Sunday, June 30, 2024

Terms in Collision; Classical Tumor Naming vs Molecular Tumor Naming

 Interesting set of articles.  How should we name & classify tumors today?

In January 2024, Andre' et al. argued in NATURE that tumor nomenclature must change, must be "unstuck" from traditional organ types.   Here.

In a new June 2024 article, Ilie' et al. respond in J Clin Oncol that Andre's viewpoint raises as many issues as it solves, and a path forward should encompass both tissue and molecular features.  Here.  (It's not direct open access, but see next.)

One of the Ilie authors, David Horgan, describes the two papers in a detailed posting on Linked In.  Here.  It includes a Linked-In, downloadable, copy of the Ilie' PDF.  


See an article by Fasola 2022 that modern molecular tumor boards  shouldn't just "tack on molecular" but require a rethinking from the ground up.  Here. [abstract]. On systems redesign, see also Dagogo-Jack 2023 (here).


Tumor classifications adjust greatly over time.   I recall that while in neuropathology fellowship in the 1990s, I got a copy of Percival Bailey's major textbook, Classification of Tumors of the Glioma Group (here) which felt as hard to read as Chaucer, from the perspective of a student trying to decipher tumor nomenclature.  That much had changed in 65 years,1926-1991.


AI Corner.

Chat GPT4 discusses the articles. [ /Start AI text. ]

Let's begin by outlining each article's core arguments and contributions to the ongoing debate about tumor naming and classification in oncology.

ANDRE Article: "Tumor Classification Must Go Molecular"

The "ANDRE" article published in Nature argues for a radical shift from traditional organ-based cancer classifications to molecular-based classifications. Here are the key points:

  • Organ-based Limitations: The authors highlight that the current system, which classifies cancers based on the organ of origin, is outdated and restricts access to effective treatments for patients with metastatic cancers.
  • Advancements in Molecular Profiling: The article points out that molecular profiling has allowed for a more precise understanding of tumors, which can significantly enhance targeted therapy.
  • Impact on Clinical Trials: The reliance on organ-specific classifications has led to delays in clinical trials for drugs that could be effective across multiple cancer types but are only tested sequentially for each organ-specific type.
  • Educational and Clinical Implications: There is a call for educational reform to focus more on molecular mechanisms rather than memorizing organ-specific trials, which would streamline medical education and improve clinical practice.

Ilie Article: "Navigating Change in Tumor Naming: Exploring the Complexities and Considerations"

The "Ilie" article published in the Journal of Clinical Oncology provides a counterpoint to the "ANDRE" article by proposing a more cautious approach to the shift towards molecular classifications:

  • Acknowledgment of Complexity: It agrees that molecular profiling enhances cancer treatment but emphasizes the complexity of completely transitioning to molecular classifications.
  • Balanced Approach: The authors advocate for a dual approach that integrates both organ-based and molecular classifications, acknowledging the merits and limitations of each.
  • Practical Considerations: Concerns are raised about the feasibility of universal molecular profiling, particularly regarding the availability of targeted therapies for all patients and the challenges in clinical trial design.

Horgan's LinkedIn Blog: Reflections on Molecular Classifications

The blog by Denis Horgan on LinkedIn provides additional context and support for the molecular approach:

  • Advocacy for Change: Horgan supports the shift towards molecular classifications, pointing out the potential for this approach to revolutionize personalized medicine and reduce inequities in treatment access.
  • Real-world Implications: The blog brings our attention to the practical implications for regulatory frameworks and healthcare systems, stressing the need for adaptive policies that can accommodate rapid advancements in cancer treatment.


In the evolving landscape of oncology, the debate over tumor classification—whether to stick with traditional organ-based methods or transition to molecular-based classifications—has never been more pertinent. 

  • The "ANDRE" article from Nature makes a compelling case for a shift towards molecular classifications, arguing that this approach aligns better with the advancements in precision medicine and can significantly enhance the efficacy of targeted therapies. 
  • On the other hand, the "ILIE" article in the Journal of Clinical Oncology urges caution, advocating for a balanced approach that integrates both molecular and organ-based classifications to address the practical and clinical challenges of a complete transition.

Denis Horgan’s reflections on LinkedIn further enrich this debate by emphasizing the potential of molecular classifications to revolutionize patient care and reduce treatment disparities. However, he also highlights the need for adaptive regulatory and healthcare infrastructures to support these changes. 

Today we stand at this critical juncture, and it is clear that while the promise of molecular profiling is vast, its implementation must be approached with a nuanced understanding of its impacts on clinical practice, trial design, and educational frameworks. As cancer experts, pathologists, and biotech or genomic developers, our roles in navigating and shaping this transition will be crucial in ensuring that the benefits of molecular insights are realized across all aspects of patient care. [ /End AI text ]

Friday, June 28, 2024

Policy Watcher Again: New SCOTUS Cases Continue to Support ACLA v FDA Case

UPDATE: Sen. Cassidy writes the FDA, says Chevron ruling should call into question the FDA's LDT land grab.  Here.


 A few days ago, I wrote a blog that as a long-time policywatcher - not an attorney - I felt several recent Supreme Court cases were supportive of the ACLA's case against FDA.  See June 21 blog here.

For June 28, I've added an update.

I would suggest that the court's anti-Chevron ruling in Loper v Raimond also supports ACLA v FDA.   Loper finds that courts are not obligated to defer to an agency when a statute is ambiguous.  Rather, a court should review the statute in question de novo.   

Here, that would mean a court does not any longer need to defer to FDA whether the 1976 intention of Congress was for [a laboratory of people and items conducting processes] to be collectively commonly and normally understood to be "a medical device.”  For example, according to 1976 dictionaries or usage at the time.


