Commutable patient-matrix reference material panel for validating tumor-informed ctDNA MRD assays (clinical QC material, pivot P2) — the story so far

How to read this. This is a synthesis, not a new source of truth: every claim it cites carries the tag the candidate file gives it right now, copied not paraphrased, and every line quoted from a deck is quoted verbatim. See docs/narrative-spec.md.

What changed since the 2026-09-15 build. The previous narrative was built against the P0 the original shape shape: an RUO (research use only) reagent for assay development, with a 46-claim file and five verified claims. The candidate is now at pivot P2 clinical QC material at launch. Its intended_use: is a quality control material for molecular laboratories to monitor the analytical performance of tumor-informed ctDNA (circulating tumour DNA) MRD (molecular residual disease) assays "during assay validation and in routine run-by-run quality control". Its launch designation is Class II (FDA device class II) under 21 CFR (Title 21 of the Code of Federal Regulations) 866.5910 alongside product code SCU (subcutaneous). Its mechanism now value-assigns levels below 0.1% VAF (variant allele frequency) by gravimetric dilution and labels them "as derived by dilution rather than measured", and it is supplied as a multi-level panel plus a daily-use single-level control. This narrative was rebuilt from the file as it stands, not patched.

Sources read for this build, and how current each is.

• The candidate file, updated: 2026-09-26: 192 claims, 32 verified, 159 unconfirmed, 1 refuted, 0 unverified. Its Verified Composite and Gate Check are dated 2026-09-26 and reason from the P2 clinical QC material at launch shape. Its ## Wrap section is not current. It still describes the product as "a laboratory research product, not a medical device", counts "5 of 46 claims verified", and names a cheapest next test (restoring the CMS (Centers for Medicare & Medicaid Services) caches) that the Verified Composite shows has since been run. This narrative reads the Wrap as a record of the P0 the original shape packet, not as the present state.
• One dossier, ctdna-mrd-assay-validation (built 2026-08-27, refreshed 2026-09-25). Its opening adaptation note still frames the candidate as RUO. It carries no confidence tags, and it is cited below only for lines that have not become claims.
• No seeds. The candidate's frontmatter lists no seeds:, so this narrative names no "why now" seed context.
• Three decks, all rebuilt on 2026-09-26 against P2 clinical QC material at launch. **None of the three passes check_decks.py today.** Their stamps match the candidate, so they are not stamped STALE. But each cites claims 179-189 with a tag of unverified, and the Verifier has since tagged all fourteen of claims 179-192 unconfirmed. The details are in the lens section below.

The one-paragraph version

This is a set of reference samples for laboratories running tumor-informed ctDNA minimal residual disease tests. It is built on pooled human donor plasma, and characterized tumor-derived fragments are spiked in at very low allele fractions. Under pivot P2 clinical QC material at launch it is sold as a Class II clinical QC (quality control) material under an FDA (US Food and Drug Administration) product code verified to exist as 510(k) (premarket notification) exempt claim 131verified. That route carries a verified quality-system burden before first distribution claim 159verified. Whether the exemption applies to this material is still an open FDA judgment claim 52unconfirmed claim 157unconfirmed. The measurement anchor P2 clinical QC material at launch rests on is verified at and above 0.1% VAF claim 100verified. The levels below that, which are the product's reason to exist, are derived by dilution, and nobody has shown they can carry the uncertainty budget P2 clinical QC material at launch sets for them claim 134unconfirmed. The file clears Stage 5 at 7/9 verified factors with TAM (total addressable market) sitting exactly on its floor. Every condition the three decks name for the business itself (who would use the daily vial, what it earns, and whether the derived levels hold) rests on unconfirmed claims.

