Utility of multiparametric MRI in Caesarean section scar characterization and preoperative prediction of scar dehiscence: a prospective study

Abstract

Background

Post-Caesarean uterine scar rupture during vaginal birth after Caesarean section (VBAC) is a potentially life-threatening complication. Prediction of scar dehiscence and scar rupture is vital in treatment planning and selecting candidates of trial of labor after a Caesarean section (CS).

Purpose

To assess the accuracy of magnetic resonance imaging (MRI) for evaluation of post-Caesarean uterine scar and to predict scar dehiscence during repeat CS.

Material and Methods

Thirty patients with a history of at least one previous CS underwent pelvic MRI for assessment of uterine scar during a subsequent gestation, all of whom underwent lower segment Caesarean section (LSCS) subsequently due to one of the established indications of CSs. Thickness, T1, T2 signal intensity ratio (SER), and apparent diffusion coefficient (ADC) value of scar site were charted. The lower uterine segment was assessed and graded intraoperatively and findings were correlated with MRI findings.

Results

A total of 30 participants were included in this study, of which nine were classified as having an abnormal scar (of various grades) based on surgical observations. T2 SER with a cutoff value of 0.935 showed the highest sensitivity of 100% and scar thickness value of 3.45 mm showed highest specificity of 91% in prediction of abnormal scar. On drawing a receiver operating characteristic (ROC) curve, T2 signal intensity ratio showed the highest area under the curve (AUC) closely followed by scar thickness values.

Conclusion

MRI derived parameters may be utilized for differentiation of an abnormal post-Caesarean uterine scar from a normal one. Both scar thickness and T2 SER measured on MRI can be used to predict scar dehiscence. However, T2 SER can serve as a more standardized and objective criterion.

Keywords

Magnetic resonance imaging (MRI)Caesarean section genital reproductive tissue characterization obstetrics

Introduction

Post-Caesarean uterine scar rupture during vaginal birth after Caesarean section (VBAC) is estimated to occur in 0.2–3% of cases and has a potentially devastating nature (1). This has led to a decrease in the rates of trial of labor after a Caesarean section (TOLAC) with a consequent increase in elective obstetric surgeries. Though a relatively safe procedure, Caesarean section (CS) also carries various intra- and postoperative risks such as placenta previa, placenta accreta, and placental abruption, which increase during the repeat surgery (2). The old adage “once a Caesarean, always a Caesarean” (3) has been challenged and a recent push from the American College of Obstetrics and Gynecology (ACOG) calls for an accurate and standardized diagnostic modality with an objective criterion that provides a reliable prediction of a satisfactory outcome after TOLAC. The existing body of evidence suggests that the thickness of the lower uterine segment (LUS) measured by transabdominal sonography (TAS) and transvaginal sonography (TVS) can successfully predict the risk of scar rupture and remains the gold standard. Many studies, however, confirm the fact that scar thickness is not the only viable predictor of rupture. Composition of scar tissue may also play a significant role in the phenomenon of scar dehiscence. Magnetic resonance imaging (MRI) is known to have the best soft tissue contrast resolution among all imaging modalities and some early work has been proposed for evaluation of uterine scar using this technology. The present study aims at the evaluation of post-Caesarean uterine scar morphology using multi-parametric MRI. We further tried to correlate the imaging morphology with the grades of uterine scar and hence with the probability of scar ruptures.

Material and Methods

Patients

This Institutional review board (IRB) approved prospective observational study was carried out at a tertiary level university-based teaching hospital over a period of two consecutive years. A total of 30 patients were recruited from those referred for sonographic evaluation of the fetus and the LUS in women who had undergone at least one lower segment Caesarean section (LSCS) in the past. Informed consent was obtained from all the patients in this study. All patients subsequently underwent LSCS due to one of the established indications of CSs (Table 1).

Table 1.

Indications for repeat CS.

Indication of repeat CS	Patients, n (n = 30)
Previous 2 LSCS	9
Placental abruption	1
Meconium stained liquor	1
Scar tenderness	7
Breech presentation	2
Chorioamnionitis	2
Severe pre-eclampsia	2
Fetal distress	2
Elective repeat CS on maternal request	2
Maternal diabetes	1
Uncontrolled intrauterine growth retardation	1

CS, Caesarean section; LSCS, lower segment Caesarean section.

Patients with confounding factors like multiple pregnancies, preterm deliveries, polyhydramnios or oligohydramnios, low lying placenta, patients with history of uterine surgery other than CS or unavailable previous Caesarean details, and with a contraindication to MRI were excluded from the study. The investigation was performed after obtaining a written informed consent from the patient.

