Using the Progression Independent of Relapse Activity Framework to Unveil the Pathobiological Foundations of Multiple Sclerosis.

Progression independent of relapse activity (PIRA), a recent concept to formalize disability accrual in multiple sclerosis (MS) independent of relapses, has gained popularity as a potential clinical trial outcome. We discuss its shortcomings and appraise the challenges of implementing it in clinical settings, experimental trials, and research. The current definition of PIRA assumes that acute inflammation, which can manifest as a relapse, and neurodegeneration, manifesting as progressive disability accrual, can be disentangled by introducing specific time windows between the onset of relapses and the observed increase in disability. The term PIRMA (progression independent of relapse and MRI activity) was recently introduced to indicate disability accrual in the absence of both clinical relapses and new brain and spinal cord MRI lesions. Assessing PIRMA in clinical practice is highly challenging because it necessitates frequent clinical assessments and brain and spinal cord MRI scans. PIRA is commonly assessed using Expanded Disability Status Scale, a scale heavily weighted toward motor disability, whereas a more granular assessment of disability deterioration, including cognitive decline, using composite measures or other tools, such as digital tools, would possess greater utility. Similarly, using PIRA as an outcome measure in randomized clinical trials is also challenging and requires methodological considerations. The underpinning pathobiology of disability accumulation, that is not associated with relapses, may encompass chronic active lesions (slowly expanding lesions and paramagnetic rim lesions), cortical lesions, brain and spinal cord atrophy, particularly in the gray matter, diffuse and focal microglial activation, persistent leptomeningeal enhancement, and white matter tract damage. We propose to use PIRA to understand the main determinant of disability accrual in observational, cohort studies, where regular MRI scans are not included, and introduce the term of "advanced-PIRMA" to investigate the contributions to disability accrual of the abovementioned processes, using conventional and advanced imaging. This is supported by the knowledge that MRI reflects the MS pathogenic mechanisms better than purely clinical descriptors. Any residual disability accrual, which remains unexplained after considering all these mechanisms with imaging, will highlight future research priorities to help complete our understanding of MS pathogenesis.

An efficient semi-supervised quality control system trained using physics-based MRI-artefact generators and adversarial training.

Large medical imaging data sets are becoming increasingly available. A common challenge in these data sets is to ensure that each sample meets minimum quality requirements devoid of significant artefacts. Despite a wide range of existing automatic methods having been developed to identify imperfections and artefacts in medical imaging, they mostly rely on data-hungry methods. In particular, the scarcity of artefact-containing scans available for training has been a major obstacle in the development and implementation of machine learning in clinical research. To tackle this problem, we propose a novel framework having four main components: (1) a set of artefact generators inspired by magnetic resonance physics to corrupt brain MRI scans and augment a training dataset, (2) a set of abstract and engineered features to represent images compactly, (3) a feature selection process that depends on the class of artefact to improve classification performance, and (4) a set of Support Vector Machine (SVM) classifiers trained to identify artefacts. Our novel contributions are threefold: first, we use the novel physics-based artefact generators to generate synthetic brain MRI scans with controlled artefacts as a data augmentation technique. This will avoid the labour-intensive collection and labelling process of scans with rare artefacts. Second, we propose a large pool of abstract and engineered image features developed to identify 9 different artefacts for structural MRI. Finally, we use an artefact-based feature selection block that, for each class of artefacts, finds the set of features that provide the best classification performance. We performed validation experiments on a large data set of scans with artificially-generated artefacts, and in a multiple sclerosis clinical trial where real artefacts were identified by experts, showing that the proposed pipeline outperforms traditional methods. In particular, our data augmentation increases performance by up to 12.5 percentage points on the accuracy, F1, F2, precision and recall. At the same time, the computation cost of our pipeline remains low - less than a second to process a single scan - with the potential for real-time deployment. Our artefact simulators obtained using adversarial learning enable the training of a quality control system for brain MRI that otherwise would have required a much larger number of scans in both supervised and unsupervised settings. We believe that systems for quality control will enable a wide range of high-throughput clinical applications based on the use of automatic image-processing pipelines.

The temporal event-based model: Learning event timelines in progressive diseases.

