Targeted prostate biopsy using mathematical optimization

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Prostate cancer is the second leading cause of death for American men. However, there is currently no imaging modality that can reliably detect cancer in the majority of cases. Therefore, needle biopsy of the prostate has been widely used as a gold standard for the diagnosis and staging of prostate cancer, when elevated prostate specific antigen (PSA) levels are measured. Following the widespread use of sextant biopsy, several enhanced random systematic biopsy methods have been adopted by different groups in an effort to reduce the significant number of cases remaining undetected at initial biopsy, mainly by using additional needles, albeit with limited success. The need to more thoroughly understand the performance of all these random systematic sampling methods has led to several computer simulation studies that utilize whole-mounted histological stained sections from prostatectomy specimens in order to estimate the performance of different biopsy approaches. However, to date there has been no mathematically rigorous attempt to precisely determine where the needles should be placed in order to maximize probability of cancer detection. The overall goal of this project is to develop and clinically test a computer-based methodology for optimal sampling of the prostate during biopsy, so that the probability of cancer detection is maximal, based on statistics obtained by applying an advanced image analysis methodology to whole-mounted sections of radical prostatectomy specimens. We thus propose to develop and clinically test a targeted prostate biopsy method. By this we mean that the exact spatial locations of biopsy sites will be determined using mathematical optimization methods, rather than approximate biopsy locations being defined in terms of a rough subdivision of the prostate, which is the current practice. Histological sections from 281 diverse specimens will be used to determine the spatial statistics of prostate cancer. A methodology for elastic shape transformation will be developed and used to accurately overlay images from different specimens, by removing inter-individual morphological variability. Optimal biopsy sites will be determined by mathematical optimization, i.e., by finding the needle coordinates that maximize probability of cancer detection, taking into consideration expected errors in needle placement. Deformable registration algorithms will be developed for overlaying the optimal biopsy sites on a patient's MR images. The optimal sampling method will be validated on an independent patient population, under precise intra-operative MR guidance.