A feasible roadmap for developing volumetric probability atlas of localized prostate cancer

TL;DR

This paper presents a statistical volumetric prostate cancer probability atlas derived from optical imaging data, enabling improved biopsy strategies through detailed distribution modeling.

Contribution

It introduces a novel non-rigid warping scheme for constructing a prostate cancer probability atlas from surgical specimens, enhancing biopsy accuracy.

Findings

Effective tumor distribution characterization

Improved biopsy sampling strategies

Validated through pilot studies

Abstract

A statistical volumetric model, showing the probability map of localized prostate cancer within the host anatomical structure, has been developed from 90 optically-imaged surgical specimens. This master model permits an accurate characterization of prostate cancer distribution patterns and an atlas-informed biopsy sampling strategy. The model is constructed by mapping individual prostate models onto a site model, together with localized tumors. An accurate multi-object non-rigid warping scheme is developed based on a mixture of principal-axis registrations. We report our evaluation and pilot studies on the effectiveness of the method and its application to optimizing needle biopsy strategies.