Timelines of events, such as symptom appearance or a change in biomarker value, provide powerful signatures that characterise progressive diseases. Understanding and predicting the timing of events is important for clinical trials targeting individuals early in the disease course when putative treatments are likely to have the strongest effect. However, previous models of disease progression cannot estimate the time between events and provide only an ordering in which they change. Here, we introduce the temporal event-based model (TEBM), a new probabilistic model for inferring timelines of biomarker events from sparse and irregularly sampled datasets. We demonstrate the power of the TEBM in two neurodegenerative conditions: Alzheimer's disease (AD) and Huntington's disease (HD). In both diseases, the TEBM not only recapitulates current understanding of event orderings but also provides unique new ranges of timescales between consecutive events. We reproduce and validate these findings using external datasets in both diseases. We also demonstrate that the TEBM improves over current models; provides unique stratification capabilities; and enriches simulated clinical trials to achieve a power of 80% with less than half the cohort size compared with random selection. The application of the TEBM naturally extends to a wide range of progressive conditions.

Networks of microstructural damage predict disability in multiple sclerosis.

BACKGROUND: Network-based measures are emerging MRI markers in multiple sclerosis (MS). We aimed to identify networks of white (WM) and grey matter (GM) damage that predict disability progression and cognitive worsening using data-driven methods. METHODS: We analysed data from 1836 participants with different MS phenotypes (843 in a discovery cohort and 842 in a replication cohort). We calculated standardised T1-weighted/T2-weighted (sT1w/T2w) ratio maps in brain GM and WM, and applied spatial independent component analysis to identify networks of covarying microstructural damage. Clinical outcomes were Expanded Disability Status Scale worsening confirmed at 24 weeks (24-week confirmed disability progression (CDP)) and time to cognitive worsening assessed by the Symbol Digit Modalities Test (SDMT). We used Cox proportional hazard models to calculate predictive value of network measures. RESULTS: We identified 8 WM and 7 GM sT1w/T2w networks (of regional covariation in sT1w/T2w measures) in both cohorts. Network loading represents the degree of covariation in regional T1/T2 ratio within a given network. The loading factor in the anterior corona radiata and temporo-parieto-frontal components were associated with higher risks of developing CDP both in the discovery (HR=0.85, p<0.05 and HR=0.83, p<0.05, respectively) and replication cohorts (HR=0.84, p<0.05 and HR=0.80, p<0.005, respectively). The decreasing or increasing loading factor in the arcuate fasciculus, corpus callosum, deep GM, cortico-cerebellar patterns and lesion load were associated with a higher risk of developing SDMT worsening both in the discovery (HR=0.82, p<0.01; HR=0.87, p<0.05; HR=0.75, p<0.001; HR=0.86, p<0.05 and HR=1.27, p<0.0001) and replication cohorts (HR=0.82, p<0.005; HR=0.73, p<0.0001; HR=0.80, p<0.005; HR=0.85, p<0.01 and HR=1.26, p<0.0001). CONCLUSIONS: GM and WM networks of microstructural changes predict disability and cognitive worsening in MS. Our approach may be used to identify patients at greater risk of disability worsening and stratify cohorts in treatment trials.

Emerging Perspectives on MRI Application in Multiple Sclerosis: Moving from Pathophysiology to Clinical Practice.

MRI plays a central role in the diagnosis of multiple sclerosis (MS) and in the monitoring of disease course and treatment response. Advanced MRI techniques have shed light on MS biology and facilitated the search for neuroimaging markers that may be applicable in clinical practice. MRI has led to improvements in the accuracy of MS diagnosis and a deeper understanding of disease progression. This has also resulted in a plethora of potential MRI markers, the importance and validity of which remain to be proven. Here, five recent emerging perspectives arising from the use of MRI in MS, from pathophysiology to clinical application, will be discussed. These are the feasibility of noninvasive MRI-based approaches to measure glymphatic function and its impairment; T1-weighted to T2-weighted intensity ratio to quantify myelin content; classification of MS phenotypes based on their MRI features rather than on their clinical features; clinical relevance of gray matter atrophy versus white matter atrophy; and time-varying versus static resting-state functional connectivity in evaluating brain functional organization. These topics are critically discussed, which may guide future applications in the field.

HLA-DRB1*1501 influences long-term disability progression and tissue damage on MRI in relapse-onset multiple sclerosis.

