Phase enhanced PSIR T1 weighted imaging improves contrast resolution of the nucleus basalis of Meynert at 7 T: a preliminary study

Deep Learning Segmentation of the Nucleus Basalis of Meynert on 3T MRI

BACKGROUND AND PURPOSE

The nucleus basalis of Meynert is a key subcortical structure that is important in arousal and cognition and has been explored as a deep brain stimulation target but is difficult to study due to its small size, variability among patients, and lack of contrast on 3T MR imaging. Thus, our goal was to establish and evaluate a deep learning network for automatic, accurate, and patient-specific segmentations with 3T MR imaging.

MATERIALS AND METHODS

Patient-specific segmentations can be produced manually; however, the nucleus basalis of Meynert is difficult to accurately segment on 3T MR imaging, with 7T being preferred. Thus, paired 3T and 7T MR imaging data sets of 21 healthy subjects were obtained. A test data set of 6 subjects was completely withheld. The nucleus was expertly segmented on 7T, providing accurate labels for the paired 3T MR imaging. An external data set of 14 patients with temporal lobe epilepsy was used to test the model on brains with neurologic disorders. A 3D-Unet convolutional neural network was constructed, and a 5-fold cross-validation was performed.

RESULTS

The novel segmentation model demonstrated significantly improved Dice coefficients over the standard probabilistic atlas for both healthy subjects (mean, 0.68 [SD, 0.10] versus 0.45 [SD, 0.11], P = .002, t test) and patients (0.64 [SD, 0.10] versus 0.37 [SD, 0.22], P < .001). Additionally, the model demonstrated significantly decreased centroid distance in patients (1.18 [SD, 0.43] mm, 3.09 [SD, 2.56] mm, P = .007).

CONCLUSIONS

We developed the first model, to our knowledge, for automatic and accurate patient-specific segmentation of the nucleus basalis of Meynert. This model may enable further study into the nucleus, impacting new treatments such as deep brain stimulation.

New Developments in Cholinergic Imaging in Alzheimer and Lewy Body Disorders

PURPOSE OF REVIEW

This paper aims to review novel trends in cholinergic neuroimaging in Alzheimer and Lewy body parkinsonian disorders.

RECENT FINDINGS

The spectrum of cholinergic imaging is expanding with the availability of spatially more precise radioligands that allow assessment of previously less recognized subcortical and cortical structures with more dense cholinergic innervation. In addition, advances in MRI techniques now allow quantitative structural or functional assessment of both the cholinergic forebrain and the pedunculopontine nucleus, which may serve as non-invasive prognostic predictors. Multimodal imaging approaches, such as PET-MRI or multiligand PET offer new insights into the dynamic and interactive roles of the cholinergic system at both local and larger-scale neural network levels.

SUMMARY

Our understanding of the heterogeneous roles of the cholinergic system in age-related diseases is evolving. Multimodal imaging approaches that provide complimentary views of the cholinergic system will be necessary to shed light on the impact of cholinergic degeneration on regional and large-scale neural networks that underpin clinical symptom manifestation in neurodegeneration.

Use of Multiplied, Added, Subtracted and/or FiTted Inversion Recovery (MASTIR) pulse sequences

