̽��ֱ�� of Cambridge - diagnostic

Handheld device could transform heart disease screening

sc604 — Tue, 08 Apr 2025 08:10:47 +0000

̽��ֱ��researchers, from the ̽��ֱ�� of Cambridge, developed a device that makes it easy for people with or without medical training to record heart sounds accurately. Unlike a stethoscope, the device works well even if it’s not placed precisely on the chest: its larger, flexible sensing area helps capture clearer heart sounds than traditional stethoscopes.

̽��ֱ��device can also be used over clothing, making it more comfortable for patients – especially women – during routine check-ups or community heart health screening programmes.

̽��ֱ��heart sound recordings can be saved on the device, which can then be used to detect signs of heart valve disease. ̽��ֱ��researchers are also developing a machine learning algorithm which can detect signs of valve disease automatically. ̽��ֱ��results are reported in the IEEE Journal of Biomedical and Health Informatics.

Heart valve disease (valvular heart disease or VHD) has been called the ‘next cardiac epidemic,’ with a prognosis worse than many forms of cancer. Up to 50% of patients with significant VHD remain undiagnosed, and many patients only see their doctor when the disease has advanced and they are experiencing significant complications.

In the UK, the NHS and NICE have identified early detection of heart valve disease as a key goal, both to improve quality of life for patients, and to decrease costs.

An examination with a stethoscope, or auscultation, is the way that most diagnoses of heart valve disease are made. However, just 38% of patients who present to their GP with symptoms of valve disease receive an examination with a stethoscope.

“ ̽��ֱ��symptoms of VHD can be easily confused with certain respiratory conditions, which is why so many patients don’t receive a stethoscope examination,” said Professor Anurag Agarwal from Cambridge’s Department of Engineering, who led the research. “However, the accuracy of stethoscope examination for diagnosing heart valve disease is fairly poor, and it requires a GP to conduct the examination.”

In addition, a stethoscope examination requires patients to partially undress, which is both time consuming in short GP appointments, and can be uncomfortable for patients, particularly for female patients in routine screening programmes.

̽��ֱ��‘gold standard’ for diagnosing heart valve disease is an echocardiogram, but this can only be done in a hospital and NHS waiting lists are extremely long – between six to nine months at many hospitals.

“To help get waiting lists down, and to make sure we’re diagnosing heart valve disease early enough that simple interventions can improve quality of life, we wanted to develop an alternative to a stethoscope that is easy to use as a screening tool,” said Agarwal.

Agarwal and his colleagues have developed a handheld device, about the diameter of a drinks coaster, that could be a solution. Their device can be used by any health professional to accurately record heart sounds, and can be used over clothes.

While a regular or electronic stethoscope has a single sensor, the Cambridge-developed device has six, meaning it is easier for the doctor or nurse – or even someone without any medical training – to get an accurate reading, simply because the surface area is so much bigger.

̽��ֱ��device contains materials that can transmit vibration so that it can be used over clothes, which is particularly important when conducting community screening programmes to protect patient privacy. Between each of the six sensors is a gel that absorbs vibration, so the sensors don’t interfere with each other.

̽��ֱ��researchers tested the device on healthy participants with different body shapes and sizes and recorded their heart sounds. Their next steps will be to test the device in a clinical setting on a variety of patients, against results from an echocardiogram.

In parallel with the development of the device, the researchers have developed a machine learning algorithm that can use the recorded heart sounds to detect signs of valve disease automatically. Early tests of the algorithm suggest that it outperforms GPs in detecting heart valve disease.

“If successful, this device could become an affordable and scalable solution for heart health screening, especially in areas with limited medical resources,” said Agarwal.

̽��ֱ��researchers say that the device could be a useful tool to triage patients who are waiting for an echocardiogram, so that those with signs of valve disease can be seen in a hospital sooner.

A patent has been filed on the device by Cambridge Enterprise, the ̽��ֱ��’s commercialisation arm. Anurag Agarwal is a Fellow of Emmanuel College, Cambridge.

Reference:
Andrew McDonald et al. ‘A flexible multi-sensor device enabling handheld sensing of heart sounds by untrained users.’ IEEE Journal of Biomedical and Health Informatics (2025). DOI: 10.1109/JBHI.2025.3551882

Researchers have developed a handheld device that could potentially replace stethoscopes as a tool for detecting certain types of heart disease.

