Inter‐and intrarater reproducibility in detecting deep myometrial invasion and cervical stromal invasion using 2D‐TVU and 3D‐VCI

Investigate the inter- and intrarater reproducibility between 2D-TVU and 3D-VCI in staging of endometrial cancer (EC). 15 gynecologists, with an median 7 years (range 2-22 years) experience in EC staging, assessed off-line de-identified TVS video-clips and 3D-VCI volumes regarding presence/absence of deep (>50%) myometrial invasion (MI) and/or cervical stromal involvement (CSI) in the same set of patients. The patient material was retrospectively collected from 58 women, with biopsy confirmed EC, examined by a single examiner (consecutive cases, selecting cases to achieve an adequate mix of deep MI (n=22) and CSI (n=9)). Reproducibility was measured using Cohen's and Fleiss' kappa (κ). κ values were denoted ‘Poor’ (≤0.2), ‘Fair’ (0.21-0.4), ‘Moderate’ (0.41-0.6), ‘Good’ (0.61-0.8) and ‘Very good’ (0.81-1). Intra-rater reproducibility was generally lower in the assessment of MI than CSI (for MI; 7/15 (47%) raters had ‘Poor’-‘Fair’ Cohen's κ and 8/15 (53%) had a ‘Moderate’-‘Very good’ κ, for CSI all raters 15/15 (100%) had ‘Moderate’-‘Very good’ κ). Fleiss' κ was ‘Fair’ (0.34) for MI and ‘Moderate’ (0.49) for CSI. Interrater reproducibility was lower for 3D-VCI than for 2D-TVU. For MI with 2D-TVU; 59 out of 105 pairs (56%) had ‘Moderate’-‘Very good’ Cohen's κ. With 3D-VCI only 29 of 105 (28%) had ‘Moderate’-‘Very good’ Cohen's κ (see image). In the assessment of CSI with 2D-TVU, 90 of 105 pairs (86%), and with 3D-VCI 67 of 105 pairs (64%), had a κ ‘Moderate’-‘Very good’. Supporting information can be found in the online version of this abstract Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.