Missed miscarriage, also known as late miscarriage, is a pregnancy loss in which a dead fetus or embryo is retained in the maternal uterine cavity without timely spontaneous expulsion, which occurs in approximately 15% of clinical pregnancies [1]. A curettage cannot be easily performed in this condition because the dead embryo or placenta is degenerated or mechanized and adheres closely to the uterine wall.

Conventional treatment of induced abortion includes medication, surgery, and expectant treatment. Surgical abortion has a higher rate of complete abortion than medication [2]. However, surgical abortion can damage the basal layer of the endometrium, resulting in fibrous connective tissue hyperplasia and dysplasia of endometrial regeneration, resulting in endometrial repair disorder or cavity adhesions and thereby severely affecting embryo implantation [3]. Endometrial damage is exacerbated by curettage in patients with missed miscarriage, and a high number of curettages ($\ge$2 times), secondary curettage, and combined pelvic inflammatory disease may be independent risk factors for uterine adhesions after medical abortion [4]. The number of previous curettages can cause failure of subsequent pregnancy in patients with medical abortion [5]. Medication involves the use of drugs to simulate the process of natural abortion; this approach can avoid surgical trauma. The conventional medication treatment is mifepristone combined with misoprostol. However, the success rate of medical abortion using medication is approximately 72–93%, and the risk of residual tissue is high [6, 7].

Presently, although many studies have reported on the factors affecting incomplete abortion, the prediction of the probability of successful expulsion of residual tissue from the uterine cavity has not yet been established [8, 9, 10, 11]. Hamel et al. [12] developed and evaluated a clinical model that can predict the probability of successful medical abortion for patients with missed miscarriage, which is based on body mass index, misoprostol pretreatment, presence of symptoms of abdominal pain or vaginal bleeding, and number of previous curettages. Lu et al. [13] investigated the factors that contribute to complete abortion after medical abortion in patients with missed miscarriage, including the size of residual tissue and blood supply. By using recursive partitioning analysis, patients with similar probabilities of complete abortion were classified into individual groups, and the probability of each group was calculated. However, the precise probability could not be derived for each patient.

Nomograms are a common predictive tool for assessing patient prognosis. By integrating diverse predictive and decision variables, individual numerical probabilities of clinical events are generated to meet the requirement for personalized medicine [14]. Based on recent methodological guidelines for prognostic modeling [15, 16], in the present study, we developed a scoring system through logistic regression analysis and a nomogram and assessed the feasibility of prediction with receiver operating characteristic (ROC) curves and calibration curves. This approach can enable to quantitatively evaluate the individual probability of expulsion of residual tissue from the uterus and to provide personalized guidance to patients with missed miscarriage.

Many studies have reported the predictors of success of medical abortion or incomplete abortion. However, inconsistencies were observed in defining the precise predictors of success of medical abortion, mainly because of differences in study populations, dosage and administration route of misoprostol, administration regimen, and the timing and criteria for the ultrasonographic assessment of a complete abortion [22]. In some studies, an endometrial thickness of 15 mm determined by ultrasound was assessed as the primary factor for complete abortion. However, a growing body of evidence suggests that this limit is too stringent, and the disappearance of the intrauterine gestational sac suggests no requirement for further clinical intervention [23, 24].

In the present study, ultrasonography was performed on the first day after medical abortion in patients with missed miscarriage. Based on ultrasonography findings and clinical factors, univariate and multivariate logistic regression analyses were conducted to determine the predictive factors for the expulsion of intrauterine residual tissue from the uterine cavity of these patients. The possibilities for the expulsion of the residual tissue from the uterine cavity were then explored. We found that three factors, namely gravidity, residual tissue size, and residual tissue with blood supply, significantly affected the expulsion of residual tissue from the uterine cavity after medical abortion for patients with missed miscarriage (all p 0.05). We then developed a scoring system based on the nomogram and quantified the probability of expulsion of residual tissue from the uterine cavity. At the total score of 77.78 points and the AUC value of 0.849, the probability of expulsion of residue tissue from the uterine cavity was 71%.

The current study suggests that low gravidity, smaller residual size, and residue tissue with no blood supply are the favorable factors for the expulsion of residual tissue from the uterine cavity after medical abortion. A multicenter randomized controlled clinical trial also showed that nulliparity was associated with the success rate of medical abortion, thus suggesting that the responsiveness of the uterus to misoprostol varies with the growth and extension of the uterus in past pregnancies [25]. Reeves et al. [26] showed that five or more deliveries (absolute risk: 11.90%, OR: 4.60%) and gestational age of $\ge$8 weeks (absolute risk: 4.10%, OR: 2.10%) were significantly associated with the uterine expulsion of residual tissue.