This device could become an affordable and scalable solution for heart health screening, especially in areas with limited medical resources

Anurag Agarwal

Acoustics Lab

Person demonstrating use of a handheld device for heart disease screening

̽��ֱ��text in this work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Images, including our videos, are Copyright © ̽��ֱ�� of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Yes

Cambridge researchers develop urine test for early detection of lung cancer

Anonymous — Fri, 06 Dec 2024 05:49:55 +0000

Researchers hope that early detection, through the simple urine test, could enable earlier treatment interventions, significantly improving patient outcomes and prognosis. Around 36,600 lives are saved from lung cancer in the UK every year, according to new analysis from Cancer Research UK.

Professor Ljiljana Fruk and Dr Daniel Munoz Espin and their teams at the ̽��ֱ�� of Cambridge are leading on the research, funded by Cancer Research UK.

̽��ֱ��work, at Cambridge’s Department of Chemical Engineering and Biotechnology, and the Early Cancer Institute, will provide a cheap, affordable sensor that uses urine samples to help doctors detect lung cancer before the disease develops.

Lung cancer has a poor prognosis for many patients because often there are no noticeable symptoms until it has spread through the lungs or into other parts of the body. ̽��ֱ��new urine test will allow doctors to spot the disease before it develops.

To create the test, scientists looked at proteins excreted by senescent cells: “zombie” cells which are alive but unable to grow and divide. It’s these cells that cause tissue damage by reprogramming their immediate environment to help promote the emergence of cancer cells.

Now, researchers have developed an injectable sensor that interacts with zombie cell proteins and releases easily detectable compound into urine, signalling their presence.

“Early detection of cancer requires cost-effective tools and strategies that enable detection to happen quickly and accurately,” said Fruk. “We designed a test based on peptide-cleaving proteins, which are found at higher levels in the presence of zombie cells, and in turn appear in the early stages of cancer.

“Ultimately, we want to develop a urine test that could help doctors identify signs of the early stages of cancer – potentially months or even years before noticeable symptoms appear.”

As well as targeting lung cancer, Fruk hopes her research, along with joint efforts across other university departments, will result in the development of probes capable of detecting other cancers.

“We have almost completed a functional urine test to detect ‘zombie' cells in lung cancer, which will spot cancer earlier and avoid the need for invasive procedures, but this test does have potential for other cancers,” she said. “Developing more efficient cancer treatments requires earlier detection and better therapies, but also work with other disciplines for a more holistic view of the disease, which is an essential part of my research.”

From uncovering the causes of lung cancer to pioneering drugs to treat it, Cancer Research UK has helped power progress for people affected by lung cancer. Over the last 10 years, the charity has invested over £231 million in lung cancer research.

“Cancer Research UK has played a key role in advancing lung cancer research and improving survival,” said Dr Iain Foulkes, Cancer Research UK’s executive director of research and innovation. “This project being led by Professor Fruk is another example of our commitment to driving progress so that more people can live longer, better lives, free from the fear of cancer.”

Adapted from a Cancer Research UK media release.

Cambridge scientists have developed a urine test for early detection of lung cancer. ̽��ֱ��test, the first of its kind, detects ‘zombie’ cells that could indicate the first signs of the disease.

koto_feja via Getty Images

Close-up of cancer cells

Yes

Deep learning and disease detection

skbf2 — Thu, 07 Apr 2022 08:58:53 +0000

Marcel Gehrung, co-founder and CEO of rapidly growing Cambridge biotech company, Cyted, on revolutionising disease diagnostics and the challenges of building a business during a pandemic.

Personalised blood test can detect persistent lung cancer

cjb250 — Thu, 17 Mar 2022 13:00:14 +0000

Scientists at the Cancer Research UK Cambridge Institute used a personalised blood test for patients, which is a type of liquid biopsy that can pick up tiny fragments of DNA that are released into the blood as tumours grow. This DNA, called circulating tumour DNA (ctDNA), can reveal the state of the tumour, its location and potentially its weaknesses, which could be used to select the best treatments.

̽��ֱ��results from the Lung Cancer Circulating Tumour DNA (LUCID-DNA) study, funded by Cancer Research UK, have been published today in the Annals of Oncology.

