Domain dossier

Validating and quality-controlling tumor-informed ctDNA MRD assays

Adaptation note, read this first. idea-003's target: is RUO (research use only), not a patient population, and its buyer is a molecular laboratory, not a clinician or a patient. The _template.md sections were written for a patient-population domain (epidemiology, care pathway, user groups, standard of care) and are adapted here rather than forced: Section 1 covers testing volume and the number/growth of tumor-informed MRD (molecular residual disease) assays and the labs running them, not disease prevalence. Section 2 ("care pathway") is read as the lab workflow step a reference-material panel sits in — assay development bench work, analytical validation, then ongoing daily/lot QC (quality control) — not a clinical site of care. Section 3 ("user groups") separates the molecular lab director/assay developer (buyer) from the CLIA (Clinical Laboratory Improvement Amendments) lab operator (daily user) from the assay manufacturer (the downstream commercial customer whose own analytical-validation needs define what the reference material must match). Section 6 ("reimbursement") is genuinely not applicable to an RUO (research use only) laboratory consumable — nobody bills a payer for it — and that is stated plainly rather than left silently blank, the same disposition idea-003's own Scorer and Verifier already recorded for its Factor 2.

This dossier is thinnest exactly where a patient-population dossier would be thickest (epidemiology, care pathway, reimbursement) and thickest exactly where this domain's real decision evidence lives: Section 5, the reference-material vendors and the MRD assays that consume their output, worked hardest per the research-lane spec.

This file carries no confidence tags. Every line below is context with a source attached. To affect a score, a line must be copied into a candidate's ## Claims as unverified via engine/scripts/append_research_claims.py and pass the Verifier or a future Corroborator. See README.md in this directory.

---

1. Condition & epidemiology

as_of: 2026-09-25 (refreshed; first pass 2026-08-27, whose text is kept below unchanged) · sources searched: connectors literature (multiple queries, see §10), directly-fetched company investor-relations press releases and an earnings-call transcript · not searched: a systematic enumeration of every US laboratory offering a tumor-informed MRD assay; non-Medicare/non-commercial test volumes

Adapted meaning of "epidemiology" here: testing volume and market growth, not disease prevalence. The clinical prevalence of the underlying cancers (largely stage II-III colorectal cancer to date) is Section 4's subject, not this section's.

Verbatim:

"Clinical molecular residual disease (MRD) oncology units grew 34,000 units over the first quarter of 2026, the largest sequential increase to-date." — Natera, Inc., Q2 2026 financial results press release, dated 2026-08-06 [18]

"Reveal continues to perform at an extremely high level with volume growth accelerating for the third quarter in a row to well above 100% year-over-year." — Guardant Health Q2 2026 earnings call, 2026-07-30 [19]

Read this narrowly. Neither figure is an absolute quarterly MRD test count independently confirmed by this pass — the 1,043,900/1,030,100 figures are Natera's total company test volume across all product lines (women's health, organ health, oncology, MRD combined), and the specific absolute Signatera-MRD-unit count reported by secondary financial media (approximately 283,000 units) was not independently confirmed by a direct fetch of a primary Natera document in this pass — it is recorded here as not retrieved, not as a finding, and must not be treated as one. The Guardant figure is a growth rate stated verbatim by management on an earnings call, not an absolute unit count; Guardant's own Q2 2026 press release (fetched directly) does not mention Reveal or MRD volumes at all, which is itself worth recording — the growth figure lives in the call transcript, not the press release.

Gaps: no enumeration of how many US laboratories run or are developing a tumor-informed MRD assay was found (the "30-60 laboratories" figure idea-003 claim 12 uses is the Generator's own unsourced estimate, already flagged unconfirmed/permanently unverifiable in that file); no absolute, independently-confirmed Signatera or Reveal MRD-specific test count for any quarter.

Update, as_of: 2026-09-25 — SEC filings and the CMS (Centers for Medicare & Medicaid Services) utilization file, folded in. Two things the first pass recorded as gaps have moved. Both are company-reported (SEC) or Medicare-FFS (fee-for-service)-only (CMS), and neither is a market size.

SEC filings (read out of the fleet report on this topic; its sources are cited here, not the report). The absolute clinical-MRD count this section called "not retrieved" is not in any Natera Form 10-Q or 10-K at all: an EDGAR full-text search over CIK 0001604821, filings 2023-01-01 to 2026-09-25, matched "oncology units processed" in all 11 Natera 10-Qs and 4 10-Ks and matched "clinical MRD" in none of them — only in Form 8-K Exhibit 99 earnings materials [32]. What the periodic filings and the 8-K exhibits say:

How these were read, stated because it limits them: the fleet's fetch of the full 10-Q HTML truncated before MD&A, so the 10-Q figures are exact phrases matched by EDGAR full-text search inside the named filing, not whole sentences read in context; a wrong-figure negative control ("…approximately 299,700") returned zero hits [32]. The "~283K" figure is therefore now retrieved from a primary company document (the 8-K exhibit), which corrects the first-pass note above that it was known only from secondary financial media — but it remains a rounded, company-reported number from an earnings presentation, not an audited periodic-filing figure.

Inference, labelled as such: ~283K clinical MRD tests would be roughly 95% of Natera's 296,700 total oncology units in the same quarter, so the 10-Q oncology figure is plausibly a close upper bound on Signatera clinical volume. The filings do not say this.

CMS Medicare FFS utilization (typed connector, cache warm). python3 -m connectors procedures 0340U --jurisdiction US → national, office/non-facility row, 25,746 services for 0340U; connector detail: "Medicare fee-for-service only … excludes Medicare Advantage (Medicare Advantage), commercial, Medicaid, VA, and cash-pay. This is a FLOOR on US volume, not a market size" [35]. The release is CMS "Medicare Physician & Other Practitioners — by Geography and Service", file MUP_PHY_R26_P05_V10_D24_Geo.csv (CY2024 claims per idea-003 claim 68; the connector itself says "year not stated in the file"), sha256 ce9e4f30…be634 in data/provenance.json. Two other codes listed alongside 0340U in the MolDX solid-tumor MRD billing articles (§6) return, from the same file: 0356U 4,481 services, 0422U 243 services; 0569U has no utilization row (the connector reads this as a finding for a code of that shape — most plausibly a code newer than the CY2024 claims year, which is an inference).

Do not ratio these. The SEC figures are quarterly, all-payer, company-reported; the CMS figure is an annual Medicare-FFS claims count for a different year. No ratio between them is computed here, for the reason the fleet report itself gives: the periods and payer bases do not match.

Gaps still open after this update: an absolute Reveal-specific count (not in any Guardant document read; the Reveal ">100% year-over-year" statement stays earnings-call-only [19]); Natera's 2023-2024 quarterly 10-Q figures and full-year 10-K figures (not searched); the numeric content of Guardant's 10-Q sentences containing "Reveal tests" (located, not read).

2. Care pathway & clinical setting

as_of: 2026-09-25 (refreshed; first pass 2026-08-27, addenda 2026-09-15 — all kept below)

Adapted meaning: a laboratory workflow step, not a clinical site of care. A commutable reference-material panel sits at three points in a molecular laboratory's workflow, not in a clinical pathway: (a) assay development bench work, where the developer establishes a limit of detection and characterizes the analytical measurement range; (b) formal analytical validation ahead of clinical launch or a CAP/CLIA inspection; (c) ongoing lot-to-lot and daily quality control once the assay is running clinically (the CLIA-mandated control-procedure requirement below). No connector or search in this repo addresses laboratory bench workflow directly; what is established instead is the regulatory floor that makes step (c) recurring rather than optional.

Regulatory context carried here because it recurs across this domain's candidates, not because it is un-established: idea-003's own Claims and Verifier passes already established, via the same regulation (eCFR) connector re-run fresh for this dossier, that 42 CFR 493.1256 (CLIA) requires "for each test system, the laboratory is responsible for having control procedures that monitor the accuracy and precision of the complete analytic process," and that "[a]t least once each day patient specimens are assayed or examined," each molecular amplification procedure must include control materials [14]. That regulation says nothing about whether the control material must be purchased from a third party or built in-house — a limit already recorded in idea-003 and not re-litigated here.

Gaps: no direct source on turnaround time, staffing, or bench sequencing for a tumor- informed MRD assay run (tumor-tissue whole-exome/whole-genome sequencing to identify a patient-specific variant panel, then serial plasma draws against that panel) was searched this pass.

Update, as_of: 2026-09-15 — FDA (US Food and Drug Administration)'s 2013 RUO/IUO distribution guidance, retrieved. The previous pass of this dossier (2026-08-27) recorded this document as unretrieved, matching idea-003 claims 2-3's own structural gap (see §10). This pass fetched the actual PDF directly from fda.gov (https://www.fda.gov/files/medical%20devices/published/Distribution-of-In-Vitro-Diagnostic-Products-Labeled-for-Research-Use-Only-or-Investigational-Use-Only---Guidance-for-Industry-and-FDA-Staff.pdf, via curl, since the WebFetch tool's own request to the same URL returned HTTP 401 — the fda.gov access flakiness this repo's conventions already anticipate) and read it in full. It is FDA CDRH (Center for Devices and Radiological Health)/CBER guidance, Docket No. 2011-D-0305, issued 2013-11-25 [21].

What Section IV.B (pp. 9-10) actually names, verbatim, as evidence — under a totality-of-the-circumstances test, not a bright-line rule — that a manufacturer's actual intended use conflicts with RUO/IUO labeling:

"Written or verbal statements in any labeling, advertising, or promotion of the IVD (in vitro diagnostic) product by or on behalf of the manufacturer, including any performance claims, instructions for clinical interpretation, clinical information, product names, or descriptors that claim or suggest that the IVD product may be used for any clinical diagnostic use... Written or verbal statements in any labeling, advertising, or promotion of the IVD product by or on behalf of the manufacturer that suggest that clinical laboratories can validate the test through their own procedures and subsequently offer it for clinical diagnostic use as a laboratory developed test. Solicitation of business from clinical laboratories; for example, a manufacturer who produces only products labeled RUO whose sales force makes routine calls to clinical laboratories that do not perform research or clinical studies may be viewed as demonstrating its intent that its products be used for clinical purposes. Provision of certain types of specialized technical support (e.g., assistance in performing clinical validation) to clinical laboratories."

Section IV.B.2 (pp. 10-11) adds, specifically on validation assistance:

"FDA views the activities of a manufacturer that aid the clinical laboratory in validation or verification of a test that incorporates RUO or IUO labeled IVD products as evidence of the manufacturer's intended use. If the manufacturer of an IVD product labeled RUO or IUO were to assist in the validation or verification of the performance of a test for clinical diagnostic use that uses its RUO or IUO labeled IVD, that assistance would be considered to be evidence of a non-research or non-investigational intended use."

And Section II (p. 6), bearing on idea-003 claim 1's "premarket half":

"Mere placement of an RUO or IUO label on an IVD product does not render the device exempt from otherwise applicable clearance, approval, or other requirements. FDA may determine that the device is intended for use in clinical diagnosis based on other evidence, including how the device is marketed."

How this refines, rather than simply confirms, idea-003 claim 3's hypothesis ("selling specifically to CLIA laboratories to validate clinical tests"). The guidance does not name bare arm's-length sale to a CLIA laboratory, by itself, as conflicting evidence. What it names is more specific and compound: (a) a sales force's routine solicitation of clinical laboratories that do not themselves perform research or clinical studies; (b) statements suggesting a lab can self-validate and then offer the test as an LDT (laboratory-developed test); (c) the manufacturer itself providing specialized technical support or assistance in performing clinical validation; and (d) any of the above assessed together under a "totality of the circumstances" standard, not as independently sufficient triggers. Whether idea-003's specific business model — marketing a reference panel directly to "molecular laboratories" for "assay development and ongoing performance monitoring" (the file's own intended_use line) — falls inside bullet (a) or (c) is an application of this text to a specific fact pattern, which is exactly the FDA intended-use judgment call this dossier does not make and the Verifier does not resolve either; the text itself, now retrieved, is what changed. See idea-003 claims 64-65 for the claims drafted from this reading.

