Abstract
Objective:
(a) To comparatively evaluate the performance of grayscale ultrasound features, power Doppler (PD) blood flow characteristics, and gel infusion sonography (GIS) in diagnosing endometrial cancer during real-time examination, (b) to compare the performance of real-time diagnosis of endometrial cancer by experienced observers with offline analysis by blinded observers using similar sonographic criteria during review of cine loop clips.
Methods:
152 females with post-menopausal bleeding (PMB) had ET ≥ 4 mm at first-line ultrasound were included. Two experienced radiologists evaluated endometrial patterns at real-time evaluation (grayscale ultrasound, PD, and GIS), then examinations were stored as video clips for later evaluation by two less-experienced radiologists. The reference standard was hysteroscopy (HY) and/or hysterectomy with the histopathological examination. The area under (AUC) the receiver operating characteristic (ROC) curve was calculated to assess the diagnostic performance for the prediction of endometrial cancer.
Results:
Among 152 females with ET ≥ 4 mm at first line TVUS, 88 (57.9%) patients had endometrial cancer on final pathologic analysis. Real-time ultrasound criteria (ET ≥ 5 mm with the presence of irregular branching endometrial blood vessels or multiple vessels crossing EM or areas with densely packed color-splash vessels with non-intact or interrupted EMJ at the grayscale ultrasound and/or GIS) correctly diagnosed 95% of endometrial cancers with 92% diagnostic efficiency.
There is comparable accuracy of real-time evaluation (96%) and offline analysis (92%) after the exclusion of poor quality videos from the analysis. The diagnostic criteria showed good to an excellent agreement between real-time ultrasound and offline analysis.
Conclusion:
When real-time ultrasound is performed with good technique, utilizing multiple parameters, it is possible to diagnose endometrial cancer with a high degree of accuracy and reproducibility.
Advances in knowledge:
when real-time ultrasound is performed with good technique, utilizing multiple parameters, it is possible to diagnose endometrial cancer with a high degree of accuracy and reproducibility.
Introduction
Endometrial cancer represents the most common cancer of the female genital tract in developed countries and the second most common in developing countries. Post-menopausal bleeding (PMB) may represent an alarming symptom of endometrial malignancy and early diagnosis is recommended.1
TVUS is the most cost-effective first-line examination in the diagnostic work-up of PMB. Endometrial thickness (ET) < 4 mm by TVUS can effectively exclude most of the endometrial malignancy.2 Post-menopausal females with ET ≥ 4 mm should receive second-line diagnostic techniques,3–5 which include infusion sonography, hysteroscopy, and endometrial sample.5 Up-to-date, there is no evidence or consensus regarding the most optimal second-line strategy for PMB with endometrial thickness at least 4 to 5 mm.4
In a meta-analysis,6 the endometrial sample has a failure rate ranging from 0 to 58%. Cancer cannot be excluded in cases with insufficient endometrial samples,7 and hysteroscopy is often added because hysteroscopic samples have higher diagnostic accuracy compared with endometrial samples.8,9
Evaluation of endometrial pattern at TVUS, saline infusion sonography, and PD10–14 yield promising results in the recognition of endometrial cancer in high-risk patients.14–17 Females with suspected cancer should receive fast-track diagnostic work-up including endometrial sample and fast-track reference to oncology units. This fast-track strategy will bypass all waiting lists (including hysteroscopy), which are particularly costly, challenging, and maybe unavailable in all settings. Ideally, first-line investigations should be performed with an experienced assessor to ensure accurate diagnosis; sometimes they are rarely available to perform these procedures. Usually, TVUS is performed in real-time and requires high skill in image optimization and pathology recognition. Modern ultrasound machines and software facilitate image optimization, but pathology identification still required an expert. To compensate for the lack of local experts, offline analysis of video clips by an expert could provide an efficient alternative.
We conducted this study in our setting to separately study the morphologic and blood flow characteristics of EC, as a basis to having it diagnosed by experienced observers at the real-time exam and have it read later by the inexperienced people.
Methods
Institutional review board approval was obtained for this study conducted from January 2017 to January 2019. Informed consent was taken from all patients participating in the study.
243 females with PMB (defined as bleeding >1 year after menopause) were referred from the gynecology clinic to the radiology department.
Two radiologists with 10 years' experience in gynecologic imaging performed a TVUS examination, real-time evaluation results were reported at the end of each examination, then systematic videos were stored to be analyzed offline by another two less-experienced radiologists (3 years' experience in gynecologic imaging).
Cut-off level of 4 mm ET was recommended by the European Menopause and Andropause Society.5 ET measures <4 mm was demonstrated in 60 patients with a normal clinical gynecological examination, these females underwent follow-up for 18 months for detection of missed cancer at the initial evaluation. ET measures ≥4 mm was demonstrated in 183 patients. Of them, patients referred after endometrial removal (by HY) (n = 14), and females without indication for further work-up including comorbidity or old age (n = 9), were excluded. 160 females had ET ≥ 4 mm were eligible for multiparametric ultrasound (grayscale ultrasound, PD and GIS). GIS has successfully performed in (95%, 152/160). GIS was attempted in 8 (5%) females with endometrial thickness >4 mm and they underwent hysteroscopic biopsy. Causes of GIS failure were cervical stenosis (n = 4), attenuation from myomas (n = 2), and extreme obesity (n = 2).
The final analysis was calculated for 152 females who underwent multiparametric ultrasound (grayscale ultrasound, PD and GIS).
