Predictive value of a nomogram based on multiparameter quantitative ultrasound for pregnancy outcomes in infertile women attempting natural conception

Infertility, defined as the failure to achieve pregnancy after 12 months or more of regular, unprotected intercourse, has emerged as a critically important and escalating public health issue worldwide, with continuously rising prevalence rates that significantly impact the reproductive health and quality of life of individuals of reproductive age ( 1 , 2 ). In the diagnostic work-up of female infertility, ultrasound has become an indispensable tool; advanced techniques such as four-dimensional hysterosalpingo-contrast sonography (4D-HyCoSy) now allow comprehensive assessment of tubal patency and uterine cavity morphology, effectively ruling out major anatomical barriers to conception ( 3 ). For couples in whom such structural abnormalities have been excluded, natural conception often represents the preferred initial approach owing to its affordability and lower risks. However, successful natural conception depends critically on endometrial receptivity (ER). ER refers to the endometrium’s capacity to accept and support embryo implantation during a specific, brief “implantation window” ( 4 ). ER is the crucial factor determining successful embryo implantation, with approximately two-thirds of implantation failures associated with it ( 5 , 6 ). Current methods for assessing ER, such as the ER array, are invasive and costly, making routine implementation during natural cycles challenging ( 7 , 8 ). Although ultrasonography offers a potential non-invasive alternative ( 9 ), the predictive value of conventional two-dimensional (2D) ultrasound parameters (e.g., endometrial thickness) remains contentious ( 10 - 12 ). Although three-dimensional (3D) power Doppler (PD) can provide quantitative vascular indices, conclusions remain inconsistent ( 13 - 16 ). Emerging technologies such as shear wave elastography (SWE) enable quantitative assessment of endometrial tissue stiffness, with preliminary studies indicating that reduced stiffness may correlate with better receptivity ( 17 , 18 ). Concurrently, the morphological characteristics of the junctional zone (JZ) are increasingly recognized for their relevance ( 19 , 20 ). Nevertheless, the application of these novel techniques in endometrial evaluation remains exploratory. Furthermore, existing research predominantly focuses on populations undergoing assisted reproductive technology (ART), resulting in a scarcity of systematic data for infertile women attempting natural conception. Most notably, there is a lack of a comprehensive, multi-parametric assessment framework capable of integrating traditional parameters with novel indicators such as tissue stiffness and JZ features. This gap underscores the pressing need for more reliable, non‑invasive indicators that can accurately reflect endometrial function and guide the timing of natural conception. Therefore, this study aimed to develop a multiparameter quantitative ultrasound assessment model specifically for infertile women pursuing natural conception. By integrating conventional ultrasound, 3D PD, and SWE, we systematically analyzed multidimensional parameters related to morphology, blood flow, and tissue stiffness. The objective was to establish a more accurate and reliable system for ER evaluation, thereby providing a clinically feasible, non-invasive basis for decision-making and optimizing the diagnostic and therapeutic pathways for infertility. We present this article in accordance with the TRIPOD reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0331/rc ).

This study prospectively enrolled infertile women who visited Shenzhen Second People’s Hospital between September 2023 and December 2024, at which time data collection was completed. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Shenzhen Second People’s Hospital (No. 2023-219-02PJ). Informed consent was provided by all participants. The inclusion and exclusion criteria for this study are illustrated in Figure 1 . All 102 patients in the final analytical cohort completed the study assessments with no missing data, constituting a complete-case dataset. Flowchart of the study population. The inclusion criteria were as follows: (I) aged 20–44 years; (II) regular menstrual cycles (28–35 days); (III) no use of any ovulation-inducing medications during the follow-up period; (IV) all patients voluntarily signed informed consent forms after fully understanding the study content. The exclusion criteria were as follows: (I) congenital uterine anomalies or acquired uterine pathologies such as endometrial polyps, submucosal myoma, intrauterine adhesions, uterine cavity fluid accumulation, or adenomyosis; (II) uterus in midposition; (III) endometriosis; (IV) male factor infertility; (V) use of hormonal medications or history of uterine/ovarian surgery within two months prior to examination; (VI) incomplete laboratory or clinical data. Sample size was estimated using PASS 21.0 (NCSS, Kaysville, UT, USA). Based on an anticipated area under the curve (AUC) of 0.698 and specificity of 62.54% ( 21 ) (α=0.05, power =0.80), the minimum required number of participants was 90. Accounting for a 20% dropout rate (target N=113), the final inclusion of 102 patients ensured sufficient statistical power. Clinical data collected included age, body mass index (BMI), menarche age, menstrual history, fertility history, duration of infertility, factors of infertility (ovulation dysfunction, tubal infertility, unexplained infertility, immunological infertility), and so on. Laboratory data included estradiol (E2), progesterone (P), and serum human chorionic gonadotropin (HCG). Examinations were performed using GE Voluson E8 and E10 ultrasound systems (GE Healthcare, Chicago, IL, USA) and a SuperSonic Aixplorer ultrasound scanner (Hologic, Marlborough, MA, USA), equipped with an RIC 