Comparison of four imaging modalities for cesarean scar defect: a retrospective study.

We included 61 individuals, of whom 9 had a history of 2 or more cesarean sections, accounting for 14.8%. The median menstrual bleeding duration was 14 days (IQR: 12–15 days). Patient characteristics are shown in Table  1 . As shown in Fig.  1 , all 61 enrolled patients underwent MRI and SCE_MRI analysis. Among them, 53 patients received TVU examination at our hospital, while the remaining 8 underwent TVU examination at other institutions (data from these 8 patients were not included in the analysis). Additionally, 60 patients underwent HSC and TAU analysis. Representative images of the four detection methods are shown in Fig.  2 . Table 1 Patient characteristics characteristic Value ( n  = 61) Age, year; (mean ± SD) 31.92 ± 3.82 Age of the last CS, year; (mean ± SD) 26.59 ± 3.88 Time interval between the last CS and this detection, year; median (Q1, Q3) 4 (3, 7.5) Duration of menstruation, day; median (Q1, Q3) 14 (12, 15) Menstrual cycle, day (mean ± SD) 28.35 ± 7.21 Number of CSs  1 52  ≥ 2 9 CS Cesarean section, SD Standard Deviation Patient characteristics CS Cesarean section, SD Standard Deviation Fig. 1 Four diagnostic modalities for CSD examination. This study enrolled 61 patients with a personal history of Cesarean section presenting with Abnormal Uterine Bleeding. Within the 7–14 day window of a single menstrual cycle (MC), each patient sequentially underwent four diagnostic modalities: Transvaginal Ultrasound (TVU), Magnetic Resonance Imaging (MRI), Saline contrast-enhanced MRI (SCE_MRI), and Hysteroscopy combined with Transabdominal Ultrasound (HSC_TAU). Measurements included Cesarean Scar Defect (CSD) length, width, depth, and Thickness of the Residual Myometrium (TRM). Among the 61 patients examined by TVU, data from 8 patients referred from external institutions were excluded from the final analysis. Additionally, HSC_TAU data for one patient was lost Four diagnostic modalities for CSD examination. This study enrolled 61 patients with a personal history of Cesarean section presenting with Abnormal Uterine Bleeding. Within the 7–14 day window of a single menstrual cycle (MC), each patient sequentially underwent four diagnostic modalities: Transvaginal Ultrasound (TVU), Magnetic Resonance Imaging (MRI), Saline contrast-enhanced MRI (SCE_MRI), and Hysteroscopy combined with Transabdominal Ultrasound (HSC_TAU). Measurements included Cesarean Scar Defect (CSD) length, width, depth, and Thickness of the Residual Myometrium (TRM). Among the 61 patients examined by TVU, data from 8 patients referred from external institutions were excluded from the final analysis. Additionally, HSC_TAU data for one patient was lost Fig. 2 Multimodal imaging of cesarean scar diverticulum. Representative images of a CSD show by transvaginal ultrasound (TVU), hysteroscopy-guided transabdominal ultrasound (HSC_TAU), magnetic resonance imaging (MRI), and intrauterine fluid saline contrast-enhanced MRI (SCE_MRI). The dashed line indicates the uterine contour. The green, blue, and red lines indicate the length, depth, and TRM measurement of the diverticulum, respectively.CSD: Cesarean scar diverticulum; TRM: Thickness of residual myometrial muscle Multimodal imaging of cesarean scar diverticulum. Representative images of a CSD show by transvaginal ultrasound (TVU), hysteroscopy-guided transabdominal ultrasound (HSC_TAU), magnetic resonance imaging (MRI), and intrauterine fluid saline contrast-enhanced MRI (SCE_MRI). The dashed line indicates the uterine contour. The green, blue, and red lines indicate the length, depth, and TRM measurement of the diverticulum, respectively.CSD: Cesarean scar diverticulum; TRM: Thickness of residual myometrial muscle HSC_TAU measured longer CSD length (median 15.4 mm, IQR: 10.7–17.7 mm) than TVU (6.5 mm, IQR: 5.0–9.3 mm), MRI (8.0 mm, IQR: 6.0–10.0 mm), and SCE_MRI (8.2 mm, IQR: 6.5–12.1 mm) (Table  2 ). All pairwise comparisons between HSC_TAU and other methods were statistically significant (all P  < 0.001, Fig.  3 A). The biases ranged