Highlights of Adaptive Radiation Therapy and Improved Image Guidance for Head and Neck Cancer at the 2020 ASTRO Annual Meeting

By Molly McCulloch, PhD, and Kristy Brock, PhD, DABR, FAAPM

Presenting author:
Hans Paul van der Laan, PhD

A popular topic for this year’s ASTRO Annual Meeting is adaptive radiation therapy (ART) and improved image guidance for head and neck cancer (HNC). Many groups have studied and reported results about normal tissue complication probability (NTCP) and tumor control probability (TCP) models for use in plan optimization. One study in particular, Quality of Life and Toxicity Based Treatment Plan Optimization for Head and Neck Cancer (van der Laan et al., University of Groningen), aimed to compare organ at risk (OAR) doses, NTCPs and quality of life (QOL) between traditional OAR-based optimization and QOL-weighted NTCP-based optimization. Findings showed that QOL-weighted NTCP-based VMAT treatment plan optimization resulted in significant reductions in NTCP for dysphagia, aspiration and hoarseness, but an increase in the NTCP for xerostomia, overall, was expected to result in improved QOL. As the outcomes of this study indicate that QOL-weighted NTCP-based VMAT treatment plan optimization is feasible clinically, this presentation demonstrated how to further advance automated treatment planning using tools already employed in most clinics, with the anticipated outcome of improving to the QOL of patients.

Presenting author:
Mariluz De Ornelas, PhD

A similar study, Application of Tumor Control (TCP) and Normal Tissue-Complication Probabilities (NTCP) to Determine the Best Robust Optimization (RO) Approaches for Proton Head and Neck Radiotherapy, (Ornelas et al., University of Miami Sylvester Comprehensive Cancer Center) considered TCP and NTCP models to be used as relative metrics to determine the most robust optimization approach for proton head and neck (HN) radiotherapy (RT), to improve clinical outcomes (TCP and NTCP). This study found that uncoupled robust optimization is preferred for HN treatment planning, as high HU streaking from metal implants is the most common artifact for HN proton planning and potentially lead to overestimation of stopping powers. This work presents the ideal robust optimization (RO) approach for HN treatment planning for intensity-modulated proton therapy (IMPT), based on common artifacts. Previous to this study, no clinical consensus existed for the optimal RO approach. With the application of these research findings, clinicians now have grounds for choosing an RO approach for IMPT HN planning.

Presenting author:
Arthur Lalonde, PhD

Another report, Anatomic Changes in Head and Neck Intensity-modulated Proton Therapy: Comparison between Robust Optimization and Daily Adaptation, (Lalonde et al., Massachusetts General Hospital and Harvard Medical School) also considered proton HN RT, evaluating which strategy ― increasing plan robustness a priori or adapting the plan throughout delivery ― could achieve the highest treatment quality in HN IMPT. Results indicated that daily online adaptive IMPT, based on CBCT, performed better than anatomical RO and classical RO in providing target coverage and lower integral dose for HN. Namely, the proposed online adaptive workflow reduces the mean dose to the parotid glands and oral cavity, compared to the RO approaches. This research implies that the proposed online adaptive IMPT workflow may be superior to the previously validated RO approaches for the treatment of HNC. 

Presenting author:
Baher Elgohary, MD, MS, MBBS

In addition to research investigating optimal planning and adaptation strategies, researchers investigated imaging biomarkers to predict and identify OAR toxicity risk, such as correlating xerostomia risk with imaging biomarkers. In particular, one team, in Mid-treatment Apparent Diffusion Coefficient Predicts Late Xerostomia following Head and Neck Cancer Radiotherapy (Elgohary et al., The University of Texas MD Anderson Cancer Center), investigated if pre-, mid-, and post-apparent diffusion coefficient (ADC) imaging biomarkers from the contralateral parotid gland can predict xerostomia six months post-RT. They found that mid-RT mean ADC and volume change post-RT from the contralateral parotid gland were significantly associated with xerostomia six months post-RT. This work demonstrates the possibilities of diffusion-weighted magnetic resonance imaging (DW-MRI) in the clinic. With the clinical deployment of DW-MRI software, ADC can be acquired for individual patients and potentially aid in the prediction of post-treatment xerostomia for HN RT patients.

Presenting author:
Seng (Gary) Lim, PhD

A second team, in Comparison of Patient Reported Xerostomia Risk and the Fluence-Based Decision Support Metric (Lim et al., Memorial Sloan Kettering Cancer Center), focused on xerostomia prediction, hypothesizing that the photon fluence exiting the patient during RT (transit fluence) is correlated with xerostomia risk. They investigated the use of photon exit fluence as a replanning trigger decision support metric and evaluated this hypothesis by comparing the change in transit fluence and patient reported QOL surveys in 21 HNC patients. This work found that there was indeed a significant correlation between this metric, which can be easily measured in the clinic, and the patient’s risk of xerostomia. Implementing the use of transit fluence as an aid for clinicians selecting patients for ART can potentially improve QOL for patients while reducing the resources needed for ART.

And finally, a particularly interesting conclusion was derived from a team in Evaluation of Metric Factors for Initiation of Adaptive Radiotherapy (ART) in Head and Neck IMRT (Rachi et al., National Cancer Center Hospital East, Japan) that aimed to develop a replanning metric based on the dose changes of the target, spinal cord and brainstem for ART using replanning CT obtained during treatment. They found a replanning metric for ART using clinical target volume (CTV) based on deformable image registration (DIR) with a Dice coefficient ≤ 0.8 and Hausdorff distance ≥ 10 mm. This research demonstrates the use of DIR results to aid in the decision to use ART, based on changes in the CTV. A threshold model is provided to use as a metric to indicate the potential need to replan a HN RT patient.

This collection of abstracts highlights the exciting advancements in adaptive radiation therapy and image guidance for head and neck cancer. The methods investigated have the potential to be implemented in the clinic using common tools and software and may improve the quality of life of patients.

Quality of Life and Toxicity Based Treatment Plan Optimization for Head and Neck Cancer was released onDemand on Wednesday, October 28 in the Science Center, as part of Scientific session (SS) 33.
Presenting author: Hans Paul van der Laan, PhD

Application of Tumor Control (TCP) and Normal Tissue-complication Probabilities (NTCP) to Determine the Best Robust Optimization (RO) Approaches for Proton Head and Neck Radiotherapy was presented on Sunday, October 25 in the Poster Hall, as part of Poster Viewing (PV) session 02.
Presenting author: Mariluz De Ornelas, PhD

Anatomic Changes in Head and Neck Intensity-modulated Proton Therapy: Comparison between Robust Optimization and Daily Adaptation was released onDemand on Tuesday, October 27 in the Science Center, as part of Scientific session (SS) 27.
Presenting author: Arthur Lalonde, PhD

Mid-Treatment Apparent Diffusion Coefficient Predicts Late Xerostomia following Head and Neck Cancer Radiotherapy was released onDemand on Monday, October 26 in the Science Center, as part of Scientific session (SS) 21.
Presenting author: Baher Elgohary, MD, MS, MBBS

Comparison of Patient Reported Xerostomia Risk and the Fluence-based Decision Support Metric was released onDemand on Tuesday, October 27 in the Science Center, as part of Scientific session (SS) 27.
Presenting author: Seng (Gary) Lim, PhD

Evaluation of Metric Factors for Initiation of Adaptive Radiotherapy (ART) in Head and Neck IMRT was presented on Sunday, October 25 in the Poster Hall, as part of Poster Viewing (PV) session 02.
Presenting author: Toshiya Rachi, MS

Published on: October 29, 2020