5-9-D probe (5–9 MHz) and an SE 12-3 probe (3–12 MHz), respectively. All ultrasound examinations were performed transvaginally by a senior sonographer with over 15 years of experience during the luteal phase (days 19–24 of the menstrual cycle). Uterine imaging strictly adhered to the Morphological Uterus Sonographic Assessment (MUSA) consensus ( 22 ). Sonographers and clinical outcome assessors were blinded to each other’s findings. Furthermore, all dynamic images and cineloops were reviewed by a second expert to ensure the consistency and standardization of scanning plane selection. Endometrial thickness was measured vertically along the endometrial midline at the thickest portion of the endometrium within the mid-sagittal section. According to Gonen’s ( 23 ) criteria, the endometrium was classified into three types: Type A was a typical three-line structure, with strong echogenic lines on both sides and in the center, and a hypoechoic zone in between; Type B was a uniformly isoechoic endometrium, but the midline showed discontinuity; Type C was a uniformly hyperechoic endometrium without midline echoes in the uterine cavity. Following a 3-minute resting period to ensure baseline stability, endometrial peristalsis was observed continuously for 2 minutes. According to Ljland’s ( 24 ) criteria, the direction of endometrial peristalsis was classified into five types—Type I: endometrial peristalsis from the cervix to the uterine fundus; Type II: peristalsis from the uterine fundus toward the cervix; Type III: endometrial peristalsis originating simultaneously from the uterine fundus and cervix; Type IV: irregular movement; Type V: rest. The frequency of peristalsis was calculated as the total number of peristalsis observed over 2 minutes divided by 2. The proportion of endometrial peristalsis duration was defined as the cumulative duration of peristalsis divided by the total observation time (120 seconds). Uterine artery blood flow was measured bilaterally at the level of the internal cervical os where the angle of insonation was <60°. Continuous and stable spectral Doppler wave forms over ≥3 consecutive cycles were recorded, and the pulsatility index (PI), resistance index (RI), and systolic/diastolic ratio (S/D) were measured and averaged for both sides. On the coronal view where the JZ was most clearly visualized in 3D mode, the maximum (JZ max ) and minimum (JZ min ) thicknesses of the JZ at both lateral walls and the fundus were measured ( 25 ). These values were then summed and divided by six to calculate the mean JZ thickness (JZ mean ). In PD mode, the plane with the richest endometrial and subendometrial vascularity was selected, and vessels were counted and classified into four types—Type I: no vessels; Type II: 1–2 vessels; Type III: 3–4 vessels; Type IV: ≥5 vessels ( 26 ). Endometrial and subendometrial blood flow distribution were categorized as follows ( 27 )—Type 0: no blood flow in the endometrium or subendometrial area; Type I: blood flow detected only in the subendometrial area; Type II: blood flow occupying no more than half of the endometrium; Type III: blood flow occupying more than half of the endometrium and even reaching the endometrial canal. After PD assessment, 3D volume scans were acquired and stored for offline analysis. Endometrial and JZ volume, vascularization index (VI), flow index (FI), and vascularization FI (VFI) were calculated using built-in machine software ( Figure 2 ). The mode was then switched to SWE. A standardized “no-pressure” technique was strictly applied to avoid artificial tissue stiffening caused by probe-induced pressure. Quantitative analysis was performed using the Q-BOX tool to automatically calculate the mean Young’s modulus (Emean) within regions of interest (ROI) placed over the JZ and endometrium. The size of the ROI was adjusted according to the average thickness of the endometrium and JZ ( 28 ) ( Figure 3 ). 3D power Doppler evaluation of the endometrium and JZ. (A) Endometrial 3D vascularization parameters. (B) JZ 3D vascularization parameters. 3D, three-dimensional; JZ, junctional zone. SWE evaluation of the endometrium and JZ. (A) Endometrial stiffness. (B) JZ stiffness. JZ, junctional zone; SD, standard deviation; SWE, shear wave elastography. Clinical pregnancy was defined as the presence of an intrauterine gestational sac with cardiac activity, confirmed by transvaginal ultrasound at 6 weeks following conception. This study used the software SPSS 26.0 (IBM Corp, Armonk, NY, USA) and Empower (X&Y Solutions, Inc., Boston, MA, USA). Continuous variables were described as mean (standard deviation) or median (interquartile range) depending on whether they follow a normal distribution, whereas categorical variables were described as frequency (percentage). Next, the Mann-Whitney U test, X 2 test, or Fisher’s exact test was used to compare differences between two groups. Intra-observer reliability was assessed using the intraclass correlation coefficient (ICC). For this purpose, 30 participants were randomly selected for repeated measurements by the same sonographer after a one-month interval. This reliability analysis focused on the key predictive variables utilized in the final nomogram. Variable selection was performed using least absolute shrinkage and selection operator (LASSO) regression with 10-fold cross-validation, and the optimal lambda (λ) was determined based on the minimum criterion (λ.min). During model development, all continuous variables were retained as original values to preserve data integrity. A nomogram was constructed, and receiver operating characteristic (ROC) curves were plotted to assess predictive efficacy. The model was internally validated using 500 bootstrap samples. Calibration curves were plotted to verify the model’s predictive ability, and decision curve analysis (DCA) was used to assess the model’s clinical utility. A P<0.05 was considered statistically significant.

