Using AI for Brain Tumour Treatment
Radiation therapy (RT) is a common treatment for brain tumours (the second most common cancer in childhood after leukemia). The goal is to deliver a focused dose of radiation to the tumour and surrounding at-risk regions while minimizing radiation to normal brain tissues. This is particularly crucial for young patients in whom excess RT to normal brain tissues can lead to significant long-term side effects.
The standard method of manual radiation therapy planning is time-consuming and can vary in quality depending on the experience of the radiation therapist and oncology staff. Artificial intelligence (AI)-assisted planning has been shown to improve the speed and quality of RT plan design and reduce unnecessary doses to normal brain tissues.
In a project involving the Princess Margaret Cancer Centre in Toronto, as well as the London Regional Cancer Program (Ontario), researchers evaluated the potential benefit of collaborative AI-RT planning for children, youth and young adults with brain tumours. The group intends to disseminate the cutting-edge AI technology and create new collaborations to improve RT for thousands of cancer patients across Canada.
For this study, 61 patients were enrolled, with approximately two RT plans created by expert RT planners (standard practice) and one AI-assisted plan created for each patient. In a blind evaluation by the treating oncologists, the AI-assisted plans were deemed to be as good or better than the manual plans. AI-assisted plans delivered a slightly smaller dose of radiation to normal brain tissues and were designed in less time.
The conclusion: AI-assisted radiation therapy planning creates high-quality radiotherapy plans for children and young adults with brain tumours. This is the first study of AI-assisted RT planning to include children, and creates an opportunity to rapidly create high quality plans for patients needing urgent treatment, and facilitate planning for smaller centres.
Read the full published article: A Prospective Study of Machine Learning − Assisted Radiation Therapy Planning for Patients Receiving 54 Gy to the Brain
Derek S. Tsang, MD, MSc; Grace Tsui, BSc, MRT(T); Anna T. Santiago, MSc, MPH; Harald Keller, PhD; Thomas Purdie, PhD, FAAPM; Chris Mcintosh, PhD; Glenn Bauman, MD; Nancy La Macchia, MRT(T); Amy Parent, BSc, MRT(T); Hitesh Dama, BSc, MRT(T); Sameera Ahmed, MSc; Normand Laperriere, MD; Barbara-Ann Millar, MBChB; Valerie Liu, BSc; David C. Hodgson, MD, MPH, FASTRO
Study Examines Late Effects and Treatment Protocols for Hodgkin Lymphoma in Children
Childhood cancer and its treatment come at a risk of late effects for survivors. In fact, the evidence shows that two of every three childhood cancer survivors are at increased risk of at least one late effect due to their cancer or its treatment, including heart disease, second cancers and cognitive challenges.
A recent study published in JAMA (The Journal of the American Medical Association) looks at the treatment of Hodgkin lymphoma in children and the resulting risk of cardiac disease to survivors.
POGO Medical Director, Dr. David Hodgson, is the lead author of the study titled Late Cardiac Toxic Effects and Treatment Protocols for Hodgkin Lymphoma in Children.
“Cardiac disease as a late effect of treating Hodgkin lymphoma is a serious risk that is well documented,” he says. “One of the challenges of improving treatment is that we don’t want to wait for over 20 years to observe the late effects of treatments we’re currently giving today. The goal of this study was to estimate the risk of late effects of current treatment to better understand how to advise current patients and families, and also explore what changes to treatment would have the greatest benefit to reduce these risks.”
Understanding the extent to which treatment can be modified for improved health outcomes is critical for several reasons.
- Reducing treatment intensity is a trade-off against the risk of relapse.
- Understanding which components of treatment can be adjusted, and to what extent, can lead to better decisions about treatment modifications.
- Care for survivors is based on their treatment exposure making the burden of follow-up care an important consideration.
The study evaluated the treatment of 2,563 patients with Hodgkin lymphoma treated in four consecutive Children’s Oncology Group clinical trials between 2002 and 2022. Patients were treated with different doses of radiation therapy and doxorubicin chemotherapy – both of which are known to increase the risk of cardiac late effects. Based on the doses of these treatments, the 30-year cumulative incidence of severe or life-threatening heart disease was estimated to decrease from 10% in the first trial to 6% in the last trial. The findings were favourable, suggesting that evolutions in treating Hodgkin lymphoma will lead to a net reduction in late cardiac disease. Moreover, findings suggest that for adolescent and young adult patients, increasing the use of “cardioprotective” medications during treatment would be one of the best ways to reduce this risk further, without compromising cure rates.
