This study was designed as a retrospective, descriptive, and analytical research. The clinical records of patients who underwent hysterectomy due to AUB at Bandirma Onyedi Eylül University Hospital between 2020 and 2025 and who had not received a pre-operative diagnosis of gynaecological malignancy were examined. Clinical information and pathology reports (EB and hysterectomy) were retrospectively evaluated. After applying inclusion and exclusion criteria, a total of 296 patients were included in the study (Figure 1).

Pre-operative endometrial sampling was performed based on a standardized clinical protocol. The Pipelle cannula was used as the primary first-line method in the office setting for patients with suitable cervical patency and tolerance. Fractionated dilatation and curettage (D&C) under sedation was reserved for cases where office biopsy failed, cervical stenosis prevented catheter insertion, or the initial office sampling yielded insufficient tissue for diagnosis. All procedures were performed blindly without hysteroscopy guidance. The samples obtained were fixed in 10% neutral formalin and sent to the pathology laboratory, where they were reported according to World Health Organisation (WHO) criteria. Regarding data handling, cases with ‘insufficient material' (n = 23) were excluded from diagnostic accuracy calculations (sensitivity and specificity) to ensure statistical reliability.

The study data were collected using a patient information form prepared in Excel format. This form recorded age, EB result, indication for hysterectomy, final hysterectomy pathology result, treatments administered for AUB, gynaecological examination and radiology findings, and surgical method information obtained from the hospital information management system. Patient identities were kept confidential and data were recorded using anonymous patient codes.

The data were analysed using the SPSS 23.0 statistical package. The distribution of demographic and clinical characteristics of the study group, the distribution of AUB and PMB according to the PALM–COEIN classification, and the distribution of histopathological findings in hysterectomy specimens according to anatomical region were summarised in terms of numbers and percentages and presented in graphs. In the evaluation of the diagnostic performance of pre-operative endometrial biopsy, hysterectomy pathology was accepted as the gold standard, and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy (ACC) were calculated. ROC curves were constructed to evaluate the diagnostic discriminatory power of pre-operative endometrial biopsy, the area under the curve (AUC) was calculated for each pathology category, and AUC values were interpreted as moderate (0.70–0.80), good (0.80–0.90), and excellent (≥0.90) diagnostic accuracy (16). The agreement between pre-operative endometrial biopsy and hysterectomy pathology results was assessed using Cohen's kappa coefficient. Kappa values were categorised to describe the level of agreement as follows: 0–0.10, virtually no agreement; 0.11–0.40, slight agreement; 0.41–0.60, fair agreement; 0.61–0.80, moderate agreement; and 0.81–1.00, substantial agreement (17). A significance level of P < 0.05 was accepted for statistical tests.

The study was conducted in accordance with the ethical principles outlined in the Helsinki Declaration. Ethical approval for the study was obtained from Bandirma Onyedi Eylul University Health Sciences Non-Interventional Research Ethics Committee (Date: 01.12.2025, Decision No: 2025-209). Due to the retrospective design of the study, the requirement for written informed consent was waived by the Ethics Committee in accordance with national legislation and institutional requirements.

The demographic structure of our study largely reflects the population undergoing hysterectomy due to AUB. Consistent with the literature, 69.9% of patients (n = 207) were in the perimenopausal period and concentrated in the 41–50 age range (21). AUB in this age group is generally associated with hormonal fluctuations and prolonged exposure to oestrogen. Therefore, particular attention should be paid to endometrial hyperplasia and endometrial cancer in this age group (22). The low prevalence among women under 40 years of age (6.4%) indicates the clinical priority of fertility-preserving treatment in young patients (5).

Post-menopausal patients have been classified as a separate category from the PALM-COEIN classification due to their different aetiological spectrum, high likelihood of endometrial carcinoma, and being outside the reproductive age. Over 90% of malignancies seen in this patient group present with AUB (23).

In our study, the most frequently preferred surgical method was total abdominal hysterectomy (TAH) at 64.9%. This was followed by total laparoscopic hysterectomy (TLH) at 34.5% and vaginal hysterectomy (VH) at 0.7%.

Although the literature emphasises that the vaginal approach (VA) should be preferred for benign hysterectomies due to its minimally invasive nature, cost-effectiveness and rapid recovery process (24, 25), in our study, the VH ratio was limited to 0.7%. One of the main reasons for this low rate is that patients who underwent surgery solely for uterine prolapse were excluded from our study design. Furthermore, the detection of leiomyoma in 75.3% of the final pathology reports in our series suggests that the size and volume of the uterus in this patient group may have made vaginal extraction technically challenging. Being a tertiary centre, the frequency of referred complex cases with a high risk of adhesions and large uterine volumes has shifted the surgical preference towards abdominal or laparoscopic approaches rather than the vaginal route.

