A surgeon, a physicist and a chemist walk into a bar – how cross disciplinary research can push forward early detection

“Overdiagnosis is a big problem. But I’d argue that underdiagnosis is worse,” says clinician scientist Professor Richard Shaw – an oral-maxillofacial surgeon who researches head and neck cancer at the University of Liverpool.

“In the worst-case scenario, we wrongly discharge a patient. Then a few years down the line they’re diagnosed with oral cancer and need major surgery, lymph node dissection, facial reconstruction, radiotherapy – you name it,” he says.

In recent years, we’ve seen impressive technological advancements that support clinicians to make important clinical decisions. The iKnife, for example, provides real-time information about surgical margins for breast and ovarian cancer. The Cytosponge can distinguish between patients who have innocuous heartburn and those with Barrett’s oesophagus who have a higher risk of oesophageal cancer.

“I’m all ears if we can develop a technology that allows us to better stratify our patients at an earlier stage, without the need for repeat biopsies,” Richard says. We don’t yet have an effective predictive tool when it comes to the red and white patches that are found in the mouths of 4% of the population and make a person around 40-fold more likely to develop oral cancer.

And with UK cases of oral cancer rising by 34% over the past decade, intervening early – when treatment is more likely to be successful – is becoming increasingly important.

New tools tackle long-standing-challenges

Enter Professor Peter Weightman, professor of Physics at the University of Liverpool.

Described by Richard as an “absolute legend”, Peter has spent the past 50-plus years developing scientific tools to solve the most complex of problems. “One could argue,” muses Peter, “that scientific advances are more often driven by the development of a new instrument than by a new concept.”

He was introduced to the challenge of oral cancer patient stratification in the early 2010s by Dr Janet Risk, a biologist and senior lecturer also based in Liverpool. The trio began exploring whether precancerous mouth patches could be detected using Fourier-transform infrared (FTIR) spectroscopy – a technology long-recognised for its potential in cancer diagnostics. A major obstacle has been making sense of the vast amounts of data it produces: analysing a 500×500 pixel patch of tissue with FTIR can give rise to a staggering 10¹¹ datapoints.

Recent developments in machine learning have changed that, however, says Peter. “We’re finally at a point where we can deduce useful information from all this data and see how FTIR could be used in the clinic.”

With CRUK and NIHR funding, the team developed a machine learning algorithm that can predict malignant transformation from FTIR data, translating it into a prototype device that can predict prognosis in retrospective patient samples. The hand-held, pen-sized device – called the Liverpool Diagnostic Infrared wand (LDIR; pronounced “elder” wand) – importantly relies on only a small number of infrared lasers, rather than a full spectrometer.

The eventual aim is for the LDIR wand to be pressed against the tissue of interest – such as a patch inside a person’s mouth – to give a “yes/no” response as to whether the tissue is potentially dangerous. The hope is it’ll prove a useful, portable tool in both the clinic and lower resource settings, like community screening centres.

Earlier this year, the team were awarded a CRUK Early Detection Project Award to validate the LDIR wand using a large cohort of patient samples with long-term follow-up data. “Our early data suggests the LDIR wand can predict prognosis with 80% accuracy,” says Richard. “If that holds true in validation, that’s a significant improvement on existing tools that could really begin to pay off for our patients.”

Lessons from physical chemistry

The Liverpool team isn’t the only multidisciplinary group utilising IR spectroscopy that have received funding for cancer detection research this year.

Physical chemist Professor Neil Hunt, at the University of York, received a CRUK Early Detection Primer Award to explore two-dimensional infrared (2DIR) spectroscopy in liquid biopsies. Like the LDIR wand, part of liquid biopsies’ appeal comes from the low invasiveness of sample collection – welcomed by many patients as an alternative to repeat solid biopsies, if they can be made as effective.

This, combined with recent developments in laser technologies and a collaboration with a clinical colleague sparked the question: could 2D-IR be a useful addition for liquid biopsies?

“We hope our work will complement what others are already doing with liquid biopsies, by making it possible to zoom in on the protein region of the infrared spectrum,” says Neil.

Learning new languages

“Falling into cancer research was quite accidental actually,” laughs Neil. “But we’ve got the technology and the capability now to look at proteins in complex fluids. With our Primer award, we’ve now got 12 months to test the potential of ultra-fast 2D-IR for liquid biopsies.”

Often accidental, this entry into cancer research from researchers outside of life sciences isn’t just a bonus – it’s a necessity to tackle many research questions in cancer. Fellow physical chemist and Cancer Grand Challenges team lead Professor Josephine Bunch entered the world of cancer research in a similar way, after hearing about 3D tumour mapping on the Radio 4 Today Programme and considering how her mass spectrometry work could be applied.

And applying your skills and research acumen to a field other than your own can have personal benefits too. “Transitioning to cancer and learning to speak new languages has been a learning curve,” says Neil. “But I do think that doing something a bit different, stretching yourself a bit, will always have a positive impact.”

We often talk of the scientific language barrier in multidisciplinary research, and it can almost sound cliché. But like Neil, the key to Richard, Janet and Peter’s success was learning to speak each other’s languages. “Ultimately, we got fed up with failed grant applications,” laughs Richard. “We knew we were onto a good idea – but if I couldn’t really understand Peter’s side and he couldn’t really explain mine, how could we explain it to a funder?” As Janet previously described, “the time and effort we invested was necessary for us all to appreciate the nuances of the biological and physical sciences aspects of the project.”

“I knew nothing about cancer when we first started collaborating,” echoes Peter. “I must say, it’s one of the most complicated problems I’ve come across. Much more complex than electron spectroscopy.”