Clinical Characteristics of Unilateral versus Bilateral Ovarian Endometriomas and Risk Stratification in Stage IV Disease Using the rASRM Score: A Retrospective Cohort Study from a Tertiary Medical Center

Endometriosis is a chronic inflammatory disorder characterized by the ectopic growth of endometrial tissue outside the uterine cavity. 1 Globally, it affects approximately 176 million women, predominantly of reproductive age, and is closely associated with pelvic pain and infertility. 1 Recent epidemiological data indicate a rising incidence, potentially linked to environmental pollution, lifestyle changes, and advancements in diagnostic technology. 2 Although its etiology remains incompletely understood, the main hypotheses include retrograde menstruation, coelomic metaplasia, immune dysregulation, and genetic predisposition. 3 , 4 Emerging evidence highlights the role of epigenetic and molecular dysregulation in disease progression, including abnormal DNA methylation, miRNA expression, and aberrations in HOX genes and Wnt signaling pathways. 5 , 6 However, how this molecular heterogeneity translates into distinct clinical phenotypes—particularly anatomical staging and ovarian reserve impairment—remains incompletely characterized. Clinically, based on anatomical location, endometriosis is classified into ovarian, peritoneal, deeply infiltrating, and rare subtypes. 7 , 8 Symptoms encompass dysmenorrhea, dyspareunia, dyschezia, dysuria, cyclic rectal bleeding or hematuria, scar-related pain and swelling, fatigue, and infertility. 8 Endometriomas can directly compromise ovarian reserve by damaging cortical follicular units, disrupting local blood supply, and inducing stromal oxidative stress. 9 , 10 Bilateral cysts, involving both ovaries, theoretically double this parenchymal damage, providing a biological rationale for comparing unilateral vs bilateral disease in terms of antral follicle count and fertility scores. Beyond disrupting ovarian tissue architecture, ovarian endometriomas are strongly associated with deeply infiltrating lesions involving the bowel, vagina, or ureters, underscoring their clinical complexity and severity. 11 Prior research has indicated that the presence of endometriomas correlates with more extensive deeply infiltrating nodules in the pelvis and bowel compared to non-cystic cases. 11 Existing studies primarily focus on cyst size, pain intensity, or fertility outcomes, while the relationship between cyst laterality and disease stage remains controversial. For instance, Sznurkowski et al 12 reported a predominance of endometriomas in the left ovary, whereas Ulukus et al 13 found that right-sided cysts were more frequently associated with cul-de-sac obliteration. Therefore, it is worth investigating whether bilateral cysts indicate more extensive pelvic disease or greater impairment of ovarian reserve. This large-scale retrospective analysis aims to elucidate the clinical differences between unilateral and bilateral ovarian endometriomas, providing a basis for formulating individualized management strategies. Meanwhile, although rASRM Stage IV represents the most severe disease category, 14 it is clinically heterogeneous, encompassing patients with widely varying surgical complexity and reproductive potential. The current binary staging does not capture this intra-Stage IV variability. Thus, this study also explores whether the total rASRM score—using the median cutoff of 74 points—can identify a high-risk subgroup within Stage IV patients, providing hypothesis-generating stratification that may inform surgical planning and fertility counseling.