Our study found that 66.2% of patients underwent surgery directly without receiving any medical treatment for AUB. This suggests that urgent hysterectomy may be necessary due to severe symptoms, structural pathologies, or contraindications to medical treatment (26). The most commonly administered treatments among patients receiving medical treatment were oral progesterone (20.3%), combined oral contraceptives (4.7%) and levonorgestrel-releasing intrauterine system (LNG-IUS; 4.4%). It is known that these treatment methods eliminate the need for hysterectomy in some women (27–29). However, the fact that hysterectomy is ultimately required despite these treatments indicates that medical treatment has limited long-term efficacy in a significant proportion of women with AUB, particularly those with underlying structural pathology (30, 31).

According to the FIGO PALM-COEIN classification system, surgery is significantly more frequently performed in cases with structural aetiologies (PALM). Hysterectomy remains an important option in the treatment of cases with identifiable structural abnormalities (3, 32).

In our study, leiomyoma (AUB-L) was the most common indication for hysterectomy, accounting for 35.5%. Leiomyoma could be isolated or observed in conjunction with both adenomyosis (AUB-L, A) and polyps (AUB-P, L). When all categories containing leiomyoma are included, this rate reaches 60%. Similar results have been reported in many studies in the literature (33–35). These findings emphasise the importance of leiomyoma aetiology in the pathogenesis of severe AUB requiring surgical intervention (31). In our study, AUB-A (3.4%) and AUB-P (3%) were observed as other common indications for isolated hysterectomy after AUB-L. Although prevalence rates vary in studies conducted in different regions, adenomyosis and endometrial polyps play a significant role in the aetiology of hysterectomy (35, 36).

Non-structural causes (COEIN) generally account for a small proportion of hysterectomy cases associated with AUB (8). In our study, the rate of hysterectomies performed for non-structural reasons was found to be only 4.4% (n = 13). The clear pre-dominance of structural aetiologies supports the view that hysterectomy should be used in the treatment of patients with complex and identifiable structural causes of AUB, rather than purely for functional disorders (37, 38).

The histopathological examination of hysterectomy specimens in the studies conducted also confirms the pre-dominance of structural lesions (39, 40). In our study, leiomyoma was detected in 75.3% of hysterectomy pathologies and adenomyosis in 26%. In the literature, the prevalence of leiomyoma ranges from 22 to 80%, depending on age, hysterectomy indication, and ethnic origin (41–43), The prevalence of adenomyosis is reported to be between 8 and 61.5% (44, 45). These two pathologies constitute the majority of uterine findings in hysterectomy specimens.

When evaluating the final pathology results in our study, functional or benign histological patterns were observed in a significant portion of the endometrium (82.3%; n =273), while chronic cervicitis was the most common cervical diagnosis (71.3%; n = 211). Incidental cervical lesions detected in the final histopathological examination demonstrate that the cervix should never be neglected during the pre-operative evaluation process. The detection of high-grade intraepithelial lesions (CIN II-III) in 2.4% (n = 7) and invasive cancer in 0.3% (n = 2) of the cervix in our study demonstrates that the risk of occult malignancy or pre-malignancy cannot be disregarded even in AUB cases operated on for benign indications. These findings are consistent with the rates of occult cervical cancer reported in the large-scale studies by Desai et al (46). In light of the data we have obtained, pre-operative cervical smear screening should be considered not merely as a ‘recommended' test, but as a mandatory step in AUB management that can directly influence the scope of the surgical procedure (whether the cervix will be preserved or not).

Similarly, in the histopathological evaluation of ovaries, non-benign pathologies were incidentally detected in 0.6% of cases (n = 2). Desai et al. reported the rate of occult ovarian cancer as 0.19% (46). Therefore, it is critically important to perform advanced pre-operative investigations for ovarian pathologies in suspected patients (47). Our study found benign pathologies in all salpiks, and similar results have been reported in the literature (48–50). Our data indicate that AUB requiring hysterectomy is generally caused by multifactorial reasons involving uterine, myometrial, cervical and adnexal pathology rather than isolated endometrial disease.

The comparison between pre-operative EB findings and hysterectomy pathology is of critical importance for evaluating diagnostic accuracy and constitutes one of the primary focal points of clinical research. The accuracy of EB diagnosis largely depends on the pathological threshold value used to define the disease and demonstrates a classic balance between sensitivity and specificity (51).

When the diagnostic threshold is broadened to include atypical hyperplasia, a moderate increase in sensitivity is generally observed, allowing for the detection of a wider range of endometrial changes. However, this trade-off inherently leads to a decrease in specificity and often results in a low positive predictive value (PPV), thereby increasing the false-positive rate (52).

