CATMoS: Collaborative Acute Toxicity Modeling Suite.

BACKGROUND: Humans are exposed to tens of thousands of chemical substances that need to be assessed for their potential toxicity. Acute systemic toxicity testing serves as the basis for regulatory hazard classification, labeling, and risk management. However, it is cost- and time-prohibitive to evaluate all new and existing chemicals using traditional rodent acute toxicity tests. In silico models built using existing data facilitate rapid acute toxicity predictions without using animals. OBJECTIVES: The U.S. Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM) Acute Toxicity Workgroup organized an international collaboration to develop in silico models for predicting acute oral toxicity based on five different end points: Lethal Dose 50 (LD50 value, U.S. Environmental Protection Agency hazard (four) categories, Globally Harmonized System for Classification and Labeling hazard (five) categories, very toxic chemicals [LD50 (LD50≤50mg/kg)], and nontoxic chemicals (LD50>2,000mg/kg). METHODS: An acute oral toxicity data inventory for 11,992 chemicals was compiled, split into training and evaluation sets, and made available to 35 participating international research groups that submitted a total of 139 predictive models. Predictions that fell within the applicability domains of the submitted models were evaluated using external validation sets. These were then combined into consensus models to leverage strengths of individual approaches. RESULTS: The resulting consensus predictions, which leverage the collective strengths of each individual model, form the Collaborative Acute Toxicity Modeling Suite (CATMoS). CATMoS demonstrated high performance in terms of accuracy and robustness when compared with in vivo results. DISCUSSION: CATMoS is being evaluated by regulatory agencies for its utility and applicability as a potential replacement for in vivo rat acute oral toxicity studies. CATMoS predictions for more than 800,000 chemicals have been made available via the National Toxicology Program's Integrated Chemical Environment tools and data sets (ice.ntp.niehs.nih.gov). The models are also implemented in a free, standalone, open-source tool, OPERA, which allows predictions of new and untested chemicals to be made. https://doi.org/10.1289/EHP8495.

Building a Quantitative Structure-Property Relationship (QSPR) Model.

Knowing the physicochemical and general biochemical properties of a compound is critical to understanding how it behaves in different biological environments and to anticipating what is likely to happen in situations where that behavior cannot be measured directly. Quantitative structure-property relationship (QSPR) models provide a way to predict those properties even before a compound has been synthesized simply by knowing what its structure would be. This chapter describes a general workflow for compiling the data upon which a useful QSPR model is built, curating it, evaluating that model's performance, and then analyzing the predictive errors with an eye toward identifying systematic errors in the input data. The focus here is on models for the absorption, distribution, metabolism, and excretion (ADME) properties of drugs and toxins, but the considerations explored are general and applicable to any QSPR.

Improvement of quantitative structure-activity relationship (QSAR) tools for predicting Ames mutagenicity: outcomes of the Ames/QSAR International Challenge Project.

The International Conference on Harmonization (ICH) M7 guideline allows the use of in silico approaches for predicting Ames mutagenicity for the initial assessment of impurities in pharmaceuticals. This is the first international guideline that addresses the use of quantitative structure-activity relationship (QSAR) models in lieu of actual toxicological studies for human health assessment. Therefore, QSAR models for Ames mutagenicity now require higher predictive power for identifying mutagenic chemicals. To increase the predictive power of QSAR models, larger experimental datasets from reliable sources are required. The Division of Genetics and Mutagenesis, National Institute of Health Sciences (DGM/NIHS) of Japan recently established a unique proprietary Ames mutagenicity database containing 12140 new chemicals that have not been previously used for developing QSAR models. The DGM/NIHS provided this Ames database to QSAR vendors to validate and improve their QSAR tools. The Ames/QSAR International Challenge Project was initiated in 2014 with 12 QSAR vendors testing 17 QSAR tools against these compounds in three phases. We now present the final results. All tools were considerably improved by participation in this project. Most tools achieved >50% sensitivity (positive prediction among all Ames positives) and predictive power (accuracy) was as high as 80%, almost equivalent to the inter-laboratory reproducibility of Ames tests. To further increase the predictive power of QSAR tools, accumulation of additional Ames test data is required as well as re-evaluation of some previous Ames test results. Indeed, some Ames-positive or Ames-negative chemicals may have previously been incorrectly classified because of methodological weakness, resulting in false-positive or false-negative predictions by QSAR tools. These incorrect data hamper prediction and are a source of noise in the development of QSAR models. It is thus essential to establish a large benchmark database consisting only of well-validated Ames test results to build more accurate QSAR models.

