Real-time prostate cancer detection

It’s a sad fact that around 14 New Zealanders die every week from prostate cancer and that many are subjected to invasive and inaccurate diagnostic and surgical procedures, which is why Dr Claude Aguergaray wants to improve the medical outcomes.

“You have issues with screening, and MRI still has limitations despite recent improvements. You also have issues when you take the biopsies and when you try to analyse them,” says Claude. “So this is a bleak picture for the diagnosis of prostate cancer right now.”

What’s also bleak is New Zealand’s current prostate cancer rate of more than 100 cases per 100,000 people. Amongst the world’s highest, it means that almost 4,000 are diagnosed each year and more than 700 die from the disease.

In his role as a senior research fellow in the Department of Physics, and a Dodd-Walls Centre principal investigator, Claude has spent the past seven years developing a diagnostic tool that could revolutionise future prostate treatment by using light to accurately identify cancer cells.

“Our goal is to demonstrate an ability to detect cancer and to classify benign and cancerous tissue – and also to grade the cancer in vivo and provide feedback to clinicians in real time during procedures.”

Initially funded by an MBIE Smart Ideas grant in 2018, which helped to develop instruments and conduct pilot studies on animal tissue, a clinical trial was then conducted on fresh prostate biopsies removed from a cohort of 180 patients.

Using Raman spectroscopy, which involves the use of scattered light from a laser beam, researchers were able to “interrogate the tissue”, as Claude puts it, and obtain biomedical information which was then compared with pathology reports to create a classification system.

“We just shine a laser light on the tissue and use our detector to measure the subtle changes in the properties of light, says Claude.

And because the measurements contain information linked to molecules in the tissue, such as DNA, RNA, proteins and lipids, any changes in the signal can be used to create an optical ‘biopsy’ of healthy and cancerous tissue. “We have direct access to what we call the molecular fingerprint of the tissue.”

Over time, Claude’s classification model is being ‘trained’ to predict future measurements, and the research has now progressed to a new clinical trial in late 2025 when measurements will be taken in vivo just before biopsy cores are taken.

“Our instrument will predict benign or cancerous tissue, and we will compare our prediction with what the pathologist says to improve our ability to diagnose cancer.” 

The next step in 2026 will be another clinical trial at the time of surgery when clinicians will be provided with a laser light probe in the form of a ‘pen’ to measure the surgical margin and ultimately help determine how much tissue should be removed.

“The goal is to test our instrument and demonstrate its potential for in-surgery use. The clinician will use our device to read the surgical margin at precise locations, and then take biopsies from these locations to compare the results and continue training our system.”

Beyond that, the plan is to organise multi-site clinical trials at different locations in New Zealand and overseas to show the strength of the classification system and further demonstrate scientific validity.

The ultimate goal is to develop a new hand-held tool that is capable of analysing tissue very accurately in real-time to address existing gaps at the time of diagnosis and surgery – and the practical implications of Claude’s research are wide ranging.

Once someone has presented with persistently high prostate specific antigen (PSA) blood tests, which he says has led to “a huge amount of overtreatment in the past few decades”, they normally progress to an MRI scan – if one is available close to where they live.

“The MRI is good but has limitations nonetheless. The MRI will miss a significant amount of the smaller cancer clusters, and it can still miss 15 to 20% of aggressive cancers.”

Biopsies come next, which Claude says are “not a fun moment” because the procedure is invasive and involves the taking of up to 14 cores which represent a large amount of prostate tissue. There’s also the risk of subsequent infection as well as urinary and rectal dysfunction.

Despite the number of biopsies, the cancer could still be missed, and there is also potential for the misclassification of cancer in understaffed pathology labs. “There is a big bottleneck worldwide in the pathology lab, and an overwhelming amount of tissue having to be analysed.”

By using a laser probe that’s only 800 microns (0.8mm) thick to minimise invasiveness, Claude’s vision is to reduce the need for biopsies to perhaps only two or three if clinicians are informed in real time whether cells are benign or cancerous.

“There is a significant reduction in the trauma to the prostate,” says Claude. “Furthermore, fewer biopsies would lead to dramatic financial savings for the health sector and a significant step improvement in the standard of care for the patient. That’s how we see it.”

When it comes to prostate removal, real-time analysis of surgical margins and the detection of any remaining cancerous tissue would help to improve outcomes. One possibility is to communicate with the clinician using a traffic light system, to identify what tissue should be removed, or is safe to remove, near nerves or other important functional structures.

“You will have a significant improvement in the standard of care, because 30 percent of men today have a positive margin – have cancer left. We can reduce that to five percent. So it’s a very compelling improvement.”

Along the way, Claude is also involved in a separate study in partnership with a team at the Auckland Bioengineering Institute which is using AI to more accurately analyse data from MRI prostate scans – something that could work in parallel with the optical biopsies to further minimise the risk of misdiagnosis.

“Can we use the information from the MRI and feed this into the classification model of the Raman spectroscopy to try to obtain even better classification, even better recognition of what is healthy or cancerous?”

In addition to two MBIE Smart Ideas grants, the project has also been supported by the Dodd-Walls Centre and MedTech whose $80,000 RAP Stage II grant has gone toward instrument development and improving the team’s understanding of the regulatory environment, particularly in the United States.

“We must understand those landscapes for commercialisation, and shape our studies to ensure they are FDA compliant and avoid having to redo them later on.”

To fund future development, Claude has established a company called Probentis Limited with the help of Uniservices and the ‘Return on Science’ commercialisation programme and is about to venture into the sometimes daunting world of capital raising.

Last but not least, he’s grateful for the ongoing support that he’s had from academic colleagues and staff at the Manukau Super Clinic, where the clinical trials have taken place, and Middlemore Hospital pathology lab.

“For this research to happen, you need to have a strong clinical network with people in hospitals who are willing to work on this project and help provide the expertise that you need.”

Researchers from the University of Auckland testing the new photonics sensing solution in a clinical setting.

The new biosensing technology provides real-time tissue analysis and cancer detection to guide clinicians during surgical procedures.