Protocol for MRI

MRI was done on a 1.5 Tesla (Siemens Avanto, Erlangen, Germany) system with an actively shielded whole-body superconducting magnet. All the MR scans were performed in late third trimester of pregnancies. Imaging was done using an eight-channel Torso phased-array body coil with the patient in the supine position and a moderately full urinary bladder. Signal improvement was done by addition of integrated spine elements. Saturation bands were applied over the abdomen to eliminate bowel peristalsis and fetal movement artifacts. Pre-designated standard protocols were followed consisting of T1-weighted (T1W) and T2-weighted (T2W) imaging sequences in axial and sagittal planes remaining perpendicular to the long axis of the scar. Initial single shot localizers were taken to define the uterine scar (similar to the method followed by sonography) followed by oblique images, which were exactly perpendicular to the scar. This exercise was done to eliminate errors of over- and underestimation due to foreshortening or widening of the region in either plane. The scar site was identified as the thinnest portion of LUS and having the lowest signal intensity on T2W imaging (Figs. 1 and 2). Scar thickness was calculated at this site. Region of interest were taken on T1W and T2W sequences at the scar site to measure the signal intensity and signal intensity ratio (SER). ADC value of the scar tissue was also charted in each case (Fig. 3).

Fig. 1.

Sagittal T2W showing severely thinned T2 dark scar (arrows). On surgery, it was Grade III scar.

Fig. 2.

Sagittal T2W image of a Grade II CS scar shows a mildly thinned scar with higher T2 signal intensity (*) compared to a Grade III scar.

Fig. 3.

Sagittal ADC map of a patient with a Grade II CS scar (*).

Protocol for surgical scar grading

The LUS was assessed and graded intraoperatively according to system developed by Qureshi et al. (4): Grade I, well developed LUS; Grade II, thin LUS, uterine contents not visible; Grade III, scar dehiscence – subperitoneal separation of uterine scar, with uterine contents (chorio-amniotic membrane) visible through the peritoneum of the LUS; and Grade IV, a well circumscribed defect present in LUS.

Grades III and IV were considered as an abnormal LUS, while Grades I and II were considered normal.

Statistical analysis

Statistical analysis was performed using SPSS software (Version 22.0. IBM Corp. Armonk, NY, USA). The scar thickness, ADC value, T1 and T2 SER observed on MRI were stratified according to surgical grades assigned during surgery. The significance of difference in mean between normal and abnormal scars was evaluated using one-way ANOVA. Receiver operating characteristics (ROC) curves were constructed to evaluate individual utility of these parameters in predicting surgical Grade III scar (abnormal scar) with the help of the area under the curve (AUC). Cutoff values were determined based on ROC curves that optimized sensitivity, specificity, and positive predictive value (PPV) of each parameter for the purpose.

Results

The mean maternal age of the cohort was 31 years (age range, 26–35 years). Grand multigravida (gravidity ≥4) constituted 53.7% of the study (range, 2–7) with most women having ≥1 previous LSCS (n = 29) and only seven (23.3%) having a past VBAC. Of the total 30 patients, nine were intraoperatively classified as Grade III scars, eight as Grade II scars, and 13 as Grade I scars. None of the patients had Grade IV scars intraoperatively.

Table 2 charts the mean values of T1 SER, T2 SER, ADC, and scar thickness, which were calculated on MRI in Grade I, II, and III scars. One-way analysis of variance (ANOVA) showed that this difference of means was statistically significant for T2 SER and scar thickness and non-significant for T1 SER and ADC values.

Table 2.

Mean values of T1 SER, T2 SER, ADC, and scar thickness which was calculated on MRI in Grade I, II, and III scars.

	Grade I (n = 13)	Grade II (n = 8)	Grade III (n = 9)	Total (n = 30)	P value
T1 SER	0.86	0.94	0.82	0.87	0.702
T2 SER	1.33	1.49	0.71	1.19	0.003
ADC (×10³mm²/s)	0.79	0.84	0.82	0.82	0.762
Scar thickness (mm)	4.14	3.86	2.64	3.62	0.001

ADC, apparent diffusion coefficient; SER, signal intensity ratio.

Fig. 4 shows the ROC curve drawn for each parameters measured on MRI with AUC charted in Table 3.

Fig. 4.

ROC curves to evaluate individual utility of various parameters (T1, T2 signal intensity ratio, ADC values, and scar thickness) in predicting surgical Grade III scar (abnormal scar).

