Table 1 Voxel-level dose rate errors (mean ± std) of static patient images (n = 80) and their differences to the ground truth were statistically analyzed with paired t-test.

From: Deep-dose: a voxel dose estimation method using deep convolutional neural network for personalized internal dosimetry

\({\bf{M}}{\bf{e}}{\bf{a}}{\bf{n}}\,{\bf{v}}{\bf{o}}{\bf{x}}{\bf{e}}{\bf{l}}\,{\textstyle \text{-}}\,{\bf{l}}{\bf{e}}{\bf{v}}{\bf{e}}{\bf{l}}\,{\bf{d}}{\bf{o}}{\bf{s}}{\bf{e}}\,{\rm{ \% }}\,{\bf{d}}{\bf{i}}{\bf{f}}{\bf{f}}{\bf{e}}{\bf{r}}{\bf{e}}{\bf{n}}{\bf{c}}{\bf{e}}=\,{{\boldsymbol{\sum }}}_{1}^{{\boldsymbol{n}}}|\frac{{\boldsymbol{M}}{\boldsymbol{e}}{\boldsymbol{t}}{\boldsymbol{h}}{\boldsymbol{o}}{{\boldsymbol{d}}}_{{\boldsymbol{n}}}^{\ast }-{\boldsymbol{D}}{\boldsymbol{M}}{{\boldsymbol{C}}}_{{\boldsymbol{n}}}^{{\rm{\S }}}}{{\boldsymbol{D}}{\boldsymbol{M}}{{\boldsymbol{C}}}_{{\boldsymbol{n}}}}|/{\boldsymbol{n}}\times {\bf{100}}\)

Method

VSV

Deep-dose

Mean ± Std (%)

Mean ± Std (%)

Gallbladder wall

1.38 ± 1.37

2.13 ± 1.70

NS

NS

Heart wall

5.83 ± 2.16

1.88 ± 1.27

p < 0.005

NS

Kidney

1.44 ± 1.14

0.86 ± 0.67

NS

NS

Liver

1.13 ± 0.68

1.10 ± 0.67

NS

NS

Lung

57.09 ± 2.07

1.41 ± 1.29

p < 0.001

p < 0.05

Pancreas

1.05 ± 0.98

1.47 ± 1.21

NS

NS

Spleen

0.94 ± 0.68

0.85 ± 0.61

p < 0.01

NS

Stomach wall

4.01 ± 2.02

2.01 ± 1.61

p < 0.05

NS

Whole-body (avg)

9.97 ± 1.79

2.54 ± 2.09

  1. The degree of significance is reported in the below of the corresponding differences (NS = nonsignificant).
  2. *Method: VSV or Deep-dose (CNN). §DMC: direct Monte Carlo simulation.