Many patients who are treated for early-stage non-small cell lung cancer can be cured with either surgery, radiotherapy or sometimes (chemo)radiotherapy.

After treatment, lung cancer patients are carefully followed up with tests including CT scans to find out if the treatment has removed the tumour, but scans won’t pick up tiny quantities of cancer cells known as minimal residual disease (MRD) which could still regrow into further tumours.

By finding signs that lung cancer cells might still be present and active after treatment, using methods such a liquid biopsy, doctors might be able make better choices about treating patients, aiming to improve the chances of survival for patients who are at higher risk while reducing side effects for patients who are at a lower risk group.

̽��ֱ��LUCID-DNA study aimed to find out if circulating DNA can be detected in early stage lung cancers. It used a liquid biopsy, called RaDaR™, which analyses up to 48 different mutations that are unique to each patient’s tumour. RaDaR™ was developed by Inivata, a biotech company co-founded by Dr Rosenfeld, and is based on technologies developed initially by his lab at the Cancer Research UK Cambridge Institute.

Dr Nitzan Rosenfeld, group leader at the Cancer Research UK Cambridge Institute, Chief Scientific Officer of Inivata and co-lead author of the study, said: “If cancer cells remain in the body after treatment a tumour can regrow. If that happens, it is a big setback for patients and the doctors treating them.

“Liquid biopsy can be used to detect tiny amounts of residual cancer after treatment, flagging those patients who have signs that their tumour may not have been eradicated completely with treatment. We’re hoping that this technology could help doctors decide when additional rounds of treatment are needed, and could save lives.”

To find out if liquid biopsy could find lung cancer patients with MRD, the LUCID-DNA study team enrolled 88 patients who were treated at Royal Papworth Hospital and Addenbrooke’s Hospital for early stage non-small cell lung cancer. This type of cancer accounts for over 85% of all lung cancer cases.

̽��ֱ��research team extracted DNA from tumour samples provided by the patients and sequenced the DNA to find combinations of mutations unique to each patient’s lung cancer. Using this genetic “fingerprint”, Inivata created a blood test which was personalised to the patient’s tumour.

̽��ֱ��liquid biopsies were then used to detect tumour DNA in blood samples collected before treatment, and for up to 9 months after treatment. ̽��ֱ��researchers found that patients who had tumour DNA present between two weeks and four months after treatment were much more likely to have their lung cancer come back or to die from it.

Professor Robert Rintoul, Professor of Thoracic Oncology at the ̽��ֱ�� of Cambridge, Honorary Respiratory Physician at Royal Papworth Hospital and co-lead author of the study, said: “We need to study these liquid biopsies further to find the best ways to deploy them, but these results clearly show that they can potentially be an effective tool to help decide which patients need further treatment.

“Being able to offer patients personalised monitoring and treatment will ultimately save more lives and help us to beat cancer sooner.”

Lung cancer is the third most common type of cancer in the UK. Every year, around 48,500 people are diagnosed with lung cancer and every year around 35,100 people die from the disease in the UK.

Aart Alders participated in the LUCID-DNA observational clinical study at Royal Papworth Hospital following his lung cancer surgery.

“I was first diagnosed with an early-stage lung cancer about five years ago and underwent a surgical operation to remove it,” he said. “Although some people need further treatment with chemotherapy, I have been very fortunate and my original lung cancer has not returned.

“I was very pleased to be able to help with the LUCID-DNA research study. By trying to develop a blood test to help doctors predict whether a lung cancer might come back or not, we will increase the chance of curing more people.”

Michelle Mitchell, Chief Executive of Cancer Research UK, said: “Lung cancer is one of the biggest killers in the UK. ̽��ֱ��earlier it is caught, the more likely it is to be treated successfully.

“Detecting signs of cancer before or after treatment without the need for invasive surgery has huge potential for both patients and doctors.

“I look forward to seeing more research that will develop liquid biopsy further, which will and ultimately make it much easier for doctors to offer treatment that best matches the patient’s needs, increasing their chance of survival.”

Adapted from a press release by Cancer Research UK

Patients who are at a higher risk of their lung cancer returning can be identified by a personalised blood test that is performed after treatment, according to researchers at the ̽��ֱ�� of Cambridge.

