Google LM Reads my "AI Teaching Guide" to DCIS LCD

I uploaded the new MolDx DCIS LCD to Chat GPT for a teaching analysis.

Here, below, I upload that teaching analysis (with the LCD) to Google LM for a summary and an audio version.

See the original teaching case by Chat GPT here.  Hear a 17 minute podcast by Google here.  

###  Google LM Below

Briefing Document: MolDx Non-Coverage Policy for DCIS Risk Stratification Molecular Tests

Date: July 18, 2025

To: CEOs and Chief Science Officers at Molecular Diagnostic Startups (Series B/C Stage)

From: Bruce Quinn MD PhD, Expert on Health Reform, Innovation, and Medicare Policy

Subject: Key Lessons from MolDx's Non-Coverage Policy on DCIS Breast Cancer Molecular Tests (DCISionRT and Oncotype DCIS Score)

Executive Summary

MolDx has issued a non-coverage Local Coverage Determination (LCD) for molecular and proteomic tests designed to risk-stratify Ductal Carcinoma in situ (DCIS) patients, specifically those aiming to identify "sufficiently low-risk" individuals who may safely forego adjuvant radiation therapy (RT). This policy, DL40142, is a critical blueprint for any company developing molecular diagnostics seeking Medicare coverage. MolDx's extensive review found the data supporting current DCIS tests, such as DCISionRT and Oncotype DCIS Score, to be consistently lacking in statistical power, clear clinical utility, and superiority over existing clinicopathologic tools. 

This document provides a detailed analysis of MolDx's rationale, highlighting their stringent evidentiary thresholds and expectations for future test development.

I. MolDx's Core Expectations and Evidentiary Thresholds for Molecular Diagnostics

MolDx's non-coverage decision for DCIS risk stratification tests provides unprecedented insight into their review process. Their primary objective is to ensure that new molecular diagnostics offer meaningful, validated improvement in patient care, particularly within the Medicare population.

A. Direct Comparison to Existing Tools: Superiority, Not Just "Added Insight"

• MolDx explicitly requires new molecular diagnostics to demonstrate clear superiority over established clinicopathologic tools, such as the MSKCC nomogram or Van Nuys Prognostic Index (VNPI). "Adds another data point" is insufficient.
• 
  
• MolDx Criticism: "MolDx criticized DCISionRT and Oncotype DCIS Score for failing to show they outperform these established models." In one study comparing MSKCC to Oncotype DCIS Score, a "92% concordance" was observed, indicating the molecular test did not meaningfully differentiate risk groups.
• Implication for Startups: "You must benchmark your test directly against the current standard of care — and show where and why you are better. 'Adds another data point' is not sufficient." Your analytical and clinical validation must isolate the molecular contribution and demonstrate its independent predictive power.

B. Statistically Robust and Replicated Data

A recurring theme in MolDx's critique is the inadequacy of sample sizes and the inconsistency of results, especially in low-risk cohorts.

• MolDx Criticism: "MolDx repeatedly criticizes the small sample sizes and inconsistent results in the low-risk cohorts for these tests." For DCISionRT, "single-digit recurrence events (e.g., 4 vs. 12 recurrences)" in low-risk groups led to "wide confidence intervals and inconclusive findings (p > 0.05)."
• MolDx Criticism: MolDx noted that "newer cutoffs (e.g., DS2 = 2.8) were derived from re-analysis of old data without sufficient external validation. Studies lacked independent validation cohorts, leaving questions of reproducibility."
• Implication for Startups: "Design your validation studies for statistical power within the subpopulations that will drive clinical decisions. Ensure external validation in independent cohorts is complete before seeking coverage." This includes thorough analysis and concordance studies if multiple cohorts are used.

C. Proven Clinical Utility Tied to Meaningful Outcomes

MolDx requires evidence that the test changes management in a way that affects health outcomes, especially in Medicare populations. A test's impact on radiation decisions is insufficient if the benefit is statistically trivial.

• MolDx Criticism: "DCISionRT’s 'low-risk' patients still showed relative risk reduction with RT; MolDx noted that the observed benefit (~5%) wasn’t statistically conclusive due to small numbers, and therefore claims about safely omitting RT are not substantiated." The Contractor Advisory Committee (CAC) generally agreed that "at a 5% absolute risk reduction of IBTR (or less) confers no significant clinical impact from RT therapy."
• Implication for Startups: "You need robust clinical utility studies showing how your test leads to safer or more efficient care, not just changes in provider behavior." This means demonstrating that the test enables identification of a patient population for whom there is no reasonable benefit from RT, defined by MolDx as an "absolute reduction in recurrence risk from RT of ≤5%."

D. Clear Definition of "Low Risk" Tied to Clinical Impact

MolDx demands a precise, defensible definition of "low risk" that directly correlates with a quantifiable clinical impact.

• MolDx Criticism: "MolDx criticized studies for inconsistent definitions of 'low risk,' varied inclusion of factors like age or tumor size, and inconsistent recurrence rates across cohorts. This muddles the claim that molecular tests are identifying patients who can safely avoid RT."
• Implication for Startups: "You must define and defend your clinical risk thresholds in terms aligned with clinical practice impact and Medicare's benefit-to-risk frameworks." This standardizes the definition, potentially based on an expected 10-year Ipsilateral Breast Tumor Recurrence (IBTR) rate of approximately 10% where RT offers minimal benefit (absolute risk reduction of ≤5%).

