If of interest, two instantly available uses of AI, doing tasks in an eyeblink that were impossible a couple years ago.
In the first example, it gives instructions for the cryptic symbols on a European clothes washer. In the second, I uploaded a photo of an article in an Italian newpaper about precision medicine, and it answers with an instant translation.
I posted the results on a separate blog -
https://bqwebpage.blogspot.com/2024/05/remarkable-uses-of-ai-cryptic-euro.html
On May 6, FDA published a 160-page complex new regulatory regime for lab-developed tests, to be rolled out (with many carveouts and complications) by 2027.
Now, on May 29, FDA has filed a whopping 450-page case, including supporting documents As I predicted at several conferences, it filed the case in Texas, where I think an initial ruling against the government is likely, which will start the ball rolling on various appeals and counter-attacks. The Texas filing, I think, raises the chance of an injunction while issues simmer. Grounds for this include the $101M in projected spending by labs to comply, just in the first months of the 5-year plan.
Much may depend on the next administration: if it is a change of regime they might provide a lukewarm defense, or less, of the regulation. For example, some LGBT-related regulations have gone through three literal 180 degree flip-flops across the three Obama, Trump, and Biden administrations.
Read About It
ARUP notes, as does ACLA, that the regulation makes LDTs illegal, except only for the FDA's discretionary non-prosecution.
360Dx Quotes ACLA Press Conference
360Dx quotes a press conference held by ACLA. Here's an important remark they provide from ACLA's President, Susan van Meter:
"Indeed, the fact that the FDA found it necessary to establish a complex regulatory system with multiple carve-outs only serves to emphasize that the [medical] device regime is a terrible fit for laboratory testing services."
(Begin AI)
The lawsuit filed against the FDA by the American Clinical Laboratory Association and HealthTrackRX focuses on challenging the FDA's plan to regulate laboratory-developed tests (LDTs) as medical devices. Here are the main points, key arguments, and requests for action from the plaintiffs:
Historical Regulation: LDTs have historically been regulated under a separate framework, the Clinical Laboratory Improvement Amendments (CLIA), which did not involve the FDA's device-oriented regulatory framework. The plaintiffs argue that LDTs are services performed by skilled professionals, not tangible medical devices that are manufactured and sold.
FDA Overreach: The FDA's new rule seeks to redefine LDTs as medical devices under the Federal Food, Drug, and Cosmetic Act (FDCA), a move that plaintiffs argue exceeds the FDA's statutory authority. They contend that Congress has never granted the FDA the authority to regulate LDTs in this manner.
Impact on Healthcare: Plaintiffs argue that the FDA's regulation would dramatically disrupt laboratory operations by imposing device approval requirements on tests that are currently under a different regulatory regime. This could hinder innovation, increase costs, and delay the availability of new LDTs, adversely affecting patient care.
Statutory Overreach: Plaintiffs claim the FDA's actions are contrary to law and exceed its statutory jurisdiction, arguing that LDTs are not "devices" under the definitions provided in the FDCA. They assert that these are professional services rendered using various tools, not products distributed in commerce.
Administrative Overreach: The rule is described as arbitrary and capricious under the Administrative Procedure Act. The plaintiffs argue that the FDA has not justified the reclassification with substantial evidence and that the move does not consider the actual operation and context of LDTs in clinical settings.
Impact and Practicality: The imposition of medical device regulations on LDTs is argued to be not only beyond the FDA's authority but also impractical and harmful, risking critical diagnostic services and potentially harming public health due to reduced access to essential tests.
Injunction and Vacatur: Plaintiffs request the court to vacate and set aside the FDA’s final rule. They seek a declaratory judgment that the rule is unlawful and an injunction to prevent the FDA from implementing or enforcing the rule.
Declaratory Relief: The plaintiffs are seeking a judicial declaration that the FDA does not have the authority to regulate LDTs as medical devices and that its attempt to do so via the final rule is invalid.
Judicial Review: The plaintiffs invite the court to review the FDA's rule under the standards of the Administrative Procedure Act to determine its legality and appropriateness under federal law.
In summary, the lawsuit focuses on defending the historical regulatory distinction between LDTs and medical devices, challenging the FDA’s statutory interpretation and its practical implications for healthcare services. The plaintiffs are seeking a court order to halt the FDA’s reclassification of LDTs as medical devices, arguing it is beyond the agency’s authority and would negatively impact the healthcare system. (end AI)
More About Major Questions
As to the Major Questions Doctrine, ACLA writes,
FDA’s attempt to regulate laboratory testing services as devices also implicates the major questions doctrine. See West Virginia, 597 U.S. at 724. FDA’s rule would mean that the entire clinical laboratory sector, which is a significant part of the U.S. healthcare system, has been breaking the law for nearly 50 years, and possibly much longer. And it would mean that going forward, the entire profession is operating unlawfully and can be subject to civil and criminal penalties at any time, with its only protection coming from a policy of enforcement discretion that FDA insists it is free to revoke at any time. The rule would also wreak havoc on clinical laboratories and the doctors and patients they serve, imposing billions of dollars in immediate, unnecessary costs....
An agency cannot impose massive costs and place an entire profession under its thumb in this manner without, at minimum, a clear statement from Congress. The Supreme Court has repeatedly warned that agencies should not attempt, and courts should not abide, such drastic expansions of the agency’s authority under a “‘long-extant statute’”—especially where, as here, Congress has “conspicuously and repeatedly declined to enact” such an expansion itself.
