Skip to main content
EJNMMI Physics

Table 3 The automatic thyroid segmentation outcomes according to SPECT input conditions in addition to the attenuation map (μ-map) input (n = 36)

From: CT-free quantitative SPECT for automatic evaluation of %thyroid uptake based on deep-learning

Metric

μ-map + pg

μ-map + psh

μ-map + ACSCRRf

μ-map only

DSCc

VDd (mL)

95% HDe (mm)

DSC

VD (mL)

95% HD (mm)

DSC

VD (mL)

95% HD (mm)

DSC

VD (mL)

95% HD (mm)

Mean

0.767

 − 0.720

9.416

0.768

2.345

10.030

0.769

1.423

10.441

0.727

 − 5.239

10.050

STDa

0.072

6.661

3.604

0.077

6.952

3.624

0.075

7.363

5.530

0.067

8.067

2.916

SERb

0.012

1.110

0.607

0.013

1.159

0.604

0.012

1.227

0.922

0.011

1.344

0.486

  1. aStandard deviation
  2. bStandard error
  3. cDice similarity coefficient
  4. dVolume difference
  5. eHausdorff distance
  6. fQuantitative SPECT with attenuation correction, scatter correction, and resolution recovery
  7. gPrimary emission SPECT
  8. hPrimary emission SPECT + scattering SPECT
Back to article page