Author Of 2 Presentations
P0008 - Divergent patterns of ventral attention network centrality relate to cognitive conversion in MS (ID 473)
Cognitive impairment (CI) is common in multiple sclerosis (MS), but due to a lack of longitudinal data it remains unclear which mechanisms relate to conversion to mild or even severe CI. Previous cross-sectional work has suggested the importance of cognition-related resting-state networks, such as the default-mode and attention networks.
To characterize the functional network changes related to conversion to CI in a large sample of MS patients over a period of 5 years.
A total of 233 MS patients and 59 healthy controls (HC), all part of the Amsterdam MS cohort, underwent extensive neuropsychological testing and resting-state fMRI at baseline and follow-up (mean time-interval 4.9±0.9 years). At baseline, MS patients were categorized as being cognitively impaired (scoring ≤-2 SD on ≥2 domains, N=74), mildly impaired (MCI, being impaired on 1 domain or scoring between -1.5 and -2SD on ≥2 domains, N=33) or preserved (CP, not fulfilling the CI or MCI criteria, N=126). In addition, these groups were categorized according to the group to which they converted at follow-up (e.g. CP to CI). Network function was quantified using eigenvector centrality, a measure of network importance, which was averaged over established resting-state networks at both time-points. Correlations with brain volumes were calculated.
Over time, 26.2% of CP patients deteriorated and developed MCI (66.7%) or CI (33.3%) and 73.8% remained CP. 23.5% of MCI patients, progressed to CI. Centrality analysis showed that patients who were CI at baseline demonstrated a higher cross-sectional DMN centrality compared to controls (P=.05). Longitudinally, patients who remained CP and CP-to-MCI converters showed increasing ventral attention network (VAN) centrality over time time (P=.017 and .008, respectively), , whereas in the MCI and CI converter groups this increase was absent. Patients with less severe deep gray matter atrophy at baseline showed stronger increases in VAN centrality over time.
We showed that conversion from intact cognition to impairment in MS is related to an increase in centrality of the VAN, which is absent when overt impairment has manifested, then shifting towards DMN dysfunction. As the ventral attention network is known to normally relay information to the DMN, our results suggest that developing cognitive impairment is related to a progressive loss of control over the DMN by means of VAN dysfunction.
P0539 - Artificial double inversion recovery images for cortical lesion visualization in multiple sclerosis (ID 817)
Cortical lesions (CLs) in multiple sclerosis (MS) are clinically important, but highly inconspicuous on conventional clinical MRI. Double inversion recovery (DIR) is sensitive to CL detection, but difficult to implement in clinical practice and research settings, as it is difficult to set up and proper acquisition may take significant time due to the required inversion times (i.e.,~8 to 10 minutes). This work examines whether artificial intelligence can mitigate this dilemma through generation of artificial DIR images from –readily available– conventional clinical MR sequences.
To determine whether artificially generated DIR (aDIR) images can be used for CL detection in MS and assess how this compares to conventionally acquired DIR (cDIR) images.
In this retrospective study, aDIR images were generated from conventional 1.5 Tesla 3D-T1 and 2D-proton density/T2 images in 73 patients with MS (49 RRMS, 20 SPMS, 4 PPMS) and 42 controls. A fully convolutional 3D conditional adversarial network following an adapted U-Net design with skip-connections was trained, using images of 58 patients and 34 controls. The remaining subjects were assigned to the test set for which artificial 3D-DIR images were generated. To determine detection reliability, precision and recall, the aDIR and cDIR images of subjects in the test set were blindly scored for CLs.
A total of 626 CLs were detected on 15 aDIR images versus 696 on cDIR images (ICC=0.92, 95% confidence interval 0.68-0.98 (F(32.755)). Compared to cDIR images, CLs were detected on aDIR images with an average precision and recall of 0.84±0.06 and 0.76±0.09, respectively. The largest difference in CL discernibility was observed in frontal and temporal regions.
Artificially generated DIR images showed excellent reliability, precision and recall in detected cortical lesions when compared to conventionally acquired DIR images. The technique has the potential to broaden DIR availability and to enable retrospective implementation of cortical lesion detection with DIR. Histopathological and multi-center validation are necessary to formally compare sensitivity and specificity and cross-scanner robustness.