A flexible framework for semi-automatic inverse planning of abdominal organ tumor ablation

Abstract

Globally, the fourth most common cause of cancer related death is hepatocellular carcinoma (HCC). The first line treatments for HCC are surgical, including the gold standard treatments, liver transplantation and surgical lobectomy. However, both treatments are invasive and require long recovery times. Access to liver transplantation is further restricted by the limited availability of transplantable organs, and roughly 80% of patients are ineligible for either surgical treatment due to reduced liver function or other comorbidities. For this large group of patients, the best remaining options are locoregional therapies, most commonly radiofrequency or microwave ablation.

Procedure planning for ablations of small tumors can be accomplished using the standard tools included with many DICOM viewers. However, to treat larger tumors multi-needle treatment plans are becoming more common. Planning for multi-needle procedures is difficult due to challenges visualizing thermal dose regions while simultaneously considering anatomical restrictions on needle placement. Therefore, planning for these procedures is subject to inter- clinician variability and is dependent on the treating clinician’s experience with the procedure.

This work develops a library called ABPlan (the Ablation Planner) to support clinicians in the planning of ablative treatments. The library is designed to support semi-automatic inverse needle planning, where the clinician defines the treatment region, and the software provides assistance in selecting trajectories to treat this region. ABPlan is focused is on being efficient, and flexible enough to support a variety of planning scenarios and future improvements. This document presents the architecture of ABPlan, followed by an example integration of the library into the 3D Slicer platform for medical image computing.