Excessive and burdensome prior authorizations are a hotspot for the AMA, have gotten attention both in Medicare regulations and in Congress, and are increasingly discussed in trade journals for the lab industry, such as 360Dx this month (Adam Bonislawsky). See also Dark Daily.
Now the New York Times weighs in with a feature article about burdens of Medicare Advantage, in a detailed article appearing over the Memorial Day Weekend. Find it here:
https://www.nytimes.com/2024/05/25/science/medicare-seniors-authorization.html
Written by Paula Span, the article got over 400 comments in the first couple days.
Here are some key points:
A new review article in JAMA on the use of surrogate biomarkers (eg by FDA in clinical trials) takes a grim position. Wallach et al. take all the surrogate biomarkers for non-cancer conditions, as listed by FDA, and perform a systematic literature review. Their main index is whether there are systematic meta-analyses of each marker, and if so, how many.
They note they exclude cancer surrogate markers (like progression-free survival) while dropping a quick comment they view that data as poor (as "weakly associated" with clinical outcomes.)
Their table 1 lists all 37 FDA surrogate biomarkers (see also their 241-page online supplement.) 17 are viewed by FDA as s strong enough for traditional approval (when not so honored, surrogates would generally be for accelerated approval.)
Their bottom line - "Most surrogate markers used as primary end points in clinical trials to support FDA approval of drugs treating nononcologic chronic diseases lacked high-strength evidence of associations with clinical outcomes from published meta-analyses."
(Most surrogate biomarkers are lab tests, though some are physiological or functional, like forced expiratory air volume or progression-free survival or blood pressure.)
According to Genomeweb, quoting San Diego Union-Tribune, Cue Health is shuttering all operations and laying off all remaining employees. Here.
It's remarkable in light of its recent past history. According to Tracxn, the company was founded in 2010 and raised $406M, including a $235M Series D in May 2021. Just months later, in September 2021, its IPO brought in $200M at a $2B valuation, per MobiHealthNews and Fierce Biotech.
From September 2021 to yesterday, share price was down from $19 to 8 cents. (It was $1 a year ago, in May 2023).
Due to COVID testing, total revenue fell from $615M in 2021, to $474M in 2022, to $64M in 2023. Pretax income peaked at $120M in 2021, falling to -$368M last year in 2023.
Tracking "What's New" on the MolDx website...
For several years, 2021-2024, MolDx staff did not regulate proteomic tests. (Think of it like FDA not regulating LDTs.) However, some labs were caught in a game of keepaway between whether they could be assessed for coverage by MolDx or by Noridian. Basically, this was like being in a corridor with an exit door on either end, one marked "Noridian" and one marked "MolDx," and both of them locked.
At the beginning of 2024, MolDx published a list - "Article A59636" - describing what proteomic tests would fall under its jurisdiction. This is "good news, bad news" - MolDx would consider the listed tests for coverage under new LCDs, but, MolDx might ensure the listed tests under its control weren't paid in the meantime, either.
On May 17, 2024, MolDx published a notice that this enforcement is delayed until January 31, 2025.
Simple (single or double protein) proteomic tests remain outside of MolDx; check the code listing in A59636 for updates. Info here.
Don't make any plans for July 15, 2024, from 2-4 ET, because you'll probably want to be at your local sports bar with your gang, watching the live telecast of this MolDx Contractor Advisory Meeting. The topic is, "Biomarker Risk Stratification Testing in DCIS."
The format is typically expert presentations and then discussion. Public questions aren't entertained. After a few months, a transcript is posted. Results may inform future LCDs. Info here.
The meeting was October 9, 2023, and this was a public comment meeting. Recordings and transcript are posted here. Transcript 26pp.
##
AI Corner
Here's an AI-generated meeting report from October 9.
Date: October 9, 2023
Duration: 1 hour, 1 minute, 57 seconds
Moderator: Dr. Angella Charnot-Katsikas
Conclusion: The meeting provided valuable insights into the current and potential uses of molecular diagnostics in managing thyroid and lung nodules, as well as metastatic prostate cancer. Speakers advocated for policy shifts to embrace less invasive diagnostic practices and highlighted the need for precise risk stratification tools in clinical settings. The discussions underscored the critical role of advanced molecular testing in improving patient outcomes through tailored treatment strategies. -- Chat GPT 4.
####
######
Sports Version
Venue: MolDx Public Comment Meeting
In an electrifying display that could rival any sports championship, the MolDx Arena was ablaze with cutting-edge discussions on the frontiers of cancer diagnostics! On October 9, 2023, the stage was set for a powerhouse lineup of scientific titans clashing over the latest in molecular diagnostics. The battleground? Thyroid, lung, and prostate cancer management, where each speaker brought their A-game, aiming for a knockout with every fact and figure.
Champion: Dr. Ashley Ross, Northwestern's Maestro of Urology In the left corner, wearing the lab coat and armed with the Decipher test, Dr. Ross dove into the fight against metastatic prostate cancer. With the precision of a seasoned fighter, she laid out how gene expression profiling isn't just a tool—it's a weapon that's revolutionizing how we attack prostate cancer. She deftly explained how this approach helps pinpoint aggressive treatments for those in dire need while protecting others from the harsh side-effects of unnecessary therapy. Her data-packed punches highlighted a survival strategy that left the audience in awe.