Update, as_of: 2026-09-15 — current status of FDA's 2024 laboratory developed test (LDT) final rule, retrieved. idea-003 claim 15 records the March 2025 vacatur but flags the current status as a structural gap ("no connector in this repo queries federal court dockets or opinions"). No connector does, and none was substituted here for the docket itself — but the practical consequence of the vacatur (whether FDA appealed, and what it did instead) is on the public record in a document a typed connector does reach: FDA's own implementing rule in the Federal Register.

python3 -m connectors policy 2025-18239 --expect "vacated" --jurisdiction US → verified, source_ref "Federal Register document_number=2025-18239 (90 FR 45134)" [22]. The connector's own retrieved quote, verbatim from the document text:

"...issued a final rule amending the definition of ``in vitro diagnostic products'' in FDA's regulations. On March 31, 2025, a federal district court vacated that rule. This final rule reverts to the text of the regulation as it existed prior to the effective date of the May 2024 final rule." — 90 FR 45134, "Regulation Identification Number 0910-AJ05 Medical Devices; Laboratory Developed Tests; Implementation of Vacatur," published 2025-09-19, effective 2025-09-19 [22]

What this settles and what it does not. The document, retrieved by a typed connector against a US federal registry (a class-1 registry fact), establishes that (a) the March 2025 vacatur was not reversed or stayed as of the document's September 19, 2025 publication, and (b) FDA's own response was to revert its regulation text to the pre-May-2024 state, not to appeal. It does not itself use the words "CLIA" or "enforcement discretion" — re-queried with --expect "enforcement discretion" and --expect "CLIA", both returned unconfirmed (phrase not found; a miss is not a refutation, per the connector's own stated behavior for free prose) — so the claim that LDTs consequently "remain under CLIA/CAP oversight" is not established by this document's text and is carried here as the pre-existing regulatory baseline (LDTs were CLIA-regulated laboratory services before the 2024 rule and this rule reverts to that state), not as something the Federal Register document asserts in those words.

Corroborating, non-registry source, read separately rather than substituted for the above: a legal-industry publication dated after the Federal Register document states plainly that no appeal was ever filed: "The government did not appeal the district court's decision to the Fifth Circuit," and "In September 2025, FDA officially implemented the vacatur and rescinded the final rule." — Papagiannopoulos, T., "American Clinical Laboratory Association v. FDA," Food and Drug Law Institute, published 2026-06-24 [23]. This is a class-2 published finding, not a registry fact, and is recorded as corroboration of the registry document's own implication rather than as the primary source for the no-appeal fact — the Federal Register document is silent on why FDA reverted the text (no appeal vs. an appeal later abandoned), and [23] is the source for that specific fact.

As of 2026-09-15, the most recent primary-source check available in this repo (a Federal Register document dated 2025-09-19, fourteen months old) shows no reversal. No connector here can confirm anything more recent than that document's own date; a docket search for any 2026 appellate activity was not performed (no connector reaches PACER or a circuit court's docket, and none was substituted). See idea-003 claim 66 for the claim drafted from this reading.

Update, as_of: 2026-09-25 — what CLIA, CAP, the AMP/CAP guidelines and ISO (International Organization for Standardization) 17034 require, accept, or leave unsaid about the material used to establish a limit of detection. This is the documentary twin of the class-3 question in §9 ("would an inspector accept a third-party panel?") and the question idea-003's open request REQ-0028 asks. Read out of the fleet report on this topic, with its sources cited directly; the two CLIA sections were re-run through the typed regulation connector this pass.

What this establishes, stated narrowly. Across the texts actually read, a laboratory's own LoD establishment is required (CLIA and CAP, two independent bodies), third-party traceable reference material for it is required by none, recommended by one professional guideline (in "should" language, and as cell lines rather than as a certified product), and accepted as one option by one CAP accuracy item. ISO 17034 binds a producer; a laboratory would meet it only indirectly, by buying from an accredited producer — and that reading rests on the standard's public scope description, not its clause text.

What it does not establish. The current CAP MOL checklist text (downloads only for CAP-accredited laboratories [39]), MOL.36118's current wording, MOL.31015/MOL.31025, the 2023 AMP/CAP cfDNA guideline's full recommendations, and ISO 17034's clause text were all not retrieved — a paywall or access restriction, not an absence. Any of them could change the "required by none" reading. This is not an answer to whether a CAP inspector would accept a third-party panel; it is what the documents say they require.

Update, as_of: 2026-09-25 — ISO 17034 accreditation: who accredits, what is published about cost and time, and who holds a nucleic-acid scope. Bears on idea-003 claims 20 and 33 (the capital estimate and "ISO 17034 achievable on the stated budget and timeline"). Read out of the fleet report on this topic.

• Fees are quote-only wherever looked. A2LA: "The cost to obtain accreditation will vary significantly depending on the size of the organization seeking accreditation, the size of the scope(s) of accreditation, and the organization's readiness for assessment" and offers a "free, no-obligation estimate" [47][48]. PJLA: "We offer competitive pricing plans that fit the needs of our customers", no rates [49]. ANAB's pages returned HTTP 403 and were **not retrieved**; its AR 2258 requirements document (effective 2025-10-29) has no fee or timeline text [50]. A2LA's R101 names fee components without amounts: after the first year, "annual fees (including any applicable program surcharge fees) and assessor fees" [46].
• Timeline — A2LA's figures, all programs, not ISO 17034 alone: "In general, a well-prepared applicant can complete the accreditation process in three to six months"; "On average, new applicants complete the accreditation process within 5 months, from start to finish" — no n, no date range, no program breakdown [47].
• Cycle obligations (A2LA R101, publication date 2026-09-01): accreditation runs two years with "a full reassessment of all accredited CABs every two years" and "a one-day surveillance assessment" after the first year; initial applicants must resolve all nonconformances within four months of the exit briefing [46]. PJLA states annual assessment [49].
• ILAC recognition: "In July 2020, A2LA extended signatory status with the International Laboratory Accreditation Cooperation (ILAC) mutual recognition arrangement (MRA) to include reference material producers (RMPs) under the standard ISO 17034" [45].
• A US producer with a nucleic-acid ISO 17034 scope exists: ATCC, ANAB certificate AR-1384 (scope version 011, issued 2025-03-10, valid to 2027-03-17), "Reference Materials and Certified Reference Materials — Nucleic Acids", characterized by methods including "Digital PCR (ddPCR) for quantitative testing" — the scope does not name human, cell-free or tumour DNA [51]. Microbiologics' A2LA scope (2655.02) is microbial cultures only [52].
• The named ctDNA (circulating tumour DNA) incumbents' pages read do not state ISO 17034: SeraCare ctDNA Complete AF0.5% (0710-0672): "Manufactured within cGMP compliant, ISO 9001 and ISO 13485 certified facilities" [53]; LGC Standards' own accreditations page lists ISO 17034 sites and no SeraCare, Milford or Gaithersburg site among them [54]; Horizon/Revvity lists cell-free DNA in buffer under "13485:2016" [55] and its Multiplex I cfDNA page makes no ISO 17034 statement [56]. This is not found on the pages read, not "not accredited" — non-US accreditation directories (UKAS, DAkkS) were not searched.

Consequence for any capital or calendar figure: no published fee schedule exists in what was read, so an ISO 17034 cost line in idea-003 claim 20 or 33 cannot be traced to a published number; the only published calendar figure is A2LA's generic three-to-six months, which is not program-specific.

3. User groups & incentives

as_of: 2026-08-27 (role table, not re-checked since; a purchase-channel update dated 2026-09-25 is appended at the end of this section)

Four separate roles, adapted from the patient-population template's five:

Gaps: the purchasing role — whether the buyer is the same person who runs the daily control procedure, and whether purchasing this consumable requires a capital-procurement process or is a bench-level consumable order — was not independently sourced this pass; idea-003's deck- surfaced assumption on this remains unverified and unresourced here.

Update, as_of: 2026-09-25 — the purchase channel, as the incumbents actually expose it. Whether a scientist would buy a panel as a consumable without capital procurement (idea-003 claim 26) and whether the segment can be reached without a distributor (claim 28) are class 3. Their revealed-behaviour twin is how the incumbents' products are in fact offered for sale, which is observable. Read out of the fleet report on list prices, retrieved 2026-09-25:

• SeraCare/LGC's own store shows ctDNA product pages with "Please sign in to view pricing" (e.g. 0710-0672, 0710-0673, 0710-2146) [53][92][99]; Horizon/Revvity's Multiplex I page shows "Select your product to view pricing" and says orders go through the online catalogue or a quote [56]; Twist's cfDNA Pan-Cancer v2 product sheet gives catalogue numbers and "sales@twistbioscience.com" and no price [100].
• Seraseq ctDNA items are nonetheless orderable, with a displayed per-unit web price, through a general laboratory-supply distributor (Fisher Scientific) — which labels them "Non-distribution item offered as a customer accommodation; additional freight charges may apply" and states "Your contract pricing may differ" [94].
• Thermo Fisher's AcroMetrix ctDNA plasma control is priced per vial on thermofisher.com [101].

Read narrowly: this shows that the incumbents' ctDNA reference materials are sold as single-unit catalogue items with a web or quote price, some also through a distributor as an accommodation. It does not show who in a laboratory places the order, whether a capital process is involved, or what anyone would pay for a different product.

4. Standard of care & clinical evidence

as_of: 2026-08-27 (§4a, not re-checked since; §4b and Gaps partly updated 2026-09-25 — see the dated update block) · sources searched: connectors literature + literature-detail on 13 PMIDs (see §10) · not searched: guideline society statements (ASCO, NCCN) directly; CAP/CLIA proficiency-testing program specifics (paywalled/member-only, not attempted)

What "standard of care" means in this domain: whether tumor-informed ctDNA MRD testing changes management, and what analytical performance the assays that use this reference material claim for themselves. Two distinct evidence threads: (a) clinical validity of ctDNA-guided management in colorectal cancer, and (b) analytical validation of the assays themselves — the latter is what tells a reference-material maker what range it needs to cover.

4a. Clinical validity — colorectal cancer surveillance and ctDNA-guided treatment

Verbatim:

"Among 968 evaluable patients, 702 (72.5%) were ctDNA negative... ctDNA-negative patients experienced significantly fewer recurrences than ctDNA-positive patients (3-year RFS 87% versus 49%; P < 0.001)... In ctDNA-negative patients, de-escalation reduced oxaliplatin use (34.8% versus 88.6%) and hospitalizations (8.5% versus 13.2%) but yielded slightly lower RFS than standard management (85.3% versus 88.1%), not meeting the non-inferiority margin." — PMID 41115959 (DYNAMIC-III) [8]

"Median DFS was 9.30 months with FTD/TPI and 5.55 months with placebo (hazard ratio = 0.79, 95% confidence interval: 0.60-1.05, P = 0.107), and the primary endpoint was not met." — PMID 42260101 (ALTAIR) [9]

Known negative/equivocal findings, recorded with the same weight as the positive ones: ALTAIR's negative primary endpoint [9] and DYNAMIC-III's failure to meet its de-escalation non-inferiority margin [8] are exactly the kind of finding a title alone would not surface — both are read at full abstract depth here, not from a headline. ctDNA status is validated as strongly prognostic in this evidence base; whether acting on a positive result (escalating or intervening) improves outcomes is, on the two largest randomized trials located this pass, not yet demonstrated.

4b. Analytical validation of the assays this reference material would serve

This is the load-bearing evidence for what an assay-validation reference panel must be able to do — the LOD and VAF (variant allele frequency) range each commercial or published tumor-informed assay claims for itself.