A new subscription article at 360DX, by Adam Bonislawsky, triages the opinions of several attorneys re the ACLA legal case.


The presidential election could also be a factor.


WSJ notes FDA LDT is one of the policies now in play because of the Anti-Chevron decision - here.

Wednesday, June 26, 2024

Brief Blog: "Happy Birthday to TCET"... Said Exactly Nobody.

Early in 2021, the Biden administraton deep-sixed the recently-completed "MCIT" breakthrough coverage regulation for Medicare.   In 2021 and 2022, CMS began talking about its new plan, "TCET," transitional coverage for emerging technology.  But it existing only as a term, or as a powerpoint deck.

On June 22, 2023, CMS announced the full proposed plan for TCET, with a notice & comment Federal Register publication.  It's never been finalized, and its one-year anniversary as a "proposal" attracted zero attention.

Several bills still float around the Hill to restore more-or-less the automatic breakthrough coverage period offered by MCIT.  The topic was mentioned during some 2024 hearings. See coverage at MedPage Today.

See a February 2024 letter from AdvaMed to CMS asking for quick action on TCET.

Tuesday, June 25, 2024

Brief Blog: Auto Transcript of CMS Annual Lab Meeting (June 25, 2024)

On June 25, 2024, CMS held its annual lab meeting for public comment on pricing of new lab tests.  There were almost 60 presentations from 9 am to 430 pm ET.

It's difficult for an auto-transcript to master this complex audio material, but for all its typos, the 97-page auto-transcript provides an imperfect though skimmable record.   Time-stamps in each of 4 transcript sections will match fairly closely to timepoints within four video uploads I made of the meeting.   

  • For example, 
    • If you skim the agenda of the PDF and want #34, Natera, 
    • search the transcript for "Natera" (it will be page 56 of the PDF), 
    • and it corresponds to about "minute 39" in "Video 3."

VIDEO ONE - #1 MelMont to #18 OncoHost

VIDEO TWO - #18 OncoHost (cont) to #28 CytoVale [lunch]

VIDEO THREE - #29 C2N DX to #44 Exact Sci

VIDEO FOUR - #45 Phenomics to #56 Wash Univ

Find the  table of contents and auto transcript at Google Drive here, with links to YouTube for each part.

Brief Blog: FDA Issues an Explanation of FDA LDT Regulation that's Manageable in Size

Header: FDA produces a 20-page guide to its hundreds of pages of FDA LDT reguatory plans.


On May 6, 2024, the FDA published its massive plan to regulate laboratory-developed tests over five years, a policy decathalon which is like an upside-down pyramid balancing on just a few words of regulatory change (LDT labs are now device manufacturers, quote unquote).

Amid the plethora of FDA rulemaking, guidance documents, attachments, and webinar transcripts, the agency has produce a concise, 20-page guide to it all.


Find at article at RAPS:

Find the FDA webpage for the guide:

Find the actual download PDF:

See the Federal Register page here:

AI Corner

Here's a Chat GPT 4 reading of the document:

The FDA's new document, a Small Entity Compliance Guide, provides a detailed overview for small laboratories on adhering to the new rules for in vitro diagnostic products (IVDs), specifically laboratory developed tests (LDTs). Here’s a concise summary:

  1. Regulatory Background and Purpose:

    • The guide addresses the FDA’s amendment to classify IVDs as devices under the Federal Food, Drug, and Cosmetic Act, regardless of the manufacturing entity. This aligns with a move to phase out the previous enforcement discretion for LDTs, aiming to standardize the regulatory approach across all IVDs.
  2. Phaseout Policy Overview:

    • A phaseout of the previous discretionary approach will occur over four years, after which IVDs must comply with the same regulations as other medical devices. This includes meeting quality standards, reporting requirements, and registration needs.
  3. Compliance Timelines and Requirements:

    • Stage 1 (Starting May 6, 2025): Compliance with Medical Device Reporting and specific Quality System requirements.
    • Stage 2 (Starting May 6, 2026): Compliance with registration, listing, and labeling requirements.
    • Stage 3 (Starting May 6, 2027): Compliance with broader Quality System regulations.
    • Stage 4 (Starting November 6, 2027): Premarket review requirements for high-risk IVDs.
    • Stage 5 (Starting May 6, 2028): Premarket review for moderate and low-risk IVDs.
  4. Special Considerations:

    • Certain categories of IVDs, such as those used solely for forensic purposes or within certain federal health facilities (like the Veterans Health Administration), will continue to benefit from enforcement discretion.
  5. Resources and Support:

    • The FDA provides additional resources to assist labs in navigating these new requirements, including specific guidance documents and web-based resources.

This guide is aimed at ensuring that all laboratory-developed tests meet the same safety and effectiveness standards as commercially manufactured tests, intending to protect public health and standardize diagnostic testing across platforms.

I see terms of art in phrases like "compliance with medical device reporting" and "compliance with registration and listing' and "compliance with quality system regulations." Will this PDF document help small labs understand, at some level, what those opaque phrases mean?

Yes, the FDA's Small Entity Compliance Guide is specifically designed to help small laboratories understand the complex regulatory landscape associated with medical devices, including in vitro diagnostic products (IVDs). Here's how the guide aids in understanding the specific terms and requirements:

  1. Medical Device Reporting (MDR):

    • The guide explains that laboratories must comply with MDR requirements, which involve reporting to the FDA any incidents where the device may have caused or contributed to a serious injury or death. The guide likely provides links or references to detailed FDA guidelines on how to meet these reporting standards.
  2. Registration and Listing:

    • It details the need for laboratories to register with the FDA and list their devices. This section would typically outline the process of registration, the type of information needed for listing, and the annual requirements to maintain compliance. This helps ensure that the FDA has a current database of all active medical devices and diagnostics.
  3. Quality System Regulations (QSR):

    • The guide covers compliance with QSR, which includes manufacturing standards, design controls, and other procedural safeguards to ensure products are consistently produced and controlled according to quality standards. It probably provides guidance on the specific aspects of QSR relevant to small labs, such as design controls, purchasing controls, and corrective and preventive actions.

