CMS is required to hold an advisory expert panel to help it price new laboratory tests. This year, the panel brought together 10 members on July 23, 2025, and worked through 90 agenda items in one televised workday.
Find the panel results here (PDF, 27pp):
https://www.cms.gov/files/document/clfs-advisory-panel-recommendations-2025.pdf
Find the home page for the panel here:
(You find the PDF by scrolling down to, "Panel Recommendations.")
##
What happened?
The panelists vote individually, and there can be several pricing options to vote on per code. Most of the votes are either 10/10 (unanimous for one pricing option) or 9/10 (one member varies, or one member abstains). There were a couple codes where no single option got more than 5 or 6 votes of 10.
About 70 of the votes were for "crosswalk" pricing and about 20 for gapfill pricing. About 90% of the crosswalk votes are for a single crosswalk price x1. (A few votes were for multiples or fractions).
##
What happens next?
CMS is required to "consult with an expert outside advisory panel." However, when CMS releases its proposed federal prices for 2026, it's under no obligation to match the panel vote on any particular code, and often its prefered choices vary from where the panelists settled.
Look for CMS proposed prices sometime in the first half of September, followed by a comment period, and final prices in the second half of November.
Example
One of the more complex votes was for item #9, code X159U, transplantation medicine. All 10 panelists voted, 1 for "gapfill," 3 for 0018U x 85%, and 6 for crosswalk to 0069U.
AI CORNER
I asked Chat GPT to see what it could make of the 27 page, 92-item PDF. I can't or haven't checked all these numbers, so the AI discussion is presented "as is."
###
Big picture.
The panel favored crosswalk over gapfill by roughly 4:1. In 90 items, 19 winners were gapfill (21%), while 71 winners were crosswalks (79%). Of those crosswalk wins, 63 (70% of all items) were single-code, straight crosswalks (e.g., 0538U → 81459), and only 8 (9%) used fractions/multiples (e.g., 87185 × 3; 87811 + 87804 × 2). Unanimity was common: 49 items (54%) were 10-0 wins—broken down as 36 straight crosswalks, 1 composite crosswalk, and 12 unanimous gapfills. The median winning vote count was 10, with a tight distribution (only four items won with ≤6 votes—the narrowest was 5 votes). Abstentions occurred on 12 items (13%), typically on less familiar or edge-case tests.
Gapfill—when did it prevail?
Gapfill was the majority (≥6 votes) and top choice in 19 items and accounted for ~23% of all individual votes cast across the day. It dominated where novel analytic paradigms or poorly comparable predecessors made crosswalks awkward—especially infectious-disease NGS and new biomarkers:
• Unanimous gapfill for plasma metagenomic ID sequencing (0531U), urine NGS for 44 organisms (X177U), sepsis marker PSP (X182U), and TB CFP-10 (0574U).
• In neurology, pTau217 (0551U) was 10–0 gapfill.
• Oncology had at least one notable gapfill: 0569U (ctDNA methylation) was 10–0 gapfill, while most other oncology codes crosswalked (see below).
Crosswalks—how “clean” vs composite?
When crosswalks won, they were usually clean. Only 8 winners used fractions/multiples (11% of crosswalk wins), and across all crosswalk votes (not just winners), ~11% were for composite options. Typical composite winners included:
• 871XX → 87185 × 3 (9–1) and 871X1 → 87150 × 5 (8–2) (phenotypic/genotypic resistance assays).
• 8XXXX (COVID/Flu antigen) → 87811 + 87804 × 2 (8–1–1).
• 0550U (prostate algorithm) → 81539 × 0.5 (10–0).
• 0420U (urothelial RNA+ddPCR signature) → 0012M + 0356U × 0.35 (9–1).
• X184U (APOE peptides) was the day’s closest call: 5 votes for 0412U × 0.33, with 2 for “0412U” and 2 for gapfill.
How fractured were crosswalk preferences?
Not very. In 64 items, all crosswalk votes went to a single comparator; only 22 items split crosswalk votes across ≥2 options, and just one item split across three composite variants (0567U, constitutional WGS). Net: when panelists chose to crosswalk, they converged on one target most of the time.
