How a Cancer’s Genome Can Tell Us How to Treat it

By Gesa Junge, PhD

 

Any drug that is approved by the FDA has to have completed a series of clinical trials showing that the drug is safe to use and brings a therapeutic benefit, usually longer responses, better disease control, or fewer toxicities.

Generally, a phase I study of a potential cancer drug will include less than a hundred patients with advanced disease that have no more treatment options, and often includes many (or all) types of cancer. The focus in Phase I studies is on safety, and on finding the best dose of the drug to use in subsequent trials. Phase II studies involve larger patient groups (around 100 to 300) and the aim is to show that the treatment works and is safe in the target patient population, while Phase III trials can involve thousands of patients across several hospitals (or even countries) and aims to show a clinical benefit compared to existing therapies. Choosing the right patient population to test a drug in can make the difference between a successful and a failed drug. Traditionally, phase II and III trial populations are based on tumour site (e.g. lung or skin) and/or histology, i.e. the tissue where the cancer originates (e.g. carcinomas are cancer arising from epithelial tissues, while sarcomas develop in connective tissue).

However, as our understanding of cancer biology improves, it is becoming increasingly clear that the molecular basis of a tumour may be more relevant to therapy choice than where in the body it develops. For example, about half of all cutaneous melanoma cases (the most aggressive form of skin cancer) have a mutation in a signalling protein called B-Raf (BRAF V600). B-Raf is usually responsible for transmitting growth signals to cells, but while the normal, unmutated protein does this in a very controlled manner, the mutated version provides a constant growth signal, causing the cell to grow even when it shouldn’t, which leads to the formation of a tumour. A drug that specifically targets and inhibits the mutated version of B-Raf, Vemurafenib, was approved for the treatment of skin cancer in 2011, after trials showed it lead to longer survival and better response rates compared to the standard therapy at the time. It worked so well that patients in the comparator group were switched to the vemurafenib group halfway through the trial.

While B-Raf V600 mutations are especially common in skin cancer, they also occur in various other cancers, although at much lower percentages (often less than 5%), for example in lung and colorectal cancer. And since inhibition of B-Raf V600 works so well in B-Raf mutant skin cancer, should it not work just as well in lung or colon cancer with the same mutation? As the incidence of B-Raf V600 mutations is so low in most cancers, it would be difficult to find enough people to conduct a traditional trial and answer this question. However, a recently published study at Sloan Kettering Cancer Centre took a different approach: This study included 122 patients with non-melanoma cancers positive for B-Raf V600 and showed that lung cancer patients positive for B-Raf V600 mutations responded well to Vemurafenib, but colorectal cancer patients did not. This suggests that the importance of the mutated B-Raf protein for the survival of the tumour cells is not the same across cancer types, although at this point there is no explanation as to why.

Trials in which the patient population is chosen based on tumour genetics are called basket trials, and they are a great way to study the effect of a certain mutation on various different cancer types, even if only very few cases show this mutation. A major factor here is that DNA sequencing has come a long way and is now relatively cheap and quick to do. While the first genome that was sequenced as part of the Human Genome Project cost about $2.7bn and took over a decade to complete, a tumour genome can now be sequenced for around $1000 in a matter of days. This technological advance may make it possible to routinely determine a patient’s tumour’s DNA code and assign them to a therapy (or a study) based on this information.

The National Cancer Institute is currently running a trial which aims to evaluate this model of therapy. In the Molecular Analysis for Therapy Choice (MATCH) Trial, patients are assigned to a therapy based on their tumour genome. Initially, only ten treatments were included and the study is still ongoing, but an interim analysis after the 500th patient had been recruited in October 2015 showed that 9% of patients could be assigned to therapy based on mutations in their tumour, which is expected to increase as the trial is expanded to include more treatments.

This approach may be especially important for newer types of chemotherapy, which are targeted to a tumour-specific mutation that usually causes a healthy cell to become a cancer cell in the first place, as opposed to older generation chemotherapeutic drugs which target rapidly dividing cells and are a lot less selective. Targeted therapies may only work in a smaller number of patients, but are usually much better tolerated and often more effective, and molecular-based treatment decisions could be a way to allow more patients access to effective therapies faster.