Abstract
Objective
To assess the concordance between pre‐ and postoperative pathology in patients with endometrial intraepithelial neoplasia (EIN) undergoing hysterectomy, and identify predictors of malignancy, including sampling method, progestin therapy, and surgical timing.
Methods
A retrospective cohort study was conducted at a tertiary care academic center in Montreal, Canada, from January 2010 to July 2023. Patients undergoing hysterectomy for EIN were included. Demographics, progestin therapy, sampling method, time from biopsy to surgery, operative details, and final pathology were collected. Logistic regression was used to identify factors associated with malignant or benign outcomes.
Results
A total of 319 patients were included; 76 received preoperative progestins (medicated group), and 243 did not (unmedicated group). Pathology concordance was 71.2% overall (72.8% unmedicated, 65.8% medicated). Malignancy was found in 11.6% (12.8% unmedicated, 7.9% medicated); benign pathology in 17.2% (14.4% unmedicated, 26.3% medicated). Sampling method was not associated with concordance.
In progestin‐naïve patients, lower body mass index (BMI) (odds ratio [OR] 2.91, 95% confidence interval [CI]: 1.18–7.13) and premenopausal status (OR 2.20, 95% CI: 1.06–4.58) were associated with benign pathology. In progestin‐treated patients, older age (OR 1.10, 95% CI: 1.01–1.22) and menopausal hormone therapy use (OR 16.85, 95% CI: 1.85–218.46) predicted malignancy, while longer time to surgery was associated with benign pathology (OR 2.50, 95% CI: 1.04–6.41).
Conclusion
Concordance between preoperative EIN diagnosis and final hysterectomy pathology was 65.8%–72.8%, depending on progestin therapy. These findings highlight the importance of personalized management strategies in patients with EIN, with attention to patient preference, their risk factors, and treatment response.
Keywords: endometrial cancer, endometrial intraepithelial neoplasia, endometrial sampling, pathology concordance, progestin therapy
1. INTRODUCTION
Endometrial intraepithelial neoplasia (EIN), formerly referred to as complex atypical hyperplasia, is a premalignant lesion of the endometrium and a recognized precursor to endometrioid‐type endometrial adenocarcinoma. 1 The pathogenesis of endometrial adenocarcinoma is closely associated with prolonged unopposed estrogen exposure, which drives endometrial hyperplasia and may progress to EIN before developing into endometrial adenocarcinoma. 2 This risk is particularly heightened in the setting of common patient‐specific factors such as obesity, advanced age, nulliparity, and unopposed estrogen therapy, which further emphasize the need for early detection and timely management of EIN. 1 , 3 , 4
Reported risks of endometrial cancer following a diagnosis of EIN vary widely, with untreated EIN associated with a 30%–50% 5 risk of progression to adenocarcinoma and concurrent malignancy identified in up to 47% of hysterectomy specimens, 6 , 7 , 8 while other studies report substantially lower rates with stability or regression, 9 likely due to differences in sampling, timing of surgery, treatment exposure, and patient‐specific risk factors.
Given the heterogeneity of the findings, stratified evaluation by preoperative progestin therapy status is essential to address whether progestin‐mediated histologic changes and patient selection bias meaningfully influence diagnostic congruence and malignancy detection. Therefore, the objective of this current study was to evaluate, in our tertiary care center among patients with EIN, the institutional concordance rates between preoperative endometrial sampling and final pathology following hysterectomy, with an emphasis on sampling technique, interval to surgery, and adjuvant progestin therapy.
2. MATERIALS AND METHODS
2.1. Study design
A retrospective cohort study was conducted between January 1, 2010, and July 31, 2023 in the Department of Gynecology at the Royal Victoria Hospital of the McGill University Health Center (MUHC), a tertiary care center in Montreal, Quebec, Canada. A comprehensive chart review was performed encompassing all patients undergoing hysterectomy for suspected EIN on preoperative pathology.
The search was carried out using the electronic operative data management system linked to the MUHC (Opera, GE HealthCare 2023). Relevant charts were identified using institution‐specific preoperative diagnostic codes for the following: EIN/endometrial hyperplasia with atypia, suspected endometrial cancer, or specific endometrial changes (hyperplasia, thickness, lesion, neoplasia, polyp, mass, bleeding). Electronic medical charts were then individually screened for inclusion. The primary endpoint of this study was concordance between the preoperative diagnosis of EIN and final hysterectomy pathology.
We included patients older than 18 years of age, and with at least one preoperative biopsy report confirming the diagnosis of EIN, sampled either by endometrial biopsy in office or hysteroscopy/dilatation and curettage (D&C). In the case that a pathology specimen was not native to the MUHC, the biopsy specimen was re‐read by a pathologist within our center. If multiple biopsies confirmed EIN, the first recorded positive biopsy date was used. Additionally, patients were included if their preoperative pathology report resulted in questionable diagnoses including at least EIN, cannot rule out carcinoma, suspicious for carcinoma, EIN with foci approaching adenocarcinoma, suspicious of EIN.
