Christos Davatzikos
· Ph.D.University of Pennsylvania · Rehabilitation Medicine
Active 1992–2024
About
Christos Davatzikos is the Wallace T. Miller Sr. Professor of Radiology at the University of Pennsylvania and serves as the Director of the Center for Biomedical Image Computing and Analytics. He holds secondary appointments in Electrical and Systems Engineering, Bioengineering, and Applied Mathematics graduate groups at Penn. His educational background includes a B.S. in Electrical Engineering and Computer Science from the National Technical University of Athens in 1989 and a Ph.D. in Electrical and Computer Engineering from Johns Hopkins University in 1994, obtained on a Fulbright scholarship. Dr. Davatzikos's research interests focus on medical image analysis, neuroimaging, machine learning, and biomedical image computing. He has overseen a diverse research program that spans from fundamental problems of imaging pattern analysis to clinical studies involving aging, Alzheimer’s Disease, schizophrenia, brain cancer, and brain development. He founded and directed the Neuroimaging Laboratory and the section of biomedical image analysis at Penn. His contributions include developing computational neuroanatomy techniques using brain deformations, pattern classification of MRI data for disease diagnosis, and biophysical models for brain tumor interaction. Dr. Davatzikos has served on various scientific journal editorial boards and grant review committees, and he is recognized as an IEEE fellow and a fellow of the American Institute for Medical and Biological Engineering.
Research topics
- Computer Science
- Artificial Intelligence
- Medicine
- Machine Learning
- Data Mining
- Radiology
- Pathology
- Internal medicine
- Psychology
- Biology
- Neuroscience
- Medical physics
- Cognitive psychology
- Bioinformatics
- Psychiatry
- Cancer research
- Computational biology
- Oncology
- Cardiology
- Computer vision
- Database
- Anatomy
- Developmental psychology
- Data science
Selected publications
Image segmentations produced by BAMF under the AIMI Annotations initiative
arXiv (Cornell University) · 2024 · 369 citations
- Computer Science
- Artificial Intelligence
- Medical physics
The Imaging Data Commons (IDC)(https://imaging.datacommons.cancer.gov/) [1] connects researchers with publicly available cancer imaging data, often linked with other types of cancer data. Many of the collections have limited annotations due to the expense and effort required to create these manually. The increased capabilities of AI analysis of radiology images provide an opportunity to augment existing IDC collections with new annotation data. To further this goal, we trained several nnUNet [2] based models for a variety of radiology segmentation tasks from public datasets and used them to generate segmentations for IDC collections. To validate the model's performance, roughly 10% of the AI predictions were assigned to a validation set. For this set, a board-certified radiologist graded the quality of AI predictions on a Likert scale. If they did not 'strongly agree' with the AI output, the reviewer corrected the segmentation. This record provides the AI segmentations, Manually corrected segmentations, and Manual scores for the inspected IDC Collection images. Only 10% of the AI-derived annotations provided in this dataset are verified by expert radiologists . More details, on model training and annotations are provided within the associated manuscript to ensure transparency and reproducibility. This work was done in two stages. Versions 1.x of this record were from the first stage. Versions 2.x added additional records. In the Version 1.x collections, a medical student (non-expert) reviewed all the AI predictions and rated them on a 5-point Likert Scale, for any AI predictions in the validation set that they did not 'strongly agree' with, the non-expert provided corrected segmentations. This non-expert was not utilized for the Version 2.x additional records. Likert Score Definition: Guidelines for reviewers to grade the quality of AI segmentations. 