By content area (illustrative patterns).
-
Oncology. Consistent alignment with the NGS families: tissue CGP to 81459 (e.g., 0538U, 0543U), plasma CGP to 81464 (0539U, 0530U), MRD setup/follow-up to 0306U/0307U (0560U/0561U). Protein/RNA signatures commonly crosswalked to established algorithm codes (81525, 81539, 81503). An outlier: 0569U (ctDNA methylation) was 10–0 gapfill. Unanimity was frequent in this cluster.
-
Infectious disease. Gapfill swept the novel ID NGS and new biomarker entries (see above), while common modalities crosswalked cleanly: respiratory NAAT panels to 87633 (e.g., 0563U, 0564U), joint-space multiplex to 0505U, and HSV/VZV NAAT to 87631. The COVID/Flu combo antigen used a composite crosswalk (87811 + 87804 × 2). Lyme IgG (X166U) was 10–0 gapfill.
-
Transplantation. Donor-derived cfDNA tests 0540U and 0544U crosswalked to 0493U; the combined cfDNA + expression test X160U crosswalked to 81558 (with 1 vote for adding 0118U). The liver miRNA rejection test (X159U) gives a good snapshot of how the panel weighs comparators: 0069U won (6 votes) over a fractional crosswalk to 0018U × 0.85 (3) and gapfill (1).
-
Neurology. pTau217 (0551U) was 10–0 gapfill; NfL and GFAP single-analyte assays crosswalked to 0361U; a TBI panel crosswalked to 0358U. The APOE prototype (X184U) was the most divided item (noted above).
What a payer or submitter should take from this.
-
If there’s a clear incumbent analog, expect a straight crosswalk. The panel rarely resorted to arithmetic constructions; composites were exceptions, not the rule.
-
Use gapfill arguments when the analytic paradigm is genuinely new (e.g., broad untargeted ID NGS, emerging protein biomarkers, or methylome approaches without obvious cousins). That’s where the panel repeatedly embraced gapfill—often with 10–0 calls.
-
Consensus was strong (median 10 votes for the winner). If you anticipate a crosswalk, aligning your dossier to the dominant comparator the panel tends to pick (e.g., 81459/81464 families; 0306U/0307U; 0493U) is a defensible strategy. If you need composite pricing, be prepared: only ~11% of all crosswalk votes—and 8 winners—went that route.
A few precise stats for your slide:
-
Winners: 19 gapfill (21%); 63 single-code crosswalk (70%); 8 composite crosswalk (9%).
-
Unanimous (10–0): 49 items (36 single-code CW, 1 composite CW, 12 gapfill).
-
Gapfill share of all panelist votes: ~22.7% (204 of 900 votes, including abstentions).
-
Crosswalk vote consolidation: 64 items had all crosswalk votes on one target; 22 items split across ≥2 targets.
-
Closest call: X184U (APOE peptides): 0412U × 0.33 won 5–2–2 over gapfill and straight 0412U.
###
Prompt.
These pages show the votes of the CMS laboratory advisory expert panel, for pricing of new lab codes for 2026. The panel met on July 23, 2025, and reviewed 92 agenda items. The items generally had 10 respones from 10 unique participants. (Occassionally one panelist abstained). Panelists could vote for "gapfill" pricing, or for crosswalk to a particular code or fractional code or multiple codes. For example, on agenda item 9, which is 0159U, transplantation medicine, we see that 1 panelist voted for gapfill, 3 for a crosswalk to 85% of 0018U, and 6 for a crosswalk to the price of 0069U. Please conduct a carefully analytical assessment of the panelist results and discuss in illuminating and insightful ways. For example, a reader might want to know how often "Gapfill" was the majority vote, or how often panelists voted for a single crosswalk as opposed to some sort of fraction or multiple. Take your time.
Items 22, 61, were numbered at one time but not shown on the final output (skips from 21 to 23, from 60 to 62).
