Prediction of reproductive outcomes in unicornuate uterus women based on Three-dimensional MRI radiomic features

Prediction of reproductive outcomes in unicornuate uterus women based on Three-dimensional MRI radiomic features | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Prediction of reproductive outcomes in unicornuate uterus women based on Three-dimensional MRI radiomic features Hui Luo, Qing Zhou, Xinyu Pan, Guofu Zhang, Keqin Hua, Jingxin Ding This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9157364/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 11 You are reading this latest preprint version Abstract Purpose: To develop and validate a radiomics nomogram integrating three-dimensional (3D) MRI features for predicting ≥35 weeks' gestation in women with unicornuate uterus. Methods: This retrospective study enrolled 170 unicornuate uterus patients who underwent pelvic 3D MRI between 2005-2022. Patients were categorized into four groups based on reproductive outcomes: primary infertility, <24 weeks, 24-35 weeks, and ≥35 weeks gestation. The uterus was segmented and reconstructed in 3D using ITK-SNAP. Radiomics features (n=1,834), including volume, surface area, sphericity, and diameters, were extracted. Feature selection employed LASSO regression, and eight machine learning algorithms were compared for model construction. A radiomics nomogram integrating selected features with clinical variables was developed. Predictive performance was assessed using ROC and decision curve analyses. Results: Uterine volume was significantly larger in the ≥35 weeks gestation group compared to infertility and <24 weeks groups (P<0.05). Surface area was also greater in the ≥35 weeks group versus the infertility group (P<0.05). No significant differences in axis lengths were observed among groups. The radiomics nomogram demonstrated robust discrimination for predicting ≥35 weeks gestation, achieving AUCs of 0.86 (95% CI: 0.79-0.92) in the training cohort and 0.84 (95% CI: 0.69-0.98) in the validation cohort. Decision curve analysis confirmed favorable clinical utility. Conclusion: Uterine volume measured by 3D MRI reconstruction serves as a reliable prognostic factor for predicting term delivery in unicornuate uterus patients. The developed radiomics nomogram integrating radiomics signatures with clinical indicators enables individualized prediction of reproductive outcomes. Unicornuate uterus Reproductive outcomes Three-dimensional MRI Radiomic features Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Congenital uterine malformation arises from abnormal development, fusion or absorption of the Mullerian duct [ 1 ] , affecting approximately 5.5% to 6.7% of women in the general population. These malformations are commonly associated with infertility (2% to 8%) and miscarriage (5% to 30%) [ 2 ] . A unicornuate uterus, a type of mullerian duct anomaly, occurs when one side of the Mullerian duct develops normally, while the opposite side remains underdeveloped [ 3 ] . Clinically, a unicornuate uterus is often a incidental and asymptomatic finding, but it is frequently associated with complex pathology of the reproductive system, leading to infertility and recurrent miscarriages [ 4 ] . The American Society for Reproductive Medicine (ASRM) classifies the unicornuate uterus as class II, indicating a complete or partial failure in Mullerian duct development [ 5 ] , while the European Society of Human Reproduction and Embryology (ESHRE) and the European Society for Gynecological Endoscopy (ESGE) classify it as U4 [ 6 ] . Depending on the presence or absence of a functional rudimentary cavity, the hemi-uterus is further subdivided into two subtypes: U4a and U4b [ 7 ] . Women with hemi-uterus increased risks of obstetric complication, including extreme impact on fertility, obstetrical outcomes and gynecological health [ 8 ] . It is reported that women with unicornuate uterus is associated with higher incidence of infertility, endometriosis, recurrent pregnancy loss, preterm delivery, and increased perinatal morbidity and mortality [ 9 ] . It is often diagnosed incidentally in women during her workup for infertility. Recently, recent researches have revealed the uterine length or uterine cavity length could influence the implantation and clinical pregnancy rates in IVF. Egbase et al [ 10 ] showed that the cavity length was a critical factor in the highest implantation and clinical pregnancy rates. Similarly, Hawkins et al [ 11 ] detected that women with extreme uterine lengths ( 9.0 cm) were less likely to achieve live birth and those with urine lengths (< 6.0 cm) were also more likely to experience spontaneous abortion. Our previous study also indicated that women with longer uterine lengths were more likely to achieve full-term delivery. Both uterine length and uterine cavity length were found to be independent protective factors for favorable obstetric outcomes [ 12 ] . Although uterine volume is typically measured by radiologists using geometric formulas and standard techniques, measurement errors may still exist. In these years, Magnetic resonance imaging (MRI) seems to be a valuable imaging modality for the diagnosis of unicornate uterus [ 7 , 13 ] . With its multiplanar reformatting capabilities, detailed delineation of unicornate uterus anomalies can be detected. It is noninvasive, provides excellent soft-tissue characterization, and highlight clear anatomic details of the uterus and its zonal anatomy [ 13 ] . The 3D-T2WI sequence of pelvis enhances multiplanar reformatting, facilitating uterine characterization and 3D reconstruction, with serves as a critical reference for preoperative planning. For precise morphological assessment, radiologists manually outlined and segmented the morphology of the uterine cavity using the open-source software ITK-SNAP software (version 4.0., ITK-SNAP Home itksnap.org). In parallel, machine learning technology has been gradually applied in clinical medicine and emerged as a promising approach to develop new prediction models in recent years [ 14 ] . In this study, we employed open-source Python Package Pyradiomics to extract radiomic features, while the least absolute shrinkage and selection operator (LASSO) regression model was applied to the discovery dataset for signature construction. After feature screening, the final selected features were incorporated into eight machine learning models to develop a risk prediction model. Based on this preliminary study, our objective was to identify the relationship between the uterine sized measured by 3D MRI imaging and reproductive outcome in women with unicornuate uterus. In addition, we aimed to developed a radiomics nomogram incorporating radiomics signatures and clinical indicators, which could potentially aid in providing individualized predictions of fertility outcomes for women with unicornuate uterus. Methods and materials In this study, we collected our MRI clinic database from January 2005 to December 2022 to identify the patients diagnosed with unicornuate uterus with Three-dimensional MRI data at the Obstetrics and Gynecology Hospital of Fudan Unversity. The hemi-uterus classification was based on the guidelines set by the European Society of Human Reproduction and Embryology (ESHRE) and the European Society for Gynecological Endoscopy (ESGE). The inclusion criteria were women over 18 years old with an unprotected sexual intercourse for about 12 months, who were willing to comply with schedule for visits. Exclusion criteria were detailed in the Fig. 1 . 170 patients diagnosed with unicornuate uteri through pelvic 3D MRI were included in our study. This study was approved by the institutional Ethics Committee in the Obstetrics and Gynecology Hospital of Fudan University. This study collected maternal characteristics, including age, menstrual history, pregnancy and delivery history, gynecological surgeries, and pregnancy outcomes. Based on an initial analysis of the data, we categorized the 170 with unicornuate uterus women into four groups according to the their reproductive outcomes: Group 1 (primary infertility, n = 23), Group 2 (< 24 gestational weeks, n = 18), Group 3 (24–35 gestational weeks, n = 9) and Group 4 (≥35 gestational weeks n = 120), Preterm birth was defined as birth between 24 weeks and 35 weeks of gestation, as births before 23 weeks are unlikely to survive and those after 35 weeks are more likely to survive. MR acquisition and segmentation MR examinations were conducted using a 1.5-Tesla MR unit (Magnetom Avanto, Siemens) with a phased-array coil. During scanning, patients were in the supine position and breathed calmly. MRI evaluation was performed using 3D T2WI SPACE sequence with the following acquisition parameters: repetition time/echo time [TR/TE], 2000/126 msec; slice thickness, 1 mm; matrix size, 256 \(\:\times\:\) 256 or 320 \(\:\times\:320\) ; flip angle, 150 deg; bandwidth, 651 Hz/pixel. The uterine cavity was manually segmented using the open-source software ITK-SNAP software (version 4.0., ITK-SNAP Home itksnap.org). Cases that resulted in term delivery were assigned to the experimental group, while those involving primary infertility, adverse pregnancy outcomes, first trimester spontaneous abortion or preterm birth were categorized as the control group. Radiomics feature extraction and selection Radiomics features were extracted using the open-source Python package Pyradiomics ( https://pypi.org/project/pyradiomics/ ). All images were resampled to a resolution of 1 × 1 × 1 mm to generate isotropic voxels, followed by cubic spline interpolation. To mitigate variations in imaging across different MRI scanners, image normalization was performed, ensuring all gray level values were standardized within the range of 0–600. The extracted features categorized into several types: first-order features, gray level co-occurrence matrix (GLCM), gray-level dependence matrix (GLDM), gray-level size-zone matrix (GLSZM), gray-level run- length matrix (GLRLM), and neighboring gray tone difference matrix (NGTDM). In total, 1,834 radiomics features were extracted and subsequently standardized using Z scores. To identify the most relevant features, we conducted Mann-Whitney U test for statistical analysis and feature selection, retaining only those features with a p value < 0.05. For features exhibiting high repeatability, Spearman's rank correlation coefficient was computed to assess inter-feature correlations, Features with a correlation coefficient greater than 0.9 were subjected to feature reduction, retaining only one feature from each highly correlated pair. For signature construction, the least absolute shrinkage and selection operator (LASSO) regression model was applied to the discovery dataset. This approach yielded a radiomics score for each patient, derived