Real-time imaging (grayscale ultrasound, PD, GIS)
All examinations were performed according to a pre-determined scanning protocol using LOGIQ 8 s X declare, GE (General electric medial system) with endovaginal (5–9 MHz) transducer.
The analysis of endometrial structure and vessels on power Doppler was evaluated using standard IETA criteria.18 The following elements were evaluated: endometrial echo pattern (isoechoic/hyperechoic/ or hypoechoic), presence of cystic areas (regular/irregular). Echo texture (homogeneity/heterogeneity); subendometrial halo: visualization and interruption of subendometrial halo (regular/irregular, homogeneity/ heterogeneity).
For the assessment of endometrial vascularization, PD settings were set to achieve maximum sensitivity for the detection of low-velocity flow without noise.
Analysis of PD parameters included the comment on color Doppler and vascular pattern. Color Doppler is evaluated subjectively by assessment of the amount of blood flow (no flow, minimal, moderate, and abundant).
The comment on endometrial vascular pattern included the presence or absence of dominant vessels (defined as one or more distinct vessels passing the endomyometrial junction). The dominant vessel may present as single or multiple and may show branching within the endometrium. Multiple dominant vessels may have a unifocal or multifocal origin at the EMJ. Scattered vessels within the endometrium, defined as dispersed color signals within the endometrium but without visible origin at the EMJ.
After evaluation of the different parameters at grayscale ultrasound and PD, the main diagnosis of malignancy was made subjectively in the presence of (heterogeneous endometrium with irregular or interrupted EMJ/ indistinct endomyometrial border with multiple or densely packed vessels on PD).
Gel infusion sonography
The cervix was cleaned with povidone-iodine, and a sterile neonatal suction catheter (with a diameter of about 2.0 mm) was introduced for about 2–7 cm beyond the external cervical os. Then, speculum was removed and the catheter was held in place. Instillagel (Farco-Pharma Gmbh, Germany) was instilled and the vaginal probe was reinserted. During gel installation, the pressure was adjusted manually (until the entire uterine cavity became expanded and visible). Evaluation of the uterine cavity was performed using the same parameters on TVUS. The endometrial thickness was measured also in the sagittal plane (as the sum of the maximum endometrial thickness at the anterior and the posterior walls).
The analysis of GIS parameters was performed using a standardized coding sheet proposed by the IETA system.18 GIS parameters include endometrial surface (regular, polypoid, irregular), lesion size (<or>25% of the endometrial surface), and the surface of the focal endometrial lesion (regular or irregular). Finally, endometrial cancer was diagnosed in the presence of cancer signs at TVUS and/or irregular endometrial surface (diffuse or focal) at GIS ± interruption to the EMJ.
Multiparametric ultrasound for the prediction of endometrial cancer was designed based on different image parameters (at grayscale ultrasound, PD, GIS) and analyzed by logistic regression in this population.
After real-time evaluation, grayscale ultrasound, PD, and GIS video recordings were saved to be evaluated 3 months later in a personal computer by a different observer (who had 10 years’ experience in the evaluation of endometrial pathology by TVUS). He was blinded to prior evaluations, the patient’s identity, and pathology to measure interobserver variability.
Reference standard
Hysteroscopy and/or hysterectomy with the histopathological examination was the reference standard. Hysteroscopic removal of all focal changes was attempted. Resectoscopic biopsies were performed in cases with large, diffuse, or localized changes (biopsy were taken from the area of the endometrium with the largest changes). In each case, three to five biopsies were sampled. If hysteroscopic findings were normal, biopsies were taken from the anterior and posterior walls of the uterine cavity followed by curettage. All specimens were processed, embedded in paraffin and sectioned at 4 µm then stained with hematoxylin-eosin (H&E) and evaluated using Olympus microscope. Photomicrographs of representative regions were taken using digital camera (Nikon DMX1200).
Offline analysis
For each real-time examination, video clips were obtained (including one sweep in an axial plane and two in the sagittal plane). For each examination, three clips were stored for their analysis in a separate file (using the patients’ ID number), clip duration ranged from 5 to 10 s. All examinations were given a new ID and were evaluated by two less-experienced radiologists (using a personal computer with virtual organ computer-aided analysis software).
A standard form including the same parameters as those obtained at the real-time evaluation was used for the offline analysis of the stored clips. To score the quality of the stored clips, a Lickert scale ranged from 1 to 5 was used (5 representing perfect quality, 1 representing poor quality). The quality of the clips depends on sharpness, brightness of the image, the presence of artifacts including air bubbles, and adequate distension of the endometrial cavity. These categories were weighted and rendered points, poor quality clips were assigned if scored <3.
Statistical analysis
Continuous data are presented as (means ± standard deviation) and categorical data are presented as (percentage). Continuous variables were compared using Student’s t-test, while categorical variables were compared using χ2 test. Statistical tests were considered statistically significant when p < 0.05.
Multivariate logistic regression was performed to detect the most accurate parameters that could be used to predict cancer. ROC-curve analysis was performed to calculate the sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of the most optimal cut-off points at real-time evaluation. The diagnostic performance of multiparametric ultrasound (grayscale, PD, GIS) at real-time evaluation and offline analysis for the prediction of endometrial cancer was calculated and compared using ROC-curve analysis. The agreement between real-time evaluation and offline analysis was calculated using the κ (k) index of agreement. IBM SPSS Statistics v. 21 (IBM Corp., Armonk, NY) was used for data analysis.
