PRIVILEGED AND CONFIDENTIAL — ATTORNEY-CLIENT COMMUNICATION / ATTORNEY WORK PRODUCT This memorandum has been prepared by outside counsel at the direction of and for the exclusive use of the client. It is protected from disclosure by the attorney-client privilege and the work product doctrine. Any unauthorized review, use, disclosure, or distribution is strictly prohibited.
CALDERWOOD & HARKNESS LLP 1700 K Street NW, Suite 900 Washington, DC 20006
MEMORANDUM
Date: April 15, 2024
To:
Theresa Nakamura, General Counsel & Vice President of Regulatory Affairs, Genome Dx Therapeutics Inc.
Kevin Driscoll, Vice President of Market Access & Reimbursement, Genome Dx Therapeutics Inc.
Dr. Priya Sundaram, Vice President of Clinical Development, Genome Dx Therapeutics Inc.
From:
Miriam Calloway, Partner, Healthcare & Life Sciences Practice Group, Calderwood & Harkness LLP
Sean Morrissey, Senior Associate, Healthcare & Life Sciences Practice Group, Calderwood & Harkness LLP
Re: Clinical Evidence Gap Analysis — Impact of CLARITY-LUNG Interim Results on OncoPrecise GXP-47 Medicare Coverage Strategy
Matter No.: 1043-00002
I. EXECUTIVE SUMMARY
This memorandum presents our analysis of the March 29, 2024 interim analysis results from the CLARITY-LUNG clinical validation study (ClinicalTrials.gov ID: NCT05891247) and their implications for Genome Dx Therapeutics Inc.'s ("GDT") planned Local Coverage Determination ("LCD") application with Cascade Fiscal Intermediaries, LLC ("CFI") and concurrent MolDx Technical Assessment submission to Ridgeline Administrative Services, LLC ("RAS"). This memorandum was prepared at the direction of the GDT Board of Directors following its April 9, 2024 meeting, at which the Board requested a formal reassessment of GDT's LCD filing strategy in light of the interim clinical data.
The central finding of this memorandum is that the CLARITY-LUNG interim results for the OncoPrecise GXP-47 47-gene expression panel fell materially below the pre-specified performance thresholds established in the trial protocol. Specifically, the interim analysis yielded a positive predictive value ("PPV") of 0.68 against a pre-specified threshold of ≥ 0.75, and a negative predictive value ("NPV") of 0.71 against a pre-specified threshold of ≥ 0.80. These shortfalls — representing deficits of 9.3% and 11.25% respectively relative to target — create material evidence gaps that, in our assessment, substantially increase the risk of a non-coverage determination if the LCD application were filed based on the currently available clinical evidence.
We have conducted a detailed analysis of each evidence gap, assessed the probability of adverse coverage outcomes under multiple filing scenarios, evaluated five supplemental evidence strategies that could partially address the identified gaps, and considered the downstream implications for GDT's PLA code application, MolDx Technical Assessment Dossier preparation, and the health economics and outcomes research ("HEOR") analysis being conducted by Larkfield Health Economics, LLC.
Our principal conclusions are as follows:
1.  The current clinical validation evidence is insufficient to support an LCD filing with a reasonable probability of favorable coverage. We assess the probability of a non-coverage determination at approximately 65–75% if the LCD is filed based on the interim data as currently available, and approximately 50–60% even with the addition of supplemental evidence strategies described herein. These risk levels are unacceptably high given the significant downstream consequences of an adverse LCD.
2.  Supplemental evidence strategies may partially address the identified gaps but are unlikely to fully substitute for final CLARITY-LUNG data meeting or approaching the pre-specified performance endpoints. Several of the strategies discussed in this memorandum — including retrospective validation sub-studies, pre-specified sub-group analyses, publication of interim results, and external expert clinical utility assessments — are worth pursuing as components of a comprehensive evidence-building program, but none, individually or collectively, can overcome the fundamental shortfall in prospective clinical validation performance.
3.  A detailed revised strategy memorandum with filing timeline recommendations will follow separately. A comprehensive Revised LCD/MolDx Strategy Memorandum is in preparation and is anticipated for delivery on or about April 22, 2024. That memorandum will present alternative filing timeline scenarios — including the option of deferring the LCD application until after the final CLARITY-LUNG data readout, now projected for Q2 2025 — and will address the financial, strategic, and corporate governance implications of each scenario in greater detail.
This memorandum should be read in conjunction with the forthcoming Revised LCD/MolDx Strategy Memorandum. Together, these two documents are intended to provide the GDT Board of Directors, management team, and in-house counsel with a comprehensive analytical framework for making informed decisions regarding the OncoPrecise GXP-47 Medicare coverage strategy.
II. BACKGROUND
A. OncoPrecise GXP-47 Test Overview
OncoPrecise GXP-47 is a laboratory-developed test ("LDT") performed exclusively at GDT's reference laboratory located at 4720 Torrey Pines Road, Suite 300, San Diego, California 92130. The laboratory holds CLIA certification (CLIA No. 05D2187432) and accreditation from the College of American Pathologists ("CAP") (CAP No. 9074821). The laboratory's National Provider Identifier is NPI 1902847365. All testing is performed on-site at this single reference laboratory facility; no testing is performed at satellite locations or through subcontracted laboratories.
The test employs a 47-gene RNA expression panel utilizing RNA extracted from formalin-fixed, paraffin-embedded ("FFPE") tumor tissue specimens obtained from patients with non-small-cell lung cancer ("NSCLC"). The assay measures the expression levels of 47 genes implicated in immune response pathways, tumor microenvironment characteristics, and treatment resistance mechanisms. Through a proprietary algorithm developed by GDT's bioinformatics team, these gene expression measurements are combined to produce a composite Therapy Response Index ("TRI") score ranging from 0 to 100.
The TRI score is designed to stratify patients into three risk categories based on their predicted likelihood of responding to immune checkpoint inhibitor ("ICI") therapy:
•  High Likelihood of Response: TRI ≥ 70
•  Intermediate: TRI 40–69
•  Low Likelihood of Response: TRI < 40
The intended clinical use of OncoPrecise GXP-47 is to guide oncologist decision-making regarding the initiation of first-line ICI therapy — including pembrolizumab, nivolumab, and atezolizumab — versus platinum-based chemotherapy in patients with Stage IIIB/IV NSCLC. In its intended clinical workflow, the test report would be provided to the treating oncologist along with the patient's PD-L1 expression status and other standard pathology results, with the TRI score serving as an additional predictive biomarker to inform the first-line treatment selection.
