Minimal Residual Disease (MRD) Negativity as an FDA Surrogate Endpoint

FDA Regulatory Guidance and the April 2024 ODAC Multiple Myeloma Precedent

FDA's regulatory framework is codified in the January 2020 Guidance: Hematologic Malignancies: Regulatory Considerations for Use of Minimal Residual Disease in Development of Drug and Biological Products for Treatment. That guidance separates prognostic biomarker labeling from surrogate or intermediate clinical endpoints supporting accelerated approval under 21 CFR 314.510 and 21 CFR 601.41.

While durable MRD-negative remission previously supported accelerated approval in acute lymphoblastic leukemia (ALL), multiple myeloma relied on Overall Response Rate (ORR) or PFS until April 12, 2024, when FDA's Oncologic Drugs Advisory Committee (ODAC) voted 12–0 that MRD-negative Complete Response (MRD-negative CR) is acceptable as an intermediate endpoint reasonably likely to predict clinical benefit for accelerated approval in multiple myeloma (FDA April 12, 2024 ODAC Materials).

Both the pooled trial meta-analyses and FDA's briefing book established strict guardrails around that precedent:

• Individual Prognosis vs Trial-Level Surrogacy: Proving that MRD-negative patients live longer than MRD-positive patients (patient-level odds ratios) establishes prognosis, not surrogacy. Regulatory surrogacy requires trial-level correlation (R^2_trial)—showing across randomized studies that treatment effects on MRD odds ratios reliably predict treatment effects on PFS and OS hazard ratios.
• Composite MRD-Negative ≥CR Standard: Because serum monoclonal paraprotein clears slowly and extramedullary plasmacytomas can persist outside the iliac crest biopsy needle, ODAC endorsed MRD negativity at 10^-5 or deeper specifically within patients achieving ≥Complete Response at a prespecified 9- or 12-month landmark.
• Randomized Trial Architecture: MRD accelerated approval is intended primarily for randomized controlled trials that continue controlled follow-up to confirm clinical benefit on PFS and monitor OS—not to replace randomized trials with small single-arm cohorts when therapies exist.

Assay Sensitivity Denominators: 10^-4 vs 10^-5 vs 10^-6 Across NGS and Flow

An MRD cutoff of 10^-5 detects 1 malignant cell among 100,000 nucleated cells (0.001%), whereas 10^-6 detects 1 malignant cell among 1,000,000 nucleated cells (0.0001%). Cross-trial comparisons fail unless analysts verify the underlying platform registered on ClinicalTrials.gov:

1. Next-Generation Sequencing (NGS) of Ig/TCR Clonal Rearrangements: FDA-cleared central diagnostic assays track patient-specific V(D)J sequences identified from a baseline diagnostic sample (~90%–95% baseline clone trackability in myeloma and B-ALL). Validated 10^-6 sensitivity requires extracting sufficient genomic DNA (≥1.5–2.0 × 10^6 cell equivalents) from the aspirate.
2. Multiparameter / Next-Generation Flow Cytometry (MFC / EuroFlow NGF): Detects aberrant cell-surface antigen phenotypes on viable cells without needing a baseline tumor sample. Standardized 8- to 10-color EuroFlow panels achieve 10^-5 sensitivity when acquiring ≥5 × 10^6 cells, but require fresh sample processing within 24–48 hours to prevent cell lysis across global sites.

FDA's 2020 Guidance warns specifically against first-pull bone-marrow hemodilution: when clinicians reserve the first marrow pull for local morphology and send later pulls to the central MRD lab, peripheral blood dilutes marrow cellularity and manufactures false-negative MRD calls. Likewise, FDA companion-diagnostic reviewers verify that a claimed 10^-5 cutoff reflects the assay's analytically validated Limit of Quantification (LoQ) or Limit of Detection (LoD) at 95% confidence—not merely a theoretical Limit of Blank.

The Missing-Sample Denominator Trap: ITT vs Evaluable-Only MRD Rates

The largest numerical gap between early conference decks and FDA package inserts sits in the MRD denominator. Under FDA review conventions, landmark MRD negativity (e.g., Month 12 post-randomization) must be evaluated across the full Intention-to-Treat (ITT) randomized denominator. Any randomized participant who:

• progresses, dies, or discontinues prior to Month 12,
• misses the scheduled bone-marrow biopsy,
• lacks a trackable baseline NGS clonotype,
• yields a clotted, hemodiluted, or insufficient-DNA sample, or
• clears marrow MRD without confirmatory immunofixation and required PET-CT imaging,

is counted strictly as MRD-positive (non-responder)!

Illustrative denominator reconciliation (hypothetical counts, not company data): Observe how one 200-patient arm yields three different headline rates:

Reporting Population (Illustrative N = 200 Arm)	Numerator (Confirmed 10^-5 MRD-Negative CR at Month 12)	Denominator Definition	Headline MRD Rate	FDA Primary Validity
Evaluable-Sample Subset	96 patients	145 patients with valid Month-12 DNA + trackable clone	66.2%	Rejected for primary efficacy; drops 55 missing/discontinued failures
Best-Overall ITT MRD (Any Response)	112 patients	200 randomized (any visit, including VGPR with residual M-spike)	56.0%	Exploratory/secondary; mixes transient clearance & non-CR
FDA Landmark ITT MRD-Negative ≥CR	96 patients	200 randomized (strict Month-12 window + confirmed ≥CR)	48.0%	Primary ODAC-aligned regulatory benchmark

Whenever a press release cites an MRD percentage without disclosing both the ITT randomized count and missing/untrackable aspirate totals, the FDA-equivalent MRD rate remains UNKNOWN.

Sustained MRD Negativity and Confirmatory PFS/OS Conversion Risk

Single-timepoint tumor clearance does not guarantee durable remission; early marrow debulking at Month 6 can relapse by Month 15 via antigen escape or T-cell exhaustion. Consequently, International Myeloma Working Group (IMWG) and FDA reviewers emphasize sustained MRD negativity—two consecutive MRD-negative assessments ≥6 or ≥12 months apart without intervening progression.

Investors must also price confirmatory conversion risk:

• Single-Trial Dual-Primary Design: Splitting alpha between Month-12 ITT MRD-negative CR (for accelerated filing) and event-driven PFS (for regular conversion) in one fully enrolled RCT avoids post-approval enrollment stalls.
• Fatal Infection Offset: Intensive bispecific or CAR-T combinations can double ITT MRD-negative CR rates while elevating fatal Grade 5 opportunistic infections from 1% to 7%. Because every death counts as a PFS and OS event, treatment-related mortality can erase survival hazard-ratio gains despite deep marrow clearance (see our PFS vs OS oncology guide).

Track hematology panels on the AdCom Calendar, action dates on the PDUFA Calendar, and confirmatory readouts on the Phase 3 Readouts Calendar.