Click here to


Are you sure ?

Yes, do it No, cancel

Novel Analysis of Flawed Transducer Element Detection in Ultrasound Has Advantage Over Existing Method

Z Li*, J Zagzebski , University of Wisconsin, Madison, WI


(Monday, 7/30/2018) 3:45 PM - 4:15 PM

Room: Exhibit Hall | Forum 8

Purpose: An important aspect of Ultrasound Quality Assurance(QA) is performing uniformity assessments (UA) and detecting defective elements for transducers. QA experts typically use B-mode images of tissue phantoms for these assessments. A common approach employs image loops and analysis tools to reduce speckle noise and increase the conspicuity of shadows emanating from defective regions of the transducer. Here we demonstrate a novel way that simplifies this task. It employs an analysis of a single image while the transducer is suspended in air. The sensitivity and specificity for detecting transducer defects is compared to that obtained using a commercially available software that analyzes images of phantoms.

Methods: Our Matlab-based computerized analysis program (MBCAP) analyzes the ringdown area of the air-scan image, computing values vs. lateral position. A “defect cutoff threshold� is taken as 3 standard deviations below the mean values. The detection performance was compared with that of UltraIQ (Cablon Industries) which analyzes phantom image data from a ROI having a 12 mm axial extent in the gray scale image. Air-scan and phantom-scan data were collected during annual QA testing. All transducers were GE linear arrays. Of the eight probes assessed, four were judged defective, with seven defects visually identified on images.

Results: After applying the cutoff threshold, our MBCAP applied to the air-scan images detected all 7 dropout areas from the 4 transducers, with no false positive detection. On the other hand, UiQ detected all of the dropout areas with 4 false detection. Nothing was detected using the ringdown signal for the 4 non-defective transducers, while UiQ has 5 false positives.

Conclusion: The MBCAP method with the air-scan image shows good sensitivity and specificity for QA compared to UiQ. We will enlarge our sample pool by acquiring air-scan data routinely to facilitate this analysis for all probes.

Funding Support, Disclosures, and Conflict of Interest: James Zagzebski serves as a consultant to Gammex/SunNuclear


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email