Monday, May 6, 2024

FDA Final LDT Rule: Federal Register, 160pp: Here's an Unlocked Copy, with AI Table of Contents

 A few days ago, the FDA released a typescript version of its Lab Developed Tests final rule.  Now, on May 6, 2024, FDA has released the final typeset Federal Register version (89 FR 37286, 160pp).

Find the web page here:

https://www.federalregister.gov/documents/2024/05/06/2024-08935/medical-devices-laboratory-developed-tests

Find the Fed Reg locked (unmarkable) version here:

https://www.govinfo.gov/content/pkg/FR-2024-05-06/pdf/2024-08935.pdf


SPECIAL FOR READERS:  UNLOCKED COPY

The Fed Reg version is locked, meaning you can't mark it up.

Here is an unlocked version so you can highlight and comment:

https://drive.google.com/file/d/1ZHsUP00Bjov40XALG2cVUV2MWtjvOGt6/view?usp=sharing


AI CORNER

Chat GPT was able to make a paginated table of contents, although the first try, it didn't get it correct.

ChatGPT
Thank you for your patience. Here is the updated and accurate paginated index for each section of the document, both in Federal Register and PDF page numbers:




Table of Contents (Federal Register Page Numbers)

Executive Summary - Starts on page 37286
Table of Abbreviations/Commonly Used Acronyms in This Document - Page 37287
Background
  A. FDA’s Current Regulatory Framework - Page 37288
  B. Need for the Rule - Page 37289
  C. Summary of Comments on the Notice of Proposed Rulemaking - Page 37290
  D. General Overview of the Final Amendment to the Definition of In Vitro Diagnostic Products - Page 37291
Legal Authority - Page 37292
Phaseout Policy
  A. Scope - Page 37293
  B. Enforcement Discretion Policies - Page 37294
  C. Stages - Page 37295

Comments on the Notice of Proposed Rulemaking and FDA Responses - Page 37296
Effective Date - Page 37433
Economic Analysis of Impacts - Page 37434
Analysis of Environmental Impact - Page 37435
Paperwork Reduction Act of 1995 - Page 37436
Federalism - Page 37436
Consultation and Coordination With Indian Tribal Governments - Page 37437
References - Page 37438

Table of Contents (PDF Page Numbers)
Executive Summary - PDF Page 1
Table of Abbreviations/Commonly Used Acronyms in This Document - PDF Page 2

Background
A. FDA’s Current Regulatory Framework - PDF Page 3
B. Need for the Rule - PDF Page 4
C. Summary of Comments on the Notice of Proposed Rulemaking - PDF Page 5
D. General Overview of the Final Amendment to the Definition of In Vitro Diagnostic Products - PDF Page 6
Legal Authority - PDF Page 7
Phaseout Policy
A. Scope - PDF Page 8
B. Enforcement Discretion Policies - PDF Page 9
C. Stages - PDF Page 10
Comments on the Notice of Proposed Rulemaking and FDA Responses - PDF Page 11
Effective Date - PDF Page 148
Economic Analysis of Impacts - PDF Page 149
Analysis of Environmental Impact - PDF Page 150
Paperwork Reduction Act of 1995 - PDF Page 151
Federalism - PDF Page 151
Consultation and Coordination With Indian Tribal Governments - PDF Page 152
References - PDF Page 153
This detailed index should now correctly align the sections with their respective starting pages in both the Federal Register and PDF formats. If you need further adjustments or additional details, feel free to ask!

Sunday, May 5, 2024

Nerdy Blog: Gapfill 2024 & Algebra-AI for Genomic Sequencing Procedures (GSP)

 A few days ago, I produced a fairly standard assessment of the 29 codes under the "Gapfill" process in 2024.  Find it here:

https://www.discoveriesinhealthpolicy.com/2024/05/cms-releases-preliminary-gapfill-rates.html

Key takeaways are that the gapfill process for new genomic sequencing procedures (GSP) will be impactful, and that extra attention will probably be paid to the first-ever liquid biopsy GSP codes.  

This is the first shake-up to the GSP coding and pricing system in a decade.  Generally, the old system was simply with 5-50 tumor genes paying about $600 and 51 or more genes paying about $2900.

I also noted that there is probably more discrepancies among the MACs than ever before.  But regardless of how low or high other MACs set gapfill prices, the only prices that matter are MolDx's, because CMS sets final prices to the median, and that is controlled by MolDx.  Still, for the six genomic sequencing procedure codes at the top of the report, the MolDx prices add to $10,762, the Novitas prices to $19,741, and the NGS MAC prices to S11,203.



###

In this new blog, we use Chat GPT to analyze the MAC prices across the GSP codes.

First, I took the GSP codes under gapfill, and put them in a single table with pricing offered by MolDx, by Novitas, and by NGS MAC.  

Next, I added columns for the key variables (features) of the new codes, copying this from an AMA CPT table.   For example, some codes include Tumor Mutation Burden, some don't.

Next, I assigned algebraic letters to each test feature, as follows.

  • A, DNA only.
  • B, DNA-RNA.
  • C, SEQ
  • D, copy number variation.
  • E, MSI.   
  • F, TMB,   
  • G, rearrangements.
  • H, liquid biopsy format.

That results in the following equations:

  • A+C+E=897 81457
  • A+C+D+E=1046 81458
  • B+C+D+E+F+G=$2990 81459
  • H+B+C+D=$1196 81462
  • H+A+C+D+E=$1345 81463
  • H+B+C+D+E+F+G=$3290 81464
click to enlarge

This is 6 equations in 8 unknowns, which does not have a general solution.  However, you can still solve for some variables, or elucidate contradictions.  Basically, what CHATGPT did is exactly the same as a kid solving A+B=10 and B=6 therefore A=4.  In the larger system just shown, ChatGPT comes to this impasse:  It appears that the system of equations is overdetermined, meaning that it has more constraints than variables to solve for. This could imply either a contradiction in the equations or the need to remove redundant equations.


Find Google Cloud Spreadsheet here [grid]:

https://docs.google.com/spreadsheets/d/16fCJh0bF92KDycLmT8ADGG63SGgg_bsdGPdJc4Z5gVY/edit?usp=sharing

Find Chat GPT Algebraic Analysis here [text]:

https://docs.google.com/document/d/1OQBA_LuKgYm3ylctjFmyGF0TNicgFn8ev9YIudZ9bR8/edit?usp=sharing

Here is a summary lifted the extensive CHATGPT4 analysis to try and parse the MolDx prices vs the AMA CPT features.  (Read, or skim, the GPT4 analysis at the prior link.)