The story so far

The problem. A laboratory introducing a test that is not FDA-cleared has to establish analytical sensitivity before it reports patient results claim 76verified. For molecular amplification procedures, CLIA (Clinical Laboratory Improvement Amendments) requires two control materials at least once on each day patient specimens are assayed claim 7verified. Tumor-informed MRD assays claim limits of detection far below where commercial reference material is characterized. Two examples on record are an LOD95 of 3.45 ppm claim 36unconfirmed and 0.008% VAF at 60 ng input claim 37unconfirmed. The one verified figure of this kind is 0.7 ppm for a B-cell malignancy assay at 120 ng input claim 127verified. The file's premise is that laboratories build their own contrived controls and find them inadequate below 0.01% VAF, and that premise is unchecked claim 23unconfirmed. The dossier looked for a published record of a laboratory switching away from a named commercial material and giving a reason, and records it as not found [dossier: ctdna-mrd-assay-validation].

What incumbents sell. Commercial contrived ctDNA material at 0.1%-5.0% VAF is verified to exist, and concordance across nine laboratories falls at 0.1% VAF claim 94verified. The incumbent product lines on record stop at 0.1% in a "plasma like matrix" claim 45unconfirmed or a synthetic plasma matrix claim 46unconfirmed. The one incumbent product found below 0.1% is a buffer-format MRD panel mix, four vials of purified DNA in Tris/EDTA at nominal levels down to 0.005% claim 111unconfirmed. It reports dPCR (digital polymerase chain reaction) for one blend only and states no uncertainty claim 101unconfirmed. The file's white-space claim is that nothing commercially available supplies a characterized sub-0.1% level in a human plasma matrix claim 184unconfirmed.

How the shape got here. The P0 the original shape launch was an RUO reagent. The rationale was that RUO labelling sat outside premarket review, and that half of claim 1unconfirmed was never verified claim 1unconfirmed. The file records that Regulatory blocked the RUO framing (IR-23 Regulatory blocks the P0 launch framing in the loop record) and that pivot P2 clinical QC material at launch changed intended_use:. The regulatory home is now FDA product code SCU, "Dna And Rna Oncology Controls", Class II under 866.5910, 510(k) exempt claim 131verified. The older 862.1660 Class I route claim 4verified claim 5verified and FDA's NZB code under 866.5910 claim 22verified still stand in the file as fallback routes. The P0 the original shape feasibility problem was the Poisson limit at the lowest levels claim 16unconfirmed. P1 derived-value bottom levels and P2 clinical QC material at launch answer it by design rather than by measurement. Levels at or above 0.1% are dPCR-quantified. Levels below 0.1% are value-assigned by gravimetric dilution into a dPCR-quantified diluent and labelled as derived claim 134unconfirmed.

Why that design, and why it is unproven. The dPCR anchor is the best-evidenced part of the file. In an interlaboratory study, ddPCR agreed with the gravimetric value from 53.9% down to 0.1%, with an LoQ of 0.1% claim 100verified. A dPCR primary reference procedure reported an LoQ of 0.2% claim 99verified. The Verified Composite reads those two claims together: this is why sub-0.1% levels "can be derived but not independently measured". The dossier reports that no national metrology institute or vendor value assignment below about 0.1% VAF with a stated expanded uncertainty was found, and that spiking into native pooled donor plasma with a dPCR uncertainty budget was not found as a published process [dossier: ctdna-mrd-assay-validation]. So the step below the anchor has no published precedent in the file. It is a bench question.

What the matrix buys, and what it risks. The commutability case is mixed on the verified record. The authors of a six-material comparison concluded that similarity to native cfDNA (cell-free DNA) aids protocol development, but that material with well-defined variants is preferable for sensitivity and precision claim 90verified. The verified hazards all point at the matrix premise. Some extraction kits preferentially recover spiked mononucleosomal DNA over endogenous plasma cfDNA claim 92verified. Spiking into fetal bovine serum, and the choice of digestion enzyme, change dPCR output claim 110verified. Fragmentation improved commutability for one viral standard and not another claim 93verified. The dossier found no head-to-head published comparison of native-plasma and synthetic-matrix material on the same assay [dossier: ctdna-mrd-assay-validation].