Patient demographics and medical history (age, body mass index [BMI, calculated as weight in kilograms divided by the square of height in meters], parity, menopausal status, previous use of tamoxifen or menopausal hormone therapy [MHT], and relevant comorbidities including respiratory disease, diabetes mellitus type 2, hypertension, and coronary artery disease) were identified. Additionally, preoperative data including systemic progestin therapy or a levonorgestrel‐releasing intrauterine device (LNG‐IUD), sampling method either by office‐based endometrial biopsy or hysteroscopy/D&C, time from endometrial biopsy to hysterectomy were included. Operative data including admission type, date of surgery, hysterectomy characteristics, total duration of procedure, complications (blood transfusion, conversion to open, postoperative infection, return to emergency department <30 days, bladder injury, other) and length of stay were recorded. Additional procedures at time of hysterectomy, including sentinel lymph node mapping, appendectomy, omental biopsies, were included. It is important to note, however, that these procedures were undertaken at the discretion of the treating physician. All final pathology diagnoses along with the pathologist who analyzed the specimen were obtained from the patients' medical records. Patients were categorized based on one of three final pathology groups (benign disease, persistent EIN, malignancy).
Patients with a preoperative diagnosis of hyperplasia with no overt atypia, non‐atypical hyperplasia, carcinoma, or with no evidence of EIN were excluded. Moreover, patients that had no recorded preoperative endometrial biopsy results were excluded. In the event that a patient had a remote history of EIN with at least two subsequent negative biopsies, the patient was excluded from our study. Lastly, patients were excluded for coding errors (e.g., uterus not removed at time of surgery) and security overrides within the charting system (e.g., patient deceased at time of chart review with further investigation underway). Intraoperative frozen‐section examination was not routinely performed and was requested at the discretion of the treating physician. Given the retrospective nature of this study, the precise clinical rationale for selection of surgical versus medical management could not be consistently determined. Treatment decisions were made at the discretion of the treating physician and patient, without a standardized institutional protocol.
2.2. Statistical and data analysis
A Kruskal‐Wallis nonparametric test, Fisher exact test or Chi‐square test was used for continuous or categorical variables as appropriate to compare demographic, clinical, and operative data among the three final pathology groups, A two‐sided P value less than 0.05 was considered as statistically significant. To search for significant predictive factors related to the final pathology, univariate logistical regression analyses were performed using variables selected for their established association with EIN. Odds ratios (OR) and 95% confidence intervals (CIs) were calculated for each covariate. The R2 statistic was calculated for each model for goodness‐of‐fit. Natural logarithm transformation was applied for normalizing the distribution for time from biopsy to hysterectomy, after testing the normality. All analyses were conducted using Statistical Analysis Software (SAS version 9.4, Cary, NC) and R‐package (version 4.3.0).
2.3. Ethics considerations
This project was reviewed by the McGill University Health Center (MUHC) Research Ethics Board and was determined to meet the definition of Quality Assurance/Quality Improvement under Article 2.5 of the Tri‐Council Policy Statement (TCPS2, 2022). As such, formal REB approval and individual informed consent were not required.
3. RESULTS
During the study period, 440 patients were identified using the institutional codes listed above undergoing hysterectomy for a preoperative diagnosis of EIN. After individual chart review and application of inclusion/exclusion criteria, 319 patients remained within our cohort (Figure 1).
FIGURE 1.
Flow chart illustrating the construction of the study cohort. EIN, endometrial intraepithelial neoplasia; EMB, endometrial biopsy; MUHC, McGill University Health Center.
The average age of patients was 57.9 ± 9.9 years, with a mean BMI of 32.5 ± 7.9. Mean parity was 1.6 ± 1.3. Of the study population, 38.2% were premenopausal and 61.8% were post‐menopausal. Comorbid conditions were common, with 13.8% having a respiratory disease, 14.1% with type 2 diabetes mellitus (DM2), 31.4% with hypertension, and 3.1% with coronary artery disease (CAD); overall, 45.8% of patients had at least one comorbidity. In terms of hormonal therapy exposure, 5.6% of patients had previous use of tamoxifen and 8.8% were managed with MHT. Regarding diagnostic interventions, 41.7% underwent office‐based endometrial sampling and 58.3% underwent hysteroscopy/D&C for diagnosis of EIN. The mean time from biopsy to hysterectomy was 224.4 ± 180.4 days. Among the 319 patients, 243 (76.2%) were progestin‐naïve and 76 (23.8%) received a form of progestin therapy prior to hysterectomy.