BACKGROUND: Whether genetic factors influence the long-term course of multiple sclerosis (MS) is unresolved. OBJECTIVE: To determine the influence of HLA-DRB1*1501 on long-term disease course in a homogeneous cohort of clinically isolated syndrome (CIS) patients. METHODS: One hundred seven patients underwent clinical and MRI assessment at the time of CIS and after 1, 3, 5 and 15 years. HLA-DRB1*1501 status was determined using Sanger sequencing and tagging of the rs3135388 polymorphism. Linear/Poisson mixed-effects models were used to investigate rates of change in EDSS and MRI measures based on HLA-DRB1*1501 status. RESULTS: HLA-DRB1*1501 -positive (n = 52) patients showed a faster rate of disability worsening compared with the HLA-DRB1*1501 -negative (n = 55) patients (annualised change in EDSS 0.14/year vs. 0.08/year, p < 0.025), and a greater annualised change in T2 lesion volume (adjusted difference 0.45 mL/year, p < 0.025), a higher number of gadolinium-enhancing lesions, and a faster rate of brain (adjusted difference -0.12%/year, p < 0.05) and spinal cord atrophy (adjusted difference -0.22 mm2/year, p < 0.05). INTERPRETATION: These findings provide evidence that the HLA-DRB1*1501 allele plays a role in MS severity, as measured by long-term disability worsening and a greater extent of inflammatory disease activity and tissue loss. HLA-DRB1*1501 may provide useful information when considering prognosis and treatment decisions in early relapse-onset MS.

Serum neurofilament light and MRI predictors of cognitive decline in patients with secondary progressive multiple sclerosis: Analysis from the MS-STAT randomised controlled trial.

BACKGROUND: Cognitive impairment affects 50%-75% of people with secondary progressive multiple sclerosis (PwSPMS). Improving our ability to predict cognitive decline may facilitate earlier intervention. OBJECTIVE: The main aim of this study was to assess the relationship between longitudinal changes in cognition and baseline serum neurofilament light chain (sNfL) in PwSPMS. In a multi-modal analysis, MRI variables were additionally included to determine if sNfL has predictive utility beyond that already established through MRI. METHODS: Participants from the MS-STAT trial underwent a detailed neuropsychological test battery at baseline, 12 and 24 months. Linear mixed models were used to assess the relationships between cognition, sNfL, T2 lesion volume (T2LV) and normalised regional brain volumes. RESULTS: Median age and Expanded Disability Status Score (EDSS) were 51 and 6.0. Each doubling of baseline sNfL was associated with a 0.010 [0.003-0.017] point per month faster decline in WASI Full Scale IQ Z-score (p = 0.008), independent of T2LV and normalised regional volumes. In contrast, lower baseline volume of the transverse temporal gyrus was associated with poorer current cognitive performance (0.362 [0.026-0.698] point reduction per mL, p = 0.035), but not change in cognition. The results were supported by secondary analyses on individual cognitive components. CONCLUSION: Elevated sNfL is associated with faster cognitive decline, independent of T2LV and regional normalised volumes.

Designing Multi-arm Multistage Adaptive Trials for Neuroprotection in Progressive Multiple Sclerosis.

There are few treatments shown to slow disability progression in progressive multiple sclerosis (PMS). One challenge has been efficiently testing the pipeline of candidate therapies from preclinical studies in clinical trials. Multi-arm multistage (MAMS) platform trials may accelerate evaluation of new therapies compared to traditional sequential clinical trials. We describe a MAMS design in PMS focusing on selection of interim and final outcome measures, sample size, and statistical considerations. The UK MS Society Expert Consortium for Progression in MS Clinical Trials reviewed recent phase II and III PMS trials to inform interim and final outcome selection and design measures. Simulations were performed to evaluate trial operating characteristics under different treatment effect, recruitment rate, and sample size assumptions. People with MS formed a patient and public involvement group and contributed to the trial design, ensuring it would meet the needs of the MS community. The proposed design evaluates 3 experimental arms compared to a common standard of care arm in 2 stages. Stage 1 (interim) outcome will be whole brain atrophy on MRI at 18 months, assessed for 123 participants per arm. Treatments with sufficient evidence for slowing brain atrophy will continue to the second stage. The stage 2 (final) outcome will be time to 6-month confirmed disability progression, based on a composite clinical score comprising the Expanded Disability Status Scale, Timed 25-Foot Walk test, and 9-Hole Peg Test. To detect a hazard ratio of 0.75 for this primary final outcome with 90% power, 600 participants per arm are required. Assuming one treatment progresses to stage 2, the trial will recruit ≈1,900 participants and last ≈6 years. This is approximately two-thirds the size and half the time of separate 2-arm phase II and III trials. The proposed MAMS trial design will substantially reduce duration and sample size compared to traditional clinical trials, accelerating discovery of effective treatments for PMS. The design was well-received by people with multiple sclerosis. The practical and statistical principles of MAMS trial design may be applicable to other neurodegenerative conditions to facilitate efficient testing of new therapies.

