Friday, March 28, 2025

Bartolomucci et al. review ctDNA for Therapeutic Monitoring; open access

 In a new article at npj Precision Oncology, Bartolomucci et al. comprehensively review the use of ctDNA for monitoring therapeutic response.

https://www.nature.com/articles/s41698-025-00876-y


The abstract is brief:

  • Circulating tumor DNA (ctDNA) has emerged as a dynamic biomarker in cancer, as evidenced by its increasing integration into clinical practice. Carrying tumor specific characteristics, ctDNA can be used to inform treatment selection, monitor response, and identify drug resistance. 
  • In this review, we provide a comprehensive, up-to-date summary of ctDNA in monitoring treatment response with a focus on lung, colorectal, and breast cancers, and discuss current challenges and future directions.
Note that while the title highlights "treatment response" the scope also includes treatment selection (e.g. LBx mutation CDx panels).

###

AI CORNER

###

Given the brief abstract, a longer summary by Chat GPT 4o.

###

Here’s a blog-ready article that serves as a detailed TL;DR of the new review paper by Bartolomucci et al., published in NPJ Precision Oncology in 2025:


TL;DR: Circulating Tumor DNA (ctDNA) to Monitor Treatment Response in Solid Tumors

A comprehensive new review by Bartolomucci et al. (NPJ Precision Oncology, 2025) offers one of the most in-depth, up-to-date examinations of how circulating tumor DNA (ctDNA) is transforming the way we monitor treatment response, resistance, and minimal residual disease (MRD) across solid tumors. The paper systematically explores ctDNA’s utility in lung, colorectal, and breast cancers, while highlighting key methodologies, clinical trials, and future directions.

๐Ÿ”ฌ What’s ctDNA and Why Does It Matter?

ctDNA refers to tumor-derived DNA fragments found in the blood, offering a non-invasive, real-time molecular snapshot of tumor dynamics. Unlike traditional imaging or tissue biopsies, ctDNA can:

  • Reflect systemic disease, including metastases.

  • Provide early insight into treatment efficacy or failure.

  • Detect residual disease and emerging resistance mutations long before clinical progression.

The authors argue that ctDNA is best positioned right now to serve as a treatment response tool, even more than for early detection or initial diagnostics.


๐Ÿงช Techniques and Tools

The review outlines two broad categories of ctDNA detection:

  • Targeted assays: PCR-based methods (qPCR, dPCR, BEAMing) to detect known mutations (e.g., EGFR, KRAS).

  • NGS-based assays: Broader approaches (e.g., CAPP-Seq, Safe-SeqS, Duplex Sequencing, CODEC) capable of detecting unknown or low-frequency variants.

Emerging areas include:

  • Methylation profiling

  • Fragmentomics (size, end motifs, nucleosome positioning)

  • Multi-analyte liquid biopsy (e.g., integrating CTCs, EVs, ctDNA)

  • Non-plasma biofluids: urine, saliva, CSF


๐Ÿซ Lung Cancer: Leading the Way

  • FDA has approved multiple ctDNA-based tests for NSCLC (e.g., Cobas EGFR, Guardant360, FoundationOne Liquid CDx).

  • ctDNA has shown strong correlation with tumor burden, MRD, and response to EGFR/ALK inhibitors and immune checkpoint therapies.

  • Studies like TRACERx, APPLE, and BR.36 highlight how ctDNA dynamics can guide treatment selection and timing, even predicting progression before RECIST imaging criteria show changes.


๐Ÿฆ  Colorectal Cancer: A Model for MRD Detection

  • ctDNA is highly sensitive in CRC, especially post-surgery, for detecting MRD and recurrence risk.

  • In stage II CRC, ctDNA positivity post-resection strongly predicts recurrence, guiding adjuvant therapy decisions.

  • In stage III CRC, ctDNA clearance correlates with survival, and persistent ctDNA flags treatment failure.

  • In metastatic CRC, ctDNA tracks drug resistance (e.g., RAS mutations after anti-EGFR therapy) and is being tested for guiding rechallenge strategies.

  • Trials like CIRCULATE-US, TRACC, and DYNAMIC are helping define ctDNA's role in personalized CRC management.


๐ŸŽ— Breast Cancer: Emerging Applications

  • ctDNA tracks treatment response in neoadjuvant and metastatic settings, especially in triple-negative and HR+/HER2- subtypes.

  • Studies show ctDNA can predict relapse even in patients with a pathologic complete response.

  • Trials like SOLAR-1 have helped secure ctDNA-driven FDA approvals (e.g., alpelisib for PIK3CA mutations).

  • ctDNA is also being used to track ESR1 mutations and guide hormone therapy adjustments (e.g., PADA-1 trial).


๐Ÿšง Challenges Ahead

Despite the promise, ctDNA still faces real-world barriers:

  • Low abundance in early-stage or low-shedding tumors.

  • Technical variation and lack of assay standardization.

  • Confounding from CHIP (clonal hematopoiesis).

  • Cost and turnaround time of NGS-based assays.


๐Ÿ“ˆ Final Takeaway

This review makes a compelling case that ctDNA is no longer just an experimental tool—it's an increasingly clinical-grade biomarker for monitoring cancer treatment, particularly in lung, colorectal, and breast cancers. With growing clinical trial evidence and technical innovations (e.g., CODEC sequencing, fragmentomics, and methylation-based detection), ctDNA may soon become a standard of care in precision oncology.

๐Ÿ“„ Read the full paper: doi.org/10.1038/s41698-025-00876-y


Let me know if you’d like a version with bullet points, or with a different tone for your audience.

###

The paper also notes that multimodal panels are starting to bridge the gap, combining mutation analysis, methylation, fragmentomics, and sometimes expression signals—blurring the line between bespoke and off-the-shelf testing.

###

Do they address concerns that MDR treatment monitoring doesn't perhaps really improve patient outcomes?   That it just makes everybody's PFS half as long? 

###

Great question—and yes, the paper does acknowledge that concern, albeit somewhat indirectly and diplomatically.

In discussing Minimal Residual Disease (MRD) detection using ctDNA, the authors point out that:

“Although most ctDNA studies to date have been observational in nature, the potential role of ctDNA monitoring to help guide clinical decision-making cannot be overlooked.”