“While there is still work to be done to monitor the long-term gains in reducing cardiac toxic effects, this is encouraging news that bodes well for the quality of life of childhood cancer survivors,” says Dr. Hodgson. “And, as with every study, we need to ensure that the guidelines for follow-up care of survivors, particularly in POGO AfterCare Clinics, are updated to reflect the latest evidence.”
Read the study here.
Related Content
Identifying Increased Risk for a Cancer Predisposition Syndrome
This POGO-funded research is helping children with its early detection of cancer predisposition syndromes.
Approximately 10% of children diagnosed with cancer have an underlying cancer predisposition
syndrome (syndrome), a condition (often inherited) that increases their risk of developing one or many cancers throughout their life. Genetic testing can be used to identify children with a syndrome, but to avoid unnecessary testing and possible delays, physicians need a way to determine which children to test.
There are over 125 syndromes known to be associated with cancer in children. In some cases, knowing whether a child has a specific syndrome will change the treatment. Identifying a syndrome also allows the healthcare team to develop a care plan to monitor the child for future cancers once treatment is complete, and to recommend monitoring for siblings.
In 2017, Dr. Catherine Goudie was awarded a POGO fellowship for her project “Identifying Children at Increased Risk for a Cancer Predisposition Syndrome: The McGill Interactive Pediatric Oncogenetic Guidelines.” Overseen by principal investigators Dr. David Malkin and Dr. William Foulkes, the project aims to develop an e-Health tool that doctors can use to help them decide which children should be referred for genetic testing.
Related Story : A Genetic Mutation: A Lifetime of Cancer Screening
Prior to becoming a POGO Fellow, Dr. Goudie and her team in Montreal spent three years building
approximately 90 algorithms for each type of childhood cancer. These yes/no questions relate to
items such as the age of the child, features of the tumour and family history. The first version of the
tumour algorithms was reviewed by genetic and oncology experts from Canada, the United States
and the United Kingdom.
The aim of the POGO-funded research was to support Dr. Goudie in her work to evaluate the performance of the tool with children at 11 Canadian hospitals. The evaluation was split into two streams.
“Our first priority was making sure this tool would do no harm,” says Dr. Goudie. “In order to do this, we looked at patients in Canada who have already been diagnosed with a cancer predisposition syndrome and who developed a cancer in childhood, meaning their physicians successfully identified the syndrome without our tool. We confirmed that our algorithms would have identified these children for genetic referrals at the time of their cancer diagnosis, therefore confirming that the tool performs at least as well as clinicians have.”
The second stream of the evaluation is ongoing and leverages precision medicine research at The Hospital for Sick Children and other hospitals in Canada and the US. These institutions are doing comprehensive genetic sequencing on pediatric cancer patients, meaning that all children undergo genetic evaluation regardless of whether they are suspected to have a syndrome.
“We will test our algorithm on the medical profiles of these children to determine if the tool indicates that the child should be referred for genetic assessment. We will then compare our results with those obtained from the patients who participated in the comprehensive genetic sequencing. Therefore, we are prospectively testing the performance of our tool.”
Dr. Goudie explains that, in most hospitals worldwide, the facilities, human resources, infrastructure, expertise, or funding to offer comprehensive genetic testing in all kids diagnosed with cancer are not available. Ultimately, the goal of the tool is to give children with a syndrome the opportunity to be diagnosed and treated appropriately. In addition, the tool can be used to educate medical professionals around the world.
Dr. Catherine Goudie received a prestigious clinical investigator award from the FRQ-S in Quebec and returned to Montreal in July 2018 to continue this research as a collaboration between the Montreal Children’s Hospital and The Hospital for Sick Children. The FRQ-S is a government organization that, annually, funds select clinician investigators in Quebec to do health research.
“For someone young like me, who doesn’t have a lot of experience, having POGO to support my work was a huge stepping stone. I am certain that POGO’s support and the credibility that this organization has internationally, helped me get this next award in Quebec. I am really grateful for this opportunity.”
This story was featured in POGO's 2018 Community Impact Report.
Check out and download the MIPOGG App now.
