Paving the Way to Better Detection, Novel Treatments and Artificial Intelligence

A look back at the 2018 ASTRO Annual Meeting Presidential Symposium

By: Sanjay Aneja, MD; Michael Dominello, DO; Tim Lautenschlaeger, MD, PhD; Marc Mendonca, PhD; Stephen Shiao, MD, PhD, on behalf of the Promoting Science through Research and Training (PSRT) subcommittee of the ASTRO Science Council

Last October, nearly 11,000 radiation oncologists, residents, physicists, dosimetrists, nurses, industry representatives, other staff and students attended ASTRO’s 60th Annual Meeting. The Presidential Symposium centered around the overall meeting theme of “Translating Discovery to Cure,” with a decidedly forward-looking slate of topics. The Symposium was organized into four sessions, each with a unique theme. With planning well underway for this year’s Annual Meeting and a full slate of exciting research to be presented, we wanted to look back at the 2018 symposium and what the themes raised mean for the field.

Theme 1) Immunotherapy

Dr. Silvia Formenti opened the immunotherapy session with a nod to the past, reminding us that the critical role of the immune system in regulating the response to radiation was recognized more than 40 years ago and that many preclinical models later, we have finally begun to see those primordial seeds bear fruit. She described how early work with anti-CTLA-4 in murine models culminated in a successful human trial of anti-CTLA-4 (ipilimumab) and radiation in non-small cell lung cancer (Formenti et al. Nat Med 2018).

Surgical oncologist Dr. Jennifer Wargo told the incredible success story of immunotherapy in melanoma, particularly anti-PD-1 therapy, all while emphasizing the current challenges of employing immunotherapy including identifying biomarkers of response/toxicity and understanding immunotherapy resistance. Though biomarkers such as PD-L1 expression in tumors, intratumoral immune cell infiltrates and mutational burden have begun to emerge, these remain imperfect and imprecise. However, new biomarkers, namely the gut microbiome, were identified by her group and others as being the next frontier of both biomarker and therapeutic exploration (Gopalakrishan et al. Science 2018). Other therapeutic avenues being explored include combination therapies with immunotherapies, chemotherapy and radiation with toxicity being the key hurdle to overcome.

The final speaker, Dr. Zachary Morris, made the case for the potential of radiation as the perfect partner to immunotherapy. He showed preclinical data that demonstrated the multiple mechanisms by which radiation can interface with the immune system including reducing intratumoral immune suppression and activating inflammatory pathways to produce an anti-tumor immune response. This radiation-induced immune response could be augmented by combining it with immunotherapies, particularly checkpoint blockade (Twyman-St Victor et al. Nature 2015, Morris et al. Cancer Immunol Res 2018). The preclinical data strongly supports the notion that radiation and immunotherapy can synergize, but also highlights questions about how best to combine radiation and immunotherapy.

In sum, the potential for radiation and immunotherapy is vast, but we need to support and promote future studies to answer key issues including site selection, dose, fractionation and timing that will be critical to successfully incorporating radiation into immunotherapy paradigms.

Theme 2: Virally induced cancers

From hepatocellular carcinoma to anal cancer, to Merkel cell carcinoma, it is estimated that less than 10% of human cancers are virally induced. This statistic however does not begin to describe the morbidity and mortality of viral oncogenesis in terms of global impact. For example, cervical cancer represents the fourth most common cancer in women worldwide. Per the World Health Organization, of the estimated more than 270,000 deaths from cervical cancer every year, more than 85% of these occur in less developed regions. While we hope and expect that continued focus on vaccination, screening and decreasing HPV/HIV coinfection will improve these statistics, the current numbers are staggering.

Beyond recognizing this problem and its impact, the following questions were proposed:

  • Is it clinically important to confirm a viral association and perform viral typing?
  • What is the role of radiotherapy in the management of these disease processes and how is it modulated once there is a known viral association?
  • Do we understand and recognize the pitfalls and limitations in our knowledge in this arena?