This signals an important limitation: we still lack definitive evidence that ctDNA-guided MRD detection and early intervention actually improve overall survival (OS) or quality of life. The authors do not shy away from this and emphasize that randomized interventional trials are ongoing to address this gap.

They also highlight:

  • That ctDNA-positive patients often relapse, but it's unclear whether earlier detection and treatment change the ultimate outcome or merely start the clock earlier, shortening PFS without improving OS.

  • The CIRCULATE-US, PEGASUS, and DYNAMIC trials are specifically designed to test whether ctDNA-driven escalation or de-escalation of therapy leads to meaningful benefits.

So while the tone is optimistic about ctDNA’s future, the review is careful not to overstate current clinical utility and calls for stronger evidence from prospective trials to justify acting on MRD signals.  [On which theme see Normanno.]

23andMe: From $3.5B Valuation in 2021 to Bankruptcy in 2025

 Numerous articles are asking, how did 23andMe get from a $3.5B (SPAC) valuation in 2021, to bankruptcy today?

I noted at the time, this was their strategic business plan and valuation explanation, from their investor deck:


________

This isn't the only slide in the deck but it's the last one, then, an appendix.  So the "business strategy," one explicitly directed to people planning to invest money, appeared to be "drawing 3 circles." Which seemed comic to me then, and now.  Here.

Hanna Forster's Discussion of MRD Acceptance

Over at Linked In, consultant Hanna Forster has a nice discussion of progress & factors in the widening usage of MRD testing in oncology.   Find it here:

https://www.linkedin.com/posts/activity-7310933641553993729-gqy2/


You can pair her essay with Decibio consultant Amal Thommil, who reguarly updates on coverage for MRD testing at Linked In:

https://www.linkedin.com/posts/amalthommil_current-reimbursement-coverage-for-solid-activity-7300250398542413824-LBVG 


Forster on Market Cap Changes

Forster also has a new blog on year on year market cap changes for companies involved in NGS Dx testing.   Note the left column of her graphic is relative market cap growth.   Absolute market cap is the bar size.   Note that she includes the full market cap of companies, like Roche, for which NGS Dx is only a sliver of their market cap value.  

In the pure play NGS space, Natera is up 178% with a market cap of $20B.





Inside and Outside the Bubble

As balance, I would just point back to my "outside the bubble" blog a few weeks ago.   "Inside the bubble," everything is full steam ahead for MRD testing, new applications, MRD thought leaders, etc.  But "outside the bubble," progress in guidelines or consensus review articles suggests slower progress.

See an article "more data is needed" in lung cancer by Normanno, here.   For another new review, see Bartolomucci in npj precision oncology, a review of ctDNA for treatment response.  Here, here.

See a new subscription article in Genomeweb, "more MRD CU outcomes are needed," here.   See also a new subscription article at  Genomeweb on precision medicine investments and barriers at instutions.


  

Thursday, March 27, 2025

STAT: 8 Key People to Know at HHS

On March 18, 2025, STAT publishes a list of 8 key personnel at HHS, below the levels of Secretary or agency heads.

https://www.statnews.com/2025/03/18/rfk-jr-hhs-who-to-know-series-profiles-eight-key-players-federal-health-policy/


  1. Heather Flick – HHS Chief of Staff
    A Trump-era legal adviser and former GOP election integrity lawyer, Flick now serves as a key aide to Secretary Kennedy and attends high-level meetings like the MAHA Commission.

  2. Hannah Anderson – Deputy Chief of Staff (Policy)
    Former director at the America First Policy Institute and ex-Senate health policy adviser, she brings extensive experience in drug pricing and private insurance policy.

  3. Shana Weir – Principal Deputy Assistant Secretary
    Known for challenging 2020 election results on Trump’s behalf, Weir now oversees administration duties and has ties to several MAGA PACs.

  4. Stefanie Spear – Principal Deputy Chief of Staff & Senior Counselor
    A longtime Kennedy ally and former environmental blogger, she’s influential at HHS but controversial due to her progressive background and vaccine skepticism.

  5. Matthew Memoli – Acting NIH Director
    A respiratory virus researcher critical of vaccine mandates, Memoli has led NIH through major changes and is expected to be replaced soon by Trump ally Jay Bhattacharya.

  6. Drew Snyder – Director of Medicaid
    A former Mississippi Medicaid chief known for cost control and expanding maternal care, Snyder is now rolling back Biden-era social support rules at the federal level.

  7. Christopher Carroll – Principal Deputy Assistant Secretary at SAMHSA
    A longtime HHS finance and policy specialist, Carroll is a surprise acting SAMHSA leader facing criticism over recent staff cuts and lack of public health focus.

  8. Michael B. Stuart – HHS General Counsel (Pending Confirmation)
    A West Virginia state senator and former U.S. Attorney, Stuart is a vaccine policy critic and MAHA supporter awaiting Senate confirmation to lead HHS legal affairs


See also an AGENCY-IQ article on how mid level HHS is continually being populated with conservative executives from Trump Admin 01.

HHS: Downsizing Announced, At Least 10,000 Employees

On March 27, 2025, HHS released a notice that HHS would be substantially reorganized, with departments and divisions merged.   There will 10,000 positions directly cut, and HHS estimates that early retirement and similar initiatives will result in a 20,000 decrease in personnel - from 82,000 to 62,000.

https://www.hhs.gov/about/news/hhs-restructuring-doge.html

Regional offices drop from 10 to 5; operating divisions drop from 28 to 15.   

'Policy" will be reorganized.  HHS writes that '"H.R., I.T, procurement, external affairs, and 'policy' " will be in a new core division, "Administration for a Health America" or AHA.

Secretary RFK Jr said, “We aren't just reducing bureaucratic sprawl. We are realigning the organization with its core mission and our new priorities in reversing the chronic disease epidemic."

He added, "The department will do more - a lot more."