Contender: Dr. Ehab Billatos, Boston University’s Pulmonary Strategist - Switching gears to the lung nodule competion, Dr. Billatos entered with a flurry of stats and screens, focusing on the perilous world of indeterminate pulmonary nodules. As he navigated the court, through lung cancer screening and risk stratification, Dr. Billatos championed less invasive diagnostics that promise to be a knock-out punch for unnecessary bronchoscopies. His vision offered a glimpse into a future where lung cancer management is as strategic as a chess match.
Veteran: Dr. Paul Ladenson from Johns Hopkins. The crowd cheered for the main competition, with Dr. Ladenson stepping up to discuss the molecular testing tactics that could outmaneuver the risks of thyroid surgery. With the finesse of a veteran, he argued for the necessity of reliable molecular tests that can discern benign from malignant nodules, potentially saving thousands from the scalpel. His detailed review of molecular testing underscored a proactive defense strategy against thyroid cancer, emphasizing precision in diagnosis and treatment.
As the dust settled on this high-stakes event, it was clear that each speaker not only shared groundbreaking data but also painted a vision of a future where cancer might just meet its match. Like any top-tier sports event, the MolDx meeting left fans—er, participants—buzzing with excitement and anticipation for what's next in the playbook of cancer diagnostics. Dr. Charnot-Katsikas, moderating this epic showdown, wrapped up the proceedings with a nod to the relentless pursuit of medical innovation, a field where the goal isn't just to win, but to save lives.
The echo of the day’s discussions resonated like the cheers in a stadium: loud, hopeful, and a testament to the progress in the relentless battle against cancer. As the attendees filed out, the message was clear: in the fight against cancer, the power of molecular diagnostics is a game-changer and it's just getting started. - Chat GPT 4
As widely noted in the biotech media, TEMPUS filed for an IPO on May 20, 2024. The IPO value may be up to $100M. Prior funding rounds had raised over $1B in total.
News at Genomeweb, Bloomberg, Chicago Business, Fierce Healthcare.
Most of the articles don't link to the actual IPO document; it's online here:
https://www.sec.gov/Archives/edgar/data/1717115/000119312524142956/d221145ds1.htm
###
The Word "Reimbursement" Occurs 120 Times
Reimbursement occurs about 120 times, Medicare about 70 times.
MACs. There's substantial discussion of their "local MAC, NGS [MAC]" on PDF page 44 [paginated 40]. They recite steps towards coverage with NGS MAC (e.g. "March 25, 2021...July 29, 2021..."). They then pivot to discussing the MolDx Palmetto jurisdiction, which processes claims that come from their alternate labs in Raleigh and Atlanta. They note they submitted a tech assessment to MolDx for their xT assay in 2022, receiving approval in October 2023. They note they submitted a tech assessment for the xF assay in 2023, but are still awaiting response as of May 2024.
14 Day Rule. The same part of the S-1, PDF page 44 [paginated 40], also discusses situations where "we cannot bill Medicare directly for tests provided for Medicare beneficiaries." This hinges on a discussion of Date of Service aka 14 Day Rule considerations. Memdicare policies are also discussed at 217-219 [paginated 209-211].
Also tied to 14-day rule, they discuss a Civil Investigative Demand (CID) regarding compliance with 14-day rule and other issues (72-73, paginated 67-68; March 2024).
See also this remark: "Our status as both a healthcare company and a technology company presents unique complexities when attempting to comply with these myriad laws and regulations. For example, certain data services we provide as a technology company may result in compensating other healthcare providers for access to data or the right to commercialize de-identified data. While such services, standing alone, appear routine, the compliance issues become more complex when considering our status as a healthcare provider that performs clinical diagnostic testing on behalf of healthcare providers." (72, paginated 68).
CPT Codes. There is a discussion of new AMA CPT codes for some tests in their genomics line of business, and how these tests are in gapfill (preliminary prices May 1), because "CMS rejected the recommendations from experts" in 2023.
There is a discussion of new capabilities to bill DNA and RNA studies under separate CPT codes (PDF 44, paginated p 40).
LBMs. Reimbursement discussion continues, into page 45 [paginated 41], discussing "third party benefit managers" and "laboratory benefit management programs" and how these can cause "additional pricing pressure" and "substantial administrative burden." (See an article on the same topic in 360Dx on May 16.)
Payment. Average reimbursement was $916 in 2022 and $1452 in 2023; CY2021-CY22 payments were 50% of cases (PDF 202, paginated 194).
On May 14, 2024, the FDA held a webinar to educate the public on its intentions for rapidly regulating lab-developed tests. FDA has posted materials:
It may seem daunting to write - or read - a review article with a scope as broad as "Cancer therapy with antibodies," but that's what we have this month from Paul et al. Find it in Nature Reviews Cancer, https://www.nature.com/articles/s41568-024-00690-x
It's a major achievement, coming in at 28 pages and nearly 300 citations.
The focus of Paul et al. is on cancer, but see also a new article in Nature by Desautels et al., "Computationally restoring the potency of a clinical antibody against Omicron," here. Writing, "...Our results suggest that computational approaches can optimize an antibody to target multiple [viral] escape variants, while simultaneously enriching potency."
AI Corner
Here's an AI summary of Paul et al. provided by Chat GPT 4o:
One of the key skills of large language models like ChatGPT is their ability to rapidly read and summarize texts, such as journal articles or, with training, doctor-patient interviews.
