Skip to main content

Table 4 Type of ground truth segmentation

From: Automatic segmentation of prostate zonal anatomy on MRI: a systematic review of the literature

First author, year of publication

Annotation

Annotators

Type

Tool

Qualification

Number

Type of reading

Experience*

Cuocolo et al. [43]

Software

itk-SNAP

(A)

4

Splitted

2 to 5

Bardis et al. [42]

Software

In house

(A)

12

Stratified

10

Lai et al. [41]

Manual

(A)

1

10

Nai et al. [18]

Software

MITK

Medical physicist

4

Stratified

2 to 10

Sanford et al. [40]

Software

pseg

(A)

1

10

Aldoj et al. [39]

Manual

(A)

1

"Expert"

Zavala-Romero et al. [6]

Manual

(A) + (B)

3

Stratified

10

Lee et al. [38]

Manual

(A)

2

?

4

Liu et al. [37]

Software

Osirix

(C) + (A)

More than 2

Stratified

10 and 19

Qin et al. [36]

Manual

(A)

?

?

Motamed et al. [35]

Manual

(A)

2

?

4 and 6

Zabihollahy et al. [13]

Software

itk-SNAP

(A)

4

Splitted

5 and 14

Padgett et al. [8]

Manual

(B)

2

Blinded§

10 and 26

Rundo et al. [15]1

Manual

?

Multiple

?

"Expert"

Meyer et al. [34]

Software

3DSLICER

Medical student + urologist + (A)

4

Stratified

"Expert"

Liu et al. [16]

Software

Osirix

(C) + (A)

7

Stratified

10–15

Rundo et al. [33]2

?

?

(A)

Multiple

?

?

Hambarde et al. [32]

?

?

(A)

Multiple

?

?

Jensen et al. [31]

Software

?

(A)

1

"Expert"

Khan et al. [17]

Manual

(A)

3

?

?

Cheng et al. [30]

Software

pseg

(A)

1

10

Zhu et al. [12]

Manual

?

2

?

More than 5

Mooij et al. [29]

Manual

?

?

?

?

Can et al. [28]

Manual

(A)

3

?

?

Clark et al. [14]

Manual

(A)

1

?

?

Chilali et al. [9]

Manual

(A)

1

15

Makni et al. [10]

Manual

(A)

3

Blinded

"Expert"

Chi et al. [27]

Manual

(A)

1

5

Toth et al. [26]

Software

3DSLICER

(A)

1

"Expert"

Litjens et al. [7]

Manual

(A)

3

?

?

Moschidis and Graham [25]

Manual

(A)

2

?

?

Yin et al. [24]

Manual

"Radiologist-trained operators"

2

Splitted

?

Makni et al. [23]

Manual

(A)

3

Blinded

4, 6 and 9

  1. (A) Radiologist
  2. (B) Radiation oncologist
  3. (C) Research fellow
  4. ? Data not reported
  5. *Experience of reader(s), in years
  6. Unclear for PROMM, in-house data
  7. Splitted but consensus per binome resident-senior
  8. Only one reading for ground truth segmentation but evaluation of intra and inter observator variability on some masks
  9. §Measure of inter- observator variability for 10 masks
  10. Splitted: database is divided such as each set of images is read only once, resulting in an equivalent of single reader
  11. Stratified: first reading (mostly by a less experienced reader) subsequently corrected by a more experience reader
  12. Blinded: blinded reading by at least 2 readers
  13. 1Rundo et al., USE-Net: incorporating Squeeze-and-Excitation blocks into U-Net for prostate zonal segmentation
  14. of multi-institutional MRI datasets [30]
  15. 2Rundo et al., CNN-based Prostate Zonal Segmentation on T2-weighted MR Images: A Cross-dataset Study[27]