The document serves as a compliance framework, offering guidance on understanding these regulations in the context of their operations. It also likely includes links to additional resources and FDA contact information for labs that require further clarification or assistance. This approach helps demystify the legal and technical jargon that can be a barrier to compliance for smaller entities.

Brief Blog: FOIA Requests Require "Years" Again

Since I first heard of FOIA, I've heard it takes "years and years."  I've done at least a dozen FOIA requests, and some of them I got great responses in a few months, others a year.

CMS FOIA seems to be running slower now.  I got an email this week from CMS about my 2021 FOIA request, saying it is still in queue.  They ask me to confirm if it's still of interest, to help them clear their backfiles.  I noticed the letter remarks that they are taking their thousands of back-requests on a strict first in first out basis.

See my earlier blog on user-friendly FOIA tips here.  CMS letter, click to enlarge.

click to enlarge


One thing I learned over time, it was usually better to route a FOIA request - even for MAC documents - via CMS FOIA, because they would log it, track it, pursue it, and get back to me.  A denial from CMA FOIA, if it occurred, came packaged with precise appeal instructions, which I've used successfully.

In contrast, in my experience, sending direct-to-MAC FOIA requests could be very hit and miss, and often would get oblique answers and no guidance to appeals rights.

Sunday, June 23, 2024

Brief Blog: MolDx Updates Instructions for Medicare Advantage Plans

On June 18, 2024, MolDx updated its instructions for Medicare Advantage plans which have, in some cases,  begun requiring "Z codes" for molecular claims processing.   See a 360DX article here.  See an April 2024 "Executive War College" presentation by Dr. Bien-Willner of MolDx (audio + PPT) here.  See a slightly earlier Dark Daily article here.

Find the  home page here:

See a 3-page table of codes falling under this program, as of June 18. It's divided in 2 tables, one for DNA-RNA tests and one for proteomics tests.$FILE/Medicare_Advantage_CPT_Codes.pdf 


Friday, June 21, 2024

A Policy-Watcher's Thoughts: Court Cases Bode Well for ACLA vs FDA [Updated]

Update (June 28):  I would suggest that the court's anti-Chevron ruling in Loper v Raimond also supports ACLA v FDA.   Loper finds that courts are not obligated to defer to an agency when a statute is ambiguous.  Rather, a court should review the statute in question de novo.   

Here, that would mean a court does not have to any longer defer to FDA whether the 1976 intention of Congress was for a laboratory of people and items conducting processes was collectively commonly understood to be "a medical device," for example, according to dictionaries or usage at the time.   


In late April, FDA filed a legal case against FDA, arguing its final regulation giving it authority over LDTs was flawed and impermissible.  (My original blog here, my ChatGPT based analysis here.)

Opinions about the ACLA-FDA case vary widely.   Some think it's a momentary flash in the pan, others think it has a good short of winning.  While I'm not an attorney, several recent cases cast a favorable light on different parts of the ACLA's action, from my ringside seat as a policy-watcher.


The first issue is whether ACLA (plus its co-plaintiff, a laboratory) have standing to sue.  In the recent case about banning mefipristone,  SCOTUS decided unanimously that the defendants (a few physicians) had no standing to sue.   But the court provides an updated analysis of all the different ways you CAN have standing to sue, and they clearly encompass the ACLA and the lab in the case of mefipristone.   Also, there are many HHS cases where the American Hospital Association (AHA) has standing to act as a plaintiff against HHS - a new one about HIPPA just this week.

So yes, it's possible for cases to fizzle out over standing, but I see this as unlikely in the ACLA FDA case.  FDA itself says its rule will costs labs $100M within 12 months and $4B within a few years, a weighty enough topic to complain about.

Texas (Where ACLA Filed)

Cases against government agencies, accusing them of incompetence or overreaching, do well in Texas federal courts.   See the new HIPPA case this week (overreach in new HIPAA regulations).  The mefipristone case won in Texas and in the overlying 5th Circuit appeals court.   There are even "court-shopping" articles about this type of case in Texas (here).

Federal Agencies Corraled by Conservative Courts

Two recent cases hinged on judges taking a magnifying glass to the wording of statues and what agencies can, and can't, do.   

  • In the bump-stock gun ruling, the SCOTUS judges took a microscope to the terminology of "machine gun" versus other types of guns, and dealt a setback to the ATF.   
  • In the HIPAA case I've just cited, Judge Pittman in Texas took a microscope to Congress's definition of what health information is, and concluded that a recent HIPAA regulation was coloring outside the lines.   
  • These types of cases (where regulatory agencies lose) support that ACLA could win in its several concerns that FDA is "coloring outside the lines" in defining interstate sale of medical devices versus a series of human-machine operations from place to place inside a laboratory.

Thursday, June 20, 2024

Brief Blog: CMS Releases Part B Cloud Database for CY2022 by Physician & CPT Code

For a decade now, CMS has been released Part B calendar-year data by physician (or lab) and CPT code. The data comes out about 18 months after the end of a calendar year - June 2024 for CY2022.

On June 4, 2024, CMS released this database for Part B CY2022.  Find it here:


The data has 10M rows, so filtering it in the cloud is key.