GDT completed analytical validation of the OncoPrecise GXP-47 assay on December 14, 2022. The analytical validation study results were published in the Journal of Molecular Oncology Diagnostics, Volume 18, Issue 4, pages 412–429, in December 2022. That publication demonstrated acceptable analytical performance characteristics, including reproducibility, repeatability, accuracy, and limit of detection, in accordance with established guidelines for molecular diagnostic assay validation. The analytical validation study remains, as of the date of this memorandum, the only peer-reviewed publication specifically addressing the OncoPrecise GXP-47 test.
B. Medicare Coverage Strategy Context
GDT's reimbursement strategy for OncoPrecise GXP-47 encompasses four sequential components, each building upon the preceding step:
1.  PLA Code Assignment. GDT is seeking a Proprietary Laboratory Analyses ("PLA") code through the American Medical Association ("AMA") CPT Editorial Panel. A draft PLA code application was delivered to Calderwood & Harkness for review on March 22, 2024. The application targets the AMA CPT Editorial Panel's September 2024 meeting, for which the submission deadline is June 28, 2024. The PLA code, once assigned, would provide a unique CPT code specific to OncoPrecise GXP-47, enabling clean billing and claim submission to Medicare and commercial payers.
2.  Local Coverage Determination (LCD) and MolDx Technical Assessment. GDT planned to submit an LCD application to CFI (Jurisdiction E, covering California) with a concurrent MolDx Technical Assessment submission to RAS. The original target LCD filing date was August 2024, following PLA code assignment. An LCD from CFI would establish Medicare coverage for OncoPrecise GXP-47 in Jurisdiction E, which includes the state of California — the jurisdiction in which GDT's reference laboratory is located and from which it would bill Medicare for all tests performed. The LCD application and supporting MolDx Technical Assessment Dossier have been under development by Calderwood & Harkness in coordination with GDT's regulatory affairs and clinical development teams.
3.  Establishment of Medicare Reimbursement Rate. GDT's target Medicare reimbursement rate for OncoPrecise GXP-47 is $3,850 per test. This rate has been determined through a combination of cost analysis, comparable test pricing, and value-based pricing considerations. Establishment of the final reimbursement rate would follow LCD issuance and would be subject to standard CMS pricing processes applicable to tests reimbursed under the Clinical Laboratory Fee Schedule ("CLFS") or, if applicable, the Protecting Access to Medicare Act ("PAMA") framework.
4.  National Coverage Determination (NCD). GDT's longer-term objective is to pursue a National Coverage Determination through CMS, which would establish Medicare coverage for OncoPrecise GXP-47 on a nationwide basis. An NCD application would likely not be pursued until after GDT has secured at least one favorable LCD and has accumulated a more robust body of clinical evidence, including final CLARITY-LUNG results and, ideally, additional clinical utility data.
The projected Medicare reimbursement revenues associated with OncoPrecise GXP-47 are a critical element of GDT's financial planning and growth strategy. Based on GDT's commercial projections, Year 1 Medicare revenue is estimated at $20,290,500, calculated as follows: 8,500 projected tests multiplied by 62% Medicare patient mix, multiplied by the target reimbursement rate of $3,850 per test (8,500 × 0.62 × $3,850 = $20,290,500). Year 1 total projected revenue, inclusive of commercial payer and self-pay volume, is estimated at $34,502,500. These revenue projections are integral to GDT's operating budget, debt service capacity, and investor return models.
GDT is not a newcomer to the MolDx and LCD process. GDT's existing product, OncoPrecise GXP-12 — a 12-gene expression panel for early-stage breast cancer prognosis — has an existing favorable LCD issued by CFI (LCD L38472, effective August 15, 2021). GDT's prior experience with the CFI and MolDx process provides institutional familiarity with the application requirements and evidentiary standards, and GDT's existing relationship with CFI may be beneficial in future interactions. However, the existence of a favorable LCD for a different test product does not create any presumption of favorable coverage for OncoPrecise GXP-47; each test is evaluated independently on the strength of its own evidence base.
C. The CLARITY-LUNG Trial
The CLARITY-LUNG trial (ClinicalTrials.gov ID: NCT05891247) is the cornerstone of GDT's clinical validation program for OncoPrecise GXP-47. It is a prospective, multi-site, blinded clinical validation study designed to evaluate the predictive performance of the TRI score in identifying NSCLC patients likely to respond to first-line ICI therapy.
The trial is being conducted across 14 clinical sites located throughout the United States. The Principal Investigator is Dr. Priya Sundaram, GDT's Vice President of Clinical Development. Patient enrollment commenced on January 18, 2023, with a target enrollment of 1,200 patients with newly diagnosed Stage IIIB/IV NSCLC. Eligible patients undergo standard-of-care diagnostic workup, including tumor biopsy. An additional FFPE tissue sample is obtained for OncoPrecise GXP-47 testing, but the TRI score results are blinded — that is, not disclosed to the treating oncologist or patient — so as not to influence treatment decisions. Patients receive first-line therapy as determined by their treating oncologists in accordance with standard-of-care clinical practice. Patient outcomes, including progression-free survival ("PFS"), are tracked prospectively.
The primary endpoint of the CLARITY-LUNG trial is the positive predictive value ("PPV") and negative predictive value ("NPV") of the TRI score for 12-month progression-free survival on ICI therapy. Specifically:
•  PPV measures the proportion of patients classified as "High Likelihood of Response" (TRI ≥ 70) who actually achieve 12-month PFS on ICI therapy.
•  NPV measures the proportion of patients classified as "Low Likelihood of Response" (TRI < 40) who actually do not achieve 12-month PFS on ICI therapy (i.e., the test correctly identifies non-responders).
The pre-specified performance thresholds for the primary endpoints, as established in the CLARITY-LUNG study protocol, are:
•  PPV ≥ 0.75
•  NPV ≥ 0.80
These thresholds were determined through statistical modeling, input from KOLs in thoracic oncology, and alignment with MolDx evidentiary expectations for predictive biomarker tests in oncology. The thresholds were reviewed and approved by GDT's Clinical Advisory Board prior to trial initiation.