Liquid biopsy can be used to detect tiny amounts of residual cancer after treatment, flagging those patients who have signs that their tumour may not have been eradicated completely with treatment

Nitzan Rosenfeld

Belova59

Blood test

̽��ֱ��text in this work is licensed under a Creative Commons Attribution 4.0 International License. Images, including our videos, are Copyright © ̽��ֱ�� of Cambridge and licensors/contributors as identified. All rights reserved. We make our image and video content available in a number of ways – as here, on our main website under its Terms and conditions, and on a range of channels including social media that permit your use and sharing of our content under their respective Terms.

Yes

Licence type:

Public Domain

Scientists can detect brain tumours using a simple urine or blood plasma test

Anonymous — Fri, 23 Jul 2021 15:05:30 +0000

̽��ֱ��team say that a test for detecting glioma using urine is the first of its kind in the world.

Although the research, published in EMBO Molecular Medicine, is in its early stages and only a small number of patients were analysed, the team say their results are promising.

̽��ֱ��researchers suggest that in the future, these tests could be used by GPs to monitor patients at high risk of brain tumours, which may be more convenient than having an MRI every three months, which is the standard method.

When people have a brain tumour removed, the likelihood of it returning can be high, so they are monitored with an MRI scan every three months, which is followed by biopsy.

Blood tests for detecting different cancer types are a major focus of research for teams across the world, and there are some in use in the clinic. These tests are mainly based on finding mutated DNA, shed by tumour cells when they die, known as cell-free DNA (cfDNA).

However, detecting brain tumour cfDNA in the blood has historically been difficult because of the blood-brain-barrier, which separates blood from the cerebrospinal fluid (CSF) that surrounds the brain and spinal cord, preventing the passage of cells and other particles, such as cfDNA.

Researchers have previously looked at detecting cfDNA in CSF, but the spinal taps needed to obtain it can be dangerous for people with brain tumours so are not appropriate for patient monitoring.

Scientists have known that cfDNA with similar mutations to the original tumour can be found in blood and other bodily fluids such as urine in very low levels, but the challenge has been developing a test sensitive enough to detect these specific mutations.

̽��ֱ��researchers, led by Dr Florent Mouliere who is based at the Rosenfeld Lab of the Cancer Research UK Cambridge Institute and at the Amsterdam UMC, and Dr Richard Mair, who is based at Cancer Research UK Cambridge Institute and the ̽��ֱ�� of Cambridge developed two approaches in parallel to overcome the challenge of detecting brain tumour cfDNA.

̽��ֱ��first approach works for patients who have previously had glioma removed and biopsied. ̽��ֱ��team designed a tumour-guided sequencing test that was able to look for the mutations found in the tumour tissue within the cfDNA in the patient’s urine, CSF, and blood plasma.

A total of eight patients who had suspected brain tumours based on MRIs were included in this part of the study. Samples were taken at their initial brain tumour biopsies, alongside CSF, blood and urine samples.

By knowing where in the DNA strand to look, the researchers found that it was possible to find mutations even in the tiny amounts of cfDNA found in the blood plasma and urine.

̽��ֱ��test was able to detect cfDNA in 7 out of 8 CSF samples, 10 out of the 12 plasma blood samples and 10 out of the 16 urine samples.

For the second approach the researchers looked for other patterns in the cfDNA that could also indicate the presence of a tumour, without having to identify the mutations.

They analysed 35 samples from glioma patients, 27 people with non-malignant brain disorders, and 26 healthy people. They used whole genome sequencing, where all the cfDNA of the tumour is analysed, not just the mutations.

They found in the blood plasma and urine samples that fragments of cfDNA, which came from patients with brain tumours were different sizes than those from patients with no tumours in CSF. They then fed this data into a machine learning algorithm which was able to successfully differentiate between the urine samples of people with and without glioma.

̽��ֱ��researchers say that while the machine learning test is cheaper and easier, and a tissue biopsy from the tumour is not needed, it is not as sensitive and is less specific than the first tumour-guided sequencing approach.

MRIs are not invasive or expensive, but they do require a trip to the hospital, and the three-month gap between checks can be a regular source of anxiety for patients.

̽��ֱ��researchers suggest that their tests could be used between MRI scans, and could ultimately be able to detect a returning brain tumour earlier.

