Clinical differentiation of parkinsonian syndromes such as the Parkinson variant of multiple system atrophy (MSA-P) and progressive supranuclear palsy (PSP) from Parkinson's disease is difficult in the early stage of the disease. In order to identify objective markers for differential diagnosis, we studied these three groups of patients with diffusion-weighted MRI (DWI). Sixteen MSA-P patients, 16 with PSP, 16 with Parkinson's disease and 15 healthy volunteers were studied. Regional apparent diffusion coefficients (rADC) were determined in different brain regions including basal ganglia, thalamus, white matter, pons and middle cerebellar peduncles (MCPs). rADC calculated in the MCP completely differentiated MSA-P patients (median: 0.93 × 10⁻³ mm²/s) from PSP patients (median: 0.82 × 10⁻³ mm²/s, P < 0.001), Parkinson's disease patients (median: 0.79 × 10⁻³ mm²/s, P < 0.001) and healthy volunteers (median: 0.81 × 10⁻³ mm²/s, P < 0.001). Other regions considered showed an overlapping among groups. DWI discriminates MSA-P from PSP and Parkinson's disease and healthy volunteers on the basis of MCP rADC values. These in vivo results confirm the pathological findings that the majority of MSA-P patients have moderate or severe degenerative changes not only in the nigrostriatal but also in the olivopontocerebellar systems. Our findings indicate that, in order to substantially contribute to the in vivo differential diagnosis of MSA-P, PSP and Parkinson's disease, rADC measurements should not be limited to the basal ganglia but should also include the MCP.

Optic neuropathy is common in mitochondrial disorders, but poorly characterized in Friedreich's ataxia (FRDA), a recessive condition caused by lack of the mitochondrial protein frataxin. We investigated 26 molecularly confirmed FRDA patients by studying both anterior and posterior sections of the visual pathway using a new, integrated approach. This included visual field testing and optical coherence tomography (OCT), pattern visual evoked potentials (P-VEPs) and diffusion-weighted imaging. The latter was used to study optic radiation by calculating water apparent diffusion coefficients (ADC). All patients suffered optic nerve involvement with their disorder. Different patterns of visual field defects were observed and a variably reduced retinal nerve fiber layer thickness was seen by OCT in all cases. P-VEPs were abnormal in approximately half of the patients. Decreased visual acuity and temporal optic disc pallor were present in advanced stages of the disease, but only five patients were symptomatic. Two of these patients suffered a sudden loss of central vision, mimicking Leber's hereditary optic neuropathy (LHON), and of the other three symptomatic patients two were noted to be compound heterozygotes. ADC values of optic radiations in patients were significantly higher than controls (P< 0.01). Retinal nerve fiber layer thickness at OCT and P-VEPs correlated with age at onset and ICARS total score. ADC values correlated with age at onset, disease duration, GAA triplet expansion size, ICARS total score and P-VEPs. Visual pathway involvement is found consistently in FRDA, being previously underestimated, and we here document that it also involves the optic radiations. Occasional LHON-like cases may occur. However, optic neuropathy in FRDA substantially differs from classic mitochondrial optic neuropathies implying a different pathophysiology of visual system degeneration in this mitochondrial disease.

Objective: Mounting evidence links neurodegenerative disorders such as Parkinson disease and Alzheimer disease with mitochondrial dysfunction, and recent emphasis has focused on mitochondrial dynamics and quality control. Mitochondrial dynamics and mtDNA maintenance is another link recently emerged, implicating mutations in the mitochondrial fusion genes OPA1 and MFN2 in the pathogenesis of multisystem syndromes characterized by neurodegeneration and accumulation of mtDNA multiple deletions in postmitotic tissues. Here, we report 2 Italian families affected by dominant chronic progressive external ophthalmoplegia (CPEO) complicated by parkinsonism and dementia.Methods: Patients were extensively studied by optical coherence tomography (OCT) to assess retinal nerve fibers, and underwent muscle and brain magnetic resonance spectroscopy (MRS), and muscle biopsy and fibroblasts were analyzed. Candidate genes were sequenced, and mtDNA was analyzed for rearrangements.Results: Affected individuals displayed a slowly progressive syndrome characterized by CPEO, mitochondrial myopathy, sensorineural deafness, peripheral neuropathy, parkinsonism, and/or cognitive impairment, in most cases without visual complains, but with subclinical loss of retinal nerve fibers at OCT. Muscle biopsies showed cytochrome c oxidase‐negative fibers and mtDNA multiple deletions, and MRS displayed defective oxidative metabolism in muscle and brain. We found 2 heterozygous OPA1 missense mutations affecting highly conserved amino acid positions (p.G488R, p.A495V) in the guanosine triphosphatase domain, each segregating with affected individuals. Fibroblast studies showed a reduced amount of OPA1 protein with normal mRNA expression, fragmented mitochondria, impaired bioenergetics, increased autophagy and mitophagy.Interpretation: The association of CPEO and parkinsonism/dementia with subclinical optic neuropathy widens the phenotypic spectrum of OPA1 mutations, highlighting the association of defective mitochondrial dynamics, mtDNA multiple deletions, and altered mitophagy with parkinsonism. Ann Neurol 2015;78:21–38

Convolutional neural networks (CNNs) constitute a widely used deep learning approach that has frequently been applied to the problem of brain tumor diagnosis. Such techniques still face some critical challenges in moving towards clinic application. The main objective of this work is to present a comprehensive review of studies using CNN architectures to classify brain tumors using MR images with the aim of identifying useful strategies for and possible impediments in the development of this technology. Relevant articles were identified using a predefined, systematic procedure. For each article, data were extracted regarding training data, target problems, the network architecture, validation methods, and the reported quantitative performance criteria. The clinical relevance of the studies was then evaluated to identify limitations by considering the merits of convolutional neural networks and the remaining challenges that need to be solved to promote the clinical application and development of CNN algorithms. Finally, possible directions for future research are discussed for researchers in the biomedical and machine learning communities. A total of 83 studies were identified and reviewed. They differed in terms of the precise classification problem targeted and the strategies used to construct and train the chosen CNN. Consequently, the reported performance varied widely, with accuracies of 91.63–100% in differentiating meningiomas, gliomas, and pituitary tumors (26 articles) and of 60.0–99.46% in distinguishing low-grade from high-grade gliomas (13 articles). The review provides a survey of the state of the art in CNN-based deep learning methods for brain tumor classification. Many networks demonstrated good performance, and it is not evident that any specific methodological choice greatly outperforms the alternatives, especially given the inconsistencies in the reporting of validation methods, performance metrics, and training data encountered. Few studies have focused on clinical usability.