from a linear combination of retained features, with each feature weighed by its corresponding model coefficient. The Python scikit-learn package was employed for LASSO regression modeling. Model Construction Following feature selection, we input the final features into eight machine learning models for risk model construction. The models evaluated included Logistic Regression (LR), Support Vector Machines (SVM), K-Nearest Neighbor (KNN), Random Forest, ExtraTrees, Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM) and Multi-Layer Perception (MLP). To ensure robust model performance, 5-fold cross-validation was employed. The diagnostic accuracy of each model was assessed using Receiver operating characteristic (ROC) curves. The model with best Area Under Curve (AUC) value was chosen as the optimal model for further analysis. Figure 3 showed Workflows for this study. Clinical signature and radiomics nomogram model The construction of the clinical signature followed a methodology similar to that of the radiomics signature. Initially, baseline statistics were used to select clinical features with a P value < 0.05. The clinical signature included variables such as age, history of rudimentary horn removal (with or without salpingectomy), severe dysmenorrhea, a history of recurrent spontaneous abortions over two instances, type and associated renal anomalies, endometriosis, and the voxel volume of the uterine cavity as measured by ITK-SNAP software. To ensure a fair comparison, fivefold cross validation was performed and a fixed test cohort was used. A radiomics nomogram was then developed by combining both the radiomics and clinical signatures. The diagnostic performance of the radiomics nomogram was evaluated using the test cohort, and Receiver Operating Characteristic (ROC) curves were fenerated to assess its accuracy. Calibration efficiency was examined using calibration curves, and the Hosmer‒Lemeshow test was also applied to evaluate the nomogram’s calibration. Finally, Mapping decision curve analysis (DCA) was employed to evaluate the clinical utility of the predictive models. Statistical analysis SPSS software (SPSS Inc., version 13.0) was used to perform all statistical analyses. Continuous variables are given as mean ± standard deviation or median ± numerical ranges, while categorical variables were expressed as numbers (percentages). One-way analysis of variance (ANOVA) was employed to compare of continuous variables across groups. For normally distributed continuous variables, an unpaired t-test was used, whereas the Mann–Whitney U test was applied for variables with non-normal distributions. The AUC values were calculated to assess the discriminatory power of various parameters in differentiating between groups. ROC analysis was performed to evaluated the performance of the radiomics models. with AUC used as a key metric. Additionally, the accuracy (ACC), sensitivity (SEN) and specificity (SPE) of the models were ccomputed to compare the effectiveness of the different methods. The radiomics score (rad-score) was derived from a linear combination of the selected features. A P value of less than 0.05 was considered statistically significant for all analyses. Results Baseline characteristics. A total of 170 patients diagnosed with unicornuate uterus were included in the study as shown in the flowchart in Fig. 1 . After applying the exclusion criteria, the maternal characteristics of 170 patients are summarized in Table 1. The average maternal age was 34.3±4.8 (SD) years old (range 23–49 years), Among these patients, 70% were diagnosed with U4bC0V0 while 30% were classified as U4aC0V0. Additionally, 14 patients had undergone rudimentary horn removal with or without salpingectomy (8.2%). Four patients (1.9%) had a history of endometriosis and 15 patients (8.8%) reported experiencing severe dysmenorrhea. Furthermore, 16 patients (9.4%) had a history of two or more spontaneous abortions. Among these patients, nearly 125 patients (73.5%) were natural conception while 23 patients (13.5%) underwent assisted reproductive technologies. Notably, 3 patients (1.8%) were diagnosed with associated renal agenesis based on MR image or ultrasound findings. Table 2 presents the reproductive outcomes of the study cohort. 120 patients delivered at ≥ 35 gestational weeks (70.6%), twenty-three patients delivered at < 24 gestational weeks (13.5%), while 9 patients delivered at 24–35 gestational weeks. In terms of fertility status, 23 patients were diagnosed with primarily infertility (13.5%). Among the 170 patients, 101 patients underwent chosen cesarean delivery (59.4%), while 48 patients achieved successful vaginal delivery (28.2%). Pelvic MRI data were obtained from the MR imaging clinic database. The distribution of uterine volume, surface area, major axis length, and minor axis length extracted using Pyradiomics is showed in Figure.2. The uterine volume of women who delivered at ≥ 35 gestational weeks (2603.3±1705.5mm 3 ) was significantly larger than that of women who delivered at < 24 gestational weeks (1760.5±758.3mm 3 ) and the primarily infertility group (1748.4±588.9mm 3 , P < 0.05). Similarly, the surface area of women delivering at ≥ 35 gestational weeks (1283.6±468.7mm 2 ) was significantly greater compared to the primarily infertility group (1046.7±207.9mm 2 , P 0.05). Feature reduction and model comparation The Rad-score was derived by selecting nonzero coefficients through a Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression model. Figure.4 illustrates the distribution of the Rad-score based on the selected features. label = 0.7234693330455401 - 0.030326 * gradient_glcm_Idn + 0.010146 * lbp_3D_k_glcm_ClusterProminence -0.030402 * lbp_3D_m1_gldm_SmallDependenceEmphasis -0.047697 * lbp_3D_m2_glcm_ClusterShade -0.069864 * log_sigma_2_0_mm_3D_firstorder_Maximum -0.022318 * log_sigma_2_0_mm_3D_glcm_DifferenceVariance + 0.038496 * log_sigma_3_0_mm_3D_glcm_ClusterShade -0.024944 * log_sigma_3_0_mm_3D_glcm_Idm -0.060778 * original_shape_MajorAxisLength -0.000853 * original_shape_Maximum2DDiameterSlice -0.019278 * squareroot_glcm_ClusterProminence -0.104702 * wavelet_HLH_glszm_ZoneVariance + 0.011928 * wavelet_LLH_firstorder_RobustMeanAbsoluteDeviation + 0.105183 * wavelet_LLH_glszm_ZoneVariance -0.004972 * wavelet_LLH_ngtdm_Coarseness The optimal predictive model was selected by evaluating the performance of radiomic features across a range of classifiers, including LR, SVM, KNN, Decision Tree, Random Forest, Extra Trees, XGBoost and LightGBM. LR achieved the best value of AUC on the training and test cohort reached 0.852 and 0.863 for predict ≥35 weeks’ gestation of unicornuate uterus respectively. Figure.5 presents the AUC values for each radiomics-based signature model on the test cohort. Nomogram In the training cohort, both the clinical and radiomics signature demonstrated strong model fitting. However, in the test cohort, the clinical signature showed signs of overfitting, while the radiomic signature remained robust and well-fitted. To combine both clinical and radiomics signature, a nomogram was developed using the Logistic Regression algorithm, which yielded the best performance. The radiomics nomogram, incorporating both clinical and radiomics features, effectively distinguished between patients with different fertility outcomes associated with unicornuate uterus. In the training cohort, the nomogram achieved an area under the curve (AUC) of 0.86 [95% CI, 0.79–0.92], and in the validation cohort, the AUC was 0.84 [95% CI, 0.69–0.98]. To compare the predictive accuracy of the clinical signature, radiomics signature and nomogram, the Delong test was conducted. The Decision Curve Analysis (DCA) revealed that the radiomics nomogram provided substantial clinical benefit compared to scenarios without any predictive model. Additionally, the nomogram outperformed both the clinical signature and the radiomics signature in term of prediction probability, highlighting its superior clinical utility. These findings suggest that the radiomics nomogram is more effective in predicting the likelihood of achieving a gestational age of ≥35 weeks in patients with unicornuate uterus. Figure.5 and Figure.6 display the AUC and DCA for both the training and test cohorts, as well as the graphical representation of the nomogram for clinical use. Discussion Genital tract anomalies in women often arise from genetic mutations or environmental factors that affect the development of the Mullerian or paramedian ducts. Among these anomalies, the unicornuate uterus is particularly notable for its impact on reproductive outcomes. Research has consistently shown that women with a unicornuate uterus face a notably higher risk of miscarriage compared to matched controls [ 15 , 16 ] , and they are also more likely to experience preterm delivery. Reichman et al. reported an incidence of unicornuate uterus occurring in approximately 1 in 4020 individuals in the general population. Their review highlighted that the rates of first-trimester abortion, second-trimester abortion, preterm delivery, and live birth were 24.3%, 9.7%, 20.1% and 51.5%, respectively [ 9 ] . Additionally, several studies have demonstrated that women with a unicornuate uterus have an elevated risk of early miscarriage, though specific miscarriage rates are often not directly reported [ 15 ] . Moreover, other studies have discovered various adverse obstetrics and neonatal outcomes associated with the condition [ 17 , 18 ] . In our study, pregnancy outcomes in women with a univornuate uterus included 9.4% spontaneous abortion (with more than two occurrence), 13.5% primary infertility, 10.6% first trimester abortion, 5.3% preterm delivery and 70.6% full-term delivery. It is hypothesized that primary infertility and early miscarriage in these women may be caused by uterine cavity dysplasia. The malformed uterus could hinder embryo implantation, growth and development. Furthermore, insufficient blood supply to the uterine myometrium and relative cervical insufficiency may increase the risk of preterm delivery [ 19 , 20 ] . Consistent with previous findings, the rate of cesarean delivery (59.4%) was significantly higher in women with a unicornuate uterus, which is largely attributed to malpresentation during delivery [ 21 ] . Wang et al. [ 22 ] found that the rates of in vitro fertilization and embryo transfer (IVF-ET) were significantly higher (12.6%) and the incidence of uterine rupture was (4.7%) among patients with a uncirnuate uterus. In our study, with the assistance of IVF, 23 patients (13.5%) with a unicornuate uterus successfully achieved pregnancy. It is generally believed that the unicornuate uterus is slightly smaller, with asymmetric polarity, which may influence pregnancy outcome [ 23 ] . Therefore, women with a