The group of Multiplied, Added, Subtracted and/or fiTted Inversion Recovery (MASTIR) pulse sequences in which usually two or more inversion recovery (IR) images of different types are combined is described, and uses for this type of sequence are outlined. IR sequences of different types can be multiplied, added, subtracted, and/or fitted together to produce variants of the MASTIR sequence. The sequences provide a range of options for increasing image contrast, demonstrating specific tissues and fluids of interest, and suppressing unwanted signals. A formalism using the concept of pulse sequences as tissue property filters is used to explain the signal, contrast and weighting of the pulse sequences with both univariate and multivariate filter models. Subtraction of one magnitude reconstructed IR image from another with a shorter TI can produce very high T₁ dependent positive contrast from small increases in T₁. The reverse subtracted IR sequence can provide high positive contrast enhancement with gadolinium chelates and iron deposition which decrease T₁. Additional contrast to that arising from increases in T₁ can be produced by supplementing this with contrast arising from concurrent increases in ρ_m and T₂, as well as increases or decreases in diffusion using subtraction IR with echo subtraction and/or diffusion subtraction. Phase images may show 180º differences as a result of rotating into the transverse plane both positive and negative longitudinal magnetization. Phase images with contrast arising in this way, or other ways, can be multiplied by magnitude IR images to increase the contrast of the latter. Magnetization Transfer (MT) and susceptibility can be used with IR sequences to improve contrast. Selective images of white and brown adipose tissue lipid and water components can be produced using different TIs and in and out-of-phase TEs. Selective images of ultrashort and short T₂ tissue components can be produced by nulling long T₂ tissue components with an inversion pulse and subtraction of images with longer TEs from images with ultrashort TEs. The Double Echo Sliding IR (DESIRE) sequence provides images with a wide range of TIs from which it is possible to choose values of TI to achieve particular types of tissue and/or fluid contrast (e.g., for subtraction with different TIs, as described above, and for long T₂ tissue signal nulling with UTE sequences). Unwanted tissue and fluid signals can be suppressed by addition and subtraction of phase-sensitive (ps) and magnitude reconstructed images. The sequence also offers options for synergistic use of the changes in blood and tissue ρ_m, T₁, T₂/T₂*, D* and perfusion that can be seen with fMRI of the brain. In-vivo and ex-vivo illustrative examples of normal brain, cartilage, multiple sclerosis, Alzheimer's disease, and peripheral nerve imaged with different forms of the MASTIR sequence are included.

Are Linear Measurements of the Nucleus Basalis of Meynert Suitable as a Diagnostic Biomarker in Mild Cognitive Impairment and Alzheimer Disease?

BACKGROUND AND PURPOSE

Cell loss within the nucleus basalis of Meynert is an early event in Alzheimer disease. The thickness of the nucleus basalis of Meynert (NBM) can be measured on structural MR imaging. We investigated NBM thickness in relation to cognitive state and biochemical markers.

MATERIALS AND METHODS

Mean bilateral nucleus basalis of Meynert thickness was measured on coronal T1-weighted MR imaging scans from the Alzheimer's Disease Neuroimaging Initiative dataset. Three hundred and fifteen scans (80 controls, 79 cases of early mild cognitive impairment, 77 cases of late mild cognitive impairment and 79 cases of Alzheimer disease) were assessed. Alzheimer's Disease Assessment Scale-Cognitive scores, CSF tau, and amyloid quantification were extracted. Group differences in NBM thickness, their correlates and measurement reliability were assessed.

RESULTS

Mean NBM thickness ± SD progressively declined from 2.9 ± 0.3, 2.5 ± 0.3, and 2.3 ± 0.3 to 1.8 ± 0.4 mm in healthy controls, patients with early mild cognitive impairment, late mild cognitive impairment and Alzheimer disease respectively (P < .001). NBM thickness was negatively correlated with Alzheimer's Disease Assessment Scale-Cognitive scores (r = -0.53, P < .001) and weakly positively correlated with CSF amyloid (r = 0.250, P < .001) respectively. No association with CSF tau was found. NBM thickness showed excellent diagnostic accuracy to differentiate Alzheimer disease (area under the curve, 0.986) and late mild cognitive impairment from controls (area under the curve, 0.936) with excellent sensitivity, but lower specificity 66.7%. Intra- and interrater reliability for measurements was 0.66 and 0.47 (P < .001).

CONCLUSIONS

There is progressive NBM thinning across the aging-dementia spectrum, which correlates with cognitive decline and CSF markers of amyloid-β pathology. We show high diagnostic accuracy but limited reliability, representing an area for future improvement. NBM thickness is a promising, readily available MR imaging biomarker of Alzheimer disease warranting diagnostic-accuracy testing in clinical practice.