It was suggested that while in the past we have treated malignancy as a function of location, how might this change when we consider a viral association? For example, should one treat a p16 positive nasopharyngeal cancer like a nasopharyngeal cancer or like any other HPV related head and neck cancer?

To grow as a field, suggestions moving forward included close observation and scrutiny of the data combined with a spirit of open-mindedness, communication and collaboration.

Theme 3: Artificial Intelligence

In addition to being a focus of the presidential symposium, Artificial intelligence (AI) was prominently featured among multiple educational sessions and abstracts within the newly established Digital Health and Informatics tract. Although AI has been an established field since the 1950s, there has been a renewed focus on the applications of AI over the last five years, primarily driven by three factors:

  • Increased digitization of healthcare data has provided a wealth of information which can be used to train AI algorithms to perform healthcare tasks.
  • Advances in computation have allowed more ease in storing and analyzing large amounts of heterogenous digitized healthcare data.
  • Advances in the methods in which we analyze data, specifically in the fields of machine learning and deep learning, have improved the ability of AI solutions to mimic human intelligence.

Dr. Kristy Brock from MD Anderson Cancer Center discussed the differences between AI, machine learning and deep learning, providing a background to AI for those who were unfamiliar. AI is a subfield of computer science which employs any rule-based learning. Machine learning is a sub-field with AI where algorithms attempt to learn from user-defined variables/features. Deep learning is a more nuanced form of AI where algorithms still learn from data, but do not require any user-defined variables and features.

Applications of AI within radiation oncology are already present in auto-segmentation programs, image reconstruction and image classification. We can expect even more AI driven solutions within radiation oncology which will attempt to improve outcome prediction and bottlenecks in practice workflows.

Theme 4: Liquid biopsies and cancer care

Blood-based circulating tumor DNA (ctDNA) testing is routinely used in patients with advanced solid tumors for mutation detection to establish eligibility for certain targeted therapies. Use of these tests can often avoid invasive tissue biopsies and its spatial heterogeneity issues. While the percentage of tumor DNA in the blood of patients with widely metastatic disease is often 10% or greater, tumor DNA is less reliably picked up in the blood of patients with early stage cancers using current detection approaches. However, as introduced by Dr. Park and expertly narrated by Drs. Rosenfeld and Diehn, there are many potential applications for ctDNA testing in patients with non-metastatic disease, including cancer screening, therapy response monitoring and recurrence monitoring.

As assay technology improves, some of these applications are expected to become feasible and ready for clinical evaluation. Efforts by the speakers’ laboratories have demonstrated approaches to create ever more sensitive ctDNA assays that soon might approach the levels of performance potentially useful for early stage patients. For example, exploiting the slight size differences of ctDNA versus other cell-free DNA in blood can significantly enrich for ctDNA, which in turn reduces sequencing cost and allows for deeper analysis.

Research demonstrated that ctDNA based minimal residual disease determination after completion of therapy is highly prognostic in many solid tumors, including nasopharyngeal, non-small cell lung, breast, pancreas and colon cancer. Presented unpublished data suggests that ctDNA assessment during radiation therapy might be of prognostic significance, which could be used for adaptive radiation therapy in the future.

Overall, liquid biopsy technology has great promise and new applications are likely to be developed for patients with non-metastatic disease in the near future, several of which might be of relevance to the radiation oncologist. Prospective studies are needed to establish clinical utility and more research is necessary to further develop technologies for the most challenging applications.

The 2018 ASTRO Annual Meeting Presidential Symposium highlighted areas of rapid change and opportunity for radiation oncology. As we move through 2019, we look forward to assessing the progress and impact of these four areas on daily radiation oncology practice and seeing the latest research presented at ASTRO’s 61st Annual meeting in Chicago.

Stay tuned for a series of blogs featuring the exciting changes coming to the 2019 Annual Meeting and Presidential Symposium!

Posted: April 30, 2019 | with 0 comments

Blog post currently doesn't have any comments.
 Security code