Results
At final pathological analysis, 64 (42.1%) patients had non-malignant endometrium. Of them, 24 (37.5%) had non-pathologic endometrium, 20 (31.3%) had typical endometrial hyperplasia, 13 (20.3%) had endometrial polyps, and 7 (10.9%) had atypical endometrial hyperplasia. 88 (57.9%) patients had endometrial cancer, demographics and clinical parameters were illustrated in Table 1.
Table 1.
Demographic and clinical characteristics in patients with endometrial cancer (n =88)
| Parameter | Number |
|---|---|
| Age (years) (mean ± SD) | 65 ± 8 |
| <60 | 39 (44.3%) |
| ≥60 | 49 (55.7%) |
| BMI, kg/m2 (mean ± SD) | 27 ± 8 |
| BMI <30 kg/m2 | 48 (54.5%) |
| BMI ≥ 30 kg/m2 | 40 (45.5%) |
| Hypertension,n(%) | 9 (10.2%) |
| Diabetes, n(%) | 10 (11.4%) |
| Hormonal treatment | 20 (22.7%) |
| Prior breast cancer, tamoxifen treatment | 0 (0%) |
| Cancer stagea | |
| IA n(%) | 33 (37.5%) |
| IB n(%) | 25 (28.4%) |
| II n(%) | 21 (23.9%) |
| IIIA–IIIC n(%) | 7 (8%) |
| IVA–IVB n(%) | 2 (2.3%) |
| Cancer type | |
| Endometrioid n(%) | 68 (77.3%) |
| Serous n(%) | 18 (20.5%) |
| Carcinosarcomas n(%) | 2 (2.3%) |
BMI, body mass index; SD, standard deviation.
International Federation of Gynecology and Obstetrics (FIGO) 2009 staging criteria.19,20
At TVUS, 152 females showed ET ≥ 4 mm. The mean ET was significantly higher in malignant endometrial lesions (21 mm ± 2.3) compared with benign ones (endometrial polyp 10 mm + 2.1, endometrial hyperplasia 13 mm ± 2.7) (p = 0.01).
At real-time evaluation, ET ≥ 14.6 mm with non-intact or interrupted EMJ on grayscale ultrasound had the highest diagnostic efficiency. AUC was 0.781 (CI: 0.72–0.85), and 0.791 (CI: 0.74–0.87), respectively, Specificity was 93.9%, and 96.1%, respectively.
At PD, the presence of irregular branching endometrial blood vessels or multiple vessels crossing EMJ or areas with densely packed color-splash vessels had the highest diagnostic efficiency, AUC was 0.79 (CI:0.71–0.80), 0.811 (CI:0.71–0.88), and 0.802 (CI:0.65–0.81), respectively. Specificity was 88.95, 91.9%, and 94.1%, respectively (Table 2).
Table 2.
Diagnostic efficiency of multiparametric ultrasound (grayscale ultrasound, PD and GIS) at real-time evaluation for the prediction of endometrial cancer in patients with post-menopausal bleeding and endometrial thickness≥4 mm
| AUC | Correctly classified% | Sensitivity | Specificity | LR+ | LR– | |
|---|---|---|---|---|---|---|
| (95% CI) | % | % | ||||
| Grayscale ultrasound | ||||||
| ET ≥ 14.6 | 0.767 (0.70–0.85) | 76.8 | 61.7 | 92.2 | 6.6 | 0.44 |
| Heterogeneous echogenicity | 0.582 (0.52–0.64) | 53.2 | 86.1 | 74.9 | 1.3 | 0.52 |
| Non-cystic echogencity | 0.634 (0.570.72) | 62.1 | 72.3 | 53.9 | 1.4 | 0.56 |
| Irregular endomyometrial junction | 0.733 (0.77–0.80) | 73.1 | 74.2 | 70.9 | 2.4 | 0.39 |
| Iintrupted endomyometrial junction | 0.781 (0.72–0.85) | 81.2 | 61.4 | 93.9 | 11.1 | 0.43 |
| Non-intact endomyometrial junction | 0.791 (0.74–0.87) | 82.1 | 63.4 | 96.1 | 13.2 | 0.42 |
| GIS | ||||||
| ET ≥ 12.8 | 0.811 (0.73–0.89) | 76.9 | 77.8 | 77.2 | 3.5 | 0.31 |
| Iintrupted endomyometrial junction | 0.811 (0.79–0.85) | 84.1 | 72.2 | 95.3 | 11.9 | 0.33 |
| Non-intact endomyometrial junction | 0.866 (0.75–0.90) | 85.1 | 72.9 | 95.1 | 13.1 | 0.33 |
| Irregular endometrial surface | 0.799 (0.75–0.82) | 79.3 | 82.2 | 76.4 | 3.6 | 0.24 |
| PD | ||||||
| Color Doppler | ||||||
| (No flow/ minimal flow) | 0.588 (0.55–0.70) | 52.1 | 92.1 | 28.3 | 1.2 | 0.29 |
| (Moderate/abundant flow) | 0.765 (0.69–0.82) | 76.1 | 69.2 | 82.1 | 3.4 | 0.39 |
| Vascular pattern | ||||||
| Single dominant vessel ± branching | 0.755 (0.62–0.83) | 73.9 | 65.2 | 81.1 | 3.5 | 0.41 |
| Multiple dominant vessel ± branching | 0.677 (0.61–0.80) | 70 | 69.1 | 68.9 | 2.1 | 0.49 |
| 0.42 | ||||||
| Scattered vessels | 0.811 (0.71–0.88) | 84.1 | 62.2 | 91.9 | 8.2 | |
| 0.6 | ||||||
| Areas with densely packed or color-splash vessels | 0.802 (0.65–0.81) | 76.2 | 49.9 | 94.1 | 7.6 |
AUC, area under the curve; CI, confidence interval; GIS, gel infusion sonography; LR, likelihood ratio; PD, power Doppler.