A total of 496 patients were ultimately analyzed, with 372 cases (75.0%) in the unilateral cyst group and 124 cases (25.0%) in the bilateral cyst group. The mean age at surgery in the bilateral cyst group was significantly higher than that in the unilateral group (32.31 ± 5.86 years vs 29.67 ± 4.93 years, P 0.05). See Table 1 . Significant differences were found between the unilateral and bilateral groups in the ultrasound-assessed indicators of pelvic adhesions. The proportion of patients with a positive sliding sign of the posterior uterine wall (indicating fixation of the posterior uterine wall to the bowel) was significantly higher in the bilateral cyst group compared to the unilateral group (82.9% vs 51.5%, P < 0.001). Similarly, the proportion with a positive sliding sign of the pouch of Douglas (indicating obliteration of the pouch of Douglas) was also higher in the bilateral group (68.9% vs 51.5%, P < 0.001). These ultrasonographic signs suggest that patients with bilateral cysts had a more severe state of pelvic adhesions preoperatively. See Table 2 . After adjusting for age, total AFC remained significantly lower in patients with bilateral endometrioma compared with those with unilateral endometrioma (estimated marginal mean 10.77 ± 0.68 vs 14.69 ± 0.39; mean difference 3.92, 95% confidence interval [CI] 2.36–5.47, F = 24.601, P < 0.001; partial η 2 = 0.064; adjusted R 2 = 0.138). Similarly, modified intraoperative EFI/fertility scores were significantly lower in the bilateral group than in the unilateral group (estimated marginal mean 5.95 ± 0.23 vs 7.35 ± 0.13; mean difference 1.40, 95% CI 0.88–1.92, F = 27.864, P < 0.001; partial η 2 = 0.070; adjusted R 2 = 0.106). The assumptions of ANCOVA were satisfied: Levene’s tests indicated homogeneity of error variances (total AFC: F = 0.426, P = 0.514; EFI/fertility score: F = 0.096, P = 0.756), and the laterality-by-age interactions were non-significant for both total AFC (F = 0.008, P = 0.931) and EFI/fertility score (F = 2.798, P = 0.095). There were no significant differences between the two groups in anti-Müllerian hormone (AMH), serum CA125, CA199, and HE4 levels (P > 0.05). See Table 1 . The bilateral group demonstrated significantly worse/more severe disease indicators. Their total rASRM score was significantly higher (84.81 ± 30.94 vs 51.32 ± 26.35, P < 0.001), and the surgical stage was more advanced: the proportion of patients with stage IV disease was as high as 88.6% in the bilateral group, compared to 47.8% in the unilateral group (P < 0.001). The incidence of deep infiltrating nodules was significantly higher in the bilateral group (83.6% vs 65.8%, P < 0.001), and the maximum diameter of these nodules was also larger (1.39 ± 0.85 cm vs 0.99 ± 0.98 cm, P < 0.001). Furthermore, the distribution of the status of the pouch of Douglas differed extremely significantly between the two groups (P < 0.001), with a much higher proportion of patients having complete obliteration of the pouch of Douglas in the bilateral group (61.3%) compared to the unilateral group (41.4%). There were no statistically significant differences between the two groups in terms of the co-existence of endometrial polyps, uterine fibroids, adenomyosis, or ureter/pelvic wall involvement. Ultrasound diagnosis of laterality (unilateral/bilateral) showed high agreement with the actual surgical grouping (Kappa index = 0.793), and the difference was statistically significant (P < 0.001). See Table 1 and Table 2 . Among the 372 patients with unilateral cysts, a comparison of clinical characteristics between patients with left-sided cysts (n=215) and right-sided cysts (n=157) showed no statistically significant differences in all examined indicators, including age, BMI, total rASRM score, fertility score, maximum diameter of deep nodules and ovarian cysts, total AFC, AMH, CA125, CA199, HE4 levels, as well as surgical stage, status of the pouch of Douglas, adenomyosis, endometrial polyps, and uterine fibroids (P > 0.05). See Table 3 and Table 4 . To further quantify the independent impact of cyst laterality (unilateral vs bilateral) on disease severity, multiple linear regression analysis was performed with the total rASRM score as the dependent variable. After controlling for age, BMI, presence/absence of deep infiltrating nodules, and status of the pouch of Douglas, the results showed that bilateral cysts were an independent factor associated with an increase in the total rASRM score (B=22.394, 95% CI: 18.342–26.446, P<0.001), indicating