Reveal_PresentationSlides_OUS_cn.ppt

* The differences between two successive ΔRR intervals are calculated and plotted in the Lorenz plot where X = ?RRi = RRi – RRi-1 and y = ?RRi-1 = RRi-1 - RRi-2. These coordinates {x,y} are then plotted on the Lorenz graph in order to build the foot print of the Cluster Signature Metric. Detailed info: The Lorenz Plot is made up of 961 bins, each encompassing a 40 ms-by-40 ms section. The extent of the plot is from -620 ms to +620 ms in both the x- and y-axis direction * * The differences between two successive ΔRR intervals are calculated and plotted in the Lorenz plot where X = ?RRi = RRi – RRi-1 and y = ?RRi-1 = RRi-1 - RRi-2. These coordinates {x,y} are then plotted on the Lorenz graph in order to build the foot print of the Cluster Signature Metric. Detailed info: The Lorenz Plot is made up of 961 bins, each encompassing a 40 ms-by-40 ms section. The extent of the plot is from -620 ms to +620 ms in both the x- and y-axis direction * The main point in the slide is to explain the performance numbers in the manual Duration Sensitivity = If there was a total of 100 hrs of AF, we will detect 98 hrs (98%) Duration Specificity = If there was a total 100 hrs of sinus rhythm, we will over-detect AF for 3 hrs (97%) Duration PPV = If we detect for 100 hrs, 84 hrs will be truly AF Episode Sensitivity = On an average we will detect 90 of 100 true AF episodes (2 min) Episode PPV = On an average every 6 out of 10 (60%) detected episodes will be true AF PPV = If Reveal XT detects AF, how likely is it to be correct? Data is based on Holter RR intervals (noise issues may further degrade performance) Note that, in a given category, there may be different numbers of patients which contribute to episode and durations totals since 1) a single detected AF episode can span a true AF period and a non-AF period and 2) a single detected AF episode need not span the entire duration of a true AF episode to be counted as a correctly identified episode. 61% of these episodes c