Last month, a collection of some of the brightest minds in cancer research were brought together under one roof at the second ever Molecular Analysis for Personalised Therapy conference.
Taking place in London, the two-day event saw a fascinating series of talks, engaging discussions and thought-provoking debates all centred on one topic: making cancer medicine more personal.
This aspiration is shared by doctors and scientists alike, and it’s their hope that this approach will become the future of cancer treatment.
“No man ever steps in the same river twice,” said one researcher as they reflected on the complexity of cancer.
Just as the water flowing in a river is unique and changing, so too are tumours. Exploiting this individualism with tailored treatments is the crux of personalised, or precision medicine.
And while the conference showcased the progress made so far, it also highlighted the challenges that researchers are facing.
“There is a major gap between our understanding of cancer biology and our basic approach to patient management,” said Professor Charles Swanton, a Francis Crick Institute scientist, part-funded by Cancer Research UK and one of the conference founders.
This isn’t through lack of trying, but it does reflect the fact that more needs to be added to the ever-growing cancer knowledge pot. And key to this is data; lots of it.
A problem shared
It turns out that scientists only know a fraction of the genetic faults that can give rise to cancer. Without filling in the gaps, treatments can’t be developed that target these flaws, meaning personalised treatments can’t be realised.
This lack of data is at the heart of the work of Professor Fabien Calvo, Chief Scientific Officer for Cancer Core Europe. He helped to establish the International Cancer Genome Consortium (ICGC), a global collaborative effort that aims to identify the faulty genes inside cancer cells by reading the DNA in tumour samples.
Our scientists are involved in ICGC projects focusing on prostate and oesophageal cancer. And by 2018, the Consortium want to have collected and studied both normal and tumour tissue from 500 cancer patients, for 50 different types of cancer. That’s 25,000 tumour samples and 25,000 healthy tissue samples in total.
If we understand the barriers, we can overcome them over time
– Dr Lillian Siu, University of Toronto
“The ICGC has already identified targets for new or old drugs in a range of cancers,” Calvo said. “And it will accelerate research and discovery for new targeted drugs.
“But we have a lot to understand still.”
Making sense of all this data won’t be easy, but many hands make light work. That’s why collaboration is crucial to the success of this project, and why the data will be made freely available to researchers.
“The interest in sharing data has become a natural phenomenon over the last few years,” said Dr Lillian Siu from the University of Toronto, who gave a complementary talk on data sharing in personalised medicine.
“It means you can learn from information that would otherwise be buried.”
And while sharing data may sound obvious, for many researchers it isn’t that simple. Through surveys, Siu has been identifying the challenges in data sharing in cancer biology, which boil down to restrictions ranging from the law to funding and privacy issues.
“If we understand the barriers,” said Siu, “we can overcome them over time.”
Target identification – spotting weaknesses in cancer cells that drugs can exploit – is the first step in personalised medicine. Finding ways to lock onto these targets is the next. But even if scientists find promising possible target molecules in the lab, that doesn’t necessarily mean that experimental drugs will work well in patients.
This frustration was exemplified by Professor Bart Vanhaesebroeck from University College London, who works on drugs that block a key molecule linked to cancer called PI3K. The molecule triggers a series of signals inside cells that control cell growth and survival. And these signals frequently go wrong in cancer.
Early versions of possible drugs have had disappointing results in patients, Vanhaesebroeck explained, which could be due to the fact that cancer cells are happy to keep growing even when these signals are dampened down. That’s because the cells can survive with only a fraction of normal PI3K activity, so it’s very difficult to keep the signals low enough with drugs to achieve any sort of effect.
But one successful drug has made it into the clinic, Idelalisib (Zydelig), which is used to treat certain blood cancers. And it turns out that it works in a rather surprising way: not by directly killing cancer cells as was first thought, but by boosting the immune system’s response against cancer cells. This means that, unexpectedly, the drug is actually a type of immunotherapy.
Understanding how drugs work is an important part of personalised cancer treatment, because if scientists know what they’re targeting, then a drug’s uses could potentially be extended. That’s why Dr Sheila Singh, from McMaster University in Canada, is developing new ways to study how certain tumours respond to treatment in the lab.
She’s growing brain tumours in mice from patients’ cells and closely monitoring them by making certain important molecules in the cell glow. “This allows us to see what happens as the tumour evolves over time,” she said. And that could help scientists develop more precise ways to target these changes and treat certain brain tumours more effectively.
Digging for detail
Studying cancer in mice offers researchers the opportunity to explore tumours in ways that wouldn’t be possible in patients. But what happens in mice may not reflect what happens in people. So what if there were a way to make detailed studies of tumours in patients, but without the need for invasive tests like traditional biopsies?
That’s where Professor Philippe Lambin’s work comes in. Based at Maastricht University, his team is using sophisticated computer software to create pictures of tumours that reveal more information than traditional scans. This is called radiomics.
“If cancer medicine is based on a single biopsy, then we have a problem,” Lambin said. He questioned where a biopsy might be taken from, saying that without knowing the exact location, it might not reflect the variety of cells that are seen across a tumour.
His team has already shown that certain features like tumour shape and texture can help predict how patients with head and neck cancers will fare, and that it’s possible to see how a patients’ tumour is responding to treatment over time.
Filling the empty glass
The idea of shifting the focus of personalised medicine away from genetic faults is interesting. Since radiomics is a relatively new field, finding out whether or not it can give researchers enough information to tailor treatments effectively will require many more studies. But it could potentially address an issue raised by Dr Levi Garraway, based at the Dana-Farber Cancer Institute: will it ever be feasible to profile every cancer patient’s genetic faults in the clinic?
He emphasised the fact that the majority of cancer patients have genetic faults that are rare, and that only around five in every 100 have cancer-causing genes that can be targeted with existing drugs.
“But that’s taking a glass half empty view,” he said. “If you have a glass half full view, then that means a significant proportion of cancers contain at least one [faulty gene] that may be actionable.”
I really believe it’s not the time to stop the efforts
– Professor Fabian Calvo, Cancer Core Europe
Though of course, with further research into drug discovery more could become actionable in the future. But even with the numbers we currently have, evidence from patients is now showing that personalised medicine is worth pursuing.
Announced by project leader Professor Jean-Charles Soria from South-Paris University, the Moscato trial found that 199 out of 1,110 patients with advanced cancer benefitted from therapy tailored to faults in their tumours.
In these patients, they lived around 30 per cent longer before their tumours started growing again compared with previous treatments they had been given.
That may sound small, but it means that tumours were kept under control in around 1 in 5 patients whose previous treatment had failed. And precision medicine is still in its infancy, so with further research hopefully that number will increase.
And optimism for the future was abundant among those at the event, with some making it clear that now is not the time for research to slow down.
“We haven’t yet done enough to help patients,” said Calvo. “I really believe it’s not the time to stop the efforts.”