Table 3.

AUC of the various MRI parameters in prediction of Grade III (abnormal) scar.

Test result variable(s)	AUC
T1 signal intensity ratio	0.640
T2 signal intensity ratio	0.886
ADC	0.476
Scar thickness	0.862

ADC, apparent diffusion coefficient; AUC, area under ROC curve; ROC, receiver operating characteristics.

T2 signal intensity ratio shows the highest AUC closely followed by scar thickness values. Table 4 enlists the various cutoff values for each parameter derived from the ROC curve and sensitivity and specificities of these cutoff values for predicting abnormal scar. T2 SER with a cutoff value of 0.935 showed the highest sensitivity of 100% and a scar thickness value of 3.45 mm showed the highest specificity of 91%. The PPVs of these cutoff values of T2 SER and scar thickness were 92.45% and 95.61%, respectively.

Table 4.

Sensitivities and specificities of various MRI parameters based on ROC curve derived cutoff values.

	Cutoff	Sensitivity (%)	Specificity (%)
T1 signal intensity ratio	0.77	66.7	66.7
T2 signal intensity ratio	0.935	100	81
ADC	0.805	66.7	53
Scar thickness (mm)	3.45	88.9	91

ADC, apparent diffusion coefficient; ROC, receiver operating characteristics.

Discussion

With the increasing CS rates in modern obstetric practice, more and more women are becoming pregnant with a scar on the uterus. The relatively avascular and thin LUS are ideal to place the surgical incision during a CS (5). However, presence of a post-surgical scar over the LUS predisposes it to various complications such as scar dehiscence or rupture, abnormal uterine bleeding, scar pregnancy, and subfertility (6). Owing to these risks, there has been steep increase in repeat CS rates and a consequent decrease in VBAC over the past two decades. However, these are costly and invasive surgeries and incur significant financial impact on the patients, especially in developing nations. The most commonly cited scar rupture rate of approximately 0.5%, i.e. one in every 200 women, cannot be justified to deny a woman an attempt to VBAC (7). What is needed is prudent application of the available imaging technology to classify these patients according to the relative risk of rupture so that an informed decision can be made.

Various studies have assessed the appearance of LUS on ultrasound in a pregnant woman, especially in the late third trimester, concluding that ultrasonography can be used effectively to predict the risk of intrapartum rupture. Both TAS and TVS have been used to evaluate scar thickness with a strong index of correlation between the two (8). A recent study construes a possible relationship between risk of rupture and the non-pregnant appearance of a CS scar as well (9). Although most of the studies have assessed the thickness of the LUS (scar thickness), some have evaluated morphological appearance on sonography and vascularity on color Doppler. However, the detection rate of scar on ultrasound is in the range of 7–89% in various studies (10). Moreover, the methods used to calculate “scar thickness” in various studies vary (11 –17) and so do the suggested cutoff values to predict uterine rupture (Table 5).

Table 5.

Methods used to calculate “scar thickness” in various studies and the suggested cutoff values for prediction of the risk of uterine rupture.

	Imaging modality	Assessed parameter	Cutoff value predicting risk of uterine rupture
Rosenberg et al. (11)	TAS	LUS thickness with UB wall thickness	3.5 mm
Fukuda et al. (12)	TAS	LUS thickness (not clearly defined site of measurement	2 mm
Gotoh et al. (13)	TVS	LUS thickness	2 mm
Bujold et al. (14)	TAS and TVS	LUS thickness	2.5 mm
Jastrow et al. (15)	TAS and TVS	LUS thickness (varied according to degree of fullness of bladder)	No cutoff value
Asakura et al. (16)	TVS	Myometrial layer of LUS	1.6 mm
Lebedev et al. (17)	TAS	LUS	3–4 mm
Lebedev et al. (17)	3D USG	Scar volume	10 cc

LUS, lower uterine segment; TAS, transabdominal sonography; TVS, transvaginal sonography; UB, urinary bladder.

Morphological sonographic parameters of the scar such as homogeneity of the scar and triangular shape has been suggested to predict successful VBAC whereas a balloon-like shape and areas of increased echogenicity in the scar area have been described as features of poor scar performance (17). Basic et al. suggest color Doppler as the gold standard in evaluation of scar behavior. The presence of so many ultrasound techniques in the literature signifies that none of these techniques have been unequivocally optimal (5).

