# Comparative evaluation of normal tissue objective functions in robotic radiosurgery planning for solitary brain tumors

## TL;DR

This study compares two planning methods for robotic radiosurgery in brain tumor treatment, finding that one reduces healthy tissue dose exposure without increasing treatment time.

## Contribution

The paper introduces and evaluates a novel normal tissue objective function as an alternative to traditional Auto-shells in CyberKnife radiosurgery planning.

## Key findings

- NTO reduced healthy brain tissue dose exposure (V12Gy and V5Gy) by up to 14% compared to Auto-shells.
- For meningiomas and metastases, NTO reduced the gradient index (GI) by up to 7%.
- Conformity (nCI) remained comparable between the two methods, within 2%.

## Abstract

The normal tissue objective (NTO) is an inverse planning approach in radiosurgery, also available for the CyberKnife system. By employing a model function, it aims to achieve precise control over the global dose fall‐off in healthy tissue. As a novel technique, NTO can serve as an alternative to the established method, which utilizes layered contours around the target to shape dose gradients and enhance conformity, referred to as Auto‐shells in CyberKnife systems.

This study compares the dose distribution achieved with NTO and Auto‐shells to evaluate their respective advantages in CyberKnife treatment planning.

A total of 45 patients with brain tumors—including 15 vestibular schwannomas, 15 meningiomas, and 15 metastases, all of whom had previously been treated using an Auto‐shells‐generated plan, were analyzed. For each case, an alternative NTO‐based plan was generated and compared with its Auto‐shells counterpart. Key treatment parameters—including nodes, beams, total monitor units (MU), treatment time, new conformity index (nCI), gradient index (GI), and dose exposure volumes to healthy brain tissue (V12Gy and V5Gy)—were evaluated.

Both methods resulted in comparable plans across many indices. Significant differences were particularly in terms of healthy brain tissue dose exposure. With the NTO method, V12Gy and V5Gy were reduced by up to 14%, and in the case of meningiomas and metastases, the GI was reduced by up to 7%. The conformity, described by the nCI, was within 2%. No significant difference was observed in MU.

NTO optimization presents a viable option to the Auto‐shells method for CyberKnife treatment of brain tumors. By reducing healthy brain tissue exposure without increasing monitor units, it enhances dose‐sparing efficiency. However, maintaining optimal conformity remains an important issue, highlighting the trade‐offs between precision and tissue preservation.