Verbatim:

"Haystack MRD demonstrated a 95% limit of detection of 0.063 mean ctDNA molecules/mL and an analytical measurement range of 0.25-1,000 mean ctDNA molecules/mL." — PMID 42624249 [1]

"The analytical validation result of CancerDetect™ showed limit of detection successfully reached down to 0.001% (10-5) with 99.9% specificity." — PMID 41212866 [2]

"Analytical measurements demonstrated a detection threshold of 1.67 parts per million (PPM) with a limit of detection at 95% (LOD95) of 3.45 PPM." — PMID 38484152 [3]

"Limit of detection ranged from 0.008% allele frequency when utilizing 60 ng of cfDNA input with 18-50 variants in the patient-specific panels (> 99.9% sensitivity)... to 0.05% allele frequency when using 10 ng of cfDNA input with an 18-variant panel." — PMID 37632661 [4]

"umiVar enabled variant detection at a limit of detection as low as 0.0017%, with no false positive calls in mutation-free reference samples." — PMID 40866952 [13]

Read this against idea-003's own stated target range. The candidate's mechanism targets "0.001%-0.5% VAF." The published assay landscape above spans 0.000345% to 0.5%+ across five distinct commercial/research methods, using inconsistent units (molecules/mL vs. VAF percent vs. ppm) that are not directly interconvertible without knowing the total cfDNA input mass per assay — a real complication for any single reference-material panel claiming to serve "the MRD assay market" as a monolithic target. The bottom of the candidate's target range overlaps the more aggressive published assays' claimed LODs; the top of its range (0.5%) is comfortably inside what commercial reference vendors already characterize (see §5).

Gaps: no guideline (NCCN, ASCO) position statement was independently retrieved this pass — the trial evidence above is read directly, not through a guideline's summary of it. CAP/CLIA proficiency-testing program specifics for ctDNA MRD assays (as opposed to the general EQA literature in §5) were not searched — CAP program materials sit behind a member paywall this pass did not attempt to breach.

Update, as_of: 2026-09-25 — two more US-offered assays' analytical claims, from the fleet's laboratory-enumeration report.

The PhasED-Seq figure (0.7 ppm at 120 ng) sits below NeXT Personal's 3.45 ppm [3] and is stated at a named input mass — the only assay row in this section that pairs an LoD with its input, which is what the Poisson arithmetic in §7 needs. No analytical-validation abstract was found for Signatera, RaDaR, Oncodetect, Pathlight or TrueMRD in PubMed title searches (7-year window) — a result of that search, not proof no paper exists.

Guideline gap, partly closed: the AMP/CAP guidelines on NGS panel and cfDNA assay validation are now read at the depth §2's update records (2017 full text; 2023 abstract only). NCCN/ASCO clinical-management guideline text remains not searched.

5. Reference products — and why people stop using them

as_of: 2026-09-25 (refreshed; first pass 2026-08-27 kept below unchanged, and partly superseded — see the correction at the head of the 2026-09-25 update) · sources searched: connectors literature/literature-detail; direct WebFetch of vendor product pages, FAQ pages, and one vendor flyer; direct WebFetch of company investor-relations press releases and one earnings-call transcript · not searched: MAUDE (Manufacturer and User Facility Device Experience) (no connector — RUO/LDT materials and assays are not device-cleared, so MAUDE would be silent by construction the same way idea-003 already reasoned about clearances/recalls); independent third-party head-to-head commutability studies comparing the three named vendors

The load-bearing section, worked hardest per the research-lane spec. The "incumbent" here is not one product but three named commercial reference-material vendors (idea-003's predicate_or_analog), plus the assay manufacturers who are the panel's actual downstream economic beneficiary.

Seraseq® ctDNA Reference Materials (LGC Clinical Diagnostics, formerly SeraCare)

• What it is: a contrived reference-material product line for validating NGS- and ddPCR-based ctDNA assays. LGC acquired SeraCare Life Sciences in November 2018; SeraCare's NGS-genomics reference-material brand became part of "LGC Clinical Diagnostics."
• Matrix: the current v4 AF0.1% product's own page describes the matrix as a **"plasma like matrix"** — not stated to be pooled human donor plasma.
• Lowest characterized allele fraction (vendor-labeled): 0.1% VAF for the ctDNA v4 line (four levels: WT/0%, 0.1%, 0.5%, 5%); the ctDNA Complete™ line's lowest non-zero level is also 0.1% (six levels: WT, 0.1%, 0.5%, 1%, 2.5%, 5%).
• Evidence behind it: described on the vendor page as "orthogonally validated truth-set materials"; no independent third-party validation study of this specific product was located this pass.
• Adoption: not established by this pass — no independent sales, installed-base, or citation-count figure was found.
• Abandonment record — why people stop:

• Verbatim:

"The product uses a 'Plasma like matrix' rather than actual human donor plasma or other specified contrived formulations." — WebFetch summary of LGC/SeraCare Seraseq ctDNA Reference Material v4 AF0.1% product page, fetched 2026-08-27 [15]

"Seraseq ctDNA v4 is manufactured with a new fragmentation process which results [in] lower background than previous Seraseq ctDNA offerings to support greater confidence at low variant allele frequencies." — same product page [15]

"We are excited to welcome SeraCare to our Standards division." — Euan O'Sullivan, Managing Director, LGC's Standards division, LGC press release dated 2018-11-26 [20]

Gaps: no independent (non-vendor) commutability study of Seraseq ctDNA v4 against real patient plasma was located. Whether "plasma like matrix" retains native cell-free DNA fragmentomics (the specific differentiator idea-003 claims) or is a wholly synthetic construct is not resolved by the vendor's own page language.

Horizon Discovery / Revvity HDx cfDNA Reference Standards

• What it is: a cfDNA reference-standard product family (Mimix™/Multiplex I, Multiplex II, OncoSpan, and others) manufactured by Horizon Discovery, now part of Revvity.
• Matrix: the vendor's own FAQ page states the standards are **"derived from engineered human cell lines"** — fragmented human genomic DNA, average fragment size ~160 bp — **spiked into a synthetic plasma matrix** for the Matrix I/II product formats. Not pooled human donor plasma with native cell-free DNA.
• Lowest characterized allele fraction (vendor-labeled): "as low as 0.1 percent," matched with a wild-type control.
• Evidence behind it: vendor page states the standards support development, validation, and "demonstration of the performance of cfDNA assays and platforms on a routine basis"; no independent third-party validation study specific to this product was located this pass.
• Adoption: not established this pass.
• Abandonment record — why people stop: see the shared cross-vendor finding [12] above (precision below 0.5% VAF); no vendor-specific discontinuation record was located.
• Verbatim:

"Derived from engineered human cell lines" and provided as "fragmented human genomic DNA (average size 160 bp) resembling cell tumor DNA derived from human plasma." — Horizon Discovery "What are cfDNA Reference Standards?" FAQ page, fetched 2026-08-27 [16]

"Available at allelic frequencies down to 0.1% across a range of variants with a matched wild type." — same page [16]

"No commutability statement is provided" — this pass's own reading of the fetched page content; recorded as an absence, not a finding [16]

Gaps: no independent commutability study located; matrix composition (synthetic plasma, not native donor plasma) is stated by the vendor itself, which is directly relevant to idea-003's claimed differentiator but is not itself evidence that the synthetic matrix under-performs a native-plasma matrix — that comparison was not located this pass.

Twist Bioscience cfDNA Reference Standards

• What it is: the Twist cfDNA Pan-Cancer Reference Standard (v2), covering 400+ variant sites across 84 genes, marketed for liquid-biopsy and MRD assay development.
• Matrix: a Twist-published flyer describes the company's own contrived-specimen methodology (for a related MRD panel product, not necessarily identical to the reference standard itself) as **"blending synthetic variant sequences with fragmented cell line gDNA"** — again not pooled native donor plasma.
• Lowest characterized allele fraction: the reference standard is sold across multiple VAF levels; the specific lowest vendor-labeled level was not independently confirmed by a directly-fetched primary document this pass (a technical variant-list PDF exists but did not render as readable text to this pass's tools — recorded as not retrieved, not as "no low level exists"). Independently, a peer-reviewed benchmarking study reported reaching 0.0017% VAF using Twist's reference standard as one of three inputs to a research assay's own dilution/method, which is a researcher's achieved detection floor using the material, not the vendor's own labeled specification [13].
• Evidence behind it / Adoption: not established this pass beyond the benchmarking use in [13].
• Abandonment record — why people stop: not found; no vendor-specific discontinuation record located. Shared cross-vendor precision-degradation finding [12] applies here as general domain context, not as a Twist-specific finding (that study used a different, independently-manufactured multiplex cfDNA reference material, not a Twist product).
• Verbatim:

"[creating] contrived specimens by blending synthetic variant sequences with fragmented cell line gDNA" — WebFetch summary of a Twist Bioscience flyer, fetched 2026-08-27 [17]

Gaps: the vendor's own lowest labeled VAF for this specific SKU was not independently confirmed (PDF rendering failure, recorded honestly as not retrieved); no independent commutability study located.

Cross-vendor finding — precision degradation below 0.5% VAF, independent of any single vendor

• What it establishes: across eight domestic NGS-based cancer IVD manufacturers tested against a common multiplex cfDNA reference material (not one of the three vendors above, but directly on point for the mechanism question idea-003 claim 16 raises), "[t]he coefficient of variation (CV) was increased with decreasing mutant allelic frequency, and poor repetition occurred when the allelic frequency was lower than 0.5%."
• Source: [12], Journal of Cancer, 2018, n=8 manufacturers, 4 NGS platforms (NextSeq, HiSeq, Ion Proton, BGISEQ), reference materials at serial allelic frequencies from ~2% down to 0.1%.
• Verbatim:

"The coefficient of variation (CV) was increased with decreasing mutant allelic frequency, and poor repetition occurred when the allelic frequency was lower than 0.5%." — PMID 30405853 [12]

Read this against idea-003's claim 16 (the Poisson-limited-imprecision risk, previously recorded as entirely unsourced — "no connector verifies a physical claim"). This is an independently published, quotable finding that a comparable reference-material class shows degraded repeatability specifically below 0.5% VAF, across multiple platforms and manufacturers. It does not resolve claim 16's specific 0.001%-0.01% VAF sub-range (this study's lowest tested level was 0.1%, not the candidate's target floor), and it is not a bench study of this candidate's own proposed material — but it is real, independently retrieved, class-2 evidence bearing directly on the mechanism-risk factor that was previously recorded as having no evidence in either direction.

External quality assurance — a related but distinct signal

• What it establishes: in an 89-laboratory pilot EQA program using contrived reference material at predicted VAFs of 0.5%-2.5%, "88 of the 89 laboratories" returned results, and "Z-score proficiency testing found that >92% of clinical laboratories were concordant."
• Source: [6], Pathology Oncology Research, 2020, n=89 laboratories (China, Australia, New Zealand), reference material at 6 genotypes across 4 genes (EGFR, KRAS, BRAF, NRAS).
• Verbatim:

"Predicted genotypic variant allelic frequencies ranged between 0.5% - 2.5%... The average consensus data for each cfDNA genotype identified allelic frequencies ranging between 0.39%

• 4.4%... Z-score proficiency testing found that >92% of clinical laboratories were

concordant." — PMID 31487000 [6]

Read this narrowly, the discipline this section requires. This is the closest thing found this pass to independent evidence of the white space idea-003 claim 17 asserts (no product characterized in the sub-0.01% VAF MRD range): a real, published multi-laboratory EQA program existed and its reference material's operating range (0.39%-4.4% VAF) sits entirely above the candidate's target range (0.001%-0.5%), and above two of the three named vendors' own lowest labeled level (0.1%). This is evidence the general EQA/reference-material field currently operates well above MRD-range sensitivity, which supports the white-space claim without answering the class-3 question of whether any lab would buy a product that closed it.

