Ehud Karavani

Ehud Karavani

Auto-generated profile, based on publicly available information

Ehud Karavani

AI for Accelerated Healthcare and Life Sciences Department, IBM Research - Israel, Haifa, Israel

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 August 2019 in Radiology
Verified
Journal Article
Closed access
AAAyelet Akselrod-Ballin
Ayelet Akselrod-Ballin
MCMichal Chorev
Michal Chorev
YSYoel Shoshan
Yoel Shoshan
ASAdam Spiro
Adam Spiro
AHAlon Hazan
Alon Hazan
MGMichal Guindy
Michal Guindy
A deep learning algorithm trained on a linked data set of mammograms and electronic health records achieved breast cancer detection accuracy comparable to radiologists as defined by the Breast Canc... BackgroundComputational models on the basis of deep neural networks are increasingly used to analyze health care data. However, the efficacy of traditional computational models in radiology is a ma...

19 April 2023
Verified
Conference Paper
Open access
GGGrace Guo
Grace Guo
EKEhud Karavani
Ehud Karavani
AEAlex Endert
Alex Endert
BKBum Chul Kwon
Bum Chul Kwon
Causal inference is a statistical paradigm for quantifying causal effects using observational data. It is a complex process, requiring multiple steps, iterations, and collaborations with domain experts. Analysts often rely on visualizations to evaluate the accuracy of each step. However, existing visualization toolkits are not designed to support the entire causal inference process within computational environments familiar to analysts. In this paper, we address this gap with Causalvis, a Python visualization package for causal inference. Working closely with causal inference experts, we adopted an iterative design process to develop four interactive visualization modules to support causal inference analysis tasks. The modules are then presented back to the experts for feedback and evaluation. We found that Causalvis effectively supported the iterative causal inference process. We discuss the implications of our findings for designing visualizations for causal inference, particularly for tasks of communication and collaboration.

6 December 2024 in Alzheimer's & Dementia
Verified
Journal Article
Open access
JXJielin Xu
Jielin Xu
WSWenqiang Song
Wenqiang Song
ZXZhenxing Xu
Zhenxing Xu
MDMichael M. Danziger
Michael M. Danziger
EKEhud Karavani
Ehud Karavani
FCFeixiong Cheng
Feixiong Cheng
INTRODUCTION: High microglial heterogeneities hinder the development of microglia‐targeted treatment for Alzheimer's disease (AD). METHODS: We integrated 0.7 million single‐nuclei RNA‐sequencing transcriptomes from human brains using a variational autoencoder. We predicted AD‐relevant microglial subtype‐specific transition networks for disease‐associated microglia (DAM), tau microglia, and neuroinflammation‐like microglia (NIM). We prioritized drugs by specifically targeting microglia‐specific transition networks and validated drugs using two independent real‐world patient databases. RESULTS: We identified putative AD molecular drivers (e.g., SYK , CTSB , and INPP5D ) in transition networks of DAM and NIM. Via specifically targeting NIM, we identified that usage of ketorolac was associated with reduced AD incidence in both MarketScan (hazard ratio [HR] = 0.89) and INSIGHT (HR = 0.83) Clinical Research Network databases, mechanistically supported by ketorolac‐treated transcriptomic data from AD patient induced pluripotent stem cell–derived microglia. DISCUSSION: This study offers insights into the pathobiology of AD‐relevant microglial subtypes and identifies ketorolac as a potential anti‐inflammatory treatment for AD. Highlights: An integrative analysis of ≈ 0.7 million single‐nuclei RNA‐sequencing transcriptomes from human brains identified Alzheimer's disease (AD)–relevant microglia subtypes. Network‐based analysis identified putative molecular drivers (e.g., SYK, CTSB, INPP5D ) of transition networks between disease‐associated microglia (DAM) and neuroinflammation‐like microglia (NIM). Via network‐based prediction and population‐based validation, we identified that usage of ketorolac (a US Food and Drug Administration–approved anti‐inflammatory medicine) was associated with reduced AD incidence in two independent patient databases. Mechanistic observation showed that ketorolac treatment downregulated the Type‐I interferon signaling in patient induced pluripotent stem cell–derived microglia, mechanistically supporting its protective effects in real‐world patient databases.

