A Standardized MR Imaging Protocol for Parkinsonism

Despite well-established Parkinson's disease (PD) diagnostic criteria,1 diagnosis can still be challenging, with a high rate of misdiagnosis in atypical cases, especially in early disease stages.2, 3 The majority of misdiagnoses are related to the differential diagnosis of PD with atypical parkinsonian disorders (APD), such as multiple system atrophy–parkinsonian type (MSA-P), progressive supranuclear palsy (PSP), or corticobasal degeneration, as well as essential tremor (ET) and other forms of parkinsonism (drug-induced, vascular, functional). Early diagnosis is of crucial importance for both individual patients in clinical practice and for research purposes, as it enables not only the implementation of adequate therapeutic plans and a definition of prognosis, but also a more appropriate selection of patients for clinical trials, particularly with the increase of disease-modifying trials for PD, MSA, and PSP.4, 5 The extensive development of neuroimaging in recent years has profoundly changed the study of PD. In the past, magnetic resonance imaging (MRI) was almost exclusively confined to the detection of secondary causes of parkinsonism.4 However, high-field MR magnets, more sophisticated head coils, and improved sequences have revolutionized the in vivo study of the brain parenchyma, and MRI studies are becoming increasingly relevant in the evaluation of PD. Furthermore, being a widely available, noninvasive, safe, and cost-effective method, repeated MRI scans can be performed and contribute to the follow-up of different pathological changes that occur throughout the course of the disease, from preclinical to late stages, supporting and guiding possible interventions. Imaging now plays a pivotal role in aiding diagnosis in PD early or premanifest stages and in enabling a differential diagnosis with APDs.6, 7 Adopting a standardized imaging protocol for parkinsonism can be very relevant for patient evaluation, with both clinical and research implications. In clinical settings, we wish to provide the highest amount of information to support clinical decisions and increase diagnostic accuracy as well as reproducibility, which is also very important for patient follow-up. In clinical research, accuracy and reproducibility of imaging results are key aspects that can only be attained with dedicated and specific protocols. Standardization also facilitates efficient data collection. In imaging acquisition, the heterogeneity of MR scanners and parameters (eg, field strength, gradient system, manufacturer, sequences) may lead to different results. Minor differences in hardware or sequence timing may result in significant changes in image information. The magnetic field strength of the scanner may also influence image interpretation, with higher field strength allowing higher signal-to-noise ratio. Also, for image interpretation, it is important to have a systematic approach and to evaluate specific features. Without a dedicated protocol, information loss and inability to value pathological findings may increase the number of false-negatives because of missing or disregarded data. MRI can now be of supportive value for a specific possible or probable diagnosis, with particularly high diagnostic value in patients whose clinical diagnosis uncertainty is greater. We suggest that brain MRI be performed in all patients presenting with parkinsonism when there is clinical doubt about the diagnosis, preferably in a 3.0T scanner when available, with a protocol designed to provide the highest amount of information possible in a routine clinical setting. The selection of sequences should focus on: wide availability, scientifically robust evidence support, contribution to diagnostic accuracy and follow-up, and reproducibility. The postprocessing methods must not be too time-consuming or need expensive/sophisticated tools. MRI is a very cost-effective means to increase diagnostic accuracy in parkinsonism. Compared with nuclear medicine techniques, MRI is less costly, widely available, and allows differentiation of atypical parkinsonian syndromes. Although a devoted MR protocol as proposed can increase examination time, the value of the additional information obtained can save money and time of health care systems and patients because with only one examination, sufficient information is gathered that avoids additional examinations. Our recommended MRI acquisition protocol comprises: On simple visual inspection, conventional T1 and T2 sequences' main role is the exclusion of secondary causes of parkinsonism.6 Several imaging signs and distinctive