that bilateral cysts per se are significantly associated with a more advanced disease stage. See Table 5 . Among the 227 patients with stage IV endometriosis, the total rASRM score was 77.53±18.67, with a median of 74.00. Accordingly, they were divided into a low-score group (113 cases, 49.78%) and a high-score group (114 cases, 50.22%). Patients in the high-score group were older (32.70±5.99 years vs 30.49±5.69 years, P < 0.01) and had significantly longer operative times (170.11±83.51 minutes vs 136.34±44.12 minutes, P < 0.001). After adjusting for age, the Endometriosis Fertility Index (EFI) score was significantly lower in the high-score group compared to the low-score group (6.01±1.35 vs 7.09±1.32, P < 0.001). There was no statistically significant difference in preoperative Anti-Müllerian Hormone (AMH) levels between the two groups. Regarding pelvic lesion characteristics, the high-score group exhibited a more severe disease state: a higher incidence of bilateral ovarian cysts, a larger average diameter of left ovarian cysts (6.31±3.66 cm vs 4.90±3.44 cm, P < 0.05), and a significantly higher proportion of complete Douglas pouch obliteration. Furthermore, the proportion of patients with AAGL stage 3–4 was also higher in the high-score group (25.44% vs 12.39%, P < 0.001). There were no significant differences in preoperative serum levels of CA125, CA199, and HE4 between the two groups. See Table 6 and Table 7 . Binary logistic regression analysis (forward stepwise method) showed that prolonged operative time (OR = 1.017, 95% CI: 1.008–1.026, P < 0.001) and severe Douglas pouch involvement (OR = 354.941, 95% CI: 24.857–5068.3, P < 0.001) were independently associated with higher odds of being classified into the high-score group (≥74 points). The extremely high OR for severe Douglas pouch involvement should be interpreted cautiously, as it may reflect quasi-complete separation or sparse-cell effects. Conversely, unilateral ovarian disease was associated with lower odds of high-score classification (OR = 0.048, 95% CI: 0.013–0.179, P < 0.001), as was a higher EFI score (OR = 0.558, 95% CI: 0.385–0.811, P = 0.002). See Table 8 . The area under the ROC curve (AUC) for this logistic regression model was 0.896 (95% CI: 0.850–0.943, P < 0.001), indicating excellent discriminatory ability ( Figure 2 ). Agreement analysis revealed fair agreement between the rASRM high-score group and advanced AAGL stage III/IV classification (Kappa = 0.121, P 0.05). Figure 2 Receiver operating characteristic (ROC) curve for predicting severe endometriosis. Receiver operating characteristic curve showing the predictive performance of the nomogram for disease recurrence, with an area under the curve of 0.896. Receiver operating characteristic (ROC) curve for predicting severe endometriosis.

This study was conducted following approval by the Institutional Review Board. Retrospective cohort data were retrieved in December 2025. The primary analysis cohort comprised 496 women aged ≥18 years with regular menstrual cycles, who underwent laparoscopic surgery between March 2022 and December 2025 and were histopathologically diagnosed with ovarian endometriomas. Exclusion criteria included malignancy, endocrine disorders, prior history of chemotherapy/radiotherapy, prior surgical treatment for endometriosis (including cystectomy, oophorectomy, or DIE excision), pregnancy, or autoimmune diseases. For the subgroup analysis, we further screened 227 patients with rASRM stage IV disease from the above cohort of 496 patients, forming a subgroup analysis cohort to explore risk stratification based on the total rASRM score. A flow diagram illustrating patient selection is shown in Figure 1 . Figure 1 Patient flow diagram. A total of 496 patients with histopathologically confirmed endometriomas who underwent laparoscopic surgery were included in the main cohort. Of these, 227 met rASRM stage IV criteria and were included in the subgroup analysis for risk stratification based on the total rASRM score. Flowchart illustrating patient selection and exclusion criteria for the endometrioma study, including 496 patients undergoing laparoscopic surgery with unilateral versus bilateral subgroup analyses. Patient flow diagram. A total of 496 patients with histopathologically confirmed endometriomas who underwent laparoscopic surgery were included in the main cohort. Of these, 227 met rASRM stage IV criteria and were included in the subgroup analysis for risk stratification based on the total rASRM score. Preoperative assessment included clinical and ultrasonographic examinations. Transvaginal ultrasonography (TVS) was performed by specialized gynecologic sonologists using a standardized protocol in accordance with the International Deep Endometriosis Analysis (IDEA) consensus statement. 