RapidPlan™ Knowledge-Based Planning: Clinical Adoption Takes Off | Varian

{ "pageType": "blog-article", "title": "RapidPlan™ Knowledge-Based Planning: Clinical Adoption Takes Off", "articleDate": "7. June 2017", "category": "Centerline", "imageId": "https://varian.widen.net/content/vl311grk0z/webp" }
RapidPlan™ Knowledge-Based Planning: Clinical Adoption Takes Off

RapidPlan™ Knowledge-Based Planning: Clinical Adoption Takes Off

RapidPlan™ knowledge-based planning is helping clinicians at centers around the world achieve greater consistency, efficiency, and quality in radiotherapy treatment planning. RapidPlan streamlines the planning process by using shared clinical knowledge to provide estimated dose volume histograms that can be used as a guideline and starting point. Planners can use shared models or create their own to reflect preferred treatment methodologies and protocols. This can reduce or even eliminate the need for multiple, time-consuming iterations when producing treatment plans.

“Delivering high-quality treatments with advanced techniques like volumetric modulated arc therapy (VMAT) or image-guided radiotherapy can be a challenge,” says Catherine Khamphan, PhD, medical physicist with the Institut Sainte Catherine in Avignon, France. “RapidPlan knowledge-based planning is a powerful tool that aggregates information from existing treatment plans to estimate an optimized and achievable dose distribution for new patients. Carefully implemented, it can improve treatment quality and efficiency. The time savings are significant.”

RapidPlan also helps to bridge the experience gap. “It enables more homogeneous plan quality across different planners,” Khamphan says, adding that without such a tool, plans created manually might not do as well sparing organs-at-risk (OAR) or covering the target. “RapidPlan helps us avoid the use of such suboptimal plans.”

French Team Develops RapidPlan Models for Treating Head & Neck and Prostate Cancer

Khamphan and her colleague, medical physicist Veronique Bodez, PhD, initially focused their efforts on creating RapidPlan models for two disease sites: head & neck and prostate.

“We had in-house guidelines that we were using; but we still found inhomogeneities in our datasets, such as structure definitions and clinical trade-offs, especially for hear & neck treatments,” said Khamphan. “RapidPlan helped us to correct this. We quickly found that implementing RapidPlan amounted to a ‘quality assurance’ for our practice.”

Khamphan and her colleagues measured the efficacy of their RapidPlan models by looking at the number of optimizations needed to reach clinical goals. They discovered that, with RapidPlan, one optimization was often enough, while a planner working manually sometimes must generate five or more.

“We found that 85 percent of the hear & neck plans optimized with RapidPlan were clinically acceptable after the first attempt. “Plan quality is at least equal or improved, and the OAR sparing is significant in most cases. Also, we believe we can improve on those rates as we improve our models,” Khamphan said.

So far, about 170 patients have received treatments planned using RapidPlan at Institut Sainte Catherine.

“We see that our patients have benefited from the overall improvement in plan quality: better OAR sparing and better dose homogeneity in the target for most cases,” Khamphan said. “We have a few striking examples of plans where RapidPlan increased the target coverage while sparing OAR beyond what the planner achieved. Our physicians are impressed by both the improvement in plan quality and also by the efficiency. Indeed, a treatment can be quickly re-planned if the physician observes anatomical changes, for example, tumor size or weight loss.”

According to Robin Garcia, PhD, head of the Medical Physics Department, some dosimetrists in her group initially expressed concerns about how RapidPlan might impact their work. “In fact, we have found that the time saved with RapidPlan can be dedicated to more complex cases and it can help reduce the pressure on the planning staff,” she said. “As we decrease the amount of effort spent on planning, we will be able to make a time investment in other areas, like developing adaptive procedures.”

Developing Rectal, Cervical and Lung Cancer Models in Beijing

Yibao Zhang, Ph.D., associate professor at Beijing Cancer Hospital in China, and his colleagues have worked together to develop RapidPlan models for rectal, cervical, and lung stereotactic body radiotherapy (SABR) treatments.

