AI Imaging Biomarkers
for Multiple Sclerosis
Accurately determining the location and spread of MS lesions throughout the brain is critical to diagnose and monitor disease progression.
Visual detection, categorization, and segmentation of white matter lesions is extremely time-consuming and variable between human readers.
Multiple sclerosis (MS) is a neuroinflammatory disease that mainly affects the white matter in the central nervous system (brain + spinal cord). It is one of the most common causes of disability that affects more than 2.3 millions of people worldwide.
MS is characterized as a disease that causes degeneration of the myelin, a lipid-like structure that is crucial for fast and reliable communication between the brain regions.
MRI, with its high level of details, is the most common tool for diagnosis of MS. Typically, on an MRI scan, the MS presents with lesions in the white matter that are usually concentrated around the ventricles.
Automated Lesion Delineation
Demyelination, oedema and gliosis extend MRI T2 relaxation time. Hyperintensities in fluid-attenuated inversion recovery (FLAIR) images indicate MS lesion presence. Automated algorithms segment lesions reproducibly, reducing inter-rater bias. In large projects such as clinical trials, hundreds of datasets can be processed simultaneously using parallel computing, saving time and money.(1)
T1 weighted gadolinium-enhanced (Gd+ T1) lesion
To identify active MS lesions, gadolinium-based contrast agents are administered to enhance contrast on T1-weighted magnetic resonance imaging. This biomarker is mainly used in the diagnosis and monitoring of patients with MS because contrast-enhancing lesions reflect Blood-Brain Barrier (BBB) dysfunction and the inflammatory response. Gd enhancement is shown to correlate with the occurrence of clinical relapses in MS, and Gd based measures such as number or volume of Gd-enhancing lesions are used to monitor breakthrough disease and to evaluate the efficacy and effectiveness of anti-inflammatory agents in MS clinical trials and in clinical practice(6,7).
T2 weighted lesion detection
The standard sequences used to identify MS lesions in the brain and spinal cord are sensitive to T2 prolongation, leading to a hyperintense appearance. The most common such sequences used for brain MRI include heavily weighted fast spin-echo T2-weighted and FLAIR sequences. T2 hyperintense lesions form the cornerstone of diagnosis, are a standard supportive outcome measure to monitor therapeutic efficacy in clinical trials, and have modest but significant value in predicting conversion from a clinically isolated syndrome (CIS) to clinically definite MS (6,8).
Combined unique active (CUA) magnetic resonance imaging (MRI) lesions
A CUA lesion is defined as a new or persisting T1 weighted gadolinium-enhancing (Gd+ T1) lesion, or a new, enlarging or persistently enlarging lesion on T2-weighted images, or both without double counting. As such it is a relevant biomarker for disease activity of relapsing and progressive MS trials (9,10).
Source: Giorgio, Antonio et al. “Mapping the Progressive Treatment-Related Reduction of Active MRI Lesions in Multiple Sclerosis.” Frontiers in Neurology (2020).
Source: Caroline Köhler, et al. “Exploring individual multiple sclerosis lesion volume change over time: Development of an algorithm for the analyses of longitudinal quantitative MRI measures.” NeuroImage: Clinical (2019).
Changes in Lesion Count and Lesion Volume
Over time new and enlarging lesions are separately quantified. Lesion volume change, number of new lesions, number of enlarging lesions can be calculated between timepoints using an automated algorithm to assist in analysis of progression or remission.(2) Detection of Central Vein lesions
Detection of Central Vein lesions
The specificity and implementation of current MRI-based diagnostic criteria for multiple sclerosis (MS) are imperfect. Approximately 1 in 5 of individuals diagnosed with MS are eventually determined not to have the disease, with overreliance on MRI findings a major cause of MS misdiagnosis. The central vein sign (CVS), a proposed MRI biomarker for MS lesions, has been extensively studied in numerous cross sectional studies and may increase diagnostic specificity for MS. FLAIR* (combination of FLAIR and T2*) is generated for evaluating CVS (11).
Source: Sati, P. et al. “The central vein sign and its clinical evaluation for the diagnosis of multiple sclerosis: a consensus statement from the North American Imaging in Multiple Sclerosis Cooperative.” Nat Rev Neurol (2016).
