Early Learning Curve in the Assessment of Deep Pelvic Endometriosis for Ultrasound and Magnetic Resonance Imaging

Table 3

Overall performance in the learning curve.

r	Interobserver agreement
	Trainee/expert (1st reference)		Trainee–laparoscopy (2nd reference)
	Ultrasound	MRI	Ultrasound	MRI
Frozen pelvis	0.903 ( 0.00)	0.735 ( 0.00)	0.623 ( 0.00)	0.128 ( 0.00)
Uterosacral ligaments	0.512 ( 0.00)	0.601 ( 0.00)	0.261 ( 0.01)	0.455 ( 0.00)
Bowel (rectum, rectosigmoid)	0.633 ( 0.00)	0.699 ( 0.00)	0.539 ( 0.00)	0.598 ( 0.00)
Endometriomas	0.706 ( 0.00)	0.828 ( 0.00)	0.754 ( 0.00)	0.746 ( 0.00)
Vagina	Not computerised
Adenomyosis	0.769 ( 0.00)	0.279 ( 0.00)	Not computerised
Bladder	1.0 ( 0.00)	0.717 ( 0.00)	0.800 ( 0.00)	0.717 ( 0.00)
Rectovaginal septum	Not computerised
Pelvic DE overall	0.690 ( 0.00)	0.697 ( 0.00)	0.490 ( 0.00)	0.531 ( 0.00)

Agreement between trainees and experts and trainees and laparoscopy/histology in the overall assessment of endometriosis in all 3 blocks, expressed in Kappa value. DE: deep endometriosis; p: value; POD: pouch of Douglas; R: rectum; RS: rectosigmoid; USL: uterosacral ligament. None of the 10 vaginal lesions were detected correctly by the trainees; only 4 rectovaginal septum lesions in the cohort out of which none was identified by the ultrasound trainee and only one correctly identified by the radiology trainee; adenomyosis detection was not assessed against laparoscopy/histology since only 1 patient had a hysterectomy.