A comparative study of shear wave elastography on preterm birth risk in women with natural pregnancies and in vitro fertilization embryo transfer pregnancies.

Preterm birth is defined as a live birth before 37 weeks of gestation. Globally, approximately 15 million newborns are born preterm every year ( 1 ), the majority of which are spontaneous preterm births (SPBs). Preterm birth has serious impacts on both patients and society, including neonatal death, long-term health problems, and psychological trauma, and imposes a heavy economic burden on families and health care systems. Some studies have shown that the incidence of preterm birth among women who conceived via in vitro fertilization (IVF) is greater than that among women who conceived naturally ( 2 ), and due to the preciousness of IVF, more attention has been given to the prevention of preterm birth. Current diagnostic and therapeutic tools for preterm birth include interventions such as pregnancy monitoring, the use of medications to delay labor, and cervical cerclage, but these methods are limited in terms of effectiveness and indications. Shear wave elastography (SWE) is a technique that uses a probe to generate mechanical impulses that propagate through the tissue, resulting in tissue displacements that can be directly measured in terms of shear wave velocity (m/s) or indirectly measured in terms of Young’s modulus (kPa), resulting in tissue stiffness values that can be used to assess tissue structure. Due to its noninvasive, reproducible, and real-time nature, SWE has been applied in the fields of obstetrics and gynecology, including for the diagnosis of uterine fibroids, adenomyosis, and cervical cancer ( 3 , 4 ). The value of using SWE to evaluate cervical function during pregnancy in the prediction of preterm birth has been gradually emphasized in recent years ( 5 ). However, most studies have been limited to women with natural pregnancies, and there are fewer reports of ultrasound evaluation of cervical function in pregnant women who conceived via assisted reproduction. The aim of this study was to analyze the predictive value of the multifactorial combination of Young’s modulus values of cervical elasticity combined with the cervical length (CL), cervical angle, and endocervical os morphology in pregnant women who conceived naturally and those who conceived via IVF embryo transfer (IVF-ET) by applying a real-time SWE technique for the prediction of preterm birth occurring in different modes of conception. The results of this study will help to identify pregnant women at risk for preterm birth and to develop targeted interventions to reduce the incidence of preterm birth in pregnant women who conceived via IVF-ET and are at risk for preterm birth, thereby improving maternal and neonatal health ( 1 ). We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1223/rc ).

A total of 250 pregnant women at 16–28 weeks of gestation who attended Baoji Central Hospital for prenatal checkups from June 2022 to June 2024 were included. Among these women, 150 conceived naturally and were aged 21–44 years, with a mean age of 31.39±4.18 years, and 100 conceived via IVF-ET and were aged 23–42 years, with a mean age of 30.42±4.16 years. The inclusion criteria were as follows: (I) routine obstetric examination at our hospital during the study period at 16–28 weeks of gestation, conception via IVF-ET or naturally, singleton pregnancy, and no abnormal fetal development; (II) conception via IVF-ET or naturally, singleton pregnancy at 16–28 weeks of gestation with symptoms of preterm labor, and seeking preterm labor intervention during the study period. The exclusion criteria were as follows: (I) twin or multiple pregnancies; (II) excessive amniotic fluid, placenta previa, or vaginal bleeding during pregnancy; (III) cervical conization, loop electrosurgical excision procedure (LEEP), or cervical cerclage before the examination; (IV) maternal factors for early termination of pregnancy. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Baoji Central Hospital (Ethics Approval No. BZYL2024-31). All examiners were informed of the purpose of the examination before the examination and provided written informed consent. An Aixplorer (SuperSonic Imagine, Aix-en-Province, France) full digital color Doppler ultrasound diagnostic instrument with an intracavitary probe (SE12-3) and a probe frequency of 3–12 MHz was used. Examination was performed as follows: A pregnant woman was instructed to empty her bladder and lie supine on the examination table with legs flexed and vulva exposed, and cervix measurements were collected in accordance with the guidelines for mid-pregnancy cervical measurements recommended by the British Fetal Medicine Foundation (FMF) ( 6 ). The operator placed the intracavitary probe slowly and gently in the anterior vaginal dome, with the probe just touching the external os of the uterine cervix and not squeezing the cervix, to determine the median sagittal plane of the uterine cervix and to carry out the two-dimensional (2D) and elasticity measurements. Three measurements were taken and recorded: (I) the cervical canal length and endocervical os morphology ( Figure 1A ); (II) the cervical angle, which was the angle between the line between the endocervical opening and the external opening and the projected line between the endocervical opening and the lower part of the anterior wall more than 3 cm away from the endocervical opening ( 7 ), measuring with a protractor; and (III) elastic Young’s modulus values of cervical elasticity, applying SWE with the blue sampling area covering the cervix and the region of interest (ROI) size of the Q-BOX adjusted to 5 mm after the image was stabilized for 3–4 seconds and using the anterior lip of the endocervical os as point A, the anterior lip of the exocervical os as point B ( Figure 1B ), the posterior lip of the endocervical os as point C, and the posterior lip of the exocervical os as point D for measurements ( Figure 1C ). The patient was a 24-year-old female at 19 +1 weeks of gestation. (A) Cervical length, 4.0 cm; the internal opening of the cervix was closed in a T shape. (B) The SWE measurements of the Young’s modulus values at point A (the inner opening of the upper lip) and point B (the outer opening of the upper lip) of the cervix were all relatively low, Emean <15 kPa. (C) The SWE measurement showed that the Young’s modulus values at point C (the inner opening of the lower lip) and point D (the outer opening of the lower lip) were very low, Emean <13 kPa. The patient eventually presented with preterm birth at 30 weeks and was selected for cervical cerclage. SD, standard deviation; SWE, shear wave elastography. Telephone follow-up or inpatient record evaluation was performed to determine pregnancy outcomes. Preterm birth was defined as delivery occurring after 20 weeks of pregnancy but before 37 completed weeks of gestation. Preterm labor was defined as labor at 28–37 weeks of gestation, with regular or irregular contractions lasting more than 30 minutes and progressively increasing cervical shortening ( 8 ). In this study, women at risk for preterm birth included those with preterm labor or cervical insufficiency diagnosed by ultrasound. Excel (Microsoft, Redmond, WA, USA) was used for data entry, SPSS 26.0 (IBM Corp., Armonk, NY, USA) was used to process the baseline and clinical data, and R 4.2.6 and RStudio software (R Foundation for Statistical Computing, Vienna, Austria) were used to construct the risk prediction model. The measurements of the baseline and clinical data that conformed to the normal distribution were expressed as the means ± standard deviations, and comparisons between groups were made using the χ 2 test. The above test level was α=0.95, and differences were considered statistically significant at P<0.05.

The general clinical data of the two groups were analyzed, and the differences between the IVF-ET group and the natural pregnancy group in terms of age, gestational week, time of menarche, and history of induced abortion were not statistically significant and were comparable. The prevalence of comorbidities during pregnancy was 35.33% in the natural pregnancy group and 47% in the IVF-ET group. The prevalence of comorbidities during pregnancy was higher in the IVF-ET group than in the natural pregnancy group, with the comorbidity with highest prevalence being diabetes mellitus. There was a statistically significant difference in the history of full-term birth between the two groups, at 30.67% in the