Prognostic Significance of the Ki-67 Index with an Optimal Cut-off Value of 32.5% in Uterine Sarcoma and Uterine Carcinosarcoma: A Retrospective Cohort Study Authors: Piaopiao Ye, Fang Wang, Xinyue Dai, Xiaoli Hu, Yuanqiu Wang Published: 2026 ABSTRACT ObjectiveUterine sarcomas and uterine carcinosarcomas are rare heterogeneous malignancies with poor prognosis. Though carcinosarcomas have been pathologically reclassified as dedifferentiated endometrial carcinomas, their highly aggressive clinical behavior merits combined analysis with uterine sarcomas. This study aimed to investigate the prognostic value of Ki-67 proliferation index in major histologic subtypes and identify an optimal cut-off value for risk stratification.Materials and MethodsWe retrospectively enrolled 86 surgically treated primary uterine sarcoma and carcinosarcoma patients (2006–2019). Ki-67 labeling index was detected via immunohistochemistry. The optimal cut-off value for Ki-67 was determined using receiver operating characteristic (ROC) curve analysis. Overall survival (OS) and progression-free survival (PFS) were evaluated via Kaplan-Meier curves and Cox regression.ResultsThe optimal Ki-67 cut-off value was 32.5%. High Ki-67 expression (≥32.5%) was significantly associated with older age (P=0.001), postmenopausal status (P=0.002), and aggressive histological subtypes 0.760 (P<0.001). High Ki-67 group had worse 3-year PFS (18.5% vs 76.4%) (range, 22–73) and 3-year OS (25.5% vs 85.4%). Multivariate analysis confirmed high Ki-67 as an independent prognostic factor for poorer PFS and OS (P < 0.001). Larger tumor size (P < 0.001), postmenopausal status (PFS: P = 0.001; OS: P = 0.002), and high-risk histological type (P < 0.001) also remained significant. Notably, the 32.5% cut-off applies across major histological subtypes.ConclusionKi-67 index serves as an independent biomarker for uterine sarcoma and carcinosarcoma. The 32.5% cut-off effectively divide patients into low- and high-risk groups, with the latter demonstrating significantly inferior survival outcomes. Introduction Despite the rarity, uterine sarcomas are highly aggressive and lethal cancers, accounting for approximately 3–7% of all uterine malignancies.1,2 Historically, uterine sarcomas are broadly categorized into pure mesenchymal tumors and heterologous sarcomas. The pure mesenchymal category mainly includes uterine leiomyosarcoma (u-LMS) and endometrial stromal sarcoma (ESS), with the latter further subdivided into low-grade (LG-ESS), high-grade endometrial stromal sarcoma (HG-ESS) and undifferentiated uterine sarcoma (UUS). The heterologous sarcomas encompass uterine adenosarcomas (AS) and uterine carcinosarcomas (UCS).3–5 Although carcinosarcoma is now reclassified as a dedifferentiated or metaplastic form of endometrial carcinoma, its distinctly aggressive behavior justifies its continued inclusion in most retrospective studies on uterine sarcomas, as well as in the 2003 World Health Organization (WHO) classification.1,3 Therefore, carcinosarcoma was also included in the present study. Each of these tumors exhibits distinct epidemiological as well as clinical and pathologic features. The rarity and histopathological diversity of these tumors have impeded a clear consensus on both the risk factors for poor prognosis and optimal therapeutic management. To this end, several clinicopathological parameters have been proposed as prognostic determinants, including histopathological type, FIGO stage, patient age, tumor size, and mitotic index.6,7 Unfortunately, it remains difficult to identify patients who would benefit from surgery based on these parameters. Currently, there are no reliable indicators to predict recurrent uterine sarcoma and carcinosarcoma, it is necessary to identify novel prognostic and predictive biomarkers. Ki-67 is a nuclear non-histone protein encoded by the MKI67 gene, located on chromosome 10q25-ter, which acts as a biosurfactant to disperse mitotic chromosomes.8 First found by Gerdes et al9 in Hodgkin and Reed-Sternberg cells, Ki-67 protein can be detected in all active phases of cell cycle (G1, S, G2 and M phase), but not in resting phase G0.10–12 Ki-67 expression fluctuates dynamically throughout the cell cycle, with low levels during the G1 and S phases, peaking at