15 DIE was defined on grayscale ultrasound as hypoechoic nodules or masses infiltrating the rectovaginal septum, bladder wall, or bowel serosa, with a maximum diameter ≥5 mm. Sonologists were aware of the clinical suspicion of endometriosis but were blinded to the subsequent laparoscopic findings and staging scores. Laparoscopic surgery was performed under general anesthesia following standard procedures. For ovarian endometriomas, cystectomy using the stripping technique (blunt and sharp dissection to separate the cyst wall from the ovarian parenchyma) was the preferred approach. When complete cystectomy was technically difficult due to extensive adhesions or risk of ovarian damage, drainage followed by bipolar coagulation of the cyst wall was performed as an alternative. For deeply infiltrating lesions, excision or ablation was performed depending on the location and extent of involvement. All procedures were performed by experienced gynecologic surgeons specializing in endometriosis. Surgical staging was based on the rASRM criteria, with retrospective scoring after surgery. In the subgroup analysis, two senior attending physicians additionally recorded the AAGL stage and the Enzian classification (simplified into limited and extensive types) based on intraoperative findings. Due to missing data for specific variables (eg, preoperative ultrasound not performed in all patients, incomplete AFC counts), the effective sample size varies across tables. The exact number of cases included in each analysis is reported in the corresponding table footnotes. No imputation was performed for missing data; only complete cases were analyzed for each comparison. Multiple comparisons were not adjusted; all P-values should be interpreted as exploratory. The Enzian classification provides a detailed topographic description of deep infiltrating endometriosis (DIE) using compartments A (rectovaginal septum/vagina), B (uterosacral ligaments/parametrium), and C (rectosigmoid/rectum), with additional F modifiers for urinary tract (FU), intestinal (FI), and other extensions. However, the full Enzian coding, although developed to complement rASRM by describing DIE topography, has been noted to have limited international uptake and operational complexity, with the 2011 revision attempting to improve usability. 16 Prior studies have demonstrated that Enzian can be meaningfully collapsed into risk strata—for example, low/intermediate/high surgical-risk groups defined by A/B/C grade cutoffs. 17 To align with the binary design of the present study (unilateral vs bilateral cysts and rASRM high- vs low-score subgroups), we pre-specified a dichotomization of Enzian findings into “limited DIE” and “extensive DIE”. “limited DIE” was defined as Enzian A0–1, B0–1, C0–1 with no F modifiers (FU, FI, FO absent). “Extensive DIE” was defined as any compartment grade ≥2 or the presence of any F modifier. All Enzian assessments were based on intraoperative findings recorded by the operating surgeon. rASRM scores were assigned postoperatively based on the operative records by the attending gynecologic surgeons who performed the procedures. While the laterality of endometriomas was evident in the operative records, the rASRM scoring system primarily depends on objective parameters including lesion size, adhesions, and cul-de-sac involvement, which are independently documented. Nevertheless, we acknowledge that the retrospective, unblinded nature of rASRM scoring represents a potential source of bias, and this limitation is discussed in the Discussion section. For the AAGL endometriosis classification, stage assignment (I–IV) was based on the operative records and intraoperative findings documented by the attending surgeons. The AAGL system incorporates lesion type, location, depth, and associated adhesions to categorize disease into minimal (stage I), mild (stage II), moderate (stage III), or severe (stage IV). As with the rASRM scoring, AAGL staging in this retrospective study was performed by the operating surgeons without independent centralized review. We therefore acknowledge that the absence of blinded AAGL staging represents a similar potential source of bias, which is also noted in the Limitations section. Agreement between the