Quantitative Susceptibility Mapping (QSM)
The specificity and implementation of current MRI-based diagnostic criteria for multiple sclerosis (MS) are imperfect. Approximately 1 in 5 of individuals diagnosed with MS are eventually determined not to have the disease, with overreliance on MRI findings a major cause of MS misdiagnosis. The central vein sign (CVS), a proposed MRI biomarker for MS lesions, has been extensively studied in numerous cross sectional studies and may increase diagnostic specificity for MS. FLAIR* (combination of FLAIR and T2*) is generated for evaluating CVS (11).
Changes in Brain Volume as detected by brain MRI
Patients with the main clinical phenotypes of multiple sclerosis (MS) manifest varying degrees of gray matter atrophy beyond that of normal aging(5). The QMENTA platform’s gray matter volumetry tools allow for quantification of region-specific atrophy measures including gray matter volumes and thickness quantification, for each patient timepoint using processing tools such as FreeSurfer, ANTs and SIENAX, Mediaire, among others. The platform’s data management allows longitudinal assessment of gray matter atrophy.
Source: Huang SY, at al. “Characterization of Axonal Pathology Independent of Fiber Crossings in Multiple Sclerosis Using High-Gradient Diffusion MRI.”(2016)
Diffusion Measures Around Lesions
Diffusion measurements around lesions have been linked to re-myelination and demyelination, offering a biomarker to measure the effect of new therapies on MS.(3,4)
Lesion Load Measures on Main Fiber Bundles
Tractography allows the spatial delineation of major white matter structures within the brain, combining this information with lesion segmentation from T2 weighted hyperintensities allows the derivation of structurally specific lesion load measures.
Interested in our MS biomarkers?
Look at the full dedicated tools
All in a single and easy-to-use environment
1. Schmidt, P., Gaser, C., Arsic, M., Buck, D., Förschler, A., Berthele, A., ... & Hemmer, B. (2012). An automated tool for detection of FLAIR-hyperintense white-matter lesions in multiple sclerosis. Neuroimage, 59(4), 3774-3783.
2. Egger, Christine, et al. "MRI FLAIR lesion segmentation in multiple sclerosis: Does automated segmentation hold up with manual annotation?." NeuroImage: Clinical 13 (2017): 264-270.
3. Klistorner, Alexander, et al. "Decoding diffusivity in multiple sclerosis: analysis of optic radiation lesional and non-lesional white matter." PLoS One 10.3 (2015): e0122114.
4. Klistorner, Alexander, et al. "Diffusivity in multiple sclerosis lesions: At the cutting edge?." NeuroImage: Clinical 12 (2016): 219-226.
5. Rocca, M. A. et al. (2017). Brain MRI atrophy quantification in MS. Neurology, 88(4), 403–413. https://doi.org/10.1212/WNL.0000000000003542
6. Hemond, Christopher C., and Rohit Bakshi. "Magnetic resonance imaging in multiple sclerosis." Cold Spring Harbor perspectives in medicine 8, no. 5 (2018): a028969.
7. Coronado, Ivan, Refaat E. Gabr, and Ponnada A. Narayana. "Deep learning segmentation of gadolinium-enhancing lesions in multiple sclerosis." Multiple Sclerosis Journal 27, no. 4 (2021): 519-527.
8. Kazancli, Erol, Vesna Prchkovska, Paulo Rodrigues, Pablo Villoslada, and Laura Igual. "Multiple Sclerosis Lesion Segmentation using Improved Convolutional Neural Networks." In VISIGRAPP (4: VISAPP), pp. 260-269. 2018.
9. Moccia, Marcello, Nicola de Stefano, and Frederik Barkhof. "Imaging outcome measures for progressive multiple sclerosis trials." Multiple Sclerosis Journal 23, no. 12 (2017): 1614-1626.
10. Giorgio, Antonio, Marco Battaglini, Giordano Gentile, Maria Laura Stromillo, Claudio Gasperini, Andra Visconti, Andrea Paolillo, and Nicola De Stefano. "Mapping the progressive treatment-related reduction of active MRI lesions in multiple sclerosis." Frontiers in neurology 11 (2020): 1466.
11. Ontaneda, D., P. Sati, P. Raza, M. Kilbane, E. Gombos, E. Alvarez, C. Azevedo et al. "Central vein sign: A diagnostic biomarker in multiple sclerosis (CAVS-MS) study protocol for a prospective multicenter trial." NeuroImage: Clinical 32 (2021): 102834.