Assessing Neurofilaments as Biomarkers of Neuroprotection in Progressive Multiple Sclerosis: From the MS-STAT Randomized Controlled Trial.

BACKGROUND AND OBJECTIVES: Improved biomarkers of neuroprotective treatment are needed in progressive multiple sclerosis (PMS) to facilitate more efficient phase 2 trial design. The MS-STAT randomized controlled trial supported the neuroprotective potential of high-dose simvastatin in secondary progressive MS (SPMS). Here, we analyze serum from the MS-STAT trial to assess the extent to which neurofilament light (NfL) and neurofilament heavy (NfH), both promising biomarkers of neuroaxonal injury, may act as biomarkers of simvastatin treatment in SPMS. METHODS: The MS-STAT trial randomized patients to 80 mg simvastatin or placebo. Serum was analyzed for NfL and NfH using Simoa technology. We used linear mixed models to investigate the treatment effects of simvastatin compared with placebo on NfL and NfH. Additional models examined the relationships between neurofilaments and MRI and clinical measures of disease severity. RESULTS: A total of 140 patients with SPMS were included. There was no evidence for a simvastatin treatment effect on NfL or NfH: compared with placebo, NfL was 1.2% lower (95% CI 10.6% lower to 9.2% higher; p = 0.820) and NfH was 0.4% lower (95% CI 18.4% lower to 21.6% higher; p = 0.969) in the simvastatin treatment group. Secondary analyses suggested that higher NfL was associated with greater subsequent whole brain atrophy, higher T2 lesion volume, and more new/enlarging T2 lesions in the previous 12 months, as well as greater physical disability. There were no significant associations between NfH and MRI or clinical variables. DISCUSSION: We found no evidence of a simvastatin treatment effect on serum neurofilaments. While confirmation of the neuroprotective benefits of simvastatin is awaited from the ongoing phase 3 study (NCT03387670), our results suggest that treatments capable of slowing the rate of whole brain atrophy in SPMS, such as simvastatin, may act via mechanisms largely independent of neuroaxonal injury, as quantified by NfL. This has important implications for the design of future phase 2 clinical trials in PMS. TRIAL REGISTRATION INFORMATION: MS-STAT: NCT00647348. CLASSIFICATION OF EVIDENCE: This study provides class I evidence that simvastatin treatment does not have a large impact on either serum NfL or NfH, as quantified in this study, in SPMS.

Staging and stratifying cognitive dysfunction in multiple sclerosis.

BACKGROUND: The sequence in which cognitive domains become impaired in multiple sclerosis (MS) is yet to be formally demonstrated. It is unclear whether processing speed dysfunction temporally precedes other cognitive impairments, such as memory and executive function. OBJECTIVE: Determine the order in which different cognitive domains become impaired in MS and validate these findings using clinical and vocational outcomes. METHODS: In a longitudinal sample of 1073 MS patients and 306 healthy controls, we measured performance on multiple, consensus-standard, neurocognitive tests. We used an event-based staging approach to model the sequence in which cognitive domains become impaired. Linear and logistic mixed-effects models were used to explore associations between stages of impairment, neurological disability, and employment status. RESULTS: Our model suggested that the order of impairments was as follows: processing speed, visual learning, verbal learning, working memory/attention, and executive function. Stage of cognitive impairment predicted greater neurological disability, ß = 0.16, SE = 0.02, p < 0.001, and probability of unemployment, ß = 1.14, SE = 0.001, p < 0.001. CONCLUSION: This is the first study to introduce a cognitive staging and stratification system for MS. Findings underscore the importance of using the Symbol Digit Modalities Test in routine screening for cognitive impairment and memory testing to assess patients later in disease evolution.

B Cells in the CNS at Postmortem Are Associated With Worse Outcome and Cell Types in Multiple Sclerosis.

