Digitizing radiology: how technology is shaping a bright future for diagnostic imaging
Many medical specialities lay claim to being at the heart of healthcare, but few occupy the pivotal position of radiology.
Radiology’s influence reaches deep, informing and inspiring the diagnoses that generate positive outcomes for patients and streamline workflows for healthcare systems.
The scale is immense. In England alone, 41.4 million medical imaging tests were reported in the 12 months from December 2016 to November 2017. The capacity to save digital information has doubled every 40 months since the 1980s and Dr Amit Kharat, writing in the Indian Journal of Radiology and Imaging in 2017, observed that big data is a technology “whose time has come”.
He is not alone in believing that radiology can contribute even more to clinical decisions as technology advances. But navigating oceans of data presents challenges as well as opportunities.
Radiology departments are feeling the strain across Europe and around the world. They need sophisticated systems to support their expertise, collate digital information and ensure they synchronize with other departments along the various treatment pathways. The pressure is on to provide accurate and timely diagnoses, a theme reinforced by the UK’s Department of Health’s Getting It Right First Time project (GIRFT), which aims to save the NHS £1.5bn per year.
Projects like GRIFT are symptomatic of a desire to raise awareness of radiology and what it involves, particularly among patients.
“Diagnostic imaging and intervention are integral to virtually all patient pathways,” says Dr Caroline Rubin, a consultant radiologist at the University Hospital Southampton NHS Foundation Trust in the UK, with special interests in education and training. “We use it for diagnosis and monitoring courses of treatment and to change it depending on how patients respond.
Radiologists are an unsung part of healthcare; not enough people know what radiologists and radiographers do even though the vast majority of people will have had an image taken in their life.
The big challenge is pure volume of work. The number of scans go up 12% a year whereas the number of radiologists goes up 3% a year so we are not keeping pace.
A new radiology revolution
Technology can help radiology forge another golden era – similar to 1970s when the arrival of the CT scanner opened up new opportunities – based on clinically-validated informatics that accelerates the speed of diagnoses while adhering to strict safety protocols.
Artificial Intelligence (AI) and machine learning can create intelligent routes for data to follow that enhance decision-making, reduce variation and establish efficiencies tailored to a particular workflow. Systems can also link with patient and imaging records to deliver enriched diagnostic content for busy radiologists.
“AI or machine learning can be hugely useful in providing measurements needed in monitoring cancer treatment so that radiologists don’t have to take the measurements every time,” adds Dr Rubin, a breast screening and imaging specialist who is also a vice-president of the Royal College of Radiologists.
The promise is that new software tools will support radiologists in the current tasks and add a layer of connectivity to enhance decision-making. Algorithms, for instance, can be trained to detect anomalies even before scans are reported and therefore accelerate diagnosis and treatment in certain cases. Intelligently-designed systems can also take the repeat measurements needed in aspects such as treatment monitoring to free up clinician time.
Big data, big opportunity
The growth of technology in radiology and the potential of big data provide the discipline with the ammunition to meet future challenges, according to Professor Guy Frija, a past president of the European Society of Radiology(ESR).
Applying AI to medical imaging can generate huge improvements in the quality of care and speed of diagnosis, and comes with the potential to create systems that integrate across healthcare disciplines, giving radiologists and radiographers intelligent support in their daily workflow.
Many radiologists and radiographers are excited by the potential of integrated AI and machine learning to support their clinical decision-making, particularly as work demands are increasing.
“We cannot work without technology,” says Dr Naleem Dugar, a consultant radiologist at Doncaster and Bassetlaw Hospitals NHS Trust in the UK. “All our images are digital and we use voice recognition to help with the volume of reports we produce, so we are already using AI and are not afraid of it.
“The algorithms are getting better and there is no doubt that machines will pick up irregularities. They might even be better at picking up a lung nodule in a 1,000-slice CT scan because they don’t get fatigued, whereas a radiologist’s eyes may get fatigued by the end of a day.”
Dr Dugar, who is a former chair of the Royal College of Radiologists’ Imaging Informatics Group, added:
AI can play a crucial part and we should be excited by its potential, but we need to retain the perspective that it is a support tool for clinical judgement – which is the gold standard.
Returning to Dr Kharat’s study in the Indian Journal of Radiology and Imaging, there’s great potential for AI to improve the quality of performed scans, assist radiologists in decision support and act as a virtual quality control tool. According to Dr Kharat, over a period of time AI can “self-learn to find hidden information within the reports and images which are rather difficult to interconnect, or find a relationship using the standard routine or conventional protocols…In the near future, big data will work to assist radiologists by providing intelligent and targeted decision support, rather than replacing radiologists.”
It was less than a generation ago that radiologists had to ‘wet process’ X-ray images, passing film through developing and fixing baths before analysis. Now they, and patients, have the energizing prospect of intelligent technology that can perform tasks at lightning pace and develop alongside the treatment – improving speed, efficiency and clinical outcomes.