E. "Rhetorical Proof" Is Not Scientific Proof

MolDx rejects the notion that a test's value stems primarily from its ability to reassure patients or physicians.

• MolDx Criticism: "MolDx called out the fallacy of tests being used merely to persuade patients or doctors ('rhetorical proof') rather than meeting a robust evidentiary standard." Some subject matter experts (SMEs) at the CAC championed DCISionRT because it "reassures patients about skipping RT," but MolDx found this insufficient without statistically sound data.
• Implication for Startups: "Avoid leaning on 'decision-support narratives' unless backed by robust data. MolDx wants evidence of superior performance, not convenience in shared decision-making."

F. Multi-Factor Models Must Show Incremental Value of Molecular Component

If a test combines molecular and clinical factors, the molecular contribution must be demonstrably valuable beyond the clinical data alone.

• MolDx Criticism: "MolDx criticized both DCISionRT and Oncotype DCIS for including overlapping clinicopathologic factors already in nomograms — questioning whether the biomarkers provide independent predictive power."
• Implication for Startups: "Your analytical and clinical validation must isolate the molecular contribution and demonstrate why it matters."

G. Medicare-Centric Evidence is Required

Validation cohorts must reflect the demographic and clinical context of Medicare beneficiaries, who are the primary population for DCIS.

• MolDx Criticism: MolDx repeatedly emphasizes "the need for Medicare-relevant data, especially in older populations, which are both the predominant DCIS cohort and the primary Medicare demographic."
• Implication for Startups: "Ensure your validation cohorts reflect the age, health status, and clinical context of Medicare beneficiaries."

II. The Critical Impact of Rare Event Rates on Study Design

One of the most significant statistical lessons from this MolDx LCD, and often overlooked by startup diagnostic companies, is the profound impact of rare event rates on study design and evidentiary thresholds.

A. The Challenge of Low Event Rates in DCIS

• In DCIS, the clinical question often revolves around identifying patients with a very low risk of recurrence (typically around 5% over 10 years). This means that "out of 100 patients, only five or fewer will experience a recurrence."
• 
  
• Why Rare Events Demand Larger N: This low event rate "creates a major statistical hurdle." "In any cohort of 100 patients, observing only 5 events leaves your study underpowered to detect meaningful differences between groups." "Small absolute numbers (4 vs. 6 recurrences, or 5 vs. 7) generate wide confidence intervals and increase the risk of false negative or non-significant findings even if a real effect exists."

B. MolDx's Specific Criticism Regarding Underpowered Studies

• MolDx repeatedly highlighted that the published studies on DCISionRT and Oncotype DCIS were "consistently underpowered for their intended purpose."
• 
  
• MolDx Criticism: In low-risk groups, "there were often only single-digit numbers of recurrences." Some studies "split these few events across treatment arms (e.g., 4 vs. 12 recurrences with or without RT), leading to non-significant p-values and wide 95% confidence intervals."
• MolDx Criticism: "MolDx specifically rejected claims that the tests could identify populations with no benefit from RT, pointing out that such conclusions were built on insufficient event counts and underpowered analyses." A "1% or 2% difference in recurrence might look reassuring to clinicians, but MolDx will not consider such findings statistically reliable without appropriate sample sizes."

C. Implications for Startups Developing Tests for Rare Outcomes

• Mandatory Large Sample Sizes: If your test predicts rare events (≤ 5% incidence), "you must design studies accordingly." "Sample sizes must be large enough to accumulate a statistically meaningful number of events in each risk category and treatment arm." This means "hundreds, not dozens, of events may be necessary."
• Careful Retrospective Data Use: If using existing retrospective datasets, "ensure your sample size planning considers event rates, not just patient counts."
• The Illusion of Precision: "When event rates are low, small N studies are misleading. They create the illusion of precision without the statistical power to support regulatory or payer coverage decisions. MolDx expects mature companies to understand this."
• 
  

III. Summary and Path Forward for Molecular Diagnostic Startups

MolDx's non-coverage policy for DCIS molecular tests serves as a clear "transparent playbook" for diagnostic companies. To achieve reimbursement success and meet Medicare's "reasonable and necessary" standards, startups must:

1. Demonstrate Superiority: Prove your test is more accurate and clinically useful than existing, widely available clinicopathologic tools (e.g., nomograms).
2. Ensure Statistical Rigor: Design studies with sufficient statistical power, especially when targeting rare events or identifying very low-risk populations. This requires large sample sizes and external validation in independent, representative cohorts (including Medicare beneficiaries).
3. Prove Clinical Utility: Show that your test leads to meaningful changes in patient outcomes or significant improvements in care efficiency, not just changes in physician behavior or "rhetorical proof." Define "low-risk" explicitly as patients for whom the absolute risk reduction from RT is ≤5%.
4. Isolate Molecular Value: If combining factors, clearly demonstrate the incremental value of the molecular component beyond clinical data alone.
5. Focus on Medicare Relevance: Ensure your evidence is relevant to the Medicare population, particularly older patients.

MolDx is open to covering biomarker tests for DCIS risk stratification "should evidence be further developed to meet reasonable and necessary standards." However, this requires a fundamental shift in how studies are designed and how evidence is presented, moving beyond small, underpowered analyses and anecdotal "decision support" claims to robust, statistically sound demonstrations of superior clinical utility and meaningful patient impact.

NotebookLM can be inaccurate; please double check its responses.