Decoding cancer: hopes rise for a new age of precision medicine

Ever since scientists began decoding the human genome in 1990, doctors have predicted an age of "precision medicine" where patients are matched with drugs that target the specific genetic drivers of their disease. Nowhere would such a revolution be more keenly welcomed than in the treatment of cancer, a single word that describes a multitude of diseases caused by a vast array of mutations.

Now, 14 years after the genome project was declared complete, some scientists say the era of precision medicine is here. A new generation of targeted drugs is emerging and could, they say, usurp chemotherapy as the mainstay of cancer treatment, sparing millions of people from a brutal approach that attacks healthy cells as well as tumorous ones.

Underpinning the shift is the realisation that the molecular features of a tumour, such as its genetic traits, are more important than where it started growing, whether it be the lung or prostate, brain or breast. In the next stage of the war against cancer, biology trumps anatomy.

In May, the US Food and Drug Administration approved the first oncology medicine for a specific genetic feature regardless of the site of the tumour. Keytruda, an immunotherapy made by Merck & Co, can now be prescribed to all cancer patients as long as they meet the genetic criteria.

A second such "pan-cancer" medicine is expected to launch in early 2018, when Loxo Oncology, a biotech group, secures approval for its drug, which targets a genetic defect that is very rare but found across almost all types of tumour.

"The way chemotherapy has been developed is to just test it, and - if you get an acceptable level of activity in the average patient - you then treat everybody," says Richard Schilsky, chief medical officer at the American Society of Clinical Oncology (Asco). "We are moving away from that."

This transition is being made possible in part by technologies that can quickly scour tumours for genetic and protein-based clues known as "biomarkers", allowing doctors to match patients with drugs or treatments that are likely to work.

Many of the innovations come with a hefty price tag, prompting concern about whether healthcare systems can bear the cost, given that they are already contending with a rapidly ageing population and soaring prices for conventional drugs and treatments.

Although the terms are often used interchangeably, precision medicine should not be confused with the futuristic realm of "personalised" medicine. The latter denotes a bespoke drug manufactured specifically for an individual patient, like Novartis' chimeric antigen receptor therapy (Car-T), which re-engineers a patient's cells in a laboratory so they will attack cancer.

Conversely, precision medicine refers to the continuous interrogation of tumours to try to anticipate whether they will respond to a particular treatment. That drug might be an experimental therapy being studied in clinical trials, but it is just as likely to be an existing product, or even a decades-old medicine that was developed for an entirely different type of cancer.

Below, the Financial Times explores four developments that could turn precision medicine from an exciting but fledgling field into standard practice.

Oncologists have traditionally categorised tumours according to what they can see under a microscope. How big are the cells? What is the relative size of the nucleus to the rest of the cell? The answers to these types of questions have often dictated how a person is treated.

But doctors and executives say molecular testing will become more common as the price of genetic sequencing falls and pharmaceutical companies develop niche drugs to target specific mutations responsible for an individual cancer.

"We've been saying for decades, that the idea of classifying tumours by what they look like under a microscope is probably not sensible," says Dr Roger Perlmutter, the most senior scientist at Merck & Co, the US drugmaker. "More and more I think we'll be defining tumours in a [molecular] way."

Molecular testing is already commonplace in cancers such as lung and breast, where doctors identify each mutation in turn. But the step-by-step process takes up time when a patient could be receiving alternative treatment.

Next-generation sequencing tests can screen a single tumour tissue sample for hundreds of genes known to play a role in cancer at one stroke, offering doctors a comprehensive picture shortly after diagnosis. If oncologists discover specific mutations, they can match the patient with drugs designed to tackle that type of cancer, or try to secure them a place on a trial of an experimental medicine. "One shot and you can get a whole plethora of information," says Joydeep Goswami, president of oncology at ThermoFisher, a US diagnostics company.

Foundation Medicine, a rival diagnostics group majority owned by Swiss drugmaker Roche, has developed a broader test, which scans 324 genes known to play a role in the disease. The test is awaiting approval by the FDA.

The high cost of molecular testing helps explain why only 12 per cent of US patients receive comprehensive genome profiling. Foundation's test costs $5,800, compared with a few hundred dollars for diagnostic kits that interrogate just one or two genes.

Some doctors also worry that such tests turn up reams of data without telling them what it means.

"The argument against it is that a lot of the information that turns up in this broad-based testing is information that nobody knows how to interpret and how to act on," says Asco's Dr Schilsky.

If next-generation diagnostics offer a comprehensive snapshot of a patient's cancer, they fail to address one of the main barriers to fighting the disease: it is a notorious shape-shifter, quietly developing resistance to drugs and returning with a vengeance.

The picture developed shortly after diagnosis - no matter how detailed - might become inaccurate within months. By the time the symptoms of recurrent cancer emerge, it can often be too late.

Repeating the screening tests regularly would not only prove prohibitively expensive, it would also be impractical because patients would have to undergo a dangerous number of biopsies to provide fresh tissue samples.