uincornuate uterus should limit the number of embryos typically opting for a single embryo transfer, which aligns with our findings. Additionally, 125 patients with a uincornuate uterus were able to conceive naturally. The rudimentary horn of an unicornuate uterus arises from unilateral failure of Mullerian duct migration and is often associated with anomalies in the urinary system, particularly renal agenesis [ 24 ] . According to the ESHRE/ESGE classification, the hemi-uterus is categorized into two sub-classes: U4a (with a functional rudimentary cavity) and U4b (without a functional rudimentary cavity). it is reported that U4b accounts for 85% of cases [ 7 ] , which is consistent with our findings. In our study, 51 patients (30%) had functional rudimentary cavity. Among these, 14 patients (8.2%) underwent rudimentary horn excision. We also observed that 3 patients (1.4%) had associated renal anomalies, which is lower than the rate reported in the literature (16–38%) [ 25 ] . Sang et al. found that patients with a normal uterine cavity (7-8cm) or longer (> 8cm) had higher clinical pregnancy rates in the IVF-ICSI treatments [ 26 ] . In our previous study, we also demonstrated that women with a uincornuate uterus and longer uterine lengths (≥ 4.5cm) were more likely to achieve a full-term delivery [ 12 ] . Similarly, Sardo et al. [ 27 ] discovered that women with dysmorphic uterus shapes, who underwent the novel Hysteroscopic Outpatient Metroplasty to Expand Dysmorphic Uteri (HOME-EU) technique, achieved a clinical pregnancy rate of 57% and a term delivery rate of 65%. These improvements were linked to the expansion of uterine volume and enhancement of uterine morphology. In the present study, we analyzed data from 170 women with a unicornuate uterus uusing 3D magnetic resonance imaging (MRI) reconstruction. Our results indicate that the women with larger uterine volumes had higher chances of to achieving full-term deliveries. Additionally, the surface area of the uterus was significantly greater in women who dilivered ≥35 gestational weeks groups compared to other three groups. These findings suggest that both uterine volume and surface area serve as independent protective factors for better reproductive outcomes. We developed and validated a predictive model for reproductive outcomes in women with a unicornuate uterus, incorporating clinical factors such as uterine volume, surface area, sphericity and diameter lines of uterine body, along with radiomics signatures extracted from multiple phases of 3D MRI scans. Previous research on 3D uterine reconstruction often relied on images with a slice thickness of 2-4mm, resulting in relatively lower resolution and accuracy [ 28 , 29 ] . In contrast, our study utilized a 3D T2-weighted imaging (T2WI) SPACE sequence with a slice thickness of 1mm and a voxel size of 1⋅1⋅1mm, enabling high-resolution 3D reconstructions with greater precision. The most lesion of uterus signals is not very obvious between the surrounding tissues [ 30 ] , so the reconstruction of the uterus can only be implemented in manual sketch, which may take approximately three hours in one case. In our study, we implemented machine learning algorithms to automatically segment the uterine volume, reducing this process to just a few seconds. This innovation significantly increased work efficiency and paved the way for large-scale data analysis in future research. For constructing our risk prediction model, we employed eight different machine learning models. When applied to predict the reproductive outcomes of women with a unicornuate uterus, the radiomics nomogram incorporating radiomics signatures and clinical feature showed good discrimination between patients with fertility outcome of unicornuate uterus with an area under the curve (AUC) of 0.86 in the training cohort and 0.84 in the validation cohort, indicating that this model had an absolute dominant advantage in predicting a reproductive outcome of unicornuate uterus. This study has several limitations. First, it is a single-center retrospective analysis with a small-sample size. To enhance the robustness of our findings, future studies should incorporate larger, multicenter datasets for validation. Second, the manual delineation of the uterine cavity is time-consuming and technically demanding. Currently, no available software can automatically perform this segmentation. The development and application of more advanced AI (artificial intelligence) based segmentation techniques would greatly improve efficiency, reduce observer variability and enhance consistency across studies. Conclusion This study introduces a noninvasive 3D-reconstruction imaging technique, demonstrating that the uterine volume can serve as a reliable prognostic factor for predicting the gestational week of delivery in women with a unicornuate uterus. Furthermore, we developed a radiomics nomogram that combines both radiomic signatures and clinical factors. This cool shows promise for facilitating the individualized prediction of fertility outcome in patients with a unicornuate uterus, offering a personalized approach to their reproductive management. Declarations Funding This work was supported by grants from Project of the National Key R&D Program of China (Grant No.2021YFC2701402) Competing Interests All authors declared no conflict of interest. Author Contributions Conception and study design: Jingxin Ding, Keqin Hua; Execution: Xinyu Pan, Guofu Zhang; Analysis: Hui Luo, Qing Zhou; Data interpretation: Hui Luo, Qing Zhou, Xinyu Pan, Guofu Zhang; Manuscript drafting and revising: Hui Luo, Qing Zhou, Jingxin Ding; Project administration: Keqin Hua; All authors have read and agreed to the published version of the manuscript. All authors contributed substantially to conception and design, or acquisition of data, or analysis and interpretation of data, and approved the final version of submission. Ethics approval All procedure performed in studies involving human participants were in accordance with the ethical standard of Obstetrics and Gynecology Hospital Affiliated to Fudan University institutional review board（No.2022-081）. Consent to participate Informed consent was obtained from all individual participants included in the study. Consent to publish The authors affirm that human research participants provided informed consent for publication of the all images. References Chan YY, Jayaprakasan K, Zamora J, et al. The prevalence of congenital uterine anomalies in unselected and high-risk populations: a systematic review[J]. Hum Reprod Update , 2011,17(6):761–771. Saravelos SH, Cocksedge KA, Li TC. Prevalence and diagnosis of congenital uterine anomalies in women with reproductive failure: a critical appraisal[J]. Hum Reprod Update , 2008,14(5):415–429. Cerekja A, Dillon KC, Racanska E, et al. Unicornuate uteri associated with contralateral renal agenesis and ovarian anomalies[J]. J Turk Ger Gynecol Assoc , 2011,12(3):189–191. Akhtar MA, Saravelos SH, Li TC, et al. Reproductive Implications and Management of Congenital Uterine Anomalies: Scientific Impact Paper No. 62 November 2019[J]. BJOG , 2020,127(5):e1-e13. Pfeifer SM, Attaran M, Goldstein J, et al. ASRM mullerian anomalies classification 2021[J]. Fertil Steril , 2021,116(5):1238–1252. Grimbizis GF, Gordts S, Di Spiezio Sardo A, et al. The ESHRE/ESGE consensus on the classification of female genital tract congenital anomalies[J]. Hum Reprod , 2013,28(8):2032–2044. Grimbizis GF, Di Spiezio Sardo A, Saravelos SH, et al. The Thessaloniki ESHRE/ESGE consensus on diagnosis of female genital anomalies[J]. Gynecol Surg , 2016,13:1–16. Zyla MM, Wilczynski J, Nowakowska-Glab A, et al. Pregnancy and Delivery in Women with Uterine Malformations[J]. Adv Clin Exp Med , 2015,24(5):873–879. Reichman D, Laufer MR, Robinson BK. Pregnancy outcomes in unicornuate uteri: a review[J]. Fertil Steril , 2009,91(5):1886–1894. Egbase PE, Al-Sharhan M, Grudzinskas JG. Influence of position and length of uterus on implantation and clinical pregnancy rates in IVF and embryo transfer treatment cycles[J]. Hum Reprod , 2000,15(9):1943–1946. Grimbizis GF, Di Spiezio Sardo A, Saravelos SH, et al. The Thessaloniki ESHRE/ESGE consensus on diagnosis of female genital anomalies[J]. Hum Reprod , 2016,31(1):2–7. Li XQ, Qian HJ, Zhang XY, et al. Analysis of the reproductive outcomes and the size of the unicornuate uterus measured by magnetic resonance imaging and their relationship[J]. Arch Gynecol Obstet , 2019,299(5):1321–1330. Behr SC, Courtier JL, Qayyum A. Imaging of mullerian duct anomalies[J]. Radiographics , 2012,32(6):E233-250. Taylor RA, Pare JR, Venkatesh AK, et al. Prediction of In-hospital Mortality in Emergency Department Patients With Sepsis: A Local Big Data-Driven, Machine Learning Approach[J]. Acad Emerg Med , 2016,23(3):269–278. Chan YY, Jayaprakasan K, Tan A, et al. Reproductive outcomes in women with congenital uterine anomalies: a systematic review[J]. Ultrasound Obstet Gynecol , 2011,38(4):371–382. Chen Y, Nisenblat V, Yang P, et al. Reproductive outcomes in women with unicornuate uterus undergoing in vitro fertilization: a nested case-control retrospective study[J]. Reprod Biol Endocrinol , 2018,16(1):64. Zhang Y, Zhao YY, Qiao J. Obstetric outcome of women with uterine anomalies in China[J]. Chin Med J (Engl) , 2010,123(4):418–422. Hiersch L, Yeoshoua E, Miremberg H, et al. The association between Mullerian anomalies and short-term pregnancy outcome[J]. J Matern Fetal Neonatal Med , 2016,29(16):2573–2578. Takami M, Aoki S, Kurasawa K, et al. A classification of congenital uterine anomalies predicting pregnancy outcomes[J]. Acta Obstet Gynecol Scand , 2014,93(7):691–697. Jacquinet A, Millar D, Lehman A. Etiologies of uterine malformations[J]. Am J Med Genet A , 2016,170(8):2141–2172. Naeh A, Sigal E, Barda S, et al. The association between congenital uterine anomalies and perinatal outcomes - does type of defect matters?[J]. J Matern Fetal Neonatal Med , 2022,35(25):7406–7411. Wang S, Wang K, Hu Q, et al. Perinatal outcomes of women with Mullerian anomalies[J]. Arch Gynecol Obstet , 2023,307(4):1209–1216. Li X, Ouyang Y, Yi Y, et al. Pregnancy outcomes of women with a congenital unicornuate uterus after IVF-embryo transfer[J]. Reprod Biomed Online , 2017,35(5):583–591. Nahum GG. Uterine anomalies. How common are they, and what is their distribution among subtypes?