. 2021 Sep 26;11(10):1771. doi: 10.3390/diagnostics11101771

Table 1.

Summarized key statistical parameters for PP and CC algorithms with/without curve fitting.

Imaging Mode, Sampling Rate	Algorithm	Without Curve Fit			With GV Curve Fit			p Value ^†
Imaging Mode, Sampling Rate	Algorithm	Bias [Lower Limit, Upper Limit] (mL/min)	95% CI \|MQE\| (%)	Corr.* (r)	Bias [Lower Limit, Upper Limit] (mL/min)	95% CI \|MQE\| (%)	Corr.* (r)	p Value ^†
DSA acquisition, 6 F/s	PP	−176 [−842, 490]	[29, 43]	0.574	−246 [−676, 184]	[21, 32]	0.773	0.175 ^‡
DSA acquisition, 6 F/s	CC	−234 [−691, 223]	[25, 35]	0.731	−145 [−484, 195]	[18, 25]	0.865	0.017 ^§
Fluoroscopic, 10 P/s	PP	−374 [−1098, 351]	[42, 67]	0.545	−339 [−891, 213]	[34, 51]	0.515	0.027 ^‡
Fluoroscopic, 10 P/s	CC	−205 [−930, 519]	[43, 59]	0.534	−198 [−722, 327]	[30, 44]	0.576	0.031 ^§

* Pearson’s correlation coefficient (r) of calculated flow rate vs. in-line flow sensor measurement. ^† Statistical test of significance (Student’s t test) between algorithms (peak-to-peak vs. cross-correlation; ^‡ without curve fitting, ^§ with gamma-variate curve fitting, respectively).