Introduction

Testicular torsion (TT) is a common urological emergency in pediatric patients, involving the twisting of the spermatic cord, which leads to impaired blood flow to the testis. If not treated promptly, it can result in ischemic damage1. Therefore, timely and accurate diagnosis is crucial to preserving testicular viability and preventing complications. One of the most severe consequences of TT is orchiectomy, which can have profound physical and psychological effects on both the child and their family. Thus, the ability to predict the risk of orchiectomy is essential for guiding clinical management and ensuring better outcomes.

Several clinical factors have been identified as closely related to testicular survival after torsion, including the duration of symptoms, degree of torsion, and certain hematological parameters1,2. However, accurately predicting the risk of orchiectomy in emergency situations remains a challenge. A reliable and easily interpretable risk assessment tool could provide families with a clearer understanding of the expected surgical outcomes.

Nomograms, which integrate multiple clinical variables into a simplified chart, are widely used in medical research3. By incorporating various risk factors, nomograms offer personalized predictions of clinical outcomes and can serve as practical tools for clinicians. Although nomograms have been extensively applied in the medical field, large-scale, multicenter predictive models specifically for the risk of orchiectomy after TT in children remain limited.

This study aims to develop and validate a nomogram model to predict the risk of orchiectomy in children after TT. By collecting clinical data from two tertiary pediatric medical centers, we will identify independent risk factors associated with orchiectomy and construct a reliable model to assist clinicians in their decision-making process.

Materials and methods

Patient selection and data collection

This retrospective study collected clinical data from 468 children diagnosed with TT between January 2009 and June 2024 at the Children’s Hospital of Fudan University (training cohort) and Anhui Provincial Children’s Hospital (validation cohort). To ensure consistency across study sites, we standardized the definitions of all clinical variables before data collection and strictly adhered to these definitions throughout the study. All included patients were aged <18 years and met the diagnostic criteria for TT. Exclusion criteria were: (1) neonatal TT; (2) acute scrotal emergency without surgical exploration; (3) confirmed torsion of the testicular appendage during surgery; (4) incomplete medical records. This study was approved by the Ethics Committees of Medical Research at Anhui Provincial Children’s Hospital (No. EYLL-2024-014) and Children’s Hospital of Fudan University (No. 2024 − 366), and informed consent was waived by both committees. All procedures involving human participants in this study were conducted in accordance with the ethical standards of the relevant institutional and/or national research committees, and with the 1964 Helsinki Declaration and its subsequent amendments or comparable ethical standards. The study report follows the Strengthening the Reporting of Cohort Studies in Surgery (STROCSS) criteria4.

Demographic data for the included children consisted of age and body mass index (BMI). Clinical data included duration of symptoms, Testicular Workup for Ischemia and Suspected Torsion(TWIST) score5, affected side, season of occurrence, time of occurrence, presence of undescended testis(UDT), ultrasound findings, intraoperative type, degree, and direction of torsion. Hematological parameters recorded within two hours of admission included white blood cell count (WBC), neutrophils, lymphocytes, monocytes, eosinophils, basophils, platelet count, mean platelet volume (MPV), plateletcrit (PCT), platelet distribution width (PDW), and C-reactive protein (CRP).

Surgical management

Upon admission, all children underwent a thorough physical examination, TWIST scoring, ultrasound examination, and hematological tests. If TT was suspected, emergency surgical exploration was performed immediately. During surgery, the child was placed in the supine position, and after adequate anesthesia, the inguinal or scrotal approach was chosen for exploration based on the location of the testis. After detorsion of the twisted testis, warm saline compresses were applied for 20 min to assess its vascularity. The tunica albuginea was then incised to observe the bleeding time from the testicular parenchyma: immediate bleeding was classified as grade I, bleeding within 10 min as grade II, and no bleeding within 10 min as grade III6. For grade I and II cases, orchiopexy was performed and considered as successful salvage, whereas for grade III cases, the necrotic torsed testis was excised. Contralateral orchiopexy was performed in all patients.