A bifunctional nociceptin and mu opioid receptor agonist is analgesic without opioid side effects in nonhuman primates.

Misuse of prescription opioids, opioid addiction, and overdose underscore the urgent need for developing addiction-free effective medications for treating severe pain. Mu opioid peptide (MOP) receptor agonists provide very effective pain relief. However, severe side effects limit their use in the clinical setting. Agonists of the nociceptin/orphanin FQ peptide (NOP) receptor have been shown to modulate the antinociceptive and reinforcing effects of MOP agonists. We report the discovery and development of a bifunctional NOP/MOP receptor agonist, AT-121, which has partial agonist activity at both NOP and MOP receptors. AT-121 suppressed oxycodone's reinforcing effects and exerted morphine-like analgesic effects in nonhuman primates. AT-121 treatment did not induce side effects commonly associated with opioids, such as respiratory depression, abuse potential, opioid-induced hyperalgesia, and physical dependence. Our results in nonhuman primates suggest that bifunctional NOP/MOP agonists with the appropriate balance of NOP and MOP agonist activity may provide a dual therapeutic action for safe and effective pain relief and treating prescription opioid abuse.

Physiologically Based Pharmacokinetic Modeling in Lead Optimization. 2. Rational Bioavailability Design by Global Sensitivity Analysis To Identify Properties Affecting Bioavailability.

When medicinal chemists need to improve oral bioavailability (%F) during lead optimization, they systematically modify compound properties mainly based on their own experience and general rules of thumb. However, at least a dozen properties can influence %F, and the difficulty of multiparameter optimization for such complex nonlinear processes grows combinatorially with the number of variables. Furthermore, strategies can be in conflict. For example, adding a polar or charged group will generally increase solubility but decrease permeability. Identifying the 2 or 3 properties that most influence %F for a given compound series would make %F optimization much more efficient. We previously reported an adaptation of physiologically based pharmacokinetic (PBPK) simulations to predict %F for lead series from purely computational inputs within a 2-fold average error. Here, we run thousands of such simulations to generate a comprehensive "bioavailability landscape" for each series. A key innovation was recognition that the large and variable number of p Ka's in drug molecules could be replaced by just the two straddling the isoelectric point. Another was use of the ZINC database to cull out chemically inaccessible regions of property space. A quadratic partial least squares regression (PLS) accurately fits a continuous surface to these thousands of bioavailability predictions. The PLS coefficients indicate the globally sensitive compound properties. The PLS surface also displays the %F landscape in these sensitive properties locally around compounds of particular interest. Finally, being quick to calculate, the PLS equation can be combined with models for activity and other properties for multiobjective lead optimization.

Physiologically Based Pharmacokinetic Modeling in Lead Optimization. 1. Evaluation and Adaptation of GastroPlus To Predict Bioavailability of Medchem Series.

When medicinal chemists need to improve bioavailability (%F) within a chemical series during lead optimization, they synthesize new series members with systematically modified properties mainly by following experience and general rules of thumb. More quantitative models that predict %F of proposed compounds from chemical structure alone have proven elusive. Global empirical %F quantitative structure-property (QSPR) models perform poorly, and projects have too little data to train local %F QSPR models. Mechanistic oral absorption and physiologically based pharmacokinetic (PBPK) models simulate the dissolution, absorption, systemic distribution, and clearance of a drug in preclinical species and humans. Attempts to build global PBPK models based purely on calculated inputs have not achieved the <2-fold average error needed to guide lead optimization. In this work, local GastroPlus PBPK models are instead customized for individual medchem series. The key innovation was building a local QSPR for a numerically fitted effective intrinsic clearance (CLloc). All inputs are subsequently computed from structure alone, so the models can be applied in advance of synthesis. Training CLloc on the first 15-18 rat %F measurements gave adequate predictions, with clear improvements up to about 30 measurements, and incremental improvements beyond that.

