Skip to main content
. 2023 Aug 4;13:1213068. doi: 10.3389/fonc.2023.1213068

Table 2.

Mean time needed to correct AI-generated contours and time saving compared to manual contouring, for three patients and five different AI contouring solutions.

Mean time
Manual
Institution
Correction
Mirada
Correction
MVision
Correction
Radformation
Correction
RayStation
Correction
Therapanacea
Breast No. structures 10 8 8 10 5 10
Time for 10 structures [min] 22 7.5 1.6 7.8 3.1 1.4
Saving [min/%] 14.5/66.0% 20.4/92.8% 14.2/64.4% 18.9/86.0% 20.6/93.7%
Head and neck No. structures 19 27 27 27 26 30
Time for 19 structures [min] 97 8.2 9.8 22.7 4.6 4.4
Saving [min/%] 88.8/91.6% 87.2/89.9% 74.3/76.6% 92.4/95.3% 92.6/95.4%
Lung No. structures 6 6 6 6 5 6
Time for 6 structures [min] 26 5.2 1.2 6.0 1.5 0.4
Saving [min/%] 20.8/80.1% 24.9/95.6% 20.0/76.8% 24.5/94.4% 25.6/98.4%
Prostate No. structures 10 8 9 9 5 10
Time for 10 structures [min] 42 7.4 0.3 4.3 5.2 0.1
Saving [min/%] 34.6/82.3% 41.7/99.3% 37.7/89.7% 36.8/87.6% 41.9/99.7%

The saving refers to the time saved computed by finding the mean time correct each contour, multiplied by the number of contours drawn by the manual.