natural pregnancy group and 17% in the IVF-ET group. When the 2D ultrasound measurements were compared, there was no statistically significant difference in the CL, cervical angle, or endocervical os morphology between the two groups (P>0.05) ( Table 1 ). Data are presented as mean ± standard deviation or n (%). CL, cervical length; IVF-ET, in vitro fertilization embryo transfer; UCA, uterocervical angle. A comparison of Young’s modulus values of elasticity in different parts of the cervix between the two groups revealed that the Emean (kPa), Emin (kPa), and Emax (kPa) values of the cervix at all points were greater in the IVF-ET group than they were in the natural pregnancy group. The differences in the Emean (kPa), Emin (kPa), and Emax (kPa) values at point A and point C were statistically significant (P<0.05). Emean (kPa), Emin (kPa), and Emax (kPa) at point B were statistically significant (P0.05) ( Table 2 ). Data are presented as mean ± standard deviation. IVF-ET, in vitro fertilization embryo transfer; ROI, region of interest. The incidence of preterm birth was 6.0% in the IVF-ET group and 4.67% in the natural pregnancy group. The percentages of final preterm birth, preterm birth with medication to preserve the pregnancy, and full-term cesarean section in the IVF-ET group were slightly greater than those in the natural pregnancy group, but the difference between the two groups was not statistically significant (P>0.05) ( Table 3 ). Data are presented as n (%). IVF-ET, in vitro fertilization embryo transfer. A comparison of full-term birth with preterm birth in the natural pregnancy group revealed that the cervical canal length, point A Emean (kPa), and endocervical os morphology affected whether preterm birth occurred. A point A Emean value less than 15.3 kPa was associated with an increased risk of preterm birth ( Table 4 ). Data are presented as mean ± standard deviation or n (%). CL, cervical length; UCA, uterocervical angle. A comparison of term and preterm birth in the IVF-ET group revealed that the CL point A Emean (m/s), point A Emin (kPa), point A Emax (kPa), point C Emean (kPa), point C Emin (kPa), and endocervical os morphology affected whether preterm birth occurred. A point A Emean less than 14.7 kPa, point A Emin less than 10.2 kPa, point A Emax less than 21.6 kPa, point C Emean less than 18.8 kPa, and point C Emin less than 15.3 kPa led to an increased risk of preterm birth ( Table 5 ). Data are presented as mean ± standard deviation or n (%). CL, cervical length; IVF-ET, in vitro fertilization embryo transfer; UCA, uterocervical angle. The multifactorial logistic regression model was established by screening the statistically significant variables in Table 5 , and the results revealed that endocervical os morphology, cervical canal length, and point A Emean were independent factors affecting preterm delivery of IVF-ET. The area under the curve (AUC) for the logistics regression model was 0.871, sensitivity was 0.80, specificity was 0.86, and the Youden index was 0.70. In the multifactorial logistic regression model, the cervical morphology of “T” complete closure, large cervical canal length, and large point A Emean kPa value were protective factors for preterm delivery in patients who had underwent IVF-ET, and a point A Emean kPa value of less than 14.7 kPa led to an increased risk of preterm ( Table 6 ). CL, cervical length; IVF-ET, in vitro fertilization embryo transfer; CI, confidence interval; OR, odds ratio; SE, standard error. The variables screened by the multifactorial logistic regression model were built into a random forest model and compared with the logistic model, and the predicted receiver operating characteristic (ROC) curves were plotted. The AUC value of the random forest model was 0.909, which was greater than that of the logistic regression model, which was 0.871 ( Figures 2,3 ). Forest model for predicting the occurrence of preterm birth in patients who underwent IVF-ET. CI, confidence interval; CL, cervical length; IVF-ET, in vitro fertilization embryo transfer; OR, odds ratio. Comparison of the ROC curves of the models for predicting the occurrence of preterm birth in the IVF-ET group. IVF-ET, in vitro fertilization embryo transfer; ROC, receiver operating characteristic.