mitosis, and then declining sharply in the subsequent anaphase and telophase.13 Ki-67 labeling index, as assessed by immunohistochemistry (IHC) on paraffin-embedded tissue sections is considered a prognostic indicator in various clinical conditions. A higher Ki-67 index is generally associated with poorer patient outcomes.14 As a recognized but non-specific marker of cellular proliferation, Ki-67 expression has been reported as a prognostic indicator in a wide range of malignancies, such as breast cancer,15 NK/T-cell lymphoma,16 prostate cancer17 and malignant Pleural Mesothelioma.18 In breast cancer, its clinical utility has been formally endorsed by the International Ki-67 in Breast Cancer Working Group (IKWG), with the consensus stating that only Ki-67 scores of ≤5% and ≥30% can be used for clinical decision.19 Whereas at the St. Gallen Consensus Conference, the majority of the expert panel voted to define a threshold of ≥20% as high Ki67 status.20 Moreover, Ki-67 is also proved as an important prognostic indicator in various gynecologic tumors, including cervical cancer,21 ovarian cancer22 and endometrial cancer.23 In uterine sarcoma, accumulating evidence shows a prognostic role for Ki-67. In uterine leiomyosarcoma, K. Mayerhofer et al24 reported that Ki-67 expression was detected in 50% of cases and was significantly associated with shortened disease-free survival. In a study of 84 uterine leiomyosarcomas, Ki-67 immunoreactivity was significant by univariate analysis and was identified as an adverse prognostic factor (p = 0.01).25 The work of Y. Zhou et al26 further demonstrated that Ki-67 is overexpressed in uterine sarcomas, with positivity conferring a worse prognosis and being significantly correlated with reduced OS in LMS patients. In their 2012 study, Park et al27 investigated potential therapeutic targets and prognostic markers for low-grade endometrial stromal sarcoma (LG-ESS). They reported detectable Ki-67 activity in 21 patients (53.8%). Critically, Ki-67 expression was significantly associated with poorer disease-free survival and OS. However, its expression showed no significant correlation with other clinicopathological factors, including age, menopause status, parity, body mass index, FIGO stage, tumor size, depth of myometrial invasion, lymphovascular space invasion, lymph node metastasis, or adjuvant therapy. In another retrospective analysis of endometrial stromal sarcoma,28 the Ki-67 index was an independent prognostic factor for PFS (P = 0.001) and determined 35% as the optimal cut-off value for predicting recurrence. However, the reported cut-off value of Ki-67 in uterine sarcoma subtypes varies greatly, ranging from 10% to 35%, and the standardized scoring methodology is lacking. Most previous studies have focused on a single histological subtype, leaving the prognostic value of Ki-67 in uterine sarcomas, including carcinosarcoma, unexplored. Therefore, this study aimed to evaluate the prognostic significance of the Ki-67 proliferation index in patients with uterine sarcoma and carcinosarcoma and to determine an optimal cut-off value for risk stratification, thereby providing a practical biomarker to guide postoperative adjuvant therapy and surveillance. Materials and Methods Study Population This retrospective study was approved by the Ethics Committee of the First Affiliated Hospital of Wenzhou Medical University in Zhejiang Province, China (approval number KY2025-R390) and was performed in accordance with the Declaration of Helsinki. Informed consent was obtained from patients. This study retrospectively collected archived medical records from patients diagnosed with primary uterine sarcoma or carcinosarcoma who were admitted to the First Affiliated Hospital of Wenzhou Medical University between May 2006 and October 2019. Inclusion criteria were as follows: (1) pathologically confirmed with uterine sarcoma or uterine carcinosarcoma; (2) complete clinical records, involving age at diagnosis, menstrual status, tumor size, blood test results and ultrasonography results before surgery, surgical procedures, histological types, stage (according to International Federation of Gynecology and Obstetrics, FIGO 2009). Exclusion criteria included (1) Absence of primary surgical intervention; (2) combined