[J]. J Reprod Med , 1998,43(10):877–887. Pados G, Tsolakidis D, Athanatos D, et al. Reproductive and obstetric outcome after laparoscopic excision of functional, non-communicating broadly attached rudimentary horn: a case series[J]. Eur J Obstet Gynecol Reprod Biol , 2014,182:33–37. Chun SS, Chung MJ, Chong GO, et al. Relationship between the length of the uterine cavity and clinical pregnancy rates after in vitro fertilization or intracytoplasmic sperm injection[J]. Fertil Steril , 2010,93(2):663–665. Di Spiezio Sardo A, Florio P, Nazzaro G, et al. Hysteroscopic outpatient metroplasty to expand dysmorphic uteri (HOME-DU technique): a pilot study[J]. Reprod Biomed Online , 2015,30(2):166–174. Pan HX, Liu P, Duan H, et al. Using 3D MRI can potentially enhance the ability of trained surgeons to more precisely diagnose Mullerian duct anomalies compared to MR alone[J]. Eur J Obstet Gynecol Reprod Biol , 2018,228:313–318. Durnea CM, Siddiqi S, Nazarian D, et al. 3D-Volume Rendering of the Pelvis with Emphasis on Paraurethral Structures Based on MRI Scans and Comparisons between 3D Slicer and OsiriX(R)[J]. J Med Syst , 2021,45(3):27. Lee SR, Kim YJ, Kim KG. A Fast 3-Dimensional Magnetic Resonance Imaging Reconstruction for Surgical Planning of Uterine Myomectomy[J]. J Korean Med Sci , 2018,33(2):e12. Tables Table.1 Demographic maternal characteristics of 170 women with unicornuate uteri. Parameters Mean±SD （range or percentage） Maternal age, (year) 34.26±4.75 Gravidity 2 Parity 1 Subtypes of hemi-uterus U4aC0V0 51(30%) U4bC0V0 119(70%) Removal of rudimentary horn with or without salpingectomy 14(8.2%) Endometriosis 4(1.9%) Severe dysmenorrhea 15(8.8%) Spontaneous abortion³2 16(9.4%) Mode of pregnancy Natural conception 125(73.5%) IVF 23(13.5%) Associated renal anomalies, n (%) 3(1.4%) Table.2 Reproductive outcomes of 170 women with unicornuate uterus Pregnancy outcome Mean±SD (range or percentage) Primary infertility 23(13.5%) <24 gestational weeks 18(10.6%) 24-35gestational weeks 9(5.3%) ³35gestational weeks 120(70.6%) Vaginal delivery 48(28.2%) Cesarean delivery 101(59.4%) Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 11 May, 2026 Reviews received at journal 22 Apr, 2026 Reviewers agreed at journal 20 Apr, 2026 Reviewers agreed at journal 18 Apr, 2026 Reviews received at journal 16 Apr, 2026 Reviewers agreed at journal 13 Apr, 2026 Reviewers agreed at journal 13 Apr, 2026 Reviewers invited by journal 23 Mar, 2026 Editor assigned by journal 21 Mar, 2026 Submission checks completed at journal 19 Mar, 2026 First submitted to journal 18 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9157364","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":610745115,"identity":"3f854c33-01f4-4011-92c1-46fa528feeb3","order_by":0,"name":"Hui Luo","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Hui","middleName":"","lastName":"Luo","suffix":""},{"id":610745116,"identity":"8f8f483a-6674-4a21-9500-71fdfb01cb5f","order_by":1,"name":"Qing Zhou","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Qing","middleName":"","lastName":"Zhou","suffix":""},{"id":610745117,"identity":"956148a5-1e8c-4d90-901c-a2ded5ab8842","order_by":2,"name":"Xinyu Pan","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Xinyu","middleName":"","lastName":"Pan","suffix":""},{"id":610745118,"identity":"1bdb6680-4e6d-43aa-b258-db1969b765d7","order_by":3,"name":"Guofu Zhang","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Guofu","middleName":"","lastName":"Zhang","suffix":""},{"id":610745119,"identity":"fcb499f4-22fe-4059-952a-190ecdea9f53","order_by":4,"name":"Keqin Hua","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAvklEQVRIiWNgGAWjYBACAwYGZiAlwcDPzHzwAWlaJNvZkg1I0QJknOcxEyBKi7lE8mODD2UWecaHGcwYGGpsoglqsZyRZpw445xEsdlhhrQHDMfSchsIOuxGgvFh3jaJxG2HGY4bMDYcJkZL+ufDf4FaNjcztkkQqSXHOBmoOHEDMzMbkVrOvCk27DknkTjjMBuzQQJRfjmevlniR1ldYn//+Y8PPtTYENYCAWxQOoE45chaRsEoGAWjYBRgAwD0wj4Q0USDIwAAAABJRU5ErkJggg==","orcid":"","institution":"Fudan University","correspondingAuthor":true,"prefix":"","firstName":"Keqin","middleName":"","lastName":"Hua","suffix":""},{"id":610745120,"identity":"fe0add0b-48e4-4be9-aeec-20e1366f5a24","order_by":5,"name":"Jingxin Ding","email":"","orcid":"","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Jingxin","middleName":"","lastName":"Ding","suffix":""}],"badges":[],"createdAt":"2026-03-18 09:25:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9157364/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9157364/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105371454,"identity":"6d9d1320-452f-4b80-9e7f-d0ba6b5e0572","added_by":"auto","created_at":"2026-03-25 09:28:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":68338,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of patient selection.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-9157364/v1/183ea21528f5e4be92454738.png"},{"id":105371448,"identity":"ed8946ac-e2bf-4c51-a6cc-71a38df18e63","added_by":"auto","created_at":"2026-03-25 09:28:11","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":99017,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of uterine size measured by 3-dimensionally reconstructed uteri between four groups (a) volume of uterine, (b) surface area of uterine, (c) major axis length of uterine, (d) minor axis length of uterine. *\u003cem\u003eP\u003c/em\u003e\u0026lt;0.05, **\u003cem\u003eP\u003c/em\u003e\u0026lt;0.01, ***\u003cem\u003eP\u003c/em\u003e\u0026lt;0.001.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-9157364/v1/7a1db61c6e0350ded7253c76.png"},{"id":105371395,"identity":"a742f5d5-9a20-4973-afe2-918180754a23","added_by":"auto","created_at":"2026-03-25 09:28:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":348703,"visible":true,"origin":"","legend":"\u003cp\u003eWorkflows for this study. After imaging preprocessing and registration, uterine cavity segmentation was performed. Radiomics features were extracted by Pyradiomics. Least absolute shrinkage and selection operator (LASSO) regression analysis was used to select radiomic features.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-9157364/v1/86b3775e92fc76c48ed779a7.png"},{"id":105371418,"identity":"445ec3f6-bb61-4359-995a-f9f86fb9bfb6","added_by":"auto","created_at":"2026-03-25 09:28:10","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":125926,"visible":true,"origin":"","legend":"\u003cp\u003eA. Selected features weight coefficients. B. Rad signature model's auc on test cohort.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-9157364/v1/c3c03e04581d55b465ddbf05.png"},{"id":105371473,"identity":"3913ce17-bd04-4d40-8d7f-93af803e9ec3","added_by":"auto","created_at":"2026-03-25 09:28:14","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":208635,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristic (ROC) curves of radiomic Logistic Regression (LR) model in the training cohort and validation cohort (a). Analysis of decision curves for each model in the training and validation cohorts (b).\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-9157364/v1/7e48e4d64ca24f1d9b4cee4e.png"},{"id":105371371,"identity":"ce98f90b-90d7-48cf-9242-14033cea3d43","added_by":"auto","created_at":"2026-03-25 09:27:59","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":46202,"visible":true,"origin":"","legend":"\u003cp\u003eA radiomics nomogram was developed in the training cohort, incorporating the clinical signature and radiomics signature, and fertility outcomes.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-9157364/v1/5b48c79052257d44ea71076c.png"},{"id":105371591,"identity":"9392f2d2-5a14-4669-bda1-cdba6968fad7","added_by":"auto","created_at":"2026-03-25 09:28:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1413089,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9157364/v1/888081d8-6e4f-41b7-9d26-09ccc4cc27cb.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003ePrediction of reproductive outcomes in unicornuate uterus women based on Three-dimensional MRI radiomic features\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eCongenital uterine malformation arises from abnormal development, fusion or absorption of the Mullerian duct\u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e, affecting approximately 5.5% to 6.7% of women in the general population. These malformations are commonly associated with infertility (2% to 8%) and miscarriage (5% to 30%)\u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. A unicornuate uterus, a type of mullerian duct anomaly, occurs when one side of the Mullerian duct develops normally, while the opposite side remains underdeveloped\u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e. Clinically, a unicornuate uterus is often a incidental and asymptomatic finding, but it is frequently associated with complex pathology of the reproductive system, leading to infertility and recurrent miscarriages\u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e. The American Society for Reproductive Medicine (ASRM) classifies the unicornuate uterus as class II, indicating a complete or partial failure in Mullerian duct development\u003csup\u003e[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e, while the European Society of Human Reproduction and Embryology (ESHRE) and the European Society for Gynecological Endoscopy (ESGE) classify it as U4\u003csup\u003e[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/sup\u003e. Depending on the presence or absence of a functional rudimentary cavity, the hemi-uterus is further subdivided into two subtypes: U4a and U4b\u003csup\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWomen with hemi-uterus increased risks of obstetric complication, including extreme impact on fertility, obstetrical outcomes and gynecological health\u003csup\u003e[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e. It is reported that women with unicornuate uterus is associated with higher incidence of infertility, endometriosis, recurrent pregnancy loss, preterm delivery, and increased perinatal morbidity and mortality\u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. It is often diagnosed incidentally in women during her workup for infertility. Recently, recent researches have revealed the uterine length or uterine cavity length could influence the implantation and clinical pregnancy rates in IVF. Egbase et al\u003csup\u003e[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e showed that the cavity length was a critical factor in the highest implantation and clinical pregnancy rates. Similarly, Hawkins et al\u003csup\u003e[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e detected that women with extreme uterine lengths (\u0026lt;\u0026thinsp;7.0 or \u0026gt;\u0026thinsp;9.0 cm) were less likely to achieve live birth and those with urine lengths (\u0026lt;\u0026thinsp;6.0 cm) were also more likely to experience spontaneous abortion. Our previous study also indicated that women with longer uterine lengths were more likely to achieve full-term delivery. Both uterine length and uterine cavity length were found to be independent protective factors for favorable obstetric outcomes\u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e. Although uterine volume is typically measured by radiologists using geometric formulas and standard techniques, measurement errors may still exist.