At GIS, ET ≥ 12.8 mm with non-intact EMJ had the highest diagnostic efficiency, AUC was 0.866 (CI: 0.75–0.90) with specificity 95.1%. Among eight females with failed GIS, cancer was confirmed in one case by hysteroscopic biopsy. Figure 1 displays the diagnostic performance (AUC) of multiparametric ultrasound (grayscale ultrasound, PD and GIS) at real-time evaluation.
Figure 1.
Displays the diagnostic performance (AUC) of multiparametric ultrasound (grayscale ultrasound, PD, GIS) at real time evaluation. AUC, are under the curve; GIS, gel infusion sonography; PD, power Doppler, ROC, receiver operating characteristic.
On multivariate regression, ET ≥ 5 mm with the presence of irregular branching endometrial blood vessels or multiple vessels crossing EMJ or areas with densely packed color-splash vessels with non-intact or interrupted EMJ at the grayscale ultrasound and/or GIS had the highest diagnostic efficiency, AUC was 0.919 (CI: 0.89–0.96) (Figures 2 and 3).
Figure 2.
Female aged 52 years old presented with PMB. TVUS B-mode image (A) showed thickened endometrium measuring 18 mm with interrupted endometrial–myometrial junction (arrow). On color Doppler (B), multiple vessels seen within the endometrium, some of them cross the myometrial–endometrial junction. GIS image (C) showed an irregular endometrial focal lesion (star) with irregular surface and interruption of the endometrial myometrial gray halo (arrow). (D, E) MRI axial andcoronal T2WI showed an endometrial soft tissue lesion involving the entire endometrial cavity with infiltration of the junctional zone and entire myometrium reaching to the serosa (arrow), infiltrating the cervical stroma, no parametrial infiltration. Operative and histopathological diagnosis was Stage II endometriodadenocarcinoma. GIS, gel infusion sonography; PMB, post-menopausal bleeding; T2WI, T2 weighted imaging.
Figure 3.
Post-menopausal female aged 58 years old presented with PMB. B-mode ultrasound image (A) showed thickened endometrium measuring 21mm with interrupted endometrial myometrial gray halo (arrow). On color Doppler (B), multiple vessels seen within the endometrium, some of them cross the myometrial–endometrialjunction (arrows). GIS image (C) showed an irregular endometrial focal lesion with irregular surface and interruption of the endometrial myometrial gray halo (arrow). (D) MRI axial T2WI showed an endometrial soft tissue lesion filling the uterine cavity, infiltrating the junctional zone with thinned out myometrium (infiltrate more than 50%, arrow), no parametrial infiltration. Operative and histopathological diagnosis was Stage IB endometriod adenocarcinoma, microscopic image (E). GIS, gel infusion sonography; PMB, post-menopausal bleeding; T2WI, T2 weighted imaging.
Overall sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and AUC of real-time evaluation in the prediction of endometrial cancer were 94.1% (CI:88–95%), 95.2% (CI: 89–96%), 17.8 (9.3–22.5), 0.10 (0.03–0.19), respectively. AUC was 0.961 (CI: 0.88–0.97).
Poor-quality videos were seen in (5.9%, 9/152) at offline analysis. Causes of poor quality were: poor sharpness (n = 6), lack of distension (n = 2), and air bubbles (n = 1). After the exclusion of poor quality videos from the analysis, overall sensitivity, specificity, positive likelihood ratio, negative likelihood ratio of offline analysis in the prediction of endometrial cancer was 92% (CI: 83–92%), 92% (CI:89–95%), 11.7 (CI:7.1–20.3), 0.15 (CI:0.06–0.19), respectively. AUC was 0.922 (CI: 0.87–0.93).
There is comparable accuracy of multiparametric ultrasound (grayscale ultrasound, PD and GIS) in the real-time evaluation and offline analysis (after exclusion of poor quality videos) (Table 3).
Table 3.
Real-time ultrasound vs offline analysis: overall comparative performance of multiparametric ultrasound (grayscale ultrasound, PD and GIS)
| Sensitivity | Specificity | LR+ | LR– | AUC | |
|---|---|---|---|---|---|
| % (95% CI) | % (95% CI) | % (95% CI) | % (95% CI) | % (95% CI) | |
| Grayscale ultrasound | |||||
| Real-time evaluation | 92.7 (87–95) | 93.2 (89–96) | 6.3 (2.8–10.7) | 0.19(.05–0.20) | 0.93 (0.83–0.93) |
| Off line analysisa | 89.1 (75–89) | 89.3 (86–91) | 6.1 (2.2–9.1) | 0.17(.04–0.24) | 0.89 (0.80–0.91) |
| PD | |||||
| Real-time evaluation | 90.8 (86–95) | 93.6 (89–95) | 10.1 (2.3–11.4) | 0.15(.06–0.17) | 0.92 (0.84–0.95) |
| Off line analysisa | 88.9 (84–96) | 91.1 (90–96) | 9.2 (3.4–10.1) | 0.19(.09–0.19) | 0.88 (0.81–0.95) |
| GIS | |||||
| Real-time evaluation | 91 (82–91) | 96.1 (86–97) | 19.2 (7.9–22.9) | 0.11 (0.08–0.12) | 0.94 (0.84–0.98) |
| Off line analysisa | 90.9 (84–91) | 92.4 (88–94) | 18.4 (6.8–19.2) | 0.13 (0.09–0.13) | 0.91 (0.86–0.94) |
GIS, gel infusion sonography; PD, power Doppler.