Owing to its excellent contrast resolution, multiplanar capability, and standardized acquisition technique, MRI has gained an established role in gestational imaging. Besides providing an unmatched anatomical detail of various fetal organs, MR enables simultaneous visualization of important extrafetal organs such as uterine wall, placenta, and amniotic fluid. The role of MRI as a problem-solving modality in evaluation of a suspected invasive placental disorder, which is another complication of a scarred uterus that has undergone previous CS (18,19), has been established in the literature with firm conviction. There have been only a few studies in the literature describing the role of MRI in the evaluation of a CS scar. A study by Singh et al. compared the precision of scar thickness measured by TVS to that measured with MRI in prediction of actual scar thickness found on elective repeat CS (20). They showed that TVS has a better correlation coefficient with actual scar thickness than MRI. However, this study did not assess radiologically measured thickness to grades of scar and thus their propensity to rupture. Our study showed that mean scar thickness of Grade III scars was 2.64 mm which differed significantly from the means of normal scars (3.86 mm for grade II and 4.1.4 mm for Grade I scars) with a P value 0.001. A cutoff value of 3.45 mm had excellent sensitivity (89%), specificity (91%), and PPV (95.6%) in predicting abnormal scar.

Since there has been confusion over the correct technique of measurement of thickness and dependence over degree of fullness of bladder, we intended to assess more standardized and objective criteria such as SER on T1W and T2W sequences and ADC values. Few studies have evaluated the morphological behavior of CS scar on MRI. Dicle et al. examined CS incision scars postoperatively by MRI to assess the healing period of the scar and observed that the scar tissues showed deceased signals on both T1W and T2W SE sequences during the first 3 months (21). Subsequently, the zonal anatomy of the uterus reappeared completely by 6 months. They inferred that complete healing of the scar tissue took 6 months, which was accompanied by recovery of MR signal changes. In the present study, we observed that the mean signal intensity ratio was lesser in abnormal scars (Grade III scars) on T1W and T2W sequences compared to the normal scars (Grade I and II scars). Although this difference was statistically significant only for T2 SER, this difference of means could represent poor healing and suboptimal scar healing following the surgery in the abnormal scars. ROC curve analysis of these parameters showed that T2 SER had the highest AUC followed by scar thickness (Table 3). T2 SER of 0.935 had 100% sensitivity, 81% specificity, and 92.5% PPV in the prediction of an abnormal scar.

Fiocci et al. compared five patients with Caesarean-scarred uteri and 13 nulliparous uteri with the help of diffusion tensor imaging (DTI) and showed that the majority of the uteri with a Caesarean scar had altered orientation of fibers in the anterior isthmus compared to non-scarred myometrium (22). Besides the number of fibers and the fiber density, they also compared FA and ADC values between scarred and non-scarred uteri. The difference in FA and ADC values, though reasonable, was not statistically significant, however, the authors believed that this was due to small sample size. ADC values in the present study did not correlate well with the grades of scar and difference of means between the different grades of scar was not statistically significant. Another publication by Fiocci et al. evaluated the role of DTI using a 3T scanner in characterizing CS scars in 24 women at term pregnancy and compared them with 12 women with prior vaginal delivery who served as controls (23). They classified CS scars into linear and retracting, based on DTI morphology. They observed that the longitudinal fibers running in the anterior wall were significantly reduced compared to posterior wall in CS scars, especially in retracting scars whereas anterior and posterior fibers were similar in controls. Although these studies show that MRI can perceive changes in uterus with a CS scar, none of them have previously attempted to grade the scar which might be more valuable to the clinicians. Moreover, non-availability of DTI in many centers somewhat limits its utility in clinical domains. The present study demonstrates that the signal intensity ratio on T2W sequence can effectively predict abnormal scar and thus risk of scar rupture.

We realize that there were few limitations of the study. First, we have assumed that the risk of scar “rupture” or uterine rupture increases according to the type of scar visualized on the intraoperative scoring system, as has been done in previous studies. However, the current literature does not provide an actual risk level given for each type of scar classification and more importantly a degree of risk associated with Grade I and II scars during the trial of labor. A second limitation of this study was the small sample size. A large multicentric study with more number of patients may yield a more statistically significant result.

In conclusion, this study shows that both scar thickness and T2 signal intensity ratio measured on MRI can be used to predict scar dehiscence and rupture. However, there are a lot of discrepancies in the techniques employed to measure the thickness of scar. T2 SER can serve as a more standardized and objective criterion, thus justifying the use of this expensive investigation in the evaluation of scar behavior.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, author-ship, and/or publication of this article.