3.1. Primary outcome
Final pathology was concordant with the preoperative diagnosis of EIN in 71.2% of cases, with a higher concordance observed among progestin‐naïve patients (72.8%, n = 177) compared to those who were progestin‐treated (65.8%, n = 50). Endometrial adenocarcinoma was identified on final pathology in 11.2% of patients overall (12.8% progestin‐naïve; 7.9% progestin‐treated), while 17.2% of patients demonstrated benign disease on final pathology (14.4% progestin‐naïve, 26.3% progestin‐treated). Table 1 provides a detailed breakdown of final pathology outcomes by treatment group and individual progestin therapies.
TABLE 1.
Final pathology outcomes by treatment group and individual progestin therapies.
| Patients | Final pathology | ||
|---|---|---|---|
| EIN (n = 227) | Benign (n = 55) | Carcinoma (n = 37) | |
| Progestin‐naïve (N = 243) | 177 (72.8%) | 35 (14.4%) | 31 (12.8%) |
| Progestin‐treated (N = 76) | |||
| Total | 50 (65.8%) | 20 (26.3%) | 6 (7.9%) |
| Medroxyprogesterone (n = 24) | 20 (83.3%) | 4 (16.7%) | 0 (0.0%) |
| Megestrol acetate (n = 9) | 6 (66.7%) | 3 (33.3%) | 0 (0.0%) |
| LNG‐IUD (n = 32) | 17 (53.1%) | 10 (31.3%) | 5 (15.6%) |
| Progesterone (n = 11) | 7 (63.6%) | 3 (27.3%) | 1 (9.1%) |
Abbreviations: EIN, endometrial intraepithelial neoplasia; LNG‐IUD, levonorgestrel releasing intrauterine device.
3.2. Secondary outcomes
Patients who underwent hysteroscopy/D&C were significantly older (P = 0.02) and more likely to be post‐menopausal (P < 0.01) compared to those who had office‐based endometrial biopsies. No significant association was observed between initial biopsy technique and concordance with final pathology between office‐based endometrial sampling and hysteroscopy/D&C (progestin‐naïve, P = 0.33; progestin‐treated, P = 0.66).
When comparing progestin‐treated and progestin‐naïve patients, the latter were more likely to be older (P < 0.01), post‐menopausal (P < 0.01), have previous tamoxifen use (P < 0.05) and, on average, had shorter intervals to hysterectomy compared to the progestin‐treated cohort (210 days vs 271 days, P = 0.05). Baseline demographics and characteristics of progestin‐naïve and progestin‐treated patients are presented in Tables 2 and 3, respectively.
TABLE 2.
Baseline demographics and characteristics of progestin‐naïve patients.
| Variable a | All patients N = 243 | Final pathology | P value | ||
|---|---|---|---|---|---|
| EIN (n = 177) | Benign (n = 35) | Carcinoma (n = 31) | |||
| Patient demographics | |||||
| Age in years (mean ± SD) | 58.9 ± 9.8 | 59.1 ± 9.6 | 56.6 ± 10.2 | 59.8 ± 10.2 | 0.28 |
| BMI | 0.01 | ||||
| Overall (mean ± SD) | 32.0 ± 7.5 | 32.6 ± 7.4 | 28.0 ± 5.9 | 32.3 ± 8.3 | |
| BMI categories | 0.02 | ||||
| 18.5–24.9 | 45 (18.5) | 28 (15.8) | 11 (31.4) | 6 (19.4) | |
| 25–29.9 | 46 (18.9) | 35 (19.8) | 3 (8.6) | 8 (25.8) | |
| ≥30 | 126 (51.9) | 97 (54.8) | 13 (37.1) | 16 (51.6) | |
| Unknown | 26 (10.7) | 17 (9.6) | 8 (22.9) | 1 (3.2) | |
| Parity (mean ± SD) | 1.6 ± 1.3 | 1.5 ± 1.3 | 1.7 ± 1.3 | 1.7 ± 1.4 | 0.65 |
| Menopausal status | 0.10 | ||||
| Premenopausal | 81 (33.3) | 53 (29.9) | 17 (48.6) | 11 (35.5) | |
| Post‐menopausal | 162 (66.7) | 124 (70.1) | 18 (51.4) | 20 (64.5) | |
| Comorbidities | |||||
| Respiratory disease | 37 (15.2) | 26 (14.7) | 4 (11.3) | 7 (22.6) | 0.47 |
| Diabetes mellitus II | 33 (13.6) | 26 (14.7) | 5 (14.3) | 2 (6.5) | 0.42 |
| Hypertension | 75 (30.9) | 61 (34.5) | 8 (22.9) | 6 (19.4) | 0.18 |