from − 7.529 to 6.181, and the intraclass correlation coefficient (ICC) among the four modalities was 0.596 (Supplementary Fig. 1). Table 2 CSD measurements by imaging modality Parameter HSC_TAU TVU MRI SCE_MRI P N* 60 50 59 59 - Length, mm 15.4 (10.7–17.7) 6.5 (5.0-9.3) 8.0 (6.0–10.0) 8.2 (6.5–12.1) < 0.001 Width, mm 14.2 (10.5–18.2) 11.0 (8.0–15.0) 14.0 (10.1–17.0) 17.0 (13.0-20.8) < 0.001 Depth, mm 8.5 (7.2–11.7) 8.0 (6.0-9.3) 6.3 (4.6-8.0) 7.2 (5.6–9.2) < 0.001 TRM, mm 2.0 (1.5-3.0) 2.7 (1.9–3.9) 2.0 (1.5-3.0) 1.8 (1.0-2.4) < 0.001 Data are presented as median (IQR) *Among the 53 patients who underwent TVU at our hospital, three had no detectable CSD, yielding 50 evaluable cases. Additionally, 60 patients underwent HSC_TAU examination. All 61 enrolled patients underwent MRI and CE_MRI analysis. Two patients had no detectable CSD on these examinations, leaving 59 cases for analysis CSD measurements by imaging modality Data are presented as median (IQR) *Among the 53 patients who underwent TVU at our hospital, three had no detectable CSD, yielding 50 evaluable cases. Additionally, 60 patients underwent HSC_TAU examination. All 61 enrolled patients underwent MRI and CE_MRI analysis. Two patients had no detectable CSD on these examinations, leaving 59 cases for analysis Fig. 3 Multimodal imaging measurements of cesarean scar diverticulum parameters. Length ( A ), width ( B ), depth ( C ), and TRM ( D ) assessed using transvaginal ultrasound (TVU), hysteroscopy-guided transabdominal ultrasound (HSC_TAU), magnetic resonance imaging (MRI), and intrauterine fluid saline contrast-enhanced MRI (SCE_MRI). TRM: thickness of residual myometrial. Between-group comparisons were performed using nonparametric tests for related samples; *** P  < 0.001 Multimodal imaging measurements of cesarean scar diverticulum parameters. Length ( A ), width ( B ), depth ( C ), and TRM ( D ) assessed using transvaginal ultrasound (TVU), hysteroscopy-guided transabdominal ultrasound (HSC_TAU), magnetic resonance imaging (MRI), and intrauterine fluid saline contrast-enhanced MRI (SCE_MRI). TRM: thickness of residual myometrial. Between-group comparisons were performed using nonparametric tests for related samples; *** P  < 0.001 For CSD width, SCE_MRI showed the largest measurements (median 17.0 mm), larger than TVU (11.0 mm, P  < 0.001) and MRI (14.0 mm, P   0.0083) (Table  2 ; Fig.  3 B). The biases ranged from − 5.021 to -0.443, and the ICC among the four modalities was 0.567 (Supplementary Fig. 1). For CSD depth, HSC_TAU (median 8.5 mm) and TVU (8.0 mm) measured deeper than MRI (6.3 mm) and SCE_MRI (7.2 mm) (Table  2 ; Fig.  3 C). The biases ranged from − 1.068 to 2.069, and the ICC among the four modalities was 0.81 (Supplementary Fig. 1). For TRM, TVU measured thicker (median 2.7 mm, IQR: 1.9–3.9 mm) than HSC_TAU (2.0 mm), MRI (2.0 mm), and SCE_MRI (1.8 mm) (Table  2 ) Post-hoc tests confirmed TVU measured thicker TRM than all other methods ( P  < 0.001 for SCE_MRI) (Fig.  3 D). The biases ranged from − 0.0020 to 1.008, and the ICC among the four modalities was 0.63 (Supplementary Fig. 1). As shown in Fig.  4 , HSC_TAU detected CSD in 60 of 60 patients (100%), compared to 50/53 (94%) for TVU, 59/61 (97%) for MRI, and 59/61 (97%) for SCE_MRI. Fig. 4 TRM of CSD detected by multimodal imaging. The pie chart illustrates the distribution of TRM measurements for CSD obtained using transvaginal ultrasound (TVU), hysteroscopy-guided transabdominal ultrasound (HSC_TAU), magnetic resonance imaging (MRI), and intrauterine fluid saline contrast-enhanced MRI (SCE_MRI). CSD: Cesarean scar diverticulum; TRM: Thickness of residual myometrial muscle TRM of CSD detected by multimodal imaging. The pie chart illustrates the distribution of TRM measurements for CSD obtained using transvaginal ultrasound (TVU), hysteroscopy-guided transabdominal ultrasound (HSC_TAU), magnetic resonance imaging (MRI), and intrauterine fluid saline contrast-enhanced MRI (SCE_MRI). CSD: Cesarean scar diverticulum; TRM: Thickness of residual myometrial muscle Using TRM < 2.5 mm as the cutoff for severe CSD, TVU