Fierce Healthcare reported there would be only 300 cuts at CMS (4%), not effecting Medicaid and Medicare benefits.   Fierce Healthcare also offered this RFK quote: “Some of these little fiefdoms, for example, are so insulated and territorial that they actually hoard our patient medical data and sell it for profit to each other,” he said, adding some public health divisions are “neglecting” public health and are only interested in helping the industries they are supposed to regulate.  "

###

The full press release is detailed and can be found here at HHS:

https://www.hhs.gov/about/news/hhs-restructuring-doge.html

Open access coverage at Fierce Healthcare:

https://www.fiercehealthcare.com/regulatory/rfk-jr-prepares-10000-job-cuts-across-hhs-new-wave-worker-reductions

Open access coverage at Fierce Biotech (focus includes FDA, NIH):

https://www.fiercebiotech.com/biotech/nearly-5000-fda-and-nih-staffers-laid-hhs-eliminates-entire-alphabet-soup-departments

Biopharma Dive here:

https://www.biopharmadive.com/news/hhs-layoffs-restructuring-kennedy-fda-cms-trump/743694/

Politico:   https://www.politico.com/news/2025/03/27/hhs-to-cut-thousands-of-workers-in-sweeping-reorganization-00253359

The Hill:  https://thehill.com/policy/healthcare/5217175-hhs-workforce-reduction-kennedy/

STAT offers a projection of 90% staff cuts at AHRQ, which houses USPSTF.   STAT also recently published an article, 8 key personnel at HHS.

https://www.statnews.com/2025/03/20/hhs-ahrq-agency-responsible-for-health-care-quality-research-threatened-with-mass-layoffs/

https://www.statnews.com/2025/03/18/rfk-jr-hhs-who-to-know-series-profiles-eight-key-players-federal-health-policy/

NYT here.   WSJ here.  Endpoints here.  


##


For history, see Politico's forecast of HHS cuts from March 13   https://www.politico.com/news/2025/03/13/hhs-reorganization-00230113



Wednesday, March 26, 2025

New White Paper: Healthcare Return on Investment (Duke-Margolis Center)

The Duke-Margolis Center at Duke, for health policy, has formed a new entity called the Capital Impact Council (CIC).   It's got an illustrious board, and the CIC has just released a 30-page white paper that is a roadmap for successful investing and ROI in healthcare.

Find the CIC here:

https://healthpolicy.duke.edu/CIC

Find the white paper here:

https://healthpolicy.duke.edu/sites/default/files/2025-03/Margolis%20CIC%20Framework.pdf

Read about it at Fierce Healthcare here:

https://www.fiercehealthcare.com/finance/duke-margolis-launches-council-aims-build-health-focused-frameworks-private-equity


In a Nutshell

To grossly simplify, you know that Investment A requires $10M input, and should return $20M in 3 years.  Same financials for Investment B.   But CIC methods show that Investment A will prevent 500 strokes, and Investment B will prevent 1000 strokes.   So "B" has better "healthcare per capital" statistics.

Context:  Negative articles about private equity in healthcare - here, here, here, here.  Look for special issue on private equity in healthcare forthcoming in May 2025.

###

AI CORNER

###

The Duke-Margolis Capital Impact Council (CIC) has released a pivotal white paper introducing a new framework—Health Value Return on Investment (HV-ROI)—designed to guide private equity and venture capital investments in health care toward measurable improvements in health outcomes, access, affordability, and community resilience, alongside financial return. The HV-ROI framework emerges at a time when public funding for health care is increasingly constrained, and the role of private capital is expanding. While private equity’s track record in health care has been uneven, the CIC, composed of seasoned investors and health policy leaders, proposes HV-ROI as a structured, evidence-informed approach to ensure that capital drives value-based care transformation rather than solely profit.

The framework is built around two core components. First, the Health Value Model and Evidence component requires a baseline analysis of an investment’s potential to deliver health impact, grounded in clinical literature, real-world data, and stakeholder input. This analysis considers dimensions such as population health needs, provider experience, equity, affordability, and sustainability. Second, the framework emphasizes tracking Health Value Key Performance Indicators (HV-KPIs), which are quantifiable metrics—such as clinical outcomes, patient engagement, reduced ER visits, or improved health equity—that organizations can use to monitor progress alongside financial metrics over time. These tools are intended to help investors make more informed decisions and report transparently on their impact.

Case studies from CIC members illustrate how the framework is already being used in real-world settings. For example, Attane Health, backed by American Heart Association Ventures, has demonstrated improvements in blood pressure and self-reported health outcomes among patients with diabetes by addressing food insecurity through digital coaching. Other examples include maternal care startup SimpliFed and behavioral health app Sober Sidekick, both of which show promising health outcomes linked to strategic private investment. By applying HV-ROI in such cases, investors and operators can validate not only the business case for their innovations but also their societal value.

Ultimately, the white paper argues that embedding health value assessments into investment decision-making is not just possible but essential. In a future where value-based care becomes the norm, investors who prioritize both financial and health returns may also see the most sustainable long-term gains. The CIC calls for wider adoption of HV-ROI principles, supported by shared case studies, common metrics, and ongoing collaboration between investors, policymakers, and health care leaders. This framework positions private capital not as a threat to public health goals but as a vital partner in achieving them.



###

  What's important?   What's missing?  What are weak points?

###

Excellent questions. Here’s a deeper analysis of the Duke-Margolis HV-ROI white paper—focusing on what’s most important, what may be missing, and where the weaknesses lie.


What’s Most Important

  1. The Core Concept of HV-ROI
    The idea that health outcomes and financial returns can and should be jointly assessed is the white paper’s core contribution. It provides a practical, structured framework for evaluating whether investments actually improve health—not just generate profit.

  2. Institutional Legitimacy & Cross-Sector Buy-in
    The involvement of trusted entities—like Duke-Margolis, major health investors (Cigna, KP Ventures), and respected figures like Cheryl Pegus and Mark McClellan—gives the initiative credibility and positions it to influence norms across the private equity sector.

  3. Case Study Integration
    The inclusion of early real-world examples (e.g., Attane Health, SimpliFed) moves the framework beyond theory. These stories help validate that HV-ROI is feasible and actionable.

  4. Timing and Policy Context
    The white paper acknowledges that with federal health budgets under pressure, private investment will become more central to care delivery innovation. This makes the HV-ROI framework especially timely in shaping the next wave of capital flows in health care.


๐Ÿ•ณ️ What’s Missing

  1. Enforcement or Accountability Mechanisms
    The framework is voluntary and self-regulatory—there’s no mention of how HV-ROI assessments might be audited, made transparent, or required in funding cycles. This limits its teeth as a governance tool.

  2. Benchmarking and Comparability
    While the paper mentions tracking KPIs, it does not yet provide standardized benchmarks (e.g., What is “good” performance? How do we compare across investments?). Without these, data from different companies may be inconsistent or non-comparable.