####

GPT Says:

Having more variables than equations generally leads to an underdetermined system, which often has infinitely many solutions or might not have any. However, even in such cases, as you mentioned, it's possible to make inferences about the relationships between the variables and detect inconsistencies if they exist.

Key Learnings from the Equations Today:

Finding the Relationships:

We were able to deduce that 

𝐷 = $149 (CNV)

A + E = $149 also. (DNA anal. plus MSI)

H= $300.  (LBX)

Spotting Redundancies:

We identified that some equations were redundant, which happens when one equation can be derived from another or from a combination of others.

Highlighting Inconsistencies:

By simplifying the equations, we found that some results were contradictory, indicating an inconsistency in the given system.


For the Student: Practical Approach:

In cases where a system of linear equations has more variables than equations, it's useful to:

  • Solve for Dependencies:
  • As we did, find relationships between variables to reduce the number of unknowns.
  • Check for Redundancies:
  • Test whether some equations merely repeat information already expressed in other equations.
  • Look for Inconsistencies:
  • Simplify and combine the equations to see if contradictions arise, which can highlight any errors or issues.
  • Conclusion
    • Even without explicit solutions, you can gain insights into the relationships between variables and test for inconsistencies or redundancies, which is a valuable approach when faced with similar problems in the future.

click to enlarge

##

The equations with MolDx prices, could be replaced with NGS MAC prices, but they are generally very similar (not identical, though) to MolDx prices.   the MolDx prices could be replaced with Novitas prices, but the latter show very little "spread" among the 6 codes.

Friday, May 3, 2024

Real Humor: Pet Owners Told She Is Part Border Collie, Part Bulldog

 

See an open access article at DARK DAILY -

Pet Owner Sends Her Own Cheek Swab Samples to a Pet DNA Testing Laboratory and Gets a Report That She is Part Border Collie and Bulldog

https://www.darkdaily.com/2024/04/26/pet-owner-sends-her-own-cheek-swab-samples-to-a-pet-dna-testing-laboratory-and-gets-a-report-that-she-is-part-border-collie-and-bulldog/

Opening Extract:

In a follow-up story, investigative news team in Boston sends a reporter’s cheek swab sample to the same pet DNA testing lab: report states the reporter is part Malamute, Shar Pei, and Labrador Retriever

One pet DNA testing company returned results from human cheek swabs showing two different people were in fact part dog. The resulting local reporting calls into question the accuracy of DNA testing of our beloved furry friends and may impact the trust people have in clinical laboratory genetic testing as well...

Dalle3 via GPT4

The linked article, has several further links at the end.

Thursday, May 2, 2024

Karius Raises $100M; Genomic Diagnostics for Infectious Disease

Over the past several years, benchmarks for molecular diagnostics funding have slipped from $100M-plus (sometimes called "the COVID bubble," which also showed in flying then plunging public company valuations), to the $10M range.

Here's a $100M funding round which is worth noting.

Karius, a Bay Area lab dedicatd to genoimic diagnostics for infectious disease, raises $100M from Khosla Ventures, 5AM Ventures, and Glide Healthcare.  Existing investors like Softbank and General Catalyst also participated.

Karius writes,

The investment will enable Karius to extend its reach beyond the 400 U.S. hospitals currently using the Karius Test®, an infectious disease diagnostic test that utilizes genomic analysis and artificial intelligence to detect over 1,000 pathogens from a single blood sample. The funding enables Karius to address increasing demand from healthcare providers to expand access to the Karius Test beyond the hospital setting. Additionally, it will support research into the broader health implications of Karius’ microbial cell-free DNA technology beyond infectious diseases.

Press release here:

https://www.businesswire.com/news/home/20240502929158/en/Karius-Raises-100M-Co-Led-by-Khosla-Ventures-5AM-Ventures-and-Gilde-Healthcare-to-Expand-Access-to-Advanced-Genomic-Diagnostics-in-Infectious-Disease-Addressing-Antimicrobial-Resistance-Crisis 



New England Journal

Karius Test was also used in a current complex NEJM case study of a difficult infectious disease case.  Link here (subscription).

Wednesday, May 1, 2024

AI Corner: A ChatGPT "SWOT" Analysis of PathAI, and a Draft Biz School Case Study

Today, Quest and PathAI announced a strategic acquisition of parts of PathAI by Quest, here.

I gave ChatGPT4 a collection of recent PathAI press releases, and asked some questions.  ChatGPT4 answers below.

This must be viewed as an experiment in AI, not as a true depiction of PathAI.   We address the question, can GPT4 infer and recreate what reads like a business plan, based only on input from a handful of internet press releases.

###

Please assess these six or seven recent press releases from PATHAI.    Discuss the findings and discuss what we can learn about the strengths, weaknesses, opportunities, and [business] threats for Path AI.   Your audience is Medtech MBAs.

ChatGPT

From the recent press releases regarding PathAI, several key findings and strategic moves can be analyzed to assess the strengths, weaknesses, opportunities, and threats for the company in the realm of medical technology and AI-driven pathology. Here's a breakdown tailored for an audience of Medtech MBAs:

Strengths

Innovative AI Solutions: PathAI has developed a strong portfolio of AI-driven products such as the AISight Image Management System and PathExplore, focusing on enhancing digital pathology capabilities. These tools help in automating the pathology workflow, improving the accuracy of diagnostics, and enabling faster case reviews.

Strategic Partnerships: PathAI has established collaborations with major healthcare and technology companies including Quest Diagnostics, Roche Tissue Diagnostics, Google Cloud, and Hoag Health System. These partnerships not only enhance PathAI's technological capabilities but also expand its market reach and application spectrum in digital pathology.

Expansion in Services: The acquisition of PathAI Diagnostics by Quest Diagnostics is expected to bolster PathAI’s R&D capabilities and integrate its AI innovations into Quest’s extensive network, enhancing cancer diagnostics and digital pathology adoption.

Weaknesses

Reliance on Strategic Partners: While partnerships can be a strength, they also pose a dependency risk. Any changes in partnership terms or discontinuations could impact PathAI’s market presence and operational capabilities.

