Using Artificial Intelligence to Screen for Hip Dysplasia
ConditionDevelopmental Dysplasia of Hip
Trial statusRecruiting
Trial phaseNot applicable
Trial typeInterventional
Biological sexAll
Age4-20
SponsorMurdoch Childrens Research Institute
About this trial
The goal of this clinical trial is to learn if an ultrasound scan using artificial intelligence can accurately screen for hip dysplasia. Researchers will compare the artificial intelligence ultrasound results to the standard ultrasound measures to see if the artificial intelligence ultrasound scan can accurately screen for hip dysplasia.
It will also seek to understand how parents feel about their children undergoing this scan.
Participants will:
* Have an additional ultrasound performed on their child at their scheduled outpatient's appointment for hip dysplasia * Complete a short questionnaire about the experience of having the measurement performed on their child
Eligibility criteria
This trial accepts healthy volunteersQualifiers
Enrolled in the VicHip study
Is 4-20 weeks of age at enrolment
Is attending The Royal Children's Hospital for the purpose of the potential diagnosis of DDH
Has a diagnostic (standard) hip ultrasound on the day of their out-patient appointment
Disqualifiers
None
Trial design
Design model
Single group
Treatments tested in this trial
Artificial intelligence augmented ultrasound
DeviceThe hip ultrasound is performed using a handheld device (Exo Iris) that uses a pocket-sized ultrasound probe and is run through an application on an IoS (Apple mobile) operation system. . A real-time algorithm detects and records the anatomical landmarks. When there are enough images for analysis the operator is notified that the scan is complete.
Treatment groups
240 Participants
are divided into 1 treatment groupGroup A: All active participantsOther 1 intervention
Trial outcomes
Primary outcomes
1
Artificial Intelligence augmented ultrasound screening test capability: Sensitivity
Artificial intelligence (AI) augmented ultrasound results will be compared to standard ultrasound imaging to calculate sensitivity (\[number of true positive cases detected/(number of true positive cases detected + number of false negative cases detected)\] X 100). Groups will be defined as follows: * True positive cases: Flagged for follow-up after the AI assessment of ultrasound sweeps from the portable probe and have a diagnosis of DDH from traditional ultrasound reports. * False positive cases: Flagged for follow-up after the AI assessment of the ultrasound sweeps and do not have a diagnosis of DDH from the traditional ultrasound reports * False negative cases: Return a "normal hips" assessment after the AI assessment of the ultrasound probe and have a diagnosis of DDH from traditional ultrasound reports * True negative cases: Return a "normal hips" assessment after the AI assessment of the ultrasound probe and do not have a diagnosis of DDH from traditional ultrasound reports
Time frame
1 day, both ultrasound scans will be performed on the same day2
Artificial Intelligence augmented ultrasound screening test capability: Specificity
Artificial intelligence (AI) augmented ultrasound results will be compared to standard ultrasound imaging to calculate specificity (\[number of true negative cases detected/(number of false positive cases detected + number of true negatives cases detected\] X 100). Groups will be defined as follows: * True positive cases: Flagged for follow-up after the AI assessment of ultrasound sweeps from the portable probe and have a diagnosis of DDH from traditional ultrasound reports. * False positive cases: Flagged for follow-up after the AI assessment of the ultrasound sweeps and do not have a diagnosis of DDH from the traditional ultrasound reports * False negative cases: Return a "normal hips" assessment after the AI assessment of the ultrasound probe and have a diagnosis of DDH from traditional ultrasound reports * True negative cases: Return a "normal hips" assessment after the AI assessment of the ultrasound probe and do not have a diagnosis of DDH from traditional ultrasound reports
Time frame
1 day, both ultrasound scans will be performed on the same day3
Artificial Intelligence augmented ultrasound screening test capability: Positive predictive value (PPV)
Artificial intelligence (AI) augmented ultrasound will be compared to standard ultrasound to calculate PPV (\[number of true positive cases detected/(number of true positive cases detected + number of false positive cases predicted) X 100). Groups will be defined as follows: * True positive cases: Flagged for follow-up after the AI assessment of ultrasound sweeps from the portable probe and have a diagnosis of DDH from traditional ultrasound reports. * False positive cases: Flagged for follow-up after the AI assessment of the ultrasound sweeps and do not have a diagnosis of DDH from the traditional ultrasound reports * False negative cases: Return a "normal hips" assessment after the AI assessment of the ultrasound probe and have a diagnosis of DDH from traditional ultrasound reports * True negative cases: Return a "normal hips" assessment after the AI assessment of the ultrasound probe and do not have a diagnosis of DDH from traditional ultrasound reports
Time frame
1 day, both ultrasound scans will be performed on the same day4
Artificial Intelligence augmented ultrasound screening test capability: Negative predictive value (NPV)
Artificial intelligence (AI) augmented ultrasound will be compared to standard ultrasound to calculate NPV \[number of true negative cases detected/(number of false negatives detected + number of true negative cases detected) X 100. Groups will be defined as follows: * True positive cases: Flagged for follow-up after the AI assessment of ultrasound sweeps from the portable probe and have a diagnosis of DDH from traditional ultrasound reports. * False positive cases: Flagged for follow-up after the AI assessment of the ultrasound sweeps and do not have a diagnosis of DDH from the traditional ultrasound reports * False negative cases: Return a "normal hips" assessment after the AI assessment of the ultrasound probe and have a diagnosis of DDH from traditional ultrasound reports * True negative cases: Return a "normal hips" assessment after the AI assessment of the ultrasound probe and do not have a diagnosis of DDH from traditional ultrasound reports
Time frame
1 day, both ultrasound scans will be performed on the same daySecondary outcomes
1
Factors associated with differences in device sensitivity
Sensitivity will be calculated between groups: 1. Degree of dysplasia as defined by the Graf classification 2. Sex 3. Infant age (categorised as 4-7.99 weeks, 8-11.99 weeks, 12-15.99 weeks, 16-20 weeks) 4. Number of scans performed by device operator (less than or greater than 60 scans).
Time frame
1 day, all data will be collected from day of scan2
Factors associated with differences in device specificity
Analyses will stratified to look at differences in specificity between groups: 1. Degree of dysplasia as defined by the Graf classification 2. Sex 3. Infant age (categorised as 4-7.99 weeks, 8-11.99 weeks, 12-15.99 weeks, 16-20 weeks) 4. Number of scans performed by device operator (less than or greater than 60 scans).
Time frame
1 day, all data will be collected from day of scan3
Factors associated with differences in device positive predictive value
Positive predictive value will be compared between groups: 1. Degree of dysplasia as defined by the Graf classification 2. Sex 3. Infant age (categorised as 4-7.99 weeks, 8-11.99 weeks, 12-15.99 weeks, 16-20 weeks) 4. Number of scans performed by device operator (less than or greater than 60 scans).
Time frame
1 day, all data will be collected from day of scan4
Factors associated with differences in device negative predictive value
Negative predictive value will be compared between groups: 1. Degree of dysplasia as defined by the Graf classification 2. Sex 3. Infant age (categorised as 4-7.99 weeks, 8-11.99 weeks, 12-15.99 weeks, 16-20 weeks) 4. Number of scans performed by device operator (less than or greater than 60 scans).
Time frame
1 day, all data will be collected from day of scanSponsors and contacts
Click on the lead sponsor to view all of their trials.
Source ClinicalTrials.gov