| CAD | 8 (3.29) | 7 (4.0) | 1 (2.9) | 0 (0.0) | 0.58 |
| None of the above | 124 (51.0) | 87 (49.2) | 20 (57.1) | 17 (54.8) | 0.54 |
| ≥1 comorbidity | 112 (46.1) | 85 (48.0) | 13 (37.1) | 14 (45.2) | |
| Unknown | 7 (2.9) | 5 (2.8) | 2 (5.7) | 0 (0.0) | |
| Preoperative data | |||||
| Sampling method | 0.26 | ||||
| Endometrial biopsy in office | 101 (41.5) | 71 (40.1) | 13 (37.1) | 17 (54.8) | |
| Hysteroscopy/D&C | 142 (58.4) | 106 (59.9) | 22 (62.9) | 14 (45.2) | |
| Time from biopsy to hysterectomy in days (mean ± SD) | 209.8 ± 171.4 | 220.4 ± 187.1 | 197.3 ± 141.1 | 163.5 ± 77.6 | 0.20 |
| Tamoxifen use | 18 (7.4) | 16 (9.0) | 2 (5.7) | 0 (0.0) | 0.18 |
| MHT use | 22 (9.1) | 18 (10.2) | 3 (8.6) | 1 (3.2) | 0.77 |
Note: BMI, calculated as weight in kilograms divided by the square of height in meters. Bold type indicates statistical significance.
Abbreviations: BMI, body mass index; CAD, coronary artery disease; D&C, dilatation and curettage; EIN, endometrial intraepithelial neoplasia; MHT, menopausal hormone therapy; SD, standard deviation.
All values are n (%) unless indicated otherwise.
TABLE 3.
Baseline demographics and characteristics of progestin‐treated patients.
| Variable a | All patients N = 76 | Final pathology | P value | ||
|---|---|---|---|---|---|
| EIN (n = 50) | Benign (n = 20) | Carcinoma (n = 6) | |||
| Medication | 0.22 | ||||
| Medroxyprogesterone | 24 (31.6) | 20 (40.0) | 4 (20.0) | 0 (0.0) | |
| Megestrol acetate | 9 (11.8) | 6 (12.0) | 3 (15.0) | 0 (0.0) | |
| Levonorgestrel‐releasing intrauterine device | 32 (42.1) | 17 (34.0) | 10 (50.0) | 5 (83.3) | |
| Progesterone | 11 (14.5) | 7 (14.0) | 3 (15.0) | 1 (16.7) | |
| Patient demographics | |||||
| Age in years (mean ± SD) | 54.7 ± 9.8 | 54.6 ± 10.0 | 52.0 ± 9.2 | 64.2 ± 3.8 | 0.02 |
| BMI | 0.79 | ||||
| Overall (mean ± SD) | 34.1 ± 8.9 | 34.3 ± 9.6 | 32.8 ± 7.0 | 35.6 ± 9.5 | |
| BMI categories | 0.96 | ||||
| 18.5–24.9 | 12 (15.8) | 8 (16.0) | 3 (15.0) | 1 (16.7) | |
| 25–29.9 | 12 (15.8) | 9 (18.0) | 3 (15.0) | 0 (0.0) | |
| ≥30 | 47 (61.8) | 30 (60.0) | 12 (60.0) | 5 (83.3) | |
| Unknown | 5 (6.6) | 3 (6.0) | 2 (10.0) | 0 (0.0) | |
| Parity (mean ± SD) | 1.4 ± 1.2 | 1.4 ± 1.2 | 1.3 ± 1.2 | 1.7 ± 1.2 | 0.77 |
| Menopausal status | 0.18 | ||||
| Premenopausal | 41 (53.9) | 28 (56.0) | 12 (60.0) | 1 (16.7) | |
| Post‐menopausal | 35 (46.1) | 22 (44.0) | 8 (40.0) | 5 (83.3) | |
| Comorbidities | |||||
| Respiratory disease | 7 (9.2) | 4 (8.0) | 3 (15.0) | 0 (0.0) | 0.36 |
| Diabetes mellitus II | 12 (15.8) | 8 (16.0) | 3 (15.0) | 1 (16.7) | 0.51 |
| Hypertension | 25 (32.9) | 18 (36.0) | 5 (25.0) | 2 (33.3) | 0.50 |
| CAD | 2 (2.6) | 1 (2.0) | 0 (0.0) | 1 (16.7) | 0.13 |
| None of the above | 41 (53.9) | 26 (52/0) | 11 (55.0) | 4 (66.7) | |
| ≥1 comorbidity | 34 (44.7) | 24 (48.0) | 8 (40.0) | 2 (33.3) | 0.51 |
| Unknown | 1 (1.3) | 0 (0.0) | 1 (5.0) | 0 (0.0) | |
| Preoperative data | |||||
| Sampling method | 0.75 | ||||
| Endometrial biopsy in office | 32 (42.1) | 22 (44.0) | 7 (35.0) | 3 (50.0) | |
| Hysteroscopy/D&C | 44 (57.9) | 28 (56.0) | 13 (65.0) | 3 (50.0) | |
| Time from biopsy to hysterectomy in days (mean ± SD) | 271.0 ± 201.0 | 229.0 ± 150.6 | 342.0 ± 255.3 | 384.5 ± 287.6 | 0.09 |
| Tamoxifen use | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0.57 |
| MHT use | 6 (7.9) | 1 (2.0) | 3 (15.0) | 2 (33.3) | 0.04 |
Note: BMI, calculated as weight in kilograms divided by the square of height in meters. Bold type indicates statistical significance.
