R = \$1,920,000\)
• If \(P = \$3,500\), then \(R = \$3,360,000\)
• If \(V = 3\), then \(R = \$1,800,000\)
• If \(V = 5\), then \(R = \$3,000,000\)

Interpretation: The model shows that cash-pay scanner economics are highly sensitive to price and daily throughput. It does not calculate profit. Gross margin would require scanner depreciation, rent, staffing, radiologist reads, marketing, insurance, and other costs. Those figures are not reliably available from Tier-1 public sources at the individual-scanner level; exact profit margins must be marked “needs primary source validation”.

Parameter table

Unit economics / business model

Who monetizes this technology

There are four distinct monetization paths:

1. Imaging OEMs sell capital equipment, service contracts, and software upgrades. GE HealthCare, Philips, and Hologic monetize via hardware plus multi-year service and software arrangements. SEC filings evidence these business lines generally (https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001932393&type=10-K&dateb=&owner=include&count=40; https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000859737&type=10-K&dateb=&owner=include&count=40, accessed 2026-04-26). Exact segment-level margins require reading the specific 10-K financial statements and are not reproduced here.

1. AI/software vendors monetize through regulatory-cleared or cleared-listed algorithms sold as enterprise licenses, per-exam fees, or cloud subscriptions. FDA’s AI/ML device page is the regulatory evidence that this is a real commercial category (https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-aiml-enabled-medical-devices, accessed 2026-04-26). Pricing terms are not uniformly disclosed and should be marked “needs primary source validation” unless obtained from a specific contract or product sheet.

1. DTC full-body MRI operators monetize cash-pay scans and employer programs. This is evident from Prenuvo’s and Ezra’s consumer-facing websites (https://www.prenuvo.com/; https://ezra.com/, accessed 2026-04-26). Their pricing is self-reported and can change by market, bundle, or promotion.

1. Monitoring-device makers monetize via durable hardware plus recurring disposable sensors or subscription data services. DexCom, Abbott, ResMed, and Medtronic all provide SEC-filed disclosures that support recurring-device/consumable models (https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001093557&type=10-K&dateb=&owner=include&count=40; https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000001800&type=10-K&dateb=&owner=include&count=40, accessed 2026-04-26).

Who captures the value

• Imaging OEMs capture value from hospital capital budgets and service contracts.
• AI companies and OEM software units capture value from per-exam or enterprise software fees if their algorithms are reimbursed or adopted into workflow.
• DTC operators capture consumer/employer wellness spending.
• Monitoring companies capture recurring sensor and service revenue.
• Patients/self-insured employers currently bear cost in cash-pay DTC models, while insurers and governments capture downstream value only if the imaging or monitoring leads to avoided or more efficient care.

No public source in this note provides a reliable, granular gross-margin figure for a full-body MRI operator or an AI body-composition software vendor. Any such number would need to be labeled “needs primary source validation.”

Disconfirming Evidence

The healthspan-imaging-and-monitoring thesis would be meaningfully weakened or falsified if any of the following occur:

• Prospective, randomized, or matched-cohort evidence shows no improvement in all-cause mortality, healthy lifespan, or major healthspan endpoints from imaging-based screening programs.
• FDA or CMS reimbursement decisions reject population-level full-body MRI or AI-based body-composition screening, collapsing the cash-pay/employer economic model.
• External validation of CT/MRI body-composition AI fails across multiple scanner vendors, institutions, or demographic groups, undermining generalizability.
• Continuous-monitoring data show strong sensor performance but no change in clinical outcomes or behavior, supporting the implementation-risk critique from Smuck et al. 2021 over the commercial adoption narrative.
• Payers determine that DTC full-body MRI increases downstream unnecessary procedures without offsetting benefit, triggering restrictive coverage or regulatory action.