🧬 Molecular Tumor Board Consensus Recommendation Tracker

From Discussion to Document — How a Molecular Tumor Board Recommendation Is Issued

A molecular tumor board (MTB) is a multidisciplinary meeting — medical oncologists, pathologists, molecular geneticists, pharmacists, and often clinical trial staff — that reviews a patient's tumor sequencing results against the current evidence base and produces a single, structured recommendation. That recommendation is only useful if it is captured as a discrete, actionable artifact, not left as verbal consensus in a meeting room. The report card format — target, evidence tier, suggested therapy — is what allows the rest of the pipeline (communication, implementation, outcome tracking) to function at all.

• 1–2: MTB meetings per week (academic ctr.) (typical major cancer center cadence)
• 8–15: Median cases discussed per session (varies widely by institution)
• ~30–50%: Actionable alterations found (of sequenced tumors, tier I–IIC)
• <48 h: Report turnaround (discussion→doc) (quality-improvement target)

What an MTB actually produces

A molecular tumor board takes as input a molecular profiling report — typically a next-generation sequencing (NGS) panel covering anywhere from 50 to >500 genes, sometimes paired with RNA fusion panels, immunohistochemistry, or whole-exome/whole-genome data — and produces as output a clinical recommendation. This is not a research exercise; it is meant to directly inform the treating oncologist's next decision.

The structured output typically includes:

• Molecular target: the specific alteration considered actionable (e.g., EGFR exon 19 deletion, KRAS G12C, BRAF V600E, ERBB2 amplification, an NTRK fusion) • Evidence tier: a standardized rating of how strong the evidence is that targeting this alteration will help this patient in this tumor type — commonly using frameworks like AMP/ASCO/CAP tiers (I–IV) or OncoKB levels (1–4, R1–R2) • Suggested therapy: an FDA-approved drug, an off-label use of an approved drug, or enrollment in a matched clinical trial • Rationale and caveats: co-occurring alterations, prior therapy history, and any factors that might modify the recommendation's strength

Without this discrete documentation step, MTB recommendations exist only as meeting-minute prose or verbal impressions — nearly impossible to track, audit, or hold the system accountable to.

Evidence tiers and why they matter for what happens next

Not all molecular recommendations carry equal weight, and the evidence tier assigned at issuance strongly predicts what happens downstream. The AMP/ASCO/CAP consensus framework (Li et al., J Mol Diagn 2017) is widely used:

• Tier I (strong clinical significance): FDA-approved therapy in this exact tumor type, or professional guideline-listed biomarker — e.g., osimertinib for EGFR exon 19del/L858R in lung adenocarcinoma • Tier II (potential clinical significance): FDA-approved therapy in a different tumor type ("basket" evidence), well-powered clinical trial data, or multiple smaller studies with consensus — e.g., larotrectinib for an NTRK fusion found in a non-approved histology • Tier III (unknown significance): variant of uncertain significance without therapeutic implication yet • Tier IV (benign/likely benign): not actionable

Tier I recommendations are the easiest to implement — there is a labeled drug, an insurance-approvable indication, and established supply chains. Tier II and lower recommendations frequently require off-label prescribing, prior authorization appeals, or clinical trial enrollment — all of which introduce friction that shows up later as lower implementation rates and longer time-to-treatment.

The Handoff — Getting the Recommendation to the Person Who Can Act on It

A perfect recommendation that never reaches the treating oncologist, or reaches them buried in an unread EHR inbox message three weeks later, has zero clinical value. Communication is the most under-studied and most fixable link in the MTB chain: it depends entirely on institutional workflow design, not on the scientific quality of the recommendation itself.