Abbreviations: BMI, body mass index; CAD, coronary artery disease; D&C, dilatation and curettage; EIN, endometrial intraepithelial neoplasia; MHT, menopausal hormone therapy; SD, standard deviation.
All values are n (%) unless indicated otherwise.
Among progestin‐naïve patients, those whose final pathology showed benign disease had significantly lower BMI (OR, 2.91; 95% CI: 1.18–7.13) compared to those with persistent EIN. In the progestin‐treated cohort, patients with adenocarcinoma were more likely to be of older age (OR, 1.10; 95% CI: 1.01–1.22) and to have previously used MHT (OR, 16.85; 95% CI: 1.85–218.46), relative to those with persistent EIN. Additionally, four of the six progestin‐treated patients who were ultimately diagnosed with endometrial adenocarcinoma had undergone interval hysteroscopic sampling during progestin therapy, which continued to demonstrate EIN despite the presence of carcinoma on final pathology.
Among the 319 patients, initial sampling pathology was re‐reviewed in 37% of cases (n = 118) post‐hysterectomy, either as part of referral to gynecology‐oncology team or through internal audit processes. The original diagnoses remained unchanged in all but two cases, where the final pathology demonstrated benign histology, and the initial diagnosis of EIN was revised to benign.
When comparing the descriptive and clinical characteristics among LNG‐IUD users only, there were no significant differences among those with final diagnoses of benign endometrium (no EIN), EIN, or malignancy (Table S1).
3.3. Logistic regression analyses
A logistic regression analysis performed on progestin‐naïve patients showed that lower BMI (OR 2.91; 95% CI: 1.18–7.13) and premenopausal status (OR 2.20; 95% CI, 1.06–4.58) were significant predictors of benign histology on final pathology compared to persistence of EIN. In contrast, no factors were identified as predictors of adenocarcinoma versus persistence of EIN within this cohort.
Among progestin‐treated patients, logistic regression analysis revealed that a longer interval to surgery was significantly associated with increased odds of benign final pathology compared to persistence of EIN (OR 2.50, 95% CI: 1.04–6.41). In contrast, predictive factors for endometrial adenocarcinoma on final pathology, when compared to persistence of EIN, included older age (OR, 1.10, 95% CI: 1.01–1.22) and those with a history of MHT use (OR 16.85, 95% CI: 1.85–218.46).
3.4. Surgical outcomes
Minimally invasive surgical techniques, including laparoscopic, robotic, and vaginal approaches, were employed in 304 of the 319 cases (95.3%). Gynecology‐oncologists performed 32.3% of hysterectomies while 67.2% were performed by general gynecologists. The median length of stay from surgery to discharge was 27.0 h (interquartile range: 13.0–32.0 h). Bilateral salpingo‐oophorectomy was performed in 88.4% of patients at the time of hysterectomy. Sentinel lymph node dissection was done in 20% of cases, with two patients demonstrating positive nodal metastases on final pathology. Table 4 provides a summary of operative characteristics in detail. The overall complication rate was 16%, including unplanned return to the emergency department (n = 25), bladder injury (n = 6), conversion to open (n = 3), surgical site infection (n = 2), blood transfusion (n = 2), and other complications (n = 15), including but not limited to, bowel injury, supplemental oxygen requirements, vaginal laceration during specimen extraction requiring repair (Table S2).
TABLE 4.