5 Strongly Agree - Use-as-is (i.e., clinically acceptable, and could be used for treatment without change) 4 Agree - Minor edits that are not necessary. Stylistic differences, but not clinically important. The current segmentation is acceptable 3 Neither agree nor disagree - Minor edits that are necessary. Minor edits are those that the review judges can be made in less time than starting from scratch or are expected to have minimal effect on treatment outcome 2 Disagree - Major edits. This category indicates that the necessary edit is required to ensure correctness, and sufficiently significant that user would prefer to start from the scratch 1 Strongly disagree - Unusable. This category indicates that the quality of the automatic annotations is so bad that they are unusable. Zip File Folder Structure Each zip file in the collection correlates to a specific segmentation task. The common folder structure is ai-segmentations-dcm This directory contains the AI model predictions in DICOM-SEG format for all analyzed IDC collection files qa-segmentations-dcm This directory contains manual corrected segmentation files, based on the AI prediction, in DICOM-SEG format. Only a fraction, ~10%, of the AI predictions were corrected. Corrections were performed by radiologist (rad*) and non-experts (ne*) qa-results.csv CSV file linking the study/series UIDs with the ai segmentation file, radiologist corrected segmentation file, radiologist ratings of AI performance. qa-results.csv Columns The qa-results.csv file contains metadata about the segmentations, their related IDC case image, as well as the Likert ratings and comments by the reviewers. Column Description Collection The name of the IDC collection for this case PatientID PatientID in DICOM metadata of scan. Also called Case ID in the IDC StudyInstanceUID StudyInstanceUID in the DICOM metadata of the scan SeriesInstanceUID SeriesInstanceUID in the DICOM metadata of the scan Validation true/false if this scan was manually reviewed Reviewer Coded ID of the reviewer. Radiologist IDs start with ‘rad’ non-expect IDs start with ‘ne’ AimiProjectYear 2023 or 2024, This work was split over two years. The main methodology difference between the two is that in 2023, a non-expert also reviewed the AI output, but a non-expert was not utilized in 2024. AISegmentation The filename of the AI prediction file in DICOM-seg format. This file is in the ai-segmentations-dcm folder. CorrectedSegmentation The filename of the reviewer-corrected prediction file in DICOM-seg format. This file is in the qa-segmentations-dcm folder. If the reviewer strongly agreed with the AI for all segments, they did not provide any correction file. Was the AI predicted ROIs accurate? This column appears one for each segment in the task for images from AimiProjectYear 2023. The reviewer rates segmentation quality on a Likert scale. In tasks that have multiple labels in the output, there is only one rating to cover them all. Was the AI predicted {SEGMENT_NAME} label accurate? This column appears one for each segment in the task for images from AimiProjectYear 2024. The reviewer rates each segment for its quality on a Likert scale. Do you have any comments about the AI predicted ROIs? Open ended question for the reviewer Do you have any comments about the findings from the study scans? Open ended question for the reviewer File Overview brain-mr.zip Segment Description: brain tumor regions: necrosis, edema, enhancing IDC Collection: UPENN-GBM Links: model weights, github breast-fdg-pet-ct.zip Segment Description: FDG-avid lesions in breast from FDG PET/CT scans QIN-Breast IDC Collection: QIN-Breast Links: model weights, github breast-mr.zip Segment Description: Breast, Fibroglandular tissue, structural tumor IDC Collection: duke-breast-cancer-mri Links: model weights, github kidney-ct.zip Segment Description: Kidney, Tumor, and Cysts from contrast enhanced CT scans IDS Collection: TCGA-KIRC, TCGA-KIRP, TCGA-KICH, CPTAC-CCRCC Links: model weights, github liver-ct.zip Segment Description: Liver from CT scans IDC Collection: TCGA-LIHC Links: model weights, github liver2-ct.zip Segment Description: Liver and Lesions from CT scans IDC Collection: HCC-TACE-SEG, COLORECTAL-LIVER-METASTASES Links: model weights, github liver-mr.zip Segment Description: Liver from T1 MRI scans IDC Collection: TCGA-LIHC Links: model weights, github lung-ct.zip Segment Description: Lung and Nodules (3mm-30mm) from CT scans IDC Collections: Anti-PD-1-Lung LUNG-PET-CT-Dx NSCLC Radiogenomics RIDER Lung PET-CT TCGA-LUAD TCGA-LUSC Links: model weights 1, model weights 2, github lung2-ct.zip Improved model version Segment Description: Lung and Nodules (3mm-30mm) from CT scans IDC Collections: QIN-LUNG-CT, SPIE-AAPM Lung CT Challenge Links: model weights, github lung-fdg-pet-ct.zip Segment Description: Lungs and FDG-avid lesions in the lung from FDG PET/CT scans IDC Collections: ACRIN-NSCLC-FDG-PET Anti-PD-1-Lung LUNG-PET-CT-Dx NSCLC Radiogenomics RIDER Lung PET-CT TCGA-LUAD TCGA-LUSC Links: model weights, github prostate-mr.zip Segment Description: Prostate from T2 MRI scans IDC Collection: ProstateX, Prostate-MRI-US-Biopsy Links: model weights, github Changelog 2.0.2 - Fix the brain-mr segmentations to be transformed correctly 2.0.1 - added AIMI 2024 radiologist comments to qa-results.csv 2.0.0 - added AIMI 2024 segmentations 1.X - AIMI 2023 segmentations and reviewer scores
Scientific Reports · 2024 · 23 citations