Fernlund et al. [27] conducted a secondary analysis of published prospective clinical trials and showed that the success rate of expectant treatment for anembryonic miscarriages increased with gestational age and crown-rump length, but further increased with the decrease in the diameter of the gestational sac. The authors also found that lower levels of serum progesterone (P) or serum $\beta$-human chorionic gonadotropin ($\beta$-HCG) were associated with a higher success rate of expectant treatment. However, their study did not observe a relationship between blood flow or vascular pulsation and treatment success, and no further 3 dimensions (3D) ultrasound was performed to assess blood flow and vascular indices. Liu et al. [28] showed that the levels of serum $\beta$-HCG, P, follicle stimulating hormone (FSH), embryo transfer (ET), and peak systolic velocity (PSV) decreased significantly after medical abortion; moreover, the decrease was significantly higher in the non-residual group than in the residual group. $\beta$-HCG, P, FSH, $\beta$-HCG/ET, $\beta$-HCG/PSV, and $\beta$-HCG/resistance index (R) have high specificity and sensitivity for the diagnosis of residual tissue in the uterine cavity. Schwärzler et al. [29] found that intervillous blood flow or blood supply was a predictor of treatment success, a finding that is partially consistent with the results of our study.

In another prospective study, residual tissue width and blood supply were found to be significant predictors (p 0.001) of complete abortion after medical abortion. Specifically, the likelihood of complete abortion was high (88.24%), moderate (67.57%), and low (34.29%) based on the following respective findings: residual pregnancy tissue width 1 cm and no blood supply, residual pregnancy tissue $>$1 cm and no blood supply, and residual pregnancy tissue width $>$1 cm and blood supply [13]. The probability of expulsion of residual tissue from the uterine cavity in this study was 71%, which was similar to our study.

To date, few studies have investigated the precise probability of expulsion of residual tissue from the uterine cavity. The main advantage of using a nomogram is the ability to estimate individualized risks based on patient and disease characteristics and to generate individual probabilities of clinical events [30, 31]. Therefore, in the present study, we established a scoring system based on a nomogram. For example, a patient was assigned a value for each variable through the rating scale, and the variable was assigned a total score of 90, which corresponds to a probability of residual expulsion of 88% according to the nomogram. Thus, the total score of the influencing factors was 90, which was higher than the threshold value of 77.78, and the probability of successful natural expulsion of residual tissue from the uterus was approximately 88%. Based on these values, the patient was recommended to wait for the natural expulsion of residual tissue and to presently avoid the clearance of the uterus [32, 33].

The present study has some limitations. First, serum HCG and progesterone levels of the patients were not tested because the results of these tests require a waiting period and prompt clinical management of these patients is required for informed decision-making. Second, the sample size was too small, and we hope that subsequent studies will be conducted with a larger sample size. We also plan to use this scoring system in clinical settings and determine its clinical validity by comparing the curettage rate, complications after abortion such as endometritis and uterine adhesions, and clinical pregnancy rate.

In summary, a scoring system based on the developed nomogram was established to quantitatively evaluate the probability of expulsion. At the total score of 77.78 for the assigned weights, the AUC of the ROC curve was 0.849, and the probability of successful expulsion of residual tissue from the uterine cavity was 71% for patients with missed miscarriage after medical abortion.

Obtain availability of data and materials through corresponding author email.

HHZ: Conceptualization, data curation, data analysis, writing - original draft. MYJ: Conceptualization, data curation, obtained informed consent. JH: Data analysis, writing - review and editing. ZYL: The acquisition, analysis, or interpretation of data for the work. HYZ: Conceptualization, formal analysis, data analysis, writing - review and editing. All authors approved the final version of the manuscript to be published. All authors contributed to editorial changes in the manuscript. All authors have participated sufficiently in the work to take public responsibility for appropriate portions of the content and agreed to be accountable for all aspects of the work in ensuring that questions related to its accuracy or integrity.

The present study was approved by the Ethics Committee of Hangzhou Women’s Hospital (no.: [2021] medical ethics (007) No.12), and informed consent was obtained from all the included participants.

We thank International Science Editing for editing this manuscript.

The study was supported by the project from Zhejiang Province Science and Technology Program [2022KY1018].

The authors declare no conflict of interest.