# Assessment of the adequacy of the Fast Healthcare Interoperability Resources (FHIR) Genomics standard for the representation of somatic testing reports

## TL;DR

This paper evaluates how well the FHIR Genomics standard can represent somatic cancer testing reports and finds it mostly adequate for current needs.

## Contribution

The paper provides a structured assessment of FHIR Genomics' ability to encode real-world somatic testing reports.

## Key findings

- 47 semantic objects were analyzed, with 87% achieving full or variable full encoding using FHIR Genomics.
- Some elements like genomic study details were only partially encoded.
- Coarsely granular negatives and other elements were not encoded at all.

## Abstract

Enabling the unambiguous communication of cancer genetic testing results and the corresponding therapeutic implications of identified genetic variants is of major clinical importance. In this report, we assess the adequacy of the Health Level Seven Fast Healthcare Interoperability Resources (FHIR) Genomics Reporting Implementation Guide, version 3 (also known as “FHIR Genomics”) for the structured communication of semantic objects typically appearing in real-world somatic testing reports.

The GenomeX project team, part of the CodeX FHIR Accelerator, performed the following assessment: (1) gather a convenience sample of public somatic reports; (2) create a representative somatic report that includes a wide range of observations taken from the sample reports; and (3) structure the representative somatic report using the FHIR Genomics standard.

Attempted encodings were categorized into 1 of the 4 buckets: (1) full encoding possible (eg, DNA variants); (2) full encoding with potential variability (eg, therapeutic implications); (3) partial encoding (eg, genomic study details); and (4) not encoded (ie, FHIR Genomics has no explicit or acceptable semantic target, such as for coarsely granular pertinent negatives). In our sample somatic report, 47 semantic objects were noted, of which 87% had either full encoding or full encoding with potential variability.

While dynamic real-world requirements, as exemplified by cancer genomics, often arise more rapidly than can be standardized, we conclude that the FHIR Genomics version 3 standard is sufficiently robust for the communication of current actionable somatic testing results.