  3. Equity-Specific Metrics
    Equity is cited as a goal, but the framework would benefit from clearer definitions or metrics tied to equity performance—e.g., reductions in disparities by race/ethnicity, income, geography, etc.

  4. Exit Strategy Alignment
    There's limited discussion of how to align exit incentives (e.g., IPOs, M&A) with long-term health outcomes. What happens to HV-ROI goals when companies are sold? This is a real-world challenge in PE/VC-backed health care.

  5. Provider and Patient Voice
    The white paper is investor-led and policy-heavy. While there are references to impact on providers and patients, their perspectives aren’t centered in the development of the framework. This may limit stakeholder trust or relevance in the field.


⚠️ Potential Weaknesses

  1. Adoption Risks
    HV-ROI may appeal to mission-aligned funds, but mainstream private equity firms—especially those operating on short time horizons—may resist the extra complexity or delay in ROI measurement that HV-ROI entails.

  2. Operational Burden on Startups
    For early-stage companies, collecting health outcome data, equity metrics, and community impact KPIs can be expensive and distracting. Without funding or support for measurement infrastructure, startups may struggle to comply.

  3. Possibility of “Health-Washing”
    Without rigorous standards, investors could cherry-pick favorable health metrics to justify deals that do not meaningfully improve health. In other words, there’s a risk of superficial use of the HV-ROI label.

  4. Underdeveloped Link to Value-Based Care Payment Models
    While HV-ROI is aligned in spirit with value-based care, the white paper does not directly connect the framework to ongoing payment reforms (e.g., ACOs, bundled payments, Medicare Advantage innovation). Strengthening this tie could add strategic depth.


Would you like recommendations on how to improve or extend the framework—either from a policymaker, investor, or startup’s perspective?

YES

###

Excellent—here’s a set of tailored recommendations for how to improve or extend the HV-ROI framework, viewed through three distinct lenses: policy makers, investors, and startups.


๐Ÿ”ท 1. Policy Makers: Drive Standardization and Incentives

What to Improve

  • Standard Metrics & Benchmarks: CMS, HHS, or NIH could develop or endorse core HV-ROI KPIs (e.g., risk-adjusted readmission rates, health equity scores) to allow benchmarking and comparability across investments.

  • Public Reporting Expectations: Encourage or require public disclosure of health impact KPIs for entities receiving government-linked funding (e.g., Medicare Advantage startups, Medicaid managed care).

How to Extend

  • Link to Value-Based Payment Models: Align HV-ROI adoption with value-based payment contracts (e.g., ACO REACH, Medicare Shared Savings Program), creating a bridge between payment reform and private investment.

  • Incentivize via Tax or Grant Policy: Offer incentives such as tax advantages, CMMI innovation funds, or priority review for PE/VC funds and companies using validated HV-ROI practices.

  • Create an HV-ROI Certification or Seal: A government or quasi-government body could develop a voluntary seal of approval—akin to ENERGY STAR—for investors and companies that meet transparent HV-ROI standards.


๐Ÿ”ท 2. Investors: Strengthen Credibility and Risk Management

What to Improve

  • Independent Evaluation Partners: Collaborate with academic institutions or third-party evaluators to validate health outcomes, reducing concerns of bias or selective reporting.

  • Portfolio-wide HV-ROI Dashboards: Aggregate HV-KPIs across companies to manage portfolio-level health impact risk, much like ESG reporting for environmental/social governance.

How to Extend

  • Embed HV-ROI in Term Sheets and Exit Criteria: Make health KPIs part of investor-company agreements and carry terms, not just pre-deal due diligence. This ensures alignment over time and across leadership changes.

  • Collaborative Data Sharing Platforms: Create industry-wide consortia (similar to the Sustainability Accounting Standards Board) for HV-ROI benchmarking and shared learnings.

  • Tie GP/LP Reporting to HV-ROI: Venture firms could include HV-ROI summaries in their quarterly reports to limited partners (LPs), appealing to impact-minded institutional investors.


๐Ÿ”ท 3. Startups & Operators: Make HV-ROI Feasible and Actionable

What to Improve

  • Practical Tools & Templates: Develop lightweight HV-ROI planning kits (with sample KPIs, dashboards, and outcome calculators) that early-stage companies can use without major cost or staff burden.

  • Operational Capacity-Building: Funders or CIC members could offer grants or shared services (e.g., analytics support, access to public health data) to help companies track and report outcomes.

How to Extend

  • Narrative Use of HV-ROI in Fundraising: Founders can use the HV-ROI model to frame impact stories during fundraising—especially for Medicaid/underserved market plays—to differentiate from competitors.

  • Incorporate into Product Roadmaps: Treat HV-KPIs as product features. For example, a digital health app could build dashboards showing users’ improved health metrics, which also feed investor reports.

  • Join Learning Collaboratives: Encourage startups to participate in peer working groups or open-source forums sponsored by Duke-Margolis or leading investors, creating a real-time feedback loop for evolving HV-ROI norms.


๐Ÿงฉ Cross-Cutting Suggestion:

All three stakeholders—policy makers, investors, and operators—would benefit from a national HV-ROI clearinghouse, possibly hosted by Duke-Margolis or a neutral convener. It would provide:

  • Case studies and dashboards

  • Data submission templates

  • Peer comparisons

  • Research partnerships

Such a resource could accelerate the legitimacy, usability, and spread of HV-ROI as a norm—not just a proposal.


Would you like this adapted into a 1-pager for any of those audiences?

Tuesday, March 25, 2025

FDA and Lab-developed Collection Kits: Law Firm Article

I am a reimbursement expert, not an FDA expert, but I will point out to clients who ship proprietary (non-FDA) collection kits, the FDA has sometimes been adverse to those.  (Not commonly.)

Zegarelli and Hawana of the Mintz law firm post an article about collection kits - titled, "FDA's Backup LDT Enforcement."

Find it here:

https://natlawreview.com/article/fdas-backup-ldt-enforcement-method-specimen-collection-kits

Pointed out by Heather Thompson at Linked In (here).


##

AI CORNER

##
AI summary.

TL;DR – Careful Summary of “FDA’s Backup LDT Enforcement Method: Specimen Collection Kits” (Mintz, March 13, 2025)

Even as FDA’s new rule to regulate laboratory-developed tests (LDTs) faces legal challenges, the agency has another powerful enforcement pathway: specimen collection devices and kits. These are separately regulated medical devices under existing FDA rules, even when used with LDTs.

