MRI-based Synthetic CT Images of the Head and Neck

Artificial Intelligence Driven Synthetic CT to Substitute CT Scans of the Head and Neck Region

In case of surgical procedures in the head and neck region, MRI in combination with CT of the bone is often the standard modality to visualise bony landmarks for planning, navigation and risk assessment. An important downside of a CT scan is the associated radiation exposure, especially in children. An additional downside is the sedation or general anaesthesia needed for both the MRI and CT scan session in very young children. These downsides could be removed if the CT scan can be substituted by an MRI sequence that can provide the same information as CT. This project aims to determine the feasibility of recreating CT like images of the craniofacial bones from MRI images using machine learning techniques.

Study Overview

• Status: Completed
• Conditions: Sinusitis; Hearing Loss; Head and Neck Tumor; Cholesteatoma; Imaging of Bony Structures of the Head (Various Conditions)
• Intervention / Treatment: Diagnostic test: CT scan; Diagnostic test: MRI scan; Other: Synthetic CT scan

• Study Type: Interventional
• Enrollment (Actual): 80
• Phase: Not Applicable

Contacts and Locations

Study Locations

• Netherlands
  • Amsterdam, Netherlands
    • Amsterdam University Medical Center

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Patients from the outpatient ENT (Ear, Nose, Throat)-clinic.
• Aged 18 years or older.
• Referred for CT scan of the mastoid, sinonasal complex or face.

Exclusion Criteria:

• Pregnancy.
• Contra-indications for MRI or CT.
• Unwillingness to be informed about possibly clinically relevant, incidental findings from the MRI examination.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Diagnostic
• Allocation: Non-Randomized
• Interventional Model: Parallel Assignment
• Masking: None (Open Label)

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Other: Training; Data from 25-35 participants will be used to train an algorithm to generate synthetic CT images from MRI scans.	Diagnostic test: CT scan; Participants receive a CT scan of the head as part of their regular care. A larger part of the head will be scanned than for standard care.; Diagnostic test: MRI scan; Participants receive an MRI scan, specifically for the purpose of the study.
Other: Testing; Data from remaining participants will be used to test the synthetic CT algorithm, by comparing true CT scans to synthetic CT scans made from MRI.	Diagnostic test: CT scan; Participants receive a CT scan of the head as part of their regular care. A larger part of the head will be scanned than for standard care.; Diagnostic test: MRI scan; Participants receive an MRI scan, specifically for the purpose of the study.; Other: Synthetic CT scan; Synthetic CT scans will be generated from MRI scans, using the trained machine learning algorithm.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Geometrical accuracy.	Geometrical accuracy of the bone morphology by determining the mean surface distance in mm between the cortical edges on synthetic CT and on true CT.	Within one year after scans have been obtained.
Radiodensity accuracy.	Accuracy of the voxelwise radiodensity in Hounsfield Units and accuracy of the radiodensity contrast.	Within one year after scans have been obtained.
Visibility of landmarks.	Accuracy of the visibility of clinically relevant anatomical landmarks on the synthetic CT images compared to the corresponding true CT images in the adult population, rated by experienced physicians on a 4-point Likert scale (1 = not visible, 4 = very well visible).	Within one year after scans have been obtained.

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Usefulness.	Evaluation of potential usefulness of the synthetic CT images for surgical planning, surgical navigation and diagnostic purposes, as evaluated by experienced physicians and dichotomised into "useful" or "not useful".	Within one year after scans have been obtained.

Collaborators and Investigators

• Sponsor: Amsterdam UMC, location VUmc
• Collaborators: MRIguidance B.V.
• Investigators: Principal Investigator: Paul Merkus, MD PhD, Amsterdam UMC, location VUmc

Study record dates

Study Major Dates

• Study Start (Actual): September 22, 2022
• Primary Completion (Actual): September 12, 2023
• Study Completion (Actual): September 12, 2023

Study Registration Dates

• First Submitted: August 7, 2023
• First Submitted That Met QC Criteria: August 25, 2023
• First Posted (Actual): August 29, 2023

Study Record Updates

• Last Update Posted (Actual): December 19, 2023
• Last Update Submitted That Met QC Criteria: December 18, 2023
• Last Verified: December 1, 2023

More Information

Terms related to this study

• Keywords: MRI; Artificial Intelligence; Craniofacial; Synthetic CT
• Additional Relevant MeSH Terms: Nervous System Diseases; Skin Diseases; Infections; Respiratory Tract Infections; Respiratory Tract Diseases; Neoplasms; Neoplasms by Site; Neurologic Manifestations; Otorhinolaryngologic Diseases; Ear Diseases; Paranasal Sinus Diseases; Nose Diseases; Sensation Disorders; Hearing Disorders; Keratosis; Head and Neck Neoplasms; Sinusitis; Hearing Loss; Cholesteatoma
• Other Study ID Numbers: 2022.0234; NL80426.029.22 (Registry Identifier: CCMO Register)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No