Whether in China, Europe or the United States, clinicians are feeling the pressure of lung cancer as a major public health concern. Newer, enterprise versions of lung CAD software are making it easier than ever to apply the technology to chest x-rays and improve early detection.
Daqing Ma, MD, professor, Department of Radiology, Beijing Friendship Hospital, Capital Medical University and head of Chest Radiology for the Chinese Radiology Society, has been using IQQA-Chest V1.2 from EDDA Technologies since 2005. The software was a stand-alone workstation version with DICOM communication with both the facility’s digital and computed radiography systems. Early in 2007, IQQA-Chest V2.0, the enterprise version, was installed. Now, the x-ray lung CAD is available on any PACS workstation. Beijing Friendship was the first hospital to adopt this enterprise-wide x-ray lung CAD in China.
The hospital has about 4,000 outpatient cases a day and about 8,000 screening cases a year, says Ma, resulting in a heavy workload for the radiologists. The IQQA-Chest system is used on screening and routine outpatient cases to help in the detection of small lung nodules less than 2 centimeters (cm).
“The progress of CAD going from standalone to enterprise is significant,” says Ma. With the enterprise version, his facility now has CAD fully incorporated into its reading workflow and into PACS softcopy reading. “This is important,” he says. “One may do clinical research with a standalone workstation, but to have CAD as part of the routine, enterprise availability is a must.”
Marco Das, MD, of the Department of Radiology at University Hospital in Aachen, Germany, agrees. Before installing a lung CAD server last fall, chest CT examinations had to be specifically sent to the CAD workstation and the reading radiologist had to switch workstations as well. “That’s the reason we only [used CAD] on selected cases,” he says. “The workflow was not too convenient.”
Das uses syngo Lung CAD from Siemens Medical Solutions and was one of the first sites to use the prototype of the software about six years ago. Now, his facility performs about 30 cases a day, primarily on lung cancer screening cases. Secondarily, he uses the software to look for metastases on oncology cases. With the ability to access CAD right from PACS, they use the software for routine scanning of every patient.
Hamilton Hospital in Webster City, Iowa, has been using RapidScreen, Riverain Medical’s x-ray based CAD system, since last April, says Radiology Administrator Matt McKinney. The company made the software and computer server available on a lease basis, so McKinney said the hospital jumped at the opportunity to integrate the relatively cutting-edge technology without a large capital outlay. “Usually technology like this starts in large institutions and filters its way down,” he says. “A lot of times, it’s toward the end of that spectrum that we can use a newer technology.”
McKinney says the hospital had been using mammography CAD so they were familiar with how CAD technology can be a benefit for the radiologist. Using Riverain’s criteria, the hospital does not perform chest x-rays on patients solely to use CAD for lung cancer screening, but rather on those patients who are already having a chest x-ray for a clinical reason. Hamilton Hospital has one on-site main radiologist and the group practice covers the hospital the rest of the time. The main radiologist “was pretty much on board after reading the literature,” says McKinney. “He was excited to provide this to our patients.”
And getting going with the software was quick and easy, he says. “It was a short in-service with the staff and the radiologist on what to expect and how to proceed. From start to finish, it was no more than a one-day process.” The radiologist need only add about 30 seconds to his reading time. After reviewing the x-ray, he then looks at an identical copy of the same image run through the CAD process and then spends a few more seconds reviewing any regions of interest. “He views that as valuable time,” says McKinney.
Barriers & benefits
Adding CAD into the workflow helps cover the limitations of chest x-ray. “Chest x-ray is the most common imaging procedure, but nodule detection from chest x-ray has challenges due to the normal structure overlap, heavy workload, different primary imaging reason, and new radiologists’ lack of experience,” says Ma. “Our results showed that both experienced and less experienced radiologists could benefit from lung CAD, although the less experienced had a greater benefit. For small nodules picked up at an early stage because of the use of IQQA-Chest, and later confirmed on CT and followed through to have a positive pathology report, patient prognosis changes.”
Ma does caution that CAD brings a learning curve. “Some [physicians] get used to it faster than others. To implement, one needs to use it consistently. Organized training and sharing experience also are important to help make the whole process go faster and smoother.”
There are barriers to CAD becoming the standard of practice for lung cancer, says Ma. “Education, economic considerations, continued proof of clinical advantages and continued development of the technology all come into play, as with any new technology that’s emerging into the mainstream.”
However, “we also look forward to the continued development of CAD technology,” he says. “To have CAD tools for more lung diseases can be a direction. If the technology is extended to work on, for example, silicosis, it will be very helpful in our work.”
With cancer the leading cause of death in China, Ma says that early detection is a key to the solution. “In China, the combination of DR/CR and advanced CAD solutions for DR/CR offers an economic and effective alternative as the first tier of lung cancer screening and early lung cancer detection.”
Ma says that radiologists in this digital era need to acknowledge that they need computer tools to assist in better and more efficient practice in imaging. And while software and other tools at radiologists’ disposal can provide many benefits, CAD should never be considered a replacement for a radiologist’s interpretation of a study.
“As long as the software is not perfect, we should always use it as a second reader,” Das recommends. “We need to have our own experience on lung cancer and nodules and combine that with the additional results from the CAD.”