There is a need for biomarkers that can identify children with Congenital Anomalies of the Kidneys and Urinary Tract (CAKUT) early in life who are at high risk of future chronic kidney disease progression. Early identification would help guide trials of therapies for those most likely to benefit from early treatment and spare those patients who are at a low risk of progression from potential treatment-associated harms. As most children with structural kidney disease have frequent ultrasounds of their kidneys and bladder, measurements of the healthy kidney tissue and the health of the bladder may predict risk of kidney function decline.
Assessment of the characteristics of the kidney with imaging could provide important information about this risk. This research will apply machine learning methods to automatically extract informative features from ultrasound images and then, using the Chronic Kidney Disease in Children (CKiD) study, will validate these features as biomarkers of chronic kidney disease progression. These biomarkers may predict chronic kidney disease progression prior to the appearance of later serum or urine biomarkers, such as nadir creatinine or proteinuria, and can be measured noninvasively immediately after birth on routine clinical imaging.
Gregory Tasian, MD, MSc, MSCE
Dr. Gregory Tasian, MD, MSc, MSCE, is a pediatric urologist in the Division of Urology at Children’s Hospital of Philadelphia. Dr. Tasian’s research involves using large data analysis, behavioral economics, dietary assessment, and evaluation of urine and stool to discover determinants of kidney stone disease with the goal of translating this knowledge into new treatments for kidney stone prevention. He directs a large microbiome research program that recently published the first evidence that oral antibiotics increase the risk of stones. His group also analyzes ultrasound of the kidney to discover novel biomarkers of chronic kidney disease progression for children with congenital urologic disease. He collaborates with imaging experts at the University of Pennsylvania to use machine learning to develop automated methods to predict the risk of chronic kidney disease early in life when kidney function is still preserved. The long-term goal of Dr. Tasian’s research program is to improve the lives of children with nephrolithiasis and congenital urologic disease.
- S. Yin et al., “Multi-instance Deep Learning with Graph Convolutional Neural Networks for Diagnosis of Kidney Diseases Using Ultrasound Imaging”, Uncertainty for Safe Utilization of Machine Learning in Medical Imaging and Clinical Image-Based Procedures (CLIP), 2019, pp. 146-154 (Best Paper Award, CLIP 2019)
- S. Yin et al., "Fully-Automatic Segmentation Of Kidneys In Clinical Ultrasound Images Using A Boundary Distance Regression Network," 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019), Venice, Italy, 2019, pp. 1741-1744.
- Zheng Q et al., Computer aided diagnosis of congenital abnormalities of the kidney and urinary tract in children based on ultrasound imaging data by integrating texture image features and deep transfer learning image features. J Ped Urol. 2019 Feb;15(1):75.e1-75.e7
- Zheng Q et al., Transfer Learning for Diagnosis of Congenital Abnormalities of the Kidney and Urinary Tract in Children Based on Ultrasound Imaging Data. 2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018), Washington DC, 2018, pp. 1487-1490.
- Zheng Q et al., A Dynamic Cuts Method with Integrated Multiple Feature Maps for Segmenting Kidneys in 2D Ultrasound Images. Acad Radiology. 2018