A total of 102 infertile women who met the inclusion criteria were enrolled in this study, including 22 in the pregnancy group and 80 in the non-pregnancy group. There was a statistically significant difference in infertility duration between the two groups with different pregnancy outcomes (P=0.040), whereas other baseline data showed no significant differences (P>0.05). The summary of baseline data for both groups is presented in Table 1 . Data are presented as n (%), mean ± standard deviation or median (interquartile range). BMI, body mass index; E2, estradiol; P, progesterone. As shown in Table 2 , there were no statistically significant differences between the two groups of patients in terms of endometrial morphology, endometrial peristalsis frequency, uterine artery blood flow characteristics, and endometrial FI (P>0.05). There were significant differences between the two groups of patients in endometrial thickness, direction of endometrial peristalsis, proportion of endometrial peristalsis duration, number of endometrial and subendometrial vessels, vascular pattern and distribution of endometrial and subendometrial blood flow, endometrial volume (EV), endometrial VI, endometrial VFI, and endometrial Emean (P<0.05). Data are presented as n (%), mean ± standard deviation or median (interquartile range). † , Type A was a typical three-line structure, with strong echogenic lines on both sides and in the center, and a hypoechoic zone in between; Type B was a uniformly isoechoic endometrium, but the midline showed discontinuity; Type C was a uniformly hyperechoic endometrium without midline echoes in the uterine cavity. ‡ , Type I: no vessels; Type II: 1–2 vessels; Type III: 3–4 vessels; Type IV: ≥5 vessels. § , Type 0: no blood flow in the endometrium or subendometrial area; Type I: blood flow detected only in the subendometrial area; Type II: blood flow occupying no more than half of the endometrium; Type III: blood flow occupying more than half of the endometrium and even reaching the endometrial canal. Emean, mean Young’s modulus; FI, flow index; PI, pulsatility index; RI, resistive index; S/D, systolic to diastolic velocity ratio; UtA, uterine artery; VFI, vascularization flow index; VI, vascularization index. As shown in Table 3 , there were no statistically significant differences between the two groups of patients in terms of JZ thickness, JZ morphology, JZ-V, and JZ-FI (P>0.05). However, there were significant differences between the two groups regarding JZ-VI, JZ-VFI, EV/JZ-V, and JZ-Emean (P<0.05). Data are presented as n (%) or median (interquartile range). Emean, mean Young’s modulus; EV, endometrial volume; FI, flow index; JZ, junctional zone; JZ-V, junctional zone-volume; JZd, junctional zone difference; VFI, vascularization flow index; VI, vascularization index. Based on the LASSO regression lambda.min, six predictors associated with pregnancy outcomes were selected: proportion of endometrial peristalsis duration, direction of endometrial peristalsis, EV, endometrial Emean, JZ-Emean, and JZ-VFI, as shown in Figure 4 . The formula was as follows: −1.98341+1.04563 × proportion of endometrial peristalsis duration + 2.53596 × direction of endometrial peristalsis + 0.12443 × EV − 0.01686 × endometrial Emean − 0.02398 × JZ-Emean + 0.92751 × JZ-VFI. Feature selection using the LASSO algorithm. (A) The coefficient convergence graph for the feature selection process. (B) The ordinate is the binomial deviation, and the abscissa is log (λ). LASSO, least absolute shrinkage and selection operator. Intra-observer reproducibility of the six selected parameters was assessed in 30 randomly selected participants with repeat measurements after a one-month interval. The ICCs ranged from 0.750 to 0.851 (all P<0.001), indicating good reproducibility ( Table S1 ). A nomogram was constructed based on the six selected predictors to provide a quantitative tool for predicting pregnancy probability ( Figure 5 ). To use the nomogram, the scores for each of the six variables were determined and summed to obtain the total points, which were then projected onto the ‘Risk’ axis to estimate the individualized predicted probability of pregnancy. Based on the ROC analysis, the optimal threshold for the linear predictor was identified as –1.89, which corresponded to a total score of approximately 36 points on the nomogram. The nomogram for predicting the probability of successful pregnancy in women with infertility opting for natural conception. † , 0: nonforward; 1: forward. Emean, mean Young’s modulus; EV, endometrial volume; JZ, junctional zone; VFI, vascularization flow index. As shown in Figure 6 , the model demonstrated robust diagnostic performance with an AUC of 0.862 [95% confidence interval (CI): 0.760–0.964]. The sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) were 0.882, 0.758, 0.783, 0.484, and 0.962, respectively. The ROC of the nomogram. AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic. The nomogram was internally validated using 500 bootstrap samples to reduce overfitting. As shown in Figure 7A , after internal validation, the model achieved an AUC of 0.855 (95% CI: 0.750–0.939). Additionally, the calibration curve showed good agreement ( Figure 7B ), supported by a Brier score of 0.066 and an Omnibus test P value of 0.051. The decision curve indicated good clinical utility ( Figure 7C ). Internal validation of the predictive model. (A) The AUC of the internal validation of our model. (B) The calibration curve of the model. (C) The DCA of the model. AUC, area under the curve; CI, confidence interval; DCA, decision curve analysis.