This weekend, I used AI to summarize several podcasts and a book that I encountered. AI can instantly produce information in different formats. For example, AI can first create a traditional book review and then, in the next moment, show you the same material as a Q&A with five questions and five answers. The Q&A format can enhance engagement and retention (see Chat GPT 4o on Socratic Method).
I also asked ChatGPT-4 to summarize the reasons for reading a whole book versus a concise summary of it.
AI CORNER
Many labs are preparing their "crosswalk presentations" for CMS pricing meetings. The decks are due May 30 and will presented on June 25.
Labs face the challenge of crafting applications that resonate with diverse audiences. These audiences include (1) CMS physicians, (2) CMS staff without clinical backgrounds, and (3) the CMS lab expert panel, which is composed of expert industry professionals.
Writing a presentation that satisfies all these groups can be daunting, plus, your presentation has to "register" amidst 100 other presentations. It's like speed dating.
The CMS PowerPoint template covers several sections, but for this post, I'll focus on "Background," "Purpose and Method," and"Resources." To illustrate how AI can assist with this task, I've asked Chat GPT to create a fictional test case, written in the three sections.
####
The evolution of cancer diagnostics has witnessed a significant breakthrough with the introduction of the OncoGen 5000, a state-of-the-art molecular diagnostic test meticulously crafted by Precision Genomics Inc. This innovative test employs a highly advanced technique known as next-generation sequencing (NGS), which is capable of sequencing DNA at an unprecedented scale and speed. The OncoGen 5000 meticulously analyzes over 800 genes that are known to be associated with various types of cancers, identifying mutations, copy number variations, and gene fusions that are pivotal for accurate cancer diagnosis, prognosis, and therapeutic decision-making. By generating an extensive genomic profile of the tumor, this test provides unparalleled insights into the genetic underpinnings of cancer. The results are compared against a vast and meticulously curated reference database, encompassing genetic data from a diverse cohort of cancer patients worldwide. This comparison not only enhances the accuracy of the diagnosis but also facilitates the development of personalized treatment regimens tailored to the unique genetic makeup of each patient's tumor. The significance of comprehensive genomic profiling in modern oncology is underscored by endorsements from leading medical organizations such as the American Society of Clinical Oncology (ASCO) and the National Comprehensive Cancer Network (NCCN), which advocate for the integration of such advanced diagnostic tools into clinical practice to optimize patient outcomes and advance the field of personalized medicine.
Background of the Test: The OncoGen 5000 test helps doctors understand the genetic changes in cancer cells. It looks at over 800 genes related to cancer to find important mutations and changes. This test uses next-generation sequencing (NGS) to compare a patient's tumor DNA with a database of genetic information from other cancer patients. The clinician can diagnose the cancer better and decide on the best treatment plan.
Background of the Test: The OncoGen 5000 is an advanced molecular diagnostic test designed to analyze genetic mutations in cancer cells using next-generation sequencing (NGS). It examines over 800 genes related to cancer, identifying mutations, copy number variations, and gene fusions. The test compares the tumor's genetic profile with a reference database containing genetic data from a diverse group of cancer patients. This approach, recommended in 2024 guidelines from ASCO and NCCN, aids in precise cancer diagnosis and personalized treatment planning, enhancing therapeutic outcomes.
Precision Genomics Inc. has developed the OncoGen 5000 with the primary objective of revolutionizing the precision of cancer diagnosis and the personalization of treatment strategies. This test leverages the capabilities of next-generation sequencing (NGS) to conduct a comprehensive analysis of genetic alterations present in tumor DNA. The workflow begins with the extraction of DNA from a tumor biopsy, followed by NGS, which sequences the DNA to identify genetic mutations, copy number variations, and gene fusions that are critical for understanding the molecular characteristics of the cancer. The extensive data generated from sequencing is processed through a sophisticated bioinformatics pipeline, which includes steps such as data alignment, variant calling, and annotation of genetic alterations. The test's advanced machine learning algorithms then compare the patient's genomic profile to an extensive and meticulously curated reference database. This comparison yields a detailed report that provides insights into the molecular underpinnings of the tumor, highlighting actionable mutations that can inform targeted therapy decisions. The precise genomic profiling enabled by the OncoGen 5000 is essential for tailoring treatment strategies to the individual patient, thereby optimizing therapeutic efficacy and minimizing adverse effects. This approach is in line with the principles of personalized medicine, which aims to provide tailored healthcare solutions based on the unique genetic makeup of each patient.
Purpose and Methodology of the Test: The OncoGen 5000 test helps doctors diagnose and treat cancer more accurately. The method analyzes genetic changes in tumor DNA using next-generation sequencing (NGS). DNA is extracted from the tumor and sequenced to find mutations. Results are compared to a large database. This helps doctors understand specific characteristics of the cancer and choose the best treatment.
Purpose and Methodology of the Test: The OncoGen 5000 is designed to improve cancer diagnosis and treatment by providing a detailed analysis of genetic mutations in tumors using next-generation sequencing (NGS). The process involves extracting DNA from a tumor sample, running quality controls, sequencing the DNA to identify mutations, and comparing the results to a validated reference database. Advanced algorithms then interpret the data, offering insights into the tumor's molecular profile and guiding personalized treatment strategies. The process is both CAP- and New York State-validated. This precise genomic profiling helps doctors make informed decisions about the best treatment options for each patient.