For example, if you want all Genomic Sequencing Procedure codes, you can filter for codes that start with 814.   (Click Filter, click "basic filtering = " HCPCS-Cd, and "starts with" 814.  Push "apply filters."  You get 468 rows of data, which you can 'export' in a downloadable spreadsheet for Excel.

Or if you want PLA codes, click Filter, "basic filtering=" HCPCS-Cd again, but now filter for "ends with" and "U".   (You get 2737 rows, most are for for pathogen panels).  

Screen Shot: PLA Codes

For PLA codes, the top 10 lines out of 2737 lines, have 75% of all PLA code payments.  A Guardant test (0242U) and a Foundation Medicine test (0037U) have 25% and 19% of PLA payments, respectively. (Below, click to enlarge). (Top 10, $254M of $327M).

click to enlarge

Screen Shot: The Death of 81408

I began reporting several years ago that in Novitas/FCSO states, the $2000 genomic code 81408 was treated like a magic ATM machine, rapidly skyrocketing to the largest or near-largest codes in genomics.   It was literally a billion dollars, and many labs that billed this code, billed it always in multiples of 2 and always in full panels of all the rare gene codes (81401-81408). 

That's gone.   Only a few dozen payments for 81408 in Medicare in 2022, and only 3 data lines.

81408 2022

See the June 2023 OIG report on 81408 fraud here.   
See my 2021 blog on the severity of the situation, when 2019 data was released - I coined the term, "fraudomatic."  
See my June 2023 update.

So 81408 rose out of nowhere (in 2017) to be a super high paying code in 2019, 2020, and 2021.  By 2022 it was nowhere to be found, like it started.


Screen Shot:  Usage of 81479 "Other Genomic Test"

This code totaled $470M of usage, nearly all of it in MolDx states.  The top 3 billers had 61% of 81479 payments, and the top 10 had 90% of 81479 payments.  Click to enlarge.

81479 click to enlarge


Timeline for This Database

The data for CY2022 was released in June 2024.  I believe CMS allows 12 months for the claims to come in (last date of service 12/31/2022, so last claim deadline 12/31/2023).   After that "12 months to file' date passes, the data is released 5-6 months later.  Much briefer national summary data for CY2022 was released in simple excel spreadsheets around 11/2023.

Keeping Up With MACs: Noridian Medical Director List; Palmetto & MolDx Lists

This morning I ran across a detailed and up-to-date list of medical directors at Noridian, with brief bios.

See a June 10 2024 listing for Palmetto, but not including MolDx medical directors such as Dr Bien-Willner.

There is a listing of MolDx medical directors, but not entirely up to date:


In August 2023, Dr Anitra Graves moved from Noridian to Novitas/FCSO.

Wednesday, June 19, 2024

Nerd Note: JAMA Article Highlights the "AUPRC" Statistic

 I'm highlighting the bottom section of a blog this week on the big-data-driven PHOENIX criteria for pediatric sepsis.   I'm posting the novel statistic here, so it can be found via a scan of blog titles.


For the original blog on big data and sepsis, here:

I rarely mention statistics, but this popular and important JAMA paper hinges on a statistic called "AUPRC" which was new and confusing to me.  Here's my short text about it, from my own "Phoenix" blog.


Nerd Note - Novel Statistics Right Here in JAMA

The Phoenix paper makes much use of the AUPRC metric - area under the precision recall score, which is an alternative to the far, far more common "AUROC" or area under the receiver operating curve.  I'm not sure I've got my head around this yet.   
  • Whereas the virtue of the AUROC is that it looks at clean simple sensitivity and specificity as if in a vacuum, and AUROC deliberately leaves base rate and PPV-NPV for a later day, the AUPRC builds PPV into itself, so it's only valid for a population with a similar base rate.
Read about AUPRC in the Phoenix paper, and for more about AUPRC, they cite a "super paper," Saito & Rehmsmeier 2015, in PLOS One. Saito has nearly 3000 citations.  

Beyond reading Saito, I will leave a public link to a long Chat GPT dialog where I tried to teach myself more about PRC - here.  

You can also take yourself to a YouTube video - here.

Tuesday, June 18, 2024

New World for Alzheimer Diagnostics: (1) Davos Initative and (2) Blood Biomarker Consensus Paper

This week, the Davos Alzheimer Collaborative (DAC) issued a press release about a major new blood biomarker initative, and held a 90 minute webinar on the topic on June 18.   

Second, a global forum released an important position paper on performance standards for blood-based testing, as Schindler et al., Nature Reviews Neurology.  

As an experiment, I've included an AI-generated article about the two events below.  (I'm familiar with both events, and I have proofed the AI essay for accuracy).



Advancements in Alzheimer’s Diagnostics: 

Insights from Davos Alzheimer Collaborative Webinar and Schindler et al. Position Paper

A recent webinar hosted by the Davos Alzheimer Collaborative, alongside the new academic insights provided by Schindler et al. in their paper, "Global Standards for Alzheimer's Blood Diagnostics," published in Nature Reviews Neurology,  aim to support blood-based biomarkers as a new standard for Alzheimer diagnostics.

Davos Alzheimer Collaborative (DAC) Webinar: The webinar showcased a collaborative effort among international experts to discuss the integration of new diagnostic technologies into healthcare systems. Here are the key topics covered:

  1. International Cooperation: The need for a coordinated global response to Alzheimer’s diagnostics was highlighted, emphasizing the importance of sharing data and insights across borders.
  2. Technology Integration: Discussions focused on how to effectively incorporate new diagnostic tools into existing clinical frameworks, ensuring that these innovations are both scalable and practical. These issues include EHR integration.
  3. Barriers to Clinical Implementation: The webinar explored real-world challenges in the adoption of new diagnostics, including regulatory hurdles, lack of reimbursement, the need for provider education, and system readiness.
  4. Future Implications: The potential for new diagnostic tools to improve early detection and management of Alzheimer’s disease was discussed, underscoring the importance of continued innovation and research.