What P2 clinical QC material at launch added to the business. The daily-use control gives the product a recurring leg. The file carries bottom-up ranges: 2-400 panels a year claim 150unconfirmed, 500-5,000 vials a year claim 151unconfirmed, and $1,500-$3,000 per panel and $400-$900 per vial claim 149unconfirmed. At its own planning points that comes to about $875,000 a year, and whether that exceeds fully loaded cost plus the recurring 866.5910 obligations is unchecked claim 178unconfirmed. The older top-down sizing rested on more than 100,000 Medicare claims for 0340U, and that figure is refuted claim 10refuted. The fresh count is 25,746 Medicare fee-for-service services in CY2024 claim 68verified, at a CY2026 national limitation amount of $3,590.00 claim 71verified. Those are floors on the customers' test revenue, not on this product's market.

What has been checked, and what has not. Twenty-seven more claims came back verified since the P0 the original shape narrative, and one came back refuted. The verified set is regulatory text, CMS records, patent bibliography and published findings. Nothing verified touches a customer's behaviour, this product's price, a bench result on this material, or the reading of any patent's claims. The file's own Stage 5 disposition is cleared at 7/9 factors. It says so plainly: "Cleared is not an approval", and TAM is "not killed on TAM yet", not established.

What's solid

Only claims tagged verified. Each is stated at the scope the file gives it. Several verify that something exists, never that it applies.

• FDA product code SCU, "Dna And Rna Oncology Controls", is Class II under 21 CFR 866.5910 and 510(k) exempt. It is defined as a QC material for NGS (next-generation sequencing)-based IVD (in vitro diagnostic) test systems, including synthetic DNA and RNA controls claim 131verified.
• FDA classifies a genetics/DNA quality control material as Class II under 866.5910, product code NZB claim 22verified.
• 21 CFR 862.1660 defines quality control material as a Class I device claim 4verified. It is 510(k) exempt subject to the limitations of 862.9 claim 5verified, and product codes exist under it claim 6verified.
• SCU, NZB and QIG are not GMP-exempt. 21 CFR 820.10 requires a documented quality management system before distribution claim 159verified.
• CLIA requires control procedures over the complete analytic process, which drives recurring control use claim 7verified. A laboratory introducing a test not subject to FDA clearance must establish analytical sensitivity, and the section names no required material, source or traceability claim 76verified. The only nearby traceability language applies to calibrators, and only "if possible" claim 77verified.
• The March 2025 vacatur of FDA's 2024 LDT (laboratory-developed test) rule stood as of FDA's own 2025-09-19 Federal Register notice claim 66verified.
• CPT (Current Procedural Terminology) 0340U had 25,746 national Medicare fee-for-service services in CY2024, not more than 100,000 claim 68verified. Its CY2026 national limitation amount is $3,590.00 claim 71verified. Four active MolDX billing articles list it claim 72verified claim 73verified. That establishes listing, not coverage.
• Two further MRD PLA codes carry $1,800.00 and $1,943.21, and a third is contractor-priced claim 74verified. Their Medicare volumes are 4,481 and 243 services, and the third has no utilization row claim 75verified.
• ddPCR agreed with the gravimetric value from 53.9% down to 0.1%, with LoB/LoD/LoQ of 0.01%/0.02%/0.1% claim 100verified. A dPCR primary reference procedure reported a false-positive rate of 0-0.003% and an LoQ of 0.2% claim 99verified.
• MNase digestion of cell-line nucleosomes gave fragment peaks at 128 bp and 143 bp claim 107verified. Enzymatic digestion gave fragments of about 154-173 bp claim 109verified.
• Commercial contrived ctDNA materials at 0.1%-5.0% VAF showed high concordance above 0.1% and lower concordance at 0.1% across nine laboratories claim 94verified.
• Similarity to native cfDNA aids protocol development, but material with well-defined variants is preferable for sensitivity and precision, per the authors of a six-material study claim 90verified.
• Verified hazards to commutability: extraction kits that preferentially recover spiked mononucleosomal DNA claim 92verified, matrix and digestion effects on dPCR output (in fetal bovine serum) claim 110verified, and a mixed viral-standard analogy claim 93verified.
• A PhasED-Seq assay reported a 95% LoD of 0.7 ppm at 120 ng of input claim 127verified.
• Three US patents in this product's space are live on naive term to 2034, 2036 and 2038, verified on title and dates only claim 102verified claim 103verified claim 104verified.
• The file's capital estimate names no production volume claim 58verified and no supplier claim 59verified. It does not itemize the donor infectious-disease screening that plasma used to manufacture a device requires claim 61verified.