An emerging solution is a type of blood test for cancer known as a liquid biopsy, which can detect tiny fragments of DNA discarded by tumours.

"What we now know is that cancer cells are constantly turning over and dying," explains Chris Bischoff, head of immuno-oncology at Pfizer. "When they die, DNA sheds into the circulation, so you can pick it up from the peripheral blood even if it's 1 per cent or less. If there is an actual mutation, you can potentially detect it."

Pfizer recently announced a partnership with Guardant Health, a Californian company that is seen as having the most accurate liquid biopsy on the market. Guardant has also signed deals with other large drugmakers like Merck and AstraZeneca, which want to use the test in their clinical trials.

The product, Guardant 360, can detect alterations in 73 genes associated with cancer, although the company recently announced plans to increase the size of the panel to 500 genes.

Given that such tests can be performed using just two vials of blood, there are few restrictions on how regularly they can be performed, according to Dr Schilsky.

"You have to interrogate the tumour each time after the treatment plan because the tumour will change," he says. "With liquid biopsies you can much more easily test on a repetitive basis and actually monitor what's happening to the cancer over time."

However, cost is a major sticking point. Guardant 360 has a pricetag of $5,400 and the company has struggled to convince insurers and healthcare systems to pay for the product in large quantities.

If sophisticated genetic testing were to become more common, it could herald an era of hundreds of ultra-targeted drugs developed for specific mutations that drive cancer.

One such medicine, Larotrectinib, developed by Connecticut biotech group Loxo Oncology, has recently captured the attention of both oncologists and Wall Street, with the latter group sending the company's shares up by 251 per cent over the past 12 months.

The drug is designed for a tiny subset of cancer patients, measured in the low thousands, whose tumours have a mutation in the so-called TRK gene. The mutation is rare - some estimates put its prevalence at 0.5 per cent of cancers - but it does occur across all tumour types regardless of their location in the body.

Hyper-targeting has made the drug hugely effective, producing the kind of results unheard of in cancer studies. In a recently published clinical trial, the medicine either shrank tumours or stopped them growing in 76 per cent of patients, putting the medicine on track to win speedy FDA approval next year.

"Hopefully there are another five of these [drugs] next year," says Jacob Van Naarden, chief business officer of Loxo. "At some point you're carving out so many slices of the pie and they're all on the 'long tail': in other words, they are all rare, the only way to find these things is to look comprehensively."

If the FDA approves Larotrectinib, it would be one of the first "pan-cancer" drugs given to patients regardless of where their first tumour is located as long as they have a particular mutation.

But precision medicine is not just about new drugs: it is about finding those who are most likely to respond to existing products. An Asco trial of more than 300 patients uses drugs that have been approved for one type of cancer to treat patients with a different kind, if they have a biomarker that suggests they might respond. Roche's Herceptin has been used in breast cancer patients with an abnormality in the HER2 gene since 1998, but is not approved for patients with other forms of the disease - even if they have the same mutation.

"We have many patients already who have tumour types other than breast cancer who are getting HER2 directed therapy on the trial," says Dr Schilsky. "There could well be lung cancer patients or colon cancer patients or really anything else."

The theory that the genetic features of a tumour are more important than its location in the body could signal a need to overhaul the field of oncology, which has based its research and treatment on the original site of a cancer.

"We have historically thought about cancers in an incredibly anatomically based way," says Gary Gordon, vice-president of oncology at AbbVie. "At what point do we start moving away from anatomy to start thinking about biology? So, rather than being a breast cancer specialist, you're a specialist in the HER2 [gene abnormality] instead."

Some oncologists and executives say such a day is still a long way off: drugs are but one way of tackling cancer and there are still reasons for focusing on the site of the tumour. "This is not going to be a revolution in the way we suddenly reorganise every cancer centre in the country, take down the sign that says 'breast clinic' and put up one that says 'HER2 clinic'," says Dr Schilsky.

It is already difficult for oncologists, many who trained before the human genome was decoded, to keep up with the huge body of research on the molecular traits of cancer. If hundreds of niche drugs like Loxo's Larotrectinib replace the comparatively small number of blockbusters used today, picking the right therapy might become harder.

Some large cancer treatment centres have tried to address the problem by holding regular tumour board reviews, where oncologists from a range of specialities meet to discuss cancers from outside their field of expertise. A breast cancer doctor familiar with HER2 abnormalities might spot something that a prostate cancer expert has not.

Replicating this model at scale is not feasible, but some believe that technology could perform much the same role. The makers of next-generation diagnostics like Foundation Medicine and ThermoFisher are developing a kind of "Bloomberg for cancer" that would take a patient's test results and automatically produce a list of available options, from existing drugs to experimental trials for which they might be a candidate.

"Simply having a great test is not enough," says Vince Miller, chief medical officer of Foundation Medicine. "Because sometimes doctors don't know what they're sitting on. They may not have the time, they may not know all the trials, they're too busy. We've developed a lot of decision support tools that are taking down those barriers."

Copyright The Financial Times Limited 2017