Co-registered photoacoustic and ultrasound tomographic imaging of human colorectal and ovarian cancer: light delivery, system development, and clinical study
Date of Award
Doctor of Philosophy (PhD)
Ovarian cancer remains the deadliest of all the gynecological malignancies. Conventional screening tests, including pelvic examination, transvaginal ultrasound (TVUS), and blood testing for cancer antigen 125 (CA-125), lack sufficient specificity for early ovarian cancer diagnosis. Imaging modalities such as computed tomography (CT), positron emission tomography (PET), and magnetic resonance imaging (MRI) have been used for surgical guidance. However, all of these modalities have limitations in detecting small lesions. Globally, colorectal cancer is the second most commonly diagnosed malignancy and the fourth most common cause of cancer mortality. Accurate staging and post-treatment surveillance of this prevalent disease are critical because treatment strategies are predicated upon the stage at presentation and the patient’s response to pre-surgical therapy – in some instances, detailed imaging allows certain patients to avoid surgery altogether. While colonoscopy and biopsy are the gold-standard diagnostic tests for colorectal cancers, multiple imaging modalities are also utilized, including optical imaging, endoscopic ultrasound (EUS), pelvic magnetic resonance imaging (MRI), computed tomography (CT), and positron emission tomography (PET). Unfortunately, each of these modalities has critical weakness in evaluating colorectal tumors. In staging colorectal tumors and evaluating their therapeutic response, more precise imaging modalities could transform the standard of care. This dissertation explores co-registered photoacoustic and ultrasound tomographic imaging of two fatal cancers: ovarian cancer and colorectal cancer. It is composed of three main parts: light delivery optimization/fiber optics, system development and optimization, and pilot patient and sample study. To begin, we designed, optimized, and evaluated an hand-held photoacoustic and ultrasound probe suitable for endo-cavity subsurface tumor imaging. Compared to previous designs, the prototype probe, consisting of four 1 mm multi-mode optical fibers attached to 1.5 mm diameter ball-shaped fiber tips sandwiched between a transvaginal ultrasound transducer and a custom-made sheath, demonstrated a higher light output and better beam homogeneity on the tissue subsurface. Second, we developed a novel fiber diffuser tip using microspheres dispersed within an ultraviolet (UV) adhesive to scatter light. This diffuser keeps the skin surface fluence under the maximum permissible exposure (MPE), while enabling higher laser energy injection to enhance the photoacoustic (PA) signal generated from the tissue. Third, we proposed an improved beamformer, named lag-based delay multiply and sum combined with coherence factor (DMAS-LAG-CF). Simulations and phantom experiments demonstrate that compared with conventional delay and sum (DAS), the proposed algorithm can provide 1.39 times better resolution and 10.77 dB higher contrast. For patient data, similar improvements in contrast ratios have been observed. However, since diagnostic accuracy in distinguishing between cancer and benign/normal groups is the significant measure, we have extracted the photoacoustic histogram features of mean, kurtosis, and skewness. When mean and skewness are used as features, DMAS-LAG-CF can improve cancer diagnosis, with an AUC of 0.91 in differentiating malignant from benign ovarian lesions. Fourth, to investigate the ability of co-registered photoacoustic and ultrasound tomographic imaging to assess human colorectal cancer, we conducted a pilot study on 23 ex-vivo human colorectal tissue samples that were collected immediately after surgical resection. Co-registered photoacoustic images of malignancies showed significantly increased PAT signals compared to normal regions of the same sample. We found statistically significant differences between untreated colorectal tumors and normal tissues, based on the quantitative relative total hemoglobin concentration (rHbT) computed from four optical wavelengths, spectral features such as the mean spectral slope and 0.5 MHz intercept extracted from PAT and US spectral data, and image features such as the first and second order statistics along with the standard deviation of the mean radon transform of PAT images. Using either a logistic regression model or a support vector machine, the best set of parameters of rHbT and PAT intercept achieved AUC values of 0.97 and 0.95 for the training and testing data sets, respectively, in predicting histologically confirmed invasive carcinoma. One limitation of the current system is its poor image resolution (~ 250 μm axial resolution) limited by the commercial endo-cavity ultrasound transducer array (6 MHz central frequency, 80% bandwidth). For better image resolution in high frequency PAT/US imaging, we first decoded the pin configuration of a high-frequency transducer array (15 MHz central frequency, 9-18 MHz bandwidth) and adapted it to our home-made 128 channel ultrasound pulsing and receiving system (sampling rate: 40 MHz). To further improve the performance, we are building a 64-channel FPGA-based high frequency photoacoustic imaging system with a sampling rate of 80 MHz and signal-to-noise ratio (SNR) of 40 dB. For in-vivo patient study, this system will be integrated with an endo-rectal probe with a side-firing fiber tip. Fifth, we investigated the laser safety of photoacoustic imaging, in preparation for its use in clinical reproductive medicine. Using preimplantation mouse blastocyst stage embryos, we assessed potential DNA damage from photoacoustic laser exposure. Different embryo groups were exposed to either 5- or 10- minute 15-Hz laser doses (typical clinical doses), or a 1-minute 1-kHz laser dose (a significantly higher dose). We demonstrated that typical lasers and exposure times used for photoacoustic imaging do not induce increased cell death in mouse blastocysts. Sixth, we demonstrated a novel fiber endface photoacoustic generator using IR 144 laser dye dispersed within an ultraviolet (UV) adhesive. The generator provides wide acoustic bandwidth in the transducer frequency range of 2-7 MHz, high thermal conversion efficiency (> 90%), good PA intensity controllability (via the easily controlled IR 144 concentration), and high feasibility (simple procedures). Through a series of experimental validations, we show this fiber-based endface photoacoustic generator can be a useful tool for a broad range of biomedical applications, such as calibrating the local absorption coefficient of biological tissue for quantitative photoacoustic tomography.
Samuel Achilefu, Song Hu, Joan Riley, Chao Zhou,