Unmet Need
More than 1 in 7, that is 15% of US adults, or 37 million people, are estimated to have chronic kidney disease (CKD). As many as 9 in 10 adults with CKD do not know they have CKD, and about 2 in 5 adults with severe CKD do not know they have CKD (CDC). Furthermore, the prevalence of autosomal dominant polycystic kidney disease (ADPKD) is 1:1000. It is estimated that less than one-half of these cases will be diagnosed during the patient's lifetime, as the disease is often clinically silent. In clinical practice, renal damage is generally detected by proteinuria/albuminuria on urinalysis or quantitative measurement, changes in serum creatinine concentration for estimation of glomerular filtration rate, or both. However, these methods have major limitations as they are nonspecific and frequently are also late manifestations of renal damage (Good et al.), as the start of renal failure may not give any symptoms initially (Polat et al.). A method for detection and diagnosis of kidney diseases during routine kidney checks would therefore aid in earlier detection of these diseases, thereby reducing the number of cases that go unnoticed, allowing for treatment. Therefore, there is a strong need for an automated, end-to-end tool that uses deep learning for the characterization of kidney and renal lesion on routine magnetic resonance imaging (MRI) and tracking growth over time in order to address this problem.
Technology Overview
Researchers at Johns Hopkins have developed Artificial Intelligence Longitudinal Growth In Kidneys (Al LOGIK): an automated, end-to-end tool that uses deep learning for the characterization of kidney and renal lesion on routine magnetic resonance imaging (MRI) and tracking growth over time. Changes in kidney volume over time have been implicated in progression of chronic kidney disease (CKD) and polycystic kidney disease (PKD). In PKD, fibrosis and extensive cyst formation leads to significantly enlarged renal volume, which can be characterized by MRI. By automatically segmenting whole kidney volume, AI LOGIK therefore provides an objective, quantitative assessment of disease progression. Importantly, identifying patients with rapidly progressing disease would facilitate selection for clinical trials and treatment. Similarly, CKD is caused by a change in renal function and/or structure over time. Structural changes as assessed by MRI are important for accurate staging and inform prognosis. Beyond CKD and PKD, AI LOGIK also segments and tracks the progression of renal lesions. The system accurately localizes the lesion, classifies lesion as benign vs. malignant, and tracks lesion volume over time to assist in diagnosis, prognosis, and treatment planning.
Stage of Development
The researchers are currently nearing a prototype of this product with additional validation and improvement of user interface pending.
Publication
N/A