with other malignant tumors; (3) unavailability of formalin-fixed paraffin-embedded (FFPE) tissue specimens for Ki-67 IHC staining; (4) complete lack of postoperative follow-up data (absolutely no return visits recorded after surgery). All patients in the study group underwent a definitive primary surgery, with total hysterectomy serving as the standard core procedure, with or without bilateral salpingo-oophorectomy or lymph node dissection. A total of 126 patients were initially identified. Among them, 40 cases were excluded due to no primary surgery (n=12), combined with other malignant tumors (n=2), unavailability of tumor paraffin (n=21) and complete loss of follow-up after surgery (n=5). The remaining 86 patients were included in the final analysis (Supplementary Figure 1). The unavailability of paraffin blocks for Ki-67 IHC in 21 cases and the lack of complete follow-up data in 5 cases may have introduced selection bias. Follow-up was conducted every 3–6 months for the first 2 years, every 6–12 months for the subsequent 3 years, and annually thereafter. All of the patients were followed up with vagino-rectal palpation, serum tumor marker tests, and ultrasound, supplemented by CT or MRI when necessary. All patients were followed up until October 2024, with a minimum follow-up period of 5 years. Overall survival (OS) was defined as time interval from primary surgery to all-cause death. Progression-free survival (PFS) was measured from time of surgery until disease progression or death of any cause, whichever occurred first. Patients lost to follow-up during the observation period were treated as censored at their last valid clinical visit. Immunohistochemistry FFPE tissue blocks were retrieved for all eligible individuals. Ki-67 IHC was performed on all 86 tissue blocks using a standardized protocol in the Department of Pathology of our hospital. Ki-67 IHC staining was performed using a CONFIRMTM anti Ki-67 (30–9) Rabbit Monoclonal Primary Antibody on a Ventana BenchMark IHC/ISC automated immunostainer according to the manufacturer’s instructions. Regions with optimal tumor fixation and high cell density were selected from paraffin sections. Immunohistochemical results of Ki-67 were independently evaluated by two senior pathologists in gynecological tumors, who were blinded to patients’ outcomes and other clinicopathological characteristics. Ki-67 proliferative activity was quantified using the hotspot counting method. Under low-power magnification, the region with the highest density of Ki-67-positive neoplastic cells was identified; a minimum of 400 tumor cells were counted in the hotspot area at high power magnification. The Ki-67 proliferation index was defined as the percentage of tumor cells displaying distinct nuclear immunoreactivity, expressed as a value ranging from 0% to 100%. If the evaluation of the two pathologists is inconsistent, the paraffin block needs to be reevaluated and discussed to reach a consensus; consistency is considered achieved when the pathologists’ assessment for the proportion of positive tumor cells differs by ≤10%. Statistical Analysis Data were statistically analyzed using IBM SPSS Statistics 27.0 software (IBM Statistics, Chicago, IL, USA). Continuous variables and categorical variables were analyzed by using Student’s t-tests and Chi-Square tests (χ2-test), respectively. The optimal cut-off values of the Ki-67 labeling index, tumor size, age and Body Mass Index (BMI) were calculated by receiver operating characteristics (ROC) curve analysis and Youden Index (Youden index = sensitivity + specificity −1) regarding patients’ PFS data.29 Using OS as the outcome, the same ki-67 cut-off value was obtained. The Kaplan–Meier method and Log rank test were performed to analyze the survival of the patients. Univariable and multivariable survival analyses were performed using Cox proportional hazards models. Prognostic factors significantly associated with PFS or OS in the Cox regression model were evaluated in multivariate Cox regression analysis by backwards stepwise selection. Hazard ratios (HRs) estimated from the Cox-regression analysis were reported as relative risks with corresponding 95% confidence intervals (CIs). The proportional risk hypothesis was tested by Schoenfeld residual, and no