Skip to main content
. 2022 Jun 29;12:898771. doi: 10.3389/fonc.2022.898771

Table 1.

Prediction accuracy with matching uncertainties based on 4Hz motion waveforms.

Patient Diaphragm motion (mm) 4DMRI libraries AR accuracy (mm) b LSTM accuracy (mm) b
Inhale phases Exhale phases Prediction Matched Prediction Matched
AVG a STD t (s) AVG a STD AVG a STD t (s) AVG a STD
1 15.8 89 122 0.24 0.20 0.44 0.30 0.20 0.42 0.28 1.47 0.47 0.32
2 12.9 97 118 0.26 0.20 0.44 0.31 0.17 0.59 0.55 1.46 0.62 0.58
3 20.7 108 110 0.44 0.36 0.44 0.47 0.36 0.88 0.70 1.50 0.93 0.70
4 14.4 108 111 0.45 0.32 0.44 0.48 0.35 0.61 0.39 1.49 0.64 0.38
5 9.6 101 117 0.25 0.19 0.44 0.31 0.18 0.24 0.21 1.45 0.27 0.21
6 15.8 114 102 0.33 0.28 0.44 0.47 0.49 0.64 0.61 1.47 0.65 0.63
7 11.6 95 122 0.42 0.24 0.44 0.42 0.26 0.75 0.48 1.45 0.78 0.51
8 29.1 92 125 0.47 0.49 0.44 0.63 0.49 0.81 1.34 1.45 0.97 1.24
AVG 16.2 100.5 115.9 0.36 0.29 0.44 0.42 0.31 0.62 0.57 1.47 0.67 0.57
STD 6.2 8.8 7.7 0.10 0.10 0.00 0.11 0.13 0.21 0.35 0.02 0.23 0.32

The AR (autoregression) and LSTM (long short-term memory) prediction accuracy and performance, diaphragm motion, two TR-4DMRI libraries in the inhalation and exhalation, and the library matching errors are provided. The AR method provides more accurate prediction results using less time than that of the LSTM method.

ap-value <0.02 for both prediction and matched accuracy between AR and LSTM.

bNo significant difference between prediction and matched accuracy using either AR or LSTM.