The successful implementation of the OncoGen 5000 test necessitates the availability of several critical resources, which are essential for conducting a comprehensive genomic analysis. Initially, a tumor tissue sample must be obtained through a biopsy, and high-quality DNA must be extracted from this sample. The extracted DNA undergoes next-generation sequencing (NGS) using highly specialized sequencing platforms, which are capable of generating vast amounts of genetic data in a relatively short time frame. The sequencing data is then processed through a series of bioinformatics pipelines that include crucial steps such as data alignment, variant calling, and the annotation of genetic mutations. The bioinformatics analysis is further supported by a proprietary reference database that contains genomic information from thousands of cancer patients, ensuring that the comparison is robust and clinically relevant. Additionally, the test requires advanced computational infrastructure to handle the large-scale data processing and the application of machine learning algorithms, which are vital for interpreting the genomic data. The entire workflow is conducted in a certified laboratory equipped with state-of-the-art technology and staffed by personnel with expertise in molecular biology, genetics, and bioinformatics. This combination of resources and expertise is paramount to ensuring the accuracy, reliability, and clinical utility of the OncoGen 5000 test results.
Resources Required to Run the Test: To conduct the OncoGen 5000 test, a sample of the tumor tissue is needed. The DNA is extracted from this tissue and analyzed using next-generation sequencing (NGS) technology. The test requires specialized sequencing machines, a large database for comparison, and advanced software to interpret the results. This process is done in a lab with skilled scientists.
Resources Required to Run the Test: The OncoGen 5000 test requires specific laboratory capabilities, including the extraction of high-quality DNA from tumor samples, next-generation sequencing (NGS) technology for analyzing the DNA, comprising a 500 MB data output, and advanced bioinformatics tools for data interpretation. The test also relies on a comprehensive reference database and sophisticated computational infrastructure. This analysis is conducted in certified laboratories staffed with experts in molecular biology, genetics, and bioinformatics. The combination of these extensive and regularly updated resources ensures the test's accuracy and reliability in providing clinically useful results.
##
Results were perhaps more impressive starting with "real" client text, which I couldn't share. Similar approaches to re-writing might be used for tech assessment dossiers. Also, when you've seen AI do these re-writes, authors might learn to write better first drafts the next time.
Much health insurance is directly employer funded, through systems called "administrative services only" ASO or "third party administrators" TPA. In a new podcast (transcript also available), Elizabeth Mitchell of PBGH, Purchaser Business Group on Health, discussing how this approach to health benefits could be improved.
She's interviewed by Stacy Richter for the Relentless Health Value podcast (Ep. #436). Find a summary, a transcript, and the 40 minute podcast here:
Note that TPA/ASO employer funding falls under ERISA benefits, and so escape most state insurance laws (such as new state biomarker laws - here).
Are employer-driven (ASO) plans a good place for you to market innovation?
###
AI Corner
A ChatGPT4o Summary.
In episode 436 of the podcast "Relentless Health Value," Stacey Richter interviews Elizabeth Mitchell from the Purchaser Business Group on Health (PBGH). The discussion centers on the inertia present in the healthcare system, specifically related to third-party administrators (TPAs), administrative services only (ASOs), and health plans, and how they impact jumbo employers.
The episode highlights the following key points:
Role of TPAs and ASOs: TPAs and ASOs often act like full health plans, which is not what self-insured employers always need. This leads to additional charges for services that may not be required.
Market Gaps: There is a need for independent TPAs that are not owned by health plans to provide more transparent and efficient services. The acquisition of TPAs by health plans often leads to a lack of true independence and transparency.
Inertia in the System: Despite the challenges, some jumbo employers are taking proactive steps, such as direct contracting with providers, to improve healthcare access, quality, and outcomes. This approach can result in significant cost savings and better care.
Transparency and Accountability: The conversation emphasizes the need for transparent pricing and data sharing between health plans and employers. Federal actions like the Consolidated Appropriations Act (CAA) aim to enforce this transparency.
Direct Contracting: Direct contracts between employers and providers are becoming more common, leading to better alignment and improved healthcare delivery. This bypasses some of the inefficiencies and costs associated with traditional health plan administration.
Challenges and Opportunities: The episode discusses the resistance from health plans to adapt to employer needs and the importance of finding aligned partners who prioritize high-quality care. There is also a call for new market entrants to provide transparent and responsive TPA services.
Mitchell concludes with the interview for greater accountability from health system and health plan executives and emphasizes the importance of collaboration between employers and providers to drive better healthcare outcomes.
[AI Corner - let's see how AI would summarize this paper!]
https://www.nature.com/articles/s41568-024-00694-7
"A Guide to Artificial Intelligence for Cancer Researchers," by Raquel Perez-Lopez et al., published in Nature Reviews Cancer, provides a comprehensive overview of how AI can be leveraged in cancer research. It emphasizes the practical applications of AI, aiming to demystify its usage for non-computational cancer researchers. The article covers several key areas, including AI for image analysis, natural language processing (NLP), and drug discovery.Key Points:
AI's Evolution in Cancer Research:
AI has transitioned from a specialized tool to a widely accessible resource that enhances research productivity and enables new discoveries.
Researchers can use off-the-shelf software for basic applications or develop custom AI solutions for more advanced needs.
Understanding Deep Learning:
The article explains three types of deep learning: supervised, unsupervised, and reinforcement learning.
Supervised learning involves training models on labeled data, while unsupervised learning finds patterns in unlabeled data. Reinforcement learning involves agents learning optimal behaviors through interaction with their environment.
AI in Biomedical Image Analysis:
Deep learning has revolutionized image analysis, with convolutional neural networks (CNNs) and transformers being central to advancements.