This study developed and validated a multiparameter ultrasound model for predicting pregnancy outcomes in infertile women attempting natural conception. We found that a nomogram integrating the proportion of endometrial peristalsis duration, the direction of endometrial peristalsis, EV, endometrial Emean, JZ-VFI, and JZ-Emean demonstrated good predictive performance. These findings suggest that a multidimensional comprehensive assessment of endometrial functional characteristics, tissue structural integrity, and biomechanical properties represents a crucial approach to achieving a holistic evaluation of ER. Our findings regarding conventional 2D ultrasound parameters align with existing controversies in the field ( 5 , 29 ). Although the endometrial thickness in the pregnant group was significantly greater than that in the non-pregnant group, its limited value as an independent predictive factor has been widely recognized ( 5 ). This suggests that adequate endometrial thickness is more likely a necessary but insufficient condition. Similarly, endometrial pattern showed no significant association with pregnancy outcomes in this study, likely because our assessment occurred during the mid-luteal phase, which had already passed the optimal window for observing the typical “triple-line sign”. Regarding blood flow assessment, this study showed that local microcirculation holds greater predictive value for pregnancy outcomes than main vessel perfusion. In contrast to uterine artery blood flow parameters lacking predictive significance, endometrial and subendometrial microvascular characteristics demonstrated substantial predictive value. The pregnant group exhibited a greater number of blood flow branches and a higher grade of blood flow classification, consistent with previous studies ( 15 , 26 ). Notably, the JZ-VFI demonstrated unique predictive value for pregnancy outcomes. As a specialized transitional zone between the endometrium and myometrium, the JZ serves as the critical entry point for spiral arteries after their winding course ( 30 ). Its vascular status may reflect endometrial blood supply earlier and more sensitively than the endometrium itself. Therefore, for patients with inadequate microcirculatory perfusion, the possibility of poor receptivity should be considered even with normal main vessel parameters, warranting timely clinical intervention. The findings of our study indicate that a positive direction of endometrial peristalsis combined with a higher proportion of peristalsis duration collectively creates a dynamic environment conducive to embryo implantation. Specifically, positive peristalsis primarily facilitates the initial directional transport of embryos within the uterine cavity, whereas subsequent positioning and adhesion processes rely more heavily on sustained peristaltic activity ( 31 , 32 ). Consequently, the proportion of peristalsis duration emerges as a critical factor influencing the ultimate efficiency of implantation. Further analysis revealed that even with a lower per-unit-time contraction frequency, a longer duration per contraction (i.e., a higher proportion of peristalsis duration) can still provide a more sustained and stable microenvironment for embryo-endometrium interactions. This sustained support is crucial for the embryo to complete a series of complex biological steps, including positioning, recognition, and adhesion. Therefore, given the presence of positive peristalsis, the proportion of peristalsis duration can serve as a reliable indicator for assessing ER. This discovery offers new insights for optimizing diagnostic and therapeutic strategies for infertility patients. Notably, SWE revealed that lower endometrial stiffness promotes implantation, as evidenced by the significantly lower Emean in the pregnancy group, consistent with findings by Li et al. ( 28 ). More importantly, this study is the first to quantify JZ stiffness via SWE, revealing that lower JZ stiffness is a robust predictor of pregnancy outcomes. This specific correlation has not been previously documented in the literature. We hypothesize that a more compliant JZ promotes embryo implantation through several synergistic physiological