Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article


Doctor of Philosophy


Medical Biophysics


Jeffrey Carson


Breast cancer is one of the most common cancers to affect women, presenting a lifetime risk of 1 in 8. Treatment of stage 1 and 2 cancers usually involves breast conserving surgery (BCS). The goal of BCS is to remove the entire tumour with a surrounding envelope of healthy tissue, referred to as a negative margin. Unfortunately, up to 50% of surgeries fail to remove the whole tumour. To minimize the risk of cancer recurrence, a second surgery, must therefore be performed. Currently, there is no widely accepted intraoperative tool to significantly mitigate this problem. Employed systems are usually based on imaging, such as x-ray or ultrasonography. Unfortunately, sensitivity and specificity deficits, especially related to breast density, reduce the effectiveness of these methods. Photoacoustic tomography (PAT) is a relatively new imaging modality which uses safe near-infrared laser illumination to generate 3-D images of soft tissues to a depth of up to several cm. We used a custom designed and built intraoperative PAT system, called iPAT, to perform a 100 patient study on freshly excised breast lumpectomy specimens within the surgical setting. The system enabled the evaluation of tumour extent, shape, morphology and position within lumpectomy specimens measuring up to 11 cm in diameter. Scan results were used to compare iPAT-derived tumour size to the gold-standard pathologic examination, and when available, to x-ray, ultrasonography and DCE-MRI. Imaging results were also used to classify specimen margins as close or wide, and positive predictive values (PPV), negative predictive values (NPV), sensitivity and specificity were then calculated to estimate the effectiveness of the iPAT system at predicting lumpectomy margin status. With a close margin prevalence of 35%, the PPV, NPV, sensitivity and specificity of iPAT were found to be 71%, 83%, 69%, and 84%, respectively. Information provided by the iPAT system identified 9 out of the 12 positive specimens, potentially reducing the positive margin rate by 75%. . Contrary to expected photoacoustic contrast mechanisms, iPAT images of hemoglobin distribution correlated poorly with US and X-ray tumour imaging, while hypo-intense regions in lipid-weighted iPAT images were in excellent agreement.