A UK first for deep learning in medical imaging. Kheiron receives CE marking for its breast cancer screening machine learning software
Landmark step for UK health tech industry expected to enable faster and more accurate breast cancer detection
Kheiron Medical Technologies, a UK start-up, has received CE marking for its new deep learning-based breast cancer screening software. Kheiron is the first UK company to receive the CE mark for deep learning software in radiology, which enables this cutting-edge technology to be launched within the UK’s National Health Service (NHS) and European healthcare systems.
In a clinical study, Kheiron’s regulatory approved software has demonstrated indications of performance above the average national benchmarks for breast screening radiologists. Application for FDA approval of Kheiron’s software is underway. CE regulatory approval means that health care providers in Europe can use Kheiron’s deep learning software as a second reader of mammographic images in a breast cancer screening setting. Clinicians are able to receive results within seconds, directly into their existing workflows, incorporating case-wise recall decision support and lesion localisation.
Deep learning uses powerful artificial neural networks and high performance computing to analyse complex medical images with tremendous precision. With the approval of Kheiron’s software for use as a second reader, the screening workload of overstretched clinical staff in all screening settings is potentially reduced, allowing them to focus on other more complex modalities and tasks. For the NHS, Kheiron’s software presents a new technology tool that could benefit the global ongoing radiology workforce shortage, starting with the NHS Breast Screening Programme.
In addition, CE marking allows for intelligent triaging of imaging studies prior to review, enabling radiologists to prioritise studies based on the algorithm’s findings. Together, the approved software is aimed at facilitating more efficient and accurate breast cancer detection at an earlier stage of the disease and reducing the number of women incorrectly subjected to invasive biopsies, unnecessary radiation and detecting more cancers at the crucial early stages.
Kheiron’s approach to clinical validation for CE marking has set a high bar. The company commissioned an independent multi-center clinical study to evaluate their software’s performance prior to submission. This retrospective study demonstrated that the software surpassed recognised standards for single radiologist reporting accuracy. Kheiron’s focus on deep clinical rigor means that it is already launching further multi-phase clinical evaluations across the UK, Europe and US to further assess and improve performance, and the potential impact on breast screening outcomes.
“Regulatory approval for Kheiron comes just as the new Health Secretary announces his vision for a more technologically driven NHS,” said Kheiron’s CEO, Dr Peter Kecskemethy. “As a UK SME we welcome the forward-thinking of our healthcare leadership and are excited to be working closely with both clinicians and service providers to ensure that patients receive the benefits of this new wave of technology.”
Kheiron’s Clinical Director and Consultant Radiologist, Dr Hugh Harvey said “CE marking of our deep learning technology is a huge step for the health tech industry, and demonstrates that the UK can really lead globally on diagnostic software in medicine. Of course, introducing deep learning tools into clinical practice is not just about passing through these regulations, and that’s why we are conducting multiple phases of clinical trials as well as change management projects and health economics analyses to ensure this cutting-edge technology is integrated in an ethical, safe and meaningful way within the existing breast screening programme so that both women and clinical staff can receive the maximum benefit.“
“These are exciting times for the breast imaging community and in particular breast screening,” said Dr Nisha Sharma, Consultant Breast Radiologist, Director of Breast Screening & Clinical Lead for Breast Imaging at Leeds University Hospitals Trust, and Secretary for the British Society of Breast Imaging. “Software such as this will have a significant impact in managing the global workforce shortage in mammography by augmenting single reading and ensuring that the sensitivity and specificity of double reading are at least maintained and potentially improved. This would reinforce the aim that deep learning tools are there to support the clinician and work alongside them.”
Kheiron’s Chief Medical Officer, Dr Christopher Austin said “With UK radiology workforce shortages in the spotlight, regulatory approval could actually position the UK and Europe as the world’s earliest adopters of AI at scale. European national breast screening programmes can now take the lead and demonstrate how to leverage high quality, proven, and safe machine learning technologies to run screening programmes better, smarter and more cost-effectively for millions of women.”
“As someone who’s both looked after patients with breast cancer and experienced it personally, I know that the one thing every woman is worried about is whether they will need chemotherapy,” said Consultant Breast Surgeon and Breast Cancer Patient, Dr Liz O’Riordan. “The aim of breast screening is to pick up cancers years before a woman might feel a lump, and therefore avoid treatments like chemo. Currently, women wait up to two weeks to get the results of their mammogram, and the time spent waiting –the ‘scanxiety’ they feel– can be harrowing to say the least. If Kheiron can help accurately speed up that process, and pick up cancers early, it would make a dramatic impact on improving the screening patient experience.”