• ~10–20%: Recommendations lost to poor handoff (never reach acting clinician (est.))
• ~60–75%: EHR inbox message read within 48h (institution-dependent)
• +15–25 pts: Direct verbal/phone follow-up effect (implementation rate uplift (reported))
• growing: Dedicated MTB navigator programs (shown to improve follow-through)

Communication channels and their reliability

Institutions vary widely in how MTB output reaches the treating oncologist, and the channel chosen has a measurable effect on whether anything happens next:

• Passive EHR documentation: the MTB note is filed in the chart; the oncologist must actively look for it during the next visit. Lowest-friction to produce, highest risk of being missed entirely, especially if the next visit is weeks away. • EHR inbox message / task: a direct message or task assignment routed to the ordering or treating oncologist. Better than passive filing, but subject to inbox overload — oncologists routinely manage hundreds of inbox items per week. • Direct verbal or phone communication: an MTB coordinator or the presenting oncologist calls or messages the treating physician directly. Highest reliability, highest resource cost — not scalable to every case at high-volume centers without dedicated staff. • Patient-facing communication: increasingly, some programs also loop in the patient directly (patient portal summary), which can create a second channel of accountability and patient-driven follow-up.

Many quality-improvement initiatives now formalize this step with a designated "MTB navigator" or "precision oncology coordinator" role, whose explicit job is closing the loop — confirming the treating oncologist received, read, and acted (or explicitly declined to act) on the recommendation.

Why the handoff gap is so consequential

Retrospective studies of MTB programs consistently identify communication and follow-through — not lack of actionable findings — as a major bottleneck. A molecular result can be actionable, arrive on time, and still fail to change care simply because it never crossed the desk of the person with prescribing authority in a form they registered as urgent and specific to this patient.

This matters because the treating oncologist is often not the same physician who presented the case at the MTB — the sequencing may have been ordered by one clinician, discussed by a tumor board composed of specialists who may never see the patient in clinic, with the resulting recommendation needing to be relayed to a community oncologist, a different subspecialist, or a covering physician. Each handoff point is a place the signal can degrade.

Followed or Not — Measuring Whether MTB Recommendations Actually Change What Happens to the Patient

Implementation tracking is the single most cited and most sobering metric in the molecular tumor board literature. Across published series, only a minority of MTB recommendations are actually implemented — commonly cited in the range of roughly 20–50% depending on institution, patient population, and how strictly "implementation" is defined. Understanding why recommendations go unimplemented is essential to improving the whole system.

• 20–50%: Implementation rate range (literature) (wide variance across published MTB cohorts)
• ~25–35%: Top reason: patient factors (decline, deterioration, death, comorbidity)
• ~20–30%: Top reason: access/insurance (denial, prior-auth delay, drug cost)
• ~15–20%: Top reason: drug/trial unavailable (no local access, trial closed/full)

What the published implementation literature actually shows

A number of retrospective single- and multi-institution studies have quantified MTB recommendation implementation, and although absolute numbers vary considerably by study design, patient population, and how "implementation" is operationally defined, a consistent pattern emerges: a substantial fraction — often more than half — of recommendations are never actually implemented.

Reported implementation rates cluster roughly in the 20–50% range across the literature, with academic centers with dedicated precision oncology infrastructure and clinical trial access tending toward the higher end, and community-facing or resource-limited settings tending toward the lower end. Some series restricted to Tier I / highest-evidence recommendations report higher implementation (sometimes >50%), reflecting that stronger, FDA-labeled evidence is simply easier to act on than an off-label or basket-trial suggestion.

Importantly, "not implemented" does not always mean "recommendation failed" — many non-implementations are clinically appropriate (e.g., the patient's disease progressed too rapidly to wait for drug access, or the patient chose hospice) rather than system failures. Careful implementation studies try to separate "clinically appropriate non-implementation" from "missed opportunity."

The four major barrier categories

Across studies, barriers to implementation cluster into a recurring set of categories:

• Patient decline or preference: the patient chooses not to pursue the recommended therapy — due to toxicity concerns, quality-of-life priorities, travel burden for a trial, or simply preferring to stop active treatment. • Clinical deterioration: the patient's performance status declines, or the disease progresses so rapidly, that by the time the recommendation is actionable the patient is no longer a candidate for further systemic therapy — this is one of the most time-sensitive barriers and directly motivates why time-to-implementation tracking (Stage 4) matters so much. • Insurance / financial / access denial: prior authorization denial, the drug not being covered for this off-label indication, or out-of-pocket cost being prohibitive. This barrier is disproportionately concentrated in Tier II–III (off-label) recommendations, since Tier I on-label indications are far more reliably covered. • Drug or trial unavailability: the recommended agent is not commercially available (investigational only), the matched clinical trial has no open slot, has geographic restrictions, or has closed enrollment by the time the patient is ready.

