Improving the consistency and reproducibility of first-trimester nasal bone assessment: a new classification in a low-risk population

The reproducibility of fetal nasal bone assessment is limited by subjective definitions of its absence (NBA). This study evaluated a new, stricter classification for NBA (complete non-visualization), aiming to improve consistency between the nuchal translucency (NT) and retronasal triangle (RNT) views. We also assessed the incidence of NBA by this new classification and its associated outcomes in a low-risk population. In this retrospective study of 354 low-risk singleton pregnancies, the nasal bone was assessed in both NT and RNT views using a new, stricter definition and compared against the traditional definition. Inter-rater reliability and inter-view consistency were evaluated using Cohen’s Kappa (κ) and the Phi coefficient, respectively. The McNemar test was then used to formally compare the agreement rates between the two classification. The new classification enhanced assessment consistency. Inter-view consistency improved, with the Phi coefficient rising from 0.268 to 0.497. Moreover, the new classification achieved perfect inter-rater reliability (κ=1.0), an increase from the substantial agreement of the traditional classification (κ=0.829, 95% CI: 0.655-1.000). However, a direct statistical comparison found no significant difference in the agreement rates between the two methods (McNemar test, p=0.500). Our stricter definition of NBA shows promise for enhancing assessment consistency and reproducibility and has the potential to reduce false positives. However, due to the low incidence of NBA in our sample, validation of this new criterion's predictive value for aneuploidy in a large, high-risk population is essential before its clinical adoption. The definition of first-trimester nasal bone absence is inconsistent and subjective, leading to poor agreement between ultrasound views and uncertain clinical interpretation. This study introduces and evaluates a stricter classification for nasal bone absence (complete non-visualization). This new classification improved consistency between ultrasound views (Phi coefficient increased from 0.268 to 0.497) and achieved perfect inter-rater reliability (κ=1.0), a numerical increase from the traditional classification’s substantial agreement (κ=0.829). However, a direct statistical comparison did not confirm the superiority of this improvement (McNemar test, p=0.500). Our findings present this as a promising, more objective approach that warrants validation in larger studies.