Several studies have reported that the risk of preterm delivery is greater in pregnancies that were conceived via IVF-ET than in natural pregnancies ( 9 - 11 ), and in this study, we analyzed and compared the general data and singleton pregnancy outcomes of women who underwent IVF-ET with those who conceived naturally. A retrospective analysis of various ultrasound measurements of pregnant women with actual preterm birth and those at risk of preterm birth as determined via ultrasound diagnosis was performed, using cervical SWE to evaluate the cervix of women who underwent IVF-ET and to identify pregnant women with potential risk for preterm birth or cervical insufficiency. The significance of this study is that it has the potential to improve the clinical management of IVF-ET pregnancies and provide a basis for targeted interventions and monitoring programs. In this study, ultrasound SWE revealed that although there were no statistically significant differences between the IVF-ET group and the natural pregnancy group in terms of traditional cervical assessment indices, including the CL, endocervical os morphology, and cervical angle, the Young’s modulus of cervical elasticity of the pregnant women in the IVF-ET group was significantly greater than that of women in the natural pregnancy group; the cervical elasticity of pregnant women with preterm birth was significantly lower than that of those with full-term births, regardless of whether they were in the IVF-ET group or the natural pregnancy group. In both the IVF-ET group and the natural pregnancy group, Young’s modulus of cervical elasticity was significantly lower in the preterm birth group than it was in the full-term birth group; the combination of the CL and endocervical os morphology with the value of the elasticity Young’s modulus of the upper lip of the endocervical os could better predict the occurrence of preterm birth in different modes of conception. The occurrence of preterm birth is the result of a multifactorial and multimechanism combination of many potential etiologic factors and risk factors, and the assessment of cervical maturity is the key to predicting SPB, in which changes in cervical morphology and texture play important roles in the development of SPB. The cervical tissue consists of a collagen-rich extracellular matrix, and the collagen concentration in the pregnant cervix decreases with the number of deliveries ( 12 ). The significantly greater number of second or multiple pregnancies in the natural pregnancy group than in the IVF-ET group in the present study may have contributed to the lower value of Young’s modulus of cervical elasticity in the natural pregnancy group than in the IVF-ET group. Thus, the greater the number of births is, the lower the value of cervical elasticity. The incidence of combined hypertension, diabetes mellitus, and thyroid dysfunction in singleton IVF-ET pregnancies is higher than that in singleton natural pregnancies ( 13 ), especially the incidence of combined diabetes mellitus, which occurs at the highest rate ( 14 ). In this study, the prevalence of gestational comorbidities such as diabetes mellitus, hypertension, abnormal thyroid function, and intrahepatic cholestasis were greater in the IVF-ET group than in the natural pregnancy group, and the prevalence of diabetes mellitus was the greatest, which is in line with the results of previous studies. However, through the follow-up of the pregnancy outcomes of the two groups, it was found that although there was no statistically significant difference in the incidence of preterm birth between the two groups, the rates of preterm birth and preterm birth with final cesarean section in the IVF-ET group were slightly higher than those in the natural pregnancy group. For the study of cervical elasticity, Yamaguchi et al. first reported that elastography of the gestational cervix could be used to predict preterm birth ( 15 ). Earlier studies of cervical elasticity to predict preterm birth focused on static strain cervical elastography. Strain imaging involves inducing tissue deformation using quasi-static methods such as manual pressure, probe compression, or cardiovascular/respiratory stimulation. By measuring this deformation, it displays the distribution of strain within ROIs. Semiquantitative elasticity scores can be performed for analysis via strain imaging ( 16 ), and a number of studies have confirmed that the strain value of the endocervical os is strongly associated with spontaneous preterm birth ( 17 , 18 ). Additionally, studies have demonstrated that cervical strain elasticity imaging can quantitatively assess cervical remodeling sensitivity following exogenous prostaglandin administration ( 19 ). SWE evaluates tissue structure by generating mechanical pulses through a probe. Shear wave propagation induces tissue displacement, enabling direct measurement of shear wave velocity (m/s) or indirect determination of tissue Young’s modulus (kPa), thereby assessing tissue elasticity. Compared to strain imaging, SWE is less susceptible to operator subjectivity. It not only enables the direct visualization of color distribution but also provides direct statistical data on shear wave velocity or Young’s modulus. Sandrin et al. ( 20 ). Applied SWE to measure the elasticity of different parts of the cervical canal and reported that the value of elasticity of the endocervical os and exocervical os had different