rASRM high/low-score grouping and the dichotomized Enzian classification (limited vs extensive DIE, defined above) and AAGL stage III/IV was assessed using cross-tabulation and Cohen’s kappa test. The Endometriosis Fertility Index was originally developed by Adamson and Pasta (2010) to predict postoperative spontaneous pregnancy rates, incorporating intraoperative rASRM scores, functional assessments of the fallopian tubes, fimbriae, and ovaries, age-related deductions, and postoperative pregnancy outcomes. In the present study, because comprehensive long-term postoperative pregnancy follow-up data were not yet available for all patients, we calculated a modified intraoperative EFI based solely on the intraoperative components—namely, the rASRM total score, functional scores of the fallopian tubes, fimbriae, and ovaries, and age-related deductions (maximum 2 points). This modified intraoperative EFI has been widely used in the literature as a valid surrogate for assessing reproductive potential at the time of surgery. The modified EFI was calculated for all 496 patients to enable consistent fertility-related comparisons between the unilateral and bilateral groups. Statistical analyses were performed using SPSS software (version 21.0). Continuous variables with a normal distribution are presented as mean ± standard deviation (Mean ± SD), and comparisons between groups were made using the independent samples t -test. Continuous variables that did not follow a normal distribution are presented as median (interquartile range), and comparisons between groups were made using non-parametric tests. Categorical variables are presented as number (percentage), and comparisons between groups were made using the chi-square test. The agreement between ultrasound diagnosis and surgical findings was assessed using the Kappa coefficient; a Kappa > 0.75 indicated excellent agreement. To explore the independent impact of cyst laterality (unilateral vs bilateral) on disease severity, multiple linear regression analysis was performed with the total rASRM score as the dependent variable. Independent variables included age, body mass index, presence/absence of deep infiltrating nodules, status of the cul-de-sac, and cyst laterality. In the subgroup analysis targeting stage IV patients, we used the median of the total rASRM score in this subgroup (74 points) as the cutoff to categorize patients into a low-score group (<74 points) and a high-score group (≥74 points). Inter-group differences were compared using the chi-square test and independent samples t -test. To explore the independent factors associated with the high-score group, binary logistic regression analysis (forward stepwise method) was employed, and a receiver operating characteristic (ROC) curve was plotted to evaluate the model’s discriminatory ability. To assess the agreement between the rASRM grouping and the AAGL stage and Enzian classification, cross-tabulation and Cohen’s kappa test were used. For comparisons of total AFC and modified intraoperative EFI/fertility scores between unilateral and bilateral groups, analysis of covariance (ANCOVA) was performed with age as a covariate. Assumptions for ANCOVA were checked, including homogeneity of error variances (Levene’s test) and homogeneity of regression slopes. The laterality-by-age interaction was not significant for total AFC (F = 0.008, P = 0.931) or modified intraoperative EFI/fertility score (F = 2.798, P = 0.095), supporting the use of ANCOVA with age as a covariate. Levene’s tests also indicated no violation of error variance homogeneity (total AFC: F = 0.426, P = 0.514; EFI/fertility score: F = 0.096, P = 0.756). All tests were two-sided, and a P-value <0.05 was considered statistically significant.Given the exploratory nature of this study and the multiple comparisons performed across Tables 1–8 , readers should interpret P values with caution. No formal correction for multiple comparisons was applied; thus, some statistically significant findings may represent type I errors. Table 1 Comparison of Demographics, Cyst Characteristics, Ovarian Reserve, and Biochemical Indicators Between Patients with Unilateral and Bilateral Endometriomas Variable Unilateral Group (N=372) Bilateral Group (N=124) P-value Age (years) 29.67 ± 4.93 32.31 ± 5.86 0.05 Weight (kg) 55.19 ± 13.60 55.35 ± 15.05 >0.05 BMI 21.27 ± 5.17 21.52 ± 6.07 >0.05 rASRM total score 51.32 ± 26.35 84.81 ± 30.94 <0.001 EFI score 7.38 ± 2.29 5.84 ±1.85 <0.001 DIE maximum diameter (cm) 0.99 ± 0.98 1.39 ± 0.85 <0.001 