significant violation was found (all P>0.05). Post-hoc power analysis using the Schoenfeld formula estimated the power to detect Ki-67’s effect on PFS, with ≥80% deemed sufficient. Continuous data were expressed as mean ±SD. Two-sided tests of P<0.05 were regarded to demonstrate statistical significance. Internal validation was performed using 5-fold cross-validation with R software (version 4.4.3). A two-sided p-value < 0.05 was considered statistically significant. Results Patient Characteristics A total of 86 patients with uterine sarcoma or carcinosarcoma admitted to the First Affiliated Hospital of Wenzhou Medical University were enrolled in this study. All patients underwent hysterectomy as the primary surgical procedure, with or without bilateral salpingo-oophorectomy or lymph node dissection. The median age at diagnosis was 48.5 years (range: 22–73) and most of the patients (57, 66.28%) were premenopausal. The cohort was primarily comprised of patients with LG-ESS (n=36, 41.86%) and u-LMS (25, 29.07%). Other subtypes include AS (n=13, 15.12%), HG-ESS (n=6, 6.98%), UUS (n=5, 5.81%) and UCS (n=7, 8.14%). In addition, the majority of patients (70, 81.4%) were initially diagnosed at FIGO stage I, while 12 (13.95%) were at stage II. Of all entire cohort, 39 patients underwent lymph node dissection. By the end of the follow-up period, 39 patients had relapsed and 31 patients had died. The estimated 3-year and 5-year PFS rates were 58.1% and 54.5%, respectively. The corresponding 3-year and 5-year OS rates were 67.2% and 60.2%, respectively. Relationship Between Ki-67 and Patients’ Clinicopathological Characteristics The expression of Ki-67, which exhibited a nuclear staining pattern, was assessed. The Ki-67 labeling index was quantified as the percentage of positive tumor cells. Figure 1 illustrates examples of low and high Ki-67 expression levels in u-LMS. The optimal cut-off value for the Ki-67 index was determined using the receiver operating characteristic (ROC) curve and the Youden index (Figure 2). Based on PFS data, the cut-off value was 32.5%. The area under the curve (AUC) was 0.760 (95% CI: 0.657–0.863), with a sensitivity of 71.8% and a specificity of 87.3% (P<0.001). Using OS as the outcome, the same cut-off value was obtained (AUC = 0.747; 95% CI: 0.655–0.866, sensitivity 77.4%, specificity 78.2%). To evaluate model generalizability, a 5-fold cross-validation was performed. The cross-validated mean AUC was 0.712 (95% CI: 0.584–0.825), compared with an apparent AUC of 0.760 (95% CI: 0.657–0.857) for the full-dataset model. The difference was only 0.048, well below the conventional overfitting threshold of 0.15, indicating no evidence of severe overfitting (Supplementary Figure 2). Based on the 32.5% cut-off, patients were stratified into a low Ki-67 expression group (<32.5%) and a high Ki-67 expression group (≥32.5%). The associations between Ki-67 expression status and clinical features are summarized in Table 1. Ki-67 levels were significantly correlated with age at diagnosis (P=0.001), menopausal status (P=0.002) and histological types (P<0.001). However, there was no association of Ki-67 expression with other clinical features such as BMI, tumor size, and FIGO stage.Table 1Comparison of Clinicopathological Characteristics Between Patients with Low Ki-67 and High Ki-67 ExpressionVariablesNo. of Patients (%)Ki-67 < 32.5% (n = 50)Ki-67 ≥ 32.5% (n = 36)PAge (years)0.001 <52.5058 (67.44%)4117 ≥52.5028 (32.56%)919BMI (kg/m2)0.377 <24.0049 (56.98%)2623 ≥24.0037 (43.02%)2413Size(mm)0.194 <61.5041 (47.67%)2714 ≥61.5045 (52.33%)2322Menopausal status0.002 Premenopausal57 (66.28%)4017 Postmenopausal29 (33.72%)1019Histological types<0.001 AS13 (15.12%)49 LG-ESS36 (41.86%)324 u-LMS25 (29.07%)1213 UUS5 (5.81%)05 UCS7 (8.14%)07 HG-ESS6 (6.98%)24FIGO stage0.776 Stage I70 (81.40%)4229 Stage II–IV15 (17.44%)87Lymphadenectomy0.570 Yes48 (55.81%)2820 No38 (44.19%)2215Ovariectomy0.064 Yes73 (84.88%)3934 No13 (15.12%)112Notes: Data are presented as n (%) unless otherwise specified. The optimal cut-off values for Ki-67 (32.5%), age (52.5 years), BMI (24.0 kg/m2), and tumor size (61.5 mm) were determined by ROC curve analysis based on progression-free survival.Abbreviations: BMI, Body Mass Index; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Figure 