BACKGROUND AND OBJECTIVES: To define the clinical and pathologic correlations of compartmentalized perivascular B cells in postmortem progressive multiple sclerosis (MS) brains. METHODS: Brain slices were acquired from 11 people with secondary progressive (SP) MS, 5 people with primary progressive (PP) MS, and 4 controls. Brain slices were immunostained for B lymphocytes (CD20), T lymphocytes (CD3), cytotoxic T lymphocytes (CD8), neuronal neurofilaments (NF200), myelin (SMI94), macrophages/microglia (CD68 and IBA1), astrocytes (glial fibrillary acidic protein [GFAP]), and mitochondria (voltage-dependent anion channel and cytochrome c oxidase subunit 4). Differences in CD20 immunostaining intensity between disease groups and associations between CD20 immunostaining intensity and both clinical variables and other immunostaining intensities were explored with linear mixed regression models and Cox regression models, as appropriate. RESULTS: CD20 immunostaining intensity was higher in PPMS (Coeff = 0.410; 95% confidence interval [CI] = 0.046, 0.774; p = 0.027) and SPMS (Coeff = 0.302; 95% CI = 0.020, 0.585; p = 0.036) compared with controls. CD20 immunostaining intensity was higher in cerebellar, spinal cord, and pyramidal onset (Coeff = 0.274; 95% CI = 0.039, 0.510; p = 0.022) compared with optic neuritis and sensory onset. Higher CD20 immunostaining intensity was associated with younger age at onset (hazard ratio [HR] = 1.033; 95% CI = 1.013, 1.053; p = 0.001), SP conversion (HR = 1.056; 95% CI = 1.022, 1.091; p = 0.001), wheelchair dependence (HR = 1.472; 95% CI = 1.108, 1.954; p = 0.008), and death (HR = 1.684; 95% CI = 1.238, 2.291; p = 0.001). Higher immunostaining intensity for CD20 was associated with higher immunostaining intensity for CD3 (Coeff = 0.114; 95% CI = 0.005, 0.224; p = 0.040), CD8 (Coeff = 0.275; 95% CI = 0.200, 0.350; p < 0.001), CD68 (Coeff = 0.084; 95% CI = 0.023, 0.144; p = 0.006), GFAP (Coeff = 0.002; 95% CI = 0.001, 0.004; p = 0.030), and damaged mitochondria (Coeff = 3.902; 95% CI = 0.891, 6.914; p = 0.011). DISCUSSION: Perivascular B cells were associated with worse clinical outcomes and CNS-compartmentalized inflammation. Our findings further support the concept of targeting compartmentalized B-cell inflammation in progressive MS.

Spatial patterns of brain lesions assessed through covariance estimations of lesional voxels in multiple Sclerosis: The SPACE-MS technique.

Predicting disability in progressive multiple sclerosis (MS) is extremely challenging. Although there is some evidence that the spatial distribution of white matter (WM) lesions may play a role in disability accumulation, the lack of well-established quantitative metrics that characterise these aspects of MS pathology makes it difficult to assess their relevance for clinical progression. This study introduces a novel approach, called SPACE-MS, to quantitatively characterise spatial distributional features of brain MS lesions, so that these can be assessed as predictors of disability accumulation. In SPACE-MS, the covariance matrix of the spatial positions of each patient's lesional voxels is computed and its eigenvalues extracted. These are combined to derive rotationally-invariant metrics known to be common and robust descriptors of ellipsoid shape such as anisotropy, planarity and sphericity. Additionally, SPACE-MS metrics include a neuraxis caudality index, which we defined for the whole-brain lesion mask as well as for the most caudal brain lesion. These indicate how distant from the supplementary motor cortex (along the neuraxis) the whole-brain mask or the most caudal brain lesions are. We applied SPACE-MS to data from 515 patients involved in three studies: the MS-SMART (NCT01910259) and MS-STAT1 (NCT00647348) secondary progressive MS trials, and an observational study of primary and secondary progressive MS. Patients were assessed on motor and cognitive disability scales and underwent structural brain MRI (1.5/3.0 T), at baseline and after 2 years. The MRI protocol included 3DT1-weighted (1x1x1mm3) and 2DT2-weighted (1x1x3mm3) anatomical imaging. WM lesions were semiautomatically segmented on the T2-weighted scans, deriving whole-brain lesion masks. After co-registering the masks to the T1 images, SPACE-MS metrics were calculated and analysed through a series of multiple linear regression models, which were built to assess the ability of spatial distributional metrics to explain concurrent and future disability after adjusting for confounders. Patients whose WM lesions laid more caudally along the neuraxis or were more isotropically distributed in the brain (i.e. with whole-brain lesion masks displaying a high sphericity index) at baseline had greater motor and/or cognitive disability at baseline and over time, independently of brain lesion load and atrophy measures. In conclusion, here we introduced the SPACE-MS approach, which we showed is able to capture clinically relevant spatial distributional features of MS lesions independently of the sheer amount of lesions and brain tissue loss. Location of lesions in lower parts of the brain, where neurite density is particularly high, such as in the cerebellum and brainstem, and greater spatial spreading of lesions (i.e. more isotropic whole-brain lesion masks), possibly reflecting a higher number of WM tracts involved, are associated with clinical deterioration in progressive MS. The usefulness of the SPACE-MS approach, here demonstrated in MS, may be explored in other conditions also characterised by the presence of brain WM lesions.

pySuStaIn: a Python implementation of the Subtype and Stage Inference algorithm.