Total AFC 14.90 ± 6.72 10.16 ± 6.18 <0.001 5.99 ± 2.82 Right cyst diameter (cm) 4.32 ± 3.71 0.05 AMH (ng/mL) 3.11 ± 2.21 2.71 ± 1.97 >0.05 CA125 (U/mL) 82.05 ± 138.53 90.27 ± 82.41 >0.05 85.80 ± 255.19 CA199 (U/mL) 140.12 ± 289.01 >0.05 HE4 (U/mL) 33.44 ± 24.95 34.49 ±10.49 >0.05 Notes : Data are presented as mean ± SD or n (%). P values are from independent samples t -test, χ 2 test, or ANCOVA, as specified. Numbers may not sum to the total cohort (n = 496) because of missing data. Missing data were not imputed; only complete cases for the variables in each analysis/table are included. Table 2 Comparison of Surgical, Ultrasonographic, and Comorbidity Findings Between Patients with Unilateral and Bilateral Endometriomas Variable Group N(%) P-value rASRM stage IV, n (%) Unilateral (N=368) 176 (47.8%) <0.001 Bilateral (N=123) 109 (88.6%) Deep infiltrating nodule, n (%) Unilateral (N=365) 240 (65.8%) <0.001 Bilateral (N=122) 102 (83.6%) Complete cul-de-sac obliteration, n (%) Unilateral (N=365) 151 (41.4%) 0.05 Bilateral (N=124) 36 (29.0%) Endometrial polyp, n (%) Unilateral (N=369) 64 (17.3%) >0.05 Bilateral (N=123) 22 (17.9%) Uterine fibroid, n (%) Unilateral (N=369) 68 (18.4%) >0.05 Bilateral (N=123) 29 (23.6%) Sliding sign of posterior uterine wall, n (%) Unilateral (N=361) 186 (51.5%) <0.001 Bilateral (N=123) 102 (82.9%) Sliding sign of pouch of Douglas, n (%) Unilateral (N=361) 186 (51.5%) <0.001 Bilateral (N=122) 84 (68.9%) Notes : Data are presented as mean ± SD or n (%). P values are from independent samples t -test, χ 2 test, or ANCOVA, as specified. Numbers may not sum to the total cohort (n = 496) because of missing data. Missing data were not imputed; only complete cases for the variables in each analysis/table are included. Table 3 Comparison of Demographics, Cyst Characteristics, Ovarian Reserve, and Biochemical Indicators Between Patients with Left-Sided and Right-Sided Endometriomas within the Unilateral Group Variable Left-Sided Group (N=215) Right-Sided Group (N=157) P-value Age (years) 29.56 ± 4.81 29.82 ± 5.10 >0.05 Height (cm) 160.20 ± 12.23 161.20 ± 5.40 >0.05 Weight (kg) 54.58 ± 14.31 56.04 ± 12.57 >0.05 BMI 21.03 ± 5.29 21.57 ± 5.46 >0.05 rASRM total score 49.65 ± 24.89 53.58 ± 28.13 >0.05 EFI score 7.39 ± 2.56 7.37 ± 1.88 >0.05 Nodule maximum diameter (cm) 0.95 ± 0.97 1.03 ± 1.00 >0.05 Total AFC 14.53 ± 6.44 15.47 ± 7.25 >0.05 Right cyst diameter (cm) 0 6.74 ± 2.25 >0.05 Left cyst diameter (cm) 6.63 ± 1.84 0 AMH (ng/mL) 3.03 ± 2.07 3.21 ± 2.41 >0.05 CA125 (U/mL) 78.82 ± 104.18 86.70 ± 177.08 >0.05 CA199 (U/mL) 73.18 ± 166.21 104.50 ± 347.42 >0.05 HE4 34.66 ± 31.16 33.49 ± 10.62 >0.05 Notes : Data are presented as mean ± SD or n (%). P values are from independent samples t -test, χ 2 test, or ANCOVA, as specified. Numbers may not sum to the total cohort (n = 496) because of missing data. Missing data were not imputed; only complete cases for the variables in each analysis/table are included. Table 4 Comparison of Surgical Stage and Comorbidity Findings Between Patients with Left-Sided and Right-Sided Endometriomas within the Unilateral Group Variable Group N(%) P-value rASRM stage IV Left-Sided (N=212) 94 (44.3%) 0.159 Right-Sided (N=156) 82 (52.6%) Deep infiltrating nodule Left-Sided (N=211) 138 (65.4%) 0.869 Right-Sided (N=154) 102 (66.2%) Complete cul-de-sac obliteration Left-Sided (N=209) 79 (37.8%) 0.250 Right-Sided (N=156) 72 (46.2%) Adenomyosis Left-Sided (N=214) 46 (21.5%) 0.141 Right-Sided (N=155) 37 (23.9%) Endometrial polyp Left-Sided (N=214) 32 (15.0%) 0.154 Right-Sided (N=155) 32 (20.6%) Uterine fibroid Left-Sided (N=214) 36 (16.8%) 0.350 Right-Sided (N=155) 32 (20.6%) Notes : Data are presented as mean ± SD or n (%). P values are from independent samples t -test, χ 2 test, or ANCOVA, as specified. Numbers may not sum to the total cohort (n = 496) because of missing data. Missing data were not imputed; only complete cases for the variables in each analysis/table are included. Table 5 Results of Multiple Linear Regression Analysis with Total rASRM Score as the Dependent Variable Variable B S.E. Beta t Sig. Age 0.291 0.167 0.050 1.743 >0.05 BMI 0.00006 0.000 0.042 1.531 >0.05 Unilateral vs Bilateral 22.394 0.061 0.312 10.867 0.05 Cul-de-sac Status 26.860 1.286 0.655 20.890 <0.001 Table 6 Comparison of Clinical Characteristics Between <74 Group and ≥74 Group Variable Group N Mean Std. Deviation P Age (years) <74 113 30.48 5.69 <0.01 ≥74 114 32.70 5.99 Surgical duration (min) <74 113 136.34 44.11 <0.001 ≥74 114 170.11 83.50 EFI <74 80 