Clinical Bioavailability of the Novel BACE1 Inhibitor Lanabecestat (AZD3293): Assessment of Tablet Formulations Versus an Oral Solution and the Impact of Gastric pH on Pharmacokinetics.

The relative bioavailability of lanabecestat administered as 2 tablet formulations versus an oral solution was investigated. This phase 1 single-center, open-label, randomized, 3-period crossover study involved healthy male and nonfertile female subjects aged 18-55 years (NCT02039180). Subjects received a single 50-mg lanabecestat dose as solution, tablet A, or tablet B on day 1 of each crossover period; 14 of 16 subjects completed the study. Relative bioavailability based on plasma lanabecestat AUC0-∞ (area under the plasma drug concentration-time curve from zero to infinity) geometric mean ratio versus oral solution (primary variable) was: tablet A, 1.052 (90% confidence interval [CI], 1.001-1.106); tablet B, 1.040 (0.989-1.093). These 90%CIs for geometric mean ratios are within accepted standard bioequivalence boundaries for all other pharmacokinetic (PK) parameters for both lanabecestat and metabolite (AZ13569724). All 3 formulations had similar plasma lanabecestat concentration-time profiles. Six adverse events were reported by 6 subjects (37.5%, all mild). GastroPlus modeling predicted a negligible impact of gastric pH changes on systemic PK (up to pH 7.4). Both tablet formulations fall within standard accepted bioequivalence criteria versus the oral solution. A single 50-mg lanabecestat dose was well tolerated as a solution or tablet formulation in this population.

Probing ligand recognition of the opioid pan antagonist AT-076 at nociceptin, kappa, mu, and delta opioid receptors through structure-activity relationships.

Few opioid ligands binding to the three classic opioid receptor subtypes, mu, kappa and delta, have high affinity at the fourth opioid receptor, the nociceptin/orphanin FQ receptor (NOP). We recently reported the discovery of AT-076 (1), (R)-7-hydroxy-N-((S)-1-(4-(3-hydroxyphenyl)piperidin-1-yl)-3-methylbutan-2-yl)-1,2,3,4-tetrahydroisoquinoline-3-carboxamide, a pan antagonist with nanomolar affinity for all four subtypes. Since AT-076 binds with high affinity at all four subtypes, we conducted a structure-activity relationship (SAR) study to probe ligand recognition features important for pan opioid receptor activity, using chemical modifications of key pharmacophoric groups. SAR analysis of the resulting analogs suggests that for the NOP receptor, the entire AT-076 scaffold is crucial for high binding affinity, but the binding mode is likely different from that of NOP antagonists C-24 and SB-612111 bound in the NOP crystal structure. On the other hand, modifications of the 3-hydroxyphenyl pharmacophore, but not the 7-hydroxy Tic pharmacophore, are better tolerated at kappa and mu receptors and yield very high affinity multifunctional (e.g. 12) or highly selective (e.g. 16) kappa ligands. With the availability of the opioid receptor crystal structures, our SAR analysis of the common chemotype of AT-076 suggests rational approaches to modulate binding selectivity, enabling the design of multifunctional or selective opioid ligands from such scaffolds.

IMI - oral biopharmaceutics tools project - evaluation of bottom-up PBPK prediction success part 1: Characterisation of the OrBiTo database of compounds.

