Artificial Intelligence in Diagnosis of DFNA9

Sensorineural Hearing Loss Clinical Trial

— DFNA9  

Official title:

Positive Predictive Value of Machine Learning Tools (Audiogene v4.0) for Diagnosing DFNA9 in a Large Series of p.Pro51Ser Variant Carriers in COCH.

Clinical Trial Details — Status: Completed

Administrative data

• NCT number: NCT04331015
• Other study ID #: JessaH-5-ORL
• Status: Completed
• Start date: February 1, 2020
• Est. completion date: March 28, 2020

Study information

• Verified date: April 2020
• Source: Jessa Hospital
• Contact: n/a
• Is FDA regulated: No
• Study type: Observational

Clinical Trial Summary

To study the positive predictive value of Audiogene v.4.0 open source online machine learning tool in accurately predicting DFNA9 (DeaFNess autosomal dominant ninth) as top 3 gene loci in a large series of genetically confirmed c.151C>T,p.Pro51Ser (p.P51S) variant carriers in COCH (coagulation factor C Homology).

Description:

DFNA9 is an autosomal dominant hereditary adult-onset and progressive sensorineural hearing loss which is associated wit vestibular deterioration.

Today, artificial intelligence plays an increasing role in diagnosis of Mendelian hearing losses and in fitting of cochlear implants. An application of this kind is the open source program, Audiogene v4.0, which was elaborated by the Center for Bioinformatics and Computational Biology, University of Iowa City, Iowa, USA. The shape of the audiogram (audioprofile) is easily recognizable in many autosomal dominantly inherited hearing losses. Machine learning based software tools, such as Audiogene v4.0, which was originally developed for prioritizing loci for the Sanger sequencing, could help the clinicians in early diagnosis of DFNA9. This tool only need subjects' age and hearing thresholds (decibel hearing loss (dB HL)) at frequency range of 0.125 - 8 kHz (kiloHerz), left, right or binaural average in order to predict top 3 gene loci according to the data entered in the program.

Goal: to use auditory data of a large series of genetically confirmed p.P51S variant carriers causing DFNA9, which were previously collected for the genotype-phenotype correlation study which terminated recently.

All individual left and right sided hearing thresholds (ranging from 0.125 to 8kHz, with the exception of 1.5 kHz) as well as binaural averaged thresholds were run through Audiogene v4.0.

Descriptive statistics were assessed and statistical analysis was carried out to check for possible differences between age or hearing thresholds between the carrier group with accurate prediction against the carrier group with inaccurate prediction.

Recruitment information / eligibility

• Status: Completed
• Enrollment: 111
• Est. completion date: March 28, 2020
• Est. primary completion date: March 28, 2020
• Accepts healthy volunteers: No
• Gender: All
• Age group: 18 Years and older

Eligibility

Inclusion Criteria:

• at least 18 years

• genetically confirmed c.151 C>T, p.Pro51Ser variant carrier in COCH gene

• not contra-indication for audiometric testing

Exclusion Criteria:

• <18 years

• no carrier status for c.151C>T, p.Pro51Ser

• no auditory data available

Related Conditions & MeSH terms

• Deafness
• DFNA9
• Hearing Loss
• Hearing Loss, Sensorineural
• Sensorineural Hearing Loss

Intervention

Diagnostic Test:
• Pure tone audiometry
pure tone audiometry

Locations

Country	Name	City	State
Belgium	University of Antwerp	Antwerpen
Belgium	Jessa Hospital	Hasselt	Limburg

Sponsors (1)

Lead Sponsor	Collaborator
Jessa Hospital

Country where clinical trial is conducted

Belgium, 

References & Publications (1)

Taylor KR, Deluca AP, Shearer AE, Hildebrand MS, Black-Ziegelbein EA, Anand VN, Sloan CM, Eppsteiner RW, Scheetz TE, Huygen PL, Smith RJ, Braun TA, Casavant TL. AudioGene: predicting hearing loss genotypes from phenotypes to guide genetic screening. Hum M — View Citation

Outcome

Type	Measure	Description	Time frame	Safety issue
Primary	hearing threshold	audiometry (pure tone) decibel hearing level (dB HL) left, right ear , binaural average	1 hour
Primary	age	years, age at time of audiometry	1 hour
Primary	prediction gene locus	top 3 gene loci as predicted by Audiogene v4.0 machine learning tool	1 hour

External Links

•   audiogene v4.0