classified only 43% of cases as severe, while HSC_TAU classified 68%, MRI 66%, and SCE_MRI 77% (Fig.  3 ). Research indicates a significant correlation between diverticulum length and menstrual bleeding duration. We performed linear regression analyses on diverticulum length measured by the four imaging methods and the patients’ menstrual bleeding duration. As shown in Fig.  5 , CSD length measured by HSC_TAU demonstrated a strong positive correlation with menstrual bleeding duration ( P  = 0.003). Finally, based on the HSC -TAU results, 42 patients who expressed a willingness for repair underwent transvaginal reparied, during which the presence of diverticula was confirmed (Data not displayed). Fig. 5 Linear regression analysis of menstrual duration and CSD length. Linear regression analysis of menstrual duration and cesarean scar diverticulum (CSD) length assessed using TVU ( A ), HSC_TAU ( B ), MRI ( C ), and SCE_MRI ( D ). The red and blue bar charts represent the frequency of CSD length and menstrual bleeding duration, respectively. TVU: transvaginal ultrasound; HSC_TAU: hysteroscopy-guided transabdominal ultrasound; MRI: magnetic resonance imaging; SCE_MRI: intrauterine fluid saline contrast-enhanced MRI; MBD: menstrual bleeding duration Linear regression analysis of menstrual duration and CSD length. Linear regression analysis of menstrual duration and cesarean scar diverticulum (CSD) length assessed using TVU ( A ), HSC_TAU ( B ), MRI ( C ), and SCE_MRI ( D ). The red and blue bar charts represent the frequency of CSD length and menstrual bleeding duration, respectively. TVU: transvaginal ultrasound; HSC_TAU: hysteroscopy-guided transabdominal ultrasound; MRI: magnetic resonance imaging; SCE_MRI: intrauterine fluid saline contrast-enhanced MRI; MBD: menstrual bleeding duration

We enrolled 61 consecutive patients at our Hospital between January 2023 and December 2024. Patients were included if they had at least one prior cesarean delivery, prolonged menstrual bleeding (> 7 days), age 24–40 years, and regular menstrual cycles. We excluded patients with other uterine surgeries, current pregnancy, known uterine pathology (fibroids, adenomyosis, polyps), contraindications to MRI or hysteroscopy, or inability to complete all examinations within one menstrual cycle. The study protocol was reviewed and approved by the ethics committee and conducted in accordance with the Declaration of Helsinki. The Ethics Committee of Xinhua Hospital approved this study (XHEC-H-2018-002). All patients gave written informed consent. All examinations were done during days 7–14 of the same menstrual cycle to minimize the effect of endometrial thickness. The order was: TVU, then MRI, then SCE_MRI, then HSC_TAU, all completed within 7 days. For TVU, we used a Philips EPIQ-5 W system with a 4–9 MHz probe. One experienced sonographer performed all measurements. CSD was identified as an anechoic defect at the anterior lower uterine segment. We took three measurements of length, width, depth, and TRM, and used the mean values. For MRI, we used a SIEMENS Prisma 1.5T system with standard sequences including axial T1-weighted, fat-suppressed T2-weighted, and sagittal T2-weighted images. For SCE_MRI, after standard MRI, we inserted a balloon catheter through the cervix and injected 5 mL saline to better visualize the defect boundaries. For HSC_TAU, we performed continuous-flow hysteroscopy (Karl Storz) using saline as distension medium (pressure 50 mmHg, flow 200 mL/min) while simultaneously measuring CSD parameters with transabdominal ultrasound (GE E10, 3.5-5 MHz probe). We also performed diagnostic curettage to rule out endometrial pathology. We measured CSD length, width, depth, and TRM with each method. We also calculated detection rates and the proportion of severe CSD (No CSD, TRM ≥ 2.5 mm and TRM < 2.5 mm). For continuous variables with normal distribution, data are presented as mean ± standard deviation; for non-normally distributed continuous variables, data are presented as median (interquartile range). Because all four imaging modalities were performed on