Skip to main content
. 2021 Nov 22;23(2):e13470. doi: 10.1002/acm2.13470

TABLE 3.

Interobserver variation of radiation oncologists (ROs) and the comparison with deep learning (DL) results

dCTV2 pCTV1
Case 1 Case 2 Case 3 Case 4 Case 1 Case 2 Case 3 Case 4 Case 5 Case 6
DSC DL 0.767 0.686 0.732 0.837 0.830 0.867 0.826 0.871 0.885 0.870
RO1 0.474 0.474 0.498 0.529 0.811 0.840 0.794 0.867 0.858 0.840
RO2 0.525 0.429 0.491 0.522 0.746 0.815 0.768 0.853 0.873 0.846
RO3 0.695 0.741 0.662 0.751 0.765 0.834 0.812 0.821 0.820 0.824
MSD (mm) DL 1.586 2.878 1.867 0.934 0.928 1.008 1.487 1.110 0.833 1.039
RO1 3.982 4.632 3.019 3.592 1.090 1.180 1.392 1.080 0.959 1.263
RO2 5.520 7.127 6.001 4.286 1.523 1.416 1.472 1.074 0.854 1.115
RO3 1.792 1.805 2.190 1.411 1.404 1.231 1.502 1.366 1.301 1.396

Abbreviations: dCTV, clinical tumor volume for definitive radiotherapy; DSC, dice similarity coefficient; MSD, mean surface distance; pCTV, clinical tumor volume for postoperative radiotherapy.