atrophy patterns were described in different APDs7-10 (Figs. 1 and 2). However, these are usually late-stage findings, and signal changes must be interpreted cautiously, as they can be influenced by the magnetic field strength6 (eg, the MSA-P hyperintense putaminal rim sign). T2-weighted imaging is also relevant for the evaluation of perivascular spaces, with described cases of parkinsonism associated with enlarged perivascular spaces.11, 12 The development of 3-D T1-weighted volumetric sequences and of several measurement tools has contributed to more objectivity in establishing imaging differential diagnosis of PD with APDs. One of the methods available is the MR parkinsonism index (MRPI), the product of the ratio pons area (pa)/midbrain area (ma) in the midsagittal plane multiplied by the ratio of the width of the middle cerebellar peduncles (MCPs) and superior cerebellar peduncles, SCP ([pa/ma] × [MCP/SCP]), which if above 13.55 strongly suggests PSP13 (Fig. 3). More recently, to improve the accuracy of the MRPI in differentiating PSP-P in the early stage from PD, a new index termed MRPI 2.0 was introduced, which is calculated by multiplying the MRPI by the third ventricle width/frontal horns width ratio.14 Other measurements include the midbrain-to-pons area (ma/pa) ratio, useful in confirming a small midbrain in the setting of PSP.15 Interestingly, new approaches with high diagnostic accuracy in discriminating PSP have been introduced, assessing midbrain diameter (md) and midbrain-to-pons diameter (md/pd).16, 17 In particular, the md seems to be the most sensitive and specific measurement.5 On the other hand, measurements of MCP width have been shown to be significantly smaller in patients with MSA.18 We propose applying visual inspection as well as measurement tools to increase sensitivity and specificity. We suggest evaluating the midbrain-to-pons area (ma/pa) ratio, the MCPd, and the SCPd and calculating the MRPI 2.0. Also, midsagittal md and pd should be assessed. T2-weighted images should be assessed for characteristic APD findings and for evaluation of perivascular spaces. Focal lesions, vascular changes ("vascular Parkinsonism") or normal pressure hydrocephalus can be better characterized with fluid-attenuated inversion recovery (FLAIR) imaging. Compared with 2-dimensional FLAIR, 3-D FLAIR imaging has higher sensitivity for detecting lesions19 and allows both visual inspection and lesion quantification. Depigmentation of the substantia nigra (SN) and the locus coeruleus (LC) related to the loss of neuromelanin (NM) is a long-known pathological hallmark of PD,20, 21 but only in recent years has in vivo visualization of NM become possible, with the development of a specific T1-weighted fast spin-echo sequence22, as well as several post-processing methods to analyze the obtained images23-27 (Fig. 4). NM measurements have reportedly high sensitivity and specificity for differentiating PD patients from healthy controls, with a high diagnostic accuracy even in early disease stages,24, 25, 27 which is only slightly lower than that obtained in [123]FP-CIT-single-photon emission computed tomography studies.28 Also, NM-MRI is able to differentiate between tremor-dominant PD and ET.29 NM-MRI visual inspection can be used to detect SN changes in early-stage PD,25 indicating its usefulness in clinical practice. We propose a simple visual inspection of the NM in the SN and LC looking at area and signal intensity performed by an experienced radiologist. Also, semiquantitative measurements of SN area and evaluation of SN width and signal intensity could be implemented. Recently, a variety of MR methods have been developed to improve sensitivity to brain iron deposits, based on relaxation rates or magnetic susceptibility techniques. Using susceptibility-weighted imaging (SWI), a dorsolateral nigral hyperintensity (DNH) area within the otherwise hypointense SN pars compacta (SNpc) has been identified in healthy individuals (swallowtail sign), corresponding to nigrosome-1, a small cluster of dopaminergic cells within the SNpc.30 This DNH is lost in PD31 (Fig. 5). In a recent meta-analysis,8 the authors concluded that visual assessment of DNH could provide excellent diagnostic accuracy for PD patients versus healthy controls. Also, it may become a potential biomarker for premotor stages of PD.32 In addition, studies using quantitative methods to investigate iron content in PD and APDs have revealed increased iron content in the SN in all groups with degenerative parkinsonism.33 Overall, patients with PSP and MSA-P seem to have significantly higher levels of iron deposition than