Point-of-Care Diagnostic for Spontaneous Bacterial Peritonitis

PROJECT SUMMARY Spontaneous bacterial peritonitis (SBP) is a frequent and severe complication of liver disease, in which an infection occurs in the ascitic fluid of the peritoneal cavity without an identifiable source of infection. Chronic liver disease is a leading cause of mortality and affects over 120 million people worldwide. Hospitalizations resulting from complications of chronic liver disease carry an economic burden of over $80 billion annually in the US, with the majority of these resources directed towards patients with cirrhosis. SBP affects around 10- 30% of patients hospitalized with cirrhosis and ascites and carries a substantial in-hospital mortality rate ranging from 17 to 30%. Prompt diagnosis and treatment are critical to prevent potential complications, including renal failure and sepsis. SBP is diagnosed when the polymorphonuclear leukocyte (PMN, also known as granulocytes) count is 250 cells/mm3 or greater. Currently, it can take days to diagnose SBP after presentation of symptoms, with testing delays leading to significantly increased mortality. Diagnosis without PMN is ineffective, as approximately a third of patients presenting with ascites lack any additional symptoms to indicate SBP. A rapid point-of-care (POC) diagnostic tool would enable immediate clinical decision-making, facilitating prompt initiation of antibiotic therapy and enhancing the overall management of SBP. We propose an innovative methodology building upon our recent breakthrough in label-free deep ultraviolet (UV) assay technology. This approach integrates microfluidic-based sample preparation and advanced algorithm-driven cell classification, facilitating on-the-spot cell counting in ascitic fluid. Our earlier research successfully utilized label-free deep UV microscopy for comprehensive 5-part white blood cell differentials, red blood cell quantification, and platelet identification in peripheral blood samples. The imaging system, characterized by a straightforward optical path, allows differential cell analysis using images captured from a single wavelength LED light source. Pseudo-colorized UV images, derived from deep-UV microscopy, have demonstrated diagnostic efficacy equivalent to the gold standard Giemsa-stained images in manual blood cell differentials. Leveraging our proprietary cell segmentation and classification algorithms, we can analyze these images to categorize various blood cell types, enabling automated differentials in ~3 minutes. This pioneering approach holds significant promise for the development of a POC SBP diagnostic with equivalent accuracy to a clinical pathology lab. In Phase I of this project, we propose to 1) visualize and identify PMNs and red blood cells (RBCs) in ascitic fluid using deep-UV microscopy and 2) automate the cell identification and classification using a machine learning-based algorithm. Phase II efforts will focus on optimization of sample preparation and translation of our prototype instrument into a user-friendly device. Successful commercialization of this device will enable faster diagnosis of SBP, which is critical for reduction of mortality.