Author Of 2 Presentations
P0007 - Detecting treatment response on T2 lesion burden in multiple sclerosis clinical trials with deep learning (ID 1171)
Abstract
Background
Deep Learning (DL) methods are promising to automate or assist the radiologist in many of the measurements currently made manually during clinical trial drug development; thereby reducing inter/intra rater variability and reading time. T2 lesion burden is a routine imaging endpoint in Multiple Sclerosis (MS) clinical trials. Ocrelizumab (OCR) is a humanized anti-CD20+ monoclonal antibody approved for the treatment of relapsing and primary progressive forms of MS. It has been shown to strongly suppress the number of new and enlarging T2 lesions (compared to control treatment) in the relapsing remitting MS (RRMS) based on manual neuro-radiologist assessments.
Objectives
Evaluate DL models to quantify T2 lesion burden and detect the therapeutic response identified by manual neuro-radiologist assessments.
Methods
We evaluated three DL models using multimodal brain MRI (T2w, FLAIR and T1w; voxel size of 1x1x3 mm3): a stack model with three consecutive slices as input; a patch model with 3D sub volumes as input; and a novel model architecture. Models were trained on Opera I (NCT01247324, n=898) datasets (baseline) and tested on Opera II (NCT01412333, n=905) datasets (baseline, 24, 48 and 96 weeks). Number of new and enlarging T2 lesions were estimated heuristically from serial lesion masks predicted by the model.
Results
Lesion size distributions were heavily skewed toward small lesions (~40% < 10 voxels, ~80% < 50 voxels with a minimum lesion size threshold of 3 voxels). All models achieved similar mean dice coefficient (0.7), mean lesion true positive rate (LTPR) of 0.87, 0.85, 0.82 and mean lesion false positive rate (LFPR) of 0.23, 0.24, 0.18. Generally, the models overestimated the number of new and enlarging T2 lesions when compared to the ground truth (GT) masks. This was driven by a high number of false positives (FP) for small lesions. All three models were able to detect a significant treatment response in favor of OCR at weeks 48 (p<0.01) and 96 (p<0.001). Only the novel model showed a significant treatment effect at week 24 (p<0.01).
Conclusions
The cross-sectional performance metrics of all 3 models were comparable to those reported in literature. All models successfully reproduced the treatment response from GT manual reads at 48 and 96 weeks, but only the novel model detected the early response at 24 weeks seen in the GT assessments. To further improve the model’s reliability, future work will be aimed to reduce the number of FPs for small size lesions.
P0553 - Bundle-specific microstructural alterations in normal appearing white matter of cognitively preserved MS patients using advanced diffusion MRI (ID 1718)
Abstract
Background
Although diffusion tensor imaging (DTI) has been widely used to investigate the microstructure of white matter (WM) in patients with multiple sclerosis (MS), it has several limitations; thus different approaches to model diffusion magnetic resonance imaging (dMRI) data are needed.
Objectives
To investigate alterations in the normal appearing WM (NAWM) of cognitively preserved relapsing remitting MS (RRMS) patients using different modelling of dMRI data.
Methods
We included 39 patients (23 females; mean age: 34±6.7 years; median EDSS: 2, range 0-4; mean disease duration: 7±4.6 years) and 39 age- and gender-matched healthy controls (HC). All subjects underwent cognitive (corsi block, n-back, PASAT) and clinical assessment (EDSS) as well as MRI using a 3 Tesla scanner (Siemens Prisma). We assessed 14 diffusion measures applying DTI, freewater volume fraction (FW) corrected tissue DTI (tDTI) and High Angular Resolution Diffusion Imaging (HARDI). Streamline-based particle filtering tractography was used to extract 33 major WM bundles with a multi-atlas/multi-parameter version of RecoBundle. For every bundle, the average of each diffusion measure was computed for the NAWM section. Groups were compared using t-tests and corrected for multiplicity (FDR, q=0.05).
Results
Patients and controls did not differ in cognitive performance. For each measure, the number of bundles showing a group difference and the effect size are reported: DTI (AD: 3 bundles, effect size r:0.3-0.47; RD: 16, r:0.29-0.44; MD: 16, r:0.29-0.48; FA: 11, r:-0.28- -0.5, MODE: 1,r:-0.29), tDTI (FW: 16, r:0.28-0.56; ADt: 2, r:0.34-0.44; RDt: 15, r:0.28-0.43; MDt: 14, r:0.29-0.41; FAt: 11, r:-0.29- -0.5) and HARDI (apparent fiber density (AFD): 9, r:-0.31- -0.41; generalized fractional anisotropy (GFA): 10, r:-0.29- -0.5; anisotropic power (AP): 10, r:-0.28- -0.51; number of fiber orientations (NuFO): 2, r:0.28-0.29). Differences were most prominent in the right arcuate fasciculus, right inferior fronto-occipital fasciculus, bilateral inferior longitudinal fasciculus, right optic radiation and right superior longitudinal fasciculus.
Conclusions
All approaches used to model dMRI data showed structural alterations of several major WM bundles associated with cognitive functions despite patients’ normal cognitive performance. Hence, diffusion MRI could be used to assess disease progression in early stages of the disease when compensatory mechanisms can still support normal cognitive performance.