NEW YORK – Researchers at the University of Strathclyde in Scotland have completed a preliminary prospective study involving the use of an infrared spectroscopy blood test that achieved high accuracy in helping clinicians decide whether to send brain cancer patients for imaging.
Reporting the results of the study today in Nature Communications, the researchers described the application of attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy combined with machine learning to analyze blood samples from patients with and without brain cancer.
If additional clinical studies continue to show promising results, clinicians may be able to use the tool in a few years to prioritize patients for rapid access to imaging that would verify the presence or absence of brain cancer, Matthew Baker, a professor of chemistry at Glasgow-based University of Strathclyde and test developer, said in an interview. Such a strategy has the potential to reduce the number of patients receiving brain scanning that don't need it, while also reducing the financial burden on health systems, he said.
In their publication, Baker and his colleagues reported converting conventional ATR-FTIR, a promising medical diagnostic technique limited in its adoption by expense and low throughput, into a more affordable, rapid approach for cancer diagnosis.
Biological samples, such as the serum specimens used in the study, emit specific biochemical vibrations under infrared light — signatures associated with the composition of the sample. Traditional ATR-FTIR spectrometers have a single point of analysis, called the internal reflection element, made from a material with a high refractive index, such as diamond, germanium, or zinc selenide.
When a serum sample, which has a lower refractive index, is deposited onto the surface of the internal reflection element, an evanescent wave is formed at the interface between the sample and internal reflection element and generates a suitable diagnostic fingerprint.
However, use of such materials in traditional ATR-FTIR systems yield low throughput at a high price, partly because they are expensive and difficult to mass produce in quantities needed for diagnostic systems, limiting the adoption of these systems in clinical applications.
Baker said that he and his colleagues have overcome limitations to adoption of the traditional systems. In February, they cofounded ClinSpec Dx, a company spun out of Baker's research at the University of Strathclyde and elsewhere. They overcame limits of the traditional approach for diagnostics by substituting silicon for diamond, germanium, or zinc selenide, and thereby introducing a more affordable material that can be mass produced and enable several measurements at one time, without losing functionality associated with the replaced materials
Further, they reported overcoming a challenge associated with silicon lattice vibrations that had previously held back use of the material for diagnostic applications. They used microfabricated silicon internal reflection elements that circumvented the unwanted spectral contributions from silicon. The silicon internal reflection elements that they fabricated are sandwiched between a printed plastic holder and a medical-grade labelled sticker that defines the optically active regions of the silicon internal refraction element.
Using samples from a 104-patient prospective cohort, the group found that the instrument differentiated cancer from healthy control patients at a sensitivity of 93.2 percent and specificity of 92.8 percent, Baker said.
Given the number of patients that could avoid having to undergo unnecessary computed tomography (CT) scanning because a blood-based tool may now become available, the diagnostic test under development by Baker and his colleagues could be "a game changer," James Livermore, a neurosurgeon at University of Oxford, UK, John Radcliffe Hospital, said in an interview. "If you can save any scanning time, that saves money for health systems and insurance companies that received reimbursement claims," he said.
Livermore is familiar with the test being developed by Baker and his colleagues but not involved in its development. Livermore is developing a Raman spectroscopy blood test used to conduct rapid genetic classification of gliomas during surgeries to guide surgical procedures.
"The difficulty with detecting brain tumors is that they are rare," Livermore said. "If you are a general practitioner you might have one or two patients in your career that has a malignant brain tumor. You may have thousands of patients who have headaches."
As a result, a major question for physicians is how to differentiate people with brain cancers from those who get headaches from other health conditions, he said. "The power of this technology is that it is fairly inexpensive and easy to use based on inserting a drop of blood into a machine."
The diagnostic platform's applications extend beyond its potential use to triage brain cancer patients. Because the tool can be used to evaluate types of tumors, clinicians may also use the test to help decide which patients need to undergo surgery versus needle biopsies, Livermore said. Further, clinicians may use the approach to decide whether a malignancy in the brain is the primary or secondary source of cancer, and if it is a secondary source, they may be able to use the new tool to identify which organ is the primary source of a malignancy.
For broad, practical use of the ClinSpec Dx test, many diagnostic labs would need to purchase a spectroscopy instrument. Although such equipment is used in many large academic centers, they are not routinely used in traditional medical diagnostic laboratories, Livermore said.
At the same time, he noted, an ATR-FTIR system costs tens of thousands of dollars, substantially cheaper than other instruments used in laboratories that cost hundreds of thousands of dollars.
The platform, which fits on a desktop, takes up very little space, an advantage for laboratories, Livermore said.
Commercial planning
Baker said that he began work on the brain cancer test in 2012 and developed it with colleagues "in a stepwise fashion." Based on an economic model that he and his colleagues developed and had published last year in BMJ Open, a serum-based spectroscopy test applied as a screening tool for brain cancers is cost-effective with sensitivities and specificities at 80 percent, or more, in UK and US clinical settings.
The Nature Communications publication today represents a milestone for his firm because it describes the first prospective clinical trial to use the tool in hospitals on patients to identify their disease status, he said.
Baker and his colleagues launched ClinSpec Dx to commercialize the test this year after receiving an undisclosed amount of High Growth Spinout Programme funding from Scottish Enterprise in 2016 that enabled it to develop the technique that they are using now and its clinical hardware.
The company raised £1.6 million ($2.0 million) in seed funding from private investors that enabled it to conduct its clinical trial. In July 2020, it anticipates opening a Series A round to obtain several million pounds to support its next stage of development, which includes conducting further clinical trials, Baker said. At around that time, the firm expects to kick off a trial involving about 1,600 people to further validate the performance of the infrared platform.
"The main aim is to achieve regulatory clearances based on successful clinical trials and try to launch the test in three to four years," Baker said. The firm will seek CE marking and IVDR approval from European regulatory authorities and clearance from the US Food and Drug Administration. It is currently discussing clinical trial requirements with the FDA and researching the need to obtain codes for reimbursement, he added.
ClinSpec Dx is also in discussion with the UK's National Health Service "about where to place the test and how to place the test," he said, adding that the firm is also talking to potential global partners about future distribution plans.
ClinSpec Dx anticipates placing instruments in hospital laboratories from which clinicians will order tests. A lab can process a test in about 10 minutes from the time a blood sample is placed in the machine, and the result can then be sent back to the patient.
"Ultimately, achieving test placements all hinges on getting good clinical data," Baker said.