Operative characteristics.
| Operative data a | All patients (N = 319) | Final pathology | P value | ||
|---|---|---|---|---|---|
| EIN (n = 227) | Benign (n = 55) | Carcinoma (n = 37) | |||
| Admission type | 0.88 | ||||
| DS | 88 (27.6) | 61 (26.9) | 17 (30.9) | 10 (27.0) | |
| IP | 1 (0.3) | 1 (0.4) | 0 (0.0) | 0 (0.0) | |
| SDA | 230 (72.1) | 165 (72.7) | 38 (69.1) | 27 (73.0) | |
| Hysterectomy | 0.35 | ||||
| Laparoscopic | 223 (69.9) | 158 (69.6) | 40 (72.7) | 25 (67.6) | |
| Open | 15 (4.7) | 10 (4.4) | 5 (9.1) | 0 (0.0) | |
| Robotic | 50 (15.7) | 35 (15.4) | 6 (10.9) | 9 (24.3) | |
| Vaginal | 31 (9.7) | 24 (10.6) | 4 (7.3) | 3 (8.1) | |
| Adnexa details | 0.54 | ||||
| 1 or both tubes only | 23 (7.2) | 17 (7.5) | 6 (10.9) | 0 (0.0) | |
| 1 ovary (+/− any tubes) | 10 (3.1) | 7 (3.1) | 2 (3.6) | 1 (2.7) | |
| Tubes and 2 ovaries | 282 (88.4) | 201 (88.6) | 46 (83.6) | 35 (94.6) | |
| None | 4 (1.3) | 2 (0.9) | 1 (1.8) | 1 (2.7) | |
| Other | 0.77 | ||||
| Appendectomy | 3 (0.9) | 1 (0.4) | 1 (1.8) | 1 (2.7) | |
| Omental biopsy | 4 (1.3) | 3 (1.3) | 1 (1.8) | 0 (0.0) | |
| Sentinel lymph node dissection | 64 (20.1) | 46 (20.3) | 6 (10.9) | 12 (32.4) | |
| Total procedure duration in minutes (mean ± SD) | 132.7 ± 44.6 | 132.6 ± 44.3 | 126.7 ± 47.2 | 142.6 ± 42.3 | 0.12 |
| Length of stay in hours (median [IQR]) | 27.0 (13.0–32.0) | 27.0 (13.0–31.0) | 28.0 (10.0–50.0) | 28.0 (24.0–31.0) | 0.58 |
| Complications | 0.84 | ||||
| No | 267 (83.7) | 190 (83.7) | 45 (81.8) | 32 (86.5) | |
| Yes | 52 (16.3) | 37 (16.3) | 10 (18.2) | 5 (13.5) | |
Abbreviations: DS, day surgery; EIN, endometrial intraepithelial neoplasia; IP, inpatient; IQR, interquartile range; SD, standard deviation; SDA, same day admission.
All values are n (%) unless indicated otherwise.
4. DISCUSSION
In symptomatic patients undergoing hysterectomy for EIN, 72.8% of progestin‐naïve patients had concurrent diagnoses of EIN at time of surgery, compared to 65.8% of those treated with progestins prior to surgery. There was no significant difference in concordance of final pathology based on the method of preoperative sampling. Among progestin‐naïve patients, younger age and lower BMI were both associated with benign histology whereas in the progestin‐treated cohort, increased age and MHT use were associated with malignancy on final pathology. Lastly, shorter interval to surgery in the progestin‐treated group was associated with a higher likelihood of EIN on final pathology compared to benign disease.
Our results show that among progestin‐naïve patients, approximately 12.8% were found to have concurrent malignancy at the time of hysterectomy, compared to 8% in the progestin‐treated group, representing a significantly lower rate than previously reported in the literature. This distinction between concurrent malignancy and malignant transformation is important, as the annual risk of progression from precursor lesions to endometrial cancer is estimated to be approximately 8%, highlighting the difference between concurrent disease and true progression over time. 10 This finding underscores the importance of evaluating patient‐specific risk factors in determining both suitability for progestin therapy and expected malignancy risk. Progestin therapy may be beneficial in patients with a higher pretest probability of malignancy based on multiple risk factors (e.g., age, BMI, metabolic comorbidities, imaging findings) or anticipated prolonged interval to surgery. Foregoing progestin therapy may be warranted in cases where definitive surgical management is preferred/indicated and will be undertaken imminently.
Our study explored whether the method of preoperative diagnosis (i.e., office‐based endometrial biopsy vs hysteroscopy/D&C) influenced the concordance of final pathology. Our data revealed no significant difference in concordance rates between the two sampling methods, consistent with literature suggesting that both office‐based biopsy and hysteroscopy/D&C are reliable methods for diagnosing EIN. 11 In our study, patients undergoing hysteroscopy/D&C were more likely to be older and post‐menopausal when compared to those undergoing office‐based endometrial biopsy. Whether hysteroscopy/D&C was offered outright or following a failed office sampling is not known and may represent a limitation of office sampling not adequately captured in this study. These findings, in conjunction with the existing literature, should provide reassurance to providers that in the evaluation of abnormal uterine bleeding or post‐menopausal bleeding, office sampling is indeed a reliable evaluation technique that does not necessarily require escalation to hysteroscopy/D&C.