AI tools can detect tumors in medical images, analyze genetic data, and assist in histopathology by distinguishing between benign and malignant cells.
AI for Language:
Large language models (LLMs) like GPT-4 have advanced NLP, enabling tasks such as summarizing text, translating, and generating new text.
These models are used in cancer research to process clinical notes, research papers, and generate structured data from unstructured text.
Emerging Uses of AI:
AI is increasingly applied in drug discovery, predicting protein structures, and identifying potential drug candidates.
Multimodal AI, which integrates various data types (images, text, genomic data), is becoming a significant trend, offering more comprehensive analysis capabilities.
Challenges and Future Directions:
The article highlights the need for better data integration, explainable AI models, and standards for data sharing.
Collaboration between AI researchers and clinicians is crucial for developing clinically relevant AI tools.
Conclusion:
This reveiw will be a valuable resource for cancer researchers looking to incorporate AI into their work. It provides practical guidance on how to start using AI tools, the types of AI applications in cancer research, and the future potential of AI in this field. Researchers may find the explanations of deep learning and the practical examples of AI applications particularly helpful.
See a new paper from UCSF reviewing the latest on state biomarker laws.
https://jamanetwork.com/journals/jama/article-abstract/2818623
Here’s the UCSF press release:
JAMA publication by TRANSPERS Collaborators Addresses Rapid Emergence of State Biomarker Testing Insurance Coverage Laws.
In JAMA article published May 13, The State of State Biomarker Testing Insurance Coverage Laws, by Lin et. al., the authors tackle the complicated topic of states legislating access to medical care. In response to perceived barriers to access to care, states have enacted legislation to mandate terms and conditions of insurance coverage for a variety of circumstances, such as cancer screening and mental health treatment parity. Recently, there have been increasing numbers of states that have passed legislation mandating that insurers cover “biomarker testing.”
This review of legislation from 15 states mandating coverage of biomarker testing finds that their impact may be limited by their lack of reach and implementation challenges. While these laws attempt to expand access to coverage, that may not always be the case since more than 50% of the population may not be covered by legislative mandates governing insurance coverage. Additionally, such insurance mandates may not affect inequities in access. Finally, there is little guidance for implementation in the legislation.
The authors conclude that current laws are likely to have limited effect, particularly concerning improving health equity, and additional approaches will likely be needed to improve patient access and outcomes.
This study was led by Grace Lin, MD, MAS (UCSF School of Medicine) and TRANSPERS (UCSF Center for Translational and Policy Research) founding director Kathryn A. Phillips, PhD, with Janet Coffman, PhD, MPP, MA (UCSF Institute for Health Policy Studies). TRANSPERS is in the Department of Clinical Pharmacy at UCSF and is affiliated with the Philip R. Lee Institute for Health Policy Studies.
###
AI Corner
You may have herd of the new AI assistant, Chat GPT 4o. It summarized the paper this way, automatically:
The landscape of state biomarker testing insurance coverage laws is rapidly evolving, driven by the need to address disparities in healthcare access. These laws mandate insurance coverage for biomarker testing, particularly for cancer subtypes, to ensure patients can access necessary treatments. However, the effectiveness and impact of these laws remain under scrutiny.
Legislative Efforts and Challenges
States have enacted various laws to improve access to cancer screenings and mental health treatments, yet the effectiveness of such mandates is debatable. For instance, mental health parity laws have had limited impact on service usage and spending. With biomarker testing, barriers like cost and inconsistent insurance coverage persist, leading to advocacy for legislative mandates.
Current State Laws
Since 2021, fifteen states have passed legislation requiring coverage for biomarker testing, with twelve including Medicaid. These laws generally define biomarkers similarly and do not exclude specific tests. Most require coverage for diagnosis, management, and monitoring, but not for screening.
Implementation and Effectiveness
The laws face challenges in implementation due to their broad language and lack of specificity. Issues like FDA approval focus on safety and efficacy rather than clinical utility, potentially complicating coverage decisions. Moreover, state mandates may not significantly impact those most affected by healthcare disparities, such as the uninsured or certain Medicaid populations.
Policy Implications
The impact of state biomarker testing laws is limited by their reach and implementation challenges. Policymakers and stakeholders must monitor these laws' effects and explore additional strategies beyond insurance mandates to enhance access and equity in healthcare. Non-legislative programs or federal legislation may be necessary to achieve broader, equitable access to biomarker testing.In conclusion, while state laws mandating biomarker testing coverage are a step toward improving access to personalized cancer treatments, their current limitations suggest that a multifaceted approach is essential to address ongoing disparities in healthcare access and outcomes.
For further details, please refer to the full article published in JAMA here.
Header: Exact Sciences May 2024 investor call: Molecular testing is capital-intensive, and sales and marketing need to be on the upswing again.
###
During COVID, there was an public equities bubble (for public companies) or a valuation bubble (for nonpublic companies) in both biotech and genomics. Since 2023, funding has been more challenging.
One example of the principal, "it takes money to make money" aka "it's capital intensive" is in the latest investors call at Exact Sciences (Cologard; Oncotype). Revenue has risen from $1.5B in 2020, to $2.5B in 2022. At the same time, operating losses have been substantial (-$558M in 2020, -835M in 2021, -577M in 20222, and -214M in 2023).
Efficiency, in dollarized terms, has improved. Revenue in 2020 ($1.5B) was less than operating expenses ($1.7B), but revenue ($2.5B) was greater than operating expenses ($2.1B) in 2023.