values in predicting preterm birth and that the value of elasticity of the endocervical os had a greater predictive value for preterm birth. The study in this paper confirms this observation and concludes that the upper lip of the endocervical os is of greater predictive value. Several studies have shown that the CL combined with endocervical os elasticity values has greater predictive value for preterm birth. During pregnancy, the cervix undergoes a remodeling process that consists of four phases, namely, softening, ripening, dilatation, and repair ( 21 ), and its elasticity decreases progressively with increasing gestation time ( 22 ). Clinically, the Bishop score is a traditional method for assessing cervical firmness. Ultrasound-measured cervical elasticity correlates with cervical texture in the Bishop score, with softer cervixes being more susceptible to dilation 16 . This can trigger preterm birth. The cervical elasticity of preterm patients is lower than that of those with full-term births ( 23 ), and Young’s modulus values of cervical elasticity at all points were lower in preterm patients than they were in those with full-term births in both groups. In addition, Young’s modulus values of the elasticity of the upper lip of the inner mouth and the lower lip of the inner mouth of the cervix were strongly associated with preterm birth in the present study ( 24 ), especially the value of the cervical elasticity of the upper lip of the inner mouth, which was the most sensitive predictor of preterm birth. The upper lip of the inner mouth was the least affected by surrounding external forces, was the most stable, had an important role in the maintenance of cervical function, and was an independent predictor of preterm birth ( 25 ). The measurement of CL is the most basic tool for screening for preterm birth ( 26 ), especially by transvaginal ultrasound, which is now becoming an important method for evaluating preterm birth, and a CL of less than 25 mm is an important cutoff value in the middle trimester ( 27 ). However, some softer cervixes may not be short in earlier periods of pregnancy. In the present study, three patients had CL measurements of 25 mm or more at the time of initial measurement, but the elasticity values of the upper and lower lips of the endocervical opening were less than 12 kPa ( Figure 1 ). Ultimately, shortening of the cervix (<15 mm) occurred, the endocervix was opened, and ultimately, cervical cerclage was clinically performed. The present study confirmed that cervical function can be more accurately assessed by the combination of CL and cervical elasticity measurements ( 28 ). This study provides valuable information to guide obstetricians in administering appropriate fertility preservation therapy in patients with cervical insufficiency before cervical shortening. The funnel-shaped opening of the endocervical os has been shown in many studies to be of great value in the prediction of preterm birth ( 29 ), but in practice, the shallower “V” shaped funnel in CLs greater than 30 mm does not seem to be of significant value. In practice, only when the cervix is less than 30 mm and the endocervix appears to be open in the shape of a funnel is there some diagnostic value for preterm birth ( Figure 4 ). The patient was a 26-year-old female at 19 +4 weeks of gestation. (A) Shortening of the cervix, measured at 1.5 cm, with a V-shaped change in the internal opening. (B) The SWE measurements of the Young’s modulus values at point A (the inner opening of the upper lip) and point B (the outer opening of the upper lip) of the cervix were all relatively low, Emean <12 kPa. (C) The SWE measurement showed that the Young’s modulus values at point C (the inner opening of the lower lip) and point D (the outer opening of the lower lip) were very low, Emean <11 kPa. The patient eventually presented with preterm birth at 19 +6 weeks and was selected for cervical cerclage. SD, standard deviation; SWE, shear wave elastography. The above indicators for predicting preterm birth in patients who underwent IVF-ET did not significantly differ from those with natural pregnancies, indicating that different modes of conception did not significantly affect cervical morphology, but the cervical texture, specifically, cervical elasticity, changed, which highlights the importance of this study. The multifactorial combination of cervical morphology combined with cervical elasticity may improve the accuracy of the prediction of preterm birth in different modes of conception. The limitations of this study include the relatively small sample size and the lack of long-term follow-up data and multicenter validation, which may affect the general applicability of the results. In addition, this study failed to incorporate experimental data analysis and lacked an in-depth exploration of the effects of biomarkers, such as fetal fibronectin attachment, on preterm birth. In future work, we will expand the sample size and increase the experimental data analysis to account for these shortcomings.

This study revealed significant differences between women with IVF-ET pregnancies and women with natural pregnancies in terms of pregnancy comorbidities, pregnancy outcomes, and ultrasound measurements, emphasizing the risk factors for preterm delivery and their importance in clinical management and providing a basis and direction for future related research.

The article’s supplementary files as