Progestin therapy was used in 24% of patients, most commonly in the form of an LNG‐IUD. A longer interval to surgery was more likely to yield benign histology on final pathology as opposed to findings of EIN or adenocarcinoma, likely reflecting the time‐dependent regression of lesions when treated with exogenous hormones. However, despite therapy, six patients (of the 76 progestin‐treated group) had a final diagnosis of adenocarcinoma from an initial EIN diagnosis, of whom four had interval sampling by hysteroscopy which showed no progression beyond EIN. These findings underscore the importance of continued close surveillance in this population, even during active medical management, as apparent stability may not preclude ultimate malignant transformation or concurrent malignancy only detected on pathologic examination of the uterus. Because the study cohort included only patients who ultimately underwent hysterectomy, individuals who achieved complete response to progestin therapy and avoided surgery were not captured in this analysis. As a result, the rates of persistent EIN or malignancy observed at hysterectomy may be overestimated relative to the broader population of patients initially diagnosed with EIN and managed conservatively.
The interval between EIN diagnosis and hysterectomy varied among patients and likely reflects a combination of clinical and system‐level factors, including referral to tertiary care or gynecologic oncology, surgical scheduling constraints, and patient‐related considerations. In our cohort, patients receiving progestin therapy generally experienced a longer interval to surgery due to the period of medical management and surveillance.
Our analysis of demographic and clinical factors found that higher BMI and post‐menopausal status were significant predictors of EIN diagnosis on final pathology in progestin‐naïve patients. These factors are consistent with the known risk factors for endometrial carcinoma, where prolonged estrogen exposure, often exacerbated by obesity, is a well‐established contributor to the development of both EIN and endometrial carcinoma. 1 , 2 , 3 Interestingly, no significant predictive factors were identified for the progression from EIN to adenocarcinoma in progestin‐naïve patients, which suggests that malignant transformation in EIN may be influenced by factors not fully captured in this cohort or that the progression is more stochastic in nature.
The setting of the study in a tertiary‐care referral center with expertise in gynecologic pathology was a strength in this study, supporting the quality and consistency of pathologic examination throughout the study period; however, the single‐center design may limit the generalizability of these findings to other institutions or patient populations with different referral patterns or practice environments. Pre‐ and postoperative pathologies were also largely read in‐house by the same pathology team, and near complete clinical information was available for all study participants. A further strength was the strong representation of both sampling techniques in the data as well as the most common progestin therapies used in the management of EIN, which contributes to the generalizability of our findings. Lastly, the long study window offered insights on the natural history of EIN when managed by hysterectomy.
Limitations of this study include its retrospective design, with susceptibility to selection and information bias, and potential unmeasured confounders. Clinical decision making and treatment selection were not standardized nor captured in our cohort and were based on physician and patient preference, which may have influenced our observations. Patients selected for progestin therapy may differ systematically from those undergoing primary surgical management (e.g., younger age, fertility desire, comorbidities, or surgical risk), introducing potential confounding by indication that cannot be fully accounted for in this analysis. The cohort also included patients with varying demographic characteristics, risk profiles, and management strategies, which may introduce heterogeneity. Details regarding progestin therapy, including dose, duration, and medication adherence, were not consistently documented in the medical record and therefore could not be systematically analyzed, although the type of progestin used was recorded. Differences in sampling technique between providers, whether in office biopsy or by hysteroscopy, and in subjective pathologic interpretation between pathologists may have influenced diagnostic accuracy. Furthermore, malignancy detected at time of hysterectomy could represent either a missed malignancy at time of preoperative sampling or an interval malignant transformation of previous EIN, which could not be determined by the design of this study. The 13‐year study period may reflect temporal changes in clinical practice, including evolving diagnostic approaches and surgical management strategies for EIN, which may have influenced treatment selection and observed outcomes; however, to our knowledge, no major institutional changes in surgical management for EIN occurred during the study period. Additionally, the absence of well‐established clinical criteria to guide treatment selection for EIN limits the ability to compare outcomes across management strategies and may introduce variability in decision making that could not be captured in this analysis.
5. CONCLUSION
In conclusion, concordance between preoperative EIN diagnosis and final hysterectomy pathology ranged from 65.8% to 72.8%, depending on progestin exposure, with no significant differences between preoperative sampling methods, supporting the reliability of both office biopsy and hysteroscopy/D&C. Although malignancy rates at hysterectomy were lower than previously reported, progression risk remains clinically relevant. Preoperative progestin therapy was associated with higher rates of benign final pathology, although surgical timing and individual risk factors remain critical consideration. Given the retrospective design of this study, the observed associations should be interpreted with caution, and our results do not allow determination of which patients should receive progestin therapy. Rather, these findings highlight clinical factors that may be associated with final pathology and may help inform future prospective studies. Overall, these findings support a personalized approach to EIN management that integrates patient preferences, risk profile, and treatment response.