...We turned down our sales and marketing investments, as you know, over the last year and really over the last couple of years, while we saw about approaching $1 billion of growth.
And the truth is we probably turned that down a little bit too much. We recognized that at the back end of last year. And starting in Q1, we increased our marketing investments...
And it's clear the more time, the more frequently our reps call on an office in a position or a nurse or physician assistant, you see a straight-line increase with no flattening from zero to six calls per quarter. That gives us confidence that by adding reps, we will see it. ,,,You will see an increase in the total number of test orders.
Professional sales staff for up to 6 calls, per quarter, is part of that substantial total SG&A in molecular pathology.
As Meehan concluded in his note, "You have to spend money to make money," and the capital intensivity is a kind of moat.
Medicare's NCD 90.2, NGS in Cancer, provides automatic coverage for FDA-approved NGS companion diagnostics.
In Transmittal R12626, CMS endorses coverage as follows:
In a purely AI-based exercise, we seek business advice for companies that currently make both IVD and RUO kits. ChatGPT4.
As open access Google doc, here.
###
You are an expert lab industry business consultant. I, too, am a business consultant for the lab industry.
In the clinical lab industry today, lab test kits may be sold as "IVD" kits, which are fully FDA approved, or "RUO" kits. RUO kits are not sold as FDA approved, but local labs can validate them and then use them in clinical care. FDA policy changes promise to radically change this system. Under the new FDA regulation, which has been finalized, FDA will not allow RUO kits to be validated and used clinically by local labs. Instead, only FDA approved IVD kits will be permitted.
My client is ABC, which is a leading manufacturer of both IVD kits (for the most popular tests, justifying the FDA costs) and RUO kits (for a wide range of less popular tests, where the IVD investment would be uneconomical.) ABC asks if it should start investing in converting its catalog of RUO kits to IVD kits, a large and time consuming investment. However, it would be worthwhile if in the future, labs are "handcuffed" to buying IVD kits only. But there are many complicating variables. The FDA regulation will not be enforced until 2027. Between now and then, the regulation may be canceled, changed, or delayed. To clarify, for example, the client ABC could invest $100M in FDA approvals, but if the FDA rule is canceled, clients will only want to buy the cheaper RUO kits (which are today's status quo.)
How can the client ABC think about the threat and opportunity to its business model, and how would you advise the commercial teams (sales), the R&D teams, and the C-level leaders of ABC.
ChatGPT
Navigating regulatory changes like the one proposed by the FDA, which would prohibit the clinical use of Research Use Only (RUO) kits unless they are FDA-approved as In Vitro Diagnostics (IVD), poses a significant strategic challenge for companies like ABC. Here’s a strategic approach you might consider in advising ABC's teams on handling this situation:
I've been enjoying the exploration of what AI can do in when tasked with summarizing documents. Here's a case study. This morning's email brought a press release from Friends of Cancer Research about a new publication, McKelvey et al., on ctDNA uses and accuracy in cancer monitoring. The press release also linked to a FOCR web page about the ongoing project and plans.
I gave the press release, the McKelvey article [open access] and the ctDNA website to ChatGPT4 and asked for a summarizing article. Result below.
You might enjoy comparing the FOCR ctDNA landscape by McKelvey with another new article, by Bronkhorst, on "the landscape of methodologies, technologies, and commercialization." Here. AI comparison of Bronkhorst and McKelvey here.
###
[Machine Generated Copy:]
Friends of Cancer Research Launches ctMoniTR Project
for Circulating Tumor DNA Research
Washington, DC – May 7, 2024
Friends of Cancer Research, a prominent cancer advocacy organization, has recently published a groundbreaking study in the journal Diagnostics examining circulating tumor DNA (ctDNA). The study, "Advancing Evidence Generation for ctDNA: Lessons Learned from A Multi-Assay Study of Baseline ctDNA Levels Across Cancer Types and Stages," provides critical insights into the use of ctDNA as a biomarker in oncology.
ctMoniTR Project Overview
The ctMoniTR Project, launched by Friends of Cancer Research, aims to evaluate the potential of ctDNA as a predictor of clinical responses to therapy across various solid tumors. The organization formed a collaboration involving multiple commercial ctDNA assay developers to assess baseline ctDNA levels across different cancer types and stages. The project focuses on establishing ctDNA as an early endpoint in oncology drug development, which can facilitate faster identification and development of effective treatments.
Key Findings
The study highlights the promise of ctDNA as a biomarker and surrogate for disease burden. The collaborative effort found that baseline ctDNA levels were similarly detected across late-stage cancers such as non-small cell lung cancer (NSCLC), breast, bladder, prostate, and head and neck squamous cell carcinoma (HNSCC). This similarity underscores the potential for ctDNA to be used as a biomarker in these cancer types. However, the study also revealed variability in ctDNA levels across different assays in early-stage NSCLC, indicating the need for standardized assay analytical performance and methodology.
Critical Insights and Future Directions
Dr. Brittany Avin McKelvey, the study's lead author and Director of Regulatory Affairs at Friends of Cancer Research, emphasized the importance of meta-analyses across multiple cancer types and ctDNA assays to support the future use of ctDNA as an early endpoint. The study underscores the necessity of developing common data standards and key performance characteristics to generate robust evidence for ctDNA's use as a biomarker and early endpoint in oncology.