The total budget for the CLARITY-LUNG trial is $28,500,000, encompassing site costs, patient recruitment, sample processing, data management, biostatistical analysis, DSMB operations, and regulatory affairs support.
The CLARITY-LUNG Data Safety Monitoring Board ("DSMB") is chaired by Dr. Evelyn Matsuda, MD, PhD, of the Carrington Cancer Institute, located at 7500 Fannin Street, Houston, Texas 77054. The DSMB is composed of independent experts in thoracic oncology, biostatistics, and clinical trial methodology who are not employed by or affiliated with GDT.
D. CLARITY-LUNG Interim Analysis Results (March 29, 2024)
The CLARITY-LUNG study protocol included a pre-specified interim analysis to be performed upon reaching the 50% enrollment milestone. That milestone was reached in early 2024, and the interim analysis was performed by the trial's independent biostatistical group on March 29, 2024, based on data from 600 enrolled patients (50% of the 1,200-patient target).
As of the date of this memorandum, actual enrollment in CLARITY-LUNG stands at 847 patients, representing 70.58% of the 1,200-patient enrollment target. Trial expenditures through the end of the most recent reporting period total approximately $19,100,000, representing 67.02% of the $28,500,000 total trial budget.
The interim analysis results for the primary endpoints were as follows:
Neither primary endpoint met the pre-specified performance benchmarks at the interim analysis.
The DSMB convened a meeting on April 5, 2024 to review the interim analysis results. Following its review, the DSMB issued a formal recommendation to continue enrollment in the trial. The DSMB did not recommend stopping the trial for futility. However, the DSMB's written communication to the sponsor characterized the interim results as "inconclusive and below pre-specified performance benchmarks." This characterization, while technically permitting trial continuation, is significant from a regulatory and coverage strategy perspective because it underscores the provisional and unsatisfactory nature of the interim data.
The final data readout for CLARITY-LUNG was originally projected for Q4 2024, based on the anticipated pace of enrollment and the 12-month PFS follow-up period required for the primary endpoint. Following the interim analysis and a reassessment of enrollment projections and data maturation timelines, the revised projection for the final data readout is Q2 2025 — representing a delay of approximately two to three calendar quarters from the original projection. This delay has cascading effects on the LCD filing timeline, commercial launch planning, and revenue projections, as discussed throughout this memorandum.
III. MOLDX PROGRAM EVIDENTIARY STANDARDS
A. Overview of MolDx Technical Assessment Requirements
The MolDx program is a collaborative initiative among several Medicare Administrative Contractors ("MACs") that provides standardized technical assessments for molecular diagnostic tests seeking Medicare coverage. The program is administered by Ridgeline Administrative Services, LLC ("RAS"), located at 1600 Hampton Street, Suite 700, Columbia, South Carolina 29201. The MolDx Technical Assessment Director is Dr. Catherine Embry, PhD.
Under the MolDx framework, molecular diagnostic tests seeking Medicare coverage through an LCD are required to undergo a rigorous technical assessment conducted by the MolDx program staff. The technical assessment evaluates the test across three evidentiary domains:
1.  Analytical Validity: Does the test accurately and reliably measure the analytes (in this case, RNA expression levels of 47 genes) that it claims to measure? This domain encompasses assay performance characteristics such as accuracy, precision, reproducibility, limit of detection, and analytical specificity and sensitivity.
2.  Clinical Validity: Does the test accurately identify the clinical condition or predict the clinical outcome it claims to identify or predict? For OncoPrecise GXP-47, clinical validity requires demonstration that the TRI score accurately predicts response to ICI therapy, as measured by PPV and NPV for 12-month PFS.
3.  Clinical Utility: Do the test results change clinical management, and do those changes lead to improved patient outcomes? For OncoPrecise GXP-47, clinical utility requires evidence — or at least a persuasive argument supported by available data — that treatment decisions guided by the TRI score lead to better patient outcomes than standard-of-care treatment selection without the TRI score.
Of these three domains, the analytical validity standard is the least onerous from GDT's perspective, as GDT has completed and published its analytical validation study and has not encountered significant challenges in this domain. The clinical validity and clinical utility domains require the most robust evidence base and are the domains in which the CLARITY-LUNG interim results create material evidentiary gaps.
The MolDx technical assessment results in a recommendation to the relevant MAC — in this case, CFI — which then issues the LCD. While the MAC medical director retains ultimate authority over the LCD decision, MolDx technical assessment recommendations carry substantial weight and are rarely overridden. A negative MolDx technical assessment recommendation virtually ensures a non-coverage LCD.
B. Clinical Validity Evidence Requirements
The clinical validity evidentiary standard under the MolDx program requires the test developer to demonstrate that the test accurately identifies the clinical condition or predicts the clinical outcome it claims to measure. For predictive biomarker tests in oncology — the category into which OncoPrecise GXP-47 falls — the standard expectation is at least one prospective clinical validation study demonstrating statistically significant PPV and NPV (or other appropriate performance metrics) relative to pre-specified thresholds.
This expectation reflects the MolDx program's long-standing position that retrospective validation studies, while considered supportive evidence, are generally insufficient as standalone evidence for clinical validity of predictive biomarker tests. Retrospective studies are susceptible to selection bias, ascertainment bias, and other methodological limitations that prospective blinded studies are specifically designed to avoid. Accordingly, MolDx assessors expect to see prospective validation data as the primary evidence of clinical validity, supplemented as appropriate by retrospective analyses, sub-group analyses, and published literature.
CFI, which would issue the LCD for OncoPrecise GXP-47 in Jurisdiction E, has a Medical Director (Dr. William Hargrove, MD, FACP) and an LCD Analyst (Sandra Phu) who work collaboratively with the MolDx technical assessment team in evaluating coverage applications. Dr. Hargrove and Ms. Phu review the MolDx technical assessment along with the complete LCD application materials and exercise independent judgment in making coverage determinations. However, as noted above, the MolDx technical assessment recommendation is highly influential in the LCD decision-making process.
