Image acquisition and review
TVUS cervical images were obtained using standardized CL Education and Review (CLER) criteria by accredited sonographers with supervision by maternal fetal medicine physicians at the NYU MFCC. Ultrasound units with 4- to 9-MHz IC5-9D (Voluson e8; GE Healthcare, Milwaukee, WI) or 4- to 8-MHz C8-4v (IU22; Phillips Healthcare, Andover, MD) wide-view transducers were used to obtain all images 20. Stored CL images were evaluated for this study by a single reviewer, who was blinded to delivery outcome. Of the patients included in final analysis, 25% were randomly selected for a second review of stored CL images by another study team member to confirm reliability of interpretation and to calculate scores of interobserver variability. Cervical images were assessed for quality, reported CL, and visibility of the CGA. Optimal cervical imaging was defined as a sagittal view of the cervix with clearly visible internal and external ostia, approximately equal thickness anterior and posterior cervical lips, complete (or near complete) visualization of the cervical canal, and minimal urine in the maternal bladder 21.