Early Learning Curve in the Assessment of Deep Pelvic Endometriosis for Ultrasound and Magnetic Resonance Imaging
Table 2
Agreement with experts and laparoscopy.
Trainee–expert agreement
Trainee–laparoscopy/histology agreement
Ultrasound
MRI
Ultrasound
MRI
Patients in 3 blocks (total number of lesions)
1st block
2nd block
3rd block
1st block
2nd block
3rd block
1st block
2nd block
3rd block
1st block
2nd block
3rd block
Frozen pelvis overall
29
0.800
0.941
0.933
0.721
0.609
0.836
0.351
0.713
0.800
0.044
0.027
0.267
Uterosacral ligaments
25
0.357
0.571
0.394
0.903
0.583
-0.023 (NS)
0.231
0.100
0.191 (NS)
0.474
0.489
0.083 (NS)
Bowel (R, RS)
19
0.330 (NS)
0.657
1.000
0.657
0.667 (NS)
0.621
0.333 (NS)
0.657
0.560
0.471 (NS)
0.833 (NS)
0.298 (NS)
Endometriomas
18
0.697
0.467
0.76
0.817
1.000
0.681
0.697
0.647
0.783
0.814
0.625
0.681
Adenomyosis
9
0.571 (NS)
0.800 (NS)
1.000 (NS)
0 (NS)
0.400 (NS)
0.421 (NS)
Not computerised
Vagina
10
Not computerised
Rectovaginal septum
4
Not computerised
Total pelvis
0.583
0.708
0.735
0.784
0.627
0.592
0.397
0.542
0.483
0.479
0.592
0.474
Patients in 2 blocks
1st block
2nd block
1st block
2nd block
1st block
2nd block
1st block
2nd block
Bladder
8
1.000
1.000
0.824
1.000
0.667
1.000
0.667
1.000
DE: deep endometriosis; MRI: magnetic resonance imaging; R: rectum; RS: rectosigmoid. NS: statistically not significant result (). Adenomyosis was not assessed against surgical reference standard because only 1 patient had a hysterectomy, vaginal DE not was computerised since none of the 10 lesions were detected on the trainee imaging, and rectovaginal septum DE not was computerised due to low prevalence (2 nodules in the 1st block, 1 lesion in the 2nd block, 1 lesion in the 3rd block).