Poor quality videos were seen in (5.9%, 9/152) at offline analysis and were excluded from the final anlysis.
After the exclusion of poor quality videos, the agreement between real-time evaluation (experienced radiologist) and offline analysis (unexperienced radiologist) ranged from good to excellent (Table 4). It was excellent in the evaluation of ET (k = 0.88, CI: 0.82–0.90), and vascular pattern (k = 0.84, CI: 0.81–0.88). It was good in the evaluation of and EMJ (k = 0.77, CI: 0.76–0.80).
Table 4.
Agreement coefficient between real-time ultrasound and offline analysis: detailed analysis of diagnostic criteria in grayscale ultrasound, PD and GIS †
| Parameter | ka (95% confidence interval) |
|---|---|
| Grayscale ultrasound | |
| ET | 0.90 (0.78–0.90) |
| Iintrupted endomyometrial junction | 0.80 (0.75–0.83) |
| Non-intact endomyometrial junction | 0.76 (0.75–0.81) |
| GIS | |
| ET | 0.82 (0.79–0.88) |
| Iintrupted endomyometrial junction | 0.79 (0.72–0.82) 11 |
| Non-intact endomyometrial junction | 0.80 (0.79–0.86) |
| PD (vascular pattern) | |
| Scattered vessels | 0.81 (0.79–0.83) |
| Areas with densely packed or color-splash vessels | 0.88 (0.80–0.89) |
GIS, gel infusion sonography; PD, power Doppler.
κ value of 0.81–1.00 indicate excellent agreement, a k-value of 0.61–0.80 indicate good agreement; a k-value of 0.41–0.60 indicate moderate agreement; a k-value of 0.21–0.40 indicates fair agreement; a k-value of < 0.20 indicates poor agreement.
Poor quality videos were seen in (5.9%, 9/152) at offline analysis and were excluded from the final anlysis
Discussion
Our results showed that real-time ultrasound criteria (ET ≥ 5 mm with the presence of irregular branching endometrial vessels or multiple vessels crossing the EMJ or areas with densely packed/color-splash vessels with non-intact or interrupted EMJ at the grayscale ultrasound and/or GIS) correctly diagnosed 95% of endometrial cancer patients (age was ≥60 years in 55.7%, BMI ≥30 kg/m2 in 45.5%, Stage I tumors was demonstrated in 60%, and 22.7% received hormonal treatment) with highest diagnostic efficiency (92%).
Previous studies used fewer parameters including ET, heterogeneous echogenicity, and multivessel vascular pattern16,21 reported lower accuracy.22 An interrupted EMJ23 and endometrial surface irregularity at GIS24 are good predictors for endometrial cancer.
Concordant with our findings, the presence of irregular branching endometrial vessels,11,13 densely packed or color splash blood vessels,13 and multiple vessels crossing EMJ15 increased the like hood of malignancy. Also, the presence of a single prevalent vessel penetrating the endometrium from the myometrium is characteristic of endometrial polyp.15
For the prediction of endometrial cancer in females with PMB, TVUS algorithm has been developed.22–25 In the present study, the highest diagnostic efficiency was achieved at ET ≥ 5 mm +PD (irregular branching endometrial vessels, multiple vessels crossing the EMJ, areas with densely packed or color-splash vessels) +non-intact EMJ at GIS. These criteria also had the highest efficiency in prior studies.23,25,26
Both real-time evaluation and offline analysis (after the exclusion of bad quality clips) had comparable accuracy in the prediction of endometrial cancer. Bad image quality is the main problem with offline analysis; high accuracy was obtained when analysis was restricted to high-quality images. We think image optimization is essential for high-quality images when the offline analysis is required.
Interestingly, the offline analysis had high sensitivity in the prediction of cancer, similarly.27 Five endometrial cancer cases were falsely diagnosed as endometrial hyperplasia at offline analysis, (GIS showed regular diffusely thickened endometrium, PD showed multiple dominant branching vessels). At the real-time evaluation, one of these cases was correctly diagnosed by the detection of focal interruption of EMJ. In PMB, TVUS appearance is not specific,24,28 and also GIS is not reliable for the diagnosis of endometrial hyperplasia.29 Opolskiene et al24 reported that diffuse thickening of the endometrium was not predictive of diagnosis and differential diagnosis including endometrial hyperplasia, metaplasia, and cancer should be considered. It represents a real problem in clinical practice and even experienced hysteroscopists may found difficulty in differentiation between hyperplasia and other endometrial abnormalities.30,31 We think detection of cancerous focus is more important than a diagnosis of hyperplasia itself, adding DWI sequence to routine conventional MRI may help to detect occult carcinoma within endometrial hyperplasia.32
In our study, real-time evaluation and offline analysis had high specificity in the prediction of endometrial cancer. Seven cases with simple endometrial hyperplasia were falsely diagnosed as cancer at offline analysis (GIS showed diffuse irregular thickening of the endometrium with non-intact EMJ and scattered vessels were detected on PD), three of these case were correctly diagnosed at real-time evaluation with intact EMJ. The specificity of cancer diagnosis differs according to the morphological appearance of the endometrial lesion23,24,27 and the integrity of the EMJ.13,23 A lower specificity was described in cancer cases presented as diffuse lesions23,27 while higher specificity (97 %) was achieved in cancer cases presented as a focal endometrial lesion with an irregular surface.24 In agreement with,24 high specificity (95.2%) was achieved in our study because 83% of endometrial cancer cases presented as an endometrial focal lesion with an irregular surface. Also, the presence of irregular EMJ,13,23 or an interrupted EMJ (due to the interruption of the subendometrial halo)23 are more specific signs in the diagnosis of cancer. In our study, 86% of cancer cases showed interrupted EMJ, this sign should be evaluated while moving the probe slightly at the point at which the EMJ is irregular. However, these signs are highly discriminative; their subjective nature could affect their implementation in clinical practice.