Dr. Jeff Allen, President and CEO of Friends of Cancer Research, highlighted the importance of this unique partnership in providing foundational evidence regarding ctDNA levels across different tumor types and stages. The organization's ctMoniTR Project is an ongoing effort to generate evidence for ctDNA's potential in oncology drug development and patient care.
Conclusion
The collaborative efforts of Friends of Cancer Research and its partners have led to significant advances in the understanding of ctDNA's role in oncology. The ctMoniTR Project is poised to play a crucial role in advancing cancer research, potentially leading to faster drug development and improved patient outcomes.
[End Machine-generated Copy]
Moldx has an elaborate article regarding billing and coding of tumor panels - A54795, which is now up to "revision 16." Find it here:
https://www.cms.gov/medicare-coverage-database/view/article.aspx?articleid=54795&ver=42&bc=0
What's New?
It may take a master's degree in "MolDxology" to decipher some of this, but here's their handy guide to what you have to know about this revision.
Under Article Title revised to MolDX: Targeted and Comprehensive Genomic Profile Testing in Cancer.
Under Article Text subsection heading Next Generation Sequencing (NGS) 2nd sentence revised “tumor tissue only-based panels” to read “tumor-based panels for cancer that may be performed by NGS”. Added verbiage and hyperlink for “Refer to MolDX: Defining panel services in MolDX A59678 for further guidance on the distinction between single analyte tests and panel tests”. Under subsection heading Targeted Tumor Panels revised 1st sentence “Targeted Next-Generation Sequencing (NGS) panels” to read “Targeted tumor panels”. Revised 2nd sentence to read “Generally, these panels are limited to specific variant types at defined sites, such as single nucleotide variants (SNVs), small insertions or deletions (INDELs), single site copy number variants, or gene fusions”. Revised last sentence to delete “regions in the genes” and replaced with “targets”. Under subsection heading Comprehensive Genomic Profile (CGP) Testing revised 1st sentence “CGP” to read “CGP testing”. Revised 4th sentence “CGP” to read “CGP tests” and replaced “copy number alterations (CNAs)” with “copy number variants (CNVs)”. Revised 5th sentence to add “and chromosome abnormalities such as loss of heterozygosity (LOH)”. Revised 6th sentence “CGP” to read “CGP testing”. Added new sentences “CGP tests are expected to yield information of clinical relevance beyond a targeted panel, for example, to identify relevant clinical trials for patient management or identify possible therapeutic interventions for off-label use. It is expected that a CGP will identify all clinically relevant information attainable for the type of service performed”. Revised subsection heading “Targeted Panels” to read “Targeted Tumor Panels”. Revised 1st sentence to read “To bill for DNA-based panels that measure SNVs, INDELs, CNVs or rearrangements, review CPT codes 81445 and 81450” and deleted 2nd and 3rd sentences. Added new sentences “If a DNA-based targeted panel meeting the coverage requirements is used, and MSI is also performed, 81457 can be billed. If a DNA-based targeted test meeting coverage requirements is performed that includes MSI and CNVs, 81458 can be billed” and 2 new paragraphs. Revised last paragraph “DEX Z-Code™” to read “DEX Z-Code®”. Under subsection heading CGP deleted first paragraph and added new paragraph. Revised 2nd paragraph 1st sentence to add “81459 or” and replaced “DEX Z-Code™” with “DEX Z-Code®”. Revised 2nd sentence to replace “other” with “further”. Revised 3rd paragraph sentence “DEX Z-Code™” to read “DEX Z-Code®”. Formatting, punctuation, and typographical errors were corrected throughout the article.
Got that? Quiz Tuesday.
Notably, one standout is that (while they never accepted 81455) they will accept 81459 instead of 81479 in some circumstances. Note, though, that under separate current gapfill processes, 81459 is to be paid at $2990, basically the same as 81455.
To help, I've put in the cloud a PDF of a redline from R15 to R16:
https://drive.google.com/file/d/1lvuE9t2awnT81E1H0KOkufUYJjQgPsjl/view?usp=sharing
###
AI CORNER
Refer only to the original text of A54795 for MolDx rules and terminology.
However, for fun, I asked Chat GPT4 to compare the older and newer documents ("March" vs "April.")
##
Attached, you will find an OLD version called MARCH 2024, and a NEW version called APRIL 2024. Please explain the details and differences, focusing on what your readers need to know to transition from operations under the MARCH version and move to new rules and operations under the APRIL version. We're counting on you to make sense of this!
Famously, CMS has a regulation (42 CFR 410.32) that tests must be ordered by a physician who is treating a patient for a particular disorder. However, CMS provides some exceptions where pathologists (or radiologists) can adjust orders or, for pathologists, implement orders necessary to complete a case. (The example provided is seeing possible bacteria and running a special stain).
(For additional clarification, see Medicare Benefits Manual 80.6, and the original Fed Reg creation of 410.32 at 10/31/1997, 62 FR 59048).
MolDx has published an important public policy article about when a pathologist can order molecular tests. Find the article here:
https://www.cms.gov/medicare-coverage-database/view/article.aspx?articleid=59741&ver=2&bc=0
##
##
A59741
Documentation requirements for laboratory services require that services be ordered by a treating physician as defined in IOM 100-02 Chapter 15 §80.6.1 and meet other criteria set in 42CFR§410.32(b). Further clarifications were made in ICN MLN909221, December 2020.
In view of these requirements, this contractor provides the following supplemental clarifying information:
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:
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