the same patients during the same menstrual cycle, the data are paired by patient. Continuous variables are presented as median (interquartile range). Overall comparisons among the four modalities for each measurement parameter (CSD length, width, depth, and residual myometrial thickness) were performed using the Friedman test (nonparametric tests for related samples). When the Friedman test indicated a statistically significant difference, post-hoc pairwise comparisons were conducted using the Wilcoxon signed-rank test for dependent samples. All tests were two‑sided, and a P value < 0.0083 was considered statistically significant for pairwise comparisons to control for the risk of type I error. Linear regression analyses were conducted using Hiplot Pro ( https://hiplot.com.cn/ ), a comprehensive web platform for biomedical data analysis and visualization. Statistical analyses were performed using SPSS (version 23.0; IBM Corp., Armonk, NY, USA), while Bland–Altman plots were generated using GraphPad Prism (version 8.0; GraphPad Software, San Diego, CA, USA).

In this study, we compared four imaging methods for CSD evaluation in the same patients during the same menstrual cycle. We found that HSC_TAU had the highest detection rate (100%) and measured the longest CSD length. TVU had the lowest detection rate (94%) and measured thicker TRM than other methods, which may lead to underdiagnosis of severe cases. The high detection rate of HSC_TAU is likely because hysteroscopic irrigation washes out blood and debris from the defect, removing artifacts that can obscure the image. Also, the intrauterine pressure during hysteroscopy expands the collapsed defect, allowing more accurate measurement. The real-time combination of hysteroscopy and ultrasound helps locate the defect precisely [ 1 , 17 ]. The longer CSD length measured by HSC_TAU (median 15.4 mm vs. 6.5–8.2 mm for other methods) probably reflects the true defect size when the diverticulum is expanded rather than collapsed. TRM is the most important parameter for surgical decision-making. A TRM < 2.5 mm is commonly used to recommend surgery [ 1 , 11 , 18 , 19 ]. The same threshold has also been reported as a predictor of uterine rupture risk in subsequent pregnancies [ 20 ] and therefore we adopted it to define “severe CSD” in the present study. Our data showed that TVU measured thicker TRM (median 2.7 mm) than other methods (1.8–2.0 mm). This means TVU classified only 43% of cases as severe, while other methods classified 66–77%. If clinicians rely only on TVU, they may miss patients who would benefit from surgery. Previous studies have reported that uterine rupture in subsequent pregnancies usually occurs when TRM is ≤ 2.5 mm [ 20 ], so reliable TRM measurement is important. Our findings are consistent with previous reports that MRI is more sensitive than TVU for CSD detection [ 15 , 21 ]. A recent review noted that while TVU can diagnose CSD, there is no consensus on the best diagnostic approach [ 1 ]. Our study shows that HSC_TAU may be the best option for preoperative evaluation. In addition to facilitating endometrial biopsy during the examination, the HSC_TAU method allowed us to identify that the CSD length measured by this approach serves as a potential indicator of menstrual bleeding duration. This study has several limitations. First, it is a single-center study with a relatively small sample size ( n  = 61). Second, we did not have surgical or pathological confirmation of CSD dimensions as a gold standard. Third, HSC_TAU is invasive and may not be suitable as a first-line screening tool. Fourth, we did not have follow-up data to see if preoperative measurements predicted surgical outcomes. Based on our findings, we suggest: (1) TVU is acceptable for initial screening because it is cheap and non-invasive, but clinicians should know it may underestimate defect size and overestimate TRM; (2) for patients being considered for surgery, HSC_TAU should be performed to get accurate measurements and assess the endometrium; (3) MRI or SCE_MRI can be used when hysteroscopy is contraindicated.