Schindler et al., Nature Reviews Neurology Position Paper: The position paper sets forth proposed standards for blood-based Alzheimer’s diagnostics, emphasizing the importance of validating these tools in clinical settings.  It's titled: "Acceptable performance of blood biomarker tests of amyloid pathology — recommendations from the Global CEO Initiative on Alzheimer’s Disease." PMID 38866966.  Press here.   

Key points include:

  1. Performance Standards: Schindler et al. propose specific sensitivity and specificity thresholds that blood biomarkers should meet to be considered effective for clinical use.
  2. Use in Primary Care: The paper advocates for the use of blood biomarkers as a preliminary screening tool in primary care settings, which could streamline the diagnostic process by identifying patients who may require more extensive testing.
  3. Alignment with Existing Methods: The recommended standards aim to align blood biomarkers with the accuracy of established Alzheimer’s diagnostics such as amyloid PET scans and cerebrospinal fluid analysis.
  4. Practical Implications: The paper discusses the logistical considerations of implementing these standards in everyday clinical practice, including the need for robust validation studies and education for healthcare providers on the interpretation of test results.
Schindler et al. is paired with a policy article by Mielke and Fowler, here.  See Sidebar below.


Both the Davos webinar and the Schindler et al. position paper address critical aspects of introducing and integrating blood-based biomarkers for Alzheimer’s diagnostics. Their combined efforts underscore the importance of setting high standards for diagnostic accuracy, enhancing the feasibility of these tests in routine healthcare, and ensuring that advancements in Alzheimer’s diagnostics are both scientifically sound and practically applicable. The focus remains on making Alzheimer's diagnostics more accessible, thereby potentially improving outcomes for patients by facilitating earlier and more accurate diagnoses.


Key Takeaways from Davos Webinar and Schindler et al. Position Paper

Davos Alzheimer Collaborative Webinar:

  1. Global Collaboration: Emphasis on international cooperation to tackle Alzheimer's diagnosis challenges.
  2. Technology Integration: Discussion on integrating new diagnostic tools into clinical practice to ensure wide accessibility.
  3. Expert Discussions: Panel of global experts highlighted ongoing efforts and future directions in Alzheimer's research.
  4. Clinical Implementation: Focus on real-world implementation of blood biomarkers and the barriers to integration into healthcare systems.
  5. Educational Initiatives: Highlighted the need for educating healthcare providers on the use of new diagnostic tools.
  6. Patient-Centric Approaches: Discussions on making diagnostic processes more patient-friendly and less invasive.
  7. Research and Insights: Sharing of latest research findings and data from various studies on Alzheimer’s diagnostics.
  8. Future Outlook: Optimism about the potential of new diagnostic tools to significantly alter the management of Alzheimer’s disease.

Schindler et al., Nature Reviews Neurology Position Paper:

  1. Standardization: Introduction of global standards for the performance of blood-based Alzheimer's diagnostics.
  2. High Sensitivity and Specificity: Recommended benchmarks of ≥90% sensitivity and ≥85% specificity for these tests.
  3. Clinical Settings: Emphasis on the use of blood biomarkers in primary care as a triage step before more invasive tests.
  4. Best Practices Alignment: Aligns new blood test standards with existing diagnostic methods like amyloid PET and CSF tests.
  5. Accessibility and Cost-effectiveness: Focus on making Alzheimer's diagnostics accessible and affordable for widespread use.
  6. Early Detection: Potential for blood biomarkers to enable earlier detection of Alzheimer’s pathology.
  7. Patient-Friendly Testing: Advocating for non-invasive tests to enhance patient compliance and comfort.
  8. Impact on Treatment: Improved diagnostics could lead to earlier and potentially more effective treatment interventions.

Mielke et al., Nature Reviews Neurology Position Paper:

  1. Comprehensive Overview: Mielke et al. provide a thorough discussion on the use of blood-based biomarkers (BBMs) for Alzheimer's disease (AD), including current clinical applications and considerations for their use at the population level.

  2. Diagnostic Accuracy: The paper emphasizes the potential of BBMs to identify AD pathology with high specificity and sensitivity, noting their utility in primary care settings for initial screening.

  3. Impact of Comorbidities: Highlighted is the significant influence of chronic conditions like kidney disease, obesity, and cardiovascular diseases on BBM levels, which can complicate the diagnosis of AD.

  4. Ethical and Social Considerations: Mielke et al. discuss the ethical, social, and practical challenges in implementing BBMs, including issues of patient consent, medical record confidentiality, and potential discrimination.

  5. Technological Advances: The review notes recent advancements in BBM technologies that improve their diagnostic accuracy, such as ultra-sensitive immunoassays and mass spectrometry.

  6. Real-World Application: There's a focus on the necessity for real-world testing of BBMs to ensure they perform well in diverse, general populations, not just controlled research settings.

  7. Policy and Guidelines: The paper calls for the development of robust policies and guidelines to manage the implementation of BBMs, ensuring they complement existing diagnostic processes without replacing clinical judgment.

  8. Future Research Directions: It underscores the need for ongoing research to address unanswered questions about the long-term effectiveness and impact of BBMs on AD diagnosis and management at the population level.