We started with planning for VMAT plus simultaneous infield boost for pre-surgical rectal patients,” said Zhang. “We have a large patient population with this disease, there were not yet other reports in the literature about RapidPlan applied to rectal cancer so it was an opportunity to publish, and this type of cancer generally involves relatively few OAR. Successful models have been reported for similar anatomical sites such as the prostate. We thought we would start by mastering this one and then move on to develop more models for other disease sites.”

Zhang worked with Hao Wu, M.Sc., chief physicist, deputy chair, and associate professor; and Fan Jiang, B.S., assistant professor, to create the rectal cancer model. The process involved taking plans that had been fine-tuned by a senior physicist and choosing only the most optimal plans to make sure excellence was incorporated into the model.

“Our rectal model is now very mature after much validation and fine-tuning,” Zhang said. “It produces better-than-manual plans in terms of quality, consistency, and efficiency.”

During validation, the Beijing-based team chose to evaluate over 200 RapidPlan plans to assess and improve the model, and then they applied it clinically. “Since September 2016, over 200 patients were treated with plans created semi-automatically using RapidPlan. We saw gains in both productivity and quality,” Zhang said. “Patients benefit directly from dosimetric improvement including lower OAR exposure and better target dose conformity and homogeneity, without extra cost.”1

Michigan Researchers Create Spinal SBRT RapidPlan Model

Martha Matuszak, PhD, medical physicist and associate professor at the University of Michigan (UM) Medical School in Ann Arbor, defines two major challenges when it comes to producing quality treatment plans for radiotherapy. “First, there is the challenge of defining what to treat to give the patient the best possible outcome,” she said, pointing out that current research on imaging, segmentation, big data, and outcomes is addressing that question.

“Once you have defined what you need to treat, you are challenged with creating the best possible plan for that patient,” she continued. “Knowing when you have achieved the best plan can be difficult, especially when making tradeoffs between the best plan possible as opposed to an acceptable plan given time and resource considerations.”

Matuszak, who is also the director of advanced treatment planning in the UM Department of Radiation Oncology, says that she was skeptical when she first heard about RapidPlan. “I thought it might be adequate to create simple plans for common body sites but would have a lot of limitations in complex geometries. After seeing the software perform and creating our first in-house models, I am pleased with how well it works.”

Funded by a grant from Varian, Matuszak headed a RapidPlan research project that assessed modeling for the treatment of spinal cancer with stereotactic body radiation therapy (SBRT). In the resulting journal article, the research team concludes: “RapidPlan is a robust technique that can improve planning efficiency in spine SBRT while maintaining or potentially improving plan quality and standardization across planners and centers.”2

Matuszak and her research group have also developed RapidPlan models for a number of other body sites and have now deployed several of these clinically, finding that automation is possible in some cases while others require a little tweaking by the user.

"As part of model creation and validation, we have created hundreds of plans using a number of models,” Matuszak said. “Clinically, we are just getting started. I would estimate that RapidPlan has been used for some 50 clinical plans so far, focusing on spine SBRT VMAT, liver SBRT IMRT/VMAT, hear & neck VMAT, and prostate IMRT.

"Well-designed RapidPlan models are valuable to the treatment team in that they help planners achieve a high level of quality, and they can do this quickly," Matuszak said. “For patients, this means potentially better plans with fewer delays and perhaps the potential for an improved outcome.”

1Wu H et al.  A dosimetric evaluation of knowledge-based VMAT planning with simultaneous-integrated-boosting for rectal cancer patients.  J Appl Clin Med Phys 2016 17(6) 78-85.

2Foya JJ et al. An analysis of knowledge-based planning for stereotactic body radiation therapy of the spine.  Practical Radiation Oncology

______________________________________________________________________________________

Note: The RapidPlan™ knowledge-based planning and its models are not intended to replace clinical decisions, provide medical advice, or endorse any particular radiation plan or treatment procedure. The patient’s medical professionals are solely responsible for and must rely on their professional clinical judgment when deciding to plan and provide radiation therapy. Also, all contributors to this article have received grants from Varian Medical Systems related to research on the clinical applications of RapidPlan™ and are either planning to or have published papers on this topic in peer-reviewed journals.