What's still open

Refuted, stated plainly.

• More than 100,000 annual Medicare claims for CPT 0340U. This was the base of the Stage 2 TAM sizing, and the record shows 25,746 claim 10refuted.

Whether the regulatory route actually applies.

• Whether a QC material for MRD assays is "intended for use in the diagnosis, monitoring, or screening of neoplastic diseases", which would withdraw the exemption on both candidate regulations claim 52unconfirmed.
• Whether 866.5910's cystic-fibrosis generic type is read as a different intended use under 866.9(a) claim 53unconfirmed, and whether "assay validation" falls within 866.5910(a) claim 156unconfirmed.
• Whether a native-plasma matrix is a "reasonably foreseeable characteristic" when 866.5910(a) names only recombinant, synthetic and cell-line controls claim 157unconfirmed.
• What the 2007 special-controls guidance asks for claim 158unconfirmed, and whether FDA would accept performance testing on LDTs or dPCR when no tumor-informed MRD assay holds FDA clearance claim 163unconfirmed claim 189unconfirmed.
• The nearest precedent, a third-party oncology-monitoring control listed under QIG with no 510(k) found claim 155unconfirmed, and who lists under SCU claim 132unconfirmed.
• The file names a drafted 513(g) request as a human spend decision.

Whether the derived levels hold.

• The gravimetric value assignment below 0.1% within a 35% relative expanded uncertainty claim 134unconfirmed. The Poisson arithmetic at 10-30 ng inputs claim 133unconfirmed claim 16unconfirmed. Vial-to-vial heterogeneity claim 138unconfirmed. An orthogonal sequencing check on the pilot lot claim 169unconfirmed.
• The daily control's specification, at a mean per-variant VAF of 0.03% claim 166unconfirmed. It has to deliver at least 3 expected molecules per extraction claim 160unconfirmed, and precision is claimed only where at least 100 molecules are expected claim 161unconfirmed. The file's Verifier flags an arithmetic slip inside claim 166unconfirmed, and the feasibility deck repeats it.
• Donor-pool background at the certified positions claim 139unconfirmed claim 167unconfirmed, extraction equivalence across three methods claim 135unconfirmed, and whether one pool can be both commutable and clean claim 63unconfirmed.
• The pre-distribution bench programme: a replicate series claim 62unconfirmed, three-lot consistency claim 170unconfirmed, and cross-chemistry commutability claim 31unconfirmed.
• Published warning signs held on one element each: every assay in the FNIH QC study dropped out at 0.1% claim 96unconfirmed, and the lowest KRISS value-assigned level was 0.25%-0.32% at 33%-90% relative uncertainty claim 97unconfirmed.

Whether anyone would use it.

• Inadequacy of in-house controls claim 23unconfirmed. Acceptance by a CAP inspector or CLIA assessor claim 24unconfirmed. Lot-to-lot continuity risk claim 25unconfirmed.
• Adoption of the daily vial by one in ten laboratories at $400-$900, in place of material they could make by diluting the buffer-format incumbent for about $80 a run claim 177unconfirmed. Use on every run-day claim 187unconfirmed.
• Designing and maintaining a patient-specific panel against the control's certified variant list claim 185unconfirmed claim 162unconfirmed. Directors accepting derived-labelled values for a published LoD claim 186unconfirmed.
• The layers above CLIA: CAP COM.40400 claim 78unconfirmed and MOL.36118 claim 144unconfirmed, BLOODPAC claim 164unconfirmed, and the MolDX technical assessment claim 137unconfirmed claim 165unconfirmed.
• The dossier's lab enumeration gives floor counts by evidence type: 8 companies in the MolDX article, 5 US-offered assays with a published analytical validation, and others. It says these sit "well below 30" but do not test the file's 30-60 range. It also flags that Guardant Reveal and NavDx are not tumor-informed [dossier: ctdna-mrd-assay-validation].