7.03 1.31 <0.001 ≥74 80 6.01 1.35 Height (cm) 0.05 ≥74 111 159.75 5.26 Weight (kg) 0.05 ≥74 111 52.99 7.83 BMI 0.05 ≥74 111 20.80 3.21 Diameter of infiltrating nodule (cm) 0.05 ≥74 107 1.41 0.92 Diameter of the left ovarian cyst (cm) <74 100 4.90 3.43 <0.05 ≥74 99 6.31 3.65 Diameter of the right ovarian cyst (cm) 0.05 ≥74 104 5.49 3.42 AMH (ng/mL) 0.05 ≥74 64 2.42 1.74 CA125 (U/mL) 0.05 ≥74 96 98.94 79.59 CA199 (U/mL) 0.05 ≥74 96 88.14 145.09 HE4 (U/mL) 0.05 ≥74 96 33.30 9.35 Table 7 Distribution of Clinical and Pathological Variables in the rASRM Low-Score (<74 Points) and High-Score (≥74 Points) Groups Variable Classification <74 Group ≥74 Group P-value Laterality of ovarian cysts Unilateral 84 51 0.05 1 74 81 2 3 9 Number of right ovarian cysts 1 48 27 <0.05 2 59 75 3 6 12 Douglas’ pouch Normal 8 4 0.05 No 82 87 AAGL score I–II 99 85 0.05 Extensive 58 64 Table 8 Results of Binary Logistic Regression Analysis Variable B S.E. Wald p Exp (B) 95.0% C.I. for Lower Upper Operative time 0.017 0.005 12.966 0.000 1.017 1.008 1.026 Douglas’ pouch 5.872 1.357 18.736 0.000 354.941 24.857 5068 Unilateral −3.035 0.670 20.539 0.000 0.048 0.013 0.179 Infiltrative nodule 0.973 0.539 3.257 0.071 2.647 0.920 7.618 EFI −0.583 0.190 9.387 0.002 0.558 0.385 0.811 Comparison of Demographics, Cyst Characteristics, Ovarian Reserve, and Biochemical Indicators Between Patients with Unilateral and Bilateral Endometriomas Notes : Data are presented as mean ± SD or n (%). P values are from independent samples t -test, χ 2 test, or ANCOVA, as specified. Numbers may not sum to the total cohort (n = 496) because of missing data. Missing data were not imputed; only complete cases for the variables in each analysis/table are included. Comparison of Surgical, Ultrasonographic, and Comorbidity Findings Between Patients with Unilateral and Bilateral Endometriomas Notes : Data are presented as mean ± SD or n (%). P values are from independent samples t -test, χ 2 test, or ANCOVA, as specified. Numbers may not sum to the total cohort (n = 496) because of missing data. Missing data were not imputed; only complete cases for the variables in each analysis/table are included. Comparison of Demographics, Cyst Characteristics, Ovarian Reserve, and Biochemical Indicators Between Patients with Left-Sided and Right-Sided Endometriomas within the Unilateral Group Notes : Data are presented as mean ± SD or n (%). P values are from independent samples t -test, χ 2 test, or ANCOVA, as specified. Numbers may not sum to the total cohort (n = 496) because of missing data. Missing data were not imputed; only complete cases for the variables in each analysis/table are included. Comparison of Surgical Stage and Comorbidity Findings Between Patients with Left-Sided and Right-Sided Endometriomas within the Unilateral Group Notes : Data are presented as mean ± SD or n (%). P values are from independent samples t -test, χ 2 test, or ANCOVA, as specified. Numbers may not sum to the total cohort (n = 496) because of missing data. Missing data were not imputed; only complete cases for the variables in each analysis/table are included. Results of Multiple Linear Regression Analysis with Total rASRM Score as the Dependent Variable Comparison of Clinical Characteristics Between <74 Group and ≥74 Group Distribution of Clinical and Pathological Variables in the rASRM Low-Score (<74 Points) and High-Score (≥74 Points) Groups Results of Binary Logistic Regression Analysis

In this retrospective cohort, bilateral endometriomas were associated with higher rASRM scores and lower AFC as a marker of ovarian reserve after adjusting for age; however, preoperative AMH levels did not differ significantly between unilateral and bilateral groups. This association may be partly influenced by the structural overlap between laterality and rASRM scoring criteria, and the findings should be interpreted in light of the retrospective, unblinded nature of disease severity assessment. The rASRM total score ≥74 cutoff—derived as the median score of this specific cohort—identified a subgroup with distinct clinical characteristics, including longer operative time and more severe pelvic adhesions. Nevertheless, given the weak agreement with AAGL stage III/IV and negligible agreement with the simplified Enzian classification, this threshold should be considered exploratory and hypothesis-generating rather than a validated addition to existing staging systems. External validation in independent, multicenter cohorts is required before any clinical application.