Predicting oral bioavailability (Foral) is of importance for estimating systemic exposure of orally administered drugs. Physiologically-based pharmacokinetic (PBPK) modelling and simulation have been applied extensively in biopharmaceutics recently. The Oral Biopharmaceutical Tools (OrBiTo) project (Innovative Medicines Initiative) aims to develop and improve upon biopharmaceutical tools, including PBPK absorption models. A large-scale evaluation of PBPK models may be considered the first step. Here we characterise the OrBiTo active pharmaceutical ingredient (API) database for use in a large-scale simulation study. The OrBiTo database comprised 83 APIs and 1475 study arms. The database displayed a median logP of 3.60 (2.40-4.58), human blood-to-plasma ratio of 0.62 (0.57-0.71), and fraction unbound in plasma of 0.05 (0.01-0.17). The database mainly consisted of basic compounds (48.19%) and Biopharmaceutics Classification System class II compounds (55.81%). Median human intravenous clearance was 16.9L/h (interquartile range: 11.6-43.6L/h; n=23), volume of distribution was 80.8L (54.5-239L; n=23). The majority of oral formulations were immediate release (IR: 87.6%). Human Foral displayed a median of 0.415 (0.203-0.724; n=22) for IR formulations. The OrBiTo database was found to be largely representative of previously published datasets. 43 of the APIs were found to satisfy the minimum inclusion criteria for the simulation exercise, and many of these have significant gaps of other key parameters, which could potentially impact the interpretability of the simulation outcome. However, the OrBiTo simulation exercise represents a unique opportunity to perform a large-scale evaluation of the PBPK approach to predicting oral biopharmaceutics.

Multiple Orderings of Events in Disease Progression.

The event-based model constructs a discrete picture of disease progression from cross-sectional data sets, with each event corresponding to a new biomarker becoming abnormal. However, it relies on the assumption that all subjects follow a single event sequence. This is a major simplification for sporadic disease data sets, which are highly heterogeneous, include distinct subgroups, and contain significant proportions of outliers. In this work we relax this assumption by considering two extensions to the event-based model: a generalised Mallows model, which allows subjects to deviate from the main event sequence, and a Dirichlet process mixture of generalised Mallows models, which models clusters of subjects that follow different event sequences, each of which has a corresponding variance. We develop a Gibbs sampling technique to infer the parameters of the two models from multi-modal biomarker data sets. We apply our technique to data from the Alzheimer's Disease Neuroimaging Initiative to determine the sequence in which brain regions become abnormal in sporadic Alzheimer's disease, as well as the heterogeneity of that sequence in the cohort. We find that the generalised Mallows model estimates a larger variation in the event sequence across subjects than the original event-based model. Fitting a Dirichlet process model detects three subgroups of the population with different event sequences. The Gibbs sampler additionally provides an estimate of the uncertainty in each of the model parameters, for example an individual's latent disease stage and cluster assignment. The distributions and mixtures of sequences that this new family of models introduces offer better characterisation of disease progression of heterogeneous populations, new insight into disease mechanisms, and have the potential for enhanced disease stratification and differential diagnosis.

Simulated field maps for susceptibility artefact correction in interventional MRI.

PURPOSE: Intraoperative MRI (iMRI) is a powerful modality for acquiring images of the brain to facilitate precise image-guided neurosurgery. Diffusion-weighted MRI (DW-MRI) provides critical information about location, orientation and structure of nerve fibre tracts, but suffers from the "susceptibility artefact" stemming from magnetic field perturbations due to the step change in magnetic susceptibility at air-tissue boundaries in the head. An existing approach to correcting the artefact is to acquire a field map by means of an additional MRI scan. However, to recover true field maps from the acquired field maps near air-tissue boundaries is challenging, and acquired field maps are unavailable in historical MRI data sets. This paper reports a detailed account of our method to simulate field maps from structural MRI scans that was first presented at IPCAI 2014 and provides a thorough experimental and analysis section to quantitatively validate our technique. METHODS: We perform automatic air-tissue segmentation of intraoperative MRI scans, feed the segmentation into a field map simulation step and apply the acquired and the simulated field maps to correct DW-MRI data sets. RESULTS: We report results for 12 patient data sets acquired during anterior temporal lobe resection surgery for the surgical management of focal epilepsy. We find a close agreement between acquired and simulated field maps and observe a statistically significant reduction in the susceptibility artefact in DW-MRI data sets corrected using simulated field maps in the vicinity of the resection. The artefact reduction obtained using acquired field maps remains better than that using the simulated field maps in all evaluated regions of the brain. CONCLUSIONS: The proposed simulated field maps facilitate susceptibility artefact reduction near the resection. Accurate air-tissue segmentation is key to achieving accurate simulation. The proposed simulation approach is adaptable to different iMRI and neurosurgical applications.

