Ulas Bagci
Northwestern University, IL·Computer Science
Medical scans contain more information than doctors can easily read by eye, and patients' health records add still more. This lab builds artificial intelligence (AI) that combines scans with clinical data to make earlier and more reliable predictions.
- Fields
Computer Science
Biomedical Engineering
Medicine & Clinical Research
- Topics
explainable AI
deep learning
medical image analysis
- Grants
- 2 known grants, through May 2028
What they work on
- Identify CT imaging and electronic health record features linked to post-COVID pulmonary fibrosis, using automated lung and body composition analysis of about 1,150 scans from four medical centers.
- Build Transformer-embedded explainable capsule networks (X-TCaps), image-analysis neural networks, with the IBA visual explanation algorithm to predict which COVID-19 patients will develop lung fibrosis.
- Select the most predictive imaging and clinical biomarkers of post-COVID fibrosis using the lab's optimal biomarker (OBM) feature-selection method.
- Develop a pre-processing framework that corrects prostate MRI artifacts and noise, and a visual bias-estimation tool (ViBeT) that flags imbalances in training data.
- Create and validate HIT-PIRADS, capsule-based AI that combines multiparametric MRI with PSA, clinical and demographic data to detect and risk-score prostate cancer, in over 7,000 patients.
More about the research
A central goal is explainability: models that show why they reached a decision, so physicians can trust them. Two projects apply this approach to lung scarring after COVID-19 and to prostate cancer.
Why it matters
Early warning of post-COVID lung scarring could let clinicians start anti-fibrotic treatment sooner for at-risk survivors. More accurate, bias-corrected prostate MRI scoring could catch more dangerous cancers while sparing men unnecessary biopsies. It could also bring reliable prostate imaging to community hospitals that lack specialist radiologists. Explainable outputs may help physicians trust and use these tools.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Flemming Forsberg
Thomas Jefferson University, PA·Biomedical Engineering
Implement multiparametric ultrasound, combining contrast-enhanced, subharmonic, dispersion and viscoelastography imaging, on a commercial 3D transrectal probe, and adapt an MRI-based fusion and registration system for it.
Jennifer A Kim
Yale University, CT·Medicine & Clinical Research
Develop an algorithm that identifies post-traumatic epilepsy patients from electronic health records, and predict future PTE risk from the first seven days after TBI using structured data and natural language processing of clinical notes.
Jie Xu
University of Florida, FL·Computer Science·Early-career
Develop transformer-based vision-language models that outline kidney masses on CT scans, extract clinical concepts from notes, and generate radiology reports, validated against expert annotations.
The school
Northwestern University: 71 labs on PI Pages, $1.3B a year on research, 536 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.