̽��ֱ��next stage of this research will see the team comparing both tests against MRI scans in a trial with patients with brain tumours who are in remission to see if it can detect if their tumours are coming back at the same time or earlier than the MRI. If the tests prove that they can detect brain tumours earlier than an MRI, then the researchers will look at how they can adapt the tests so they could be offered in the clinic, which could be within the next ten years.

“We believe the tests we’ve developed could in the future be able to detect a returning glioma earlier and improve patient outcomes,” said Mair. “Talking to my patients, I know the three-month scan becomes a focal point for worry. If we could offer a regular blood or urine test, not only will you be picking up recurrence earlier, you can also be doing something positive for the patient’s mental health.”

Michelle Mitchell, Chief Executive of Cancer Research UK said, “While this is early research, it’s opened up the possibility that within the next decade we could be able to detect the presence of a brain tumour with a simple urine or blood test. Liquid biopsies are a huge area of research interest right now because of the opportunities they create for improved patient care and early diagnosis. It’s great to see Cancer Research UK researchers making strides in this important field.”

Sue Humphreys, from Wallsall, a brain tumour patient, said: "If these tests are found to be as accurate as the standard MRI for monitoring brain tumours, it could be life changing.

If patients can be given a regular and simple test by their GP, it may help not only detect a returning brain tumour in its earliest stages, it can also provide the quick reassurance that nothing is going on which is the main problem we all suffer from, the dreaded Scanxiety.

̽��ֱ��problem with three-monthly scans is that these procedures can get disrupted by other things going on, such as what we have seen with the Covid pandemic. As a patient, this causes worry as there is a risk that things may be missed, or delayed, and early intervention is the key to any successful treatment.”

Reference:
Florent Mouliere et al. ‘Fragmentation patterns and personalized sequencing of cell-free DNA in urine and plasma of glioma patients.’ EMBO Molecular Medicine (2021). DOI: 10.15252/emmm.202012881

Adapted from a Cancer Research UK press release.

Researchers from the Cancer Research UK Cambridge Institute have developed two tests that can detect the presence of glioma, a type of brain tumour, in patient urine or blood plasma.

Yes

Biological ‘fingerprints’ of long COVID in blood could lead to diagnostic test, say Cambridge scientists

cjb250 — Mon, 19 Jul 2021 07:39:19 +0000

̽��ֱ��team has received funding from the National Institute for Health Research to develop a test that could complement existing antibody tests. They also aim to use similar biological signatures to develop a test and monitor for long COVID.

While most people recover from COVID-19 in a matter of days or weeks, around one in ten people go on to develop symptoms that can last for several months. This can be the case irrespective of the severity of their COVID-19 – even individuals who were asymptomatic can experience so-called ‘long COVID’.

Diagnosing long COVID can be a challenge, however. A patient with asymptomatic or mild disease may not have taken a PCR test at the time of infection – the gold standard for diagnosing COVID-19 – and so has never had a confirmed diagnosis. Even antibody tests – which look for immune cells produced in response to infection – are estimated to miss around 30% of cases, particularly among those who have had only mild disease and or beyond six months post-initial illness.

A team at the ̽��ֱ�� of Cambridge and Cambridge ̽��ֱ�� Hospitals NHS Foundation Trust has received £370,000 from the National Institute for Health Research to develop a COVID-19 diagnostic test that would complement existing antibody tests and a test that could objectively diagnose and monitor long COVID.

̽��ֱ��research builds on a pilot project supported by the Addenbrooke’s Charitable Trust. ̽��ֱ��team has been recruiting patients from the Long COVID Clinic established in May 2020 at Addenbrooke’s Hospital, part of Cambridge ̽��ֱ�� Hospitals NHS Foundation Trust.

During the pilot, the team recruited 85 patients to the Cambridge NIHR COVID BioResource, which collects blood samples from patients when they are first diagnosed and then at follow-up intervals over several months. They now hope to expand their cohort to 500 patients, recruited from Cambridgeshire and Peterborough.

In their initial findings, the team identified a biomarker – a biological fingerprint – in the blood of patients who had previously had COVID-19. This biomarker is a molecule known as a cytokine produced by T cells in response to infection. As with antibodies, this biomarker persists in the blood for a long time after infection. ̽��ֱ��team plans to publish their results shortly.

