Ioannis Delakis PhD, MSc, Presenter: Nothing to Disclose

Robert Wise, Abstract Co-Author: Nothing to Disclose

Eugenia Kulama MSc, Abstract Co-Author: Nothing to Disclose

Donald Mcrobbie, Abstract Co-Author: Nothing to Disclose

We compared the Hotelling observer SNR against the CDMAM technique. to evaluate the performance of FFDM detectors. Results showed that the Hotelling observer SNR methodology is more consistent and can be more representative of the system's performance characteristics.

Image quality evaluation plays an important role in ensuring and enhancing the diagnostic value of mammography studies. According to EUREF (www.euref.org), image quality in mammography is assessed using images acquired with the Contrast Detail mammography (CDMAM) phantom (Artinis medical systems, Netherlands). However, CDMAM analysis can suffer from intra-observer variations, limited statistics and structural differences between CDMAM phantoms. The purpose of this work was to evaluate detector performance for a range of full field digital mammography systems using Hotelling observer SNR analysis and ascertain whether it can be an alternative to CDMAM evaluation.

FFDM units used in West of London Breast Screening were evaluated. Detector performance was first analysed using CDMAM phantom methodology. As part of Hotelling observer SNR analysis, the generalised normalised noise power spectrum (GNNPS) was measured by collecting flat field images of a 5cm PMMA phantom. The generalised modulation transfer function (GMTF) was measured by placing a 0.2mm Tungsten edge in the middle of the PMMA phantom. This setup allowed for scatter and focal spot unsharpness to be incorporated in the measurements. The Hotelling observer SNR was calculated for input signal originating from gold discs of varying thicknesses and diameters.

The Hotelling SNR values were used to estimate the threshold gold thicknesses for each diameter as per CDMAM analysis. The Hotelling SNR technique was more consistent than CDMAM results. There were small differences between the two techniques, especially in small diameter details, which can be attributed to structural characteristics of the CDMAM, as confirmed by previous comparative work from our group. Overall, the Hotelling SNR technique showed variations in the performance of FFDM detectors, demonstrating the use of this metric as a differentiator.

Delakis, I, Wise, R, Kulama, E, Mcrobbie, D, Comparative Performance Evaluation of Contrast-detail in Full Field Digital Mammography Systems Using Hotelling Observer Signal to Noise Ratio versus Automated CDMAM Image Analysis.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13010275.html