Objective Bayesian fMRI analysis-a pilot study in different clinical environments.

Functional MRI (fMRI) used for neurosurgical planning delineates functionally eloquent brain areas by time-series analysis of task-induced BOLD signal changes. Commonly used frequentist statistics protect against false positive results based on a p-value threshold. In surgical planning, false negative results are equally if not more harmful, potentially masking true brain activity leading to erroneous resection of eloquent regions. Bayesian statistics provides an alternative framework, categorizing areas as activated, deactivated, non-activated or with low statistical confidence. This approach has not yet found wide clinical application partly due to the lack of a method to objectively define an effect size threshold. We implemented a Bayesian analysis framework for neurosurgical planning fMRI. It entails an automated effect-size threshold selection method for posterior probability maps accounting for inter-individual BOLD response differences, which was calibrated based on the frequentist results maps thresholded by two clinical experts. We compared Bayesian and frequentist analysis of passive-motor fMRI data from 10 healthy volunteers measured on a pre-operative 3T and an intra-operative 1.5T MRI scanner. As a clinical case study, we tested passive motor task activation in a brain tumor patient at 3T under clinical conditions. With our novel effect size threshold method, the Bayesian analysis revealed regions of all four categories in the 3T data. Activated region foci and extent were consistent with the frequentist analysis results. In the lower signal-to-noise ratio 1.5T intra-operative scanner data, Bayesian analysis provided improved brain-activation detection sensitivity compared with the frequentist analysis, albeit the spatial extents of the activations were smaller than at 3T. Bayesian analysis of fMRI data using operator-independent effect size threshold selection may improve the sensitivity and certainty of information available to guide neurosurgery.

Multi-channel registration of fractional anisotropy and T1-weighted images in the presence of atrophy: application to multiple sclerosis.

Co-registration of structural T1-weighted (T1w) scans and diffusion tensor imaging (DTI)-derived fractional anisotropy (FA) maps to a common space is of particular interest in neuroimaging, as T1w scans can be used for brain segmentation while DTI can provide microstructural tissue information. While the effect of lesions on registration has been tackled and solutions are available, the issue of atrophy is still open to discussion. Multi-channel (MC) registration algorithms have the advantage of maintaining anatomical correspondence between different contrast images after registration to any target space. In this work, we test the performance of an MC registration approach applied to T1w and FA data using simulated brain atrophy images. Experimental results are compared with a standard single-channel registration approach. Multi-channel registration of fractional anisotropy and T1-weighted images in the presence of atrophy: application to multiple sclerosis Both qualitative and quantitative evaluations are presented, showing that the MC approach provides better alignment with the target while maintaining better T1w and FA co-alignment.

Consensus between pipelines in structural brain networks.

Structural brain networks may be reconstructed from diffusion MRI tractography data and have great potential to further our understanding of the topological organisation of brain structure in health and disease. Network reconstruction is complex and involves a series of processesing methods including anatomical parcellation, registration, fiber orientation estimation and whole-brain fiber tractography. Methodological choices at each stage can affect the anatomical accuracy and graph theoretical properties of the reconstructed networks, meaning applying different combinations in a network reconstruction pipeline may produce substantially different networks. Furthermore, the choice of which connections are considered important is unclear. In this study, we assessed the similarity between structural networks obtained using two independent state-of-the-art reconstruction pipelines. We aimed to quantify network similarity and identify the core connections emerging most robustly in both pipelines. Similarity of network connections was compared between pipelines employing different atlases by merging parcels to a common and equivalent node scale. We found a high agreement between the networks across a range of fiber density thresholds. In addition, we identified a robust core of highly connected regions coinciding with a peak in similarity across network density thresholds, and replicated these results with atlases at different node scales. The binary network properties of these core connections were similar between pipelines but showed some differences in atlases across node scales. This study demonstrates the utility of applying multiple structural network reconstrution pipelines to diffusion data in order to identify the most important connections for further study.