Dr Mark Wills from the Department of Medicine at the ̽��ֱ�� of Cambridge, who co-leads the team, said: “We need a reliable and objective way of saying whether someone has had COVID-19. Antibodies are one sign we look for, but not everyone makes a very strong response and this can wane over time and become undetectable.

“We’ve identified a cytokine that is also produced in response to infection by T cells and is likely to be detectable for several months – and potentially years – following infection. We believe this will help us develop a much more reliable diagnostic for those individuals who did not get a diagnosis at the time of infection.”

By following patients for up to 18 months post-infection, the team hopes to address several questions, including whether immunity wanes over time. This will be an important part of helping understand whether people who have been vaccinated will need to receive boosters to keep them protected.

As part of their pilot study, the team also identified a particular biomarker found in patients with long COVID. Their work suggests these patients produce a second type of cytokine, which persists in patients with long COVID compared to those that recover quickly and might be one of the drivers behind the many symptoms that patients experience. This might therefore prove to be useful for diagnosing long COVID.

Dr Nyarie Sithole, also from the Department of Medicine at the ̽��ֱ�� of Cambridge, who co-leads the team and helps to manage long COVID patients, said: “Because we currently have no reliable way of diagnosing long COVID, the uncertainty can cause added stress to people who are experiencing potential symptoms. If we can say to them ‘yes, you have a biomarker and so you have long COVID’, we believe this will help allay some of their fears and anxieties.

“There is anecdotal evidence that patients see an improvement in symptoms of long COVID once they have been vaccinated – something that we have seen in a small number of patients in our clinic. Our study will allow us to see how this biomarker changes over a longer period of time in response to vaccination.”

At the moment, the team is using the tests for research purposes, but by increasing the size of their study cohort and carrying out further work, they hope to adapt and optimise the tests that can be scaled up and speeded up, able to be used by clinical diagnostic labs.

As well as developing a reliable test, the researchers hope their work will help provide an in-depth understanding of how the immune system responds to coronavirus infection – and why it triggers long COVID in some people.

Dr Sithole added: “One of the theories of what’s driving long COVID is that it’s a hyperactive immune response – in other words, the immune system switches on at the initial infection and for some reason never switches off or never goes back to the baseline. As we’ll be following our patients for many months post-infection, we hope to better understand whether this is indeed the case.”

In addition, having a reliable biomarker could help in the development of new treatments against COVID. Clinical trials require an objective measure of whether a drug is effective. Changes in – or the disappearance of – long-COVID-related cytokine biomarkers with corresponding symptom improvement in response to drug treatment would suggest that a treatment intervention is working.

Markers in our blood – ‘fingerprints’ of infection – could help identify individuals who have been infected by SARS-CoV-2, the coronavirus that causes COVID-19, several months after infection even if the individual had only mild symptoms or showed no symptoms at all, say Cambridge researchers.

Because we currently have no reliable way of diagnosing long COVID, the uncertainty can cause added stress to people who are experiencing potential symptoms. If we can say to them ‘yes, you have a biomarker and so you have long COVID’, we believe this will help allay some of their fears and anxieties

Nyarie Sithole

Annie Spratt

Tired looking woman

Yes

Licence type:

Public Domain

Artificial intelligence could be used to triage patients suspected at risk of early-stage oesophageal cancer

cjb250 — Thu, 15 Apr 2021 15:00:49 +0000

When researchers applied the technique to analysing samples obtained using the ‘pill on a string’ diagnostic tool Cytosponge, they found that it was capable of reducing by half pathologists’ workload while matching the accuracy of even experienced pathologists.

Early detection of cancer often leads to better survival because pre-malignant lesions and early stage tumours can be more effectively treated. This is particularly important for oesophageal cancer, the sixth most common cause for cancer-related deaths. Patients usually present at an advanced stage with swallowing difficulties and weight loss. ̽��ֱ��five-year overall survival can be as low as 13%.

One main subtype of oesophageal cancer is preceded by a condition known as Barrett oesophagus, in which cells in the lining of the oesophagus change shape. Barrett oesophagus occurs in patients with Gastro-oesophageal Reflux Disease (GORD), a digestive disorder where acid and bile from the stomach return into the oesophagus leading to heartburn symptoms. In Western countries, 10-15% of the adult population are affected by GORD and are hence at an increased risk of having Barrett oesophagus.

