AI-based differential diagnosis of dementia etiologies on multimodal data.

Differential diagnosis of dementia remains a challenge in neurology due to symptom overlap across etiologies, yet it is crucial for formulating early, personalized management strategies. Here, we present an artificial intelligence (AI) model that harnesses a broad array of data, including demographics, individual and family medical history, medication use, neuropsychological assessments, functional evaluations and multimodal neuroimaging, to identify the etiologies contributing to dementia in individuals. The study, drawing on 51,269 participants across 9 independent, geographically diverse datasets, facilitated the identification of 10 distinct dementia etiologies. It aligns diagnoses with similar management strategies, ensuring robust predictions even with incomplete data. Our model achieved a microaveraged area under the receiver operating characteristic curve (AUROC) of 0.94 in classifying individuals with normal cognition, mild cognitive impairment and dementia. Also, the microaveraged AUROC was 0.96 in differentiating the dementia etiologies. Our model demonstrated proficiency in addressing mixed dementia cases, with a mean AUROC of 0.78 for two co-occurring pathologies. In a randomly selected subset of 100 cases, the AUROC of neurologist assessments augmented by our AI model exceeded neurologist-only evaluations by 26.25%. Furthermore, our model predictions aligned with biomarker evidence and its associations with different proteinopathies were substantiated through postmortem findings. Our framework has the potential to be integrated as a screening tool for dementia in clinical settings and drug trials. Further prospective studies are needed to confirm its ability to improve patient care.

AI-based differential diagnosis of dementia etiologies on multimodal data.

Differential diagnosis of dementia remains a challenge in neurology due to symptom overlap across etiologies, yet it is crucial for formulating early, personalized management strategies. Here, we present an AI model that harnesses a broad array of data, including demographics, individual and family medical history, medication use, neuropsychological assessments, functional evaluations, and multimodal neuroimaging, to identify the etiologies contributing to dementia in individuals. The study, drawing on 51,269 participants across 9 independent, geographically diverse datasets, facilitated the identification of 10 distinct dementia etiologies. It aligns diagnoses with similar management strategies, ensuring robust predictions even with incomplete data. Our model achieved a micro-averaged area under the receiver operating characteristic curve (AUROC) of 0.94 in classifying individuals with normal cognition, mild cognitive impairment and dementia. Also, the micro-averaged AUROC was 0.96 in differentiating the dementia etiologies. Our model demonstrated proficiency in addressing mixed dementia cases, with a mean AUROC of 0.78 for two co-occurring pathologies. In a randomly selected subset of 100 cases, the AUROC of neurologist assessments augmented by our AI model exceeded neurologist-only evaluations by 26.25%. Furthermore, our model predictions aligned with biomarker evidence and its associations with different proteinopathies were substantiated through postmortem findings. Our framework has the potential to be integrated as a screening tool for dementia in various clinical settings and drug trials, with promising implications for person-level management.

Deep learning for risk-based stratification of cognitively impaired individuals.

Quantifying the risk of progression to Alzheimer's disease (AD) could help identify persons who could benefit from early interventions. We used data from the Alzheimer's Disease Neuroimaging Initiative (ADNI, n = 544, discovery cohort) and the National Alzheimer's Coordinating Center (NACC, n = 508, validation cohort), subdividing individuals with mild cognitive impairment (MCI) into risk groups based on cerebrospinal fluid amyloid-ß levels and identifying differential gray matter patterns. We then created models that fused neural networks with survival analysis, trained using non-parcellated T1-weighted brain MRIs from ADNI data, to predict the trajectories of MCI to AD conversion within the NACC cohort (integrated Brier score: 0.192 [discovery], and 0.108 [validation]). Using modern interpretability techniques, we verified that regions important for model prediction are classically associated with AD. We confirmed AD diagnosis labels using postmortem data. We conclude that our framework provides a strategy for risk-based stratification of individuals with MCI and for identifying regions key for disease prognosis.

Multimodal deep learning for Alzheimer's disease dementia assessment.

Worldwide, there are nearly 10 million new cases of dementia annually, of which Alzheimer's disease (AD) is the most common. New measures are needed to improve the diagnosis of individuals with cognitive impairment due to various etiologies. Here, we report a deep learning framework that accomplishes multiple diagnostic steps in successive fashion to identify persons with normal cognition (NC), mild cognitive impairment (MCI), AD, and non-AD dementias (nADD). We demonstrate a range of models capable of accepting flexible combinations of routinely collected clinical information, including demographics, medical history, neuropsychological testing, neuroimaging, and functional assessments. We then show that these frameworks compare favorably with the diagnostic accuracy of practicing neurologists and neuroradiologists. Lastly, we apply interpretability methods in computer vision to show that disease-specific patterns detected by our models track distinct patterns of degenerative changes throughout the brain and correspond closely with the presence of neuropathological lesions on autopsy. Our work demonstrates methodologies for validating computational predictions with established standards of medical diagnosis.