These insights from both the Davos webinar and the Schindler et al. position paper highlight significant advancements and the collaborative effort needed to improve access to Alzheimer’s diagnostics through innovative, accessible, and effective policies, proceesses, and tools.   ( / End ChatGPT 4 )

Monday, June 17, 2024

Sepsis: A Rational Definition Founded on Big Data


The diagnostic criteria for sepsis in adults have been in flux since the early 1990s, and are still a matter of debate.    Criteria have evolved via expert panels and consensus.   Newly, an international group has used "big data" from tens of thousands of cases to develop diagnostic criteria for pediatric sepsis.   It's called the "Phoenix" project and appeared recently in JAMA.

This blog includes, at bottom, a bit about the "AUPRC" stat used in Phoenix.


JAMA has a home page for collected sepsis articles and op eds (here).  The most recent international consensus document was "Sepsis-3," issued in 2016 by the "Third International Concensus Definitions" group, Singer et al. JAMA 315:801.  It was designed by a 19-member task force from the US Society of Critical Care Medicine and the European Society of Intensive Care Medicine.   They include a summary of prior work, such as "Sepsis-1" in 1991 and "Sepsis-2" in 2001.

There is a lot of spread between the Sepsis-2 and Sepsis-3 definitions (see Engoren 2020 here and Vermassen 2021 here.)    Debate spills over into related areas, like the CMS SEP-1 quality measure which was recently elevated to a financially impactful status under Medicare despite criticisms of the measure (IDSA here).

  • From a different field, see the colorful and stormy history of "consensus definitions" in the psychiatry handbook DSM; find the new book by Horwitz here.

Big Data Informs Pediatric Sepsis Definition

I was struck by the February 2024 paper by Sanchez-Pinto et al., in JAMA, "development and validation of Phoenix Criteria for pediatric sepsis and shock."   Here.   The data analysis project spanned 5 countries and used literally hundreds of thousands of patient records.   The resulting model out-performed prior 2005 international pediatric consensus criteria.   

The extremely large body of empirical data ["A," Phoenix study] then informed a separate pediatric consensus criteria process ("B," see Schlapbach et al., here) to adopt the Phoenix result.   These authors write: 

  • Prior pediatric sepsis criteria that used systemic inflammatory response syndrome (SIRS) criteria, which have poor predictive properties, and included the redundant term, severe sepsis. 
  • The SCCM task force recommends that sepsis in children be identified by a Phoenix Sepsis Score...

JAMA includes an op ed overview of the whole process by Carlton et al. here.

Nerd Note - Novel Statistics Right Here in JAMA

The Phoenix paper makes much use of the AUPRC metric - area under the precision recall score, which is an alternative to the far, far more common "AUROC" or area under the receiver operating curve.  I'm not sure I've got my head around this yet.   
  • Whereas the virtue of the AUROC is that it looks at clean simple sensitivity and specificity as if in a vacuum, and AUROC deliberately leaves base rate and PPV-NPV for a later day, the AUPRC builds PPV into itself, so it's only valid for a population with a similar base rate.
Read about AUPRC in the Phoenix paper, and they cite to a "super paper," Saito & Rehmsmeier 2015, in PLOS One, with some 2000 citations, for further understanding.

Beyond reading Saito & Rehmsmeier, I will leave a public link to a long Chat GPT dialog where I tried to teach myself more about PRC - here.  

You can also take yourself to a YouTube video - here.



Economics Corner: How Biosimilars Have Changed Biotech Pricing

When the FDA introduced ground rules for approving biosimilars, there was much debate over  how much impact there would be on biotech drug prices.   

Now that a number of years have passed, a team from Stanford, Penn, and USC has just brought together some important data.   

Historic data suggests that drug prices drop when several generic entrants are competing against a costly originator drug.  (With small molecules, like Prozac, multiple generic makers usually do enter the market).   New work confirms that the same seems to apply for trastuzumab (Herceptin), based on a 2023 paper by Horn et al.  

  • The first biosimilar was priced 15% below Herceptin (e.g. "shadow pricing" which avoids triggering a price war).  
  • The fifth biosimilar was 58% below the originator brand.   
  • In addition, the originator price fell 29% between 2019 and 2022.   
  • Biosimilar market shared (with 3 biosimilars) reached 32%.   

Horn et al. (2023)  Biosimilar competition and payments in Medicare: The case of trastuzumab.   JCO Oncol Pract 19:e476-83.

See a similar article assessing similar public data, from Chen et al. in Health Affairs (PMID 37276475).


FDA has a webpage on generic drug competition and pricing, including a graph of price vs number of competitors. 


Of related interest, a 2023 Health Affairs article on innovative payment models for ultra high cost drugs (Horrow & Kesselheim).   And similarly, Zhang & Shugarman write about value based "financing" for cell and gene therapies in J Med Econ 2024.


AI Corner

I asked Chat GPT to illustrate Zhang & Shugarman and it came up with this pair.

See a new short history of trastuzumab in Nature.


Friday, June 14, 2024

Tempus Up in First-Day Trading; GRAIL Steady

TEMPUS and GRAIL both became publicly-traded entities this week, TEMPUS through an IPO and GRAIL as a spin-out from Illumina.

TEMPUS (TEM) was priced at $37 earlier this week, and open trading on NASDAQ began at noon Friday.   The stock bumped from $37 to  $42 before settling at $38 in mid-afternoon.  According to news sources, the transaction raised $410M.   

One source stated that while Tempus's valuation at $37 would be $6.4B, the firm's highest valuation was in 2022 at $10.2B.   Per Pitchbook, Google bought in at $8B in 2020, and Baillie Gifford holds $350M of the stock.   (Here).

Baillie Gifford, an investment firm based in Edinborough, has a detailed and interesting online article on how evolutionary fitness can be applied to technology investments.  [See also this book.]


GRAIL (GRALV) has now traded for several days in a narrow band, around $14 to $15.50, with a market cap around $450M.  Here.