Structure-based virtual screening of the nociceptin receptor: hybrid docking and shape-based approaches for improved hit identification.

The antagonist-bound crystal structure of the nociceptin receptor (NOP), from the opioid receptor family, was recently reported along with those of the other opioid receptors bound to opioid antagonists. We recently reported the first homology model of the 'active-state' of the NOP receptor, which when docked with 'agonist' ligands showed differences in the TM helices and residues, consistent with GPCR activation after agonist binding. In this study, we explored the use of the active-state NOP homology model for structure-based virtual screening to discover NOP ligands containing new chemical scaffolds. Several NOP agonist and antagonist ligands previously reported are based on a common piperidine scaffold. Given the structure-activity relationships for known NOP ligands, we developed a hybrid method that combines a structure-based and ligand-based approach, utilizing the active-state NOP receptor as well as the pharmacophoric features of known NOP ligands, to identify novel NOP binding scaffolds by virtual screening. Multiple conformations of the NOP active site including the flexible second extracellular loop (EL2) loop were generated by simulated annealing and ranked using enrichment factor (EF) analysis and a ligand-decoy dataset containing known NOP agonist ligands. The enrichment factors were further improved by combining shape-based screening of this ligand-decoy dataset and calculation of consensus scores. This combined structure-based and ligand-based EF analysis yielded higher enrichment factors than the individual methods, suggesting the effectiveness of the hybrid approach. Virtual screening of the CNS Permeable subset of the ZINC database was carried out using the above-mentioned hybrid approach in a tiered fashion utilizing a ligand pharmacophore-based filtering step, followed by structure-based virtual screening using the refined NOP active-state models from the enrichment analysis. Determination of the NOP receptor binding affinity of a selected set of top-scoring hits resulted in identification of several compounds with measurable binding affinity at the NOP receptor, one of which had a new chemotype for NOP receptor binding. The hybrid ligand-based and structure-based methodology demonstrates an effective approach for virtual screening that leverages existing SAR and receptor structure information for identifying novel hits for NOP receptor binding. The refined active-state NOP homology models obtained from the enrichment studies can be further used for structure-based optimization of these new chemotypes to obtain potent and selective NOP receptor ligands for therapeutic development.

Susceptibility artefact correction using dynamic graph cuts: application to neurosurgery.

Echo Planar Imaging (EPI) is routinely used in diffusion and functional MR imaging due to its rapid acquisition time. However, the long readout period makes it prone to susceptibility artefacts which results in geometric and intensity distortions of the acquired image. The use of these distorted images for neuronavigation hampers the effectiveness of image-guided surgery systems as critical white matter tracts and functionally eloquent brain areas cannot be accurately localised. In this paper, we present a novel method for correction of distortions arising from susceptibility artefacts in EPI images. The proposed method combines fieldmap and image registration based correction techniques in a unified framework. A phase unwrapping algorithm is presented that can efficiently compute the B0 magnetic field inhomogeneity map as well as the uncertainty associated with the estimated solution through the use of dynamic graph cuts. This information is fed to a subsequent image registration step to further refine the results in areas with high uncertainty. This work has been integrated into the surgical workflow at the National Hospital for Neurology and Neurosurgery and its effectiveness in correcting for geometric distortions due to susceptibility artefacts is demonstrated on EPI images acquired with an interventional MRI scanner during neurosurgery.

Preventing visual field deficits from neurosurgery.