No commercial reference product's discontinuation, recall, or withdrawal from distribution was located for any of the three named vendors. This is recorded as not found, not as "none exists" — a discontinuation record for an RUO reagent (as opposed to a cleared device) is not the kind of fact any US federal registry captures, and no vendor press archive was searched specifically for a product-line sunset announcement.

Update, as_of: 2026-09-25 — sub-0.1% products, head-to-head evidence against patient cfDNA, and prices

Correction to the first pass, stated before anything else. The first pass recorded the incumbents' lowest vendor-labeled level as 0.1% and read the EQA evidence as showing the field "operates well above MRD-range sensitivity." That was incomplete. A named incumbent sells a product labeled below 0.01%: SeraCare/LGC's Seraseq ctDNA MRD Panel Mix (catalog 0710-2146) is listed at tumor-fraction levels "0%, 0.5%, 0.05% & 0.005%", format "Purified DNA in buffer (Tris/EDTA)", 4 vials × 20 µL at 10 ng/µL (200 ng per vial) — product page fetched directly by this pass on 2026-09-25 and matching the fleet's reading [92]. The product's own datasheet (MKT-00626-01, 2021) qualifies how those levels were characterized: dPCR (digital polymerase chain reaction) was run only on the "0.5% tumor content equivalent = 0.25% VAF" blend, pre-fragmentation, with a six-variant average of 0.43% VAF; at the 0.005% tumor blend, NGS observed 4 variant copies across 4 of 100 variants; no uncertainty is stated at any level [75]. So the white-space claim in idea-003 claim 17 has to be read more narrowly than it was written: a sub-0.01% labeled product exists from an incumbent, in buffer rather than plasma matrix, with no dPCR value assignment or stated uncertainty at its low levels. What was not found is a product value-assigned at 0.001%-0.01% with a stated expanded uncertainty, in any matrix (§7). Inference, labelled: at the datasheet's own 0.5%-tumor = 0.25%-VAF ratio, the 0.005% tumor level would correspond to roughly 0.0025% VAF; the datasheet does not state that conversion for the low levels.

Head-to-head against native patient cfDNA — the closest thing to an abandonment record this domain has. The first pass found no independent comparison. The fleet's commutability report found one, plus several adjacent studies. None is an IFCC/CLSI EP30-design commutability study (§9), and no study compared native-plasma-matrix and synthetic-matrix formats of the same material on the same assay — both searched for and not found.

Also on record, lower weight: IMI CANCER-ID used Seraseq ctDNA "as purified ctDNA (study setup A) or encapsulated in a plasma-like matrix format (study setup B)" but on different assays, so not a matrix head-to-head [60]; an LGC-co-authored blood-TMB reference-material paper reports fragment sizes "similar to those found within patient plasma-derived cell-free DNA" — manufacturer-authored and without a patient comparator in the abstract [62]; the 2023 ELBS expert workshop (44 experts) emphasised materials that "closely mimic patient-derived ctDNA" [64].

What this section can now say about "why people stop," and what it still cannot. No published record of a laboratory discontinuing a named ctDNA reference material was found (the first pass's twin search, still empty — §9). What now exists is the nearest revealed-behaviour evidence short of that: an independent three-laboratory comparison documenting measured shortfalls of six named commercial materials against patient cfDNA (fragment-size profile, manufacturing-derived low-VAF background), and a multi-lab ring trial showing performance dropping at the 0.1% level. These are failure modes a user of the incumbents would meet. They are not a user's stated reason for leaving, and the same study's own recommendation (well-defined variants over native-likeness for sensitivity work) is a caution against reading them as demand for a native-matrix product. Inference, labelled: in that study the one material closest to native fragment size (Twist) was also among those with excess low-VAF background, so "native-like fragmentomics" and "fit for LoD" can pull apart — this bears directly on idea-003 claims 31 and 63, on one duplex-sequencing chemistry only.

Displayed prices — the first published price anchors for this domain. None of the three named vendors shows a price on its own store without login or quote [53][56][100]. Prices located, all single-source, all retrieved 2026-09-25, all distributor or storefront web prices, not vendor list prices (Fisher's page: "Promotional price valid on web orders only. Your contract pricing may differ" [94]):

Two corrections this makes to the first pass: the Twist v2 lowest labeled VAF (recorded then as "not retrieved") is 0.1% on the product sheet [100]; and there is a fourth RUO ctDNA control vendor in a plasma format that the candidate's predicate_or_analog does not name — Thermo Fisher AcroMetrix [101]. Inference, labelled: per-ng arithmetic on these web prices (≈$7.36/ng for the 125 ng AF0.1% plasma-like vial, ≈$2.40/ng for the MRD Panel Mix) is the fleet writer's own and must not stand in for a list price. No peer-reviewed methods section printing a reference-material price was found (four PubMed queries, 10-year window).

ISO 17034 status of the incumbents' pages: see §2's 2026-09-25 update — the SeraCare and Horizon ctDNA pages read cite ISO 9001/13485, not ISO 17034 [53][54][55][56].

6. Reimbursement landscape

as_of: 2026-09-25 (refreshed; first pass 2026-08-27 and 2026-09-15 addendum kept below)

Genuinely not applicable, stated explicitly rather than left blank. This is a laboratory reagent/consumable sold to CLIA laboratories and assay manufacturers as an operating input, the same disposition idea-003's own Stage 2 Scorer and Stage 4 Verifier already recorded for its Factor 2 ("no referent... nobody bills a payer for an RUO quality-control reagent"). No HCPCS (Healthcare Common Procedure Coding System) or CPT (Current Procedural Terminology) code applies to the reference material itself. The only reimbursement-adjacent fact in this domain is indirect and belongs to the downstream MRD test the reference material validates (CPT 0340U for Signatera, and the MolDX coverage article for tumor-informed MRD testing), both of which are already fully worked in idea-003's own Claims (claims 8-13) and are not re-derived here to avoid duplicating a candidate-specific pass in a file meant to outlive one candidate.

Update, as_of: 2026-09-15 — the downstream MolDX coverage article, actually opened. idea-003 claim 13 had recorded this as unconfirmed because the coverage connector keys on a HCPCS/CPT code, not a MolDX article ID, so a populated cache would never confirm this specific article's existence — a structural gap, not a fixable one. That gap is closed for the article-identification question by opening the CMS Medicare Coverage Database page for the article directly, which no connector in this repo does but which is the primary source of record the article ID names.

Article A58456, "Billing and Coding: MolDX: Minimal Residual Disease Testing for Solid Tumor Cancers," exists at https://www.cms.gov/medicare-coverage-database/view/article.aspx?articleId=58456, fetched 2026-09-15 [24]. It is the billing-and-coding companion to LCD (local coverage determination) L38816 (a Medicare local coverage determination number), "MolDX: Minimal Residual Disease Testing for Cancer," fetched the same day [25].

Coverage stance: conditional, not unconditional, and the article itself is retired. Both documents carry a Retired status with a revision-ending date of 2026-02-05, consolidated into the JE/JF jurisdiction's unified LCD documentation — the article and LCD this dossier and idea-003 cite were the coverage policy of record from 2022-01-02 through 2026-02-05, not the current one. LCD L38816's coverage indications, as fetched, state conditional (not blanket) coverage subject to ten named conditions, including: compliance with NCD (national coverage determination) 90.2 where NGS methodology is used; a personal history of cancer within the test's intended use; identification of recurrence/progression per NCCN guidelines requiring a management change; demonstrated sensitivity/specificity comparable to standard monitoring; peer-reviewed clinical validity for therapeutic monitoring; use within the test's validated population; and equivalent-or-superior performance versus similar-methodology tests [25]. Article A58456 itself states the billing-level conditions: "Each component and timepoint of the service must be reasonable and necessary and adhere to the criteria set forth in the related LCD," a frequency limit ("[t]he service may be performed once per patient per cancer diagnosis, unless there is clinical evidence of a priori change in genetic content" for NGS-based tests), and a documentation requirement ("[t]he medical record must contain clear documentation of the medical necessity of this test, as per the conditions outlined in the LCD") [24].

What retirement means here, stated plainly rather than left implicit. A retired article and LCD are not the same finding as "MolDX does not cover MRD testing" — search results locate at least one active successor LCD for the same policy title under a different LCD ID (L38835, among others, apparently one per MAC (Medicare administrative contractor) jurisdiction), consistent with the stated retirement reason (consolidation across jurisdictions) rather than a coverage reversal. The successor's own coverage text was not independently fetched this pass — recorded as not retrieved, not as "identical to the retired version." A reader must not assume the ten conditions above still hold verbatim in the active successor without checking it directly.

Gaps: the active successor LCD/article's own text was not independently fetched and compared to the retired L38816/A58456 pair; whether Medicare's stance changed at consolidation (as opposed to only the document ID and jurisdiction grouping) is unconfirmed either way. See idea-003 claim 67 for the claim drafted from this reading.

Update, as_of: 2026-09-25 — the CMS caches are populated; the downstream test's rate and its current billing articles are now connector-checkable. Still about the downstream MRD test, never the reference panel, which has no code (first paragraph above, unchanged). Every line here is a typed-connector replay against a provenance-recorded CMS file; the exact invocation is given so a Verifier can re-run it.

Payment rate. python3 -m connectors payment 0340U --min-expected 3590 --jurisdiction US → source_ref "local cache of CMS payment file, code=0340U, schedule=CLFS (clinical laboratory fee schedule), year=2026"; detail: "0340U pays $3,590.00, clearing the claimed floor of $3,590.00. Schedule: CLFS, year 2026. National limitation amount — the CLFS pays one national rate with no locality or wage-index adjustment" [33]. The cached row reads 0340U,CLFS,03590.00,2026,,20260101,N. The file is CMS 26clabq4.zip → PUF_CLFS_CY2026_Q4V1.csv, retrieved 2026-09-25T17:01:03Z, sha256 8b6ab628…0baf60ba in data/provenance.json. This supersedes, for the current year, the 2021-22-period $3,500 figure idea-003 claim 9 cites; it does not check that historical figure. Descriptors are not carried by the cache (AMA-licensed) — nothing here says what 0340U describes; idea-003 claim 8 remains the place that question lives.

Other codes listed in the same MolDX articles, same file: 0356U $1,800.00; 0422U $1,943.21; 0569U on the CY2026 CLFS with a published rate of 0.00 and pricing indicator "L" ("local/contractor pricing" — "not a payment of zero", per the connector); 81445 $597.91 [33]. Replay: python3 -m connectors payment <code> --jurisdiction US. Which test each PLA code describes is not stated here (AMA descriptors are not reproduced).

Coverage documents naming 0340U. python3 -m connectors coverage 0340U --jurisdiction US → "10 Medicare coverage document(s) name 0340U"; the connector's own caveat: "EXISTENCE ONLY — this does not establish that coverage is positive" [34]. Reading the cached rows directly (data/medicare_coverage_policies.csv, CMS MCD current_article.zip, retrieved 2026-09-25T17:05:49Z, sha256 e449b3a2…2f63; the cache carries no article prose and only status-A articles, per its provenance record), four of the ten are MRD-specific:

The other six are general MolDX "Molecular Diagnostic Tests (MDT)" and "Molecular Pathology and Genetic Testing" billing articles (A56853, A56973, A57526, A57772, A58917, A58918). The retired A58456/L38816 pair read on 2026-09-15 is not in this cache — consistent with the cache carrying only active (status A) articles, and consistent with the retirement recorded above; one "Response to Comments" article (A58954) in the cache still names L38816 as a related LCD.

A discrepancy between two readings of A58376, recorded rather than resolved. The fleet's laboratory-enumeration report read the A58376 web page and recorded "The only CPT/HCPCS code the article lists is 81479." The CMS bulk export for the same article lists six codes including 0340U (table above). Inference, labelled: the MCD web view typically withholds CPT-coded content (PLA codes are CPT Level I) behind an AMA licence acceptance, which would explain a page reading that shows only 81479; this pass did not re-open the page to confirm. The bulk-export reading is the one carried into claims, with its connector invocation; the web-page reading was not drafted as a claim.