Integration Challenges: Integrating AI solutions with existing systems, like Epic’s Beaker LIS as seen with Hoag, can be complex and resource-intensive. Such integrations are crucial for seamless operations but pose significant execution risk.

Opportunities

Growing Market for Digital Pathology: The adoption of digital and AI-enabled pathology is increasing globally. PathAI can leverage this trend to expand further, especially in under-served markets and areas facing pathologist shortages.

Research and Development: The collaboration with Google Cloud to enhance the AISight platform with advanced AI tools offers significant opportunities for PathAI to lead in innovative research in drug discovery and biomarker quantification.

Regulatory Expansion: As PathAI continues to develop AI-enabled diagnostics, there is potential to explore new regulatory pathways for AI in clinical diagnostics, which could further establish the company as a leader in precision pathology.

Threats

Technological Advancements by Competitors: Rapid advancements in AI and digital pathology by other competitors could challenge PathAI’s market share and force continuous innovation to maintain its competitive edge.

Regulatory Hurdles: The AI-driven pathology solutions face stringent regulatory requirements. Changes in regulatory landscapes or delays in approvals for new products can impede PathAI’s ability to bring innovations to market swiftly.

Data Privacy and Security Concerns: As PathAI expands its use of cloud platforms and handles more sensitive patient data, ensuring robust data protection and compliance with global privacy regulations is crucial to maintaining trust and avoiding legal complications.

In conclusion, PathAI exhibits a robust strategic positioning in the digital pathology market, underscored by innovative AI applications and strong partnerships. However, the company must navigate technological competition, regulatory environments, and integration challenges to ensure sustained growth and market leadership.

Draft a treatment of a Harvard business school case study of PATHAI at this juncture.   You are  a Harvard business school professor for health industry and innovation.

ChatGPT

Title: PathAI: Navigating the Future of AI-Enabled Pathology

Overview:

This Harvard Business School case study explores PathAI, a leader in AI-powered pathology, at a critical juncture in its evolution. The company has established itself through significant partnerships, innovative technology, and strategic acquisitions, such as its recent interactions with Quest Diagnostics and Roche Tissue Diagnostics. This case assesses PathAI’s strategic decisions, challenges, and opportunities as it aims to redefine diagnostic pathology with AI integration.

Educational Objectives:

Understanding Strategic Partnerships: Analyze how PathAI's alliances with major healthcare and technology firms contribute to its growth and market penetration.

Innovation Management: Examine the role of continuous innovation in maintaining competitive advantage in the rapidly evolving MedTech industry.

Challenges of AI Integration in Healthcare: Discuss the regulatory, ethical, and operational challenges companies face when integrating AI into clinical settings.

Business Model Evolution: Evaluate how PathAI’s business model adapts to expanding its technology into global markets and different healthcare systems.

Case Synopsis:

Background: Introduction to PathAI, its founding vision, and its rise as a key player in AI-driven pathology.

Partnerships and Acquisitions: Detailed overview of strategic partnerships with Quest Diagnostics, Roche, Google Cloud, and others, including the nature of these partnerships and their strategic significance.

Technological Innovations: Exploration of PathAI’s product offerings such as AISight and PathExplore, focusing on technological capabilities and market impacts.

Market Dynamics: Analysis of the digital pathology market, including trends, growth drivers, and potential barriers.

Regulatory Landscape: Discussion on the complexities of AI in healthcare, focusing on regulatory challenges and how PathAI navigates these waters.

Discussion Questions:

How do PathAI’s strategic partnerships shape its competitive landscape and operational capabilities? What risks might these partnerships pose?

What role does innovation play in PathAI’s strategy, and how can the company maintain its edge over competitors in AI pathology?

Considering the regulatory challenges associated with AI in healthcare, what strategies should PathAI employ to manage regulatory risks?

As PathAI expands globally, what considerations should it keep in mind regarding market entry strategies and cultural differences in healthcare systems?

Teaching Plan:

Introductory Lecture: Brief on AI’s role in healthcare and PathAI's background.

Group Discussions: Students analyze PathAI’s business strategy, discuss the case study questions, and propose future strategies for PathAI.

Guest Speaker/Panel Discussion: Invite a PathAI executive or an industry expert to discuss real-world applications and challenges of AI in healthcare.

Conclusion and Wrap-Up: Synthesize the discussion points and strategic insights drawn from the case, focusing on broader implications for the healthcare industry.

Conclusion:

This case presents a comprehensive view of PathAI at a strategic inflection point, offering insights into how innovative companies in the healthcare sector can leverage technology, partnerships, and strategic acquisitions to drive growth and industry transformation. The case encourages critical thinking about the future of healthcare innovation and PathAI's role in shaping that future.



Of the 7 press releases, discuss the most and the least promising initiatives.  Justify your choices.

ChatGPT 

[NOTE: These are presented solely as examples of AI writing, not endorsed as business tactics]

Among the seven press releases reviewed, two stand out due to their strategic implications and potential impact on PathAI's future in the medical technology and pathology sectors.

Most Promising Initiative: Quest Diagnostics Collaboration

The collaboration with Quest Diagnostics appears to be the most promising initiative for PathAI. This partnership involves the acquisition of PathAI Diagnostics by Quest Diagnostics, which is a significant move for several reasons:

Increased Scale and Resources: Quest Diagnostics is a well-established player in the diagnostic information services industry. By integrating PathAI's AI and digital pathology technologies into Quest's broad network, PathAI gains access to extensive resources, expertise, and a larger customer base. This expands their market reach and potential impact significantly.

Strategic R&D Investment: The agreement includes turning PathAI Diagnostics' laboratory into Quest’s AI and digital R&D and solutions center. This positions PathAI at the forefront of innovation in pathology, leveraging Quest’s oncology and subspecialty pathology services to potentially develop new diagnostic solutions that could transform cancer care.

Licensing and Product Development: The licensing of PathAI’s AISight™ digital pathology image management system and algorithms by Quest means that these innovations will be integrated into a wide array of pathology laboratories across the United States, thereby enhancing their adoption and utility in real-world clinical settings.

This partnership not only boosts PathAI’s capabilities but also solidifies its position in the industry by associating with a leading company like Quest Diagnostics. It promises substantial growth in the adoption of AI-driven pathology solutions, thereby enhancing diagnostic accuracy and efficiency in oncology and beyond.

