GestaltMatcher facilitates rare disease matching using facial phenotype descriptors

Abstract

The majority of monogenic disorders cause craniofacial abnormalities with characteristic facial morphology. These disorders can be diagnosed more e ciently with the support of computer-aided nextgeneration phenotyping tools, such as DeepGestalt. These tools have learned to associate facial phenotypes with the underlying syndrome through training on thousands of patient photographs. However, this "supervised" approach means that diagnoses are only possible if they were part of the training set. To improve recognition of ultra-rare diseases, we created GestaltMatcher, which uses a deep convolutional neural network based on the DeepGestalt framework. We used photographs of 21,836 patients with 1,362 rare disorders to de ne a "Clinical Face Phenotype Space". Distance between cases in the phenotype space de nes syndromic similarity, allowing test patients to be matched to a molecular diagnosis even when the disorder was not included in the training set. Similarities among patients with previously unknown disease genes can also be detected. Therefore, in concert with mutation data, GestaltMatcher could accelerate the clinical diagnosis of patients with ultra-rare disorders and facial dysmorphism.

De novo variants of CSNK2B cause a new intellectual disability-craniodigital syndrome by disrupting the canonical Wnt signaling pathway

“…We applied GestaltMatcher 22 to compare facial phenotypes among individuals with variants in CSNK2B . The detailed method of GestaltMatcher is provided in the extended supplemental material and methods .…”

Case Report: The success of face analysis technology in extremely rare genetic diseases in Korea: Tatton–Brown–Rahman syndrome and Say–Barber –Biesecker–Young–Simpson variant of ohdo syndrome

“…Recent studies have demonstrated Face2Gene is as good as or superior to clinical assessments performed by trained healthcare providers in identifying genetic disorders ( Basel-Vanagaite et al, 2016 ; Gripp et al, 2016 ; Liehr et al, 2018 ; Pantel et al, 2018 ; Vorravanpreecha et al, 2018 ). In a very recent study, GestaltMatcher, an encoder based on deep convolutional neural networks, matched patients with other patients with the same molecular diagnosis, even if the disorder was not included in the training set ( Hsieh et al, 2022 ). This program could accelerate the clinical diagnosis of patients with extremely rare diseases.…”

Integrative analysis of transcriptomics in human craniofacial development reveals novel candidate disease genes

“…Machine learning technology has been able to take advantage of this observation and several tools have been developed to match facial phenotypes with known syndromes (Gurovich et al, 2019; Huang et al, 2008; Pantel et al, 2020; Taigman et al, 2014). Most recently machine vision tools have shown potential for applications in diagnosing and characterizing ultra-rare syndromes with the added benefit of being minimally invasive (van der Donk et al, 2019; Hsieh et al, 2022). However the majority of craniofacial abnormalities are much more common than these syndromes and occur in the absence of other abnormalities (Leslie and Marazita, 2013; Morris, 2016).…”