mechanisms. First, diminished JZ stiffness may optimize the coordination and amplitude of uterine peristalsis, thereby facilitating precise embryo transport and positioning ( 33 ). Second, a flexible JZ likely lowers the physical resistance to spiral artery remodeling, which enables more effective trophoblast invasion and increases blood flow to the intervillous space ( 34 ). Third, reduced JZ stiffness may alleviate extrinsic compression on the radial and spiral arteries, further enhancing global endometrial perfusion ( 35 ). Collectively, these biomechanical advantages converge to create a uterine milieu that is highly conducive to embryo implantation. These findings provide a novel perspective on the role of JZ stiffness in the assessment of ER. The nomogram developed in this study demonstrates favorable predictive performance by transforming complex quantitative ultrasound parameters into a user-friendly clinical tool. Clinicians can determine the corresponding points for each indicator on the nomogram and sum them to obtain a total score, where a higher score indicates a greater probability of pregnancy. Based on the ROC analysis, the optimal threshold for the linear predictor was identified as –1.89, which corresponds to a total score of approximately 36 points. This threshold serves as a practical cutoff for risk stratification; patients scoring below 36 points (predicted probability <13%) are categorized as a high-risk group for conception failure. In such scenarios, the model provides a tangible basis for clinicians to refine management, such as guiding the optimal timing of conception, recommending targeted pre-conception conditioning, or performing further clinical intervention assessments. Although the model exhibits high sensitivity, the specificity and PPV indicate a relatively high rate of false positives. This suggests that although the nomogram is effective at identifying the majority of patients who may achieve pregnancy, nearly half of those predicted to be successful may not conceive within the observed period. Nevertheless, compared to the baseline pregnancy rate of 21.6% in this cohort, the model significantly improves prognostic insight. Therefore, it is best positioned as a supportive counseling tool to help clinicians identify patients with favorable physiological conditions, rather than a standalone predictor of clinical outcome. Furthermore, although the model incorporates advanced ultrasound capabilities, its feasibility is ensured by the fact that all parameters are acquired non-invasively during routine examinations. As high-resolution ultrasound systems become increasingly standard in reproductive centers, this approach provides a practical and cost-effective tool for daily clinical practice. Notably, the nomogram was constructed primarily using quantitative ultrasound parameters. Although infertility duration differed between groups, this focused design was chosen to directly assess the predictive value of objective physical and vascular markers of the uterus. Although infertility duration provides important clinical context, the core objective of this study was to evaluate the immediate uterine physiological state as reflected by sonographic features. This study has several limitations. First, the modest sample size and imbalanced outcome groups resulted in a low events-per-variable ratio, which may introduce bias. Although LASSO and bootstrap internal validation were applied, validation in larger, multicenter cohorts is warranted. Second, the single-center design and exclusion of midposition uteri due to suboptimal acoustic windows may limit generalizability. Third, all examinations were performed by a single sonographer, and interobserver reproducibility was not assessed. Additionally, single-time-point luteal phase assessments may not capture dynamic uterine changes across the menstrual cycle, and the effects of different peristaltic directions were not explored. Finally, the findings may not directly generalize to women with irregular cycles, as the study enrolled only those with regular cycles not using hormonal medications.

This study established a multiparametric ultrasound-based nomogram that provides prognostic stratification for pregnancy outcomes in infertile women attempting natural conception. By integrating indicators of endometrial function, structure, and biomechanics, the model offers clinicians an objective reference to identify patients with favorable uterine conditions. Its primary value lies in supporting counseling and shared decision-making, particularly in determining the timing of conception attempts and the need for pre-conception intervention. All parameters are acquired non-invasively during routine examination, facilitating integration into daily clinical practice. Future prospective multicenter studies are still warranted to validate these findings.

The article’s supplementary files as