Least Promising Initiative: AISight Introduction for Research Use Only

The introduction of the AISight Image Management System labeled for "Research Use Only" seems to be the least immediately promising initiative compared to others that have broader commercial applications. While still significant, its impact is more limited because:

Limited Market Application: The designation "For Research Use Only" restricts the use of AISight to non-diagnostic applications, limiting its immediate commercial potential and revenue generation capabilities. While important for advancing research and development, the inability to apply these tools directly to clinical diagnostics can delay the realization of their full market potential.

Delayed Commercial Impact: Although essential for validation and further development, the focus on research applications means that the benefits of AISight in improving patient care directly are deferred until it can clear the regulatory hurdles for clinical use. This prolongs the timeline for seeing a direct impact on healthcare outcomes and financial return from this product line.

Regulatory Challenges: The transition from a research-only tool to a clinically approved system can be fraught with significant regulatory challenges and require substantial investment in further development and compliance activities. This can strain resources and divert focus from other potentially revenue-generating projects.

While the AISight system’s introduction for research use is a strategic step for future clinical applications and advancements in AI-driven pathology, its immediate impact and promise are overshadowed by initiatives like the Quest Diagnostics collaboration that have more direct paths to commercialization and market influence.

##

Products sold and marketed as research use only (RUO) may be validated under CLIA for medical use in a specific lab, after validation.  Does that change your assessment of the least favorable proposal.

ChatGPT

The clarification regarding the potential for "Research Use Only" (RUO) products to be validated under CLIA (Clinical Laboratory Improvement Amendments) for medical use in specific laboratories indeed shifts the perspective on the immediate and practical value of the AISight Image Management System. This possibility means that while the AISight system is initially restricted to research applications, it can be adapted for clinical diagnostics through validation processes in CLIA-certified labs. Here’s how this information might affect the assessment:

Revised Consideration of AISight as an Initiative:

Pathway to Clinical Application: The ability to validate AISight under CLIA for clinical use substantially enhances its potential impact. Once validated, it can be used in diagnostic settings, thus broadening its market applicability and enabling PathAI to capitalize on its technological advancements sooner than if it had to undergo the full FDA approval process for a clinical indication.

Strategic Market Entry: The CLIA pathway offers a strategic advantage for introducing new technologies into the healthcare market. It allows PathAI to gather real-world data and refine its AI algorithms within a clinical context, which can be leveraged to strengthen subsequent regulatory submissions for broader clinical use.

Revenue Opportunities: Even during the initial RUO phase, AISight can generate revenue through collaborations with research institutions and biopharma companies. The transition to clinical diagnostics can then open up more substantial and stable revenue streams from healthcare providers and pathology labs.

Enhanced Product Development: Early usage in a clinical environment, under CLIA validation, can provide critical feedback and data, accelerating improvements and adaptations that enhance the system’s effectiveness and market readiness for broader diagnostic use.

Conclusion with Revised Perspective:

With this additional pathway for clinical deployment through CLIA validation, the AISight Image Management System's initiative gains a more promising outlook. It suggests a more immediate and impactful application in the medical field than initially assessed. The capacity for CLIA-certified labs to validate and use AISight for diagnostic purposes not only mitigates some of the commercialization risks associated with the RUO designation but also positions PathAI to more rapidly influence and penetrate the clinical diagnostics market.

Given this revised understanding, the initiative involving AISight’s introduction for research use now appears significantly more favorable. It facilitates a practical and regulatory-savvy approach to market entry and expansion, enhancing PathAI’s ability to implement AI-driven innovations in pathology more broadly and effectively.

Brief Blog: Quest Acquires PathAI Diagnostics Service

According to a May 1, 2024, press release, Quest Diagnostics will acquire PathAI Diagnostics to advance AI and digital pathology in cancer diagnostics. This acquisition includes a licensing agreement for PathAI’s AISight™ digital pathology image management system. 

Per the press release, Quest aims to enhance diagnostic quality, speed, and cost efficiency by integrating PathAI's AI technologies with its oncology expertise. The collaboration will involve Quest's Memphis lab into an AI and digital R&D center, supporting the specialized pathology services while extending digital pathology benefits to a wider network.

Find the full press release here:

https://www.pathai.com/resources/quest-diagnostics-acquire-pathai-diagnostics/

##

PathAI raised $165M in May 2021.

##

See also an April press release about PathAI and Hoag Health System:

https://www.pathai.com/resources/hoag-health-system-selects-pathai-and-the-aisight-image-management-system-to-support-the-transition-to-digital-and-ai-enabled-pathology/

See also an April press release about PathAI and Google Cloud-driven drug discovery services:

https://www.pathai.com/resources/pathai-partners-with-google-cloud-to-transform-drug-discovery-and-precision-medicine-through-ai-powered-pathology/

____________

AI Corner

Press release summary by GPT4; reviewed.



CMS Releases Preliminary Gapfill Rates for CY2024; Opens Comment Period

On May 1, 2024, CMS runs the new proposed gapfill rates up the Maypole to see what people think.

The webpage for CLFS crosswalk and gapfill rates is here:

https://www.cms.gov/medicare/payment/fee-schedules/clinical-laboratory-fee-schedule-clfs/annual-public-meetings

Scroll down to the section, "CLFS Gapfill Determinations."  There you'll find a cloud file labeled with: 2042 Preliminary Gapfill Determinations.  The active file is an Excel called, "CLFS_Gapfill_prelim_for_CY25_Final.xlsb.

Files and Comments

These 29 codes are "gapfilled" because CMS could not determine a "crosswalk" price during summer 2023.

This Excel file has 2 tabs.  The first page is pricing by CLFS district, of which there are 57 (all the states plus a couple at two per state, plus DC zone, Puerto Rico zone.)   The second tab is "rationales."

For comments, CMS writes, Comments may be submitted to clfs_inquiries@cms.hhs.gov by by 07/01/2024.

Pricing Rule #1 and #2

#1 CMS sets the media price of the 57 CLFS zones;  and #2, this is always the MolDx price, since it has a clear majority of states and always controls the median.