Our study generally yielded results consistent with the literature. When the threshold value for pre-operative EB was set at AEH+, specificity reached 99.6%, while sensitivity was found to be 22.2%. When the threshold value was set at Non-AEH and above, sensitivity increased to 44.4%, while specificity decreased to 93.9%.

The low sensitivity rate of 22.2% observed for AEH and above lesions identified in our study indicates that the risk of ‘false negatives' in pre-operative EB is a factor that must be considered in clinical practice. The fact that the pre-operative biopsy results were reported as benign in two of the three cases in which endometrial carcinoma was detected in the final pathology supports the limitation of blind sampling methods, particularly in detecting focal malignant or pre-malignant lesions.

The diagnostic performance of pre-operative endometrial biopsy (EB) subgroups in predicting final hysterectomy endometrial pathology was assessed using ROC analysis (Table 2, Figure 3). Overall, the discriminatory ability of EB was limited across all subgroups, with AUC values below 0.700. In particular, for endometrial polyps and endometritis, the AUC estimates had 95% confidence intervals that included 0.5, indicating that discrimination beyond chance could not be demonstrated in this cohort; together with low sensitivity, these findings limit the utility of EB as a standalone diagnostic tool for these conditions. Although the AUC for AEH+ (atypical endometrial hyperplasia and above) was comparatively higher, its wide confidence interval and the low sensitivity observed for detecting AEH+ support cautious clinical interpretation. In addition, ROC curves exhibited diagonal segments, consistent with overlapping test results between outcome-positive and outcome-negative groups and reflecting limited separability. Collectively, these results suggest that pre-operative EB has a restricted role in predicting final hysterectomy pathology when used in isolation, and should be interpreted in conjunction with imaging findings, clinical characteristics, and complementary diagnostic methods during clinical decision-making.

Another noteworthy finding of our study concerns PMK cases. None of the 83 patients who underwent EB due to PMB and whose results were reported as benign or inadequate showed malignancy in the final pathology after hysterectomy. This situation supports the notion that the negative predictive value of EB in ruling out malignancy is quite high in post-menopausal patients and that a benign biopsy result in these cases provides reliable information to the clinician.

In light of these findings, a benign biopsy result may not be sufficient on its own to rule out malignancy, particularly in symptomatic patients with risk factors (obesity, advanced age, diabetes mellitus, unopposed oestrogen exposure, etc.). Therefore, even if the histopathological result is benign, in cases where AUB persists or clinical suspicion is strong, it would be a safer clinical approach not to terminate the diagnostic process and to include hysteroscopy or advanced imaging methods that allow direct evaluation of the uterine cavity.

The limited sensitivity of EB is also observed in the diagnosis of focal intrauterine lesions such as endometrial polyps and chronic endometritis. All blind endometrial sampling methods are generally unreliable in the diagnosis of endometrial polyps and have been shown to miss most focal lesions (12, 53). It has been reported that sensitivity is below 30% for endometrial polyps in particular and that there is inconsistency with the final histopathology (54, 55).

In our study, EB was found to have a sensitivity of 30.6% and a specificity of 83% in the pre-operative diagnosis of polyps. These rates indicate that EB has a very limited role in the diagnosis of endometrial polyps. These low rates can be explained by the fact that the lesions cannot be detected by blind techniques due to their focal distribution. Consequently, cases of endometrial carcinoma developing on the basis of endometrial polyps may be missed. Unlike blind techniques, hysteroscopy is currently considered the gold standard for detecting focal intrauterine pathology and obtaining a reliable histological diagnosis of endometrial polyps due to its ability to directly visualise the lesion (12, 53, 56).

In our study, the reliability of EB in the diagnosis of chronic endometritis was found to be similar to that of endometrial polyps. Although its specificity was 94.7%, higher than that of endometrial polyps, its sensitivity of 30% alone was found to be insufficient for the diagnosis of endometritis.

The finding that patients diagnosed with functional endometrium by EB ultimately harbour abnormalities such as polyps, hyperplasia, endometritis and even endometrial carcinoma in the final pathology results highlights the significant clinical consequences of false negative EB results (57).

In our study, endometrial carcinoma was detected in the final pathology in two cases that were similarly reported as benign by EB. Furthermore, in three cases, the pre-operative diagnosis was upgraded to AEH+, and in nine cases, it was upgraded to Non-AEH. These results confirm that blind endometrial sampling methods may miss concomitant malignancies, particularly in focal lesions (20, 57). Particular care should be taken in the management of patients who are resistant to medical treatment, carry risk factors, and have EB results that are inconsistent with clinical findings (58). If clinical suspicion persists despite negative EB results, it is critically important to perform advanced investigations such as transvaginal ultrasonography (TV-USG), saline infusion sonohysterography (SIS), hysteroscopy, CT, and MRI (59). Consequently, it would be more appropriate to use EB as an exclusion test rather than as a definitive diagnostic tool (60).