Date: Sunday, November 8, 2015
Session Type: ACR Poster Session A
Session Time: 9:00AM-11:00AM
Cine-loop analysis is an ultrasound technology that allows the acquisition of images in a digital form as a sequence of a determined frame number. The images can be reviewed as many time as necessary in order to identify the one with the highest level of information. Up to now, in musculoskeletal sonography, this was the prefered techniqueused to select the best Doppler frame (i.e. the image with the highest level of Doppler signal). The frame is subsequently analyzed in quantitative way ( the proportion of total colorized pixels from the total number of existing pixels in a definite region). In order to compare the quantitative Doppler data from different time points a very precise selection of the best Doppler image from a recorded cine-loop is needed. Our study examined the ability of both experienced and in-training sonographers to select the best Doppler image from such a cine-loop using a computerized analytical system as gold standard.
The study analised 480 frame selections done in two distinct exercises as following: twenty cine-loops have been recorded using standardized scans; one region of interest (ROI) was marked on each scan and the colorized pixels / total pixels ratio (CTRatio) was computed for each frame of the loop by using an independent software (Retina®). From each loop a number of 4 different frames have been extracted: the one with the highest CTR (named CTR100) and 3 with 5%, 10% and respectively 20% lower level of CTR (named CTR95, CTR90 and CTR80) . All frames emerged from the same loop have been randomly presented to six experienced and six in training sonographers; the request was to select the frame with the highest CTR from each package (twenty packages in total). A similar exercise was performed with CTRs decreasing in steps of 2% (CTR100, CTR98, CTR96, CTR94).
In the first exercise the CTR100 have been correctly identified in 79.1% cases while in 15.8% of cases CTR95 have been wrongly indicated as CTR100 and in 4.6% of cases theCTR90 was selected. In the 2ndexercise the subjects indicated as CTR100 the following frames: CTR100 in 67% of CTR98 in 18.3%, cases, CTR 96 in 11,6% and CTR94 in 3.1% of cases.
The experienced sonographers performed better than in=training colleagues in both exercises (correct selection in 95 to 75% of cases vs. 90 to 50% of cases in 1st exercise, respectively 90 to 65% of cases vs 75 to 45% of cases in the 2nd exercise).
Using the computerized analyses as the gold standard, we demonstrated a large heterogeneity across sonographers regarding their ability to identify the best Doppler image (i.e. the one with the highest CTR) even from a small group of frames. The ability is even more questionable when the differences are small. When the quantitative Doppler signal is assessed in the same patient at different timepoints this lack of ability might induce a significant error. When computerized analyses is feasible it should be used in place of the manual selection.
To cite this abstract in AMA style:Berghea F, Vlad V, Palanciuc L, Bojinca V, Vreju F, Enache L, Copotoiu M, Kosevoi A, Serban T, Stanciu D, ABOBULUI M, Borangiu A, Balanescu AR, Ionescu R. When Compared with a Computerized System Both Experienced and in-Training Sonographers Have Difficulties to Select the Best Doppler Image from a Cine-Loop [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/when-compared-with-a-computerized-system-both-experienced-and-in-training-sonographers-have-difficulties-to-select-the-best-doppler-image-from-a-cine-loop/. Accessed May 31, 2020.
« Back to 2015 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/when-compared-with-a-computerized-system-both-experienced-and-in-training-sonographers-have-difficulties-to-select-the-best-doppler-image-from-a-cine-loop/