Take Home Lesson:  Less Money to Take Home

This year there were 29 codes in gapfill.   My preliminary read is that many stakeholders may find their codes of interest are priced 20% or 30% lower than they might have expected, based on precedent codes of recent years.   I believe that this is because MolDx prices more on its "Equitable Pricing Model" set of rules (which are an unpublished business secret*] rather than older precedents.  This is not a loose paranoid guess; it's confirmed by the second "rational" tab, which states that "priced based on equitable pricing model."

Gapfillology

There are 29 codes in gapfill.  6 are Category I (80,000 series) codes.   Helpfully, this year CMS included a short descriptor column next to the code number.  One code, 0019M, is an adminstrative MAAA code.  The other 22 or so codes are PMA codes.  

5 codes are priced $2900-3500.  This includes two of the 6 Category I codes.   

7 codes are priced $890-1800, including four Category I codes in the $890-1350 range.

MAC Disparities Rise

Disparities were higher than I have ever seen.   There were NO codes where the 3 MAC systems (MolDx, Novartis, NGS MAC) agreed.   There were only 5 codes where MolDx and NGS MAC agreed (previously, NGS MAC had often matched MolDx).  

There were 15 times the MolDx price was higher than at least one of the non-MolDx prices (out of 68 comparisons).  

There were 5 cases where the Novitas prices were about $1500-2000 higher than the MolDx prices.  In these cases, the NGS MAC price was close to (but not identical to) the MolDx price.

For the six genomic sequencing procedure codes at the top of the report, the MolDx prices add to $10,762, the Novitas prices to $19,741, and the NGS MAC prices to S11,203.

click to enlarge

Here are the gapfil codes with MAC price comparisons:

click to enlarge


Alzheimer

One Alzheimer code, 0412U, C2N Precivity beta amyloid 42/40, is $339.  This is a multiplex test including ApoE and demographic facotors and an algorithm.   This is not far off from last year's price for the Fujirebio FDA CSF Amyloid test (0358U).  

But it's much higher than the Quest basic amyloid LDT (0346U).  Note that there are some Alzheimer codes at the May 2024 CPT meeting, which will enter this summer's gapfill-crosswalk public meetings in June and July  

7 Double Digit Codes

6 PLA codes and one Cat I code (87467 hep B) are priced in the double digit dollars, at $14-70.


___

 * I had a FOIA request for the EPM, which returned the reply, that EPM is an unpublished business secret.

__
Nerd note. 
CMS released the file as .xlsb, I asked ChatGPT what "b" means.  Here.


Monday, April 29, 2024

FDA LDT Rule: Quest, Labcorp Stocks Pop Up to 5%

The release of the final FDA regulation for reviewing and controlling LDT tests was released on April 29, 2024.  (Story here).

LabCorp and Quest reacted favorably.

Quest opened at $135 and popped as high as $142 before settling to a 4% gain. 

LapCorp opened at $200 and popped as high as 142 before settling to a 3% gain.

The two labs have a market cap of about $15b (Q) and $17B (L), so a 4% gain represents about $1.2B of market value across both.


___
For scale, LabCorp's range today was well inside its 365-day range of $180 to $230.

Guardant and Natera were each up several percent.  Exact Sciences floated up a couple percent but ended the day flat.  CareDx was up 8%, perhaps reflecting its experience with FDA approvals.

AI Corner: Chat GPT Discusses FDA LDT Final Rule, "Healthcare Systems" Tests for Their Own Patients

[Chat GPT; original article here.]

[The FDA exemption below is somewhat similar to a similar pathway for in-house hospital testing in the EU under the new IVD R].

[Begin Chat GPT]

The FDA's final rule regarding Laboratory Developed Tests (LDTs) specifically addresses LDTs manufactured and performed by laboratories integrated within a healthcare system. 

This policy focuses on tests designed to meet unmet needs of patients receiving care within the same healthcare system, reflecting a targeted approach to enforcement discretion. Here's a detailed summary based on the contents of the provided document:

Key Aspects of FDA's Policy on Local Healthcare System Tests

1. Enforcement Discretion

The FDA intends to exercise enforcement discretion, generally not enforcing premarket review and certain Quality System (QS) requirements for LDTs that are manufactured and performed within the same healthcare system, particularly for addressing unmet patient needs.

2. Conditions for Enforcement Discretion

This discretion applies only to LDTs where no FDA-authorized In Vitro Diagnostics (IVDs) meet specific patient needs, or where no such tests are available.

The policy does not cover all IVDs but is specific to those integrated within healthcare systems that demonstrate particular risk mitigations.

Full compliance with other applicable FDA regulations is still expected, excluding certain premarket and QS requirements.

3. Rationale for the Policy

Many healthcare systems develop unique LDTs for rare conditions or patient populations too small to be commercially viable, making the development and validation of such tests challenging.

The FDA acknowledges that the internal controls and patient feedback mechanisms within these healthcare systems can serve as effective mitigations against the risks associated with LDTs.

4. Feedback and Integration

Integrated healthcare systems often provide immediate feedback loops between laboratory professionals and clinicians, enhancing test validity and responsiveness to clinical needs.

This integration facilitates adjustments and improvements in LDTs based on direct healthcare provider inputs and patient outcomes.

5. Future Adjustments

The FDA remains open to revising this policy [through guidance not regulation] based on emerging evidence or changing circumstances, suggesting a dynamic approach to regulatory oversight.

6. Phase-out of General Enforcement Discretion

While specific premarket and QS requirements may be relaxed under this policy, the FDA plans a gradual phase-out of general enforcement discretion for these LDTs, aligning them more closely with broader FDA regulatory standards over time.

Implications for Stakeholders

This policy primarily benefits laboratories within healthcare systems that develop LDTs for specific, often underserved, patient needs. It allows these laboratories to focus on innovation and patient care without the immediate burden of stringent FDA premarket reviews. However, it also places responsibility on these systems to maintain high standards of internal review and patient safety monitoring, as they must still comply with other FDA regulations not covered by the enforcement discretion.

This summary outlines how the FDA's final rule aims to balance patient safety with the practical needs of healthcare systems, fostering an environment that encourages the development of necessary diagnostic tests while ensuring they are safe and effective.


User

Do they discuss or define what it means to be truly 'receiving care from' the same healthcare system?   I assume just getting a blood draw and test isn't a "patient" who gets care from "the healthcare system."  Does FDA get into that?  