What the article says about who is covered — from the fleet's web-page reading, not the cache. A58376 (revision effective 2026-08-06) names MRD tests from eight companies: Natera, Inc; NeoGenomics Laboratories, Inc; Exact Sciences Corp; Personalis, Inc; Guardant, Inc; SAGA Diagnostics; Veracyte, Inc; and Naveris, Inc [103]. Its parent LCD L38779 (original effective 2021-12-26) requires that an MRD test "satisfactorily completes a technical assessment (TA) that will evaluate and confirm that the analytical validity, clinical validity, and clinical utility criteria set in this policy are met" [104]. The hematologic companion article A58988 names one test, Adaptive's clonoSEQ [105]. These are coverage-document existence and naming, not a coverage stance; L38779's conditions were not compared line-by-line with the retired L38816's ten conditions recorded above.

What the rate implies for the candidate's TAM (total addressable market) arithmetic — inference, labelled, never a market size. $3,590.00 × 25,746 CY2024 Medicare FFS services ≈ $92.4M of Medicare FFS allowed-at-NLA spend on 0340U, mixing a CY2026 rate with CY2024 volume. This is arithmetic on two connector outputs from different years, excludes every non-FFS payer, and is not drafted as a claim.

7. Technology trajectory

as_of: 2026-09-25 (refreshed; first pass 2026-08-27 kept below, its closing "not found" partly superseded)

Constraint carried from the template: a counted series (clearances per year, entrants per year) cannot carry a confidence tag anywhere in this repo (docs/backlog.md 1.10); nothing of that shape is asserted here.

Gaps: no bench-level uncertainty-budget study (dPCR replicate series at stated DNA input mass) was located or is expected to exist as a published, citable source — idea-003's own Verifier already recorded this as a bench question no connector or literature scan settles, and this pass's literature search for "digital PCR uncertainty measurement ultra-low allele fraction circulating tumor DNA" returned zero PubMed records, confirmed not found, not merely not searched.

Update, as_of: 2026-09-25 — dPCR value assignment near the Poisson floor (REQ-0030; idea-003 claims 16, 62, 63)

The first pass's single zero-result query was too narrow. The fleet's report on this topic searched national-metrology-institute (NMI) and consortium dPCR literature and found the published floor of dPCR value assignment for ctDNA reference material. It still found nothing at 0.001%-0.01% — but it is now "searched for with NMI-targeted queries and not found," not "one query returned nothing."

What the published record says, narrowly. Every NMI dPCR quantitation or detection floor read sits at or above about 0.1% VAF, and no NMI or vendor document found value-assigns a ctDNA material below 0.1% with a stated expanded uncertainty. The one incumbent product labeled below that level (Seraseq MRD Panel Mix, §5) carries no dPCR assignment and no stated uncertainty at its low levels [75]. The lowest published uncertainty found (KRISS, ~0.25%-0.32%) is already 33%-90% relative. Most sources were read at abstract depth only.

The fleet writer's Poisson arithmetic — inference, labelled, carried because it is the question REQ-0030 asks. With N = genome equivalents × VAF, relative SD = 1/√N and P(0 copies) = e^−N: at 30 ng (≈9,091 GE) a 0.01% level holds ≈0.91 expected mutant copies (P0 ≈ 40%) and a 0.001% level ≈0.09 (P0 ≈ 91%); a 10% Poisson-only relative uncertainty (N = 100) needs ≈3.3 µg of DNA at 0.01% and ≈33 µg at 0.001%, before LoB, partition-volume, thresholding or heterogeneity terms. At the 0.04%-0.11% QX200 LoB reported by KRISS, or even a 0.003% false-positive rate, background at 0.001% would equal or exceed signal unless characterized and subtracted. This arithmetic is not a published finding and is not drafted as a claim. Note what it distinguishes: idea-003 claim 16 frames the floor at the assay's 10-30 ng input; the value-assignment step for a reference material is not bound to that input and can pool many wells — the arithmetic says how many (≈600-800 wells at KRISS loading, or ≈50 at an unsourced λ≈1 × 20,000-droplet loading), which is a cost and throughput question rather than an impossibility. Whether the laboratory using a 0.001% level at 10-30 ng input sees anything but Poisson noise is a separate question the same arithmetic answers unfavourably.

Update, as_of: 2026-09-25 — how variant-bearing cfDNA-like fragments are made (REQ-0033; idea-003 claims 20, 30)

• **MNase digestion of cell-line chromatin is a published route at two independent metrology groups.** Shanghai (TP53): made "by digesting nucleosomes from cultured cancer cell lines with micrococcal nuclease, followed by magnetic beads purification", "single dominant peaks at 128 bp (R175H) and 143 bp (R248W)" [80]. KRISS (EGFR): nucleosomal DNA via "the EZ Nucleosomal DNA Prep Kit (Zymo Research …), which incorporates MNase digestion", "approximately 150–170 bp", supplied in Tris-EDTA (not plasma), stable three years at −70 °C (full text) [72]. The kit is an off-the-shelf catalogue item, Zymo D5220 [93].
• Enzymatic shearing: NIM China nine-variant panel (0.33–23.9% VAF), "approximately 154-bp to 173-bp fragments using a laboratory-developed reaction system" [82]. Horizon Multiplex I: "fragmented to an average size of 160 bp", method not named [56]. Seraseq MRD Panel Mix: "Blended DNA is fragmented, sized to mimic ctDNA fragment sizes", method not named [92].
• Synthetic fragments: gravimetric BRAF material from "a 152 bp PCR amplicon and sonicated wild-type genomic DNA" [69]; FDA-led SEQC2 synthetic internal-standard spike-ins per target [83]; TÜBİTAK UME CRM 3022/3024 gravimetric 80–240 bp dsDNA for extraction recovery, not variant detection [84].
• Organoid/conditioned-media cfDNA: two independent groups report mononucleosome-sized cfDNA released by tumor organoids [85][86]; nuclease-treated conditioned media used "to simulate cfChromatin nucleosomal distributions" [87]; release shifts under chemotherapy toward "larger DNA fragments" [88].
• Matrix effect on the value itself: KRISS KRAS (2020): the choice of Atlantis dsDNase or MNase "affects DNA quantification", and spiking nucleosomal DNA into fetal bovine serum changed dPCR output significantly versus no-serum conditions [81].
• **Spiking into native pooled donor plasma as a characterized reference-material process: not found** at abstract depth (queries "healthy donor plasma spiked limit of detection circulating tumor DNA", "contrived samples ctDNA limit of detection"). Every material read is supplied in buffer or synthetic plasma. The CANCER-ID extraction study spiked into pooled donor plasma but as an extraction experiment, not a value-assigned reference material [61].
• No reference-material abstract reports a peak at exactly ~167 bp (128/143, 150–170, 154–173, avg 160, claimed 147–195 bp).

Issued US patents naming the nucleosomal-reference-material route — existence and term only, not an FTO (freedom to operate) reading. The fleet recorded that the patent connector returned "No EPO (European Patent Office) OPS record found" for these; this pass found that was an identifier-format artefact (the kind code): the same connector resolves all three when the kind code is dropped.

The connector returns title, filing date and a naive term; it does not return the assignee. Google Patents renders US9499870B2/US10577655B2 as Natera and US11215622B2 as Covaris — carried here as the fleet's reading, not as a registry fact. None of the three vendor pages read mentions a patent or licence [56][92][93]. Whether any of these claims reads on a pooled-donor-plasma matrix spiked with dPCR-quantified fragments is a claims-construction question for patent counsel, and nothing here answers it. This is three named documents in the space, not a landscape (REQ-0027's actual ask stays open — §9).

Inference, labelled (the fleet writer's): fragment production reads as an established published process with several routes and an off-the-shelf kit; the step specific to idea-003 — spiking into native pooled donor plasma with a dPCR uncertainty budget at 0.001–0.01% — was not found as a published process and on this evidence reads as development work, with matrix and fragment source as where that risk sits [81][58].

8. Market players & dynamics

as_of: 2026-09-25 (refreshed; first pass 2026-08-27 kept below)

Gaps: no funding-round, IPO, recall, or exit event for any of the three named reference-material vendors was located this pass — the search effort concentrated on product specification and analytical performance rather than corporate-transaction history beyond the one located acquisition (LGC/SeraCare, 2018).

Update, as_of: 2026-09-25 — which US laboratories have a public trace of offering an MRD test (REQ-0037's enumeration half; idea-003 claim 12's "30-60 laboratories"). Whether a lab would buy is class 3 and is not answered. Its twin — which labs have a named, public trace of offering or validating an MRD test — was researched by the fleet. Four evidence types, counted separately and never summed; every count is a floor that under-counts research-only and pharma/biotech in-house developers by construction.

Counts: MolDX A58376: 8 companies [103]; NYS CLEP: 1 registry record + 1 manufacturer statement (systematic enumeration not retrieved — the approved-LDT search page ignored filters or returned HTTP 500); published analytical validation: 5 US-offered assays; PLA on CLFS: the fleet could not retrieve it (cache empty then); now that the CLFS cache is warm, §6 shows four PLA codes (0340U, 0356U, 0422U, 0569U) listed in the MolDX MRD articles and on the CY2026 CLFS — but which laboratory holds which code is a descriptor question this repo does not answer. Inferences, labelled: "NSTX, Inc" is plausibly Natera's Texas laboratory; no source read says so. Test names in A58376 carrying "Bespoke Assay Design"/"Assay Design" suggest tumor-informed designs, while Guardant Reveal and NavDx are not tumor-informed — which qualifies the first-pass table row above that files Guardant Reveal under "tumor-informed MRD assay manufacturer" (kept as written, flagged here). The floor counts (8, 2, 5, 4) sit well below 30 but do not test the 30-60 range; the unretrieved NY CLEP enumeration and MolDX's technical-assessment (DEX) registry could close much of that gap.

Other players surfaced this pass: Thermo Fisher (AcroMetrix Multi-Analyte ctDNA Plasma Control, RUO, plasma format — a fourth ctDNA control vendor, §5) [101]; ATCC (the one US producer located with an ISO 17034 nucleic-acid scope and ddPCR value assignment, §2) [51]; Covaris and Natera (named on issued US patents over nucleosomal cfDNA reference material, §7 — assignee from Google Patents, not the registry) [89][91]. SensID (Rostock) appears as a material in the Hallermayr comparison [57]; its accreditation status was not found. nRichDx and Goffin were named by search results and not pursued.

9. Open questions — and what would settle them

Class-3 substitution performed this pass, stated plainly: the underlying question behind idea-003's claim 21 ("would labs prefer a third-party panel to in-house controls?") was not answered. Its revealed-behaviour twin — a documented vendor discontinuation or a published lab complaint driving a workaround — was searched for specifically and came up empty, which is itself the honest result of this pass, not a substitute answer to the original question. This differs from the potassium-monitoring-ckd dossier's experience with the same maneuver (§3.1 of docs/research-lane-spec.md), where a revealed-behaviour twin (RAASi discontinuation rates) was found; here, no twin was found, and that absence is recorded rather than papered over with the general precision-degradation finding, which is adjacent but is not itself a switching decision.

Update, as_of: 2026-09-25 — status of every row above, plus the questions this pass opened. The rows above are kept as the first pass wrote them; this table says where each stands now.

Class-3 substitutions this pass, stated plainly: claim 12's "addressable/obtainable" framing → twin: public traces of labs offering an MRD test (§8). REQ-0028's "a fast follower cannot skip" → twin: what the documents require/recommend/accept (§2). Claim 33's "achievable on budget and timeline" → twin: published fees, stated time-to-accreditation, cycle obligations (§2). Claims 26 and 28 (consumable purchase; no distributor) → twin: how incumbents' products are actually offered for sale (§3). Claim 21's "would labs prefer a third-party panel" → the first pass's twin (a discontinuation record) is still not found; the nearest revealed-behaviour evidence is now the measured shortfalls of six named materials against patient cfDNA (§5), which is a failure mode, not a stated reason for leaving.