Reproducibility of TVUS33–35 and infusion sonography34–36 has been previously reported. In agreement with,35 our results demonstrated excellent agreement in the measurement of ET between real-time evaluation (experienced radiologist) and offline analysis (unexperienced radiologist). Epstein et al35 described that one ET measurement at TVUS is enough in clinical practice and could provide reliable discrimination between post-menopausal females with ET ≤4.4 mm and ≥4.5 mm.
Reproducibility of infusion sonography was dependent on observer experience34 and the aim of the study (normal vs abnormal endometrium or benign vs malignant endometrium or a specific type of endometrial pathology).36,37 Our results demonstrated good to an excellent agreement at GIS even with inexperienced hands, similarly.34 We think diagnostic performance in GIS is influenced by an adequate level of training rather than years of experience and thus adequate training is essential for implementation in general practice.
It is much simpler and cheaper to assess TVUS examinations remotely than to refer the patient to a radiology clinic. This approach is time-consuming and cost-efficient and already used for MRI and ultrasound. Based on offline analysis, patients can be classified simply into four categories: 1) endometrial cancer is unlikely (examinations with thin endometrium); 2) benign pathology is likely; 3) additional imaging is indicated, and 4) cancer is very likely and fast track strategy is indicated.
Our results demonstrated that multiparametric ultrasound is warranted to optimize patient management. It could determine the best second-line diagnostic procedures and accelerate the diagnosis of endometrial cancer. The offline analysis could provide an efficient alternative if the local expert was not available.
The limitation of this study was that 60 patients with ET <4 mm underwent follow up without endometrial pathology. It could represent a limitation in our study. However, the absence of endometrial cancer in this group of patients proved the effectiveness of this management strategy in these populations. Tumor vascularity is dependent on tumor characteristics21,38 and the sample of cancer cases. Accuracy of cancer diagnosis is dependent on diagnostic signs with subjective nature (endometrial focal lesion with an irregular surface, irregular EMJ, an interrupted EMJ), and variable accuracy could be expected. Therefore, larger multicenter studies are warranted to evaluate the efficiency of multiparametric ultrasound, to determine the best second-line diagnostic procedures for early diagnosis of endometrial cancer. Implementation of continuous systematic education in-residence training may help to reduce inter observer variability in sonographic examinations.
Conclusion
When real-time ultrasound is performed with good technique, utilizing multiple parameters, it is possible to diagnose endometrial cancer with a high degree of accuracy and reproducibility.
Contributor Information
Shimaa Abdalla, Email: shimaaabdalla@aun.edu.eg.
Hisham Abou-Taleb, Email: hishamaboutaleb1@yahoo.com.
Dalia M. Badary, Email: hamasat82@yahoo.com.
Wageeh A. Ali, Email: Drwageeh_abdelhafeez@aun.edu.eg.
REFERENCES
- 1.NCCN Clinical Practice Guidelines in Oncology: uterine Cancer.. ; 2018AvailableNCCN. Available from: http://www.nccn.org.
- 2.Timmermans A, Opmeer BC, Khan KS, Bachmann LM, Epstein E, Clark TJ. Endometrial thickness measurement for detecting endometrial cancer in women with postmenopausal bleeding: a systematic review and meta-analysis, Obstet. Gynecol 2010; 116: 160–7. [DOI] [PubMed] [Google Scholar]
- 3.van Hanegem N, Prins MMC, Bongers MY, Opmeer BC, Sahota DS, Mol BWJ, et al. The accuracy of endometrial sampling in women with postmenopausal bleeding: a systematic review and meta-analysis. Eur J Obstet Gynecol Reprod Biol 2016; 197: 147–55. doi: 10.1016/j.ejogrb.2015.12.008 [DOI] [PubMed] [Google Scholar]
- 4.van Hanegem N, Breijer MC, Khan KS, Clark TJ, Burger MPM, Mol BWJ, et al. Diagnostic evaluation of the endometrium in postmenopausal bleeding: an evidence-based approach. Maturitas 2011; 68: 155–64. doi: 10.1016/j.maturitas.2010.11.010 [DOI] [PubMed] [Google Scholar]
- 5.Dreisler E, Poulsen LG, Antonsen SL, Ceausu I, Depypere H, Erel CT, et al. EMAS clinical guide: assessment of the endometrium in peri and postmenopausal women. Maturitas 2013; 75: 181–90. doi: 10.1016/j.maturitas.2013.03.011 [DOI] [PubMed] [Google Scholar]