In other news, Reuters carries a story about a legal case from Guardant versus Tempus relative to DNA-testing patents.


AI Corner.

Chat GPT 4 summarizes the investment article noted above.

The essay explores the concept of "fitness landscapes" from evolutionary biology and applies it to business strategy and growth investing. Key points include:

  1. Complex Systems Commonality: Human brains, cities, cells, ecosystems, and business markets are all complex systems with dynamic interactions between underlying components. They exhibit properties like nonlinearity, emergence, punctuated equilibrium, and path dependence.

  2. Fitness Landscapes: Introduced by geneticist Sewall Wright in 1932, fitness landscapes visualize genetic variations and their survival likelihood. Modern adaptations use three-dimensional models to represent these landscapes.

  3. Dancing Fitness Landscapes: Biophysicist Stuart Kauffman and biologist Sönke Johnsen extended the concept by introducing time variability, highlighting that fitness landscapes are not static but continuously shifting due to interaction dynamics.

  4. Adaptive Strategies in Business: Echoing natural evolutionary processes, businesses are encouraged to be restless, use parallel strategies, and balance short and long strategic moves to adapt and thrive in dynamic markets.

  5. Three Horizons of Growth: McKinsey’s model identifies three strategic horizons: extending existing businesses (short jumps), leveraging the core business to create new opportunities (medium jumps), and creating entirely new businesses (long jumps). Duolingo is presented as a case study, successfully spanning all three horizons.

  6. Founder's Edge: The essay suggests that founder-led companies are often more adaptable and better positioned to make bold strategic decisions necessary for long-term success.

  7. Investment Risks: The document reminds investors of the risks associated with mutual funds, including international investing risks like currency rate changes, political uncertainty, and market volatility.

The essay serves as a guide for investors to understand and navigate the complexities of investing in businesses by learning from evolutionary biology and complex system theory, advising a proactive, diversified approach to business strategy and investment.  [ / end chat gpt 4 ]

Thursday, June 13, 2024

MedPAC Issues Annual June Report: Software in Medicare and Hospital-at-Home

Each June, MedPAC releases a lengthy Report to Congress on a selected list of Medicare policy topics.  The June 2024 edition is out.

Particularly novel or interesting chapters include:

  • Paying for software technologies in Medicare
  • Medicare's Acute Hospital Care-at-Home Program
  • Medicare Advantage: Prior Auth & Narrow Networks
Find the 269-page report here:


AI Corner

I gave the 30-page chapter on software to GPT4 for a summary, as below.  The actual chapter defies such a short summary, with many interesting tables and examples of CMS pricing history and complicated, different evolving policies for software.   MedPAC is always highly in favor of bundling - they never saw a service they didn't think was benefited by bundling rather than paid separately, and this is clear in the last several pages of the software chapter.   But, at a higher level, here's a good AI summary.

This MeddPAC report on CMS payments for software technologies in Medicare is extensive and detailed, focusing on the evolving landscape of software as a medical device (SaMD) in healthcare, particularly how these technologies are integrated, regulated, and reimbursed within the Medicare system. Here is a detailed summary, highlighting dilemmas and key points relevant to diagnostic tests:


The increased reliance on software technologies in healthcare, including AI and machine learning applications, poses new challenges and opportunities for Medicare. Technologies such as clinical decision support software and computer-aided detection (CAD) systems are growing more sophisticated, requiring updated regulatory and payment frameworks.

FDA Approval and Medicare Coverage

Software technologies must receive FDA clearance or approval as SaMD before they are eligible for Medicare coverage. The FDA classifies these technologies based on risk, and there are different pathways for their approval: premarket notification (510(k)), De Novo classification, and premarket approval (PMA).

Medicare Payment Systems

Medicare covers and pays for SaMD under various systems:

  • Part A and Part B services: Coverage depends on the software fitting into a Medicare benefit category and being reasonable and necessary for diagnosis or treatment.
  • Outpatient Prospective Payment System (OPPS) and Physician Fee Schedule (PFS): SaMD can be separately payable or bundled with other services, with specific codes assigned for billing purposes.

Dilemmas in Payment and Coverage

  1. Balancing Innovation and Cost: Medicare must balance the encouragement of innovation with the need to control costs and prevent overpayment for technologies that could be bundled with other services.
  2. Service Integration: Deciding whether a new technology should be paid separately or as part of a bundled payment presents a significant challenge, especially as some software can perform functions independently of hardware devices.

Diagnostic Tests: Specific Insights

  • Software used in radiology, pathology, and genetics, such as CAD systems for diagnostic imaging and AI-driven analysis tools, is critical. These systems analyze medical images and data to assist in diagnosis and management but pose unique challenges for integration into Medicare’s payment frameworks.
  • Technologies like the LumineticsCore for detecting diabetic retinopathy and CAD systems used in radiological assessments are examples where Medicare provides coverage under specific conditions.

Policy Recommendations

The report suggests continued evaluation and adjustment of payment models to accommodate the rapid advancement of medical software technologies. This includes considering the implications of AI and machine skill learning, where software not only supports but potentially replaces human decision-making in clinical settings.

This summary addresses the critical points of integrating new medical software technologies into the Medicare framework, focusing on both the potential benefits and the complexities of ensuring effective coverage and reimbursement without stifling innovation or increasing costs disproportionately. [/end chatgpt 4]

Medicare MACs Finalized LCDs for KidneyIntelX (NGS MAC) and Thyroid Nodule Stratification (MolDx)

 Two new final LCDs have been posted on June 13, 2024.