Statistical analysis

Statistical analyses were performed using R (version 4.4.0) and SPSS (version 26.0.0). Categorical data were presented as counts (percentages) and analyzed using the chi-square test or Fisher’s exact test. Continuous variables that followed a normal distribution were expressed as mean ± standard deviation, while non-normally distributed variables were presented as median (interquartile range). Parametric tests (t-test) and non-parametric tests (Mann-Whitney U test) were used for the analysis of normally and non-normally distributed continuous variables, respectively. Variables with more than 20% missing data were directly excluded, while the remaining missing values were imputed using the mean or median. Initially, univariate logistic regression was conducted to assess the association between each independent variable and orchiectomy. Significant factors identified in the univariate analysis (P <0.05) were included in a multivariate logistic regression model. The final model was determined using backward selection based on the Akaike Information Criterion (AIC), and the selected factors were used to construct the nomogram. The model’s discrimination ability was evaluated using sensitivity, specificity, accuracy, positive predictive value, and negative predictive value, as well as the receiver operating characteristic (ROC) curve, with the area under the ROC curve (AUC) calculated. A calibration curve was used to assess the model’s calibration, and the results of the Hosmer-Lemeshow test were presented visually. Additionally, decision curve analysis (DCA) was performed to evaluate the net clinical benefit of the model. A value of P <0.05 was considered statistically significant.

Results

Demographic characteristics

The study flowchart is shown in Fig. 1. A total of 468 cases were included, with 230 cases (49.15%) eventually undergoing orchiectomy. The clinical characteristics of children with TT in the training and validation cohorts are presented in Table 1. There were no statistically significant differences in the variables between the two groups (P > 0.05), indicating good comparability between the groups.

Fig. 1
figure 1

Flowchart of participants in this study.

Full size image
Table 1 Baseline comparison between the training cohort and validation cohort.
Full size table

Variable selection

In the training cohort, 24 variables were collected from 327 children. The variable coding was as follows: TWIST score grading (Intermediate risk = 2, High risk = 3), TT of UDT (Yes = 1, No = 2), and blood flow signal (Visible = 1, Invisible = 2). Orchiectomy status was used as the dependent variable, and the statistically significant variables were included in the univariate and multivariate analyses (Table 2). The results showed that duration of symptoms (OR, 1.02; 95% CI: 1.01–1.03; P = 0.001), neutrophil (OR, 0.84; 95% CI: 0.74–0.95; P = 0.005), eosinophil (OR, 27.51; 95% CI: 2.53–298.88; P = 0.006), degree of torsion (OR, 1.01; 95% CI: 1.01–1.01; P <0.001), TT of UDT (OR, 0.12; 95% CI: 0.04–0.43; P = 0.001), TWIST score grading (OR, 3.10; 95% CI: 1.54–6.23; P = 0.002), and blood flow signal (OR, 37.12; 95% CI: 8.82–156.24; P <0.001) were independent risk factors for orchiectomy in children with TT.

Table 2 Perform univariable and multivariable analyses on factors associated with testicular loss.
Full size table

Nomogram construction and validation

Based on the results of the multivariate logistic regression analysis, we constructed a nomogram to predict the likelihood of orchiectomy in children with TT (Fig. 2). The nomogram assigns a score to each value of the variables, allowing clinicians to quickly assess risk by vertically matching the total score of all variables to the estimated probability of orchiectomy. The higher the total score, the greater the risk of orchiectomy.

Fig. 2
figure 2

Nomogram to predict testicular loss after TT in children. The nomogram is used to add up all the points determined by each variable on the scale. The total points projected onto the bottom scale represent the probability of testicular loss in children with TT.

Full size image

The nomogram demonstrated good predictive performance in both the training and validation cohorts, with AUC values of 0.93 (95% CI: 0.91–0.96) and 0.86 (95% CI: 0.79–0.92), respectively (Fig. 3A, B). The calibration curves showed good agreement between the predicted and actual probabilities, with Hosmer-Lemeshow test P-values of 0.347 and 0.218 for the training and validation cohorts, respectively (Fig. 4A, B). Additionally, DCA was plotted for both the training and validation cohorts to assess the clinical utility of the predictive model. In the DCA curves of the training and validation cohorts, patients achieved higher net benefits when the predicted probability thresholds were between 2% and 97% for the training cohort and between 12% and 98% for the validation cohort, respectively (Fig. 5A, B).

Fig. 3
figure 3

(A) ROC curve of the training cohort, with an AUC of 0.93 (95% CI: 0.91–0.96); (B) ROC curve of the validation cohort, with an AUC of 0.86 (95% CI: 0.79–0.92).

Full size image
Fig. 4
figure 4

(A) Calibration curve of the training cohort, with the Hosmer-Lemeshow test showing a P-value of 0.347; (B) Calibration curve of the validation cohort, with the Hosmer-Lemeshow test showing a P-value of 0.218.

Full size image
Fig. 5
figure 5

(A) In the DCA curve of the training cohort, patients achieved a higher net benefit at predicted probability thresholds between 2% and 97%; (B) In the DCA curve of the validation cohort, patients achieved a higher net benefit at predicted probability thresholds between 12% and 98%.