\u003c/p\u003e \u003cp\u003eIn these years, Magnetic resonance imaging (MRI) seems to be a valuable imaging modality for the diagnosis of unicornate uterus\u003csup\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. With its multiplanar reformatting capabilities, detailed delineation of unicornate uterus anomalies can be detected. It is noninvasive, provides excellent soft-tissue characterization, and highlight clear anatomic details of the uterus and its zonal anatomy\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. The 3D-T2WI sequence of pelvis enhances multiplanar reformatting, facilitating uterine characterization and 3D reconstruction, with serves as a critical reference for preoperative planning. For precise morphological assessment, radiologists manually outlined and segmented the morphology of the uterine cavity using the open-source software ITK-SNAP software (version 4.0., ITK-SNAP Home itksnap.org). In parallel, machine learning technology has been gradually applied in clinical medicine and emerged as a promising approach to develop new prediction models in recent years\u003csup\u003e[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/sup\u003e. In this study, we employed open-source Python Package Pyradiomics to extract radiomic features, while the least absolute shrinkage and selection operator (LASSO) regression model was applied to the discovery dataset for signature construction. After feature screening, the final selected features were incorporated into eight machine learning models to develop a risk prediction model.\u003c/p\u003e \u003cp\u003eBased on this preliminary study, our objective was to identify the relationship between the uterine sized measured by 3D MRI imaging and reproductive outcome in women with unicornuate uterus. In addition, we aimed to developed a radiomics nomogram incorporating radiomics signatures and clinical indicators, which could potentially aid in providing individualized predictions of fertility outcomes for women with unicornuate uterus.\u003c/p\u003e"},{"header":"Methods and materials","content":"\u003cp\u003eIn this study, we collected our MRI clinic database from January 2005 to December 2022 to identify the patients diagnosed with unicornuate uterus with Three-dimensional MRI data at the Obstetrics and Gynecology Hospital of Fudan Unversity. The hemi-uterus classification was based on the guidelines set by the European Society of Human Reproduction and Embryology (ESHRE) and the European Society for Gynecological Endoscopy (ESGE). The inclusion criteria were women over 18 years old with an unprotected sexual intercourse for about 12 months, who were willing to comply with schedule for visits. Exclusion criteria were detailed in the Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. 170 patients diagnosed with unicornuate uteri through pelvic 3D MRI were included in our study. This study was approved by the institutional Ethics Committee in the Obstetrics and Gynecology Hospital of Fudan University.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThis study collected maternal characteristics, including age, menstrual history, pregnancy and delivery history, gynecological surgeries, and pregnancy outcomes. Based on an initial analysis of the data, we categorized the 170 with unicornuate uterus women into four groups according to the their reproductive outcomes: Group 1 (primary infertility, n\u0026thinsp;=\u0026thinsp;23), Group 2 (\u0026lt;\u0026thinsp;24 gestational weeks, n\u0026thinsp;=\u0026thinsp;18), Group 3 (24\u0026ndash;35 gestational weeks, n\u0026thinsp;=\u0026thinsp;9) and Group 4 (\u0026ge;35 gestational weeks n\u0026thinsp;=\u0026thinsp;120), Preterm birth was defined as birth between 24 weeks and 35 weeks of gestation, as births before 23 weeks are unlikely to survive and those after 35 weeks are more likely to survive.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eMR acquisition and segmentation\u003c/h2\u003e \u003cp\u003eMR examinations were conducted using a 1.5-Tesla MR unit (Magnetom Avanto, Siemens) with a phased-array coil. During scanning, patients were in the supine position and breathed calmly. MRI evaluation was performed using 3D T2WI SPACE sequence with the following acquisition parameters: repetition time/echo time [TR/TE], 2000/126 msec; slice thickness, 1 mm; matrix size, 256\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\times\\:\\)\u003c/span\u003e\u003c/span\u003e256 or 320\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\times\\:320\\)\u003c/span\u003e\u003c/span\u003e; flip angle, 150 deg; bandwidth, 651 Hz/pixel. The uterine cavity was manually segmented using the open-source software ITK-SNAP software (version 4.0., ITK-SNAP Home itksnap.org). Cases that resulted in term delivery were assigned to the experimental group, while those involving primary infertility, adverse pregnancy outcomes, first trimester spontaneous abortion or preterm birth were categorized as the control group.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eRadiomics feature extraction and selection\u003c/h3\u003e\n\u003cp\u003eRadiomics features were extracted using the open-source Python package Pyradiomics (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pypi.org/project/pyradiomics/\u003c/span\u003e\u003cspan address=\"https://pypi.org/project/pyradiomics/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). All images were resampled to a resolution of 1 \u0026times; 1 \u0026times; 1 mm to generate isotropic voxels, followed by cubic spline interpolation. To mitigate variations in imaging across different MRI scanners, image normalization was performed, ensuring all gray level values were standardized within the range of 0\u0026ndash;600. The extracted features categorized into several types: first-order features, gray level co-occurrence matrix (GLCM), gray-level dependence matrix (GLDM), gray-level size-zone matrix (GLSZM), gray-level run- length matrix (GLRLM), and neighboring gray tone difference matrix (NGTDM). In total, 1,834 radiomics features were extracted and subsequently standardized using Z scores. To identify the most relevant features, we conducted Mann-Whitney U test for statistical analysis and feature selection, retaining only those features with a p value\u0026thinsp;\u0026lt;\u0026thinsp;0.05. For features exhibiting high repeatability, Spearman's rank correlation coefficient was computed to assess inter-feature correlations, Features with a correlation coefficient greater than 0.9 were subjected to feature reduction, retaining only one feature from each highly correlated pair. For signature construction, the least absolute shrinkage and selection operator (LASSO) regression model was applied to the discovery dataset. This approach yielded a radiomics score for each patient, derived from a linear combination of retained features, with each feature weighed by its corresponding model coefficient. The Python scikit-learn package was employed for LASSO regression modeling.\u003c/p\u003e\n\u003ch3\u003eModel Construction\u003c/h3\u003e\n\u003cp\u003eFollowing feature selection, we input the final features into eight machine learning models for risk model construction. The models evaluated included Logistic Regression (LR), Support Vector Machines (SVM), K-Nearest Neighbor (KNN), Random Forest, ExtraTrees, Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM) and Multi-Layer Perception (MLP). To ensure robust model performance, 5-fold cross-validation was employed. The diagnostic accuracy of each model was assessed using Receiver operating characteristic (ROC) curves. The model with best Area Under Curve (AUC) value was chosen as the optimal model for further analysis. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e showed Workflows for this study.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eClinical signature and radiomics nomogram model\u003c/h3\u003e\n\u003cp\u003eThe construction of the clinical signature followed a methodology similar to that of the radiomics signature. Initially, baseline statistics were used to select clinical features with a \u003cem\u003eP\u003c/em\u003e value\u0026thinsp;\u0026lt;\u0026thinsp;0.05. The clinical signature included variables such as age, history of rudimentary horn removal (with or without salpingectomy), severe dysmenorrhea, a history of recurrent spontaneous abortions over two instances, type and associated renal anomalies, endometriosis, and the voxel volume of the uterine cavity as measured by ITK-SNAP software. To ensure a fair comparison, fivefold cross validation was performed and a fixed test cohort was used. A radiomics nomogram was then developed by combining both the radiomics and clinical signatures. The diagnostic performance of the radiomics nomogram was evaluated using the test cohort, and Receiver Operating Characteristic (ROC) curves were fenerated to assess its accuracy. Calibration efficiency was examined using calibration curves, and the Hosmer‒Lemeshow test was also applied to evaluate the nomogram\u0026rsquo;s calibration. Finally, Mapping decision curve analysis (DCA) was employed to evaluate the clinical utility of the predictive models.\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eSPSS software (SPSS Inc., version 13.0) was used to perform all statistical analyses. Continuous variables are given as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation or median \u0026plusmn; numerical ranges, while categorical variables were expressed as numbers (percentages). One-way analysis of variance (ANOVA) was employed to compare of continuous variables across groups. For normally distributed continuous variables, an unpaired t-test was used, whereas the Mann\u0026ndash;Whitney U test was applied for variables with non-normal distributions. The AUC values were calculated to assess the discriminatory power of various parameters in differentiating between groups. ROC analysis was performed to evaluated the performance of the radiomics models. with AUC used as a key metric. Additionally, the accuracy (ACC), sensitivity (SEN) and specificity (SPE) of the models were ccomputed to compare the effectiveness of the different methods. The radiomics score (rad-score) was derived from a linear combination of the selected features. A \u003cem\u003eP\u003c/em\u003e value of less than 0.05 was considered statistically significant for all analyses.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eBaseline characteristics.