10. Provenance & staleness

What this dossier has never looked at, as of 2026-08-27 (first pass): any non-US regulatory or reimbursement framework (this dossier is US-scoped per its frontmatter, per every connector in this repo); a systematic enumeration of US laboratories running tumor-informed MRD assays; SEC filings (10-K/10-Q) for Natera or Guardant, as opposed to press releases and an earnings-call transcript; CAP Molecular Pathology checklist or ISO 17034 text directly (both sit behind access barriers this pass did not attempt to cross); MAUDE (no connector exists, and would be silent by construction for RUO/LDT material in any case); ~~fda.gov guidance documents (the 2013 RUO/IUO distribution guidance central to idea-003 claims 2-3 remains unretrieved here exactly as it does there — this dossier does not duplicate that already-documented structural gap)~~; veterinary data (out of scope by design, no US federal dataset exists for it in any domain). This is a first-pass dossier, built in one research-lane session — it is a starting point for the next candidate in this domain, not the output of a completed research program.

Correction, as_of: 2026-09-15 — struck through above rather than deleted. The 2013 RUO/IUO distribution guidance was retrieved this pass (fetched directly from fda.gov via curl; see §2 addendum above and idea-003 claims 64-65). CAP Molecular Pathology checklist and ISO 17034 text, SEC 10-K/10-Q filings, and a systematic US-laboratory enumeration remain un-retrieved exactly as the first pass recorded.

Refresh, as_of: 2026-09-25 — what was folded in, how, and what it does not cover.

Method note that bears on every fleet-sourced line above. Except where this table says a line was replayed or re-fetched this pass, the sources in [26]-[109] were read out of the eight 2026-09-25 fleet reports, whose own section 5 records the read depth (mostly abstract depth) and the [single-source] / [manufacturer] markers. They are cited here by their primary identifiers, never by report, so an independent checker can re-retrieve them; this pass did not re-retrieve most of them, and this dossier does not vouch for them beyond what those reports and the spot-checks above recorded.

What this dossier has still never looked at, as of 2026-09-25: any non-US framework; CPT descriptors for any MRD PLA code (AMA-licensed — never reproduced); CAP MOL checklist text and ISO 17034 clause text (access barriers, not attempted); a patent landscape (no capability); MAUDE (no connector; silent by construction for RUO/LDT material); a licensed lab-economics report on QC spend; primary customer research of any kind (class 3 — no dossier line may substitute for it); veterinary data (out of scope, no US federal dataset exists).

---

Sources

[1] Highly Accurate Detection of Circulating Tumor DNA for Monitoring Minimal Residual Disease in Solid Tumors: Analytical and Clinical Validation of the Haystack MRD Assay. The Journal of Molecular Diagnostics, 2026 Aug 20. PMID 42624249.

[2] Analytical validation of a hybrid-approach combining tumor-informed and tumor-agnostic bespoke ctDNA panel assay for the sensitive detection of minimal residual disease. PLoS ONE,

1. PMID 41212866.

[3] Analytical validation of NeXT Personal®, an ultra-sensitive personalized circulating tumor DNA assay. Oncotarget, 2024 Mar 14. PMID 38484152.

[4] Personalized Cancer Monitoring Assay for the Detection of ctDNA in Patients with Solid Tumors. Molecular Diagnosis & Therapy, 2023 Nov. PMID 37632661.

[5] Harmonising ctDNA Measurement in Haematological Malignancies: Traceability, Commutability and Reporting. Diagnostics (Basel), 2026 Apr 1. PMID 41975767. (Title/abstract read; haematological-malignancy scope noted, not directly applied to a solid-tumor MRD claim in this dossier.)

[6] External Quality Assurance of Current Technology for the Testing of Cancer-Associated Circulating Free DNA Variants. Pathology Oncology Research, 2020 Jul. PMID 31487000.

[7] A pilot plasma-ctDNA ring trial for the Cobas® EGFR Mutation Test in clinical diagnostic laboratories. Clinical Chemistry and Laboratory Medicine, 2019 Apr 24. PMID 30267626. (No abstract available in the PubMed record; recorded for completeness of the EQA/ring-trial landscape search, not as a source of any quantitative claim in this dossier.)

[8] Circulating tumor DNA-guided adjuvant therapy in locally advanced colon cancer: the randomized phase 2/3 DYNAMIC-III trial. Nature Medicine, 2025 Dec. PMID 41115959.

[9] Post-adjuvant chemotherapy in ctDNA-positive patients with resected colorectal cancer: a randomized phase 3 trial (ALTAIR, CIRCULATE-Japan). Nature Medicine, 2026. PMID 42260101. ClinicalTrials.gov identifier NCT04457297 (a ClinicalTrials.gov registration number).

[10] Utility of circulating tumor DNA to detect minimal residual disease in colorectal cancer: A systematic review and network meta-analysis. International Journal of Cancer, 2025. PMID 40293388.

[11] Prognostic role of postoperative circulating tumor DNA in metastatic colorectal cancer treated with curative intent: A systematic review and meta-analysis. Cancer Treatment Reviews, 2026. PMID 42217424.

[12] Multiplex Cell-Free DNA Reference Materials for Quality Control of Next-Generation Sequencing-Based In Vitro Diagnostic Tests of Colorectal Cancer Tolerance. Journal of Cancer,

1. PMID 30405853.

[13] GeneBits: ultra-sensitive tumour-informed ctDNA monitoring of treatment response and relapse in cancer patients. Journal of Translational Medicine, 2025 Aug 27. PMID 40866952.

[14] 42 CFR 493.1256, "Standard: Control procedures" — retrieved via python3 -m connectors regulation 493.1256 --title 42 --expect "control procedures" --jurisdiction US, status verified, source_ref "eCFR versioner title-42, issue 2026-08-13, section=493.1256", re-run fresh for this dossier on 2026-08-27.

[15] LGC/SeraCare, "Seraseq® ctDNA Reference Material v4 AF0.1%" product page, https://www.seracare.com/Seraseq-ctDNA-Reference-Material-v4-AF01-0710-3992/, fetched 2026-08-27.

[16] Horizon Discovery (Revvity), "What are cfDNA Reference Standards?" FAQ page, https://horizondiscovery.com/en/frequently-asked-questions/reference-standards/what-are-cfdna-reference-standards, fetched 2026-08-27.

[17] Twist Bioscience, "High sensitivity detection of ultra low-frequency somatic mutations for minimal residual disease monitoring" (flyer), https://www.twistbioscience.com/resources/flyer/high-sensitivity-detection-ultra-low-frequency-somatic-mutations-minimal-residual, fetched 2026-08-27.

[18] Natera, Inc., "Natera Reports Second Quarter 2026 Financial Results," press release dated 2026-08-06, https://investor.natera.com/news/news-details/2026/Natera-Reports-Second-Quarter-2026-Financial-Results/default.aspx, fetched 2026-08-27.

[19] Guardant Health, Inc., Q2 2026 earnings call transcript, call held 2026-07-30, published by The Motley Fool, https://www.fool.com/earnings/call-transcripts/2026/08/07/guardant-health-gh-q2-2026-earnings-call-transcript/, fetched 2026-08-27.

[20] LGC, "LGC Acquires Seracare Life Sciences, Strengthening Its Position in Clinical Quality Control Tools," press release dated 2018-11-26, https://www.seracare.com/about-seracare/press-releases/lgc-acquires-seracare-life-sciences-strengthening-its-position-in-clinical-quality-control-tools/, fetched 2026-08-27.

[21] FDA, Center for Devices and Radiological Health and Center for Biologics Evaluation and Research, "Distribution of In Vitro Diagnostic Products Labeled for Research Use Only or Investigational Use Only — Guidance for Industry and Food and Drug Administration Staff," Docket No. 2011-D-0305, issued 2013-11-25, https://www.fda.gov/files/medical%20devices/published/Distribution-of-In-Vitro-Diagnostic-Products-Labeled-for-Research-Use-Only-or-Investigational-Use-Only---Guidance-for-Industry-and-FDA-Staff.pdf, fetched 2026-09-15 (direct curl; WebFetch on the same URL returned HTTP 401).

[22] Federal Register, "Regulation Identification Number 0910-AJ05 Medical Devices; Laboratory Developed Tests; Implementation of Vacatur," 90 FR 45134, document number 2025-18239, Health and Human Services Department / Food and Drug Administration, published 2025-09-19, effective 2025-09-19 — retrieved via python3 -m connectors policy 2025-18239 --expect "vacated" --jurisdiction US, status verified, source_ref "Federal Register document_number=2025-18239 (90 FR 45134)", run 2026-09-15. https://www.federalregister.gov/documents/2025/09/19/2025-18239/regulation-identification-number-0910-aj05-medical-devices-laboratory-developed-tests-implementation

[23] Papagiannopoulos, T. (Sidley Austin LLP), "American Clinical Laboratory Association v. FDA," Food and Drug Law Institute, published 2026-06-24, https://www.fdli.org/2026/06/acla-v-fda/, fetched 2026-09-15.

[24] CMS Medicare Coverage Database, Article A58456, "Billing and Coding: MolDX: Minimal Residual Disease Testing for Solid Tumor Cancers," status Retired (revision ending 2026-02-05), companion article to LCD L38816, https://www.cms.gov/medicare-coverage-database/view/article.aspx?articleId=58456, fetched 2026-09-15.

[25] CMS Medicare Coverage Database, LCD L38816, "MolDX: Minimal Residual Disease Testing for Cancer," status Retired effective 2026-02-05 (retirement reason stated as consolidation into unified JE/JF jurisdiction documentation), effective period 2022-01-02 to 2026-02-05, https://www.cms.gov/medicare-coverage-database/view/lcd.aspx?lcdid=38816, fetched 2026-09-15.

Sources [26]-[109] added 2026-09-25. Except where §10's refresh table says a source was replayed through a connector or re-fetched by this pass, each was read out of one of the eight 2026-09-25 fleet reports for this domain and is cited here by its own primary identifier.

[26] Natera, Inc. Form 10-Q, quarter ended 2026-06-30, filed 2026-08-07. SEC accession 0001628280-26-054525, https://www.sec.gov/Archives/edgar/data/1604821/000162828026054525/ntra-20260630.htm, accessed 2026-09-25 (phrase matched via EDGAR full-text search [32]).

[27] Natera, Inc. Form 10-Q, quarter ended 2026-03-31, filed 2026-05-08. SEC accession 0001628280-26-032478, https://www.sec.gov/Archives/edgar/data/1604821/000162828026032478/ntra-20260331.htm, accessed 2026-09-25.

[28] Natera, Inc. Form 10-Q, quarter ended 2025-06-30, filed 2025-08-08. SEC accession 0001558370-25-010829, https://www.sec.gov/Archives/edgar/data/1604821/000155837025010829/ntra-20250630x10q.htm, accessed 2026-09-25.

[29] Natera, Inc. Form 8-K Exhibit 99.2, Q2 2026 earnings presentation, filed 2026-08-06. SEC accession 0001604821-26-000015, https://www.sec.gov/Archives/edgar/data/1604821/000160482126000015/ntra-20260630xex992.htm, accessed 2026-09-25. (Exhibit 99.1 of the same filing is the press release already cited as [18].)

[30] Guardant Health, Inc. Form 10-Q, quarter ended 2026-06-30, filed 2026-07-30. SEC accession 0001576280-26-000037, https://www.sec.gov/Archives/edgar/data/1576280/000157628026000037/gh-20260630.htm, accessed 2026-09-25.

[31] Guardant Health, Inc. Form 8-K Exhibit 99.1, Q2 2026 financial results, filed 2026-07-30. SEC accession 0001576280-26-000036, https://www.sec.gov/Archives/edgar/data/1576280/000157628026000036/gh-06302026xexhibit991.htm, accessed 2026-09-25.