How Much Is That Head CT? Price Transparency and Variability in Radiology.

PURPOSE: To extend the investigation of price transparency and variability to medical imaging. METHODS: Eighteen upper-tier academic hospitals identified by U.S. News & World Report and 14 of the 100 largest private radiology practices in the country identified by the Radiology Business Journal were contacted by telephone between December 2013 and February 2014 to determine the cash price for a noncontrast head CT. The price for a noncontrast head CT was chosen to assess price transparency in medical imaging because it represents a standard imaging examination with minimal differences in quality. RESULTS: Fourteen upper-tier academic hospitals (78%) and 11 private practices (79%) were able to provide prices for a noncontrast head CT. There was no significant difference between the proportions of upper-tier academic hospitals and private practices that were able to provide prices for a noncontrast head CT (P = .96). The average total price for the upper-tier academic hospitals was $1,390.12 ± $686.13, with the price ranging from $391.62 to $2,015. The average total price for the private practices was $681.60 ± $563.58, with the total price ranging from $211 to $2,200. CONCLUSIONS: Prices for a noncontrast head CT study were readily available from the vast majority of upper-tier academic hospitals and private practices, although there was tremendous variation in the price estimates both within and between the upper-tier academic hospitals and private practices. Routine medical imaging thus appears to be more price transparent compared with other health care services.

Radiation exposure from CT-guided ablation of renal masses: effects on life expectancy.

OBJECTIVE. The purpose of this article is to project the effects of radiation exposure on life expectancy (LE) in patients who opt for CT-guided radiofrequency ablation (RFA) instead of surgery for renal cell carcinoma (RCC). MATERIALS AND METHODS. We developed a decision-analytic Markov model to compare LE losses attributable to radiation exposure in hypothetical 65-year-old patients who undergo CT-guided RFA versus surgery for small (≤ 4 cm) RCC. We incorporated mortality risks from RCC, radiation-induced cancers (for procedural and follow-up CT scans), and all other causes; institutional data informed the RFA procedural effective dose. Radiation-induced cancer risks were generated using an organ-specific approach. Effects of varying model parameters and of dose-reduction strategies were evaluated in sensitivity analysis. RESULTS. Cumulative RFA exposures (up to 305.2 mSv for one session plus surveillance) exceeded those from surgery (up to 87.2 mSv). In 65-year-old men, excess LE loss from radiation-induced cancers, comparing RFA to surgery, was 11.7 days (14.6 days for RFA vs 2.9 days for surgery). Results varied with sex and age; this difference increased to 14.6 days in 65-year-old women and to 21.5 days in 55-year-old men. Dose-reduction strategies that addressed follow-up rather than procedural exposure had a greater impact. In 65-year-old men, this difference decreased to 3.8 days if post-RFA follow-up scans were restricted to a single phase; even elimination of RFA procedural exposure could not achieve equivalent benefits. CONCLUSION. CT-guided RFA remains a safe alternative to surgery, but with decreasing age, the higher burden of radiation exposure merits explicit consideration. Dose-reduction strategies that target follow-up rather than procedural exposure will have a greater impact.

Minocycline blocks bilirubin neurotoxicity and prevents hyperbilirubinemia-induced cerebellar hypoplasia in the Gunn rat.

Encephalopathy induced by hyperbilirubinemia in infants has been described in the medical literature for over a century but neither the cellular nor molecular mechanisms underlying bilirubin neurotoxicity are well understood. In this study, we have demonstrated that minocycline potently protects primary cultured rat cerebellar granule neurons against bilirubin neurotoxicity (IC50 approximately 2 microm) and almost completely blocks cerebellar hypoplasia and the profound loss of Purkinje and granule neurons observed in homozygous Gunn rats, a genetic model of hyperbilirubinemia-induced neurotoxicity. Minocycline-treated newborn Gunn rats had nearly equivalent numbers of viable Purkinje and granule neurons in the cerebellum as did control animals. Moreover, minocycline inhibits the bilirubin-induced phosphorylation of p38 mitogen-activated protein kinase both in vivo as well as in vitro. Taken together our data demonstrate that minocycline is able to greatly reduce bilirubin-induced neurotoxicity and suggest that minocycline's neuroprotective effects may be due in part to an inhibition of p38 mitogen-activated protein kinase activity. Our findings may lead to novel approaches for treating bilirubin-induced encephalopathy.