Evaluation of an Automatic Segmentation Software (Pixyl.Neuro) to Track Lesions in Multiple Sclerosis Patients Via Cerebral MRI

Multiple Sclerosis Clinical Trial

— PIXCIS  

Official title:

Evaluation of an Automatic Segmentation Software (Pixyl.Neuro) to Track Lesions in Multiple Sclerosis Patients Via Cerebral MRI

Clinical Trial Details — Status: Completed

Administrative data

• NCT number: NCT03438357
• Other study ID #: IDIL/2017/ET-01
• Status: Completed
• Start date: January 2, 2019
• Est. completion date: July 6, 2022

Study information

• Verified date: February 2023
• Source: Centre Hospitalier Universitaire de Nimes
• Contact: n/a
• Is FDA regulated: No
• Study type: Observational

Clinical Trial Summary

Using an automatic software tool, Pixyl.Neuro, to conduct a retrospective analysis (detection of lesions + segmentation of images + tracking over time) of cerebral MRI images acquired during a prospective study of patients presenting with clinically isolated syndrome, and comparing the results against manual analysis.

Recruitment information / eligibility

• Status: Completed
• Enrollment: 200
• Est. completion date: July 6, 2022
• Est. primary completion date: July 6, 2022
• Accepts healthy volunteers: No
• Gender: All
• Age group: 18 Years to 56 Years

Eligibility

Identical to those of the study D-Lay-MS: NCT01817166:

Inclusion Criteria:

• The patient must have given his/her informed and signed consent

• The patient must be insured or beneficiary of a health insurance plan

• The patient is available for 24 months of follow-up

• The patient has had a classic CIS with the past 90 days

• Reference cerebro-medullary MRI scheduled within the 90 days after the beginning of symptoms

• With MRI (cerebro ± medullary) showing demyelination according to spatial spread criteria by Swanton (2006):

• At least 1 lesion in at least 2 of the 4 following territories: (1) Peri-ventricular; (2) Juxta-cortical; (3) Sub-tentorial; (4) Medullary

• No other suspected pathology

• Vitamin D level in blood less than 100 nmol / l at the pre-inclusion visit

• Women of childbearing potential must use very effective contraception for the duration of the study. A very effective contraceptive method is defined as a method resulting in a low failure rate (that is to say less than 1% per year) when used consistently and correctly, such as implants, injectables, combined oral contraceptives, IUDs, sexual abstinence, or partner with a vasectomy.

Randomisation stratification criteria:

• The patient can also also meet the temporal dissemination criteria defined according to McDonald criteria 2010 (Polman et al., 2011), because this condition is currently not sufficient for prescribing a background treatment: Simultaneous presence of at least one asymptomatic lesion taking on contrast and at least one asymptomatic lesion not taking on contrast after injection of gadolinium

Exclusion Criteria:

• The patient is participating in another study other than D-Lay-MS: NCT01817166 (this criteria does not apply to the POLAR study (RCB 2011-A01269-32); patients included in this study may simultaneously participate in the POLAR study)

• The patient is in an exclusion period determined by a previous study

• The patient is under judicial protection, under tutorship or curatorship

• The patient refuses to sign the consent

• It is impossible to correctly inform the patient

• The patient is pregnant, parturient, or breastfeeding

• Major medical or psychiatric illness that, according to the investigator, would result in the patient running an unnecessary risk or that could affect compliance with the study protocol

• Vitamin D insufficiency linked to currently active digestive or more general diseases (celiac disease, inflammatory bowel disease, intestinal bypass, short bowel syndrome, cirrhosis, nephrotic syndrome, hyperthyroidism, rickets, hypoparathyroidism, cancer, granulomatous diseases and lymphomas)

• Moderate or severe renal insufficiency (creatinine clearance less than 60 ml / min)

• Epilepsy not adequately controlled by treatment

• Any illness requiring chronic treatment with corticosteroids

• Patient with osteoporosis or history of osteopenia

• Pathology requiring calcium intakes greater than 1 gram per day

• Current or past history of hypercalcemia

• Medications that affect the metabolism of vitamin D other than corticosteroids; e.g. anticonvulsants [phenobarbital, primidone, phenytoin] rifampicin, isoniazid, ketoconazole, 5-FU and leucovorin, thiazide diuretics.

• Situations accompanied by increased vulnerability to hypercalcemia, e.g. arrhythmia or known heart disease, treatment with digitalis, and subjects with nephrolithiasis.

• Contraindications to vitamin D3 as mentioned in the documentation for UVEDOSE

• Known hypersensitivity to gadolinium and / or known inability to undergo an MRI (pacemaker, osteosynthesis material, intraocular metal splinter, etc ....).

• The subject is participating in an interventional study, or is in a period of exclusion determined by a previous study

• The subject refuses to sign the consent

• It is impossible to give the subject informed information

• The patient is under safeguard of justice or state guardianship

Related Conditions & MeSH terms

• Clinically Isolated Syndrome
• Multiple Sclerosis
• Sclerosis

Intervention

Other:
• Analysis of cerebral MRI with automatic segmentation software
Pixyl.Neuro software versus manual analysis of lesions

Locations

Country	Name	City	State
France	CHU Nimes	Nîmes

Sponsors (2)

Lead Sponsor	Collaborator
Centre Hospitalier Universitaire de Nimes	Roche Pharma AG

Country where clinical trial is conducted

France, 

Outcome

Type	Measure	Description	Time frame	Safety issue
Primary	Lesions detected by Pixyl.Neuro software that were missed by manual analysis of MRI images follow-up images compared to reference image (new T2 lesions, enlarging T2 lesions, new T1 gadolinium lesions)	% cases	baseline until follow-up (maximum 2 years)
Secondary	Inter-reader reproducibility	DICE similarity coefficient between readers	baseline until follow-up (maximum 2 years)
Secondary	Software versus manual analysis reproducibility for same reader	DICE similarity coefficient between software result and manual result	baseline until follow-up (maximum 2 years)
Secondary	Classification of discordances	Specific description of missed lesions according to technology at fault	baseline until follow-up (maximum 2 years)