Historical precedent strongly informs our risk assessment. Based on our review of publicly available MolDx technical assessments and LCD decisions from the 2019–2024 period, molecular diagnostic tests submitted to MolDx with inconclusive or below-threshold prospective validation data have frequently received non-coverage technical assessments. In particular, we are aware of at least three comparable molecular diagnostic tests in the oncology predictive biomarker category that were submitted with below-threshold prospective validation data during the 2021–2023 period and received non-coverage determinations, as discussed in greater detail in Section V below. The consistency of these adverse outcomes underscores the materiality of the evidence gaps identified in this memorandum.
C. Clinical Utility Evidence Requirements
The clinical utility evidentiary standard requires the test developer to demonstrate that test results change clinical management and that those changes lead to improved patient outcomes. This is the highest bar in the MolDx evidentiary hierarchy and is often the most challenging domain for emerging molecular diagnostic tests.
For OncoPrecise GXP-47, a demonstration of clinical utility would ideally show that treatment decisions guided by the TRI score — specifically, the selection of ICI therapy for high-TRI patients and platinum-based chemotherapy for low-TRI patients — lead to measurably better outcomes (including PFS, overall survival, quality of life, and reduced adverse events) compared to standard-of-care treatment selection without the TRI score. The gold standard for clinical utility evidence is a prospective, randomized controlled trial in which patients are randomized to a test-guided treatment arm versus a standard-of-care arm, with outcomes compared between groups.
GDT has retained Larkfield Health Economics, LLC to prepare the clinical utility and cost-effectiveness analyses for the MolDx Technical Assessment Dossier. Dr. Nathan Ostroff, Principal Consultant at Larkfield Health Economics (7910 Woodmont Avenue, Suite 610, Bethesda, Maryland 20814), is leading this engagement. Dr. Ostroff's team is preparing a decision-analytic model and cost-effectiveness analysis designed to demonstrate the value proposition of OncoPrecise GXP-47 to the Medicare program. However, these analyses are modeling-based and rely on assumptions derived from the clinical validity data. They do not constitute direct empirical evidence of clinical utility.
The persuasiveness of the clinical utility argument is directly and inextricably dependent on the strength of the underlying clinical validity data. If the clinical validity evidence — i.e., the CLARITY-LUNG results — demonstrates that the TRI score accurately predicts ICI response (PPV and NPV at or above threshold), then the clinical utility argument flows naturally: treatment decisions guided by an accurate predictive biomarker will, on average, lead to better outcomes than decisions made without such guidance. Conversely, if the clinical validity data shows below-threshold predictive performance, the clinical utility argument is undermined at its foundation, because directing treatment decisions based on an insufficiently accurate test may not improve — and could potentially worsen — patient outcomes.
This interdependency between clinical validity and clinical utility amplifies the impact of the CLARITY-LUNG interim results. The below-threshold PPV and NPV not only constitute direct evidence gaps in the clinical validity domain but also weaken the clinical utility case that Larkfield Health Economics is constructing.
IV. DETAILED ANALYSIS OF EVIDENCE GAPS
We have identified five distinct but interrelated evidence gaps arising from the CLARITY-LUNG interim analysis results and the current state of GDT's clinical evidence program. Each gap is analyzed below in terms of its magnitude, clinical significance, regulatory significance within the MolDx/LCD framework, and its individual and cumulative impact on the probability of a favorable coverage determination.
A. Gap 1: PPV Shortfall (0.68 vs. ≥ 0.75 Target)
Magnitude. The observed PPV of 0.68 at the interim analysis represents a shortfall of 0.07 in absolute terms from the pre-specified threshold of 0.75. In relative terms, this shortfall is 9.33% ((0.75 − 0.68) / 0.75 = 9.33%). While a shortfall of 0.07 might appear modest in absolute terms, it is significant in the context of clinical validation for a predictive biomarker test, where small differences in predictive performance can have meaningful clinical consequences across large patient populations.
Clinical significance. A PPV of 0.68 means that among patients classified by the TRI score as "High Likelihood of Response" (TRI ≥ 70) — the patient group for whom the test would recommend ICI therapy — approximately 32% would not achieve 12-month PFS on ICI therapy. In other words, roughly one in three patients whom the test identifies as likely ICI responders would, in fact, not experience the predicted therapeutic benefit. This rate of false positives raises legitimate concerns about the clinical reliability of the TRI score in its primary intended use: guiding the decision to initiate ICI therapy.
It is important to contextualize this finding within the broader clinical landscape. ICI therapies carry significant side-effect profiles, including immune-related adverse events ("irAEs") such as pneumonitis, colitis, hepatitis, and endocrinopathies, some of which can be severe or life-threatening. When a predictive biomarker test directs a patient toward ICI therapy, there is an implicit clinical promise that the patient is likely to benefit from that therapy — a promise that is not adequately fulfilled when 32% of positively classified patients do not achieve the predicted outcome. While a PPV of 0.68 is not trivially low in absolute terms and does indicate some predictive value, it falls short of the level of reliability that the CLARITY-LUNG protocol, and the MolDx evidentiary framework, expect for this class of test.
Regulatory significance for MolDx/CFI. The PPV shortfall suggests that OncoPrecise GXP-47 may not sufficiently distinguish ICI responders from non-responders in the high-TRI population, which is the test's primary claimed clinical decision point. MolDx assessors, led by Dr. Catherine Embry, will evaluate the PPV relative to both the pre-specified threshold and to the performance of comparable tests that have previously undergone MolDx assessment. Our review of published MolDx assessments for NGS-based and expression-based oncology panels in the predictive biomarker category shows that tests receiving favorable coverage determinations have typically demonstrated a PPV of at least 0.72 — and more commonly 0.75 or above — for their primary predictive claims. The observed PPV of 0.68 falls below even this informal floor observed among favorably assessed tests.
Assessment. The PPV shortfall, standing alone, would likely prompt an unfavorable or deferred coverage determination from CFI/RAS. While some degree of latitude may exist for a test that narrowly misses its pre-specified threshold — particularly if other elements of the evidence base are strong — the combination of the PPV shortfall with the other gaps identified in this section reduces the likelihood of such latitude being extended.
B. Gap 2: NPV Shortfall (0.71 vs. ≥ 0.80 Target)
Magnitude. The observed NPV of 0.71 at the interim analysis represents a shortfall of 0.09 in absolute terms from the pre-specified threshold of 0.80. In relative terms, this shortfall is 11.25% ((0.80 − 0.71) / 0.80 = 11.25%). Notably, the NPV shortfall is larger than the PPV shortfall in both absolute terms (0.09 vs. 0.07) and relative terms (11.25% vs. 9.33%). This asymmetry is significant because, as discussed below, the clinical and regulatory consequences of the NPV shortfall are more severe than those of the PPV shortfall.