1Representative images of high (A) and low (B) Ki-67 expression in u-LMS (×400).Micrograph: two tissue sections, A & B, with brown nuclear stain on pale blue.Figure 2Receiver operating characteristic (ROC) curves of the Ki-67 index for predicting progression-free survival (PFS, (A)) and overall survival (OS, (B)) in patients with uterine sarcoma and carcinosarcoma.Two ROC curve plots for Ki-67 index predicting progression-free survival and overall survival. Table 1Comparison of Clinicopathological Characteristics Between Patients with Low Ki-67 and High Ki-67 ExpressionVariablesNo. of Patients (%)Ki-67 < 32.5% (n = 50)Ki-67 ≥ 32.5% (n = 36)PAge (years)0.001 <52.5058 (67.44%)4117 ≥52.5028 (32.56%)919BMI (kg/m2)0.377 <24.0049 (56.98%)2623 ≥24.0037 (43.02%)2413Size(mm)0.194 <61.5041 (47.67%)2714 ≥61.5045 (52.33%)2322Menopausal status0.002 Premenopausal57 (66.28%)4017 Postmenopausal29 (33.72%)1019Histological types<0.001 AS13 (15.12%)49 LG-ESS36 (41.86%)324 u-LMS25 (29.07%)1213 UUS5 (5.81%)05 UCS7 (8.14%)07 HG-ESS6 (6.98%)24FIGO stage0.776 Stage I70 (81.40%)4229 Stage II–IV15 (17.44%)87Lymphadenectomy0.570 Yes48 (55.81%)2820 No38 (44.19%)2215Ovariectomy0.064 Yes73 (84.88%)3934 No13 (15.12%)112Notes: Data are presented as n (%) unless otherwise specified. The optimal cut-off values for Ki-67 (32.5%), age (52.5 years), BMI (24.0 kg/m2), and tumor size (61.5 mm) were determined by ROC curve analysis based on progression-free survival.Abbreviations: BMI, Body Mass Index; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Comparison of Clinicopathological Characteristics Between Patients with Low Ki-67 and High Ki-67 Expression Notes: Data are presented as n (%) unless otherwise specified. The optimal cut-off values for Ki-67 (32.5%), age (52.5 years), BMI (24.0 kg/m2), and tumor size (61.5 mm) were determined by ROC curve analysis based on progression-free survival. Abbreviations: BMI, Body Mass Index; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Representative images of high (A) and low (B) Ki-67 expression in u-LMS (×400). Receiver operating characteristic (ROC) curves of the Ki-67 index for predicting progression-free survival (PFS, (A)) and overall survival (OS, (B)) in patients with uterine sarcoma and carcinosarcoma. Survival Analysis The prognostic value of all clinical and histopathological parameters—including age, BMI, tumor size, menstrual status, histological type, FIGO stage, performance of lymphadenectomy, oophorectomy and Ki-67 index—was assessed using univariable analyses for both PFS and OS (Table 2). These analyses identified age at diagnosis, tumor size, menstrual status, histological type, and the Ki-67 index as significant prognostic factors for both endpoints. While FIGO stage was associated with PFS, it showed no significant association with OS. In addition to clinical parameters such as tumor size, menopausal status, and histological types, Ki-67 index was also identified as a powerful independent prognostic factor for both PFS and OS in the multivariable analysis (Table 3). Figure 3 shows the Kaplan–Meier survival curves for PFS and OS, categorized by Ki-67 status. Patients in the low-expression group had a better 3-year PFS (76.4% vs 18.5%) and OS (85.4% vs 25.5%) than those in the high-expression group (log-rank P < 0.001 for both).Table 2Univariate Analysis for Progression Free Survival and Overall Survival of Uterine Sarcoma Patients (n=86)ParameterProgression Free SurvivalOverall SurvivalHR (95% CI)P-valueHR (95% CI)P-valueAge (years) <52.501 (referent)1 (referent) ≥52.502.950 (1.567–5.556)0.0012.602 (1.281–5.283)0.008BMI (kg/m2) <24.001 (referent)1 (referent) ≥24.001.544(0.821–2.901)0.1771.430(0.705–2.900)0.322Size (mm) <61.501 (referent)1 (referent) ≥61.504.348(2.053–9.207)<0.0015.101(2.083–12.493)<0.001Menstrual status Premenopausal1 (referent)1 (referent) Postmenopausal2.702 (1.439–5.074)0.0022.857466 (1.217–4.996)0.0012Histological types<0.001<0.001 AS1 (referent)1 (referent) LG-ESS1.765(0.221–14.120)0.5920.846(0.094–7.571)0.881 u-LMS8.783(1.159–66.578)0.0366.196(0.810–47.374)0.079 UUS17.278(2.009–148.619)0.00917.958(2.075–155.429)0.009 UCS3.721(0.386–35.833)0.2562.800 (0.253–30.999)0.401 HG- ESS14.909(1.719–129.304)0.01413.175(1.512–114.824)0.020FIGO stage Stage I1 (referent)1 (referent) Stage II–IV2.439(1.166–4.733)0.0171.523(0.655–3.538)0.328Lymphadenectomy No1 (referent)1 (referent) Yes0.918(0.487–1.731)0.7921.049(0.514–2.142)0.895Ovariectomy Yes1 (referent)1 (referent) No1.774 (0.630–4.994)0.2782.479(0.707–12.444)0.137Ki-67 index <32.5%1 (referent)1 (referent) ≥32.5%6.271(3.093–12.712)<0.0017.843(3.352–18.352)<0.001Abbreviations: HR, hazard ratio; CI, confidence interval; BMI, Body Mass Index; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Table 3Multivariate Cox Proportional Hazards Regression Analysis for Progression Free Survival and Overall Survival of Uterine Sarcoma Patients (n=86)ParameterProgression Free SurvivalOverall SurvivalHR (95% CI)P-valueHR (95% CI)P-valueAge (years) <52.501 (referent)1 (referent) ≥52.501.042 (0.390–2.789))0.9340.311(0.109–1.195)0.095Size (mm) <61.501 (referent)1 (referent) ≥61.506.372(2. 682–16.896)<0.00111.512(3.804–34.838)<0.001Menstrual status Premenopausal1 (referent)1 (referent) Postmenopausal5.354 (1.975–14.509)0.0016.610 (1.945–22.465)0.002Histological types<0.001<0.001 AS1 (referent)1 (referent) LG-ESS4.078(0.464–35.876)0.2052.110(0.209–21.257)0.527 u-LMS18.823(2.269–156.112)0.00711.120(1.324–93.391)0.027 UUS15.888(1.644–153.549)0.01729.680 (2.860–313.943)0.005 UCS0.793(0.070–8.998)0.8511.871(0.136–25.733)0.639 HG- ESS47.816(4.768–479.515)0.00142.172(3.879–458.540)0.002Ki-67 index <32.5%1 (referent)1 (referent) ≥32.5%8.071(3.372–19.319)<0.0017.443(2.788–19.872)<0.001Abbreviations: HR, hazard ratio; CI, confidence interval; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Figure 3Kaplan–Meier survival curves of Ki-67 index for PFS (A) and OS (B).Kaplan–Meier survival curves showing progression free survival and overall survival by Ki-67 index groups.Image A: Survival analysis with Y-axis (Cumulative survival, 0.0-1.0) and X-axis (PFS in months, 0.00-250.00). Legend: Ki-67 index. Groups: ≤32.5% and >32.5%, both with censored data. ≤32.5% curve starts at 1.0, drops to ~0.82 by 50 months, stabilizes near 0.78 from 100 to 220 months. >32.5% curve starts at 1.0, falls sharply to ~0.48 by 10 months, ~0.31 by 30 months, ~0.24 by 80 months and ~0.19 by 110 months, staying near 0.19 through 190 months. Image B: Survival analysis with Y-axis (Cumulative survival, 0.0-1.0) and X-axis (OS in months, 0.00-250.00). Legend: Ki-67 index. Groups: ≤32.5% and >32.5%, both with censored data. ≤32.5% curve starts at 1.0, remains near 1.0 until 50 months, drops to ~0.95, then ~0.86 by 70 months, staying near 0.86 from 80 to 220 months. >32.5% curve starts at 1.0, decreases to ~0.66 by 15 months, ~0.50 by 35 months, ~0.40 by 60 months, ~0.35 by 80 months, ~0.26 by 120 months, remaining near 0.26 through 200 months. Table 2Univariate Analysis for Progression Free Survival and Overall Survival of Uterine Sarcoma Patients (n=86)ParameterProgression Free SurvivalOverall SurvivalHR (95% CI)P-valueHR (95% CI)P-valueAge (years) <52.501 (referent)1 (referent) ≥52.502.950 (1.567–5.556)0.0012.602 (1.281–5.283)0.008BMI (kg/m2) <24.001 (referent)1 (referent) ≥24.001.544(0.821–2.901)0.1771.430(0.705–2.900)0.322Size (mm) <61.501 (referent)1 (referent) ≥61.504.348(2.053–9.207)<0.0015.101(2.083–12.493)<0.001Menstrual status Premenopausal1 (referent)1 (referent) Postmenopausal2.702 (1.439–5.074)0.0022.857466 (1.217–4.996)0.0012Histological types<0.001<0.001 AS1 (referent)1 (referent) LG-ESS1.765(0.221–14.120)0.5920.846(0.094–7.571)0.881 u-LMS8.783(1.159–66.578)0.0366.196(0.810–47.374)0.079 UUS17.278(2.009–148.619)0.00917.958(2.075–155.429)0.009 UCS3.721(0.386–35.833)0.2562.800 (0.253–30.999)0.401 HG- ESS14.909(1.719–129.304)0.01413.175(1.512–114.824)0.020FIGO stage Stage I1 (referent)1 (referent) Stage II–IV2.439(1.166–4.733)0.0171.523(0.655–3.538)0.328Lymphadenectomy No1 (referent)1 (referent) Yes0.918(0.487–1.731)0.7921.049(0.514–2.142)0.895Ovariectomy Yes1 (referent)1 (referent) No1.774 (0.630–4.994)0.2782.479(0.707–12.444)0.137Ki-67 index <32.5%1 (referent)1 (referent) ≥32.5%6.271(3.093–12.712)<0.0017.843(3.352–18.352)<0.001Abbreviations: HR, hazard ratio; CI, confidence interval; BMI, Body Mass Index; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Univariate Analysis for Progression Free Survival and Overall Survival of Uterine Sarcoma Patients (n=86) Abbreviations: HR, hazard ratio; CI, confidence interval; BMI, Body Mass Index; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Table 3Multivariate Cox Proportional Hazards Regression Analysis for Progression Free Survival and Overall Survival of Uterine Sarcoma Patients (n=86)ParameterProgression Free SurvivalOverall SurvivalHR (95% CI)P-valueHR (95% CI)P-valueAge (years) <52.501 (referent)1 (referent) ≥52.501.042 (0.390–2.789))0.9340.311(0.109–1.195)0.095Size (mm) <61.501 (referent)1 (referent) ≥61.506.372(2. 