Toward Improving Treatment Planning Quality and Efficiency Using Knowledge Engineering and Autoplanning: A Study Based on NRG-HN001 Clinical Trial

A comprehensive evaluation of the quality and complexity of prostate IMRT and VMAT plans generated by an automated inverse planning tool

Introduction:

The Pinnacle3 Auto-Planning (AP) package is an automated inverse planning tool employing a multi-sequence optimisation algorithm. The nature of the optimisation aims to improve the overall quality of radiotherapy plans but at the same time may produce higher modulation, increasing plan complexity and challenging linear accelerator delivery capability.

Methods and materials:

Thirty patients previously treated with intensity-modulated radiotherapy (IMRT) to the prostate with or without pelvic lymph node irradiation were replanned with locally developed AP techniques for step-and-shoot IMRT (AP-IMRT) and volumetric-modulated arc therapy (AP-VMAT). Each case was also planned with VMAT using conventional inverse planning. The patient cohort was separated into two groups, those with a single primary target volume (PTV) and those with dual PTVs of differing prescription dose levels. Plan complexity was assessed using the modulation complexity score.

Results:

Plans produced with AP provided equivalent or better dose coverage to target volumes whilst effectively reducing organ at risk (OAR) doses. For IMRT plans, the use of AP resulted in a mean reduction in bladder V50Gy by 4·2 and 4·7 % (p ≤ 0·01) and V40Gy by 4·8 and 11·3 % (p < 0·01) in the single and dual dose level cohorts, respectively. For the rectum, V70Gy, V60Gy and V40Gy were all reduced in the dual dose level AP-VMAT plans by an average of 2·0, 2·7 and 7·3 % (p < 0·01), respectively. A small increase in plan complexity was observed only in dual dose level AP plans.

Findings:

The automated nature of AP led to high quality treatment plans with improvement in OAR sparing and minimised the variation in achievable dose planning metrics when compared to the conventional inverse planning approach.