Clinical significance. An NPV of 0.71 means that among patients classified by the TRI score as "Low Likelihood of Response" (TRI < 40) — the patient group for whom the test would recommend against ICI therapy and in favor of platinum-based chemotherapy — approximately 29% would in fact have achieved 12-month PFS on ICI therapy had they received it. In practical clinical terms, this means that nearly three in ten patients who are directed away from ICI therapy based on the TRI score would have benefited from immunotherapy.
This finding implicates potential patient harm. Unlike the PPV shortfall — where false positives result in patients receiving ICI therapy that may not provide the predicted benefit but does not necessarily cause affirmative harm — the NPV shortfall directly raises the concern that effective immunotherapy could be withheld from patients who would have responded to it. These patients, relying on the TRI score's classification, would instead receive platinum-based chemotherapy, which may be less effective for their particular disease biology. The downstream consequences could include shorter progression-free survival, earlier disease progression, reduced quality of life, and potentially shorter overall survival.
The patient safety dimension of the NPV shortfall is the most clinically troubling aspect of the CLARITY-LUNG interim results. In the clinical decision-making framework that OncoPrecise GXP-47 is designed to inform, a low-TRI classification carries a strong implicit recommendation against ICI therapy. Oncologists relying on the test may be less likely to prescribe ICI therapy for low-TRI patients, even when other clinical factors might support it. A false-negative rate of 29% in this context raises fundamental questions about whether the test's negative classification is reliable enough to influence treatment decisions.
Regulatory significance. Patient safety concerns arising from false negatives are weighted heavily in MolDx technical assessments. The MolDx program has consistently emphasized that predictive biomarker tests must demonstrate a sufficiently low false-negative rate to ensure that patients are not inappropriately denied access to effective therapies. An NPV gap of the magnitude observed in the CLARITY-LUNG interim analysis — 11.25% below the pre-specified threshold — would very likely be flagged by Dr. Catherine Embry's technical assessment team as a significant concern and would likely be highlighted by CFI Medical Director Dr. William Hargrove in any coverage deliberation.
In our experience, MolDx assessors are particularly attuned to the directional consequences of test errors. A false positive (PPV issue) results in a patient receiving a treatment that may not work; a false negative (NPV issue) results in a patient being denied a treatment that would have worked. The latter is generally viewed as the more serious error from a patient welfare standpoint, and MolDx technical assessments reflect this weighting.
Assessment. The NPV shortfall represents the more concerning of the two gaps and, in our judgment, would likely be the primary basis for a non-coverage determination if the LCD were filed based on current interim data. This gap, more than any other single factor, drives our overall assessment that the current evidence base is insufficient to support a filing with a reasonable probability of favorable coverage.
C. Gap 3: Interim (Not Final) Data Status
The CLARITY-LUNG results under consideration are from an interim analysis performed at the 50% enrollment milestone (n = 600 patients). They are not the final data readout from the fully enrolled study. This distinction has independent significance for MolDx and CFI evaluation, even apart from the below-threshold performance results.
MolDx and CMS generally do not issue favorable coverage determinations based solely on interim analysis data from an ongoing clinical trial. The rationale for this position is straightforward: interim analyses, by definition, represent incomplete datasets with lower statistical power than the fully enrolled study. Confidence intervals around point estimates of PPV and NPV are wider, and the risk of the estimates changing — in either direction — as additional data accrues is non-trivial. Coverage decisions based on interim data carry the risk that the final results may diverge materially from the interim estimates, potentially leaving the MAC in the position of having issued coverage based on data that was ultimately superseded by less favorable (or more favorable) final results.
The statistical power of the CLARITY-LUNG interim analysis, conducted at n = 600, is lower than the power that will be available at the final analysis with n = 1,200. The wider confidence intervals around the interim PPV and NPV estimates mean that the true population-level performance of the TRI score could fall within a broader range than the point estimates suggest. While the interim results provide a directional signal, they do not provide the same level of certainty as the final data readout.
The DSMB's characterization of the interim results as "inconclusive" — a word carefully chosen by a committee chaired by Dr. Evelyn Matsuda, who brings considerable experience and credibility in clinical trial oversight — further underscores the provisional nature of the data. The DSMB's decision to recommend trial continuation, while declining to recommend stopping for futility, reflects a judgment that the data do not foreclose the possibility of acceptable final results but also do not provide affirmative evidence of adequate performance. This "inconclusive" characterization would be difficult to distinguish from the MolDx assessment team's perspective and could be cited in a non-coverage determination as evidence that even the trial's own oversight body did not view the data as supportive.
Assessment. Even if the interim PPV and NPV had met the pre-specified thresholds, the interim status of the data would itself constitute an evidentiary limitation — though not necessarily a disqualifying one if the results were clearly strong and the other evidence elements were robust. In the actual scenario, where the interim results are both below threshold and characterized as "inconclusive" by the DSMB, the interim data status compounds the other evidence gaps rather than mitigating them.
D. Gap 4: Absence of Clinical Utility Data from Prospective Studies
The CLARITY-LUNG trial is a clinical validation (clinical validity) study designed to measure the PPV and NPV of the TRI score. It is not designed to measure clinical utility — i.e., whether treatment decisions guided by the TRI score lead to improved patient outcomes compared to treatment decisions made without the TRI score. The trial's blinded design, in which TRI scores are not disclosed to treating oncologists, precludes any assessment of how the test results would change clinical management in practice.
As a result, GDT does not currently have any prospective clinical utility data for OncoPrecise GXP-47. The available clinical utility evidence is limited to the following:
•  Decision-analytic modeling and cost-effectiveness analyses being prepared by Larkfield Health Economics, LLC, which use assumptions derived from the CLARITY-LUNG data and published literature to project the expected impact of TRI-guided treatment selection on patient outcomes and healthcare costs.
•  Indirect evidence from the published literature on the relationship between biomarker-guided treatment selection and outcomes in the ICI treatment setting, drawn from studies of other biomarkers (e.g., PD-L1 expression, tumor mutational burden) rather than the TRI score specifically.
