Photoacoustic model tackles bias in medical sensors
Researchers in the UK are using a combination of ultrasound and light to improve the diagnostic accuracy of medical sensors for a wide range of conditions, including cancers.
The researchers at the University of Cambridge’s Cavendish Laboratory have developed a model that improves the accuracy of photoacoustic sensors and are developing devices that go beyond the visible spectrum, using different wavelengths of light to enable earlier detection of cancer.
One of the key projects involves photoacoustic imaging that combines light and ultrasound. Normal ultrasound scanners send sound into the body and measure the echo. Photoacoustic imaging systems, however, measure the ultrasound generated when molecules absorb light and heat up.
“This is a useful feature for looking at cancer,” said Thomas Else, a research associate in the team led by Prof Sarah Bohndiek. “Cancers tend to have lower oxygen levels than healthy tissue because they grow so quickly.”
Most widely-used methods for detecting and monitoring cancer – such as MRI and X-ray mammography machines – are bulky and expensive pieces of equipment, and in the case of mammography uses ionising radiation and applies painful compression.
“The lack of ionising radiation makes photoacoustic imaging safe and suitable for long-term monitoring, and it can produce high-contrast images at the bedside, like an ultrasound,” said Else. “Photoacoustic systems are also lower cost and more portable than MRI machines, which makes it a valuable cancer monitoring technique in a wider range of medical settings.”
Photoacoustic imaging is already CE-marked in Europe and FDA-approved in the United States for breast cancer detection which could be a cost-effective alternative to traditional mammograms.
However, one of the challenges with the application of photoacoustic imaging, like several light-based medical technologies, is that it does not perform as well for people with darker skin tones. Melanin, the main pigment that affects skin tone, absorbs light, meaning less light can get through the skin to make a measurement.
Typically, photoacoustic imaging uses light from the red end of the visible spectrum to the infrared – at wavelengths around 700 to 900nm. In this range, haemoglobin in the blood is the main source of contrast. But in darker-skinned patients, melanin can overwhelm the haemoglobin signal.
“Melanin in the skin blocks some light and prevents it from getting deep into the tissue and to the cancer or the organ we want to image,” said Else. “We wanted to look at the physics behind the underperformance of light-based imaging technologies for people of colour, which could help us address it.”
The team used a computer model of skin with varying levels of pigmentation to simulate photoacoustic images. They validated the results in mice with different pigmentation patterns and in gel models, called phantoms, that mimic different skin tones.
“We found two main effects: one was if you try to measure blood oxygenation – which is important in cancer – in people with dark skin, you get incorrect measurements,” said Else. “Blood oxygenation in Black patients is overestimated compared to white patients. You also get lower quality images, with noise and distortions in the image, that could lead a clinician to misinterpret the scan, potentially making a tumour look more aggressive than it really is.”
To address these discrepancies, the researchers are conducting further studies with volunteers at Addenbrooke’s Hospital in collaboration with the Alliance for Cancer Early Detection to refine the technology and ensure its efficacy across diverse skin tones.
Bohndiek’s team is also exploring the use of photoacoustics for tailoring radiotherapy. By using photoacoustics to understand how cancers respond to treatment, clinicians might be able to customise radiation doses more effectively.
The researchers are also exploring other light-based technologies, like pulse oximeters and smartwatches, which have shown biases in blood oxygenation readings for Black patients. They aim to develop wearable devices that provide accurate measurements regardless of skin tone.
By tackling these biases head-on, the researchers believe they can create better and more inclusive medical devices that not only improve cancer detection and monitoring but also enhance the overall quality and accessibility of healthcare technology.
“Although we’re working on developing and improving technologies, our focus remains the same: we want to improve cancer detection and monitoring, so that every patient is given the best possible chance,” said Else.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
