Date of Completion
Field of Study
Doctor of Philosophy
Ovarian cancer is the fifth most common cancer among women and has the highest mortality rate of all gynecologic cancers. Current clinical imaging modalities are limited by poor sensitivity and specificity. Therefore, there is an urgent need to develop effective tools to detect ovarian cancer. In this dissertation, two imaging modalities, optical coherence tomography (OCT) and photoacoustic imaging have been investigated for ovarian cancer detection and characterization.
In the first modality, optical scattering coefficient, phase retardation and phase retardation rate were quantitatively extracted from polarization-sensitive OCT (PS-OCT) images. A highly positive correlation was found between those three parameters and collagen content, which is an indicator of ovarian tissue malignancy. Malignant ovarian tissue showed statistically significant lower scattering and birefringence property than normal ovarian tissue. A three-parameter logistic model was developed to diagnose ovaries as malignant or normal. The extracted parameters from 33 ovaries were used as input predictors to train the logistic model, and 10 additional ovaries were tested using this trained model. 100% sensitivity and specificity were achieved in the training group; 100% sensitivity and 83.3% specificity were achieved in the testing group. In the second modality, a laser pulse stretching scheme is designed to improve laser energy delivery in photoacoustic imaging in order to improve in vivo ovarian cancer detection based on optical fiber illumination. The effects of pulse width on photoacoustic detection using different ultrasound transducers were systematically investigated by simulations and experiments. In addition, an optical-resolution photoacoustic microscopy (PAM) was developed to map microvasculature networks in ovarian tissue. The feasibility of PAM to differentiate malignant from normal ovaries was explored by comparing PAM images morphologically. PAM images of both normal and malignant ovarian tissue match the histology. Based on the observed differences between PAM images of normal and malignant ovarian tissue in microvasculature features and distributions, seven parameters were quantitatively extracted and applied to a logistic model for ovarian tissue diagnosis. A specificity of 81.3% and a sensitivity of 88.2% were achieved. Those results have demonstrated the great potential of OCT and photoacoustic imaging for clinical ovarian cancer detection.
Wang, Tianheng, "Optical and Photoacoustic Imaging for Ovarian Cancer Detection" (2014). Doctoral Dissertations. 357.