Progressive disorders are highly heterogeneous. Symptom-based clinical classification of these disorders may not reflect the underlying pathobiology. Data-driven subtyping and staging of patients has the potential to disentangle the complex spatiotemporal patterns of disease progression. Tools that enable this are in high demand from clinical and treatment-development communities. Here we describe the pySuStaIn software package, a Python-based implementation of the Subtype and Stage Inference (SuStaIn) algorithm. SuStaIn unravels the complexity of heterogeneous diseases by inferring multiple disease progression patterns (subtypes) and individual severity (stages) from cross-sectional data. The primary aims of pySuStaIn are to enable widespread application and translation of SuStaIn via an accessible Python package that supports simple extension and generalization to novel modelling situations within a single, consistent architecture.

Linking immune-mediated damage to neurodegeneration in multiple sclerosis: could network-based MRI help?

Inflammatory demyelination characterizes the initial stages of multiple sclerosis, while progressive axonal and neuronal loss are coexisting and significantly contribute to the long-term physical and cognitive impairment. There is an unmet need for a conceptual shift from a dualistic view of multiple sclerosis pathology, involving either inflammatory demyelination or neurodegeneration, to integrative dynamic models of brain reorganization, where, glia-neuron interactions, synaptic alterations and grey matter pathology are longitudinally envisaged at the whole-brain level. Functional and structural MRI can delineate network hallmarks for relapses, remissions or disease progression, which can be linked to the pathophysiology behind inflammatory attacks, repair and neurodegeneration. Here, we aim to unify recent findings of grey matter circuits dynamics in multiple sclerosis within the framework of molecular and pathophysiological hallmarks combined with disease-related network reorganization, while highlighting advances from animal models (in vivo and ex vivo) and human clinical data (imaging and histological). We propose that MRI-based brain networks characterization is essential for better delineating ongoing pathology and elaboration of particular mechanisms that may serve for accurate modelling and prediction of disease courses throughout disease stages.

Ordinal SuStaIn: Subtype and Stage Inference for Clinical Scores, Visual Ratings, and Other Ordinal Data.

Subtype and Stage Inference (SuStaIn) is an unsupervised learning algorithm that uniquely enables the identification of subgroups of individuals with distinct pseudo-temporal disease progression patterns from cross-sectional datasets. SuStaIn has been used to identify data-driven subgroups and perform patient stratification in neurodegenerative diseases and in lung diseases from continuous biomarker measurements predominantly obtained from imaging. However, the SuStaIn algorithm is not currently applicable to discrete ordinal data, such as visual ratings of images, neuropathological ratings, and clinical and neuropsychological test scores, restricting the applicability of SuStaIn to a narrower range of settings. Here we propose 'Ordinal SuStaIn', an ordinal version of the SuStaIn algorithm that uses a scored events model of disease progression to enable the application of SuStaIn to ordinal data. We demonstrate the validity of Ordinal SuStaIn by benchmarking the performance of the algorithm on simulated data. We further demonstrate that Ordinal SuStaIn out-performs the existing continuous version of SuStaIn (Z-score SuStaIn) on discrete scored data, providing much more accurate subtype progression patterns, better subtyping and staging of individuals, and accurate uncertainty estimates. We then apply Ordinal SuStaIn to six different sub-scales of the Clinical Dementia Rating scale (CDR) using data from the Alzheimer's disease Neuroimaging Initiative (ADNI) study to identify individuals with distinct patterns of functional decline. Using data from 819 ADNI1 participants we identified three distinct CDR subtype progression patterns, which were independently verified using data from 790 ADNI2 participants. Our results provide insight into patterns of decline in daily activities in Alzheimer's disease and a mechanism for stratifying individuals into groups with difficulties in different domains. Ordinal SuStaIn is broadly applicable across different types of ratings data, including visual ratings from imaging, neuropathological ratings and clinical or behavioural ratings data.