control and PD groups.34 In general, pathological iron accumulation seems to be more prevalent and severe in PSP compared with MSA-P, except in the putamen.33, 34 We propose using SWI for visual inspection for the presence/absence of the swallowtail sign and for inspection of increased iron deposition in the basal ganglia. If centers have expertise with quantitative iron-sensitive sequences, a gradient-echo sequence or quantitative susceptibility mapping could be added. A recent meta-analysis concluded that not only are fractional anisotropy (FA) and mean diffusivity (MD) of the SN good indicators for identifying PD patients, but also for assessing PD progression.35 However, FA has limitations because atrophy-based partial volume with free water can bias the diffusion index,36 and measuring free water in the brain is a potential new biomarker of disease progression.36, 37 In PD differential diagnosis, diffusion MRI has been shown to be useful in discriminating ET,38 as well as in discriminating MSA-P (even in early disease stages) from PD and healthy controls based on higher putaminal MD values18, 39-41 and abnormal diffusion metrics in the MCP.18, 42 Abnormal diffusion metrics were also described in the SCP for PSP.4, 18 However, further studies are needed to ascertain diagnostic accuracy of MRI diffusivity measures in APDs. We propose simple visual inspection of FA color maps to detect abnormalities in cases of APDs. The main imaging aspects to evaluate on each MR sequence are summarized in Supplemental Table 1. In Supplemental Table 2 we summarize the key MRI findings in PD and APD for each sequence. Recent research on PD and APD MRI markers has had prolific results. However, the discovery of a single perfect marker is unlikely. Therefore, a multimodal approach to image analysis emerges as the answer to overcoming single parameter limitations. For instance, combined evaluation of iron- and NM-sensitive MRIs may detect changes in the SN with higher sensitivity.44 Simultaneous measurement of volumes, T2* relaxation rates, and diffusion in nigrostriatal structures also seem better for separating PD, PSP, and MSA.45-47 On the other hand, the level of experience of the physician reading the study can greatly impact examination's sensitivity and specificity. As with other pathologies (namely, epilepsy), an effort for subspecialization in movement disorder imaging can greatly improve accuracy. Another progress in PD imaging analysis to minimize evaluator bias and limitations is the use of automatic methods, namely, machine learning–derived classification algorithms for quantitative MRI analysis, including volumetric data sets,48-50 neuromelanin imaging,26 and diffusion imaging.51 If the results are confirmed by further large-scale studies, automated image analysis may open up another window into detecting objectively degenerative parkinsonian disorders in an operator-independent way.52 However, in current practice, a widely accessible imaging protocol should have good accuracy without being too time-consuming or having the need for expensive and technically challenging computer software, unavailable in many health care facilities. Although research and technological developments continue to improve imaging evaluation of the brain, a simple MR protocol based on widely available and easy-to-use standardized sequences can greatly improve the care of patients with parkinsonism. An expert consensus on parkinsonian MRI for clinical application, with well-defined imaging criteria and standardized data processing, can change diagnostic accuracy and therapeutic evaluation, and support clinical trials. Ultimately, we hope that standardization will benefit the individual patient, which after all is the goal of medical imaging. This protocol proposal is provided with the understanding that it will likely require modification as instrument capabilities change, new sequences are developed, and more quantitative methodologies become validated and feasible in clinical practice. (1) Research Project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript Preparation: A. Writing of the First Draft, B. Review and Critique. S.R.: 1A, 1B, 3B. C.G.: 1C, 3A. K.S.: 3B. J.J.F.: 1A, 3B. W.P.: 3B. Ethical Compliance Statement: The authors confirm that the approval of an institutional review board and patient consent were not required for this work. We confirm that we have read the journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines. Additional Supporting Information may be found in the online version of this article at the publisher's web-site. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.