How an iPhone attachment is helping streamline cancer referrals

9th January 2025

Dr Chris Coulson, ENT consultant at Queen Elizabeth Hospital Birmingham, UK, and CEO of Endoscope-i, talks to Beth Gault about how a new iPhone-compatible device being piloted in the West Midlands aims to streamline throat cancer referrals.

Beth Gault (BG): What is Endoscope-I and how did you first come up with the idea?

Chris Coulson (CC): Me and another ENT surgeon selfishly tried to solve our problems. We spend a lot of time looking in people’s ears, noses and throats and we generally use an endoscope for that, which you put to your eye to look through and see the area of concern.

About 12 years ago, we recognised we had a decent camera in our pocket, on our phone, and thought why don’t we connect it with the endoscope. So, we did. We created an adapter to align the endoscope with the phone to allow you to take an image or video of the ear, nose or throat, which is what the device is.

We progressed to thinking we could train someone to take a video, which then comes through to me and if the clinician has filled out a history and done a hearing test, then I could just sit down and look at them remotely. So, we put that system together.

ENT is an interesting field in that only 10% of people who see me in clinic need an operation, 90% of people – with appropriate information – can be managed remotely. That’s classically done through a face-to-face consultation, but we wondered how much of that information can I get through other means?

BG: How does the tool work?

CC: You just attach the aligner, which is a bit like a phone case, with the endoscope and the iPhone and then take a video or image and upload it to our cloud system via our referral app. It’s currently only on iPhone, not Android, but the iPhones are included with the service.

BG: The tool is now being used to try and detect and rule out throat cancer – what role can it play in cancer referrals?

CC: The symptoms that are concerning with head and neck cancer are also symptoms that sound like a cold, like a hoarse voice or a sore throat. They are general, non-specific symptoms. Most people with a two-week wait referral for cancer do not have cancer.

The number of people picked up from an urgent cancer referral clinic is tiny – about 3%.  So, the trouble that us head and neck surgeons have is how do we sift through these huge numbers of referrals to get to people who actually have cancer?

Capacity is a huge problem, so the number of referrals go up and up and up and actually the numbers of cancers are not. But you still have to deal with these referrals because you never know who is going to have the cancer.

We set up this referral pathway in Stoke where anyone with low risk for cancer went to a nurse-led clinic. The nurse used the endoscope with our camera system, took a video of their throat and then sent it to the head and neck surgeons. They were able to have a look and say who did not have cancer, and then who needed a further appointment in the clinic.

It’s really quick for us to look and know whether they do not have cancer, because it’s what we do every day. Within seconds you can look at a video and tell, but in a consultation, it takes 20 minutes.

We found through this that we can safely run a service where the nurse sees the vast majority of patients, but the consultant reports on it so it’s a consultant-level answer. 

BG: What funding have you received for this?

CC: We designed and delivered it ourselves without funding. But we then got funding from the Small Business Research Initiative to run an assessment of its delivery. We did this in Dorset County Hospital and Stoke University Hospital for head and neck cancer.

BG: What were the results of the test?

CC: We don’t have impressive numbers that say we’ve caught all these cancers, because our aim is not to find cancer, it’s to deal with the people who don’t have cancer. If you deal with them, you can have more capacity to deal with the cancer patients.

We put 1,800 people through the service and found around six cancers. All of these were low risk, so they’d likely go on to have a late diagnosis if they hadn’t used the service.

BG: If a system was interested, would integrated care boards (ICBs) or Trusts be the one to fund this?

CC: Up to now it’s been Trust funded, because it’s a way of increasing their capacity and saving money. But we do have a couple of ICBs around the country who are looking at it for delivery in their patch. My guess is in a few years we will have a combination of ICBs and hospitals. However, the one thing we won’t have is GPs funding this. The location of delivery could be in primary care, which I think is advantageous to the population, but I don’t think it will be primary care-led funding.

BG: What does it cost?

CC: Around £100,000 upfront for the kit, training and integration and then there’s a monthly license fee for security updates and ongoing support.

BG: How do you see the tool being used in the future in the NHS?

CC: I don’t see GPs using it, it’s one of the core things we’re trying to do is to make things easier for GPs, not delivering something else for them to do. It would work best with someone doing it in high volumes, rather than intermittent users, so it might be that trained staff will go out to practices or any healthcare facility to do it. It would work well in a diagnostic treatment centre as well.

We are also looking at artificial intelligence and if at some stage we can automate some diagnoses or letters, this would further increase efficiency.

BG: How do you think digital tools like this can make a difference to the NHS?

CC: There’s clearly a huge opportunity for digital tech to improve healthcare. But one of the challenges is how can you make sure that your digital tech doesn’t make things worse. We’ve got to be cautious that we don’t make things more complex.

But it’s also got to be clinically led. As soon as there’s tech designed by tech people with tech people in mind, most of us struggle to get on with it. It needs to solve a clinical problem.

This article was originally published by our sister publication Healthcare Leader.

Smartphone screening doubles detection rate of atrial fibrillation

5th September 2022

Atrial fibrillation detection with a smartphone-based screening app more than doubles the rate of detection compared to symptom-based, usual care according to the results of a randomised trial by researchers based in Munich, Germany.

AF is the most common cardiac arrhythmia and in 2019 was estimated to have a global burden of 59.7 million. The condition is diagnosed with an electrocardiogram tracing showing AF and while many patients experience symptoms including chest pain, palpitations, shortness of breath, and fatigue, others have no symptoms, a condition referred to as asymptomatic or “silent” AF.

The presence of AF increases the risk of a stroke and one meta-analysis found that AF might be newly detected in nearly a quarter of patients with stroke or transient ischaemic attack, highlighting the need for early detection.

Although smart device-based photoplethysmography technology could be a feasible approach for AF screening, the extent to which such methods can increase detection compared to usual methods is uncertain.

In the present study, the German team undertook an open-label, randomised trial which compared digital screening (using the smartphone app) or usual care. Eligible patients were all free of AF at baseline and if there were any abnormalities detected via the app (based on pulse irregularities), these were verified using external ECG loop recorders.

The study was conducted over a 6-month period and the primary efficacy endpoint was the first diagnosis of treatment-relevant AF and which led to the initiation of oral anticoagulant therapy. The secondary endpoints were newly diagnosed AF, stroke and newly prescribed oral anticoagulants. After the first 6 months, participants crossed over for a second phase of the study with reverse assignment.

Atrial fibrillation diagnoses

A total of 5,551 participants with a mean age of 65 years (31% female) were randomly assigned to digital monitoring (2,860) or usual care. Among the whole cohort, 15% had coronary heart disease and 6% a history of stroke.

Within the first six months, the primary endpoint was reached by 1.33% of those in the smartphone group compared to 0.63% in the usual care arm (odds ratio, OR = 2.12, 95% CI 1.19 – 3.76, p = 0.010).

A total of 4,752 individuals for whom the primary endpoint was not reached, crossed over for a second six-month period. As with the first phase, the proportion achieving the primary endpoint was 1.38% vs 0.51% (OR = 2.75, 95% CI 1.42 – 5.34, p = 0.003).

There were also significant differences for the secondary endpoints of newly diagnosed AF (1.68% vs 0.89%, p = 0.011), newly prescribed anticoagulants (1.71% vs 0.85%, p = 0.006) but not for stroke (p = 0.95).

The authors concluded that smartphone-based screening increases the detection of AF compared to routine-symptom-based screening and suggested that future studies should focus on whether digital screening for AF leads to better treatment outcomes.

Citation
Rizas KD et al. Smartphone-based screening for atrial fibrillation: a pragmatic randomized clinical trial Nat Med 2022.