Artificial Intelligence Analysis of Initial Scan Evolution of Traumatic Brain Injured Patient to Predict Neurological Outcome

Trauma, Brain Clinical Trial

— RADIOMIC-TBI  

Official title:

Artificial Intelligence Analysis of Initial Scan Evolution of Traumatic Brain Injured Patient to Predict Neurological Outcome: Pilot Translational an Exploratory Study

Clinical Trial Details — Status: Active, not recruiting

Administrative data

• NCT number: NCT04058379
• Other study ID #: 38RC19.193
• Status: Active, not recruiting
• Phase: N/A
• Start date: January 1, 2020
• Est. completion date: October 12, 2021

Study information

• Verified date: November 2020
• Source: University Hospital, Grenoble
• Contact: n/a
• Is FDA regulated: No
• Study type: Interventional

Clinical Trial Summary

We assume that an early iterative automatic CT scan analysis (D0, D1 and D3) by different AI approaches will allow an early differentiation of the tissues evolution after TBI. Our objective is to couple theses scan profiles to a neurological evolution, measured by therapeutic intensity.

Description:

Traumatic brain injury is a common and serious pathology, responsible of an important morbi-mortality. The TBI can be consider as a complex set of nosological entities of different evolution with difficult early identification whereas the main issue of this pathology depends on prevention and management of the lesions caused by the initial cerebral aggression.

Different evolutionary profiles seems to exist and sometimes coexists: edema evolution, hemorrhagic transformation and/or cerebrospinal fluid (CSF) resorption issues with hydrocephalus apparition.

Currently, there is no Imaging methods that can be used in every day clinical management that allows a visualization, quantification and prediction of these different lesional evolutions

CT scan is the reference imaging method for TBI patient monitoring. It allows a lesion description, a therapeutic adaptation and an evaluation of the prognostic.

Even if it is used as a routine examination, the analysis of cerebral scanners remains manual and a non-quantitative one, which make a little informative analysis as far as lesions evolution is concerned.

Recently it has been established the automatic MRI analysis with AI approach allows:

1. - To show aspects of images that can't be seen to the naked eye

2. - To automatically segment and quantify the different tissues (edema, hemorrhage...). First tests on this kind of analysis on CT scans shows that this technology can be transferred from MRI to CT scans and more importantly it brings out new quantitative informations on cerebral lesions evolution.

Recruitment information / eligibility

• Status: Active, not recruiting
• Enrollment: 30
• Est. completion date: October 12, 2021
• Est. primary completion date: April 12, 2021
• Accepts healthy volunteers: No
• Gender: All
• Age group: 18 Years and older

Eligibility

Inclusion Criteria:

• Age > or = 18 years old

• Closed TBI

• Primary admission in Grenoble University Hospital

• Initial CT scan with visible cerebral lesion rated at least 3 on abbreviated injury score (AIS)

• In ICU for an expected length of 48 hours

• Social security system affiliation

Exclusion Criteria:

• Life expectation <48 hours

• In ICU for more than 24h

• Transferred from another hospital

• Patients corresponding to articles L1121-5, L1121-6, L1121-7, L1121-8 (under legal protection) of French Public Health Code

• Patient in exclusion time of another study

Related Conditions & MeSH terms

• Brain Injuries, Traumatic
• Trauma, Brain

Intervention

Radiation:
• CT scan
3 ct scans : D0, D1 and D3

Locations

Country	Name	City	State
France	University Hospital Grenoble	Grenoble

Sponsors (1)

Lead Sponsor	Collaborator
University Hospital, Grenoble

Country where clinical trial is conducted

France, 

Outcome

Type	Measure	Description	Time frame	Safety issue
Primary	Clinical evolution during first 7 days in ICU with therapeutic intensity level (TILsum)	Composite criteria : Head position, depth, sort and objective of sedation, presence or absence of a CSF draining system, management of ventilation, presence or absence of a hyperosmolar therapy, management of body temperature, surgical intervention for intracranial hypertension.	7 days after TBI
Secondary	Mortality according to scan profiles		28 days after TBI and 6 month after TBI
Secondary	Morbidity (consequences of the trauma) according to scan profiles	Lenght of stay in ICU with more than 20 mmHg of intracranial pressure with no stimulation	28 days after TBI
Secondary	Morbidity (consequences of the trauma) according to scan profiles	length of stay in ICU with a therapeutic intensity level > or = 8,	28 days after TBI
Secondary	Morbidity (consequences of the trauma) according to scan profiles	Hospital stay length	28 days after TBI
Secondary	Morbidity (consequences of the trauma) according to scan profiles	ICU stay length days with mechanical ventilation	28 days after TBI
Secondary	Morbidity (consequences of the trauma) according to scan profiles	Glasgow Outcome Scale (GOSe)	6 months after TBI
Secondary	Comparison and Description of correlation between early scan profiles evolution signature by AI and to clinical evolution (with TILSum)	Analysis of main outcome (TIL sum after 7 days in ICU maximum) according to a kinetic scan evolution between D0, D1 and D3	7 days after TBI
Secondary	Neurological Pupil Index	Measure of neurological pupilla index within 1h after admission and at D1	1 day after TBI