A study published in the American Journal of Roentgenology, showed computer-aided detection (CAD) technology improves sensitivity in breast MRI for experienced and novice readers, but does not further accuracy in general.
There was also no real difference in the time it took experienced readers (9) and novice readers (11) from inferring breast MR images with or without the use of CAD.
On average, a breast MRI exam can have as many as 1,000 images before and after insertion of complementary media.
Since then CAD software has been developed to help radiologists read and interpret scans more effectively.
The 20 radiologists completed two reading sittings two months apart to avoid any recall bias.
Each session consisted of reading half of 70 cases with CAD and the other half without.
Of the 70 breast MRI cases, 27 had been determined as benign, while 43 malignant.
The images were purposefully chosen in order to gain a "representative real-world set of images on which to test the utility of the CAD system," said PhD, from the University of Washington and Seattle Cancer Care Alliance, Dr. Constance Lehman.
The assortment and order of MR images were done at random again to lessen the chances of bias amongst readers.
Color overlays and CAD tools were utilized at the finesse of the radiologist, while non-CAD readings incorporated visual assessment.
Each radiologist was asked to classify any suspicious gash and determine the propensity of malignancy.
Readers were also required to turn-in a cumulative assessment of each case as a whole, focusing on the most suspect gashes found.
Lehman and peers used a modified BI-RADS scale to gauge readers’ individual analysis, in part with a five-point possibility of malignancy scale and a percentage possibility of malignancy scale.
Readers’ efficiency was measured using ‘area under the receiver operator characteristics curve’ (AUC), while accuracy was contrasted by the use or non-use of CAD and the various experience levels.
While comparing results, CAD was found to improve sensitivity for both experienced (AUC of 0.91 with CAD; 0.84 without CAD) and novice readers (AUC of 0.83 with CAD; 0.77 without CAD); with sensitivity impressively higher for experienced readers (p=0.01), according to the study.
For diagnostic accuracy, AUC for novices was 0.77 without CAD and 0.79 with CAD.
For experienced readers, AUC was 0.80 without CAD and 0.83 with CAD.
And despite a gradual rise, readers felt the differences were not statistically significant.
"As the use of breast MRI with CAD increases, more attention to the potential contributions of CAD to the diagnostic accuracy of MRI is needed, and additional reader studies are indicated, even if larger reader studies will be needed to detect changes and improvements in reading accuracy,” said Lehman.
From the study, Lehman concluded that much needs to go toward developing educational programs on CAD to improve both overall diagnostic precision and effectiveness of breast MRI interpretation.