Quality-improvement programs that specifically track barrier category (rather than just a binary implemented/not-implemented flag) are far better positioned to intervene — for example, standing up a dedicated prior-authorization support team measurably improves the insurance-denial bucket without touching the others.

Why implementation rate is a program-level accountability metric

Tracking implementation rate over time, and stratifying it by evidence tier, tumor type, and barrier category, converts the MTB from a purely advisory forum into a program with measurable accountability. Programs that publish and review their own implementation metrics tend to identify and fix systemic issues — for example, discovering that off-label Tier II recommendations for a specific payer are being denied at unusually high rates, and building a standing appeal template in response.

The Clock That Matters Most — Time from Recommendation to Therapy Start

Even a recommendation that is ultimately implemented can arrive too late to matter. Time-to-implementation — the interval between the MTB recommendation date and the actual start of the recommended therapy — is increasingly tracked as a core quality metric, because in rapidly progressing cancers, weeks of delay can mean the difference between a patient being eligible for the recommended therapy and not.

• ~2–6 wks: Reported median time-to-implementation (varies widely by series and barrier)
• ≤14 days: QI benchmark target (leading centers) (recommendation to treatment start)
• days–weeks: Delay driver: prior authorization (single largest controllable delay)
• ~1–2 wks: Delay driver: drug procurement/shipping (specialty pharmacy logistics)

Why the clock starts at the MTB date, not the sequencing order date

Different programs define the clock's start point differently — some measure from tissue collection, some from sequencing order, some from result availability — but for tracking the MTB's own operational performance specifically, the most meaningful start point is the date the recommendation is finalized at the tumor board. This isolates the piece of the pipeline the MTB and its supporting infrastructure can actually control: everything upstream (biopsy scheduling, sequencing turnaround, pathology review) is a separate — and also important — quality metric, while everything from the MTB date forward (communication speed, prior authorization, drug procurement, patient scheduling) is what "time-to-implementation" is meant to capture.

The stop point is typically the date of first dose of the recommended therapy, or first day of protocol therapy for a clinical trial enrollment, giving a clean, auditable interval.

What drives the interval, and where the time actually goes

Breaking down a typical multi-week time-to-implementation interval, the time is rarely dominated by clinical decision-making itself — it is dominated by administrative and logistic friction:

• Communication and scheduling: several days to a week between MTB output and the patient's next scheduled oncology visit where the plan can be discussed and consented to. • Prior authorization: often the single largest controllable delay — insurance prior-authorization review for a specialty oncology drug can take anywhere from a few days (expedited/on-label) to several weeks (off-label appeal process), and denials that require peer-to-peer review or formal appeal add further delay. • Specialty pharmacy and drug procurement: oral targeted therapies are frequently dispensed through specialty pharmacies with their own intake, insurance verification, and shipping timelines, commonly adding another week. • Trial-specific delays: if the recommendation is clinical trial enrollment, screening procedures (repeat imaging, washout periods from prior therapy, additional biomarker confirmation) can extend the interval substantially beyond what a straightforward drug prescription would require.

Leading precision oncology programs that specifically track and report this interval as a dashboard metric — and assign a QI benchmark such as ≤14 days — have been able to identify and shorten the largest bottlenecks, most often by building dedicated prior-authorization support and pre-negotiated specialty pharmacy pathways for their most commonly recommended agents.

Time-to-implementation is not just an efficiency metric — for patients with rapidly progressing disease, every week of delay is a week in which clinical deterioration can convert an implementable recommendation into a non-implementable one. This is why several published MTB quality-improvement initiatives treat time-to-implementation as important as implementation rate itself, since the two metrics interact directly: shortening the clock is one of the few levers that can also raise the implementation rate.