In conclusion, A uniform method for assessment should be formulated in order to further treatment and evaluation the risk of next pregnant. TVU, MRI, HSC were good tools to detect uterine scar defects, but HSC combined with TAU could make up for the inadequacy of single ultrasound and MRI. Hysteroscopy could easily obtain endometrial pathology, at the same time, evaluated the uterine wound healing of previous cesarean section, giving guidance before next pregnancy to reduce pregnancy related complications caused by diverticula., it has important clinical value. We think HSC_TAU works better because: (1) irrigation removes blood and mucus from the defect; (2) controlled intrauterine pressure expands the defect to its true size; (3) the hysteroscope can help correct uterine position for better ultrasound visualization; and (4) the operator can correlate what they see through the hysteroscope with the ultrasound image in real time. Based on the above, we think it is worth to recommend hysteroscopy combined with transabdominal ultrasound as the potentially optimal diagnostic modality for diagnosis of cesarean section defect.

Cesarean scar defect (CSD), also called isthmocele or niche, is a pouch-like defect in the uterine wall at the site of a previous cesarean incision. As cesarean delivery rates continue to rise worldwide, CSD is now a common clinical problem, affecting 20–70% of women with prior cesarean Sects [ 1 – 4 ]. About one-third of these women experience prolonged menstrual bleeding or postmenstrual spotting [ 5 ]. Beyond abnormal uterine bleeding (AUB) (early-cycle intermenstrual bleeding in FIGO-AUB system) with CSD, CSD can cause secondary infertility and cesarean scar pregnancy [ 6 – 8 ]. Previous studies have shown that bleeding symptoms correlate with CSD size and residual myometrial thickness (TRM) [ 9 , 10 ]. For symptomatic patients, surgical repair is often needed, and TRM is the key factor in deciding whether and how to operate [ 11 , 12 ]. We established a predictive model for the prognosis of CSD repair, and found that it was closely related to the length, width and depth of the diverticulum, meanwhile, uterine position and preoperative thickness of residual myometrium are the key factors affecting postoperative thickness of residual myometrium [ 13 ]. Therefore, accurate preoperative imaging is essential. Several imaging methods can detect CSD, including transvaginal ultrasound (TVU), MRI, sonohysterography, and hysteroscopy [ 1 , 14 ]. TVU is widely used because it is inexpensive and easy to perform. MRI provides better soft - tissue contrast and may be more sensitive than TVU [ 15 ]. However, each method has limitations. TVU measurements can be affected by blood in the defect or the uterine position, while MRI cannot evaluate the endometrium. Hysteroscopy combined with transabdominal ultrasound (HSC - TAU) is a newer approach that allows direct visualization of the uterine cavity while measuring the defect with ultrasound. This combination may overcome the limitations of using either method alone [ 16 ]. In this study, we compared TVU, MRI, SCE_MRI, and HSC_TAU for CSD evaluation in the same patients during the same menstrual cycle, aiming to determine which method provides the most accurate assessment.

Supplementary Material 1: Supplementary Figure 1. Bland–Altman plots comparing HSC TAU with TVU, MRI, and CE MRI for CSD length, width, depth, and residual myometrial thickness (TRM). In each plot, the dashed lines represent the 95% limits of agreement. The intraclass correlation coefficient (ICC, two way random, absolute agreement) among the four methods is displayed at the bottom of each panel. Supplementary Material 1: Supplementary Figure 1. Bland–Altman plots comparing HSC TAU with TVU, MRI, and CE MRI for CSD length, width, depth, and residual myometrial thickness (TRM). In each plot, the dashed lines represent the 95% limits of agreement. The intraclass correlation coefficient (ICC, two way random, absolute agreement) among the four methods is displayed at the bottom of each panel.