Ultrasonic Assessment of Cervical Heterogeneity for Prediction of Spontaneous Preterm Birth: A Feasibility Study

    loading  Checking for direct PDF access through Ovid

Abstract

Background In a normal pregnancy, cervical collagen fibers remain organized in predictable patterns throughout most of the gestation. Cervical remodeling reflects a rearrangement of collagen fibers in which they become increasingly disordered and contribute to the pathogenesis of spontaneous preterm birth. Quantitative ultrasound analysis of cervical tissue echotexture may have the capacity to identify microstructural changes before the onset of cervical shortening.

Objective The primary objective of this study was to examine the utility of a novel quantitative sonographic marker, the cervical heterogeneity index (HI), which reflects the relative organization of cervical collagen fibers. Also, we aimed to determine an optimal HI cut-point to predict spontaneous preterm birth.

Study Design This retrospective cohort study employed a novel image-processing technique on transvaginal ultrasound images of the cervix in gestations between 14 and 28 completed weeks. The transvaginal sonography images were analyzed in MATLAB (MathWorks, Natick, MA) using a custom image-processing technique that assessed the relative heterogeneity of the cervical tissue.

Results A total of 151 subjects were included in the study. The mean HI in subjects who delivered preterm and at term was 8.28 ± 3.73 and 12.35 ± 5.80, respectively (p < 0.0001). Thus, decreased tissue heterogeneity was associated with preterm birth, and increased tissue heterogeneity was associated with delivery at term. In our study population, preterm birth was associated with a short cervix (<2.5 cm), history of preterm birth and lower HI, and our findings indicate that HI may improve prediction of preterm birth.

Conclusion Quantitative ultrasound measurement of the cervical HI is a promising, noninvasive tool for early prediction of spontaneous preterm birth.

Related Topics

    loading  Loading Related Articles