- 6.Dijkhuizen FP, Mol BW, Brölmann HA, Heintz AP. The accuracy of endometrial sampling in the diagnosis of patients with endometrial carcinoma and hyperplasia: a meta-analysis. Cancer 2000; 89: 1765–72. doi: [DOI] [PubMed] [Google Scholar]
- 7.van Doorn HC, Opmeer BC, Burger CW, Duk MJ, Kooi GS, Mol BWJ, et al. Inadequate office endometrial sample requires further evaluation in women with postmenopausal bleeding and abnormal ultrasound results. Int J Gynaecol Obstet 2007; 99: 100–4. doi: 10.1016/j.ijgo.2007.05.040 [DOI] [PubMed] [Google Scholar]
- 8.Bakour SH, Jones SE, O'Donovan P. Ambulatory hysteroscopy: evidence-based guide to diagnosis and therapy. Best Pract Res Clin Obstet Gynaecol 2006; 20: 953–75. doi: 10.1016/j.bpobgyn.2006.06.004 [DOI] [PubMed] [Google Scholar]
- 9.Cicinelli E. Hysteroscopy without anesthesia: review of recent literature. J Minim Invasive Gynecol 2010; 17: 703–8. doi: 10.1016/j.jmig.2010.07.003 [DOI] [PubMed] [Google Scholar]
- 10.Dueholm M, Marinovskij E, Hansen ES, Møller C, Ørtoft G. Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm. Menopause 2015; 22: 616–26. doi: 10.1097/GME.0000000000000358 [DOI] [PubMed] [Google Scholar]
- 11.Epstein E, Valentin L. Gray-scale ultrasound morphology in the presence or absence of intrauterine fluid and vascularity as assessed by color Doppler for discrimination between benign and malignant endometrium in women with postmenopausal bleeding. Ultrasound Obstet Gynecol 2006; 28: 89–95. doi: 10.1002/uog.2782 [DOI] [PubMed] [Google Scholar]
- 12.Epstein E, Skoog L, Isberg P-E, De Smet F, De Moor B, Olofsson P-A, et al. An algorithm including results of gray-scale and power Doppler ultrasound examination to predict endometrial malignancy in women with postmenopausal bleeding. Ultrasound Obstet Gynecol 2002; 20: 370–6. doi: 10.1046/j.1469-0705.2002.00800.x [DOI] [PubMed] [Google Scholar]
- 13.Opolskiene G, Sladkevicius P, Valentin L. Ultrasound assessment of endometrial morphology and vascularity to predict endometrial malignancy in women with postmenopausal bleeding and sonographic endometrial thickness >or= 4.5 mm. Ultrasound Obstet Gynecol 2007; 30: 332–40. doi: 10.1002/uog.4104 [DOI] [PubMed] [Google Scholar]
- 14.Opolskiene G, Sladkevicius P, Valentin L. Prediction of endometrial malignancy in women with postmenopausal bleeding and sonographic endometrial thickness ≥ 4.5 MM. Ultrasound Obstet Gynecol 2011; 37: 232–40. doi: 10.1002/uog.8871 [DOI] [PubMed] [Google Scholar]
- 15.Alcázar JL, Castillo G, Mínguez JA, Galán MJ. Endometrial blood flow mapping using transvaginal power Doppler sonography in women with postmenopausal bleeding and thickened endometrium. Ultrasound Obstet Gynecol 2003; 21: 583–8. doi: 10.1002/uog.143 [DOI] [PubMed] [Google Scholar]
- 16.Alcazar JL, Galvan R. Three-Dimensional power Doppler ultrasound scanning for the prediction of endometrial cancer in women with postmenopausal bleeding and thickened endometrium. Am J Obstet Gynecol 2009; 200: 44.e1–44.e6. doi: 10.1016/j.ajog.2008.08.027 [DOI] [PubMed] [Google Scholar]
- 17.Opolskiene G, Sladkevicius P, Jokubkiene L, Valentin L. Three-Dimensional ultrasound imaging for discrimination between benign and malignant endometrium in women with postmenopausal bleeding and sonographic endometrial thickness of at least 4.5 MM. Ultrasound Obstet Gynecol 2010; 35: 94–102. doi: 10.1002/uog.7445 [DOI] [PubMed] [Google Scholar]
- 18.Leone FPG, Timmerman D, Bourne T, Valentin L, Epstein E, Goldstein SR, et al. Terms, definitions and measurements to describe the sonographic features of the endometrium and intrauterine lesions: a consensus opinion from the International endometrial tumor analysis (IETA) group. Ultrasound Obstet Gynecol 2010; 35: 103–12. doi: 10.1002/uog.7487 [DOI] [PubMed] [Google Scholar]
- 19.Creasman W. Revised FIGO staging for carcinoma of the endometrium. Int J Gynaecol Obstet 2009; 105: 109. doi: 10.1016/j.ijgo.2009.02.010 [DOI] [PubMed] [Google Scholar]
- 20.Pecorelli S. Revised FIGO staging for carcinoma of the vulva, cervix, and endometrium. Int J Gynaecol Obstet 2009; 105: 103–4. doi: 10.1016/j.ijgo.2009.02.012 [DOI] [PubMed] [Google Scholar]
- 21.Epstein E, Van Holsbeke C, Mascilini F, Måsbäck A, Kannisto P, Ameye L, et al. Gray-scale and color Doppler ultrasound characteristics of endometrial cancer in relation to stage, grade and tumor size. Ultrasound Obstet Gynecol 2011; 38: 586–93. doi: 10.1002/uog.9038 [DOI] [PubMed] [Google Scholar]
- 22.Sladkevicius P, Opolskiene G, Valentin L. Prospective temporal validation of mathematical models to calculate risk of endometrial malignancy in patients with postmenopausal bleeding. Ultrasound Obstet Gynecol 2017; 49: 649–56. doi: 10.1002/uog.15941 [DOI] [PubMed] [Google Scholar]