While the MolDx framework does not categorically require prospective clinical utility data for every molecular diagnostic test — and modeling-based clinical utility arguments have been accepted in some MolDx assessments — the absence of direct prospective clinical utility data is a gap. This gap is particularly consequential when coupled with below-threshold clinical validity data. The logic is as follows: clinical utility modeling relies on assumptions about the test's predictive accuracy. When the empirical evidence of predictive accuracy (clinical validity) falls below the assumptions used in the model, the clinical utility argument is undermined. A MolDx assessor reviewing a clinical utility model calibrated to PPV = 0.75 and NPV = 0.80 would reasonably question whether the modeled outcomes are achievable when the observed PPV is 0.68 and the observed NPV is 0.71.
Assessment. The absence of prospective clinical utility data is a gap that exists independently of the CLARITY-LUNG interim results but is amplified by them. In a scenario where CLARITY-LUNG had met its pre-specified endpoints, the clinical utility modeling from Larkfield Health Economics might have been sufficient to satisfy MolDx assessors. In the current scenario, the clinical utility argument is weakened by the underlying clinical validity shortfalls, and the absence of direct prospective utility data leaves GDT without an alternative empirical basis for the utility claim.
E. Gap 5: Limited Published Peer-Reviewed Literature Specific to OncoPrecise GXP-47
The published peer-reviewed literature base specific to OncoPrecise GXP-47 is limited. As of the date of this memorandum, there is only one peer-reviewed publication specifically addressing the OncoPrecise GXP-47 test: the analytical validation study published in the Journal of Molecular Oncology Diagnostics, Volume 18, Issue 4, pages 412–429, December 2022.
The CLARITY-LUNG interim analysis results have not been submitted for publication in any peer-reviewed journal. They exist only as internal GDT data, communicated to the DSMB and to GDT's management team and Board of Directors. There is no citable, publicly available reference for the clinical validation data.
The draft MolDx Technical Assessment Dossier currently under preparation includes a bibliography of 43 peer-reviewed publications. However, our review of this bibliography reveals that the substantial majority of these publications relate to background literature on ICI therapy in NSCLC, biomarker science generally, gene expression profiling methodologies, and comparable predictive biomarker tests — not to OncoPrecise GXP-47 specifically. While this background literature provides important context and scientific foundation, it does not substitute for test-specific published evidence in the MolDx assessment framework.
MolDx assessors give greatest weight to test-specific published evidence — that is, peer-reviewed publications reporting the results of studies conducted with the specific test under review, using the specific assay platform, algorithm, and scoring methodology that the test employs in clinical practice. Background literature and evidence from comparable tests are considered supportive but cannot carry the clinical validity or clinical utility argument on their own.
Assessment. The limited test-specific publication base is a gap that pre-dated the CLARITY-LUNG interim analysis and would have been partially addressed by the planned publication of the final CLARITY-LUNG results. In the current situation, the gap persists and is compounded by the below-threshold interim results. Publication of the CLARITY-LUNG interim data — even with results below the pre-specified thresholds — in a peer-reviewed journal would strengthen the evidence base by adding a citable reference and demonstrating scientific transparency. However, publication alone would not cure the performance shortfalls reflected in the data. As discussed in Section VI.C below, we recommend pursuing publication as a matter of scientific integrity and strategic positioning, while recognizing its limitations as an evidence gap remedy.
V. RISK ASSESSMENT: PROBABILITY OF NON-COVERAGE DETERMINATION
A. Likelihood Assessment if LCD Filed Based on Current Evidence
Drawing on the five evidence gaps analyzed in Section IV, our assessment of precedent from comparable MolDx/LCD proceedings, and our knowledge of the evidentiary expectations of the MolDx program and CFI, we have assessed the probability of a non-coverage determination under three alternative filing scenarios.
Scenario 1 — File LCD Application with Interim Data as Currently Available.
Under this scenario, GDT would proceed with the original August 2024 LCD filing plan, submitting the LCD application to CFI and the MolDx Technical Assessment Dossier to RAS based on the CLARITY-LUNG interim analysis results, supplemented by the analytical validation publication, background literature, and Larkfield Health Economics' HEOR analysis.
Assessed probability of non-coverage determination: approximately 65–75%.
This assessment reflects the cumulative weight of the five evidence gaps identified in Section IV: below-threshold PPV (Gap 1), below-threshold NPV (Gap 2), interim-only data status (Gap 3), absence of prospective clinical utility data (Gap 4), and limited test-specific published literature (Gap 5). The DSMB's "inconclusive" characterization would further undermine the strength of the submission.
Our assessment is informed by a review of four comparable molecular diagnostic tests that filed MolDx dossiers with below-threshold clinical validation data during the 2021–2023 period. Of these four:
•  Three received non-coverage determinations from their respective MACs following unfavorable MolDx technical assessments.
•  One received a provisional recommendation for Coverage with Evidence Development ("CED"), which required the test developer to conduct additional clinical studies as a condition of continued coverage. However, this test ultimately did not receive a final favorable LCD; the CED pathway imposed significant additional costs and operational complexity, and the test developer ultimately withdrew its application.
None of these four comparable tests received an outright favorable coverage determination. While each test's circumstances were unique and direct comparisons should be made cautiously, the pattern is clear: below-threshold prospective clinical validation data presents a very difficult evidentiary profile for MolDx and MAC review.
Scenario 2 — File LCD Application with Interim Data Plus Supplemental Evidence.
Under this scenario, GDT would pursue the supplemental evidence strategies described in Section VI of this memorandum prior to filing — including a retrospective validation sub-study, pre-specified sub-group analyses, publication of interim results, and external expert clinical utility assessments — and would incorporate this supplemental evidence into the LCD application and MolDx dossier.
Assessed probability of non-coverage determination: approximately 50–60%.
The supplemental evidence strategies could modestly reduce the risk of a non-coverage determination by broadening the evidence base, demonstrating scientific rigor and transparency, and potentially identifying patient sub-groups in which the TRI score performs at or above threshold. However, supplemental evidence is unlikely to fully compensate for below-threshold primary endpoint results from the pivotal prospective clinical validation study. MolDx assessors will appropriately focus on the CLARITY-LUNG results as the primary evidence of clinical validity, and supplemental analyses will be evaluated in that context. The risk reduction from supplemental evidence, while meaningful, is insufficient to bring the probability of favorable coverage above 50%.