[32] SEC EDGAR full-text search, exact-phrase queries restricted to CIK 0001604821 (Natera) and CIK 0001576280 (Guardant), filing dates 2023-01-01 to 2026-09-25, queried 2026-09-25, https://efts.sec.gov/LATEST/search-index. Several queries returned HTTP 500.

[33] CMS Clinical Laboratory Fee Schedule, CY2026 Q4 public use file (26clabq4.zip → PUF_CLFS_CY2026_Q4V1.csv), https://www.cms.gov/files/zip/26clabq4.zip, retrieved 2026-09-25T17:01:03Z, cache sha256 8b6ab6288a003cb00fd1a4a4708934bc6d154c219efa5addb990000f0baf60ba (data/provenance.json). Read via python3 -m connectors payment <code> --jurisdiction US, run 2026-09-25.

[34] CMS Medicare Coverage Database, current article export (current_article.zip → current_article_csv.zip), https://downloads.cms.gov/medicare-coverage-database/downloads/exports/current_article.zip, retrieved 2026-09-25T17:05:49Z, cache sha256 e449b3a23520734dcd3c7221653cce43e2c2ad890af4c3516938e8ba7d022f63; carries no article prose and status-A articles only. Read via python3 -m connectors coverage 0340U --jurisdiction US and by opening data/medicare_coverage_policies.csv, 2026-09-25.

[35] CMS "Medicare Physician & Other Practitioners — by Geography and Service", MUP_PHY_R26_P05_V10_D24_Geo.csv, https://data.cms.gov/sites/default/files/2026-05/e534c74b-79b8-4892-8a95-5a17e2dfec9f/MUP_PHY_R26_P05_V10_D24_Geo.csv, retrieved 2026-09-24T17:30:29Z, sha256 ce9e4f30897241b328ef4e65963c9699ad53c0db64956e13dab5ac46323be634. Read via python3 -m connectors procedures <code> --jurisdiction US, run 2026-09-25.

[36] 42 CFR 493.1253, "Standard: Establishment and verification of performance specifications" — python3 -m connectors regulation 493.1253 --title 42 --expect "Analytical sensitivity" --jurisdiction US, source_ref "eCFR versioner title-42, issue 2026-08-13, section=493.1253", run 2026-09-25.

[37] 42 CFR 493.1255, "Standard: Calibration and calibration verification procedures" — python3 -m connectors regulation 493.1255 --title 42 --expect "if possible, traceable to a reference method or reference material of known value" --jurisdiction US, source_ref "eCFR versioner title-42, issue 2026-08-13, section=493.1255", run 2026-09-25.

[38] College of American Pathologists, "Test Method Validation and Verification: Excerpts from CAP All Common Checklist", checklist release dated 2017-08-21 (COM.40300, COM.40400), https://documents-cloud.cap.org/appsuite/learning/LAP/FFoC/ValidationVerificationStudies/story_content/external_files/checklistrequirements.pdf, accessed 2026-09-25.

[39] College of American Pathologists, "2020 Checklist Edition: Revised Discipline-Specific Checklist Change Summary" (MOL.36118), June 2020, https://documents-cloud.cap.org/appsuite/learning/LAP/FoC_Webinars/2020ToolkitChecklistUpdate.pdf, accessed 2026-09-25.

[40] Check W., "NGS checklist takes in infectious disease testing," CAP TODAY, October 2017, https://www.captodayonline.com/ngs-checklist-takes-infectious-disease-testing/, accessed 2026-09-25. (No DOI; secondary description of a checklist, not checklist text.)

[41] Jennings LJ et al., "Guidelines for Validation of Next-Generation Sequencing-Based Oncology Panels: A Joint Consensus Recommendation of the AMP and CAP," J Mol Diagn 19(3), 2017. PMID 28341590; PMC6941185; DOI 10.1016/j.jmoldx.2017.01.011.

[42] Lockwood CM et al., "Recommendations for Cell-Free DNA Assay Validations: A Joint Consensus Recommendation of the AMP and CAP," J Mol Diagn, 2023. PMID 37806433; DOI 10.1016/j.jmoldx.2023.09.004. (Abstract only; full text 403.)

[43] CLSI EP17-A2, "Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures," 2nd ed., 2012-06-18, product page https://clsi.org/shop/standards/ep17/, accessed 2026-09-25.

[44] BSI Knowledge, "BS EN ISO 17034:2016 General requirements for the competence of reference material producers," published 2016-11-30, https://knowledge.bsigroup.com/products/general-requirements-for-the-competence-of-reference-material-producers-1, accessed 2026-09-25.

[45] A2LA, "ISO 17034 - Reference Materials Producers Accreditation Program" (undated page), https://a2la.org/accreditation/reference-material-producer/, accessed 2026-09-25.

[46] A2LA R101, "General Requirements: Accreditation of Conformity Assessment Bodies" incl. Annex D (Q ID 5573), publication date 2026-09-01, https://portal.a2la.org/documents/serveDoc.cfm?id=32, accessed 2026-09-25. (A2LA F336 scope selection list, https://portal.a2la.org/documents/serveDoc.cfm?id=106, last revised 2020-11-17, is the companion.)

[47] A2LA, "Frequently Asked Questions" (undated), https://a2la.org/faq/, accessed 2026-09-25.

[48] A2LA, "Accreditation Estimate Requests" (undated web form), https://a2la.org/estimate-request/, accessed 2026-09-25.

[49] Perry Johnson Laboratory Accreditation, "Reference Material Producer ISO 17034" (undated), https://www.pjlabs.com/accreditation-programs/reference-material-producer, accessed 2026-09-25.

[50] ANAB AR 2258, "Accreditation Requirements: ISO 17034 Reference Material Producers," effective 2025-10-29, https://anab.qualtraxcloud.com/ShowDocument.aspx?ID=12549, accessed 2026-09-25.

[51] ANAB Certificate Number AR-1384, Certificate and Scope of Accreditation to ISO 17034:2016, American Type Culture Collection, scope version 011 issued 2025-03-10, valid to 2027-03-17, https://delivery-files.atcc.org/api/public/content/92779-ATCC17034CertScope.pdf, accessed 2026-09-25.

[52] A2LA Cert. No. 2655.02, Scope of Accreditation to ISO 17034:2016, Microbiologics, Inc., revised 2026-02-13, valid to 2028-02-29, https://www.microbiologics.com/core/media/media.nl?id=112655&c=915960&h=905e588efc9433863fb7&_xt=.pdf, accessed 2026-09-25.

[53] SeraCare / LGC Clinical Diagnostics, "Seraseq ctDNA Complete Reference Material AF0.5%," catalog no. 0710-0672, https://www.seracare.com/Seraseq-ctDNA-Complete-Reference-Material-AF05-0710-0672/, accessed 2026-09-25.

[54] LGC Standards, "Our Quality" (undated), https://www.lgcstandards.com/LS/en/ourquality, accessed 2026-09-25.

[55] Horizon Discovery / Revvity, "Reference standards product information" (undated), https://horizondiscovery.com/en/reference-standards/reference-standards-product-information, accessed 2026-09-25.

[56] Horizon Discovery / Revvity, "Mimix Multiplex I, cfDNA Reference Standard Set," catalog nos. HD776, HD777, HD778, HD779, https://horizondiscovery.com/en/reference-standards/products/multiplex-i-cfdna-reference-standard-set, accessed 2026-09-25.

[57] Hallermayr A et al., "Impact of cfDNA Reference Materials on Clinical Performance of Liquid Biopsy NGS Assays," Cancers (Basel) 15(20):5024, 2023-10-17. PMID 37894392; PMC10605119; DOI 10.3390/cancers15205024. (Abstract re-read by this pass via literature-detail.)

[58] Giroux S et al., "SeraSeq reference materials support non-invasive cfDNA screening methods," Clinical Biochemistry, online 2025-11-11. PMID 41232587; DOI 10.1016/j.clinbiochem.2025.111046.

[59] Hernandez KM et al., "Contrived Materials and a Data Set for the Evaluation of Liquid Biopsy Tests: A BLOODPAC Community Study," J Mol Diagn, 2023 Mar. PMID 36828596; DOI 10.1016/j.jmoldx.2022.12.003.

[60] Weber S et al., "Technical Evaluation of Commercial Mutation Analysis Platforms and Reference Materials for Liquid Biopsy Profiling," Cancers (Basel) 12(6):1588, 2020-06-16. PMID 32560092; PMC7352370.

[61] Lampignano R et al., "Multicenter Evaluation of Circulating Cell-Free DNA Extraction and Downstream Analyses for the Development of Standardized (Pre)analytical Work Flows," Clinical Chemistry, 2020-01-01. PMID 31628139; DOI 10.1373/clinchem.2019.306837.

[62] Raiber-Moreau EA et al., "Development and validation of blood tumor mutational burden reference standards," Genes Chromosomes Cancer, online 2022-12-07. PMID 36326821; PMC10107199. (LGC co-authored; LGC manufactured the materials.)

[63] Hayden RT et al., "Impact of Fragmentation on Commutability of Epstein-Barr Virus and Cytomegalovirus Quantitative Standards," J Clin Microbiol, 2019-12-23. PMID 31619529; PMC6935903.

[64] Deans ZC et al., "Reference materials and external quality assessment for liquid biopsy assays: Expert opinions from the European Liquid Biopsy Society ctDNA workshop," Cancer Treatment and Research Communications, online 2026-02-19. PMID 41734680.

[65] Xu J et al., "Construction of a reference material panel for detecting KRAS/NRAS/EGFR/BRAF/MET mutations in plasma ctDNA," J Clin Pathol, 2021 May. PMID 32817175; PMC8070650.

[66] "Towards standardisation of cell-free DNA measurement in plasma: controls for extraction efficiency, fragment size bias and quantification," Anal Bioanal Chem, 2014. PMID 24853859; PMC4182654.

[67] "Validation of ctDNA Quality Control Materials Through a Precompetitive Collaboration of the Foundation for the National Institutes of Health," JCO Precision Oncology, 2021-06-01. PMID 34250423; PMC8232894. (Abstract re-read by this pass via literature-detail.)

[68] "Assessment of Digital PCR as a Primary Reference Measurement Procedure to Support Advances in Precision Medicine," Clinical Chemistry, 2018-09. PMID 29903874.

[69] "Interlaboratory assessment of droplet digital PCR for quantification of BRAF V600E mutation using a novel DNA reference material," Talanta, 2020-01-15. PMID 31594564.

[70] "Establishment of primary reference measurement procedures and reference materials for EGFR variant detection in non-small cell lung cancer," Analytical Methods, 2021-05-13. PMID 33870958.

[71] "International co-validation on absolute quantification of single nucleotide variants of KRAS by digital PCR," Anal Bioanal Chem, 2022-08. PMID 35810409. (Listed for completeness: the abstract gives no FA level or uncertainty size, so no quantity here rests on it.)

[72] "Development and characterization of nucleosomal DNA-based reference materials for the epidermal growth factor receptor gene liquid biopsy," Anal Bioanal Chem, 2026. PMID 41936644; PMC13221357; DOI 10.1007/s00216-026-06475-5. (Abstract re-read by this pass via literature-detail; the level/uncertainty figures are from the fleet's full-text read.)

[73] "New candidate reference measurement procedure for NRAS p.Q61R quantification by digital PCR," Methods, 2026-10. PMID 42431334.

[74] "An Automated Correction Algorithm (ALPACA) for ddPCR Data Using Adaptive Limit of Blank and Correction of False Positive Events Improves Specificity of Mutation Detection," Clinical Chemistry, 2021. PMID 33842952.

[75] SeraCare Life Sciences, "Seraseq ctDNA MRD Panel Mix Datasheet," MKT-00626-01, 2021, catalog no. 0710-2146, https://www.seracare.com/globalassets/seracare-resources/ds-mkt-00626-0710-2146-seraseq-ctdna-mrd-panel-mix.pdf, accessed 2026-09-25.