- 23.Dueholm M, Moeller C, Rydbjerg S, Hansen ES, Oertoft G. An ultrasound algorithm for the identification of endometrial cancer, ultrasound Obstet. Gynecol 2014; 43: 557–68. [DOI] [PubMed] [Google Scholar]
- 24.Opolskiene G, Sladkevicius P, Valentin L. Two- and three-dimensional saline contrast sonohysterography: interobserver agreement, agreement with hysteroscopy and diagnosis of endometrial malignancy. Ultrasound Obstet Gynecol 2009; 33: 574–82. doi: 10.1002/uog.6350 [DOI] [PubMed] [Google Scholar]
- 25.Dueholm M, Hjorth IMD, Dahl K, Hansen ES, Ørtoft G. Ultrasound scoring of endometrial pattern for fast-track identification or exclusion of endometrial cancer in women with postmenopausal bleeding. J Minim Invasive Gynecol 2019; 26: 516–25. doi: 10.1016/j.jmig.2018.06.010 [DOI] [PubMed] [Google Scholar]
- 26.Dueholm M, Hjorth IMD, Dahl K, Pedersen LK, Ørtoft G. Identification of endometrial cancers and atypical hyperplasia: development and validation of a simplified system for ultrasound scoring of endometrial pattern. Maturitas 2019; 123: 15–24. doi: 10.1016/j.maturitas.2019.01.017 [DOI] [PubMed] [Google Scholar]
- 27.Ahmed SA, Abo-Taleb H. Diagnostic efficacy of gel infusion sonography in post-menopausal bleeding. Egypt J Radiol Nucl Med 2020; 51: 194. doi: 10.1186/s43055-020-00308-w [DOI] [Google Scholar]
- 28.Valentin L. Imaging techniques in the management of abnormal vaginal bleeding in non-pregnant women before and after menopause. Best Pract Res Clin Obstet Gynaecol 2014; 28: 637–54. doi: 10.1016/j.bpobgyn.2014.04.001 [DOI] [PubMed] [Google Scholar]
- 29.Dueholm M, Hjorth IMD, Secher P, Jørgensen A, Ørtoft G. Reproducibility of Endometrial Pathologic Findings Obtained on Hysteroscopy, Transvaginal Sonography, and Gel Infusion Sonography in Women With Postmenopausal Bleeding. J Minim Invasive Gynecol 2015; 22: 1036–44. doi: 10.1016/j.jmig.2015.05.019 [DOI] [PubMed] [Google Scholar]
- 30.Lasmar RB, Barrozo PRM, de Oliveira MAP, Coutinho ESF, Dias R. Validation of hysteroscopic view in cases of endometrial hyperplasia and cancer in patients with abnormal uterine bleeding. J Minim Invasive Gynecol 2006; 13: 409–12. doi: 10.1016/j.jmig.2006.05.002 [DOI] [PubMed] [Google Scholar]
- 31.Garuti G, Cellani F, Garzia D, Colonnelli M, Luerti M. Accuracy of hysteroscopic diagnosis of endometrial hyperplasia: a retrospective study of 323 patients. J Minim Invasive Gynecol 2005; 12: 247–53. doi: 10.1016/j.jmig.2005.03.006 [DOI] [PubMed] [Google Scholar]
- 32.Ahmed SA, El Taieb HA, Abotaleb H. Diagnostic performance of sonohysterography and MRI diffusion in benign endometrial lesion characterization. The Egyptian Journal of Radiology and Nuclear Medicine 2018; 49: 579–89. doi: 10.1016/j.ejrnm.2018.02.010 [DOI] [Google Scholar]
- 33.Epstein E, Van HC, Mascilini F, Masback A, Kannisto P, Ameye L. Grayscale and color Doppler ultrasound characteristics of endometrial cancer in relation to stage, grade and tumor size, ultrasound Obstet. Gynecol 2011; 38: 586–93. [DOI] [PubMed] [Google Scholar]
- 34.Dueholm M, Hjorth IMD, Secher P, Jørgensen A, Ørtoft G. Reproducibility of Endometrial Pathologic Findings Obtained on Hysteroscopy, Transvaginal Sonography, and Gel Infusion Sonography in Women With Postmenopausal Bleeding. J Minim Invasive Gynecol 2015; 22: 1036–44. doi: 10.1016/j.jmig.2015.05.019 [DOI] [PubMed] [Google Scholar]
- 35.Epstein E, Valentin L. Intraobserver and interobserver reproducibility of ultrasound measurements of endometrial thickness in postmenopausal women. Ultrasound Obstet Gynecol 2002; 20: 486–91. doi: 10.1046/j.1469-0705.2002.00841.x [DOI] [PubMed] [Google Scholar]
- 36.Dueholm M, Lundorf E, Sørensen JS, Ledertoug S, Olesen F, Laursen H. Reproducibility of evaluation of the uterus by transvaginal sonography, hysterosonographic examination, hysteroscopy and magnetic resonance imaging. Hum Reprod 2002; 17: 195–200. doi: 10.1093/humrep/17.1.195 [DOI] [PubMed] [Google Scholar]
- 37.Beemsterboer SN, Thurkow AL, Verstraeten R, Brölmann HAM. Reproducibility of saline contrast sonohysterography for the detection of intracavitary abnormalities in women with abnormal uterine bleeding. Ultrasound Obstet Gynecol 2008; 31: 445–9. doi: 10.1002/uog.5287 [DOI] [PubMed] [Google Scholar]
- 38.Saarelainen SK, Vuento MH, Kirkinen P, Mäenpää JU. Preoperative assessment of endometrial carcinoma by three-dimensional power Doppler angiography. Ultrasound Obstet Gynecol 2012; 39: 466–72. doi: 10.1002/uog.10103 [DOI] [PubMed] [Google Scholar]