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 170 patients diagnosed with unicornuate uterus were included in the study as shown in the flowchart in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. After applying the exclusion criteria, the maternal characteristics of 170 patients are summarized in Table\u0026nbsp;1. The average maternal age was 34.3\u0026plusmn;4.8 (SD) years old (range 23\u0026ndash;49 years), Among these patients, 70% were diagnosed with U4bC0V0 while 30% were classified as U4aC0V0. Additionally, 14 patients had undergone rudimentary horn removal with or without salpingectomy (8.2%). Four patients (1.9%) had a history of endometriosis and 15 patients (8.8%) reported experiencing severe dysmenorrhea. Furthermore, 16 patients (9.4%) had a history of two or more spontaneous abortions. Among these patients, nearly 125 patients (73.5%) were natural conception while 23 patients (13.5%) underwent assisted reproductive technologies. Notably, 3 patients (1.8%) were diagnosed with associated renal agenesis based on MR image or ultrasound findings.\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;2 presents the reproductive outcomes of the study cohort. 120 patients delivered at \u0026ge; 35 gestational weeks (70.6%), twenty-three patients delivered at \u0026lt; 24 gestational weeks (13.5%), while 9 patients delivered at 24\u0026ndash;35 gestational weeks. In terms of fertility status, 23 patients were diagnosed with primarily infertility (13.5%). Among the 170 patients, 101 patients underwent chosen cesarean delivery (59.4%), while 48 patients achieved successful vaginal delivery (28.2%).\u003c/p\u003e\n\u003cp\u003ePelvic MRI data were obtained from the MR imaging clinic database. The distribution of uterine volume, surface area, major axis length, and minor axis length extracted using Pyradiomics is showed in Figure.2. The uterine volume of women who delivered at \u0026ge; 35 gestational weeks (2603.3\u0026plusmn;1705.5mm\u003csup\u003e3\u003c/sup\u003e) was significantly larger than that of women who delivered at \u0026lt; 24 gestational weeks (1760.5\u0026plusmn;758.3mm\u003csup\u003e3\u003c/sup\u003e) and the primarily infertility group (1748.4\u0026plusmn;588.9mm\u003csup\u003e3\u003c/sup\u003e, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Similarly, the surface area of women delivering at \u0026ge; 35 gestational weeks (1283.6\u0026plusmn;468.7mm\u003csup\u003e2\u003c/sup\u003e) was significantly greater compared to the primarily infertility group (1046.7\u0026plusmn;207.9mm\u003csup\u003e2\u003c/sup\u003e, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). However, no significant differences were observed in the major axis length or minor axis length between four groups (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e\n\u003ch3\u003eFeature reduction and model comparation\u003c/h3\u003e\n\u003cp\u003eThe Rad-score was derived by selecting nonzero coefficients through a Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression model. Figure.4 illustrates the distribution of the Rad-score based on the selected features.\u003c/p\u003e\n\u003cp\u003elabel\u0026thinsp;=\u0026thinsp;0.7234693330455401 -\u003c/p\u003e\n\u003cp\u003e0.030326 * gradient_glcm_Idn\u003c/p\u003e\n\u003cp\u003e+\u0026thinsp;0.010146 * lbp_3D_k_glcm_ClusterProminence\u003c/p\u003e\n\u003cp\u003e-0.030402 * lbp_3D_m1_gldm_SmallDependenceEmphasis\u003c/p\u003e\n\u003cp\u003e-0.047697 * lbp_3D_m2_glcm_ClusterShade\u003c/p\u003e\n\u003cp\u003e-0.069864 * log_sigma_2_0_mm_3D_firstorder_Maximum\u003c/p\u003e\n\u003cp\u003e-0.022318 * log_sigma_2_0_mm_3D_glcm_DifferenceVariance\u003c/p\u003e\n\u003cp\u003e+\u0026thinsp;0.038496 * log_sigma_3_0_mm_3D_glcm_ClusterShade\u003c/p\u003e\n\u003cp\u003e-0.024944 * log_sigma_3_0_mm_3D_glcm_Idm\u003c/p\u003e\n\u003cp\u003e-0.060778 * original_shape_MajorAxisLength\u003c/p\u003e\n\u003cp\u003e-0.000853 * original_shape_Maximum2DDiameterSlice\u003c/p\u003e\n\u003cp\u003e-0.019278 * squareroot_glcm_ClusterProminence\u003c/p\u003e\n\u003cp\u003e-0.104702 * wavelet_HLH_glszm_ZoneVariance\u003c/p\u003e\n\u003cp\u003e+\u0026thinsp;0.011928 * wavelet_LLH_firstorder_RobustMeanAbsoluteDeviation\u003c/p\u003e\n\u003cp\u003e+\u0026thinsp;0.105183 * wavelet_LLH_glszm_ZoneVariance\u003c/p\u003e\n\u003cp\u003e-0.004972 * wavelet_LLH_ngtdm_Coarseness\u003c/p\u003e\n\u003cp\u003eThe optimal predictive model was selected by evaluating the performance of radiomic features across a range of classifiers, including LR, SVM, KNN, Decision Tree, Random Forest, Extra Trees, XGBoost and LightGBM. LR achieved the best value of AUC on the training and test cohort reached 0.852 and 0.863 for predict \u0026ge;35 weeks\u0026rsquo; gestation of unicornuate uterus respectively. Figure.5 presents the AUC values for each radiomics-based signature model on the test cohort.\u003c/p\u003e\n\u003ch3\u003eNomogram\u003c/h3\u003e\n\u003cp\u003eIn the training cohort, both the clinical and radiomics signature demonstrated strong model fitting. However, in the test cohort, the clinical signature showed signs of overfitting, while the radiomic signature remained robust and well-fitted. To combine both clinical and radiomics signature, a nomogram was developed using the Logistic Regression algorithm, which yielded the best performance.\u003c/p\u003e\n\u003cp\u003eThe radiomics nomogram, incorporating both clinical and radiomics features, effectively distinguished between patients with different fertility outcomes associated with unicornuate uterus. In the training cohort, the nomogram achieved an area under the curve (AUC) of 0.86 [95% CI, 0.79\u0026ndash;0.92], and in the validation cohort, the AUC was 0.84 [95% CI, 0.69\u0026ndash;0.98]. To compare the predictive accuracy of the clinical signature, radiomics signature and nomogram, the Delong test was conducted.\u003c/p\u003e\n\u003cp\u003eThe Decision Curve Analysis (DCA) revealed that the radiomics nomogram provided substantial clinical benefit compared to scenarios without any predictive model. Additionally, the nomogram outperformed both the clinical signature and the radiomics signature in term of prediction probability, highlighting its superior clinical utility. These findings suggest that the radiomics nomogram is more effective in predicting the likelihood of achieving a gestational age of \u0026ge;35 weeks in patients with unicornuate uterus. Figure.5 and Figure.6 display the AUC and DCA for both the training and test cohorts, as well as the graphical representation of the nomogram for clinical use.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eGenital tract anomalies in women often arise from genetic mutations or environmental factors that affect the development of the Mullerian or paramedian ducts. Among these anomalies, the unicornuate uterus is particularly notable for its impact on reproductive outcomes. Research has consistently shown that women with a unicornuate uterus face a notably higher risk of miscarriage compared to matched controls\u003csup\u003e[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]\u003c/sup\u003e, and they are also more likely to experience preterm delivery. Reichman et al. reported an incidence of unicornuate uterus occurring in approximately 1 in 4020 individuals in the general population. Their review highlighted that the rates of first-trimester abortion, second-trimester abortion, preterm delivery, and live birth were 24.3%, 9.7%, 20.1% and 51.5%, respectively \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. Additionally, several studies have demonstrated that women with a unicornuate uterus have an elevated risk of early miscarriage, though specific miscarriage rates are often not directly reported\u003csup\u003e[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e. Moreover, other studies have discovered various adverse obstetrics and neonatal outcomes associated with the condition\u003csup\u003e[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e. In our study, pregnancy outcomes in women with a univornuate uterus included 9.4% spontaneous abortion (with more than two occurrence), 13.5% primary infertility, 10.6% first trimester abortion, 5.3% preterm delivery and 70.6% full-term delivery. It is hypothesized that primary infertility and early miscarriage in these women may be caused by uterine cavity dysplasia. The malformed uterus could hinder embryo implantation, growth and development. Furthermore, insufficient blood supply to the uterine myometrium and relative cervical insufficiency may increase the risk of preterm delivery\u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e. Consistent with previous findings, the rate of cesarean delivery (59.4%) was significantly higher in women with a unicornuate uterus, which is largely attributed to malpresentation during delivery\u003csup\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWang et al.\u003csup\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003efound that the rates of in vitro fertilization and embryo transfer (IVF-ET) were significantly higher (12.6%) and the incidence of uterine rupture was (4.7%) among patients with a uncirnuate uterus. In our study, with the assistance of IVF, 23 patients (13.5%) with a unicornuate uterus successfully achieved pregnancy. It is generally believed that the unicornuate uterus is slightly smaller, with asymmetric polarity, which may influence pregnancy outcome\u003csup\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/sup\u003e. Therefore, women with a uincornuate uterus should limit the number of embryos typically opting for a single embryo transfer, which aligns with our findings. Additionally, 125 patients with a uincornuate uterus were able to conceive naturally.\u003c/p\u003e \u003cp\u003eThe rudimentary horn of an unicornuate uterus arises from unilateral failure of Mullerian duct migration and is often associated with anomalies in the urinary system, particularly renal agenesis\u003csup\u003e[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]\u003c/sup\u003e. According to the ESHRE/ESGE classification, the hemi-uterus is categorized into two sub-classes: U4a (with a functional rudimentary cavity) and U4b (without a functional rudimentary cavity). it is reported that U4b accounts for 85% of cases\u003csup\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]\u003c/sup\u003e, which is consistent with our findings. In our study, 51 patients (30%) had functional rudimentary cavity. Among these, 14 patients (8.2%) underwent rudimentary horn excision. We also observed that 3 patients (1.4%) had associated renal anomalies, which is lower than the rate reported in the literature (16\u0026ndash;38%)\u003csup\u003e[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eSang et al. found that patients with a normal uterine cavity (7-8cm) or longer (\u0026gt;\u0026thinsp;8cm) had higher clinical pregnancy rates in the IVF-ICSI treatments\u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003c/sup\u003e. In our previous study, we also demonstrated that women with a uincornuate uterus and longer uterine lengths (\u0026ge; 4.5cm) were more likely to achieve a full-term delivery\u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e. Similarly, Sardo et al.