Did It Work? Measuring Whether Recommended Therapy Changed Management and Improved Outcomes

The final and hardest-to-measure link in the chain asks the question that ultimately justifies the entire MTB infrastructure: among patients who received the recommended therapy, did outcomes actually improve relative to what would have happened otherwise — and, more broadly, did the recommendation change the treatment plan the patient would have received in the MTB's absence ("management change")? Retrospective cohort studies comparing MTB-guided therapy against standard-of-care or non-matched therapy consistently show favorable trends, though methodological limitations (selection bias, small numbers, heterogeneous tumor types) mean these results should be read as encouraging signal rather than definitive proof.

• ~20–40%: Cases with documented management change (of all cases reviewed, varies by series)
• often 2×: Response rate, matched-therapy cohorts (vs. unmatched/standard therapy (reported))
• favorable: PFS benefit, matched vs. unmatched (in most retrospective MTB cohort studies)
• mixed/positive: Overall survival benefit (signal present, RCT-level proof limited)

Defining "management change" as the core impact metric

"Management change" is the metric that separates a truly impactful MTB recommendation from documentation that had no bearing on what actually happened to the patient. It is typically defined as: the treatment the patient actually received differs from what would have been prescribed absent the MTB recommendation — most commonly meaning the patient received a targeted therapy, immunotherapy, or clinical trial enrollment they would not otherwise have received, in place of, or in addition to, standard cytotoxic chemotherapy.

Across published MTB program evaluations, reported management-change rates commonly fall in a wide range — roughly 20–40% of all cases reviewed — reflecting that management change is a stricter bar than implementation alone: a recommendation can be "implemented" in a narrow sense (patient received the suggested agent) while still overlapping heavily with what standard care would have offered anyway, which is not usually counted as true management change.

Retrospective outcome comparisons: response rate and progression-free survival

Because randomized controlled trials directly testing "MTB recommendation followed vs. not followed" are difficult to conduct (it would require randomizing patients to receive or not receive genomically matched therapy, which raises ethical and practical issues once an actionable target is identified), most outcome evidence comes from retrospective cohort comparisons.

A recurring finding across multiple retrospective series (including landmark work from MD Anderson's IMPACT program and similar precision oncology cohorts at other academic centers) is that patients who received genomically matched therapy following molecular profiling tend to show:

• Higher objective response rates compared with patients who received non-matched (standard) therapy for the same molecular findings • Longer progression-free survival (PFS) in the matched-therapy cohort — differences on the order of several months in various published cohorts • A trend toward improved overall survival in matched-therapy patients, though this signal is less consistent and more subject to confounding (patients well enough to receive complex matched therapy may simply be healthier at baseline — a form of selection bias that retrospective studies struggle to fully correct for)

These studies collectively support genomically matched therapy as beneficial, but the field still lacks large prospective randomized evidence directly isolating the MTB's incremental value, which is why interpretation should remain appropriately cautious even as the directional signal is consistently positive across many independent cohorts.

The MTB program impact score — rolling metrics into program-level accountability

Mature precision oncology programs increasingly report a composite "program impact" dashboard rather than any single metric in isolation, because implementation rate, time-to-implementation, and management-change rate interact and can each mask problems in the others. A composite view — such as the percentage of all reviewed cases resulting in a documented management change, tracked longitudinally alongside implementation rate and time-to-implementation — gives tumor board leadership and hospital administration a defensible way to demonstrate the clinical and resource value of continuing to invest in MTB infrastructure, molecular pathology capacity, and precision oncology navigator staffing.

Quality-improvement cycles built around this dashboard — for example, standing up a prior-authorization support service after seeing insurance denial dominate the barrier breakdown, or renegotiating specialty pharmacy turnaround after seeing procurement dominate the time-to-implementation interval — represent the practical endpoint of building a tracking system in the first place: not just measuring the pipeline, but using the measurements to shorten it.

The throughline across the MTB outcome-tracking literature is consistent: molecular tumor boards generate substantial numbers of scientifically sound, actionable recommendations, but a large fraction of that potential benefit is lost not to bad science but to operational friction — communication gaps, insurance barriers, and procurement delays. Programs that treat implementation, time-to-implementation, and management-change rate as measured, reported, and continuously improved metrics — rather than assuming a good recommendation automatically becomes good care — are the ones shown to convert molecular insight into measurable patient benefit.