Key Points:

  • Specimen collection devices and kits (e.g., saliva tubes, lancets, dried blood spot cards) are considered in vitro diagnostic (IVD) medical devices, regulated independently of the LDTs they support.

  • When multiple collection components are assembled and sold together (e.g., a branded home saliva kit), FDA considers this a “convenience kit” subject to device regulations, including establishment registration, labeling, quality systems, and sometimes 510(k) clearance.

  • For at-home collection kits, FDA requires human factors data showing safe use by laypersons, often a high regulatory hurdle.

  • Some labs and kit providers mistakenly assume these kits fall under LDT enforcement discretion—they don’t. FDA can target these kits even if LDT oversight is blocked or delayed.

  • Noncompliant kits—such as those marketed for unauthorized uses or directly to consumers without clearance—are at high risk of enforcement, especially if tied to serious health conditions like cancer or infectious diseases.

  • Clinical labs may face a Catch-22: to get a collection kit cleared, they must show it's safe with their specific LDT, but can't collect data without clearance—potentially exposing the LDT to FDA scrutiny.

Strategic Implication:

Regardless of what happens to the LDT final rule, the FDA can still enforce compliance by focusing on collection kits. Labs must ensure the kits they develop or distribute meet all applicable FDA device regulations—or obtain them from compliant third-party manufacturers. State agencies like New York CLEP are also emphasizing this requirement.

Bottom Line:

Specimen collection kits are a regulatory choke point for LDTs. Even if broader LDT regulation is stalled, FDA oversight of these kits is real, expanding, and enforceable now. Labs using or distributing them need a clear compliance strategy—or risk enforcement.

Monday, March 24, 2025

Nerd Note: MAC Contacts Table at CMS Database

 Are you looking for names and emails of CMS medical directors?   There's a report for that.

Go to the Medicare Coverage Database and click on REPORTS.   Under "Local Coverage" find the "MAC Contacts Report."    (Prior, there often had been a PDF of MAC contacts buried somewhere on the CMS website.)


###

AI CORNER

###
Chat GPT writes an excited article about the big news.

Nerd Note: 

How to Find CMS MAC Medical Directors 

(the Easy Way)

Are you on the hunt for CMS medical director names and emails like a hound dog sniffing out a scent trail? You’re in luck.

For years, Medicare stakeholders played “Where’s Waldo?” when trying to find who was in charge at their local MAC. If you were lucky, you stumbled across a crusty old PDF on the CMS website listing contact points—usually two versions out of date.

No more. CMS has given us a grown-up toy train set: a live, sortable MAC Contacts Report on the Medicare Coverage Database. This is your one-stop-shop for finding out who’s who at each MAC—from CGS to Noridian to Palmetto and beyond. Medical directors, emails, contracts, coverage areas—you name it, it’s here.

๐Ÿ‘‰ Head over to the CMS Medicare Coverage Database Reports page.
๐Ÿ‘‰ Click on “MAC Contacts Report” under the “Local Coverage” section.
๐Ÿ‘‰ Or go directly to the MAC Contacts Report here.

You’ll see contacts like:

  • Meredith Loveless, MD, Chief Medical Officer for CGS Part A and Part B (covering Kentucky and Ohio)

  • Neil Sandler, MD, leading CGS’s HHH MAC work across states from Colorado to Virginia

It’s all in there. You can even download the Excel file (the highlight of our week!) and nerd out on the full list, like any self-respecting policy wonk with a thing for coverage logic, LCDs, and email addresses ending in @cgsadmin.com.

So go ahead, bookmark it, share it, and bask in the glow of government transparency—one Excel file at a time.



Friday, March 21, 2025

AI in the Office: Chat GPT Reads NEJM Article on Burnout, Summarizes, Provides Infographic

 Interesting title in NEJM this week:

Evidence-Based Work Design — Bridging the Divide, by by Burden and Dirbye, 392:1044.

###

Chat GPT provides a simple "infographic," then a short summary, then answers a question about what might actually be envisioned.

####

AI CORNER

####

AI Infographic

Evidence-Based Work Design — Bridging the Divide (NEJM, March 2025)

The article by Burden and Dyrbye argues that corporatization and financially driven decisions in health care are fueling burnout, moral injury, and inefficiencies among health care workers (HCWs). Much of the harm comes from modifiable but poorly designed work structures, which the authors label as “administrative harm.”

They call for a shift to evidence-based work design, using data-driven and iterative methods — similar to clinical evidence-based practice — to align job demands and resources for better outcomes. This includes using EHR metadata, AI tools, and predictive analytics to monitor and improve workload and team structures.

They advocate for interdisciplinary collaboration and innovative research (e.g., simulations, rapid trials) to guide best practices. The ultimate goal is to support health systems in achieving the quadruple aim: better health outcomes, improved patient experience, lower costs, and enhanced HCW well-being.

###

Can you give some more specific examples?

###

Great question — you're right that the article lays out a compelling framework but remains high-level on specifics. Based on the text and its implications, here are some specific changes that the authors likely envision as part of implementing evidence-based work design:


๐Ÿ”ง 1. Redesigning Staffing Models

  • Shifting from RVU-based productivity metrics to evidence-based staffing thresholds (e.g., optimal patient load per hospitalist or per nurse).
  • More thoughtful deployment of team members (e.g., when to use MDs vs. nurse practitioners vs. both) based on empirical outcomes rather than cost or habit.
  • Flexible team structures based on case complexity, not flat headcount.

๐Ÿ•’ 2. Monitoring and Adjusting Workload in Real Time

  • Use of EHR metadata (e.g., message volume, after-hours charting, click fatigue) to trigger automated flags when workload is approaching burnout thresholds.
  • Early-warning systems for unsafe staffing levels, akin to patient deterioration alerts.

๐Ÿ“Š 3. Embedding Simulation or AI-Based Planning Tools

  • Simulation models to predict the downstream effects of staffing cuts (e.g., how reducing a unit’s nurse coverage increases falls or LOS).
  • AI-powered planning tools to optimize shift design, break timing, or patient handoff processes.

