CMS Part B buy-and-bill covers all four approved 1L NSCLC IO agents at ASP+6%, but PD-L1 assay requirements split coverage into three distinct biomarker-testing tiers, and only 46% of US patients receive the full guideline biomarker panel before first-line therapy begins.
All four FDA-approved first-line NSCLC immunotherapies are physician-administered infused biologics, so Medicare routes their coverage through Part B buy-and-bill (HCPCS J-code, reimbursed at ASP+6%), not the Part D formulary architecture that governs oral oncology drugs. That single fact sets the access frame: the fight for share happens at the biomarker-testing and prior-authorization layer, not the pharmacy benefit tier.
Pembrolizumab monotherapy requires PD-L1 IHC 22C3 pharmDx testing at a TPS threshold of 50% or higher, per its KEYNOTE-024 label indication. Nivolumab plus ipilimumab carries no PD-L1 threshold at all, per CheckMate-227 and CheckMate-9LA. Atezolizumab requires a third assay entirely, SP142 IHC, a different antibody clone that is not interchangeable with the 22C3 test most practices already run. Three agents, three distinct testing burdens, and the prior-authorization documentation a practice must assemble changes with each one.
Durvalumab sits outside that fight. Its approved use is Stage III unresectable NSCLC consolidation after concurrent chemoradiotherapy, a Part B-covered but functionally separate access lane from the four-way first-line metastatic competition, per PACIFIC. New entrants targeting first-line metastatic disease will be evaluated against the KEYNOTE, CheckMate, and IMpower coverage precedent already written into payer medical policy, not against a fresh read of their own trial design. Below, we map that Part B coverage architecture in full and run a gap analysis against your clinical package.
Exhibit 5 — CMS Part B buy-and-bill governs all four approved 1L NSCLC IO agents, but PD-L1 assay requirements split them into three distinct biomarker-testing tiers.
| Agent | Part B routing | Biomarker / PA requirement | Approved setting | Evidence basis |
|---|---|---|---|---|
| PembrolizumabKeytruda · Merck | Buy-and-bill, ASP+6% | PD-L1 IHC 22C3 pharmDx required for monotherapy (TPS ≥50%); no threshold for chemo-combination | 1L monotherapy (high PD-L1) or 1L chemo-combination, any histology | KEYNOTE-024 (PMID 27718847), KEYNOTE-189 (PMID 29658856), KEYNOTE-407 (PMID 30280635) |
| Nivolumab + ipilimumabOpdivo + Yervoy · BMS | Buy-and-bill, ASP+6% | No PD-L1 threshold in label; PA burden centers on combination-dosing duration limits, not biomarker testing | 1L IO-IO doublet, all histologies | CheckMate-227 (PMID 29658845), CheckMate-9LA (PMID 34126067) |
| AtezolizumabTecentriq · Roche/Genentech | Buy-and-bill, ASP+6% | SP142 IHC required — a distinct assay/clone from the 22C3 test already run for pembrolizumab-eligible patients | 1L combination regimens | IMpower110 (PMID 34280284), IMpower150 (PMID 29863955) |
| DurvalumabImfinzi · AstraZeneca | Buy-and-bill, ASP+6% | No PD-L1 threshold; does not compete in the 1L metastatic PA architecture at all | Stage III unresectable, post-concurrent chemoradiotherapy consolidation | PACIFIC (PMID 28885881) |
Sources: FDA prescribing information (pembrolizumab, nivolumab, ipilimumab, atezolizumab, durvalumab). KEYNOTE-024: Reck et al., NEJM 2016, PMID 27718847. KEYNOTE-189: Gandhi et al., NEJM 2018, PMID 29658856. KEYNOTE-407: Paz-Ares et al., NEJM 2018, PMID 30280635. CheckMate-227: Hellmann et al., NEJM 2018, PMID 29658845. CheckMate-9LA: Reck et al., Lancet Oncol 2021, PMID 34126067. IMpower110: Spigel et al., NEJM 2021, PMID 34280284. IMpower150: Socinski et al., NEJM 2018, PMID 29863955. PACIFIC: Antonia et al., NEJM 2017, PMID 28885881. Medicare Part B buy-and-bill and ASP+6% reimbursement mechanics per CMS program rules. On a commissioned brief, each coverage figure is re-verified against the client's specific plan at point of writing.
Five questions. Each section is built to answer one of them.
Every section answers a named commercial question your team is asking, scoped to your asset.
Delivers
- Coverage posture by agent · biomarker and line restrictions · policy citations
Delivers
- PA criteria · step edits · documentation burden by payer archetype
Delivers
- Evidence threshold from incumbent precedents · endpoint expectations
Delivers
- Gap analysis vs. your trial design · risk areas · mitigations
Delivers
- Economic evidence expectations · comparator framing · value-story inputs
Scoped to your trial design and proposed label — not the category average.
Scope Your WorkWhat's inside
- Why coverage precedent, not your trial, sets the bar
- The access decisions to make pre-approval
- What secondary policy review resolves vs. primary payer research
- Coverage posture by agent — broad, restricted, step-gated
- Biomarker and line restrictions in policy
- Regional and PBM variation
- PA criteria and documentation burden
- Step edits and their commercial impact
- Where PA narrows a broad label
- The threshold from incumbent precedents
- Endpoint and comparator expectations
- What payers discount or reject
- Where your clinical package clears the bar
- Risk areas and likely pushback
- Evidence-generation options to close gaps
- Economic evidence expectations
- Comparator framing and value story
- Budget-impact considerations
- Which policy positions are current vs. evolving
- Inputs that drive access-timing variance
- Sensitivity on coverage-decision timing
- Payer conversations to start 12–18 months out
- Evidence gaps requiring primary payer research
- Decisions contingent on label and trial outcomes
Included with every brief
Every figure is live-sourced before delivery. If a number cannot be verified, it does not appear.
Prepared by MoatRx analysts.
This is a field where AI confidently reproduces outdated epidemiology, superseded payer policy, and retracted analyses. AXLRx uses none of its own memory as a source. Every figure your team receives is verified against a live document at the time of writing.
A wrong number in front of your payer or your leadership team is not recoverable in the same meeting.
- Every claim cited to a live PMID, ClinicalTrials.gov ID, or URL at point of writing — uncited claims are dropped, not estimated
- PubMed metadata fetched live during authoring — model memory produces incorrect author and journal data even on correct PMIDs
- Numeric cross-check: the specific figure must appear in the cited source, not merely be consistent with its topic
- Independent audit pass after generation — broken links, unsourced claims, and numeric inconsistencies flagged before delivery
- Drop gate: any figure that cannot clear the above is removed. No confidence tiers. No exceptions.
Frequently asked questions
Tell us your asset. Your team has the intelligence in 72 hours.
We build from your asset's clinical profile: mechanism, biomarker strategy, proposed label, and target cohort. Scope confirmation takes one call.
Drug, mechanism, proposed indication, target cohort, geography. Five minutes via the intake form.
We confirm scope with your team, clarify ambiguities, and lock delivery timing.
PDF intelligence document, Excel model, and optional executive deck — with a 30-minute readout call included.