Objective
The objective of the study was to describe the assessment of lower segment uterine scar (LSCS) by transvaginal ultrasound (TVUS) during a first-trimester scan.
Study Design
Patients with a history of LSCS were prospectively enrolled over a 6 month period. Four groups were defined: type 1A, thin scar within cervicoisthmic canal (CIC); type 1B, thin above the internal os (IO); type 2A, dehiscent within the CIC; type 2B, dehiscent above the IO. Accuracy of first-trimester TVUS was investigated by blind testing a panel of 14 operators over a web-based dataset.
Results
The scar was visualized in 122 of 123 patients enrolled. Types 1A, 1B, 2A, and 2B occurred in 49.2%, 3.3%, 38.3%, and 9.2%, respectively. When blind tested, fetal medicine specialists achieved a median sensitivity of 82% and specificity of 100% for the detection of a scar. These were 83% and 87% for nonspecialists.
Conclusion
First-trimester uterine scar assessment may become a valuable tool in early recognition of patients at risk of subsequent perinatal complications.
The increase in the rate of cesarean section in most developed countries has led to a concurrent increase of potentially life-threatening related subsequent complications, such as placenta accreta, low-lying placenta, and uterine rupture, as well as surgical morbidity.
In these circumstances, efforts are made in predelivery screening for these complications. Although most of the attention has focused on second- and third-trimester diagnosis of abnormal placental insertion and risk of rupture during labor in patients with a history of lower-segment cesarean section (LSCS), it may be that first-trimester screening could allow early recognition of patients at risk for any of these perinatal complications.
Indeed, the first trimester is increasingly considered as a starting point for stratifying patients at risk of subsequent obstetric complications, such as placenta previa and preeclampsia. In this frame of work, several clinical risk factors required for first-level screening for abnormal placental insertion, including maternal age and number of uterine scars, are available at time of first-trimester screening.
The objective of this study was 2-fold: first, to describe the ultrasound anatomy of uterine scars following LSCS in pregnant women at the first trimester; and second, to assess the feasibility and accuracy of using ultrasound for uterine scar detection in daily practice. The anatomical ultrasound semeiology is described in a prospective study of ultrasound findings of patients with prior LSCS.
The ability to assess the presence of a scar by first-trimester ultrasound was studied by evaluating a panel of practitioners blinded to the true presence of a scar in a standardized design.
Materials and Methods
Over a 6 month period (January to June 2010), all consecutive patients with a history of prior LSCS were prospectively enrolled at 11-13 +6 gestational weeks. Assessment of uterine scar was performed by transvaginal ultrasound, based on a standardized plane. This standardized plane was defined as a midsagittal view of the uterus comprising the cervical canal, the uterine isthmus, and internal os, as well as the lower part of the gestational sac and the bladder.
Because the exact differentiation between the isthmus and the cervix is seldom visible, the internal os was defined as the most internal end of the cervicoisthmic canal when it reaches the gestational sac. Once the scar was visualized, its location was described either within the cervicoisthmic canal or outside the cervicoisthmic canal above the internal os. The distance between the internal os and the scar was measured. Concurrently, the scar was classified as dehiscent when presenting as a wide hypoechoic defect greater than 2 mm or as thin when presenting as a linear defect as described by Osser et al.
Four groups were therefore defined: type 1A, thin and within cervicoisthmic canal; type 1B, thin and above cervicoisthmic canal; type 2A, dehiscent and within the cervicoisthmic canal; and type 2B, dehiscent and above the cervicoisthmic canal. The relationship between a low-lying trophoblast and the scar was assessed when the scar was found above the cervicoisthmic canal and therefore exposed to placental invasion. The concurrent finding of a trophoblast overlapping the internal os and an exposed scar was specifically recorded. All other situations (ie, nonoverlapping trophoblast or protected scar within the cervicoisthmic canal) were also described.
Ultrasounds were performed using a General Electric Voluson E8 or a Voluson 730 Expert (GE Medical System Europe, Buc, France) with a 3.5-5 MHz or 6-8 MHz transvaginal transducer. Images as well as obstetric history including age, parity, previous uterine surgery, number of and indication for cesarean sections, postoperative pyrhexia, previous or subsequent vaginal deliveries, the interval between the last cesarean section, and the current estimated date of conception were recorded in our electronic database (Astraia, Munich, Germany).
The ability to detect the scar by transvaginal ultrasound was investigated by measuring the sensitivity and specificity of a panel of operators. A web-based collection of 110 images was constructed, including a 50% prevalence of uterine scars. All the ultrasound images were rated by an independent fetal medicine specialist to ensure that the abovementioned criteria for the midsagittal plane were met.
All participating raters were given a short theoretical course before starting presenting the purpose of the study as well as the semeiologic elements of uterine scars on first-trimester scans. The question was “Can you see a uterine scar on this image?” and possible answers were only yes or no. For each image, the rater was given a maximum of 1 minute. All raters were blinded to the existence of a scar as well as to the artificial prevalence of uterine scars given to the database. The purpose of blinding is to avoid overdiagnosing scars when they may not be visible just because of the knowledge of maternal history. This may be due to artifact images such as posterior shadows behind bladder folds wrongly considered as uterine scars for example. Therefore, sensitivity and specificity represent the true ability of first-trimester ultrasound to detect a scar in a clinical daily practice without interference of prior knowledge.
The statistical analysis was conducted using R ( www.r-project.org ). Quantitative variables are summarized by the median and interquartile range (25th to 75th centile) and qualitative variables are described by n (%). Comparisons of demographic and ultrasound characteristics between thin and dehiscent scars was performed using Mann-Whitney U tests for quantitative variables and Fisher exact tests for qualitative variables. Sensitivity and specificity of each operator were computed: sensitivity = (true positives)/(true positives + false negatives) and specificity = (true negatives)/(true positives + false negatives).
Because transvaginal ultrasound is offered routinely for first-trimester screening in our practice, this study did not require an institutional review board; however, written informed consent was obtained from all women. This study was supported by the Société Française pour l’Amélioration des Pratiques Echographiques in setting teaching sessions.
Results
The cohort included 123 patients with a history of prior LSCS. All patients were between 11 and 13 +6 gestational weeks as assessed by crown-rump lengths within 45 and 84 mm. Most patients had only 1 scar (79%), with 1 patient presenting a history of 3 cesarean sections. The median interval since last delivery was 2.7 years (32 months), with a 7% rate of history of vaginal birth after cesarean (VBAC). All cesarean sections were carried out using a transverse lower segment incision at longer than 37 gestational weeks.
LSCS was performed in labor in 46 cases, and 9 women reported postoperative pyrhexia. Six patients had a twin pregnancy, 2 also had a previous myomectomy, and 2 were diagnosed with a bicornuate uterus. There were 63 cases with a thin scar (type 1) (51.6%). Among these, 59 (49.2%) were located within the cervical canal (type 1A), and 4 (3.3%) were found on the uterine isthmus above the internal os (IO) (type 1B) ( Figure 1 , A and B) . Fifty-nine (48.4%) had at least 1 large scar (median thickness of 4.9 mm; interquartile range, 3.3–7.2) (type 2), including 46 (38.2%) and 11 (9.2%) within (type 2A) the cervical canal and above the IO (type 2B), respectively. Figure 2 summarizes the frequencies of each anatomical type in our population.
