B-089 Development of Nanosensor Enabled Liquid Biopsy for Ovarian Cancer Recurrence

Abstract

The vast majority of high-grade serous ovarian carcinoma (HGSOC) cases are diagnosed at an advanced stage, contributing to the high rate of mortality. Furthermore, the probability of HGSOC recurrence is high after remission. Current HGSOC monitoring paradigms, such as cancer antigen 125 (CA-125) measurements, are ineffective due to a lack of disease specificity. We approach HGSOC diagnosis and monitoring in a biomarker-agnostic manner. Rather than quantifying a single marker, we use a diverse array of non-specific nanosensor elements coupled with machine learning models to identify disease states in patient samples. Previously, we used this machine perception liquid biopsy (MPLB) approach to develop a prediction model that identified HGSOC with 87% sensitivity at 98% specificity. This investigation aims to apply MPLB to the detection and monitoring of HGSOC recurrence. Plasma samples were collected over two years from 145 HGSOC patients in remission undergoing periodic CA125 measurements. These patient samples were screened with an array of covalently modified single-walled carbon nanotube (SWCNT) sensors. Covalent modifications are introduced to modulate the near-infrared (nIR) photoluminescence of SWCNTs, creating novel emission peaks that increase the sensitivity of SWCNT-based sensors. Two new hydrogen peroxide-mediated functionalization reactions have been developed for use in this assay. These chemical modifications, as well as the choice of SWCNT wrapping, modulate which proteins adsorb onto the sensor surface when incubated with patient plasma samples, resulting in changes in nanosensor emissions. The spectral changes from different SWCNTs incubated in the patient samples were quantified, and, along with clinical information about the patient, such as recurrence status and time since diagnosis, were used to train binary prediction machine learning models to identify recurrence. This ongoing work aims to determine minimal residual disease and risk of recurrence in HGSOC patients using a nanosensor-enabled liquid biopsy tool. The novel functionalization chemistries developed for this project result in SWCNTs with bright and sensitive nIR emission. The ratio of the D band to the G band of the Raman spectra of the functionalized SWCNTs was at least 0.15, suggesting the sensors were well functionalized with sp3 defects. The resulting MPLB platform will enable risk stratification of HGSOC recurrence in surgically treated patients and may influence treatment decision-making for HGSOC. Future work will focus on proteomic analysis of adsorbed biomolecules on the sensor surface to determine what drives disease-specific sensor response.