Genomic-Adjusted Radiation Dose (GARD) is a mathematical formalism that allows radiation oncologists to translate how much dose comes out of the radiation machine and into the patient and translate that into how much that will affect an individual patient.
And it does that through a genetic signature of the tumor genome. And I think to anyone who's (patients) treated radiation patients or any patients in the past. We all know intuitively that the same (radiotherapy prescription) dose doesn't do the same thing to each patient (radiotherapy benefit). One patient may have an incredibly good response, and another patient may have a response.
That's frustrating for all of us, and I think that heterogeneity of response is something that we have not yet been able to capture in radiation oncology. So, what exactly is genomically adjusted radiation? Does it allow us to change the way we think about radiation dosing from one that is guided by a physical paradigm?
That is, how much energy should I put into the tumor in the form of radiation? And instead, use a biological plus mathematical predictor to ask the question, "How much effect do I want to deliver?" And so we can then use a genomic signature of the tumor, together with the radiation dosing decisions, to make a prediction about how that will affect a patient.
So the history of that is that it starts with a genomic signature, which was published by colleagues of mine quite some time ago in the late 2010s, or sorry, the late 2000s. Then, in 2017, we came up with the idea of Genomic-Adjusted Radiation Dose (GARD). Genomic-Adjusted Radiation Dose (GARD) takes that genomic signature and runs it through the standard equations of radio biology to get this number, which we call the GARD score.
And then just last year in 2021 and Lancet Oncology, we published a large scale PAN cancer analysis, which actually had seven different cancer types and 11 cohorts of about 2000 patients, and showed the genomic prediction of radiation effect and Genomic-Adjusted Radiation Dose (GARD) strongly outperformed radiation dose as a predictor.
So what that means overall is that in a large PAN cancer analysis we showed statistically that if you just know how much dose someone gets, more or less, that's a very poor predictor of their outcome when it comes to survival and local control. However, if you modulate that dose through the lens of (cancer) genomics, you are actually able to get a strong, statistically significant predictor of outcome.
These two studies gave us the information we needed to look into a larger group of people with head and neck cancer, which is what we did here. So, in the past, we had to combine quite small groups to get the statistical power we needed. However, in this new cohort, which is a cohort of patients from an Italian Cooperative Group who are all oropharynx cancer patients with HPV positive disease whose tissues were collected on FPE and sequenced after the fact.
We were able to have enough patients, around 250, that we were able to prove, through the statistics, that the signal was strong enough to show us that Genomic-Adjusted Radiation Dose (GARD) was a continuous predictor of outcome. And what that means is that for each unit increase of Genomic-Adjusted Radiation Dose (GARD), we saw a statistically significant increase in survival and local recurrence-free survival. So these two factors, we believe, allow us to take the next step in understanding how genomics and radiation work together to predict outcomes in patients with head and neck cancer.
Read and Share the Article Here: https://oncologytube.com/v/41381
As a result, the standard of care for patients with early stage oropharynx cancer is HIV positive. Oropharynx cancer treated, is usually a combination of chemotherapy and radiation. There have been recent studies looking at de-escalation of therapy. So I think when I was training 10 years ago, the standard was a combination of 70 g of radiation plus chemotherapy, usually in the form of a platinum agent. And people over the years have noticed that HPV positive diseases might need less therapy. And so we've done different studies looking at reductions in the dose. From 70 to 60, we've removed chemotherapy and either removed it entirely or replaced it with agents like cetuximab.
But I think that there's currently we're still struggling to understand which patients should be deescalated. And in this study, what we show is that the vast majority of patients with HIV-positive disease could be, but there still remains a group of patients who are genetically, or sorry, genomically, disinclined to have that de-escalation. And so what we're hoping is that by the incorporat. With the development of tumor biology in the form of expression profiling of tumors, we can really do a better job of picking which patients are good candidates for dose de-escalation and possibly even for dose escalation.
So the 2, or I say the 3, most exciting results from this study are that first, this is the first time that we've had a cohort large enough to show that Genomic-Adjusted Radiation Dose (GARD) is a statistically significant predictor of outcome as a continuous variable. And that's very important because most biomarker studies are done in a dichotomous manner where we ask Maybe biomarker high is different than biomarker low, but in this case, Genomic-Adjusted Radiation Dose (GARD). This is not a typical biomarker. And what we're able to do is show that for every step increase in Genomic-Adjusted Radiation Dose (GARD) that incorporates or is associated with a significant increase in our good outcomes in these clinical trials.
So that's the number one and number two results, which are significant, as we can see further down. I think it's very easy; it's a lot easier for radiation oncologists to think about dose levels that are applicable to certain patients. And what we were able to do is show that not only does Genomic-Adjusted Radiation Dose (GARD) also perform as a dichotomous variable, which is obvious given that it was significant as a continuous variable, but we were also able to identify 3 different dose levels of differential outcomes that were significantly different and identified, in particular, a very high-risk group of patients based on tumor genomics that we think would not be appropriate for dose de-escalation. And then the final result, which I think is also useful today in the clinic, is an important point. I think the dose adjustments are likely going to be things that we need to do either on registries.
In the course of my radiation oncology residency program, I really had a desire to. I did basic research, but I was woefully unprepared to do cancer biology as it's, I would say, normally practiced. But I realized that I did have skills and quantitative approaches either as a complex systems analyst or engineer. So, I took a break from my residency to study math at Oxford University. And I did a thesis on mathematical biology and radiation predictive radiation biology. And that led me to understand radiation genomics and work on mathematical models of radiation genomics. And now my clinical practice here at Cleveland Clinic and my laboratory focus heavily on cancer evolution and also predictors of outcomes from the genome and also from dynamic data such as Radiomics or Delta radiomics. And so it is. This is an extension of our previous work as a more holistic PAN cancer approach to really drill down on it in the individual disease of night cancer in order to make a difference for all of our patients.
First of all, thanks for your interest in our work, and I look forward to sharing a manuscript with all of you soon. Also, keep your eyes on the preprint servers. We'll post it on medRxiv before publication, so it should be in the next few months. And second, please feel free to reach out to me with any questions about this topic or anything else in radiation oncology.
Dr. Scott is a former member of the US Navy submarine force who is now a physician-scientist in academia. Through mathematical modeling and biological and clinical validation of these models, our team seeks to research deconstructing the complexity of cancer. His education in His background in physics, medicine, mathematics, and engineering has provided him with a unique viewpoint on cancer and systems biology, and I am able to interact and cooperate with specialists from a variety of fields. He has worked extensively on the mathematical modeling of the evolution and therapy of cancer, utilizing a number of models, such as evolutionary game theory, cellular automata, differential equations, and Markov chains. His doctoral dissertation was about how genetic and microenvironmental differences affect how cancer changes and how it responds to radiation. His research is about how cancer changes and how it becomes resistant to treatments. Since we started the Theory Division, we've started to branch out. The lab now has a large experimental part, and we're doing evolution experiments on both cancer cell lines and bacteria. His group is one of the most interdisciplinary in the fields of translational cancer evolution and evolutionary medicine because they combine math, experimental evolution, and a clinical focus. He wants to help cancer patients and has a unique point of view that can help this field move forward.