References

Ofili-Yebovi

Ben-Nagi

Sawyer

. Deficient lower-segment Cesarean section scars: prevalence and risk factors. Ultrasound Obstet Gynecol 2008; 31: 72–77.

Hamilton

Hoyert

Martin

. Annual summary of vital statistics: 2010-2011. Pediatrics 2013; 131: 548–558.

Craigin E. Conservation in obstetrics. N Y State Med J 1916;1:104.

Qureshi

Inafuku

Oshima

. Ultrasonographic evaluation of lower uterine segment to predict the integrity and quality of cesarean scar during pregnancy: a prospective study. Tohoku J Exp Med 1997; 183: 55–65.

Basic

Basic-Cetkovic

Kozaric

. Ultrasound evaluation of uterine scar after cesarean section. Acta Inform Med 2012; 20: 149–153.

Naji

Abdallah

Bij De Vaate

. Standardized approach for imaging and measuring Cesarean section scars using ultrasonography. Ultrasound Obstet Gynecol 2012; 39: 252–259.

Enkin

Keirse

MJNC

. Effective Care in Pregnancy and Childbirth, Oxford: Medical Publications, 1995.

Sen

Malik

Salhan

. Ultrasonographic evaluation of lower uterine segment thickness in patients of previous caesarean section. Int J Gynecol Obstet 2004; 87: 215–219.

Vikhareva Osser

Valentin

. Clinical importance of appearance of caesarean hysterotomy scar at transvaginal Ultrasonography in nonpregnant women. Obstet Gynecol 2011; 117: 525–532.

10.

Naji

Daemen

Smith

. Visibility and measurement of cesarean section scars in pregnancy: a reproducibility study. Ultrasound Obstet Gynecol 2012; 40: 549–556.

11.

Rosenberg

Gofnet

Phillippe

. Thickness of the lower uterine segment: its influence in the management of patients with previous caesarean section. Eur J Obstet Gynecol Reprod Biol 1999; 87: 39–45.

12.

Fukuda

Mochizuki

. Examination of previous caesarean section scars by ultrasound. Arch Gynecol Obstet 1988; 243: 221–224.

13.

Gotoh

Masuzaki

Yoshida

. Preoperative diagnosis of dehiscence of lower uterine segment in patients with a single previous caesarean section. Aust N Z J Obstet Gynaecol 2000; 40: 402–404.

14.

Bujold

Gauthier

Hamilton

. Maternal and perinatal outcomes associated with a trial of labor after previous caesarean delivery. J Midwifery Womens Health 2005; 50: 363–364.

15.

Jastrow

Demers

Gauthier

. Sonographic lower uterine segment thickness and risk of uterine scar defect: A systematic review. J Obstet Gynecol Can 2010; 32: 321–321.

16.

Asakura

Nakai

Ishikawa

. Prediction of uterine dehiscence by measuring lower uterine segment thickness prior to onset of labor: Evaluation by transvaginal Ultrasonography. J Nippon Med Sch 2000; 67: 352–356.

17.

Lebedev

Strizhakov

Zhelnezov

. Echographic and morphological parallels in the Evaluation of the condition of the uterine scar. Akush Ginekol 1991; 8: 44–49.

18.

Kumar

Verma

Ojha

. Invasive placental disorders: a prospective US and MRI comparative analysis. Acta Radiol. DOI: 10.1177/0284185116638567.

19.

Kumar

Verma

Jain

. Chemical shift artefact on steady-state MRI sequences for detection of vesical wall invasion in placenta percreta. J Obstet Gynecol India 2016; 66: 101–106.

20.

Singh

Tripathi

Mala

. Comparison of scar thickness measurements using trans-vaginal sonography and MRI in cases of pregnancy with previous caesarean section. Do they correlate with actual scar thickness? J Obstet Gynaecol 2013; 33: 810–813.

21.

Dicle

Küçükler

Pirnar

. Magnetic resonance imaging evaluation of incision healing after cesarean sections. EurRadiol 1997; 7: 31–34.

22.

Fiocchi

Petrella

Nocetti

. Transvaginal ultrasound assessment of uterine scar after previous caesarean section: comparison with 3T-magnetic resonance diffusion tensor imaging. Radiol Med 2015; 120: 228–238.

23.

Fiocchi

Nocetti

Siopis

. In vivo 3 T MR diffusion tensor imaging for detection of the fibre architecture of the human uterus: a feasibility and quantitative study. Br J Radiol 2012; 85: e1009–1017.