Face2Gene's DeepGestalt changes geneticists' testing in 17 cases

AI system

System: Face2Gene (DeepGestalt facial phenotyping)
Model: Not assessed
Release: Not assessed
Application developer: FDNA
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: Geneticists uploading facial photographs and clinical features to Face2Gene during consultations. Cases where recognition was hard: atypical presentation, overlapping dysmorphology or phenotypes across ancestries, per the abstract.
Human oversight: Not assessed
Relation to AI: The authors state that DeepGestalt's suggestions changed geneticists' testing decisions in real time in all 17 cases. The clinicians, not the software, made the diagnoses. The assessment rests on the abstract alone because the full text was unreadable. One co-author is an FDNA employee. The other 15 are at clinical institutions.
AI contribution: The authors state that DeepGestalt's suggestions changed geneticists' testing decisions in real time in all 17 cases. The clinicians, not the software, made the diagnoses. The assessment rests on the abstract alone because the full text was unreadable. One co-author is an FDNA employee. The other 15 are at clinical institutions.

Studies: Not assessed
Study design: Not assessed
Measurement: Not assessed
Setting and sample: Not assessed
Measures: 17 cases involving 19 patients
Comparison: the testing strategy the medical geneticists had planned before the software's suggestion
Results: Result: changed medical geneticists' testing decisions in 17 cases involving 19 patients.
Additional value from AI: Not assessed
Limitations: Not assessed
Denominators: Not assessed
Materials for replication: Not assessed

Participants: Developer: FDNA. Operator: Clinical genetics services at the authors' institutions in the United States, Georgia, Italy and Canada (per Crossref affiliations).
Jurisdictions: Benefit: United States, Georgia, Italy and Canada (authors' institutions; case locations not readable). Developer: United States.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Conditions for replication: The true syndrome is represented in DeepGestalt's training data. Clinicians act on the suggestion by changing the test ordered and confirm by molecular testing.
Replication attempts: 1 follow-up entries. Benefit recurred: 1.
Later findings: The abstract provides no later development.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Assessed benefits: People. Organizations.
Depth: People: 2. Organizations: 1.
Reasons: People: DeepGestalt changed medical geneticists' testing decisions in 17 cases involving 19 patients, per the abstract read on Europe PMC. That is a measurable change in the test strategy itself. What diagnoses or treatment followed is not stated. The full text is inaccessible. Organizations: The authors claim “a more efficient test strategy”. No counts of tests avoided or time saved are readable in the abstract.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: The makers' claim

Kind: Not assessed
Related entries: Not assessed
Field: Healthcare / clinical genetics. Rare disease diagnosis.
Purpose: Face2Gene is built to suggest candidate syndromes from facial photographs and clinical features so that geneticists can direct testing. Changed testing decisions were that purpose.
Dates: Date basis unassessed: 2025 (precision unassessed). Updated: 21 October 2026.
Stage: Not assessed
Version: SHA-256: 1c55708b4240…

Case report Developer

Rudy NL, Gomes A, Tkemaladze T, Abdul-Rahman O, Cratsenberg D, Pascolini G, Di Zenzo G, Castiglia D, Black E, Washington C, Chad L, Curry C, Del Campo M, Fleischer N, Bird LM, Hurst ACE (2025). Artificial Intelligence Software Changes Rare Disease Testing Strategy in Real Time: An International Case Series Using Face2Gene. American Journal of Medical Genetics Part A 197(11); online 4 July 2025. Abstract read on Europe PMC (PMID 40613735); one author at FDNA Inc.
Other Independent author

Crossref metadata for doi:10.1002/ajmg.a.64171 (author affiliations, online date 4 July 2025, CC BY-NC-ND 4.0 licence); the publisher page itself returned a security block.