Scenario 3 — File LCD Application After Final CLARITY-LUNG Data (Q2 2025), Assuming Final Data Meets or Approaches Thresholds.
Under this scenario, GDT would defer the LCD filing until after the final CLARITY-LUNG data readout, now projected for Q2 2025. This scenario assumes that the final data, based on the full 1,200-patient enrollment, meets or closely approaches the pre-specified PPV and NPV thresholds.
Assessed probability of favorable coverage: approximately 60–70% (highly dependent on final results).
If the final CLARITY-LUNG results demonstrate PPV ≥ 0.75 and NPV ≥ 0.80, the evidence profile would be substantially stronger, and the probability of favorable coverage would increase significantly — potentially to the 70–80% range. If the final results improve from the interim levels but still fall short of the pre-specified thresholds, the probability would be lower but still materially better than under Scenarios 1 or 2, because the final data would carry greater statistical power and would not bear the "interim" and "inconclusive" characterizations.
We emphasize that this scenario is contingent on assumptions about the final CLARITY-LUNG results that cannot be known at this time. The interim results provide a directional signal but do not predict the final results with certainty. It is possible that the final results could improve, remain similar to the interim estimates, or deteriorate as additional patient data accrues.
B. Consequences of a Non-Coverage Determination
The consequences of an adverse LCD from CFI would extend well beyond the immediate denial of Medicare coverage in Jurisdiction E. We have identified the following specific categories of downstream impact:
Negative Precedent. An adverse LCD from CFI would become a published, public record in the CMS Medicare Coverage Database. Non-coverage determinations create negative precedent that may influence:
(1) Other MACs, if GDT subsequently sought coverage in other Medicare jurisdictions. While MACs are not formally bound by the coverage decisions of other MACs, published non-coverage LCDs are routinely reviewed and considered by other MACs as part of their own coverage evaluation processes. An adverse LCD from CFI — particularly one based on a negative MolDx technical assessment — would create a presumption against coverage that GDT would need to overcome in any subsequent jurisdiction.
(2) Commercial payers, who frequently follow Medicare coverage decisions as a benchmark for their own coverage policies. A Medicare non-coverage determination would likely delay or prevent commercial payer coverage of OncoPrecise GXP-47, further reducing the total addressable market and revenue projections.
(3) Any future NCD proceeding. If GDT ultimately pursues a National Coverage Determination through CMS, the existence of a prior adverse LCD would be part of the administrative record and could be cited by CMS staff or public commenters as evidence against national coverage.
Procedural Hurdle for Future Applications. A non-coverage determination would need to be disclosed in connection with any future LCD reconsideration request — even after final CLARITY-LUNG data becomes available that meets the pre-specified thresholds. The reconsideration process is procedurally distinct from a de novo LCD application and imposes additional requirements, including a demonstration that new evidence is available that was not previously considered. While a favorable final CLARITY-LUNG dataset would likely satisfy this requirement, the reconsideration process adds procedural complexity, timeline uncertainty, and administrative burden that would not exist if GDT filed a de novo LCD application based on strong initial evidence.
Financial Impact. An adverse LCD would delay or eliminate the projected Year 1 Medicare revenue of $20,290,500. Given that Medicare revenue represents approximately 58.8% of GDT's total projected Year 1 revenue of $34,502,500, the financial impact would be material. The absence of Medicare coverage would also reduce commercial ordering volume, as many oncologists are reluctant to order tests that lack Medicare coverage, both because of reimbursement uncertainty for their Medicare patients and because Medicare coverage is often viewed as a proxy for clinical evidence strength.
Impact on PCNB Milestone Covenant. GDT's credit facility with Pacific Coast National Bank includes a milestone covenant (Section 7.14 of Credit Agreement Ref. PCNB-GDT-RF-2023-0922) requiring GDT to "obtain a final Medicare coverage determination" by December 31, 2024. A formal non-coverage LCD — while technically a "final Medicare coverage determination" — would not satisfy the covenant's intended purpose, which is to demonstrate commercial and reimbursement readiness. Moreover, a formal non-coverage LCD could be worse for covenant compliance purposes than no filing at all: the covenant language and its interpretation by PCNB's counsel may distinguish between "no coverage determination yet obtained" (a matter of timing) and "coverage determination obtained, and it is adverse" (a substantive negative outcome). Calderwood & Harkness recommends that GDT coordinate with Whitford & Sable LLP (Harrison Whitford, lead counsel for corporate and financing matters) regarding the interpretation and potential amendment or waiver of this covenant in light of the revised LCD filing timeline.
Impact on Investor Relationships. Pinnacle Health Ventures, L.P. ("PHV"), GDT's largest equity holder at 41.3%, has board representation through David Ashford and Elena Marchetti and holds observer rights over material commercialization and reimbursement decisions. An adverse LCD could affect PHV's assessment of GDT's portfolio value, potentially triggering enhanced oversight, revised valuation adjustments, or reduced appetite for follow-on investment. PHV's observer rights may require advance notification of the LCD filing decision, and the Board should consider the timing and content of any such notification.
Additionally, Verdant Life Sciences Capital, LLC (Managing Director Christine Lau) has incorporated OncoPrecise GXP-47 Medicare reimbursement revenue into GDT's forward financial projections for potential Series D financing or IPO planning purposes. An adverse LCD, or even a material delay in LCD filing, could affect Verdant's financial models and timeline assumptions, with downstream effects on GDT's capital markets strategy.
VI. SUPPLEMENTAL EVIDENCE STRATEGIES
We have identified five supplemental evidence strategies that GDT could pursue, individually or in combination, to partially address the evidence gaps identified in Section IV. Each strategy is analyzed below in terms of its description, potential contribution, limitations, estimated timeline and cost, and overall assessment. We emphasize at the outset that none of these strategies — individually or collectively — is likely to fully cure the primary evidence gaps arising from the below-threshold CLARITY-LUNG interim results. These strategies should be understood as components of a broader evidence-building program, not as substitutes for final CLARITY-LUNG data meeting or approaching the pre-specified performance thresholds.