\u003csup\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u003c/sup\u003e discovered that women with dysmorphic uterus shapes, who underwent the novel Hysteroscopic Outpatient Metroplasty to Expand Dysmorphic Uteri (HOME-EU) technique, achieved a clinical pregnancy rate of 57% and a term delivery rate of 65%. These improvements were linked to the expansion of uterine volume and enhancement of uterine morphology. In the present study, we analyzed data from 170 women with a unicornuate uterus uusing 3D magnetic resonance imaging (MRI) reconstruction. Our results indicate that the women with larger uterine volumes had higher chances of to achieving full-term deliveries. Additionally, the surface area of the uterus was significantly greater in women who dilivered \u0026ge;35 gestational weeks groups compared to other three groups. These findings suggest that both uterine volume and surface area serve as independent protective factors for better reproductive outcomes.\u003c/p\u003e \u003cp\u003eWe developed and validated a predictive model for reproductive outcomes in women with a unicornuate uterus, incorporating clinical factors such as uterine volume, surface area, sphericity and diameter lines of uterine body, along with radiomics signatures extracted from multiple phases of 3D MRI scans. Previous research on 3D uterine reconstruction often relied on images with a slice thickness of 2-4mm, resulting in relatively lower resolution and accuracy\u003csup\u003e[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u003c/sup\u003e. In contrast, our study utilized a 3D T2-weighted imaging (T2WI) SPACE sequence with a slice thickness of 1mm and a voxel size of 1\u0026sdot;1\u0026sdot;1mm, enabling high-resolution 3D reconstructions with greater precision. The most lesion of uterus signals is not very obvious between the surrounding tissues\u003csup\u003e[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]\u003c/sup\u003e, so the reconstruction of the uterus can only be implemented in manual sketch, which may take approximately three hours in one case. In our study, we implemented machine learning algorithms to automatically segment the uterine volume, reducing this process to just a few seconds. This innovation significantly increased work efficiency and paved the way for large-scale data analysis in future research. For constructing our risk prediction model, we employed eight different machine learning models. When applied to predict the reproductive outcomes of women with a unicornuate uterus, the radiomics nomogram incorporating radiomics signatures and clinical feature showed good discrimination between patients with fertility outcome of unicornuate uterus with an area under the curve (AUC) of 0.86 in the training cohort and 0.84 in the validation cohort, indicating that this model had an absolute dominant advantage in predicting a reproductive outcome of unicornuate uterus.\u003c/p\u003e \u003cp\u003eThis study has several limitations. First, it is a single-center retrospective analysis with a small-sample size. To enhance the robustness of our findings, future studies should incorporate larger, multicenter datasets for validation. Second, the manual delineation of the uterine cavity is time-consuming and technically demanding. Currently, no available software can automatically perform this segmentation. The development and application of more advanced AI (artificial intelligence) based segmentation techniques would greatly improve efficiency, reduce observer variability and enhance consistency across studies.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study introduces a noninvasive 3D-reconstruction imaging technique, demonstrating that the uterine volume can serve as a reliable prognostic factor for predicting the gestational week of delivery in women with a unicornuate uterus. Furthermore, we developed a radiomics nomogram that combines both radiomic signatures and clinical factors. This cool shows promise for facilitating the individualized prediction of fertility outcome in patients with a unicornuate uterus, offering a personalized approach to their reproductive management.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by grants from Project of the National Key R\u0026amp;D Program of China (Grant No.2021YFC2701402)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declared no conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConception and study design: Jingxin Ding, Keqin Hua; Execution: Xinyu Pan, Guofu Zhang; Analysis: Hui Luo, Qing Zhou; Data interpretation: Hui Luo, Qing Zhou, Xinyu Pan, Guofu Zhang; Manuscript drafting and revising: Hui Luo, Qing Zhou, Jingxin Ding; Project administration: Keqin Hua; All authors have read and agreed to the published version of the manuscript. All authors contributed substantially to conception and design, or acquisition of data, or analysis and interpretation of data, and approved the final version of submission.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll procedure performed in studies involving human participants were in accordance with the ethical standard of Obstetrics and Gynecology Hospital Affiliated to Fudan University institutional review board（No.2022-081）.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInformed consent was obtained from all individual participants included in the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to publish\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors affirm that human research participants provided informed consent for publication of the all images.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eChan YY, Jayaprakasan K, Zamora J, et al. The prevalence of congenital uterine anomalies in unselected and high-risk populations: a systematic review[J]. \u003cem\u003eHum Reprod Update\u003c/em\u003e, 2011,17(6):761\u0026ndash;771.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaravelos SH, Cocksedge KA, Li TC. Prevalence and diagnosis of congenital uterine anomalies in women with reproductive failure: a critical appraisal[J]. \u003cem\u003eHum Reprod Update\u003c/em\u003e, 2008,14(5):415\u0026ndash;429.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCerekja A, Dillon KC, Racanska E, et al. Unicornuate uteri associated with contralateral renal agenesis and ovarian anomalies[J]. \u003cem\u003eJ Turk Ger Gynecol Assoc\u003c/em\u003e, 2011,12(3):189\u0026ndash;191.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAkhtar MA, Saravelos SH, Li TC, et al. Reproductive Implications and Management of Congenital Uterine Anomalies: Scientific Impact Paper No. 62 November 2019[J]. \u003cem\u003eBJOG\u003c/em\u003e, 2020,127(5):e1-e13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePfeifer SM, Attaran M, Goldstein J, et al. ASRM mullerian anomalies classification 2021[J]. \u003cem\u003eFertil Steril\u003c/em\u003e, 2021,116(5):1238\u0026ndash;1252.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrimbizis GF, Gordts S, Di Spiezio Sardo A, et al. The ESHRE/ESGE consensus on the classification of female genital tract congenital anomalies[J]. \u003cem\u003eHum Reprod\u003c/em\u003e, 2013,28(8):2032\u0026ndash;2044.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrimbizis GF, Di Spiezio Sardo A, Saravelos SH, et al. The Thessaloniki ESHRE/ESGE consensus on diagnosis of female genital anomalies[J]. \u003cem\u003eGynecol Surg\u003c/em\u003e, 2016,13:1\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZyla MM, Wilczynski J, Nowakowska-Glab A, et al. Pregnancy and Delivery in Women with Uterine Malformations[J]. \u003cem\u003eAdv Clin Exp Med\u003c/em\u003e, 2015,24(5):873\u0026ndash;879.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eReichman D, Laufer MR, Robinson BK. Pregnancy outcomes in unicornuate uteri: a review[J]. \u003cem\u003eFertil Steril\u003c/em\u003e, 2009,91(5):1886\u0026ndash;1894.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEgbase PE, Al-Sharhan M, Grudzinskas JG. Influence of position and length of uterus on implantation and clinical pregnancy rates in IVF and embryo transfer treatment cycles[J]. \u003cem\u003eHum Reprod\u003c/em\u003e, 2000,15(9):1943\u0026ndash;1946.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrimbizis GF, Di Spiezio Sardo A, Saravelos SH, et al. The Thessaloniki ESHRE/ESGE consensus on diagnosis of female genital anomalies[J]. \u003cem\u003eHum Reprod\u003c/em\u003e, 2016,31(1):2\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi XQ, Qian HJ, Zhang XY, et al. Analysis of the reproductive outcomes and the size of the unicornuate uterus measured by magnetic resonance imaging and their relationship[J]. \u003cem\u003eArch Gynecol Obstet\u003c/em\u003e, 2019,299(5):1321\u0026ndash;1330.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBehr SC, Courtier JL, Qayyum A. Imaging of mullerian duct anomalies[J]. \u003cem\u003eRadiographics\u003c/em\u003e, 2012,32(6):E233-250.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTaylor RA, Pare JR, Venkatesh AK, et al. Prediction of In-hospital Mortality in Emergency Department Patients With Sepsis: A Local Big Data-Driven, Machine Learning Approach[J]. \u003cem\u003eAcad Emerg Med\u003c/em\u003e, 2016,23(3):269\u0026ndash;278.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan YY, Jayaprakasan K, Tan A, et al. Reproductive outcomes in women with congenital uterine anomalies: a systematic review[J]. \u003cem\u003eUltrasound Obstet Gynecol\u003c/em\u003e, 2011,38(4):371\u0026ndash;382.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Y, Nisenblat V, Yang P, et al. Reproductive outcomes in women with unicornuate uterus undergoing in vitro fertilization: a nested case-control retrospective study[J]. \u003cem\u003eReprod Biol Endocrinol\u003c/em\u003e, 2018,16(1):64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Zhao YY, Qiao J. Obstetric outcome of women with uterine anomalies in China[J]. \u003cem\u003eChin Med J (Engl)\u003c/em\u003e, 2010,123(4):418\u0026ndash;422.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHiersch L, Yeoshoua E, Miremberg H, et al. The association between Mullerian anomalies and short-term pregnancy outcome[J]. \u003cem\u003eJ Matern Fetal Neonatal Med\u003c/em\u003e, 2016,29(16):2573\u0026ndash;2578.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTakami M, Aoki S, Kurasawa K, et al. A classification of congenital uterine anomalies predicting pregnancy outcomes[J]. \u003cem\u003eActa Obstet Gynecol Scand\u003c/em\u003e, 2014,93(7):691\u0026ndash;697.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJacquinet A, Millar D, Lehman A. Etiologies of uterine malformations[J]. \u003cem\u003eAm J Med Genet A\u003c/em\u003e, 2016,170(8):2141\u0026ndash;2172.