Whether it makes money, and at what cost.

• Price claim 149unconfirmed, panel and vial demand claim 150unconfirmed claim 151unconfirmed, capture claim 47unconfirmed, and revenue against cost claim 178unconfirmed. The cost side has no figure anywhere in the file, and the bottom-up build is itself open claim 140unconfirmed.
• The P2 clinical QC material at launch fixed layer fitting the P0 the original shape budget claim 180unconfirmed claim 20unconfirmed. First distribution taking longer than the 9-15 months stated for the RUO shape claim 179unconfirmed claim 50unconfirmed.
• The direct price of the incumbent's MRD panel mix against its $1,919.59 distributor price claim 152unconfirmed claim 117unconfirmed, and a 25% premium for a plasma matrix with stated uncertainty claim 153unconfirmed.
• Whether the customers' own test is actually covered under LCD (local coverage determination) L38779 (a Medicare local coverage determination number) and not merely listed claim 136unconfirmed.
• The file flags two claims without resolving them. claim 70refuted's premise "now looks false" and needs a human retag claim 70refuted. The Verifier notes that claim 192unconfirmed, on the non-Natera share of MRD billing, "leans toward refuted", but it stays claim 192unconfirmed.

Inputs, IP (intellectual property) and competitors.

• Plasma supply and consent claim 172unconfirmed claim 148unconfirmed, cell-line terms claim 176unconfirmed, and a contract manufacturer already listed under 866.5910 claim 171unconfirmed claim 181unconfirmed.
• The scope of the three live patents' claims claim 145unconfirmed claim 146unconfirmed claim 147unconfirmed, and the broader FTO (freedom to operate) negative claim 174unconfirmed.
• No incumbent shipping a plasma-matrix sub-0.1% control first claim 188unconfirmed, and whether incumbents license rather than build claim 182unconfirmed.

Where the three lenses agree — and where they don't

How current the decks are. All three decks were rebuilt on 2026-09-26 against P2 clinical QC material at launch, with source_updated: 2026-09-26. None is stamped STALE. But check_decks.py fails all three today, 0/3, on tag agreement. The decks cite claims they put on record, or that other artifacts added the same day, as unverified. The Verifier has since tagged every one of claims 179-192 unconfirmed. The desirability deck cites claims 183, 185, 186 and 187 this way. The viability deck cites claims 179, 180, 181, 182, 184, 187 and 188. The feasibility deck cites claim 189unconfirmed. No verified or refuted tag has moved under any deck, so none of the decks' readings changes direction. But each deck's bracket for those claims is behind the file, and its "Where this deck outruns the file" slide describes them as unverified. This narrative cites them with their current tag. None of the three load-bearing sentences quoted below cites one of the drifted claims.

Desirability. This chair sees three people in one laboratory: the director who signs the LoD claim, the scientist who validates it, and the technologist who loads a control every run-day. It finds the moment is real and mandated. CLIA requires daily controls claim 7verified and an established analytical sensitivity claim 76verified. Everything past the moment is open: dissatisfaction claim 23unconfirmed, acceptance by CAP, MolDX and BLOODPAC claim 24unconfirmed claim 164unconfirmed, panel design against the control claim 185unconfirmed, and acceptance of derived labelling claim 186unconfirmed. It counts six of six conditions with nothing verified behind them, and the file records no interview, letter of intent or pilot. This deck agrees with feasibility that the matrix premise is contested by verified findings claim 90verified claim 92verified. Its sharpest point is that P2 clinical QC material at launch moved the product's centre of gravity to the daily vial.