[76] "Multilaboratory Assessment of a New Reference Material for Quality Assurance of Cell-Free Tumor DNA Measurements," J Mol Diagn, 2019-07. PMID 31055023. (NIST multilaboratory study; dPCR at 2%, 0.5%, 0.125%, 0% VAF; a commercial material "formulated to simulate the size distribution of circulating wild-type and tumor DNA in a synthetic plasma matrix".)

[77] "Evaluation of droplet digital PCR and next generation sequencing for characterizing DNA reference material for KRAS mutation detection," Scientific Reports, 2018-11-30. PMID 30504843.

[78] NIST, "NIST HER2 Genomic DNA Standard Reference Material SRM 2373," catalog no. SRM 2373, https://www.nist.gov/programs-projects/nist-her2-genomic-dna-standard-reference-material-srmr-2373, accessed 2026-09-25.

[79] NIST, "Measurement of low frequency mutations in circulating tumor DNA (ctDNA)" project page, updated 2025-03-26, https://nist.gov/programs-projects/measurement-low-frequency-mutations-circulating-tumor-dna-ctdna-through, accessed 2026-09-25.

[80] Tang Y et al., "Establishment of Reference Measurement Procedure for TP53 R175H/R248W Detection and a Novel Preparation Method for ctDNA Reference Material," Genes (Basel), 2025-05-14. PMID 40428398; DOI 10.3390/genes16050576.

[81] "Improvement of digital PCR conditions for direct detection of KRAS mutations," J Clin Lab Anal, 2020-08. PMID 32329932; DOI 10.1002/jcla.23344.

[82] "Development of a nine-variant reference material panel to standardize cell-free DNA detection," Anal Bioanal Chem, 2024-07. PMID 38782781; DOI 10.1007/s00216-024-05336-3.

[83] "Advancing NGS quality control to enable measurement of actionable mutations in circulating tumor DNA," Cell Reports Methods, 2021-11-22. PMID 35475002.

[84] "Certified Reference Materials for Standardization of Cell-Free DNA Isolation Recovery in Liquid Biopsy," J Mol Diagn, 2026-06. PMID 41936819.

[85] "Multidimensional fragmentomic profiling of cell-free DNA released from patient-derived organoids," Human Genomics, 2023-10-28. PMID 37898819.

[86] "Characterization of the cell-free DNA released by tumor organoids derived from colorectal cancer patients," European Journal of Cancer, 2025-12-09. PMID 41223563.

[87] "Simulating cell-free chromatin using preclinical cancer models for liquid biopsy applications," iScience, 2025-12-19. PMID 41497400.

[88] "Cell-Free DNA Release Kinetics and Fragmentation Reflect Treatment Response and Resistance In Vitro," Cells, 2026-08-28. PMID 42738856.

[89] US Patent 9,499,870 B2, "Cell free DNA diagnostic testing standards" — python3 -m connectors patent US9499870 --jurisdiction US, source_ref "EPO OPS published-data/publication epodoc=US9499870", run 2026-09-25 (filed 2014-09-26; naive term end 2034-09-26). Claim text: Google Patents, https://patents.google.com/patent/US9499870B2/en, read by the fleet 2026-09-25.

[90] US Patent 10,577,655 B2, "Cell free DNA diagnostic testing standards" — python3 -m connectors patent US10577655 --jurisdiction US, source_ref "EPO OPS published-data/publication epodoc=US10577655", run 2026-09-25 (filed 2016-01-14; naive term end 2036-01-14). https://patents.google.com/patent/US10577655B2/en.

[91] US Patent 11,215,622 B2, "Generation of cfDNA reference material" — python3 -m connectors patent US11215622 --jurisdiction US, source_ref "EPO OPS published-data/publication epodoc=US11215622", run 2026-09-25 (filed 2018-01-23; naive term end 2038-01-23). Claim text: Google Patents, https://patents.google.com/patent/US11215622B2/en, read by the fleet 2026-09-25.

[92] SeraCare / LGC Clinical Diagnostics, "Seraseq ctDNA MRD Panel Mix," catalog no. 0710-2146, https://www.seracare.com/Seraseq-ctDNA-MRD-Panel-Mix-0710-2146/, fetched directly by this pass 2026-09-25.

[93] Zymo Research, "EZ Nucleosomal DNA Prep Kit," catalog no. D5220, https://zymoresearch.com/products/ez-nucleosomal-dna-prep-kit, accessed 2026-09-25.

[94] Fisher Scientific, "SERASEQ CTDNA MRD PANEL MIX," catalog no. NC2587386, https://www.fishersci.com/shop/products/NC2587386/NC2587386, price displayed 2026-09-25.

[95] Fisher Scientific, Seraseq ctDNA Complete Reference Material AF0.1%, 1x5mL, catalog no. NC1626188, https://www.fishersci.com/shop/products/seraseq-reference-mat-af0-1/NC1626188, price displayed 2026-09-25.

[96] Fisher Scientific, Seraseq ctDNA Complete Reference Material AF0.5%, 1x5mL, catalog no. NC1626186, https://www.fishersci.com/shop/products/seraseq-reference-mat-af0-5/NC1626186, price displayed 2026-09-25.

[97] Fisher Scientific, Seraseq ctDNA Complete Mutation Mix AF 0.5%, 25 uL, catalog no. NC1717818, https://www.fishersci.com/shop/products/seraseq-ctdna-comp-mut-af-0-5-1/NC1717818, price displayed 2026-09-25.

[98] Fisher Scientific, ctDNA Mutation Mix v2 AF0.125%, 25 uL, catalog no. NC1814090, https://www.fishersci.com/shop/products/ctdna-mutation-mix-afo-13/NC1814090, price displayed 2026-09-25.

[99] SeraCare / LGC Clinical Diagnostics, "Seraseq ctDNA Complete Reference Material AF0.1%," catalog no. 0710-0673, https://www.seracare.com/Seraseq-ctDNA-Complete-Reference-Material-AF01-0710-0673/, accessed 2026-09-25.

[100] Twist Bioscience, "Twist cfDNA Pan-Cancer Reference Standard v2," product sheet DOC-001467 REV 2.0, catalog no. 107576 (singles 107577–107590), https://www.twistbioscience.com/content/dam/twistbioscience/resources/2024-03/DOC-001467%20cfDNA%20Pan-Cancer%20Reference%20Standard%20v2%20REV2%20SINGLES.pdf, accessed 2026-09-25.

[101] Thermo Fisher Scientific, "AcroMetrix Multi-Analyte ctDNA Plasma Control A - E (RUO)," catalog nos. 957563 (Control A) and 957566 (Control D), https://www.thermofisher.com/order/catalog/product/957563, price displayed 2026-09-25.

[102] Fisher Scientific, Thermo Scientific AcroMetrix Multi-Analyte ctDNA Plasma Control A (RUO), catalog no. 10-095-7563, https://www.fishersci.com/shop/products/acrometrix-multi-analyte-ctdna-plasma-control-a-e-ruo/100957563, price displayed 2026-09-25.

[103] CMS Medicare Coverage Database, Article A58376, "Billing and Coding: MolDX: Minimal Residual Disease Testing for Solid Tumor Cancers," companion to LCD L38779, Palmetto GBA, revision effective 2026-08-06, https://www.cms.gov/medicare-coverage-database/view/article.aspx?articleid=58376, accessed 2026-09-25 (web-page reading by the fleet; see §6 for the code-list discrepancy with [34]).

[104] CMS Medicare Coverage Database, LCD L38779, "MolDX: Minimal Residual Disease Testing for Cancer," Palmetto GBA, original effective 2021-12-26, https://www.cms.gov/medicare-coverage-database/view/lcd.aspx?lcdid=38779, accessed 2026-09-25.

[105] CMS Medicare Coverage Database, Article A58988, "Billing and Coding: MolDX: Minimal Residual Disease Testing for Hematologic Cancers," revision effective 2025-04-24, https://www.cms.gov/medicare-coverage-database/view/article.aspx?articleid=58988, accessed 2026-09-25.

[106] NYSDOH Wadsworth Center CLEP, approved laboratory developed test record, facility "NSTX, Inc" (PFI 9456, project 103554), analyte "Signatera" (record undated), https://www.wadsworth.org/node/563674/printable/print, accessed 2026-09-25.

[107] Quest Diagnostics, "Quest Diagnostics Receives New York State Approval for Haystack MRD®…," press release 2026-06-24, https://newsroom.questdiagnostics.com/2026-06-24-Quest-Diagnostics-Receives-New-York-State-Approval-for-Haystack-MRD-R-,-Broadening-Patient-Access-to-ctDNA-Minimal-Residual-Disease-Testing, accessed 2026-09-25.

[108] "Analytical Validation of NavDx, a cfDNA-Based Fragmentomic Profiling Assay for HPV-Driven Cancers," Diagnostics (Basel), 2023-02-14. PMID 36832208.

[109] "Analytical validation of a circulating tumor DNA assay using PhasED-Seq technology for detecting residual disease in B-cell malignancies," Oncotarget, 2025-05-09. PMID 40353795.

Words on this page30 abbreviation(s) and code(s), with what each stands for — copied from _shared/plain-language.md

CDRH

Center for Devices and Radiological Health — the part of the FDA that handles devices

cfDNA

cell-free DNA — DNA fragments floating in the blood, most of them from normal cells

CLFS

clinical laboratory fee schedule — the Medicare price list for laboratory tests

CLIA

Clinical Laboratory Improvement Amendments — the US rules a clinical laboratory must meet to report test results on patients

CMS

Centers for Medicare & Medicaid Services — the US agency that runs Medicare, and so decides what Medicare pays for

CPT

Current Procedural Terminology — the US billing codes for what a clinician does; owned by the American Medical Association

ctDNA

circulating tumour DNA — fragments of a tumour's DNA found in the blood

dPCR

digital polymerase chain reaction — a laboratory method that counts individual DNA molecules

EPO

European Patent Office — the European patent office

FDA

US Food and Drug Administration — the US regulator that clears or approves a device before it can be sold

FFS

fee-for-service — the traditional form of Medicare that pays per item or service

FTO

freedom to operate — whether a product can be sold without infringing someone else's live patent

HCPCS

Healthcare Common Procedure Coding System — the US billing codes for equipment, supplies and some services; a code is how a payer recognises what to pay for

ISO

International Organization for Standardization — the body whose numbered standards devices and laboratories are tested against

IVD

in vitro diagnostic — a test run on a sample taken from the body rather than on the body

L#####

a Medicare local coverage determination number — the public id of one local coverage rule; the letter L and five digits

LCD

local coverage determination — a Medicare contractor's rule, valid in its region, saying when something is covered

LDT

laboratory-developed test — a test a laboratory builds and runs itself rather than buys

MAC

Medicare administrative contractor — a regional contractor that processes Medicare claims and writes local coverage rules

MAUDE

Manufacturer and User Facility Device Experience — the FDA's public database of reported device problems

Medicare Advantage

Medicare Advantage — Medicare delivered through a private insurer; must cover at least what Medicare covers

MRD

molecular residual disease — cancer left behind after treatment, too little to see on a scan, detected through DNA in the blood

NCD

national coverage determination — a Medicare rule, valid everywhere in the US, saying whether and when something is covered

NCT########

a ClinicalTrials.gov registration number — the public id of one registered clinical trial

NGS

next-generation sequencing — the laboratory method that reads many DNA fragments at once

PMID

PubMed identifier — the public id of one published medical paper

QC

quality control — the samples and checks a laboratory runs to prove its test is still working

RUO

research use only — a labelling status that keeps a product out of patient diagnosis; no FDA review, no clinical claims

TAM

total addressable market — how much money the whole market would be worth if every possible buyer bought

VAF

variant allele frequency — what share of the DNA fragments at one position carry the tumour's mutation; the lower it is, the harder the test