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNaeh A, Sigal E, Barda S, et al. The association between congenital uterine anomalies and perinatal outcomes - does type of defect matters?[J]. \u003cem\u003eJ Matern Fetal Neonatal Med\u003c/em\u003e, 2022,35(25):7406\u0026ndash;7411.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang S, Wang K, Hu Q, et al. Perinatal outcomes of women with Mullerian anomalies[J]. \u003cem\u003eArch Gynecol Obstet\u003c/em\u003e, 2023,307(4):1209\u0026ndash;1216.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi X, Ouyang Y, Yi Y, et al. Pregnancy outcomes of women with a congenital unicornuate uterus after IVF-embryo transfer[J]. \u003cem\u003eReprod Biomed Online\u003c/em\u003e, 2017,35(5):583\u0026ndash;591.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNahum GG. Uterine anomalies. How common are they, and what is their distribution among subtypes?[J]. \u003cem\u003eJ Reprod Med\u003c/em\u003e, 1998,43(10):877\u0026ndash;887.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePados G, Tsolakidis D, Athanatos D, et al. Reproductive and obstetric outcome after laparoscopic excision of functional, non-communicating broadly attached rudimentary horn: a case series[J]. \u003cem\u003eEur J Obstet Gynecol Reprod Biol\u003c/em\u003e, 2014,182:33\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChun SS, Chung MJ, Chong GO, et al. Relationship between the length of the uterine cavity and clinical pregnancy rates after in vitro fertilization or intracytoplasmic sperm injection[J]. \u003cem\u003eFertil Steril\u003c/em\u003e, 2010,93(2):663\u0026ndash;665.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDi Spiezio Sardo A, Florio P, Nazzaro G, et al. Hysteroscopic outpatient metroplasty to expand dysmorphic uteri (HOME-DU technique): a pilot study[J]. \u003cem\u003eReprod Biomed Online\u003c/em\u003e, 2015,30(2):166\u0026ndash;174.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePan HX, Liu P, Duan H, et al. Using 3D MRI can potentially enhance the ability of trained surgeons to more precisely diagnose Mullerian duct anomalies compared to MR alone[J]. \u003cem\u003eEur J Obstet Gynecol Reprod Biol\u003c/em\u003e, 2018,228:313\u0026ndash;318.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDurnea CM, Siddiqi S, Nazarian D, et al. 3D-Volume Rendering of the Pelvis with Emphasis on Paraurethral Structures Based on MRI Scans and Comparisons between 3D Slicer and OsiriX(R)[J]. \u003cem\u003eJ Med Syst\u003c/em\u003e, 2021,45(3):27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee SR, Kim YJ, Kim KG. A Fast 3-Dimensional Magnetic Resonance Imaging Reconstruction for Surgical Planning of Uterine Myomectomy[J]. \u003cem\u003eJ Korean Med Sci\u003c/em\u003e, 2018,33(2):e12.\u003c/span\u003e\u003c/li\u003e \u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable.1\u0026nbsp;\u003c/strong\u003eDemographic maternal characteristics of 170 women with unicornuate uteri.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eParameters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eMean\u0026plusmn;SD\u003c/p\u003e\n \u003cp\u003e（range or percentage）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eMaternal age, (year)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e34.26\u0026plusmn;4.75\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eGravidity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eParity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eSubtypes of hemi-uterus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;U4aC0V0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e51(30%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; U4bC0V0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e119(70%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eRemoval of rudimentary horn with or without salpingectomy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e14(8.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eEndometriosis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e4(1.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eSevere dysmenorrhea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e15(8.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eSpontaneous abortion\u0026sup3;2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e16(9.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eMode of pregnancy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eNatural conception\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e125(73.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eIVF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e23(13.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eAssociated renal anomalies, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e3(1.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eTable.2\u0026nbsp;\u003c/strong\u003eReproductive outcomes of 170 women with unicornuate uterus\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003ePregnancy outcome\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eMean\u0026plusmn;SD\u003c/p\u003e\n \u003cp\u003e(range or percentage)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003ePrimary infertility\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e23(13.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 277px;\"\u003e\n \u003cp\u003e\u0026lt;24 gestational weeks\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e18(10.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e24-35gestational weeks\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e9(5.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e\u0026sup3;35gestational weeks\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e120(70.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eVaginal delivery\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e48(28.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003eCesarean delivery\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 277px;\"\u003e\n \u003cp\u003e101(59.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"archives-of-gynecology-and-obstetrics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"arch","sideBox":"Learn more about [Archives of Gynecology and Obstetrics](https://www.springer.com/journal/404)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/arch/default.aspx","title":"Archives of Gynecology and Obstetrics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Unicornuate uterus, Reproductive outcomes, Three-dimensional MRI, Radiomic features","lastPublishedDoi":"10.21203/rs.3.rs-9157364/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9157364/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose:\u003c/strong\u003e To develop and validate a radiomics nomogram integrating three-dimensional (3D) MRI features for predicting ≥35 weeks' gestation in women with unicornuate uterus.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e This retrospective study enrolled 170 unicornuate uterus patients who underwent pelvic 3D MRI between 2005-2022. Patients were categorized into four groups based on reproductive outcomes: primary infertility, \u0026lt;24 weeks, 24-35 weeks, and ≥35 weeks gestation. The uterus was segmented and reconstructed in 3D using ITK-SNAP. Radiomics features (n=1,834), including volume, surface area, sphericity, and diameters, were extracted. Feature selection employed LASSO regression, and eight machine learning algorithms were compared for model construction. A radiomics nomogram integrating selected features with clinical variables was developed. Predictive performance was assessed using ROC and decision curve analyses.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Uterine volume was significantly larger in the ≥35 weeks gestation group compared to infertility and \u0026lt;24 weeks groups (P\u0026lt;0.05). Surface area was also greater in the ≥35 weeks group versus the infertility group (P\u0026lt;0.05). No significant differences in axis lengths were observed among groups. The radiomics nomogram demonstrated robust discrimination for predicting ≥35 weeks gestation, achieving AUCs of 0.86 (95% CI: 0.79-0.92) in the training cohort and 0.84 (95% CI: 0.69-0.98) in the validation cohort. Decision curve analysis confirmed favorable clinical utility.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003eUterine volume measured by 3D MRI reconstruction serves as a reliable prognostic factor for predicting term delivery in unicornuate uterus patients. The developed radiomics nomogram integrating radiomics signatures with clinical indicators enables individualized prediction of reproductive outcomes.\u003c/p\u003e","manuscriptTitle":"Prediction of reproductive outcomes in unicornuate uterus women based on Three-dimensional MRI radiomic features","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-25 09:25:38","doi":"10.21203/rs.3.rs-9157364/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-11T05:58:20+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-22T16:26:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"193221478473741278954452798835224259726","date":"2026-04-20T10:41:56+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"20419911016996981513081329559883605756","date":"2026-04-18T07:24:09+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-16T08:51:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"331223151383578599649448603316667600281","date":"2026-04-13T15:44:09+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"274951354504484511168622977248448052001","date":"2026-04-13T08:12:50+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-23T06:23:49+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-21T13:55:57+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-19T15:08:25+00:00","index":"","fulltext":""},{"type":"submitted","content":"Archives of Gynecology and Obstetrics","date":"2026-03-18T09:18:38+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"archives-of-gynecology-and-obstetrics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"arch","sideBox":"Learn more about [Archives of Gynecology and Obstetrics](https://www.springer.com/journal/404)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/arch/default.aspx","title":"Archives of Gynecology and Obstetrics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"6594eac2-a641-4136-844d-9eff189b98d4","owner":[],"postedDate":"March 25th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-11T05:58:20+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[],"tags":[],"updatedAt":"2026-05-11T06:11:44+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-25 09:25:38","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9157364","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9157364","identity":"rs-9157364","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}