Senior authorCorresponding- Computer Science
- Artificial Intelligence
- Computational biology
Glioblastoma is a highly heterogeneous disease, with variations observed at both phenotypical and molecular levels. Personalized therapies would be facilitated by non-invasive in vivo approaches for characterizing this heterogeneity. In this study, we developed unsupervised joint machine learning between radiomic and genomic data, thereby identifying distinct glioblastoma subtypes. A retrospective cohort of 571 IDH-wildtype glioblastoma patients were included in the study, and pre-operative multi-parametric MRI scans and targeted next-generation sequencing (NGS) data were collected. L21-norm minimization was used to select a subset of 12 radiomic features from the MRI scans, and 13 key driver genes from the five main signal pathways most affected in glioblastoma were selected from the genomic data. Subtypes were identified using a joint learning approach called Anchor-based Partial Multi-modal Clustering on both radiomic and genomic modalities. Kaplan-Meier analysis identified three distinct glioblastoma subtypes: high-risk, medium-risk, and low-risk, based on overall survival outcome (p < 0.05, log-rank test; Hazard Ratio = 1.64, 95% CI 1.17-2.31, Cox proportional hazard model on high-risk and low-risk subtypes). The three subtypes displayed different phenotypical and molecular characteristics in terms of imaging histogram, co-occurrence of genes, and correlation between the two modalities. Our findings demonstrate the synergistic value of integrated radiomic signatures and molecular characteristics for glioblastoma subtyping. Joint learning on both modalities can aid in better understanding the molecular basis of phenotypical signatures of glioblastoma, and provide insights into the biological underpinnings of tumor formation and progression.
Neuro-Oncology Advances · 2023 · 50 citations
- Computer Science
- Artificial Intelligence
- Medicine
Background: Brain tumors are the most common solid tumors and the leading cause of cancer-related death among all childhood cancers. Tumor segmentation is essential in surgical and treatment planning, and response assessment and monitoring. However, manual segmentation is time-consuming and has high interoperator variability. We present a multi-institutional deep learning-based method for automated brain extraction and segmentation of pediatric brain tumors based on multi-parametric MRI scans. Methods: = 21) subsets. DeepMedic, a three-dimensional convolutional neural network, was trained and the model parameters were tuned. Finally, the network was evaluated on the withheld internal and external test cohorts. Results: Dice similarity score (median ± SD) was 0.91 ± 0.10/0.88 ± 0.16 for the whole tumor, 0.73 ± 0.27/0.84 ± 0.29 for ET, 0.79 ± 19/0.74 ± 0.27 for union of all non-enhancing components (i.e., NET, CC, ED), and 0.98 ± 0.02 for brain tissue in both internal/external test sets. Conclusions: Our proposed automated brain extraction and tumor subregion segmentation models demonstrated accurate performance on segmentation of the brain tissue and whole tumor regions in pediatric brain tumors and can facilitate detection of abnormal regions for further clinical measurements.
Proceedings of the National Academy of Sciences · 2023 · 73 citations
Senior authorCorresponding- Machine Learning
- Artificial Intelligence
- Computer Science
Despite the great promise that machine learning has offered in many fields of medicine, it has also raised concerns about potential biases and poor generalization across genders, age distributions, races and ethnicities, hospitals, and data acquisition equipment and protocols. In the current study, and in the context of three brain diseases, we provide evidence which suggests that when properly trained, machine learning models can generalize well across diverse conditions and do not necessarily suffer from bias. Specifically, by using multistudy magnetic resonance imaging consortia for diagnosing Alzheimer's disease, schizophrenia, and autism spectrum disorder, we find that well-trained models have a high area-under-the-curve (AUC) on subjects across different subgroups pertaining to attributes such as gender, age, racial groups and different clinical studies and are unbiased under multiple fairness metrics such as demographic parity difference, equalized odds difference, equal opportunity difference, etc. We find that models that incorporate multisource data from demographic, clinical, genetic factors, and cognitive scores are also unbiased. These models have a better predictive AUC across subgroups than those trained only with imaging features, but there are also situations when these additional features do not help.