AUTHOR CONTRIBUTIONS
All listed authors meet ICMJE criteria for authorship, having made substantial contributions to the conception and design of the work. They participated in drafting and critically revising the manuscript, approved the final version, and agreed to be accountable for all aspects of the work. JD, DBN, XZ, JPS, FWM, SK, and AZ contributed to the study's conception and design; JD and AZ prepared the material and collected, analyzed, interpreted the data. DBN, XZ, JPS, FWM, SK, and AZ contributed to methodology and project administration. DBN, XZ, JPS, FWM, SK, and AZ provided supervision. The original draft of the manuscript was written by JD and AZ. All authors commented on previous versions of the manuscript and read and approved the final manuscript.
FUNDING INFORMATION
This research did not receive any specific grant from funding agencies in the public, commercial, or not‐for‐profit sectors. Biostatistics fees were funded by a departmental Academic Enrichment Fund from the authors' clinical unit.
CONFLICT OF INTEREST STATEMENT
The authors report no conflicts of interest to declare.
Supporting information
Appendix S1.
ACKNOWLEDGMENTS
The authors would like to thank Ms. Liying Zhang, for her valuable support and expertise in conducting the statistical analyses that supported this study. An earlier version of this work was presented in abstract form at the Society of Obstetricians and Gynecologists of Canada (SOGC) Annual Clinical and Scientific Conference in June 2025. 12
DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.
REFERENCES
- 1. Nees LK, Heublein S, Steinmacher S, et al. Endometrial hyperplasia as a risk factor of endometrial cancer. Arch Gynecol Obstet. 2022;306(2):407‐421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Jarboe EA, Mutter GL. Endometrial intraepithelial neoplasia. Semin Diagn Pathol. 2010;27(4):215‐225. [DOI] [PubMed] [Google Scholar]
- 3. Schonfeld SJ, Hartge P, Pfeiffer RM, et al. An aggregated analysis of hormonal factors and endometrial cancer risk by parity. Cancer. 2013;119(7):1393‐1401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Mahdy HVE, Crotzer D. Endometrial cancer. StatPearls. StatPearls Publishing; 2024. Accessed January 4th, 2025. https://www.ncbi.nlm.nih.gov/books/NBK525981/ [PubMed] [Google Scholar]
- 5. Management of Endometrial Intraepithelial Neoplasia or atypical endometrial hyperplasia: ACOG clinical consensus No. 5. Obstet Gynecol. 2023;142(3):735‐744. [DOI] [PubMed] [Google Scholar]
- 6. Trimble CL, Kauderer J, Zaino R, et al. Concurrent endometrial carcinoma in women with a biopsy diagnosis of atypical endometrial hyperplasia: a gynecologic oncology group study. Cancer. 2006;106(4):812‐819. [DOI] [PubMed] [Google Scholar]
- 7. Dolanbay M, Kutuk MS, Uludag S, et al. Concurrent endometrial carcinoma in hysterectomy specimens in patients with histopathological diagnosis of endometrial hyperplasia in curettage specimens. Ginekol Pol. 2015;86(10):753‐758. [DOI] [PubMed] [Google Scholar]
- 8. Giede KC, Yen TW, Chibbar R, Pierson RA. Significance of concurrent endometrial cancer in women with a preoperative diagnosis of atypical endometrial hyperplasia. J Obstet Gynaecol Can. 2008;30(10):896‐901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Semere LG, Ko E, Johnson NR, et al. Endometrial intraepithelial neoplasia: clinical correlates and outcomes. Obstet Gynecol. 2011;118(1):21‐28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Doherty MT, Sanni OB, Coleman HG, et al. Concurrent and future risk of endometrial cancer in women with endometrial hyperplasia: a systematic review and meta‐analysis. PLoS One. 2020;15(4):e0232231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Sakna NA, Elgendi M, Salama MH, Zeinhom A, Labib S, Nabhan AF. Diagnostic accuracy of endometrial sampling tests for detecting endometrial cancer: a systematic review and meta‐analysis. BMJ Open. 2023;13(6):e072124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Dimakos J, Nguyen DB, Smith JP, Mansour FW, Krishnamurthy S, Zakhari A. Pre‐ and post‐operative pathology concordance in endometrial intraepithelial neoplasia. J Obstet Gynaecol Can. 2025;47(6):102856. doi: 10.1016/j.jogc.2025.102856 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1.
Data Availability Statement
Data sharing is not applicable to this article as no new data were created or analyzed in this study.