ChatGPT

The FDA's approach to defining "receiving care from" the same healthcare system in the context of enforcing discretion for Laboratory Developed Tests (LDTs) is quite specific. The policy generally applies to situations where both the laboratory producing the LDT and the healthcare providers using the test are part of the same healthcare system. This ensures that there is an integrated approach to patient care, where the laboratory services are directly aligned with the clinical services provided to the patient.

Key Elements of the Definition:

Integration and Direct Care Relationship: The policy specifies that the laboratory and the healthcare providers must be part of the same healthcare system. This implies a level of integration where the laboratory services are not just adjunct but are an integral part of the patient care workflow.

Shared Healthcare Infrastructure: Typically, this means that the hospital or healthcare facilities and the laboratory share common governance, oversight, and a unified medical records system, which facilitates seamless communication and data sharing between laboratory professionals and healthcare providers.

The FDA emphasizes that this kind of integrated care setting enhances the quality and safety of LDTs by ensuring that there is immediate and effective communication about patient responses and test results, which can be promptly integrated into patient management strategies.

Not Covered:

Casual or One-Off Services: Merely getting a blood draw or a diagnostic test from a facility does not constitute receiving care from that healthcare system under this policy. The patient needs to be engaged in a broader, continuous healthcare relationship within the system.

FDA's Rationale:

The rationale behind this specific definition is to ensure that LDTs are used in environments where they can be most effectively monitored and where their impact on patient care can be directly observed and managed. This integration allows for better quality control, immediate response to any issues with test performance, and a more personalized approach to patient care based on comprehensive data integration and analysis.

The FDA has structured this policy to encourage and support the development of LDTs in a way that directly supports integrated patient care systems, recognizing that such systems are uniquely positioned to develop and utilize LDTs effectively due to their cohesive operational structure and shared focus on patient outcomes.

[End ChatGPT]

Bruce notes - See page 7.  The test must be ordered by providers who are credentialed or employed by the same health system as OWNS the laboratory   If their hospital is affiliated, but under separate ownership, the tests are not exempt.  

The tests are described as LDTs "manufactured" by a laboratory witin the healthcare system [eg ownership].  

Bruce notes, there could also be equity issues as wealthier patients, including children, might afford to travel to a health system offering the test.   Simply fedexing the sample to the AMC would not make the poor and geographically distanced child a patient.  A workaround might be having the test ordered pro forma by an on-staff clinician as a consult to the remote clinician and child, since the FDA seems to focus on the staffing or employment status of the physician plus whether such a physician 'ordered' the test per se.

AI Corner: Chat GPT Discusses FDA LDT Rule With Regard to New York State Review

In a final rule for regulation of LDTs, on April 29, 2024, FDA announced it would not require FDA review for LDTs reviewed and approved by New York State CLEP.

See the whole 175,000 word rule for yourself, here.

Below, find a Chat GPT summary of key pages discussed NYS CLEP below.  (Pages 43-49,302-307).  Remember this is an AI-driven interpretation about what the FDA itself writes, it is not a critique of the FDA's position.

[start Chat GPT]

###

The New York State Department of Health Clinical Laboratory Evaluation Program (NYS CLEP) plays a significant role in the regulation and oversight of Laboratory Developed Tests (LDTs). The FDA acknowledges that NYS CLEP has established a robust system for evaluating the analytical and clinical validity of high and moderate risk LDTs. This system is designed to mitigate the risk of harm from inaccurate and unreliable LDTs. Based on this, the FDA intends to exercise enforcement discretion, meaning that it will not enforce certain premarket review requirements for LDTs that have either been fully approved by NYS CLEP or received conditional approval.

Key Aspects of NYS CLEP's Oversight:

High Risk LDTs: Require full technical review and approval before they can be used on specimens from New York State.

Moderate Risk LDTs: Also require full technical review and may receive conditional approval, allowing the test to be offered while the review is still pending.

Review Process: Includes an evaluation of the LDT's analytical and clinical validity, ensuring that the test is both accurate and reliable.

Conditional Approval: Allows tests to be used while under review. If a test ultimately does not receive full approval, it is removed from the enforcement discretion policy.

Exemptions: Laboratories with proven method validation can request exemptions from full technical documentation for future tests using the same methodology.

FDA's Enforcement Discretion:

The FDA's decision to not enforce premarket review requirements for NYS CLEP-approved LDTs allows the FDA to allocate its resources more effectively to other oversight activities. This policy applies regardless of whether the specimen comes from New York State or elsewhere, but only to the version of the LDT that is approved by NYS CLEP.

Impact and Implications:

Efficiency: By recognizing NYS CLEP’s review process, the FDA avoids duplicative efforts and reduces the regulatory burden on laboratories.

Public Health: This policy is expected to maintain high standards for LDT safety and effectiveness, as NYS CLEP’s review process aligns closely with the FDA’s standards.

Phase-out of General Enforcement Discretion: While premarket review requirements may be relaxed for NYS CLEP-approved LDTs, other FDA requirements such as adverse event reporting, registration, and listing will still be enforced according to the phased timeline established by the FDA.

Concerns and Considerations:

Some stakeholders have expressed concerns that leveraging NYS CLEP's program should not replace FDA regulation entirely. They emphasize that while NYS CLEP’s oversight is robust, FDA’s regulatory framework includes broader and potentially more stringent standards. The FDA acknowledges these concerns and clarifies that this policy is a form of enforcement discretion, not a complete handover of regulatory authority.

In summary, the FDA’s approach to utilizing NYS CLEP’s established programs for certain LDTs reflects a strategic use of regulatory resources to ensure safety and effectiveness while minimizing unnecessary duplication of regulatory efforts. This decision underscores the FDA's commitment to a risk-based regulatory framework that protects public health without stifling innovation in laboratory testing.

[end Chat GPT]


FDA Publishes Final Rule for LDT Regulation

On Monday, April 29, 2024, the FDA released the final rule for its regulation of LDTs.   The rule had been released in proposed form for public comment last fall.

FDA continues to plan an enforcement deadline of 3.5 years from now for high risk LDTs and 4 years from now for moderate risk LDTs.  However, as FDA had proposed, submitting a PMA (or  510k and de novo) application will "stay the clock" for the duration of its review.  