Scientific Reports · 2022 · 86 citations
Senior authorCorresponding- Computer Science
- Artificial Intelligence
- Machine Learning
Multi-omic data, i.e., clinical measures, radiomic, and genetic data, capture multi-faceted tumor characteristics, contributing to a comprehensive patient risk assessment. Here, we investigate the additive value and independent reproducibility of integrated diagnostics in prediction of overall survival (OS) in isocitrate dehydrogenase (IDH)-wildtype GBM patients, by combining conventional and deep learning methods. Conventional radiomics and deep learning features were extracted from pre-operative multi-parametric MRI of 516 GBM patients. Support vector machine (SVM) classifiers were trained on the radiomic features in the discovery cohort (n = 404) to categorize patient groups of high-risk (OS < 6 months) vs all, and low-risk (OS ≥ 18 months) vs all. The trained radiomic model was independently tested in the replication cohort (n = 112) and a patient-wise survival prediction index was produced. Multivariate Cox-PH models were generated for the replication cohort, first based on clinical measures solely, and then by layering on radiomics and molecular information. Evaluation of the high-risk and low-risk classifiers in the discovery/replication cohorts revealed area under the ROC curves (AUCs) of 0.78 (95% CI 0.70-0.85)/0.75 (95% CI 0.64-0.79) and 0.75 (95% CI 0.65-0.84)/0.63 (95% CI 0.52-0.71), respectively. Cox-PH modeling showed a concordance index of 0.65 (95% CI 0.6-0.7) for clinical data improving to 0.75 (95% CI 0.72-0.79) for the combination of all omics. This study signifies the value of integrated diagnostics for improved prediction of OS in GBM.
Federated learning enables big data for rare cancer boundary detection
Nature Communications · 2022 · 326 citations
- Computer Science
- Computer Science
- Machine Learning
Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.
Embracing the Disharmony in Medical Imaging: A Simple and Effective Framework for Domain Adaptation
Medical Image Analysis · 2021 · 58 citations
Senior authorCorresponding- Computer Science
- Computer Science
- Artificial Intelligence
Domain shift, the mismatch between training and testing data characteristics, causes significant degradation in the predictive performance in multi-source imaging scenarios. In medical imaging, the heterogeneity of population, scanners and acquisition protocols at different sites presents a significant domain shift challenge and has limited the widespread clinical adoption of machine learning models. Harmonization methods which aim to learn a representation of data invariant to these differences are the prevalent tools to address domain shift, but they typically result in degradation of predictive accuracy. This paper takes a different perspective of the problem: we embrace this disharmony in data and design a simple but effective framework for tackling domain shift. The key idea, based on our theoretical arguments, is to build a pretrained classifier on the source data and adapt this model to new data. The classifier can be fine-tuned for intra-site domain adaptation. We can also tackle situations where we do not have access to ground-truth labels on target data; we show how one can use auxiliary tasks for adaptation; these tasks employ covariates such as age, gender and race which are easy to obtain but nevertheless correlated to the main task. We demonstrate substantial improvements in both intra-site domain adaptation and inter-site domain generalization on large-scale real-world 3D brain MRI datasets for classifying Alzheimer's disease and schizophrenia.
