Abdominal organs
Body composition analysis
Most CT and MRI exams of the abdomen are done to assess the condition of abdominal organs. But body composition can help predict health and the ability to withstand great stresses such as major surgery.
In this project, the Radiology Informatics Lab uses images of the abdomen to compute specific markers of body composition. These markers include fat within and outside of the abdomen and the thickness of the abdominal musculature. Then the lab assesses these images to predict health and survival. The lab also is examining related biomarkers to see whether they can predict health.
Cholangiocarcinoma
The 3D DenseNet-121 deep learning model was developed to analyze MRI scans and detect early-stage perihilar cholangiocarcinoma in patients with primary sclerosing cholangitis. The model, which has been tested on 398 patients across multiple international centers, significantly outperformed expert radiologists in sensitivity — 87.9% versus 50%. The model caught far more curable cancers at a stage when transplantation is still an option.
Artificial intelligence: Cholangiocarcinoma care
This illustration depicts the detection of early-stage perihilar cholangiocarcinoma in patients with primary sclerosing cholangitis.
Deep learning analysis of biopsy images for prognostic markers of chronic kidney disease
This study investigates whether deep learning can be an efficient alternative to traditional manual methods of assessing chronic kidney disease risks. The lab applied a deep learning model to kidney biopsy images to automatically estimate microstructures in the kidneys and identify prognostic markers for chronic kidney disease.
The research team is matching metrics from deep learning to manual evaluations to find correlations with clinical characteristics.
Kidney cyst segmentations
This example of segmentations of kidney cysts in MRI scans of patients with autosomal dominant polycystic kidney disease was obtained manually using the lab's automatic deep learning model.
Deep learning-based decision support tool for distinguishing uterine diseases, including endometriosis, adenomyosis and fibroid disease
Diagnosing uterine diseases, particularly endometriosis, is challenging. These diagnoses require a strong connection among radiologists, surgeons and pathologists because of the many ways and locations that endometriosis can appear. This project uses artificial intelligence (AI) to improve sensitivity and consistency in how this disease is diagnosed and managed.
Deep learning for automated segmentation of polycystic kidneys
While MRI is commonly used for people with polycystic kidney disease, ultrasound is portable and cost-effective. The lab believes deep learning can efficiently automate segmentation of kidney images in people with polycystic kidney disease.
The lab is exploring how effective ultrasound is at classifying polycystic kidney disease to see whether it can be an alternative or supplementary imaging method.
Detection of multiple myeloma lesions using whole-body CT scans
Dr. Erickson's lab is developing a fast and accurate computer-assisted tool to detect multiple myeloma lesions on whole-body, low-dose CT scans.
Estimation of acute pancreatitis and clinical course
The prediction of the clinical course of a patient presenting with moderate or severe acute pancreatitis is difficult, but it has important implications. Undertreatment can result in significant morbidity or mortality, but overtreatment can be very expensive. We are investigating AI tools to assist in this task.
Liver fat quantification via endoscopic ultrasound videos
This project focuses on predicting liver fat fractions directly from endoscopic ultrasound videos, correlating findings with magnetic resonance imaging — proton density fat fraction, the established clinical reference standard. To manage the complexity of dynamic ultrasound feeds, the researchers use a video-based multiple instance learning framework with a specialized attention mechanism.
By matching the predictive utility of MRI, this framework enables opportunistic screen of metabolic dysfunction-associated steatotic liver disease, formerly known as nonalcoholic fatty liver disease, during routine endoscopy. This could potentially eliminate the need for separate, costly MRI scans and make liver fat quantification more accessible.
Quantitative liver fat estimation and informative frame identification from routine endoscopic ultrasound videos
The model encodes endoscopic ultrasound video frames and uses attention-based multiple instance learning to aggregate frame-level features, proton density fat fraction and identify informative frames.
Multimodal prognostics and functional assessment in hepatobiliary surgery
This project leverages data-driven prognostic modeling to forecast individual patient outcomes. It specifically targets the prediction of posthepatectomy liver failure by integrating clinical scores with future liver remnant metrics. Also, the research evaluates the correlation between volumetric growth and functional recovery following liver augmentation procedures and establishes volumetric thresholds to define the oncologic benefit of cytoreductive surgery in neuroendocrine tumor metastases.
By bridging traditional surgical workflows with advanced analytics, this work aims to personalize surgical decision-making and enhance postoperative care.
Predictive analysis of primary sclerosing cholangitis
This project uses deep learning and topological data analysis to examine MRI scans. The project aims to predict outcomes in people with primary sclerosing cholangitis.
Using advanced analytical techniques, the research team seeks to make diagnoses more accurate and provide insights into disease progression and outcomes. These techniques offer a new approach to understanding and managing primary sclerosing cholangitis.
Imaging analysis for primary sclerosing cholangitis
This in-depth workflow diagram shows the use of algebraic topology-based machine learning to analyze imaging signals for diagnosing primary sclerosing cholangitis.