"whether at least one in ten US laboratories running tumor-informed MRD assays would use this daily control as their run-by-run external control, at $400-$900 per vial, in place of in-house material, including material they could make by diluting the buffer-format incumbent into their own plasma for roughly $80 a run" — the load-bearing condition of the desirability deck (built 2026-09-26).

Viability. This chair sees an investor and an operator running a regulated consumable business. It agrees with desirability that the daily vial carries most of the revenue claim 178unconfirmed. It adds that P2 clinical QC material at launch's quality-system layer is the one verified change to the cost side claim 159verified. Every verified line about the capital estimate weakens it claim 58verified claim 59verified claim 61verified. On the customer side, it reads the verified CMS numbers as a Medicare fee-for-service floor on one channel claim 68verified claim 71verified, not a market size. On the moat, it finds the verified evidence "leans the wrong way": CLIA mandates no source or traceability claim 76verified claim 77verified, and a vendor already holding ISO (International Organization for Standardization) 13485 would meet the QMSR entry cost more easily. It counts seven of seven conditions unverified behind the condition itself. Where it differs from feasibility is emphasis. Viability treats the sub-0.1% white space as a pricing question that falls if an incumbent ships first claim 188unconfirmed. Feasibility treats it as the product thesis.

"whether, at the file's own planning volumes (50 panels and 2,000 daily vials a year) and the low end of its own prices, annual revenue of about $875,000 exceeds the fully loaded cost of goods plus the recurring cost of the 866.5910 obligations" — the load-bearing condition of the viability deck (built 2026-09-26).

Feasibility. This chair sees engineering: make a material with known tumor-derived molecules at levels no dPCR measures directly, and make it three times. It says the question has moved from "can it be measured" to "can the derivation be defended, and can it be built three times". It is the only lens with a verified anchor under its central condition claim 100verified claim 99verified. But the step below that anchor claim 134unconfirmed and every hardest-unknown line claim 63unconfirmed claim 167unconfirmed claim 135unconfirmed are open, and its verified lines on the matrix are all hazards. It also carries the regulatory engineering question. The exemption has to survive three 866.9 tests claim 52unconfirmed claim 53unconfirmed claim 157unconfirmed, and the special control's performance testing presumes a cleared assay that does not exist claim 189unconfirmed. It agrees with viability that if the bench programme fails, the product retreats to the range incumbents already occupy. It agrees with desirability that a control an assay was not designed against is invisible to that assay.

"whether each panel level below 0.1% VAF can be value-assigned by gravimetric dilution from a dPCR-quantified parent into a dPCR-quantified diluent with an expanded uncertainty (k=2) of no more than 35% relative that includes dilution, between-unit heterogeneity and a characterized limit-of-blank term, and be labelled as derived rather than measured" — the load-bearing condition of the feasibility deck (built 2026-09-26).

Where they converge. All three conditions meet at one object: the derived sub-0.1% level. Feasibility asks whether it can carry a defensible uncertainty. Desirability asks whether a director and a technologist will put it on the bench every run-day in place of what they make themselves. Viability asks whether doing that at the stated price pays for the 866.5910 layer. No lens has a verified line under its own load-bearing condition.

The load-bearing question

Can a sub-0.1% VAF level that is value-assigned by dilution and labelled as derived rather than measured claim 134unconfirmed carry an uncertainty budget that US laboratory directors will accept both for the limit of detection they publish claim 186unconfirmed and as the run-by-run control they load every patient run-day claim 177unconfirmed claim 187unconfirmed, at prices that cover the 866.5910 quality-system cost the file has not yet put a figure on claim 178unconfirmed?

Naming this question is not a recommendation, a gate, or a kill. The dPCR anchor above the derived levels is verified claim 100verified, and nothing below it is.