In thinking about 510K's and PMA's, remember #1 the FDA has promised to downclassify many PMA tests to 510k and #2 has early plans to completely revamp and speed the 510K process.  Consider also wider use of predetermined change plans.

FDA reserves the right to march in faster on "problematic" tests (the word problematic appears 46 times in the final rule, 6 in the draft rule).   

FDA creates several new flexibilities for itself:

  • NYS CLEP: Rule does not generally apply to NYS CLEP approved tests;
    • The term CLEP appears 145 times in the final rule, 1 time in the proposed rule.
    • See ChatGPT Sidebar here.
  • WITHIN SYSTEM: Generally not applying to LDTs "by a laboratory within a healthcare system to meet an unmet need of patients receiving care in the same healthcare system."   
    • (This is similar to a pathway in the new European IVDR system).
    • The term "same healthcare system" occurs 42 times in the final rule, 0 times in the proposed rule. 
    • See ChatGPT Sidebar here
  • GRANDFATHER: Generally enforcement discretion for LDTs marketed prior to the rule, and not altered.
  • VA-DOD: Rule does not apply to V.A. or DOD;
All of the above points are not part of the regulation; they are part of the free-flowing and shape-shifting "discretion" that FDA may use or revisit as leadership changes, without returning to rule-making.  For example, the regulatory change (several words long) says it is "effective" 60 days after publication; that follows a federal law for regulations.  But FDA says that its grandfather date is THE DAY of publication, and chooses that at is own fancy.

See ACLA comment here. ARUP here.  Congr. Rodgers (R) here.  Sen. Cassidy (R) here. AdvaMed here (wait for VALID!). Biopharma Dive here and here.  Endpoints here.  Deep dive day 2 article at Genomeweb here.  See a listing of more articles from FDA GROUP here.  Kevin Nichols here.   Deep dive from Hogan Lovells here.

###
Massive Down-Shift in Expected Costs

In the Fall 2023 projection, FDA predicted costs to industry to be $40-50B dollars in the first five years (reflecting a impossible workload in man-years).   

In the Spring 2024 final project, FDA expects costs no higher than $500M over the first 3 years, and then onward costs of $1.5B per year.   That works out to about $5B not $50B in the first five years.  Using my own ballpark figure of $200,000 per regulatory professional (fully loaded overhead), that is still 25,000 man-years of regulatory time in the first five years (down from 250,000 man-years in the proposal).




click to enlarge (shows fall table 35, spring table 40, regulatory impact document)

Benefits are still calculated in value per statistical life year (VSLY) (new regulatory impact, page 68ff).  That is, costs continue to be real costs like professional labor and benefits continue to be  projected life-years added or projected deaths delayed by IVDs.

###
Final Rule Six Times Bigger

The inspection (typescript) copy of the proposed rule was 83 pages; the inspection copy of the final rule is 528 pages, or 6X.  The 528 pages add up to 175,000 words.  

###

Regulatory Change Tiny; Satellite Rules All FDA Discretion

Most of the important stuff is still "FDA discretion," the actual rule change is only a few words long.   Everything else (like pivotal proposals like grandfathering) relates to FDA internal plans and/or FDA discretion as time goes by.  (E.g. FDA remarks on its ability to change the rules by guidance, p 42.) 

###

WHERE TO FIND IT

See the FDA webpage for the rule's release here:

https://www.fda.gov/news-events/press-announcements/fda-takes-action-aimed-helping-ensure-safety-and-effectiveness-laboratory-developed-tests?utm_medium=email&utm_source=govdelivery

See the Fed Reg location for the final rule here. The typeset rule will appear May 6.

https://www.federalregister.gov/public-inspection/2024-08935/medical-devices-laboratory-developed-tests

See the typescript or pre-publication version of the actual final rule here--528 pages.

https://public-inspection.federalregister.gov/2024-08935.pdf

See the webpage for regulatory impact analysis here:

https://www.fda.gov/about-fda/economic-impact-analyses-fda-regulations/laboratory-developed-tests-regulatory-impact-analysis-final-rule

See the 178-page final impact analysis here:

https://www.fda.gov/media/178133/download?attachment

##

See a new guidance document for "public health response" issues absent an actual declared emergency:

https://www.fda.gov/regulatory-information/search-fda-guidance-documents/enforcement-policy-certain-in-vitro-diagnostic-devices-immediate-public-health-response-absence

See a new guidance document for declared public health emergencies:

https://www.fda.gov/regulatory-information/search-fda-guidance-documents/consideration-enforcement-policies-tests-during-section-564-declared-emergency


##
More About NYS CLEP

The final rule mentions CLEP 145 times, vs. just once or twice in the proposal.  In my white paper on the proposed rule, I suggested that we should just require CLEP-like reviews - not 510(k) reviews by NYS CLEP, but the CLEP review itself.   ("I think the [FDA"s] accuracy concern could be met by less drastic measures, such as having something roughly the equivalent of New York State review of LDTs.")

FDA notes (FN29) that certain tests exempted from NYS CLEP review are NOT exempted from this FDA rule.

FDA discusses CLEP processes at p.43 ff.   Tests approved under CLEP are still subject to various listing and reporting rules, just like 510K/PMA tests  (p. 47), including labeling requirements [claims].  

AMCs are discussed at page 50ff.  See ChatGPT Sidebar here.

CLEP is discussed again, in the context of response to comments, at p.302ff.

FDA discusses phaseout policy at section V ("5") (p30ff), V.B.2 including CLEP (p. 41ff).

See a Chat GPT sidebar about the FDA's CLEP pages, here.

 



Wednesday, April 24, 2024

AI Corner: SCITE.AI Auto-Generates Short Medical Articles with Clear Citations

A year ago, the bad news around any AI-generated writing was hallucinations (such as made-up bibliography.)   The field moves fast and some sources seem to have completely licked that problem.

I sometime use SCITE.AI (subscription) which has two main modes.  

In the first mode for SCITE, you search or input a key paper ("key paper.")   SCITE instantly gives you all the papers that have cited it, AND, if full text is available, the paragraph or sentences that cite the key paper.   If the key paper has been cited 15 times, at a glance you can see how the key paper was cited, rather than you go and download and/or buy all 15 of the PDFs and read them one by one.

But here's a second mode for SCITE.   It's called SCITE Assistant.  https://scite.ai/assistant   You can give it a targeted question or request, and it instantly provides a correctly footnoted essay in response.