This comparative study of 496 patients with endometriomas revealed significant clinicopathological differences between unilateral and bilateral cysts. Our core finding is that bilateral cysts are closely associated with a series of indicators suggesting greater disease severity, which also manifested as more significant signs of pelvic adhesions on preoperative ultrasound. The positivity rates for the sliding sign of the posterior uterine wall and the sliding sign of the pouch of Douglas were significantly higher in the bilateral cyst group, corroborating the intraoperative findings of pelvic adhesions. 18–20 This suggests that ultrasound sliding signs can serve as a useful, non-invasive preoperative indicator for assessing the severity of pelvic endometriosis and predicting the possibility of bilateral cysts. The high agreement between ultrasound diagnosis of cyst laterality and actual surgical results (Kappa = 0.793) indicates the important clinical value of ultrasound examination in comprehensively evaluating the severity and extent of endometriomas. In this study, patients with bilateral cysts were older at presentation, which aligns with the hypothesis that this type of disease progresses cumulatively over time. More importantly, the bilateral group demonstrated superior performance across all objective indicators reflecting disease severity: significantly higher rASRM scores and proportion of stage IV disease, significantly increased incidence and size of deep infiltrating nodules, and a higher rate of complete obliteration of the pouch of Douglas. These findings strongly support the view that bilateral ovarian endometriomas represent a more advanced and widespread manifestation of pelvic endometriosis. 21 The significantly lower total AFC and lower EFI score in the bilateral group suggest more severe impairment of ovarian reserve, consistent with the speculation that involvement of both ovaries may cause greater cumulative damage to the ovarian cortex and blood supply. 9 , 10 In our study cohort, there were no statistically significant differences between the two groups in BMI, serum CA125, AMH levels, or the incidence of comorbidities such as adenomyosis and endometrial polyps. This differs from the conclusions of some earlier studies, 22 , 23 suggesting that these indicators may not be the key factors in distinguishing the disease behavior of unilateral vs bilateral cysts. Notably, our in-depth analysis of 227 stage IV patients further reveals that disease burden exhibits significant heterogeneity even within the same rASRM stage IV. The high-risk subgroup (≥74 points), delineated by the median total rASRM score of 74, is characterized by greater surgical complexity (longer operative time), a more extensive range of pelvic lesions (higher prevalence of bilateral cysts and severe involvement of the pouch of Douglas), and lower fertility potential scores (EFI). Logistic regression analysis strongly suggests that bilateral cysts and severe involvement of the pouch of Douglas are the core anatomical features constituting this high-risk status. While the agreement between this stratification and the AAGL and Enzian staging systems—which emphasize the specific location/depth of lesions—is limited, it precisely demonstrates that an rASRM total score ≥74 provides an additional informational dimension, independent of detailed anatomical localization, concerning the “overall disease burden” and “anticipated surgical difficulty”. This aligns with and reinforces the conclusion from our primary analysis that “bilateral cysts represent a more severe phenotype”, and offers a potential quantitative tool for clinical preoperative planning and prognostic stratification. The strengths of this study include its relatively large sample size and standardized surgical assessment at a single center. However, its retrospective design carries inherent limitations. Several limitations of this study should be acknowledged. First, this was a single-center retrospective cohort study, which inherently carries risks of selection bias and limits the generalizability of our findings. Second, rASRM scores were assigned postoperatively based on the operative records by the attending surgeons. Although the rASRM staging system relies on objective anatomical parameters—including lesion size, extent of adhesions, and cul-de-sac involvement—the laterality of endometriomas was evident in the surgical documentation. We cannot exclude the possibility of unconscious bias in severity assessment, and the inherent circularity between laterality and rASRM total scores may have contributed to the observed association between bilateral disease and higher rASRM scores. Future prospective studies with independent, blinded staging assessment are warranted to validate these findings. Third, as with the rASRM scoring, AAGL staging and Enzian classification were both assigned retrospectively by the operating surgeons without independent or blinded review, which may introduce observer bias. Fourth, the modified intraoperative Endometriosis Fertility Index (EFI) used in this study excluded postoperative pregnancy outcomes due to the lack of comprehensive long-term follow-up data for all patients. While the intraoperative EFI components alone are widely accepted as a valid surrogate for assessing intraoperative reproductive potential, the absence of pregnancy data limits the ability to correlate EFI scores with actual fertility outcomes. Fifth, the identification of an rASRM total score ≥74 as a cutoff for high-risk stratification was data-driven, based on the median score of this specific cohort. This threshold should be regarded as exploratory and hypothesis-generating, and requires external validation in independent, multicenter cohorts before any clinical application. Sixth, some tables exhibited inconsistencies in sample sizes across variables due to missing data; no imputation was performed, and complete-case analysis may have introduced attrition bias. Seventh, given the exploratory nature of this study and the multiple comparisons performed across Tables 1–8 , some statistically significant findings may represent type I errors; P values should be interpreted with caution, and no formal correction for multiple comparisons was applied. Finally, the binary logistic regression model for high-risk stratification produced an exceptionally large odds ratio for Douglas’ pouch involvement (OR = 354.941), which likely reflects quasi-complete separation—a statistical artifact rather than a clinically meaningful effect size—and should be interpreted with caution.