JAMA Neurology · 2021 · 69 citations
- Medicine
- Internal medicine
- Cardiology
Importance: Meta-analyses of randomized clinical trials have indicated that improved hypertension control reduces the risk for cognitive impairment and dementia. However, it is unclear to what extent pathways reflective of Alzheimer disease (AD) pathology are affected by hypertension control. Objective: To evaluate the association of intensive blood pressure control on AD-related brain biomarkers. Design, Setting, and Participants: This is a substudy of the Systolic Blood Pressure Intervention Trial (SPRINT MIND), a multicenter randomized clinical trial that compared the efficacy of 2 different blood pressure-lowering strategies. Potential participants (n = 1267) 50 years or older with hypertension and without a history of diabetes or stroke were approached for a brain magnetic resonance imaging (MRI) study. Of these, 205 participants were deemed ineligible and 269 did not agree to participate; 673 and 454 participants completed brain MRI at baseline and at 4-year follow-up, respectively; the final follow-up date was July 1, 2016. Analysis began September 2019 and ended November 2020. Interventions: Participants were randomized to either a systolic blood pressure goal of less than 120 mm Hg (intensive treatment: n = 356) or less than 140 mm Hg (standard treatment: n = 317). Main Outcomes and Measures: Changes in hippocampal volume, measures of AD regional atrophy, posterior cingulate cerebral blood flow, and mean fractional anisotropy in the cingulum bundle. Results: Among 673 recruited patients who had baseline MRI (mean [SD] age, 67.3 [8.2] years; 271 women [40.3%]), 454 completed the follow-up MRI at a median (interquartile range) of 3.98 (3.7-4.1) years after randomization. In the intensive treatment group, mean hippocampal volume decreased from 7.45 cm3 to 7.39 cm3 (difference, -0.06 cm3; 95% CI, -0.08 to -0.04) vs a decrease from 7.48 cm3 to 7.46 cm3 (difference, -0.02 cm3; 95% CI, -0.05 to -0.003) in the standard treatment group (between-group difference in change, -0.033 cm3; 95% CI, -0.062 to -0.003; P = .03). There were no significant treatment group differences for measures of AD regional atrophy, cerebral blood flow, or mean fractional anisotropy. Conclusions and Relevance: Intensive treatment was associated with a small but statistically significant greater decrease in hippocampal volume compared with standard treatment, consistent with the observation that intensive treatment is associated with greater decreases in total brain volume. However, intensive treatment was not associated with changes in any of the other MRI biomarkers of AD compared with standard treatment. Trial Registration: ClinicalTrials.gov Identifier: NCT01206062.
Applications of Radiomics and Radiogenomics in High-Grade Gliomas in the Era of Precision Medicine
Cancers · 2021 · 74 citations
- Computer Science
- Medicine
- Medical physics
Machine learning (ML) integrated with medical imaging has introduced new perspectives in precision diagnostics of high-grade gliomas, through radiomics and radiogenomics. This has raised hopes for characterizing noninvasive and in vivo biomarkers for prediction of patient survival, tumor recurrence, and genomics and therefore encouraging treatments tailored to individualized needs. Characterization of tumor infiltration based on pre-operative multi-parametric magnetic resonance imaging (MP-MRI) scans may allow prediction of the loci of future tumor recurrence and thereby aid in planning the course of treatment for the patients, such as optimizing the extent of resection and the dose and target area of radiation. Imaging signatures of tumor genomics can help in identifying the patients who benefit from certain targeted therapies. Specifying molecular properties of gliomas and prediction of their changes over time and with treatment would allow optimization of treatment. In this article, we provide neuro-oncology, neuropathology, and computational perspectives on the promise of radiomics and radiogenomics for allowing personalized treatments of patients with gliomas and discuss the challenges and limitations of these methods in multi-institutional clinical trials and suggestions to mitigate the issues and the future directions.
QSIPrep: an integrative platform for preprocessing and reconstructing diffusion MRI data
Nature Methods · 2021 · 315 citations
- Computer Science
- Computer Science
- Data Mining
Recent grants
NIH · $1.7M · 2010
NIH · $706k · 2017–2027
Imaging signatures of genetic mutations in glioblastoma using machine learning
NIH · $8.7M · 2002–2025
Cancer imaging phenomics software suite: application to brain and breast cancer
NIH · $3.2M · 2015–2021
NIH · $6.3M · 2017–2025
Frequent coauthors
- 354 shared
Güray Erus
- 300 shared
Susan M. Resnick
National Institute on Aging
- 215 shared
Spyridon Bakas
Indiana University School of Medicine
- 209 shared
Aristeidis Sotiras
- 203 shared
Theodore D. Satterthwaite
Children's Hospital of Philadelphia
- 195 shared
Michael W. Weiner
University of California, San Francisco
- 188 shared
Jimit Doshi
University of Pennsylvania
- 183 shared
Clifford R. Jack
WinnMed
Education
- 1989
B.S., Electrical Engineering and Comp. Science
Nat. Tech. University of Athens
- 1994
Ph.D., Electrical and Comp. Eng.
Johns Hopkins University
Awards & honors
- IEEE Fellow
- Fellow of the American Institute for Medical and Biological…
Similar researchers at University of Pennsylvania
- Resume-aware match score
- Save to shortlist
- AI-drafted outreach
See your match with Christos Davatzikos
PhdFit ranks faculty by your research interests, methods, and publications — grounded in their actual work, not templates.
- Free to start
- No credit card
- 30-second signup