NGS MAC posts final version of L39726, "KidneyIntelX and KidneyIntelX.dkd Testing."  The process reflects a public meeting (CAC meeting) on August 24, 2023 (17pp; see link within LCD). NGS MAC writes,

  • “The current evidence concerning KidneyIntelX or KidneyIntelX.dkd as a test to identify and stratify patients with T2D and early-stage CKD into low, intermediate, and high risk for near-term rapid progressive decline in kidney function, suggests that the early identification of high-risk patients by the test allows for more intensive patient management, selection of appropriate medications, and appropriate specialty referral or consultation.”
  • “Also, the clinical principles, that more proactive care leads to better health outcomes and improved quality of life for patients, including slowed disease progression, avoidance or delay of kidney failure and need for hemodialysis, were supported by our CAC Subject Matter Experts.”

See LCD L39726 here.  Response to comments A59758 here.  Coding article A59595 here.

The LDT form of the test is "Kidney Intel X" while the FDA-cleared form of the text has the FDA-provided suffix ".dkd".   The text is coded as 0407U, $950.   

  • The code text is, "Nephrology (diabetic chronic kidney disease [ckd]), multiplex electrochemiluminescent immunoassay (eclia) of soluble tumor necrosis factor receptor 1 (stnfr1), soluble tumor necrosis receptor 2 (stnfr2), and kidney injury molecule 1 (kim-1) combined with clinical data, plasma, algorithm reported as risk for progressive decline in kidney function."
Share price for Renalytix has been as high as $32 (2021) and is currently trading around 45 cents, for a market cap of $35M. For the fiscal year ending June 2023, per Yahoo Finance, revenue was $3.4M and operating income negative $44M.  


Palmetto and other MolDx MACs post the final version of L39646, a foundational (general) LCD for thyroid nodule risk stratification,

Find LCD L39646 here.  Response to comments A59734 here.   Coding article A59470 here.

The LCD, effective in August 2024, is based on a request letter from Veracyte dated January 12, 2022.

Updates to the coverage bullet points were minor.  The LCD states somewhat obliquely that "if the patient has multiple nodules, concurrent or reflex testing may be medically necessary."  The Q&A document clarifies that if a nodule is positive for potential malignancy, indicating surgery, then testing of additional nodules on that side are not necessary (which makes sense).


I believe that the Veracyte Afirma test had previously been covered under a general-purpose LCD ("Molecular Diagnostic Tests L35160") and a short billing and coding article which will soon be retired (A54356).   Veracyte's request letter (2022) had been to include the Afirma MTC test as well as creating a broader (open ended or foundational) LCD.  MTC is a test version for medullary thyroid cancer.  In early trading on Thursday, Veracyte (VCYT) was down 2% at $21.62.  Its five-year high was $73 in 2021, low at $15 in late 2022.




It can be difficult to predict a consensus for market demand and market response in diagnostics (e.g. the share price of Renalytix ranges from $32 to <$1 in a couple years, during which, they hit progressive benchmarks like FDA approval).   I found the book "Heart of Innovation" helpful in thinking about utility cases and market response.


AI Corner.

Here is a Chat GPT 4 summary of the 17-page NGS MAC renal transcript.

The CMS transcript presents a detailed Contractor Advisory Committee (CAC) meeting conducted by Dr. Gina Mullen on August 24, 2023, focusing on the KidneyIntelX test, an AI-enabled prognostic testing platform. This test is intended to guide care management for adults with type 2 diabetes and chronic diabetic kidney disease. Key points discussed include:

  1. Overview and Purpose: Dr. Gina Mullen outlined the meeting's purpose—to evaluate medical evidence regarding the KidneyIntelX test as outlined in CMS IOM Publication 100-08, Chapter 13, Section Renalytix introduced two versions of the test, KidneyIntelX and KidneyIntelX.dkd, with the latter distinguished by a suffix and targeted for different CKD-EPI equations used in clinical practice.

  2. Expert Panel: A panel of five experts—Dr. Maryam Gondal, Dr. Jeffrey Rimmer, Dr. Mike Ross, Dr. Charles Jacobs, and Dr. Catherine Kimball—introduced themselves and confirmed no conflicts of interest. They represent diverse institutions and bring significant expertise in nephrology and family practice.

  3. Scientific and Clinical Evaluation: The panel engaged in a structured discussion based on 12 key questions addressing the scientific basis, clinical utility, and potential impact of the KidneyIntelX test. Topics included the adequacy of published evidence, the prognostic capability of the test, and its role in improving health outcomes and clinical management practices.

  4. Evidence Review and Discussion: Highlights from the discussion include:

    • Prognostic Utility: Dr. Gondal supported the test's integration into current practice, highlighting its ability to stratify patients into low, intermediate, and high-risk categories but noted that it should not be used as the sole method for predicting disease progression.
    • Clinical Impact and Utility: Dr. Ross detailed a study indicating that the test led to increased referrals and medication adjustments, suggesting improvements in clinical management but not necessarily in direct health outcomes.
    • Integration Challenges: Concerns were raised about the test’s integration into electronic health records and its dependency on specific clinical variables, which might limit its applicability across diverse health systems.
  5. Future Directions and Considerations: There was a consensus on the need for more randomized controlled trials to conclusively demonstrate the test’s efficacy in improving health outcomes. The discussion also touched upon the test's potential to reduce socioeconomic disparities in CKD management, provided its broader integration into healthcare systems.

  6. Concluding Remarks: Dr. Mullen appreciated the expert inputs, indicating that these would be considered in future deliberations and potential guidelines.

This transcript provides a comprehensive account of expert opinions and discussions regarding the KidneyIntelX test, illustrating the complexity and interdisciplinary approach required to evaluate and potentially integrate new medical technologies into clinical practice.   [ /end ChatGPT4 ]