Date of Award

Summer 8-15-2015

Author's School

School of Engineering & Applied Science

Author's Department

Computer Science & Engineering

Degree Name

Doctor of Philosophy (PhD)

Degree Type

Dissertation

Abstract

Image guided surgery (IGS) utilizes emerging imaging technologies to provide additional structural and functional information to the physician in clinical settings. This additional visual information can help physicians delineate cancerous tissue during resection as well as avoid damage to near-by healthy tissue. Near-infrared (NIR) fluorescence imaging (700 nm to 900 nm wavelengths) is a promising imaging modality for IGS, namely for the following reasons: First, tissue absorption and scattering in the NIR window is very low, which allows for deeper imaging and localization of tumor tissue in the range of several millimeters to a centimeter depending on the tissue surrounding the tumor. Second, spontaneous tissue fluorescence emission is minimal in the NIR region, allowing for high signal-to-background ratio imaging compared to visible spectrum fluorescence imaging. Third, decoupling the fluorescence signal from the visible spectrum allows for optimization of NIR fluorescence while attaining high quality color images. Fourth, there are two FDA approved fluorescent dyes in the NIR region—namely methylene blue (MB) and indocyanine green—which can help to identify tumor tissue due to passive accumulation in human subjects.

The aforementioned advantages have led to the development of NIR fluorescence imaging systems for a variety of clinical applications, such as sentinel lymph node imaging, angiography, and tumor margin assessment. With these technological advances, secondary surgeries due to positive tumor margins or damage to healthy organs can be largely mitigated, reducing the emotional and financial toll on the patient.

Currently, several NIR fluorescence imaging systems (NFIS) are available commercially or are undergoing clinical trials, such as FLARE, SPY, PDE, Fluobeam, and others. These systems capture multi-spectral images using complex optical equipment and are combined with real-time image processing to present an augmented view to the surgeon. The information is presented on a standard monitor above the operating bed, which requires the physician to stop the surgical procedure and look up at the monitor. The break in the surgical flow sometimes outweighs the benefits of fluorescence based IGS, especially in time-critical surgical situations. Furthermore, these instruments tend to be very bulky and have a large foot print, which significantly complicates their adoption in an already crowded operating room.

In this document, I present the development of a compact and wearable goggle system capable of real-time sensing of both NIR fluorescence and color information. The imaging system is inspired by the ommatidia of the monarch butterfly, in which pixelated spectral filters are integrated with light sensitive elements. The pixelated spectral filters are fabricated via a carefully optimized nanofabrication procedure and integrated with a CMOS imaging array. The entire imaging system has been optimized for high signal-to-background fluorescence imaging using an analytical approach, and the efficacy of the system has been experimentally verified. The bio-inspired spectral imaging sensor is integrated with an FPGA for compact and real-time signal processing and a wearable goggle for easy integration in the operating room. The complete imaging system is undergoing clinical trials at Washington University in the St. Louis Medical School for imaging sentinel lymph nodes in both breast cancer patients and melanoma patients.

Language

English (en)

Chair

Viktor Gruev

Committee Members

Mark Anastasio, Roger Chamberlain, Yasutaka Furukawa, Richard Loomis, Robert Pless

Comments

Permanent URL: https://doi.org/10.7936/K7RR1WFM

Included in

Engineering Commons

Share

COinS