Atri-U: assisted image analysis in routine cardiovascular magnetic resonance volumetry of the left atrium

Table 2 Estimation of time saving derived from the time-saving proportions, minus the average time required for reading Atri-U outputs

	Time required per case (in seconds)	Time saved = manual time \(\) (proportion of data scored as “no need for correction” + 50% \(\) proportion of data scored as “minor corrections needed”)	Proportion of datasets from sample D₂ for each possible score
Score			0	1	2^a
ES frame detection	5	= 5 \(\) 91.3% + 50% \(\) 5 \(*\) 4.7% = 4.7 s	91.3%	4.7%	4.0%
Segmentation	80	= 80 \(\) 82.7% + 50% \(\) 80 \(*\) 11.3% = 70.7 s	82.7%	11.3%	6.0%
Landmark detection	20	= 20 \(\) 90.7% + 50% \(\) 20 \(*\) 5.3% = 18.7 s	90.7%	4.0%	5.3%
Subtotal	105	= 94.1
Minus reading	23	= 94.1–23.0
Total time saving		71.1

^aScore 2 meant “no time saving/correction from scratch”, and therefore is not included in the above calculations; ES, End-systolic

ISSN: 1532-429X