๐Ÿ’ฐ 4. Reframing Financial Incentives

  • Moving from short-term cost-cutting to total cost of care modeling (e.g., recognizing that fewer doctors per shift may lead to higher readmission rates or burnout-driven turnover costs).
  • Payer or CMS pilot programs that reimburse based on staff well-being metrics or organizational sustainability.

๐Ÿ“‰ 5. Incorporating Work Design into Safety & Quality Dashboards

  • Adding workforce strain indicators to existing quality dashboards (e.g., ratio of in-basket messages per hour; % of after-hours charting).
  • Tracking clinician attrition and burnout as key organizational metrics, not just HR stats.

๐Ÿงช 6. Redefining What Gets Studied in Health Services Research

  • RCTs or robust observational studies comparing, say, 10 vs. 15 patient loads per hospitalist in terms of safety, satisfaction, and discharge efficiency.
  • Studies assessing which team configurations (e.g., solo NP vs. NP + MD) yield better chronic disease outcomes or fewer ED visits.

๐Ÿฅ 7. Co-Designing Workflows with Frontline Staff

  • Including clinicians in administrative design decisions through structured “co-design” processes.
  • For example, redesigning the inbox triage system with those who use it daily, not just IT or finance staff.

๐Ÿšจ 8. Guardrails on AI Implementation

  • Requiring evidence of time savings and safety outcomes before rolling out AI tools that increase clinician responsibilities (like documentation).
  • Avoiding the trap of offsetting AI efficiency gains by increasing visit volume quotas.

These would represent a paradigm shift: from reactive, finance-first staffing to a data-informed, adaptive, and humane approach to organizing clinical work. It’s an attempt to bring the same level of rigor to workflow design that we expect in prescribing a medication or evaluating a clinical intervention.

Want a summary of this list in a visual format (like a table or infographic)?


##

WSJ: Doctors are as upset as patients.  Here.

Why AI-driven productivity is hard to measure in general.  Here.  Same topic but, why healthcare AI productivity is hard to measure.  Here.



Thursday, March 20, 2025

CMS Updates NCD Coverage for Cologuard, Guardant Shield, Illumina Trusight, Caris

CMS regularly publishes updates to its coding and coverage instructions for NCDs.   Find them here:

https://www.cms.gov/medicare/regulations-guidance/transmittals/2025-transmittals

On March 20, 2025, CMS released CR 13939 (T13097) updating its colorectal screening NCD and its NGS-tumor NCD.

https://www.cms.gov/medicare/regulations-guidance/transmittals/2025-transmittals/r13097otn

Bonus

At least when this PDF was released, the CMS reference to a cloud instructions file was a "dead link." ("Please follow the link below for the NCD spreadsheets ...")  This is a cloud copy I received (note it has many tabs).

https://docs.google.com/spreadsheets/d/1iYzETIFix-HfzHYxaTlU7QTdqNm9hTGc/edit?usp=sharing&ouid=110053226805181888143&rtpof=true&sd=true

##

NCD 90.2 (NGS in Cancer) is updated for:

0448U, Pillar Oncoreveal Lung/Colon (Delete)
0523U, Pillar Oncoreveal CDx. (Start)

These are a little quirky because a recent AMA PLA update (Jan 1 2025) deleted 0448U effective January 1.  That AMA PLA update also made 0523U (22 genes FFPE) effective January 1, 2025.

ILMN Trusight Comprehensive, under code 81455 from August 21 2024 and under code 0543U effective April 1, 2025.    

Caris MI Cancer Seek 0211U, effective November 5, 2024.

##

NCD 210.3 CRC Screening is also updated.

Add 74263, screening CT colonography.
Add Cologuard Plus 0464U (October 3, 2024)
Add 0537U CRC Guardant SHIELD (July 26, 2024)

There is also a comment, "remove stool-based language from policy to all for all non invasive CRC screening tests" (e.g. apparently stool or other media).  There is also a comment, "Subsequent omnibus CR with additional editing for expansion of colorectal cancer screening policy."  To me, these dangling comments are just cryptic.  NCD 210.3 actual text would be updated here.

There is a link to what I suspect should be a giant excel file, but for me today it's a dead link.

https://www.cms.gov/Medicare/Coverage/DeterminationProcess/downloads/CR13939.zip




Wednesday, March 19, 2025

The Year in AI: A Special Issue of "Sensitive and Specific" Newsletter

The newsletter "Sensitive and Specific" is worth tracking for frequent lab industry updates and insights.

I especially enjoyed their new issue, "AI Year in Review," covering a wide range of topics in the clinical lab, imaging, and beyond.

Find it here:

https://sensitiveandspecific.substack.com/p/sensitive-and-specific-the-year-in

It seemed appropriate to try an AI review of the special issue.



###

AI Corner

###

The 2024 AI Explosion in Diagnostics: Hype, Hope, and Hard Questions

The latest issue of Sensitive and Specific is a fitting encapsulation of 2024—a year in which AI didn’t just enter the diagnostic landscape; it steamrolled through it. The sheer breadth of developments covered in this special issue reinforces a point many in the molecular and diagnostic industries have sensed for some time: AI is not a tool of the future; it is the present.

While AI has already transformed radiology and imaging, the newsletter rightly highlights how digital pathology and patient-reported data analysis are rapidly gaining traction. The next challenge isn’t technical capability—it’s proving scalability, achieving physician buy-in, and navigating the regulatory maze. The comparison to the early days of genomic testing is apt. AI’s biggest barrier isn’t whether it works (it does); it’s whether clinicians trust it enough to change their workflows and decision-making.

AI Across Diagnostics: A Tipping Point?

The special issue lays out an impressive collection of real-world AI applications, from predicting pancreatic cancer 475 days before diagnosis to guiding ultrasound exams for non-specialists with expert-level accuracy. Some highlights:

  • Cancer detection: AI-assisted colonoscopy improved polyp detection by 40%, and AI-enhanced CT scans identified pancreatic tumors over a year earlier than standard screening.
  • Cardiology: Deep learning models can now diagnose coronary artery disease from retinal images—a development that could shift how we think about non-invasive screening.
  • Mental health & neurology: AI-based speech analysis is making strides in depression diagnosis, and machine learning models are showing promise in Alzheimer’s prediction.

And yet, the issue also captures AI’s limitations—notably, that AI doesn’t always improve radiology accuracy and that many AI-powered diagnostic tools remain uncovered by insurance.