At present Barrett oesophagus can only be detected by a gastroscopy and tissue biopsy. Researchers at the ̽��ֱ�� of Cambridge have developed a far-less invasive diagnostic tool called the Cystosponge – a ‘pill on a string’ that dissolves in the stomach and which, as it is withdrawn, picks up some cells from the lining of the oesophagus. These cells are then stained using a laboratory marker called TFF3 and can then by examined under a microscope.

Now, in a study published today in Nature Medicine, a team at Cambridge has applied deep learning techniques to the sample analysis, stratifying patients into eight triage classes that determine whether a patient sample requires manual review or if automated review would suffice. ̽��ֱ��algorithms were trained using 4,662 pathology slides from 2,331 patients.

Professor Rebecca Fitzgerald from the MRC Cancer Unit at the ̽��ֱ�� of Cambridge, who developed the Cytosponge and worked with the AI team, said: “Any system that supports clinical decisions needs to balance its performance against workload reduction and potential economic impact. Replacing pathologists entirely could lead to substantial workload reduction and speed up diagnoses, but such an approach would only be viable if the performance remains comparable to that of human experts and there are regulatory hurdles to overcome.”

For the analysis of Cytosponge-TFF3 samples, the triaging approach showed several benefits, substantially reducing workload and matching the sensitivity and specificity of experienced pathologists. Sensitivity is the ‘true positive’ rate – that is, how often a test correctly generates a positive result for people who have Barrett oesophagus. Specificity, on the other hand, measures a test’s ability to correctly generate a negative result for people who don’t have the disease.

̽��ֱ��researchers showed that a fully manual review by a pathologist achieves 82% sensitivity and 93% specificity. In a fully automated approach, they observed a sensitivity of 73% and a specificity of 93%. ̽��ֱ��team was able to demonstrate that using a triage-driven approach, up to two-thirds of cases can be reviewed automatically while achieving a sensitivity of 83% and specificity of 93%. ̽��ֱ��team estimates that this approach would reduce workload for the pathologists by 57%.

̽��ֱ��team were able to build into their algorithm problem-solving techniques applied by pathologists familiar with Cytosponge-TFF3 samples. This meant that the algorithms were interpretable – in other words, a clinician would be able to understand why they had reached a particular decision. This is important for accountability.

Dr Florian Markowetz from the CRUK Cambridge Institute, who led the work on the AI algorithm, said: “We’ve shown that it’s possible to use computer-aided tools to streamline identification of people at risk of Barrett oesophagus. By semi-automating the process, we can reduce the workload by more than half while retaining the accuracy of a skilled pathologist. This could potentially speed up the diagnosis of Barrett oesophagus and, potentially, the identification of those individuals at greatest risk of oesophageal cancer.”

̽��ֱ��team say that this triage-driven approach could be applied beyond the Cytosponge to a number of tests for other conditions such as pancreatic cancer, thyroid cancer or salivary gland malignancies.

̽��ֱ��research was supported by Cancer Research UK, the Medical Research Council and Cambridge ̽��ֱ�� Hospitals NHS Foundation Trust.

Reference
Gehrung, M et al. Triage-driven diagnosis of Barrett esophagus for early detection of esophageal adenocarcinoma using deep learning. Nat Med; 15 Apr 2021; DOI: 10.1038/s41591-021-01287-9

Artificial intelligence ‘deep learning’ techniques can be used to triage suspected cases of Barrett oesophagus, a precursor to oesophageal cancer, potentially leading to faster and earlier diagnoses, say researchers at the ̽��ֱ�� of Cambridge.

We’ve shown that it’s possible to use computer-aided tools to streamline identification of people at risk of Barrett oesophagus... This could potentially speed up the diagnosis of Barrett oesophagus and, potentially, the identification of those individuals at greatest risk of oesophageal cancer

Florian Markowetz

Cytosponge

Yes

Safety screens

cjb250 — Wed, 03 Feb 2021 10:52:24 +0000

Since the start of October, a dedicated team drawn from across the ̽��ֱ�� and its Colleges has been running an innovate programme to screen its students for COVID-19. Getting it up and running in time may have been a Herculean task, but its success has been remarkable.