682–16.896)<0.00111.512(3.804–34.838)<0.001Menstrual status Premenopausal1 (referent)1 (referent) Postmenopausal5.354 (1.975–14.509)0.0016.610 (1.945–22.465)0.002Histological types<0.001<0.001 AS1 (referent)1 (referent) LG-ESS4.078(0.464–35.876)0.2052.110(0.209–21.257)0.527 u-LMS18.823(2.269–156.112)0.00711.120(1.324–93.391)0.027 UUS15.888(1.644–153.549)0.01729.680 (2.860–313.943)0.005 UCS0.793(0.070–8.998)0.8511.871(0.136–25.733)0.639 HG- ESS47.816(4.768–479.515)0.00142.172(3.879–458.540)0.002Ki-67 index <32.5%1 (referent)1 (referent) ≥32.5%8.071(3.372–19.319)<0.0017.443(2.788–19.872)<0.001Abbreviations: HR, hazard ratio; CI, confidence interval; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Multivariate Cox Proportional Hazards Regression Analysis for Progression Free Survival and Overall Survival of Uterine Sarcoma Patients (n=86) Abbreviations: HR, hazard ratio; CI, confidence interval; AS, uterine adenosarcomas; LG-ESS, low-grade endometrial stromal sarcoma; u-LMS, uterine leiomyosarcoma; UUS, undifferentiated uterine sarcoma; UCS, uterine carcinosarcomas; HG-ESS, high-grade endometrial stromal sarcoma; FIGO, International Federation of Gynecology and Obstetrics. Kaplan–Meier survival curves of Ki-67 index for PFS (A) and OS (B). Discussion Uterine sarcoma and carcinosarcoma are characterized by high invasion and poor prognosis, but there are no reliable and effective biomarkers to precisely predict postoperative recurrence and death risk. Ki-67, a recognized marker of cellular proliferation, has been reported as an important prognostic indicator in various malignancies, such as breast cancer,15 NK/T-cell lymphoma,16 prostate cancer,17 cervical cancer,21 ovarian cancer22 and endometrial cancer.23 The Ki-67 proliferation threshold has become an indicator to guide clinical decision in breast cancer. Although Ki-67 has shown potential value in uterine sarcoma and carcinosarcoma, the reported cut-off value varies greatly, and the standardized scoring methodology is lacking. This retrospective study, conducted over an 18-year period with a minimum follow-up of 5 years, investigated the prognostic value of the Ki-67 index in patients with uterine sarcoma and carcinosarcoma. Our study demonstrated that Ki-67 is a powerful and independent biomarker both for PFS and OS. The optimal cut-off value of Ki-67 identified in our study is 32.5%, which is closely aligned the 35% threshold reported by Meng et al23 in endometrial stromal sarcoma.28 The optimal cut-off value stratified patients into low- and high-expression group. Patients in high-expression group showed significantly worse outcomes, with estimated 3-year PFS rates of 18.5% vs 76.4% and 3-year OS rates of 25.5% vs 85.4%. These findings confirm the value of this threshold for risk stratification in uterine sarcoma and carcinosarcoma. Our findings demonstrated that more attention should be paid to the prognostic management of patients with high expression of Ki ‑ 67. Patients with a Ki-67 index ≥32.5% may warrant a more intensive prognostic management strategy, which could include appropriately extending the duration of adjuvant therapy, increasing the number of treatment cycles, and implementing a more rigorous postoperative surveillance protocol. For example, National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology for Uterine Neoplasms (Version 3.2026)30 recommend that patients with stage II or III disease who have undergone complete surgical resection with negative margins, including those with HG-ESS, UUS, LMS, observation may be considered, as may systemic therapy and/or external beam radiation therapy(EBRT). In this context, Ki-67 may serve as a useful adjunct for individualized treatment decisions, with high expression favoring more aggressive adjuvant