Beyond the Hype: The Policy and Adoption Questions

The special issue smartly recognizes that AI’s greatest challenge is no longer whether it works—it’s how we integrate it responsibly into healthcare. The final section highlights:

  • Regulatory shifts: HHS now requires AI vendors to disclose training data—a crucial move for transparency and bias reduction.
  • Liability and reimbursement: Who is responsible when an AI-assisted diagnosis goes wrong? Will insurers pay for AI-enhanced diagnostics? These questions remain unanswered.

The Genie is Out of the Bottle

Perhaps the most important takeaway from this issue is that AI’s role in diagnostics is no longer theoretical. From early cancer detection to AI-driven sepsis prediction, 2024 cemented AI as an indispensable part of the diagnostic ecosystem. The wait-and-see approach that once applied to genomics is now being tested again, this time with AI. If history is any guide, those who embrace AI early will define the future of diagnostics.

For molecular and diagnostics professionals, this issue of Sensitive and Specific is required reading—not just to understand AI’s current capabilities, but to prepare for the seismic changes still to come.

Tuesday, March 18, 2025

AdvaMed Strongly Supports Dr. Oz for CMS; Wins Favorable Remarks; Lessons for AMA?

 In November 2024, AdvaMed's CEO rapidly released a favorable statement about the nomination of Dr. Oz to lead CMS.    Press release, news coverage:

https://www.advamed.org/industry-updates/news/advamed-statement-on-president-elect-trumps-intent-to-nominate-dr-oz-to-lead-cms/

https://www.medtechdive.com/news/Advamed-Dr-Oz-Trump-pick-CMS/733486/

At the CMS Senate hearing last Friday, Dr. Oz promptly featured his interest in rapid coverage for new technologies at Medicare.   The exchange was prompted by Sen. Daines (R-MT).

https://www.medtechdive.com/news/dr-oz-tcet-medicare-coverage-medical-devices/742611/

The sequence is, AdvaMed strongly supports the Republican nominee; then the Republican nominee gives a presentation where he in turn prioritizes a goal of AdvaMed. 

####

####

This contrasts with a crisis brewing at AMA, where for several years, and worst this year 2025, physician payments for services have badly lagged inflation.    

https://www.ama-assn.org/system/files/2025-medicare-updates-inflation-chart.pdf

For nearly the first time in history, physician pay cuts continue, and were not fixed by a Congressional budgeting process.   This is a major hit for physicians and all services valued in RVUs (e.g. non-hospital-based MRI).

It's difficult to know how much a role AMA and JAMA positions constantly disagreeing with the administration played into the pay cuts.   But, a minimalist political argument would just point to the lack of AMA and JAMA positions supporting the administration.   For example, the AMA's journal JAMA Forum currently writes, "President Trump nominated Robert F. Kennedy Jr, an advocate known for false and misleading statements..." without qualifying the assertion.   That's their business, but it must make it harder for AMA to then negotiate sitting next to the same RFK Jr a few weeks later.  

And it's not just RVU's; UCSF's Vinay Prasad asserts the HHS is looking at AMA's "monopoly" on coding, a major source of revenue for AMA:



###

###

3/20: Dr Oz in Charge of Medicare and Medicaid.

https://www.washingtonpost.com/opinions/2025/03/20/mehmet-oz-cms-medicare-trump/

3/20:  Dr Oz, the Grown-Up in the Room

https://www.theatlantic.com/health/archive/2025/03/dr-oz-senate-hearing/682102/



Chat GPT Answers a Question about U.S. ICD-10-CM

 I can't guarantee the precision of this AI answer but I believe it's directionally correct.

QUESTION

ICD-10 is the "international" classification of disease.   But it is also maintained by CDC and CMS, and edited and updated by them regularly.  Do what extent is our ICD-10 CM in the US "international" and to what extent "local" due to these CDC CMS updates?

### Chat GPT 4o:

ICD-10-CM (Clinical Modification) in the U.S. is both international and local due to how it is derived from and modified from the global ICD-10 standard.

International Aspects

  1. Core Structure from WHO – The ICD-10 system itself is maintained by the World Health Organization (WHO) as an international standard for disease classification.
  2. Fundamental Coding Categories – The broad chapters, category structures, and many codes remain consistent with WHO’s ICD-10 to ensure international comparability.
  3. Crosswalks with Global Systems – Because it is derived from ICD-10, ICD-10-CM maintains alignment for global epidemiological and statistical tracking.

Local (U.S.-Specific) Aspects

  1. Expanded Clinical Modification (CM) – The U.S. CDC and CMS developed ICD-10-CM to significantly expand ICD-10, adding thousands of additional diagnosis codes tailored to U.S. clinical and billing needs.
  2. Annual Updates by U.S. Agencies – Unlike the WHO version of ICD-10, which is updated less frequently, ICD-10-CM is updated annually by CDC and CMS based on U.S. healthcare policy, reimbursement needs, and clinical practice changes.
  3. U.S.-Specific Codes – Many codes are unique to ICD-10-CM and do not exist in the WHO’s ICD-10, such as codes for:
    • Specific social determinants of health (e.g., Z codes for housing or food insecurity)
    • U.S.-centric diseases and conditions (e.g., opioid epidemic-related codes)
    • Greater specificity for certain conditions to align with U.S. billing and compliance needs.
  4. CMS Influence on Coding Rules – The Medicare & Medicaid Services (CMS) define how ICD-10-CM codes are used for reimbursement, which adds an additional layer of U.S.-specific coding guidance that differs from how ICD-10 is used globally.

Conclusion

While ICD-10-CM is rooted in the WHO’s ICD-10, it has evolved into a heavily modified version tailored to U.S. healthcare, reimbursement, and public health tracking. This means international comparisons remain possible at the broader category level, but at the specific code level, ICD-10-CM is highly localized.





Monday, March 17, 2025

CMS Hasn't Released CY2023 Summary Part B Data (?)

 In my experience, dating back years, CMS releases Part B data by CPT code in two tranches.

