Research Projects
Isoattenuating Gallstone Detection
Approximately 15% of the adult population is affected by gallstones. Complications from gallstones include acute cholecystitis, biliary obstruction, pancreatitis, and even gallbladder carcinoma. Adding to this complication is the fact that 30% of all gallstones appear isoattenuating with bile in conventional CT images (see image above) making them difficult to detect. These patients often require further imaging with ultrasound or magnetic resonance (MRCP) that increases both the cost to the patient and the time to diagnosis. To meet this clinical need we have recently developed a Dual-Energy CT method that is capable of differentiating solid cholesterol gallstones from liquid bile (and other tissues) and segments them in a single image using colored pixels (see right image above). This allows the radiologist to quickly determine the exact location and size of gallstones that were previously invisible on conventional CT images, thus eliminating the need for subsequent imaging. For more information on this research project please contact Todd Soesbe, Ph.D.
Liver Fat Quantification
Non-alcoholic fatty liver disease (NAFLD) is highly prevalent in the United States where it affects one-third of the population. This silent epidemic usually remains asymptomatic until the late stages of the disease that include non-alcoholic steatohepatitis (NASH), fibrosis, cirrhosis, and even liver cancer (hepatocellular carcinoma). Currently, there are no approved therapies other than weight loss making early detection essential for positive patient outcomes. If the liver fat fraction could be included as a standard biometric in all abdominal CT scans regardless of iodinated contrast, then this would help to diagnose early-stage asymptomatic NAFLD. To meet this clinical need we have recently developed a Dual-Energy CT method capable of measuring the liver fat fraction in both non-contrast and contrast-enhanced scans. This method gives the same fat fraction regardless of the scan phase as seen in the conventional and fat fraction images above. For more information on this research project please contact Lakshmi Ananthakrishnan, M.D..
Artificial Intelligence
The additional x-ray attenuation information acquired by Dual-Energy CT is ideally suited for analysis using Artificial Intelligence where multiple layers of images can be used as the input. Convolutional Neural Networks like UNET that specialize in medical image processing can use this additional information to perform automatic liver segmentation (as shown above), removal of iodinated contrast (i.e., virtual non-contrast), isoattenuating gallstone detection, liver fat fraction measurements, and other opportunistic screenings. For more information on this research project please contact Yee Seng Ng, M.D..
Minimum-Iodine CTA
Contrast-enhanced CT examinations represent approximately half of all clinical CT scans and utilize the administration of intravenously injected iodinated contrast agents. Unfortunately, approximately 1% of all contrast-enhanced CT patients suffer from allergic-like or physiologic reactions due to these intravascular iodinated contrast agents. This represents a huge clinical problem for hundreds of thousands of patients who are in need of the diagnostic benefits of contrast-enhanced CT imaging but have risks associated with intravascular iodine. Therefore, creating safe options for these patients is a prevailing important clinical need. In order to meet this clinical need we have recently developed a Dual-Energy CT method that maximizes the conspicuity of intravascular contrast agents. With this method the true minimum dose of iodine contrast can be used, which is approximately one-tenth of a standard dose of iodine. Alternately, this method can also be used with gadolinium contrast using standard magnetic resonance doses (as shown above). For more information on this research project please contact Todd Soesbe, Ph.D..