Full size image

In this study, the optimal cut-off value was determined to be 0.58, based on ROC curve analysis. This cut-off value balances sensitivity and specificity, providing a threshold to classify patients at higher risk of testicular excision after torsion. When the predicted probability exceeds 58%, it indicates a significantly increased risk of testicular loss, suggesting that more aggressive clinical intervention or monitoring may be necessary. The AUC, sensitivity, specificity, accuracy, positive predictive value, negative predictive value, and cut-off values for the training and validation cohorts are listed in Table 3.

Table 3 Performance metrics of the model in both the training and validation cohorts.
Full size table

Discussion

To our knowledge, this is the first nomogram established to predict the risk of orchiectomy after TT in children through a large-scale multicenter retrospective data analysis. The results showed that the duration of symptoms, levels of neutrophils and eosinophils, degree of torsion, presence of UDT, TWIST score grading, and ultrasound blood flow signal were key factors influencing the risk of orchiectomy. The nomogram constructed based on these findings demonstrated excellent discriminatory ability, with AUC values of 0.93 and 0.86 in the training and validation cohorts, respectively, indicating satisfactory predictive performance. DCA further confirmed that the model provides significant net clinical benefit. Additionally, the indications for orchiectomy after TT were comprehensively assessed by experienced urologists from two tertiary pediatric medical centers, ensuring the accuracy of surgical management. This predictive model not only offers surgeons a valuable tool to assess testicular viability but also facilitates effective communication with parents, helping to clarify treatment expectations and optimize the clinical decision-making process.

The duration of symptoms is one of the important predictive factors in this study. Previous studies have shown that the longer the duration of symptoms, the significantly lower the likelihood of testicular survival7. It is generally believed that surgical intervention within 6 h of symptom onset is crucial for saving the testicle. If the testicle twists within 0–4 h of pain onset, the salvage rate can be as high as 95%. However, if it persists for 8–10 h, the salvage rate drops to 45–60%, and thereafter decreases sharply8. Research by Yi et al.9 indicated that the orchiectomy rate for patients undergoing surgery within 6 h of symptom onset was 5.9%, while it rose to 82.1% within 24 h. Notably, only 23.8% of patients received timely treatment within 6 h of symptom onset, emphasizing the importance of early medical intervention, especially in children and adolescents, as delays in treatment not only affect testicular survival but may also have long-term implications for future fertility1,10. Factors contributing to prolonged symptom duration include being younger than 6 years old, lack of alternative diagnoses, history of previous treatments or symptom progression, distance from major urban areas, and abdominal pain11,12. Additionally, Bayne et al.13 pointed out that delays related to distance from the hospital and transfer time are major factors in the delayed treatment of pediatric TT. Therefore, our model highlights the duration of symptoms as a key indicator for predicting outcomes, indicating that patients with prolonged symptoms should be regarded as high-risk individuals in urgent need of prompt diagnosis and surgical intervention. Future research should continue to explore mechanisms to reduce delays in care, such as through telemedicine or optimizing referral systems, to improve treatment outcomes for these high-risk patients14.

Another significant predictive indicator is the degree of TT. Numerous studies have confirmed that the higher the degree of torsion, the greater the severity of ischemic injury to the testicle, thereby increasing the risk of orchiectomy15,16. Yi et al.9 found through a cross-sectional study that a degree of TT exceeding 180° is one of the risk factors for orchiectomy. Dias et al.17 pointed out that a torsion degree between 360° and 540° is a critical point for significant obstruction of arterial blood flow, posing a risk of testicular damage. However, as the degree of torsion increases, the blood supply to the affected testicle decreases, which also means that the time required to cause damage to the testicular vasculature is significantly reduced18. Therefore, when assessing testicular viability, relying solely on the degree of torsion is insufficient; a comprehensive evaluation must include other factors such as the duration of symptoms19.

Currently, an increasing number of scholars have successfully incorporated hematological parameters into the construction of clinical predictive models19,20,21. In this study, we found that hematological parameters, particularly changes in neutrophils and eosinophils, are also important predictive indicators of the risk of orchiectomy. Although several studies have indicated a certain association between monocytes and the salvage of testicular torsion2,19, He et al.22 further discovered that MPV could serve as a marker for assessing testicular viability. While monocytes showed some correlation in univariate analysis (p = 0.030, OR = 1.99, 95% CI: 1.07 ~ 3.70) in our study, multivariate analysis results indicated that neutrophils (p = 0.005, OR = 0.84, 95% CI: 0.74 ~ 0.95) and eosinophils (p = 0.006, OR = 27.51, 95% CI: 2.53 ~ 298.88) are more effective in predicting the risk of orchiectomy. Additionally, we observed a significant correlation between lower neutrophil levels and increased risk of orchiectomy, in contrast to the eosinophil response. The literature indicates that during ischemia-reperfusion, neutrophils migrate from the bloodstream to tissues, releasing reactive oxygen species that lead to tissue damage23,24. After the detorsion of TT, the massive influx of neutrophils and the reactive oxygen species they release are primary contributors to ischemia-reperfusion injury25. Since our hematological parameters were obtained before surgery and did not include the ischemia-reperfusion process, we hypothesize that in the early stages of TT, blood can still perfuse, leading to higher levels of neutrophils; however, as the duration of torsion increases, blood perfusion gradually ceases, resulting in lower neutrophil levels. Future research could further explore the specifics of this mechanism and its clinical significance to enhance management strategies for patients with TT.

