Methods for cervical volume measurements and interobserver agreement analysis

Prior to starting research on cervical regeneration after Loop electrosurgical excision procedure (LEEP), the authors conducted a pilot study to assess reproducibility between their measurements and compared three different methods for cervical volume (CV) evaluation. Women of reproductive age undergoing gynecological ultrasound examination were recruited and gave their informed consent. Cervical 3D measurements were obtained, stored in an archive, and assessed by two examiners in a blinded manner. Measurements included cervical length (CL), cervical anteroposterior (AP), and longitudinal (LL) diameters. Cervical volume was measured three times by 3D VOCAL (Virtual Organ Computer-Aided Analysis) software (V vocal), by cylinder formula (Vformula=3.14×[AP+LL)/4]2 ×Lenght cervix) and USG devices inbuilt generic volume formula (Vformula_gen). To evaluate measurement reproducibility between methods and two examiners we calculated the Intraclass Correlation Coefficient (ICC) and 95% Confidence Intervals for intraobserver, interobserver, and inter-method agreement. ICC was calculated using a two-way mixed-effects model in SPSS software. We analyzed 35 images. Mean women's age- 34.7 ± 5.5 mm, CL 30.8 ±4.5 mm, AP 30.4 ± 3.6 mm, LL 34.7 ±4.9 mm. Average V formula_gen 17.46 cm3 ±7.60, average V vocal=25.73 cm3 ±5.40 and average V formula= 26.53 cm3 ±8.65. CV measured with Generic formula estimated volume inadequately small , due to the fact that only depth, length, and height measurements were used. All ICC showed excellent reliability. CL interobserver ICC 0.85, P< 0.000, CI 95% 0.702-0.924. V vocal interobserver ICC= 0.927, p<0.000, CI 0.855-0.965.V vocal interobserver ICC= 0.85, p<0.000, CI 0.808-0.947. All intraobserver ICC were within 0.85-0.99 interval range (p<0.000). Intermethod agreement between V vocal and V vocal ICC=0.86, p<0.000, CI 0.741-0.927. All measurements showed adequate reliability within interobserver, intraobserver, and inter method measurements. Generic formulas should not be used for the evaluation of CV. The cylinder formula for the assessment of CV is comparable to 3D VOCAL measurements.