In November, they release excel spreadsheets of national CPT usage and payment data, for the prior calendar year.   These go on a page called, "Part B National Summary Data File."   

https://www.cms.gov/data-research/statistics-trends-and-reports/part-b-national-summary-data-file

In around June, they release very detailed data by lab and/or physician, in a cloud database.  This is data for the year 18 months prior.

https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners

Specifically, "Medicare physician and other practitioners by provider and service."

https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners/medicare-physician-other-practitioners-by-provider-and-service

##

However, right now in March 2025, the most recent data I can find is for CY2022 (which is available for both types of data above, the national Excel and the cloud physician- or lab-level data).    So what is missing is, the CY2023 data that should have appeared in November 2024 at the first link shown above (National summary data file) but has not appeared.

###

There is one CMS cloud database that IS updated to 2023, which is "Physician Supplier Procedure Summary." [Link below.]  However, I haven't been able to sort and filter this 14M line database in a way that rolls up the data usefully without breaking each CPT code into too many lines.   

So I haven't published my normal Autumn summary of prior-year national data, such as fall 2024 report for CY2023.  For example, we'd ask what proportion of 2023 lab claims were under 81479, and other topics.

 https://data.cms.gov/summary-statistics-on-use-and-payments/physiciansupplier-procedure-summary/data


Sunday, March 16, 2025

CMS Releases NCD Proposals with CED; Calls it TCET; Relevant Article by Hernรกn et al

Background

Last year the Biden administration released a roadmap for coverage with evidence development (CED) called TCET - Transitional Coverage for Emerging Technologies.   The administration proposed 5 NCDs of this type per year, with review of each 2 years out.   At scale, this means 10 NCD reviews per year under TCET.   How CMS will pull this off has puzzled me, since CMS only does 2-3 NCDs per year.

Some initial data has emerged.   CMS has recently released 3 CED proposals that refer to TCET.  One is  for trans-vascular cardiac valves, a topic where CMS has done numerous CED NCDs for a decade.  In short, some types projects already being done under NCD-CED will now be tallied in the TCET column.

  • See CED for renal denervation, proposed 1-14-2025, CED for cardiac contractility, proposed 1-10-2025, CED for transcatheter tricuspid valve, proposed 12-19-2024.   
  • These all have CED, and all refer to TCET.
  • For example, we read, "CMS received a complete, formal request to provide coverage for the EVOQUE tricuspid valve replacement system (EVOQUE system). This is a Transitional Coverage of Emerging Technology (TCET) pilot. The manufacturer of this device tested the processes and concepts of TCET."
    • (CMS also proposed, on March 11, an NCD for home ventilation, a DME-like product, with no CED.)   
    • (CMS also finalized, on February 11, an NCD on pulmonary heart failure centers, that has CED, but doesn't mention TCET).

? Structure for CED:

See Hernรกn et al, 2025, "The Target Trial Framework for Causal Inference From Observational Data: Why and When Is It Helpful?"



While it's subscription-based at Annals of Internal Medicine, I was struck by the high-quality thinking in Hernรกn et al. (here).  Most CED studies have been based on registries, rather than detailed RCT's.   Hernan et al. describe an important approach to thinking about observational studies.    You should first lay out, in detail, a randomized controlled trial that answers the question that must be answered.    (This is the "Target Trial.")   Once you have done that, look closely at whether observational data (including a de factor control or comparison) can address the underlying question.    Hernan et al. argue that when this is done, and successfully, the observational data is very likely to be valid.   When the observational power falls short of key findings that an RCT would have provided, conclusions (if any) from the observational data are likely to be lacking or not be confirmable.   That's a summary; the full article lays out the logic and uses many examples.

For my money, the level of thinking in Hernรกn et al. goes beyond the level of logic brought to most discussions of CED.

Diagnostic Tests

Hernรกn et al. focus on interventional trials with therapies - you get a drug or placebo; you get a surgery or you don't.   With diagnostic tests, we more likely have accuracy data, clinical context, superiority to standard of care diagnostics, and decision impact.   You don't, for example, take a woman with a very low Oncotype score and give her chemotherapy, or a woman with a very high Oncotype score, and deny her chemotherapy.   

I've argued for years that simply evaluating diagnostics under the rubric "analytical validity, clinical validity, and clinical utility" is too vague, and more logic is required.   (It might be assumed that all the thinking and logic is just recreated for each new test assessment, pegged to AV, CV, CU).   In 2014, Frueh and I (here) wrote a paper on "defining clinical utility" where we argued that about six or seven questions are enough.   (One is too few - "do you have clinical utility" and 30 or 40 questions is too many).   These involved:

  1. What is the population?
  2. What is the standard of care test?
  3. What is the new test?
  4. What is the improvement obtained with the new test (e.g. #3 minus #2 = delta).   
  5. How much COULD that improvement affect clinical outcomes?
  6. How much DOES that improvement affect clinical outcomes?
  7.  Some measure of cost effectiveness or efficiency.
While we published this in 2014, and it's sometimes quoted, my experience remains if you badly fail some of these questions, a T.A. probably won't go well, regardless of exactly what method the T.A. uses.  If you can give a tight, logical answer in 2-3 sentences to all the questions, you'll probably do OK.

Similarly, for therapeutic questions and with observational data to look at, Hernรกn et al. provide a logical framework which should give decision-makers a lot of traction.

While I'm not at the level of Hernรกn et al., both their article and mine on diagnostics aimed to shine a light on ways of making data plans more sound and logical.

###
###
###
Hernรกn et al., Abstract

When randomized trials are not available to answer a causal question about the comparative effectiveness or safety of interventions, causal inferences are drawn using observational data. 

A helpful 2-step framework for causal inference from observational data is 1) specifying the protocol of the hypothetical randomized pragmatic trial that would answer the causal question of interest (the target trial), and 2) using the observational data to attempt to emulate that trial. The target trial framework can improve the quality of observational analyses by preventing some common biases. 

In this article, we discuss the utility and scope of applications of the framework. We clarify that target trial emulation resolves problems related to incorrect design but not those related to data limitations. [BQ - And it highlights which is which.]  

We also describe some settings in which adopting this approach is advantageous to generate effect estimates that can close the gaps that randomized trials have not filled. In these settings, the target trial framework helps reduce the ambiguity of causal questions.

###
###
AI CORNER
###
###
See a Chat GPT 4o summary of the article.

###
Final point.  This article discusses CED going forward as part of NCDs.  However, during the first Trump admininstration, HHS General Counsel took the position, CMS should not be doing CED at all, for legal reasons.   Here, here (Charrow 2021).

See an update 2025 on CMS coverage, Tunis et al.