A. Strategy 1: Retrospective Validation Sub-Study
Description. GDT could design and conduct a retrospective analysis using archived FFPE NSCLC tissue samples and linked clinical outcomes data from one or more of the 14 CLARITY-LUNG participating clinical sites. Under this approach, archived FFPE specimens from previously treated NSCLC patients would be retrieved, subjected to OncoPrecise GXP-47 testing, and the resulting TRI scores would be correlated with known clinical outcomes (12-month PFS on ICI therapy) to generate retrospective estimates of PPV and NPV.
Potential contribution. A retrospective validation sub-study could provide additional PPV and NPV estimates from a separate (retrospective) patient cohort, potentially with more favorable results. Because the retrospective cohort would be distinct from the prospective CLARITY-LUNG cohort, favorable results from the retrospective study would expand the evidence base and demonstrate consistency (or improvement) of TRI score performance across different patient populations and clinical settings. The retrospective study design also allows for control over specimen selection criteria, which could yield a more homogeneous patient population and potentially better test performance.
Limitations. Retrospective data is weighted less heavily than prospective data in MolDx technical assessments. MolDx assessors are well aware that retrospective studies are susceptible to selection bias — including the potential for cherry-picking specimens or sites with favorable outcomes — as well as ascertainment bias and variability in specimen handling and storage conditions. A retrospective sub-study would be unlikely to be considered sufficient to overcome below-threshold prospective interim results from the CLARITY-LUNG trial, which is the pivotal study in the evidence hierarchy. MolDx assessors would appropriately view retrospective data as supplementary, not primary.
Additionally, the logistics of a retrospective sub-study — including institutional review board ("IRB") approvals at participating sites, specimen retrieval and quality assessment, outcomes data extraction, and biostatistical analysis — would add time and cost to the evidence development program.
Estimated timeline. Approximately 4–6 months to design, execute, and analyze, assuming cooperation from participating clinical sites and availability of suitable archived specimens.
Estimated cost. Approximately $1.5 million to $2.5 million, which is not included in the current CLARITY-LUNG trial budget of $28,500,000 and would need to be funded separately.
Assessment. A retrospective validation sub-study may modestly improve the overall evidence profile but would not close the primary gaps. It would be most valuable as supporting evidence filed alongside final CLARITY-LUNG data — that is, as part of a comprehensive evidence package submitted after the final data readout — rather than as standalone evidence for an earlier LCD filing.
B. Strategy 2: Pre-Specified Sub-Group Analyses from CLARITY-LUNG Interim Data
Description. GDT could explore whether the TRI score performs better in specific patient sub-groups within the CLARITY-LUNG interim dataset (n = 600). Potential sub-groups of interest include stratification by histologic subtype (adenocarcinoma versus squamous cell carcinoma), PD-L1 expression level (e.g., PD-L1 ≥ 50% versus PD-L1 < 50%), line of ICI therapy, ECOG performance status, or other clinically relevant patient characteristics.
Potential contribution. If PPV and/or NPV meet or exceed the pre-specified thresholds in a clinically meaningful and well-defined patient sub-group, GDT could potentially pursue a narrower clinical indication for the initial LCD filing. For example, if the TRI score demonstrates PPV ≥ 0.75 and NPV ≥ 0.80 in patients with adenocarcinoma histology and PD-L1 ≥ 50%, GDT could seek an LCD covering only that sub-population. A narrower indication supported by above-threshold performance data could present a more compelling evidence profile to MolDx and CFI, even if the overall study population results are below threshold.
Limitations. Sub-group analyses from an interim dataset with n = 600 would have limited statistical power, particularly for smaller sub-groups. The smaller the sub-group, the wider the confidence intervals around PPV and NPV estimates, and the less persuasive the results. Post-hoc sub-group analyses — that is, analyses not pre-specified in the study protocol — are viewed with skepticism by MolDx assessors and the broader regulatory and clinical trial community, because they are susceptible to data dredging and the identification of spurious associations. MolDx would give substantially greater weight to sub-group analyses that were pre-specified in the CLARITY-LUNG protocol.
A narrower indication would also reduce the addressable patient population and, consequently, the projected test volume and revenue. The financial impact of narrowing the indication would need to be modeled and evaluated against the improved probability of coverage for the narrower population.
Estimated timeline. Approximately 6–8 weeks if the relevant sub-group analyses were pre-specified in the CLARITY-LUNG study protocol. If protocol amendment is required to formalize sub-group analyses that were not originally pre-specified, the timeline would be longer and the analyses would carry less evidentiary weight.
Assessment. This strategy is worthy of exploration and, in our view, should be pursued regardless of the LCD filing decision. We recommend that Dr. Sundaram review the CLARITY-LUNG study protocol to confirm whether relevant sub-group analyses were pre-specified. If pre-specified sub-group analyses exist, they should be conducted and reported as part of the ongoing clinical development program. Even if the results do not support a narrower-indication LCD filing, the sub-group data would be valuable for understanding the TRI score's performance characteristics and for informing future study design or algorithm refinement.
C. Strategy 3: Publication of CLARITY-LUNG Interim Results
Description. GDT could prepare a manuscript reporting the CLARITY-LUNG interim analysis results and submit it for publication in a peer-reviewed oncology or molecular diagnostics journal. The manuscript would present the study design, patient population, interim results (including the below-threshold PPV and NPV), DSMB recommendations, and a discussion of the implications for ongoing clinical development.
Potential contribution. Publication would add to the test-specific published literature base, which is currently limited to a single analytical validation paper. It would demonstrate scientific transparency — a value that MolDx assessors and the broader clinical community regard highly. A published interim analysis would also establish a citable reference that could be included in the MolDx Technical Assessment Dossier bibliography, replacing the current reliance on internal, unpublished data.
Limitations. Publication does not change the underlying data. The published interim results would still show PPV = 0.68 and NPV = 0.71, below the pre-specified thresholds. In fact, publication of below-threshold results could be cited by MolDx assessors as affirmative evidence against coverage — the assessors could point to the published results and note that the test's own pivotal study found performance below acceptable levels.
However, non-publication carries its own risks. If MolDx assessors become aware of unpublished interim data during the technical assessment process — which is likely, given that the CLARITY-LUNG trial is registered on ClinicalTrials.gov and interim analysis milestones are part of the public record — the failure to publish the results could be perceived as an attempt to suppress unfavorable data. This perception would be damaging to GDT's credibility and could negatively influence the MolDx assessment.
Estimated timeline. Approximately 4–8 months from manuscript submission to publi