Cystoscopy-augmenting tools to improve the sensitivity and specificity of detection of urothelial carcinoma in situ

About This Grant

ABSTRACT: Up to 75% of patients diagnosed with high-grade, non-muscle-invasive bladder cancer (BC) will develop recurrences. Left undetected or untreated, recurrent bladder cancer leads to muscle invasion and radical cystectomy, a major surgery with high morbidity and mortality. Thus, clinical guidelines recommend surveillance every 3–12 months, making BC the costliest cancer to treat over a patient’s lifetime. The standard- of-care tool for BC surveillance, in-office white light cystoscopy (WLC), has low sensitivity and specificity for the leading cause of recurrence: urothelial carcinoma in situ (CIS). The most promising technique to improve the sensitivity of CIS detection, blue light cystoscopy (BLC), has low clinical adoption because it is costly, time- consuming, and has low specificity for CIS. Our long-term goal is to improve the detection of CIS during in- office surveillance. The main objective of this proposal is to develop a multimodal tool with high sensitivity and specificity for CIS that is appropriate for in-office surveillance: real-time digitally stained WLC (dsWLC)–guided polarization-sensitive optical coherence tomography (PS-OCT). Digital staining uses deep learning to convert WLC images into accurate, low-cost, BLC look-alikes. PS-OCT is a microscopic imaging technique proven to improve the specificity of BLC. Our preliminary data and published literature confirm that PS-OCT has the specificity to differentiate CIS from inflammation, a primary contributor to false positives in WLC and BLC that leads to unnecessary biopsies and associated co-morbidities. The proposed two-prong strategy of dsWLC and PS-OCT will thus bring the benefits of BLC (high sensitivity) to the office while 1) removing the administrative burdens that prevent its widespread adoption and 2) adding the specificity that it lacks. Our central hypothesis is that dsWLC-guided PS-OCT has better sensitivity of detection for CIS than the standard of care for in-office surveillance (WLC), comparable sensitivity to BLC, and better specificity than both. Our specific aims are: 1) Improve the sensitivity of in-office CIS detection by developing dsWLC, 2) Improve the specificity of in-office CIS detection by building a miniature PS-OCT probe, and 3) Demonstrate the clinical effectiveness of real-time dsWLC+PS-OCT for bladder cancer detection. Aim 1 will build on our expertise in machine learning and our preliminary data showing the first application of digital staining to cystoscopy. In Aim 2, we will develop a miniature PS-OCT probe, delivering a technical innovation in size and speed that meets the clinical need for in- office surveillance. Aim 3 will assess the clinical effectiveness of dsWLC+PS-OCT by comparing its sensitivity and specificity to existing clinical standards, WLC and BLC. Individually, both dsWLC and PS-OCT bring significant innovation to the clinical workflow by improving, respectively, the sensitivity and specificity of CIS detection. Used together, they will inspire a new paradigm for the bladder cancer surveillance workflow to increase confidence in detection and eradication of high-grade disease, lower morbidity due to unnecessary biopsies, lower BC recurrence and progression, bring significant healthcare savings and raise quality of life.