OBJECTIVE: We assessed whether display of optic radiation tractography during anterior temporal lobe resection (ATLR) for refractory temporal lobe epilepsy (TLE) can reduce the severity of postoperative visual field deficits (VFD) and increase the proportion of patients who can drive and whether correction for brain shift using intraoperative MRI (iMRI) is beneficial. METHODS: A cohort of 21 patients underwent ATLR in an iMRI suite. Preoperative tractography of the optic radiation was displayed on the navigation and operating microscope displays either without (9 patients) or with (12 patients) correction for brain shift. VFD were quantified using Goldmann perimetry and eligibility to drive was assessed by binocular Esterman perimetry 3 months after surgery. Secondary outcomes included seizure freedom and extent of hippocampal resection. The comparator was a cohort of 44 patients who underwent ATLR without iMRI. RESULTS: The VFD in the contralateral superior quadrant were significantly less (p = 0.043) with iMRI guidance (0%-49.2%, median 14.5%) than without (0%-90.9%, median 24.0%). No patient in the iMRI cohort developed a VFD that precluded driving whereas 13% of the non-iMRI cohort failed to meet UK driving criteria. Outcome did not differ between iMRI guidance with and without brain shift correction. Seizure outcome and degree of hippocampal resection were unchanged. CONCLUSIONS: Display of the optic radiation with image guidance reduces the severity of VFD and did not affect seizure outcome or hippocampal resection. Correction for brain shift is possible but did not further improve outcome. Future work to incorporate tractography into conventional neuronavigation systems will make the work more widely applicable.

A data-driven model of biomarker changes in sporadic Alzheimer's disease.

We demonstrate the use of a probabilistic generative model to explore the biomarker changes occurring as Alzheimer's disease develops and progresses. We enhanced the recently introduced event-based model for use with a multi-modal sporadic disease data set. This allows us to determine the sequence in which Alzheimer's disease biomarkers become abnormal without reliance on a priori clinical diagnostic information or explicit biomarker cut points. The model also characterizes the uncertainty in the ordering and provides a natural patient staging system. Two hundred and eighty-five subjects (92 cognitively normal, 129 mild cognitive impairment, 64 Alzheimer's disease) were selected from the Alzheimer's Disease Neuroimaging Initiative with measurements of 14 Alzheimer's disease-related biomarkers including cerebrospinal fluid proteins, regional magnetic resonance imaging brain volume and rates of atrophy measures, and cognitive test scores. We used the event-based model to determine the sequence of biomarker abnormality and its uncertainty in various population subgroups. We used patient stages assigned by the event-based model to discriminate cognitively normal subjects from those with Alzheimer's disease, and predict conversion from mild cognitive impairment to Alzheimer's disease and cognitively normal to mild cognitive impairment. The model predicts that cerebrospinal fluid levels become abnormal first, followed by rates of atrophy, then cognitive test scores, and finally regional brain volumes. In amyloid-positive (cerebrospinal fluid amyloid-ß1-42 < 192 pg/ml) or APOE-positive (one or more APOE4 alleles) subjects, the model predicts with high confidence that the cerebrospinal fluid biomarkers become abnormal in a distinct sequence: amyloid-ß1-42, phosphorylated tau, total tau. However, in the broader population total tau and phosphorylated tau are found to be earlier cerebrospinal fluid markers than amyloid-ß1-42, albeit with more uncertainty. The model's staging system strongly separates cognitively normal and Alzheimer's disease subjects (maximum classification accuracy of 99%), and predicts conversion from mild cognitive impairment to Alzheimer's disease (maximum balanced accuracy of 77% over 3 years), and from cognitively normal to mild cognitive impairment (maximum balanced accuracy of 76% over 5 years). By fitting Cox proportional hazards models, we find that baseline model stage is a significant risk factor for conversion from both mild cognitive impairment to Alzheimer's disease (P = 2.06 × 10(-7)) and cognitively normal to mild cognitive impairment (P = 0.033). The data-driven model we describe supports hypothetical models of biomarker ordering in amyloid-positive and APOE-positive subjects, but suggests that biomarker ordering in the wider population may diverge from this sequence. The model provides useful disease staging information across the full spectrum of disease progression, from cognitively normal to mild cognitive impairment to Alzheimer's disease. This approach has broad application across neurodegenerative disease, providing insights into disease biology, as well as staging and prognostication.