Software for Comparing Changes in Pulmonary Nodule Size on Chest X-Rays
Posted Jan 26 2012 12:00am
Most healthcare consumers are familiar with the new digital imaging modalities such as CT, MRI, and PET scans. However, fewer may be aware of the image analysis software that is in common use for analyzing digital images in radiology. A recent article was fascinating for me in terms of cost-saving and quality opportunities that are being made available by such software, even for the common chest x-rays (see: FDA Approves New Chest X-Ray Scanning Software ). Below is an excerpt from the article:
Software that could lead to low-cost early detection of lung cancer won US Food and Drug Administration (FDA) clearance this month....Development of an inexpensive tool for early diagnosis could have a tremendous effect on cancer screening and disease survival, said [a radiation oncologist]. The new Temporal Comparison software helps radiologists better detect changes in lung tissue by improving comparisons between a new chest X-ray and one taken previously, according to a company news release. The software superimposes the new image on the old one and highlights areas of change, using pattern-recognition and machine-learning algorithms. Normally, radiologists draw conclusions about suspicious tissue changes by conducting side-by-side X-ray comparisons....Detection scores of solitary nodules improved 12.4% with the new software...However, widespread application of CT scanning [which are deemed to be superior to chest x-ray screening] has been stymied by the costs of the testing, which insurance companies are reluctant to reimburse without a prior diagnosis....[The Temporal Comparison] software would be "an order of magnitude" cheaper than CT scanning. However, ...there is a trade-off, because a CT scan provides a 3-dimensional view that an X-ray cannot....[A] single server with the new software could serve all of a facility's X-ray equipment. [The x-ray scans] go directly to the server, which searches the facility's picture archive and communication [PACS] system for prior scans. If a prior scan exists, the [two] X-rays are aligned and the difference image goes into the patient file....."If you could use a standard chest X-ray and some new algorithm to diagnose patients earlier, that's a gold mine for the patients," [an expert in the field] concluded.
There are several aspects of this news that are interesting. First of all, the Temporal Comparison software now takes on the burden of some of the professional responsibility of the radiologist. As noted above, "the software superimposes the new image [of a nodule in the lungs] on the old [x-ray image of a nodule] and highlights areas of change, using pattern-recognition and machine-learning algorithms." The comparison of a current image with priors for a patient has been the responsibility of the interpreting radiologist. Secondly, this new software seems to give new life and significance to garden-variety chest x-rays that are deemed inferior to CT scans but for which insurance companies have been reluctant to reimburse absent a prior diagnosis. This must be one of the reasons that the FDA has provided clearance for this new software. It would seem to result in better and more cost effective care.
Finally, the availability of this software makes a compelling argument for the development of larger and larger radiology image archives, beginning with all hospitals within a health system and moving to the cloud-sharing of radiology images across referring hospitals in a region. Here's an excerpt from an article that points in this direction (see: Memorial Hermann to Employ Cutting-Edge Radiology Image Management and Sharing Platform ):
In a move that expands its technological capabilities and further differentiates Memorial Hermann as an innovative leader in the greater Houston area, the hospital system has implemented a cutting-edge, cloud-based medical image sharing platform that vastly enhances image management, distribution and data exchanges between referring physicians and hospitals in southeast Texas.... Powered by DICOM Grid , provider of a cloud-based platform for medical imaging management applications, the technology greatly increases the quality of care, patient safety and business continuity between Memorial Hermann’s hospitals, Imaging Centers, affiliated physicians and other providers in the community. Memorial Hermann’s enterprise-wide technology initiative improves service and promotes tighter integration with facilities that refer patients to its Level One trauma center and physicians that utilize any of the 28 Memorial Hermann Imaging Centers. “DICOM Grid eliminates the need for CDs and traditional radiology film because authorized facilities can now easily and instantly transmit images to our Level One trauma center or receive radiology studies from our Imaging Centers, including MRI and CT scans, via the system,” said [the CIO] at Memorial Hermann.