For example, I gave it this question:

  • "How important is minimal residual disease molecular testing today, in oncology?"

Here is the automated answer:

[START SCITE AI WRITING]

Minimal residual disease (MRD) molecular testing is a crucial aspect of oncology today. MRD refers to the small number of cancer cells that persist in a patient after treatment. It is a significant prognostic factor in various types of cancer, including acute lymphoblastic leukemia (ALL) (Tierens et al., 2021; Pui et al., 2015; Brüggemann & Kotrová, 2017). Detecting MRD is essential for identifying patients at high risk of relapse and shorter survival, aiding in refining risk stratification and treatment response assessment (Ravandi et al., 2018; Pui et al., 2015). 

Advancements in molecular testing techniques such as real-time quantitative polymerase chain reaction, digital polymerase chain reaction, and next-generation sequencing have enabled the detection of MRD at levels far below what is required for morphological remission (Ball & Stein, 2019). These techniques have revolutionized the field by providing highly sensitive methods to monitor MRD and predict the risk of relapse in cancer patients (Johnston et al., 2023). 

MRD testing is not limited to hematologic malignancies but extends to solid tumors as well. Studies have shown that assessing MRD in solid tumors post-surgery or treatment can guide clinical decisions related to adjuvant therapy, emphasizing the importance of MRD detection in improving patient outcomes (Larribere & Martens, 2021). 

Furthermore, MRD monitoring is essential in acute myeloid leukemia (AML), where it serves as a critical biomarker for prognostic, predictive, monitoring, and efficacy-response assessments (Heuser et al., 2021). The ability to quantify MRD using various molecular techniques, including whole-genome sequencing and immunoglobulin clonality assays, has become integral in tailoring treatment strategies for patients with leukemia and other cancers (Subhash et al., 2021; Lee et al., 2022; Fürstenau et al., 2019).

 In conclusion, MRD molecular testing plays a vital role in oncology today by providing clinicians with valuable information to guide treatment decisions, predict relapse risk, and improve patient outcomes across a spectrum of cancers.  [END SCITE AI WRITING]

##
Then, it provides a listing of all the papers it cited, AND, for each citation, it clips several sentences to show you what facts it was pulling from that citation.

Here's the top part of the footnotes section, which explains why it chose Tierens et al. as the first reference in the AI essay above.

[SCITE.AI Continues]

Reference #1 Tierens et al

“…Measurable (minimal) residual disease (MRD) is an established, key prognostic factor in adult B-cell acute lymphoblastic leukemia (B-ALL), and testing for MRD is known to be an important tool to help guide treatment decisions.…”

[click] See full context
Section quoted: Abstract

Paper:
Consensus Recommendations for MRD Testing in Adult B-Cell Acute Lymphoblastic Leukemia in Ontario

Anne Tierens1, 
Tracy L. Stockley2, 
Clinton J. V. Campbell3 
et al. 2021
Journal: Current Oncology

8       0  7  0
The paper (Tierens) has been cited 8 times.   Of these, 7 have full text.    Of the 7 with full text, 0 "support" the Tierens, 7 "mention" Tierens, 0 "disagree" with Tierens.

[Continues bibliography #2-#11]   [End SCITE writing.]

####
I've also put the full MRD AI Essay, and the full listing of 11 annotated citations, in an open access cloud PDF file here:


##

  • I first mentioned SCITE in August 2023.
  • See JAMA 2024 on rules for use of AI assistance..
  • SCITE has full-text access to all public articles, plus licenses that allow it to present short quotes even from many firewalled articles.





Two Links: on Physician Income; on Physician Industry Payments

Two sources appeared almost at the same time, regarding physicians and finances.

INCOMES

Advisory Board issues a detailed report on 2023 physician income levels.

   https://www.advisory.com/daily-briefing/2024/04/15/physician-compensation

It traces back to a 21-page report at MedScape (registration required):

   https://www.medscape.com/slideshow/2024-compensation-overview-6017073

The report has break-outs by specialty, gender, racial/ethnic group, etc.

click to enlarge

INDUSTRY DOLLARS

Over at JAMA, see a report by Sayed et al. on industry payments to US physicians by specialty, and product type.  There are also break-outs by company. The highest-dollar drug was Xarelto, the highest device was Da Vinci Surgical System.  This week in Washington Post, former NIH director Francis Collins announced he was having a radical prostatectomy (cancer took off after a period of surveillance) and with the Da Vinci system.

   https://jamanetwork.com/journals/jama/fullarticle/2816900

click to enlarge


Tuesday, April 23, 2024

Stanford Biodesign Addresses Coding vs the Innovation Ecosystem

 Innovation policy experts from Stanford's Biodesign center have published a new paper about the coding ecosystem and the innovation system and where they can come into opposition.

Find the paper here:

https://hmpi.org/2024/04/12/current-common-procedural-terminology-cpt-coding-process-challenges-impact-on-the-healthtech-innovation-ecosystem/

Here's Stanford Byers Center overview from Linked In:

“For medical innovators, the process of applying for and achieving a CAT I CPT code is a challenging and costly process that has become fundamental to unlocking reimbursement and patient access in the United States. The CAT I requirement of “widespread use,” in particular, has become a roadblock for innovators with novel products or services and the physicians that use them, that is having a negative impact on the innovation ecosystem and patient care. There is a strong and near-term need for a more transparent, predictable, and achievable CAT I CPT code process to ensure that healthcare innovation and patient access to FDA-cleared and clinically proven therapies can flourish and be preserved for future generations.”

Read our latest research on the challenges of the CPT coding process and its impact on the innovation ecosystem.

__________

AI Corner

A GPT4 summary:

The report explores significant challenges in obtaining Category I Current Procedural Terminology (CPT) codes, which are essential for medical innovators to secure reimbursement and patient access for new medical technologies in the U.S. The process is marked by stringent criteria including "widespread use" and documented clinical efficacy, which pose barriers particularly for novel technologies. Survey data from stakeholders highlighted issues such as the high costs, resource demands, and lack of transparency in the CPT code application process. Respondents advocated for a more transparent and predictable process to foster healthcare innovation and improve patient access to new therapies. Recommendations include refining the "widespread use" criteria and enhancing collaboration between innovators, physicians, and medical societies to streamline the CPT code application and approval processes.