Sunday, December 18, 2022

CMS CCI Edits for 2023: Some New AMA DNA-RNA Codes "Mutually Exclusive"

Every fall, CMS releases new edits called "P2P" Procedure to Procedure Edits, which are pairs of codes are unpayable on the same claim.  (These are released as gigantic Excel spreadsheets).  And every fall CMS also releases its manual for correct coding in each area of AMA CPT.   

These documents have been released for 2023.  There are only tiny changes in the lab section of the manual, and regarding the new genomic sequencing tumor codes for RNA, CMS only quotes a phrase already in the code descriptor itself.  

Of more interest, I think they allow both a DNA code and an RNA code for solid tumors and for hematologic cancers (both, 5-50 gene range), but CMS pivots and doesn't plan to allow the 50+ tumor gene code to be used with its new RNA-only sibling codes.  

The New CMS Resources

NCCI home page here.

Find the new P2P edits in Excel here.

Find the new coding manual for 2023 here.

The AMA GSP CGP Changes

Basically, AMA has three familiar codes for cancer genomics, 81445 (solid cancer, 5-50 genes), 81450 (hematolymphoid cancer, 5-50 genes), and 81455 (any cancer, 51+ genes).  

These 3 codes are defined as representing either DNA only reports, or DNA/RNA reports when the DNA and RNA is done in an integrated procedure.   

New.  When RNA is done as a separate procedure, or later, AMA has new codes.  

These are 81449 (RNA analysis, 5-50 genes, solid cancer), 81451 (RNA analysis, 5-50 genes, hematolymphoid cancer), and 81456 (RNA analysis, 51+ genes, any cancer.

Let's list them as AMA does:

  • 81445 (5-50 genes, solid cancer, DNA or DNA/RNA analysis)
    • 81449, same but RNA only
  • 81450 (5-50 genes, hematolymphoid cancer, DNA or DNA/RNA analysis)
    • 81451, same but RNA only
  • 81455 (51+ genes, any cancer, DNA or DNA/RNA analysis)
    • 81456, same but RNA only
AMA adds some clarifying text to its discussion page for sequencing procedures.  Regarding these RNA codes, AMA writes,
  • Testing for somatic alternations in neoplasms may be reported differently  based on whether combined methods and analyses are used for both DNA and RNA analytes, or if separate methods and analyses are used for each analyte (DNA only, RNA only).  
    • For...a single combined method, use 81445, 81450, 81455.  
    • For targeted genomic sequence analysis of RNA when performed using a separate method, report 81449, 81451, 81456.  
  • For...DNA analysis AND RNA analysis performed SEPARATELY rather than via a combined method, report 81445, '50, '55 for DNA analysis AND REPORT 81449, '51, '56 for RNA analysis.

This fall, CMS priced the new RNA codes at the same price as their sibling DNA code, so if you are able to report both, the CMS payment would exactly double.

I would think that on the lab bench, RNA analysis is always at least partly separate from DNA analysis (reverse transcriptase etc), and there's no clear instruction for "partly separate" methods, so I am sure there are some dotted lines to be filled in at some point.  Alternately, the running of "partly separate methods" might lose out financially against a decision to shift lab chemistry towards "wholly separate" methods for the DNA and for the RNA.  

CMS P2P Codes and DNA/RNA Analysis

The P2P edits  are literally hundreds of thousands (100,000's) of lines long, and I'm not guaranteed to be reading them correctly.  

However, what I think I found, is that the 5-50 gene versions of RNA and DNA codes are NOT edited against each other (do NOT deny in tandem), but the 51+ versions ARE edited against each other (if you bill 81455, you can't bill any of the RNA codes).   See graphic: 

In other words, as I am reading this, 81445/81449 are "not" classed as mutually exclusive, but 81455/81456 "are" classified as mutually exclusive.  

Regarding the nuances of "separate method," CMS in its manual only quotes the same phrase the AMA uses:


An AMA workgroup is working in 2022, 2023, on further revisions to the tumor genomic sequencing codeset (e.g. here).

There are other examples of mutual exclusive procedures; for example, cytogenomics codes (81228, '229 (constitutional), 81277 (cancer), 81349 (germline by low pass NGS) appear to be generally blocked from use at the same time as tumor genomics codes (like 8145, '50, 55).   (Again, all the combinations and where they appear in the tables make it a bit confusing).  Shown below, if you bill 81228 microarray, they don't want to pay on the same day for tumor codes 81445 to 18456.

JAMA provides a brief review of the growing uses of RNASEQ in healthcare.