Concurrent CAD utilization speeds CT colonography interpretations
The use of advanced imaging technologies such as computer-aided detection (CAD) software have enabled better reader performance for the interpretation of CT colonography exams. In addition, adding the software tool may enhance productivity by reducing the time needed to read the study. However, the question of how best to deploy these systems for optimal clinical efficiency remains: as a second read after primary interpretation of the study, or as a concurrent read during the primary interpretation?
A multinational group of researchers from the United Kingdom, Belgium, and the United States prospectively compared the diagnostic performance and time efficiency of both second and concurrent (CAD) reading paradigms for retrospectively obtained CT colonography data sets by using consensus reading by three radiologists of colonoscopic findings as a reference standard. Their findings appear in this month’s issue of Radiology.
“A potentially more time-efficient paradigm is concurrent read CAD, which applies CAD at the start of the assessment,” the authors wrote. “Although an intuitively attractive proposition, there is some evidence from breast and pulmonary literature that concurrent application of CAD reduces observer vigilance, reducing sensitivity.”
A group of 10 radiologists (five experienced CT colonography readers and five less-experienced colleagues) took part in the study, which used Medicsight’s CAD software (Medicolon 1.2) interpreted on Vital Images’ workstations (Vitrea 2). The researchers selected 25 CT colonography data sets for review by the study participants.
“Each participant independently read the 25 studies on two occasions and each reading was separated by a period of 6 weeks to minimize recall bias,” the authors noted. “Readers were encouraged to read the studies in small blocks, mirroring normal clinical practice, rather than attempting all 25 at one sitting.”
One of the two reading paradigms (second-read CAD or concurrent CAD) was employed for the first read of the 25 studies, the other during the second read 6 weeks later. The interpreting radiologists were blinded to the prevalence of abnormalities and were not told they would be rereading the same 25 studies in the second session.
For the second-read CAD portion, the clinicians analyzed each case and recorded their interpretation time. Once their analysis was complete, they applied the CAD software, reviewed the case, and recorded the time for the CAD-assisted review. For the concurrent-CAD portion of the study, the clinicians applied the technology prior to their initial review of the cases so that interpretation was performed with CAD markers on the images. The time for interpretation using this paradigm also was recorded.
The researchers reported that mean reading times for unassisted read, second read CAD, and concurrent read CAD were 11, 15, and 12 minutes, respectively. They found that use of CAD as a second reader added 3.7 minutes to the unassisted read in their study.
“We did not find any definite evidence of a difference between the two CAD reader paradigms in the detection of polyps 6 mm or larger,” the authors wrote. “Interestingly, when small (1–5 mm) polyps were included in the model, the differences between the two CAD paradigms became more marked with odds of detection 30 percent higher, on average, when using CAD as a second read.”
On the basis of their findings, the researchers determined that concurrent application of colonographic CAD software during a single radiologist review is more time efficient and achieves similar sensitivity for polyps 6 mm or larger in comparison with application as a second reader after full unassisted radiologist review
The researchers stated that there is still a role for the use of colon CAD software as an automated second reader, particularly for lesions less than 5 mm.
“However, use of CAD as a second reader maximizes sensitivity, particularly for smaller lesions,” they noted.
A multinational group of researchers from the United Kingdom, Belgium, and the United States prospectively compared the diagnostic performance and time efficiency of both second and concurrent (CAD) reading paradigms for retrospectively obtained CT colonography data sets by using consensus reading by three radiologists of colonoscopic findings as a reference standard. Their findings appear in this month’s issue of Radiology.
“A potentially more time-efficient paradigm is concurrent read CAD, which applies CAD at the start of the assessment,” the authors wrote. “Although an intuitively attractive proposition, there is some evidence from breast and pulmonary literature that concurrent application of CAD reduces observer vigilance, reducing sensitivity.”
A group of 10 radiologists (five experienced CT colonography readers and five less-experienced colleagues) took part in the study, which used Medicsight’s CAD software (Medicolon 1.2) interpreted on Vital Images’ workstations (Vitrea 2). The researchers selected 25 CT colonography data sets for review by the study participants.
“Each participant independently read the 25 studies on two occasions and each reading was separated by a period of 6 weeks to minimize recall bias,” the authors noted. “Readers were encouraged to read the studies in small blocks, mirroring normal clinical practice, rather than attempting all 25 at one sitting.”
One of the two reading paradigms (second-read CAD or concurrent CAD) was employed for the first read of the 25 studies, the other during the second read 6 weeks later. The interpreting radiologists were blinded to the prevalence of abnormalities and were not told they would be rereading the same 25 studies in the second session.
For the second-read CAD portion, the clinicians analyzed each case and recorded their interpretation time. Once their analysis was complete, they applied the CAD software, reviewed the case, and recorded the time for the CAD-assisted review. For the concurrent-CAD portion of the study, the clinicians applied the technology prior to their initial review of the cases so that interpretation was performed with CAD markers on the images. The time for interpretation using this paradigm also was recorded.
The researchers reported that mean reading times for unassisted read, second read CAD, and concurrent read CAD were 11, 15, and 12 minutes, respectively. They found that use of CAD as a second reader added 3.7 minutes to the unassisted read in their study.
 | |
| Method of highlighting obscured polyp candidates by using CAD software during endoluminal 3D scan shows sigmoid polyp (arrow) as barely visible behind a haustral fold (arrowhead). CAD software mark alerts reader to presence of hidden polyp (triangle). Image and caption courtesy of the Radiological Society of North America. |
On the basis of their findings, the researchers determined that concurrent application of colonographic CAD software during a single radiologist review is more time efficient and achieves similar sensitivity for polyps 6 mm or larger in comparison with application as a second reader after full unassisted radiologist review
The researchers stated that there is still a role for the use of colon CAD software as an automated second reader, particularly for lesions less than 5 mm.
“However, use of CAD as a second reader maximizes sensitivity, particularly for smaller lesions,” they noted.