OpenIGTLink: a network communication interface for closed-loop image-guided interventions

点击翻译按钮获取中文摘要

Project Summary/Abstract In this cycle of the project, we will extend OpenIGTLink software to enable individualized precision guidance, where patient-specific characteristics are taken into account in the planning and control process for image- guided robotic interventions (IGRI) to achieve a better clinical outcome. In particular, we will focus on improving needle placement accuracy for IGRI of the prostate, as a simple, and yet a representative application of individualized precision guidance. Percutaneous needle placement plays a fundamental role in both the diagnosis and treatment of prostate cancer. Nearly one million prostate biopsies are performed annually in the United States, where tissues are sampled from the prostate with a biopsy needle for pathological examination. The confirmed lesions may then be focally treated with brachytherapy or thermal ablation using applicator needles depending on clinical indications and patient preference. For those procedures, accurate needle placement is essential to avoid false-negatives or ensure an optimal dose distribution. The demand for accurate needle placement is growing in recent years, as clinicians are now capable of pinpointing subregions in a heterogeneous tumor thanks to the recent progress in imaging technologies, particularly multiparametric magnetic resonance imaging (MRI). However, the in vivo accuracies achieved by previous studies have never matched the level of tumor subregion targeting due to needle deviation as a result of physical interaction between the needle and the heterogeneous tissue. To address this issue, we will incorporate needle deviation models in an IGRI system using OpenIGTLink software. Our hypothesis is that the models adapted to individual patients will enable avoidance or compensation of needle deviation, leading to an improved in vivo needle placement accuracy. The project comprises the following four specific aims: (Aim 1) Develop a biomechanical model with image-based semi-automatic model generation; (Aim 2) Develop a data-driven prediction model and hybridize with the biomechanical model; (Aim 3) Extend the IGRI platform for model- based closed-loop control using OpenIGTLink; (Aim 4) Evaluate biomechanical, data-driven, and hybrid models in vivo needle placement.