In today’s world, imaging has become highly sophisticated. The use of imaging intelligence is changing the way radiologists look at scans by providing a wealth of information beyond what meets the eye, paving the way toward better diagnosis by processing a wealth of data faster and more efficiently than ever.
“At the end of the day, making a diagnosis requires knowledge of patterns or disease processes,” said Jonathan Goldin, professor of radiology at the University of California, Santa Monica.
“And what we can do now is have an image, have the computer look at the image with us, and go into big data sets and recognize patterns that have looked like that in vast numbers of other cases, so improving our accuracy of diagnosis.”
UCLA Medical Center in Santa Monica uses standard acquisition technologies but adds its own algorithms to analyze images quickly, identifying which may be normal or abnormal, and triaging the images in the work flow, he said.
Intelligent Imaging Suites
Elsewhere, intelligent imaging suites are becoming commonly used in breast MRI, for screening the lung disease COPD, and even for creating 3D and 4D images of organs like the heart prior to surgery.
Some leading medical manufacturers in this arena include Vital — a division of Toshiba — which sells its Vitrea Advanced visualization software to enable visualization and analysis of 2D, 3D and 4D images of CT, MRI, PET scans and ultrasound data “contribute to fast, confident analysis and improved patient outcomes,” according to its website.
A second product, VitreaView Universal Viewer “provides images, reports and other key radiologic data across the medical enterprise, supporting a longitudinal and holistic view of the patient’s care record and helping care teams formulate and monitor treatment plans.”
Among Hitachi’s offerings are the Hitachi Content Platform, which offers data storage and management solutions to help deal with a growing volume of unstructured data. The platform works in concert with Pixcelldata Collibio, a web application that helps manage, store, protect and retrieve digitalized pathology images and data.
Displaying up to Eight Image Series
Siemens seeks to address the problem of managing an increasingly large and diverse number of images — and integrating treatment planning — with its syngo.via RT Image Suite.
By allowing the concurrent display of up to eight image series on up to two monitors, the system provides radiation oncologists “with a clear and comprehensive view of their patients, clinicians are empowered with a solution that results in easier and more intuitive clinical decision-making,” it said on its web site.
In September, Royal Philips received US Food and Drug Administration (FDA) clearance for its Spectral Diagnostic Suite (SpDS), a set of advanced visualization and analysis tools designed for the Philips IQon Spectral CT.
SpDS promises “a new level of flexibility and clinical information for CT users.” A key feature is capturing spectral information in each scan, without planning or set-up in advance, so that clinicians can spare the patient from returning for another scan and “analyze the spectral data in any image retrospectively, using a variety of spectral viewing tools”. The goal is to “achieve better clinical decision support without any added complexity of special modes or workstations that disrupt user workflow.”
The package includes enhanced cardiac analysis, vessel analysis and tumor tracking.
Next Step: Artificial Intelligence
While Goldin declined to comment specifically on any of the latest technologies on offer, he said the promise of the field as a whole is to use artificial intelligence to complement human abilities.
“What they lack at the moment is full intelligence, but what they can do better than a human is they can do repetitive tasks, without fatiguing, they will always do the same thing each time, they are very reproducible and they can quantitate a lot better,” said Goldin.
The art side of medicine — the ability to bring together a wealth of other knowledge bases — is something only doctors can do, he said.
“I think we are on the upslope now of a major transformation of the way imaging is handled,” said Goldin.
“We are going to move away from the fact that a picture represented patterns that perceptually we could classify as being likely to be a type of disease, to automating those processes so that many more images can be processed in a more efficient way.”