approaches. It is noteworthy that, in addition to the Ki-67 index, several clinicopathological factors appeared as independent prognostic factors, including tumor size, menopausal status, and histological type. These findings are consistent with our previous study7 and other published series,4,25,31,32 emphasizing the importance of these clinical parameters in evaluating prognosis. Notably, the correlation between ki‑67 expression and histological subtypes is particularly significant. All UCS and UUS cases, as well as the majority of HG-ESS cases, fell into the high Ki-67 group. This pattern is biologically reasonable, because UCS, UUS and HG-ESSS are considered to be more aggressive with rapid disease progression and early recurrence. Based on their aggressive behavior and poor outcomes, the World Health Organization (WHO) has reclassified UUS and HG-ESS as high-grade uterine sarcomas.33 This finding is consistent with the summary by E. D’Angelo et al,1 which indicated markedly elevated Ki-67 expression in undifferentiated endometrial sarcomas—a subtype characterized by a highly aggressive clinical course, wherein most patients succumb to the disease within two years of diagnosis. Collectively, these findings reinforce that Ki-67 expression level serves as a direct indicator of tumor aggressiveness. What needs to be explained is that although low-grade endometrial stromal sarcoma (LG-ESS) was the most prevalent subtype in the final study cohort, uterine leiomyosarcoma (u-LMS) actually constituted the largest subgroup among all uterine sarcomas prior to the exclusion of cases lacking Ki-67 immunohistochemical data or lost to follow-up. In our 2022 study,7 uterine leiomyosarcoma (u-LMS) was the most common histopathological type, accounting for 40.64% (63/155) of cases, which is consistent with the reported range of 40% to 60% in the literature.33–35 What’s more, in our study, univariate analysis revealed that age was significantly associated with PFS and OS; While, in the multivariate analysis, this association did not retain statistical significance, which may be attributable to the fact that this was a single-center study, other independent prognostic factors demonstrated stronger correlations with uterine sarcoma and carcinosarcoma. Nevertheless, this finding does not diminish the prognostic importance of age. In fact, many articles have reported that age is an important factor in predicting the prognosis of uterine sarcoma.36,37 In addition, we did not observe a significant correlation between Ki-67 and FIGO stage, which may be attributable to the predominance of early-stage patients (81.4% stage I) in our cohort, limiting the power to detect an association with advanced-stage disease. This study has several limitations that should be acknowledged. First, its retrospective, single-center design and relatively small cohort size may introduce selection bias and limit the generalizability of the results. Second, the absence of Ki-67 IHC data, mainly attributable to unavailable paraffin blocks from earlier years, together with the exclusion of five patients who never received follow-up, may have potentially introduced selection bias. Third, the optimal cut-off value of 32.5% was derived using a ROC curve analysis. Although our 5-fold cross-validation demonstrated internal stability and the cut-off closely aligns with the previously reported 35% threshold in endometrial stromal sarcoma, this does not replace formal validation. Finally, the assessment of the Ki-67 index, while performed by two experienced pathologists, remains a semi-quantitative method with inherent subjectivity; the lack of fully automated digital image analysis might impact the reproducibility of the scoring. Therefore, our findings should be confirmed in larger, prospective multicenter studies. In conclusion, this study identifies the Ki-67 index as a powerful and independent prognostic biomarker in uterine sarcoma and carcinosarcoma. High Ki-67 expression predicts poorer PFS and OS. Ki-67 assessment should therefore be incorporated into clinical practice to improve risk stratification. Patients with a high Ki-67 index should be considered for a more aggressive follow-up and treatment strategy, including prolonging adjuvant therapy and stricter surveillance.