The prevalence of TT in UDT ranges from 5.9 to 10.2% of all TT cases16,26,27. In our study, UDT cases accounted for 11.3% (53/468) of all TT cases, slightly higher than previously reported. The testicular salvage rate for UDT is generally lower; Zvizdic et al.16 reported a salvage rate of 35% in their series, which is similar to the 36.4% reported by Zhong et al.28. However, in our study, the testicular salvage rate for UDT was only 30.2% (16/53), significantly lower than the rate for normally descended testicles (53.5%, 222/415). Zilberman et al.29 also found that the surgical salvage rate for cryptorchid torsion patients was only 10%, far lower than the salvage rate for normally descended testicles (56%). The low salvage rate for UDT may be due to atypical symptoms, misdiagnosis as other diseases (such as gastrointestinal disorders), and delayed presentation16,28,30.

Another significant predictive factor for orchiectomy is the absence of blood flow signals on Doppler ultrasonography (DUS). The use of preoperative Doppler ultrasound is controversial, as the results are highly operator-dependent and may delay surgical exploration31. Pinar et al.31 evaluated 2,922 cases of TT and found that 32.7% of patients underwent DUS prior to surgery, with a sensitivity of 85.2% and a specificity of 52.7%. They concluded that, in early torsion or partial/intermittent torsion, DUS may misleadingly show arterial blood flow, and continuous blood flow does not necessarily exclude the possibility of TT. It must be compared with the contralateral side32. In late-stage torsion, if DUS shows the absence of blood flow, it indicates prolonged testicular ischemia, further decreasing the possibility of salvage and increasing the risk of orchiectomy9.

Another important predictor is the TWIST score. This 7-point scoring system is specifically used for the preoperative evaluation of scrotal emergencies and includes the following parameters: testicular swelling (2 points), hard testicle (2 points), absent cremasteric reflex (1 point), nausea/vomiting (1 point), and high riding testis (1 point). The cutoff for low-risk and high-risk categories is 2 and 5 points, respectively5. Although several studies have demonstrated the decision-making value of the TWIST score in evaluating acute scrotal symptoms5,33,34, few have applied this scoring system to the decision for orchiectomy in TT. In our study, all patients were assessed as being in the intermediate-to-high-risk range, and the results showed that the TWIST score was significantly associated with the risk of orchiectomy. A high TWIST score typically indicates a longer symptom duration and more pronounced clinical manifestations35, suggesting that the testicle may already be in an ischemic state, thus reducing the possibility of salvage. It is worth noting that, although the TWIST score was confirmed as an independent risk factor in this study, its accuracy still depends on the clinical experience and judgment of healthcare professionals34.

Although this study provides valuable insights into predicting the risk of orchiectomy after TT in children, there are several limitations. First, as a retrospective cohort study, the results may be influenced by selection bias and incomplete data, especially given potential variations in the standardization of data collection across different medical institutions. Second, the data collection spanned a long period (2009–2024), during which diagnostic and surgical techniques may have evolved. These technological advancements could introduce heterogeneity into the data, potentially affecting the results and generalizability of the model. Finally, although this study explored the salvage rate for torsion in UDT, the small number of UDT cases (only 53) may limit the generalizability of these findings. Additionally, while we identified some independent predictors, we did not thoroughly investigate factors contributing to delayed diagnosis and treatment, such as family economic status, health insurance policies, and the educational level of guardians, which could provide more information for optimizing the model. Therefore, despite the multicenter validation, large-scale prospective studies are still needed in the future to further validate and refine the predictive model.

Conclusion

This study successfully developed a nomogram model to predict the risk of orchiectomy after TT in children and identified several independent risk factors, including duration of symptoms, degree of torsion, hematological parameters, and UDT. The model demonstrated good predictive performance in both the training and validation cohorts, providing a robust decision-support tool for clinicians managing cases of TT. However, larger-scale multicenter prospective studies are needed in the future to further validate the model’s reliability and applicability, ultimately optimizing the management strategies for pediatric TT.