The drug is used for a variety of solid tumours and usually patients take this drug when they have advanced disease, when the disease has been metastatic and spread throughout their bodies. Again, a variety of tumours from colorectal to kidney cancer to hepatocellular carcinoma, ovarian cancer and a few others. This drug is usually taken in combination with other agents and so patients, in fact, take a cocktail of drugs including bevacizumab.

What were the aims and objectives of this analysis?

The analysis was really motivated by the fact that there are no markers or indicators that would let physicians understand which patients might experience the side effects of bevacizumab after they are treated with this drug. This drug causes, quite often, up to 40% of patients, causes hypertension, an elevation in blood pressure, and causes kidney toxicity that manifests itself with a leak of proteins into the urine. The kidneys cannot retain the proteins in the body. As I said, because the side effects are common and because there is no way for physicians to predict which patients might experience this toxicity and also, number three, because this toxicity might be severe in some patients, then we decided to perform a study looking at and trying to identify predictors of these side effects of the drug.

Really the main aim was to study the genetic variation in the DNA of the patients. We are using the DNA of the patients that is responsible for traits like colour of the eyes, height and many others. So we investigated all of these genetic variations in the patients, in their DNA, and then we wanted to associate the frequency of these genetic variations in the patients that had toxicities versus the patients that did not have the toxicities. By doing so we were able to then identify some important genetic variations that are associated with a risk of toxicity, of hypertension and kidney toxicity, in patients treated with bevacizumab.

We did this in a large group of patients. We had access to data from more than 1,000 cancer patients treated with this drug.

What methodology did you use?

Absolutely. The methodology is based on a genetic analysis of common variants in the genome of the patients. So DNA was extracted from the peripheral blood of the patients. We used a cheap assay, a high throughput cheap genotyping assay to interrogate all of these variants. Then, because the patients have been involved in randomised clinical trials, we have also obtained all the toxicity data that were related to the patients. Then through statistical association testing we evaluated whether patients that had a high rate of toxicities or higher severity of toxicity were carrying these genetic variations more often than patients that did not have these toxicities. That’s how the statistical analysis was performed.

What were the key results?

When we looked at these genetic variations and their occurrence in patients with and without toxicity, we have looked at hypertension and we have found a genetic variant that occurs at a pretty high frequency in the general population, about 30% in the global population, that was associated with an increased risk of hypertension in patients treated with bevacizumab.

In addition to this, we also wanted to look at this specific association for this variant in another completely different set of patients, just to provide additional external replication and validation for the findings. In fact, when we have also interrogated this variant in the patients that belong to a completely separate group of patients that were treated with bevacizumab and we looked for association between this variant and the risk of hypertension, we in fact saw the same effect. We saw the same signal. So for this variant we were able to provide a validation of the association between the variant and the risk of hypertension.

But what pertains to the kidney toxicity, we also have identified another common variant associated with increased risk of kidney toxicity in patients treated with bevacizumab. Unfortunately we have been not able so far to have another external dataset where we can validate further this association.

Regarding the variant that we have found associated with hypertension, the variant is located in one gene that codes for a potassium channel. There are several studies pointing to the fact that this potassium channel is very important for the risk of hypertension – when this protein has been knocked out in experimental models in animals the models experienced hypertension. So we believe that there is a very strong biological rationale for supporting the association that we have seen between the variant and the risk for toxicity.

How will these results influence the future use of this drug?

Physicians will have genetic tests that they can use to screen patients that are candidates for treatment with bevacizumab. According to the results of these tests, these tests will tell the physicians whether there is an increased risk or not for their patients to develop hypertension. The genetic test is a simple test that can be run at a relatively inexpensive cost. So physicians, by evaluating the results of the test, will have an additional tool, an additional marker, to evaluate the risk-benefit ratio for the patients that can be treated with this drug.

Physicians can use this test for potentially intervening in many different ways. They can spare the patients for treatment with bevacizumab if, in fact, there is a high risk of toxicity and then use another drug that might be similarly effective in that particular tumour. They can reduce the dose of the drug in the patients at risk of hypertension and then potentially increase the dose of the drug in further cycles if they don’t experience hypertension. They can start potentially treating with anti-hypertensives at the first onset of mild elevation in blood pressure in patients at higher risk by increasing the monitoring of the patients during the course of treatment.

So there is a variety of interventions that the physicians can take in the case of this genetic variation and the results that are related to screening for this genetic variation. Clearly, the results need further prospective studies that will evaluate the interventions that can be applied in patients that have been screened for this genetic variant for hypertension and then evaluate what is the result of all of these different interventions for really reducing the risk of toxicity for the patients.

I would like to thank all my collaborators and all the patients that participated in the clinical trials that provided consent for performing the genetic analyses. I want to thank you for the opportunity to speak today.