15 April 2024 in Epidemiology
Verified
Journal Article
Open access
RGRom Gutman
Rom Gutman
EKEhud Karavani
Ehud Karavani
YSYishai Shimoni
Yishai Shimoni
Theoretical guarantees for causal inference using propensity scores are partially based on the scores behaving like conditional probabilities. However, prediction scores between zero and one do not necessarily behave like probabilities, especially when output by flexible statistical estimators. We perform a simulation study to assess the error in estimating the average treatment effect before and after applying a simple and well-established postprocessing method to calibrate the propensity scores. We observe that postcalibration reduces the error in effect estimation and that larger improvements in calibration result in larger improvements in effect estimation. Specifically, we find that expressive tree-based estimators, which are often less calibrated than logistic regression-based models initially, tend to show larger improvements relative to logistic regression-based models. Given the improvement in effect estimation and that postcalibration is computationally cheap, we recommend its adoption when modeling propensity scores with expressive models.

26 March 2021 in medRxiv
Verified
Preprint
Open access
PDF file
TETal El-Hay
Tal El-Hay
EKEhud Karavani
Ehud Karavani
APAsaf Peretz
Asaf Peretz
MNMatan Ninio
Matan Ninio
SRSivan Ravid
Sivan Ravid
AJAnil Jain
Anil Jain
Background The 2019 novel coronavirus (SARS-CoV-2) is reported to result in both respiratory and non-respiratory severe health outcomes, but quantitative assessment of the risk – while adjusting for underlying risk driven by comorbidities – is not yet established. Methods A retrospective observational study using electronic health records of 9,344,021 individuals across the U.S. with at-least 1 year of clinical history and followed up throughout 2020. Results 131,329 individuals were associated with SARS-CoV-2 infection by January 6, 2021 in three distinct surges. While the age and number of preexisting conditions had decreased throughout the pandemic, the characteristics of those who experienced severe health events did not. During the second surge, between June 7 and November 18, 2020, 425,988 individuals in the base cohort were admitted to emergency rooms or hospitals. Among them, 15,486 were detected with SAR-CoV-2 within few days of admission. Significant adjusted odds ratios were observed between SARS-CoV-2 infection and the following severe health events: respiratory (4.38, 95% confidence interval 4.16– 4.62), bacterial pneumonia (3.25, 2.76–3.83), sepsis (1.71, 1.53–1.91), renal (1.69, 1.57–1.83), hematologic/immune (1.32, 1.20–1.45), neurological (1.23, 1.09–1.38). Conclusions SARS-CoV-2 infection among hospitalized patients is associated with non-negligible increased risk of severe events including multiple non-respiratory ones. These associations, which complement recent studies, are persistent even after accounting for sources of selection and confounding bias, increasing the confidence they are not spurious.

8 October 2026 in Nature Neuroscience
Verified
Journal Article
Open access
PDF file
YHYuan Hou
Yuan Hou
YLYichen Li
Yichen Li
PZPengyue Zhang
Pengyue Zhang
NLNoah Lorincz-Comi
Noah Lorincz-Comi
DGDhruv Gohel
Dhruv Gohel
FCFeixiong Cheng
Feixiong Cheng
Despite prolific genomic data from individuals with Alzheimer’s disease (AD), translation into treatments has lagged. Here we applied Mendelian randomization to multiple AD genome-wide association studies and functional multiomics datasets to nominate druggable targets, identifying 19 targets for individuals of European ancestry and 7 for African ancestry. This included epoxide hydrolase 2 (EPHX2), for which we identified a protective missense variant (p.Arg287Gln). Experiments in induced pluripotent stem cell-derived neurons and cerebral organoids from patients with AD validated the protective role of p.Arg287Gln via reduced levels of EPHX2 and phosphorylated tau (pTau-181). Pharmacological inhibition of EPHX2 improved cognitive performance in the 5xFAD mouse model of AD. Finally, by leveraging electronic health data from 111,680 patients with mild cognitive impairment, we identified 12 drugs associated with reduced AD incidence, including trazodone (targeting ADRA1A) and baclofen (targeting GABBR1). Thus, integrating multiomics data, functional validation and electronic health data identifies new approaches for treating AD.