Image-scoring system for crown-rump length measurement

Abstract

Objective: To develop and evaluate an objective image-scoring system for crown-rump length (CRL) measurements and to determine how this compares with subjective assessment.

Methods: A total of 125 CRL ultrasound images were selected from the database of the International Fetal and Newborn Growth Consortium for the 21(st) Century study group. Two reviewers, who were blinded to the operators' and to each others' results, evaluated all images both subjectively and objectively. Subjective evaluation consisted of rating an image as acceptable or unacceptable, while objective evaluation was based on six criteria. Reviewer differences for both the subjective and objective evaluations were compared using percentage of agreement and adjusted kappa values.

Results: The distribution of individual scores and differences between subjective and objective evaluation for the two reviewers was similar. Overall agreement between the reviewers was higher for objective evaluation (95.2%; adjusted κ, 0.904), than for subjective evaluation (77.6%; adjusted κ, 0.552). There was a high level of agreement for horizontal position (κ = 0.951), magnification (κ = 0.919), visualization of crown and rump (κ = 0.806) and caliper placement (κ = 0.756), while agreement for mid-sagittal section (κ = 0.629) and neutral position (κ = 0.565) were moderate and poor, respectively.

Conclusion: The proposed six-point scoring system for CRL image rating is more reproducible than is subjective evaluation and should be considered as a method of quality assessment and audit.

Keywords: crown-rump length; image scoring; objective evaluation; reproducibility; subjective evaluation.