Abnormal uterine bleeding represents the most frequent indication in women for gynecological surgical intervention. Although the differential diagnosis is broad and frequently benign, most perimenopausal and postmenopausal women undergo a diagnostic procedure to exclude endometrial carcinoma. Many consider endometrial biopsy the standard of care when screening for endometrial cancer. Although the Pipelle device was reported to have a sensitivity of 97.5%, the diagnosis of cancer already was known before the sampling.1 Since then, studies have demonstrated that endometrial biopsy is best for detecting global, not focal, processes in the endometrium.2
Evaluation of the endometrium by transvaginal ultrasonography is currently the imaging method of choice in the diagnostic workup for abnormal uterine bleeding.3–5 An endometrial thickness of 5 mm or less has a 99% negative predictive value in excluding endometrial carcinoma in postmenopausal women, but a thickness greater than 5 mm is a nonspecific finding.6,7 Saline-infusion sonography improves the characterization of endometrial thickening and the detection and description of endometrial lesions, with a sensitivity and specificity comparable with hysteroscopy.8–11
The combination of endometrial biopsy and saline-infusion sonography enhances the diagnostic accuracy of either procedure performed alone.12,13 Sampling the endometrium under direct visualization during saline-infusion sonography would seem a logical next step. The purpose of our study is to estimate prospectively whether saline-infusion sonography endometrial sampling will improve the diagnosis of benign and malignant endometrial disease compared with endometrial biopsy in perimenopausal and postmenopausal women with abnormal uterine bleeding.
MATERIALS AND METHODS
From January 2006 through November 2007, all perimenopausal and postmenopausal women (aged 40 years or older) with abnormal uterine bleeding and abnormal endometrial cavities on ultrasound examination who were referred for saline-infusion sonography were eligible to enroll in the saline-infusion sonography endometrial sampling study. Exclusion criteria included pregnancy, inability to give informed consent, having enrolled previously, and voluntarily choosing not to participate. Ninety-eight consecutive perimenopausal and postmenopausal women consented to undergo saline-infusion sonography endometrial sampling at the time of their sonohysterography procedures. The procedures were performed in the Obstetrics and Gynecology Ultrasound Unit at Parkland Health and Hospital System. The study was approved by the institutional review board of the University of Texas Southwestern Medical Center (IRB# 112004-036).
The initial clinical assessment of the patients was done by the resident and attending physicians of the gynecology service and included a physical examination and blind endometrial biopsy using a Pipelle device, which is our present standard of care. Transvaginal ultrasonography then was performed in those patients referred by the gynecology service. The ultrasonographer and sonologist were not aware of the gynecology endometrial biopsy results. The endometrial appearance and thickness were measured in the sagittal plane. A woman with an endometrium with characteristics suspicious for pathology, such as a thickened endometrium (defined as greater than 10 mm in perimenopausal women6 and greater than 5 mm in postmenopausal women), heterogeneous or echogenic texture of the endometrium, suspicion of a focal endometrial mass, or inability to visualize the endometrium, was offered further evaluation with sonohysterography, which is our current standard of care.
This sonohysterography was performed on days 5–10 of the patient’s subsequent menstrual cycle or bleeding episode to minimize the effect of the prior endometrial biopsy on the sonohysterography findings. After informed written consent was obtained, the patient was placed in the dorsal lithotomy position and a speculum was placed into the vagina to expose the cervix. The external os was cleaned with povidone-iodine (Betadine). A 7 F balloon catheter (Ackard Labs, Cooper Surgical, Trumball, CT) was advanced through the cervical canal. If the catheter could not be passed secondary to cervical stenosis, which occurred with four patients, a paracervical block with 1% lidocaine without epinephrine was performed and the cervix was dilated sequentially using lacrimal duct dilators under direct sonographic observation. The transvaginal probe was inserted, and approximately 10–20 mL of sterile saline solution was injected slowly under direct sonographic visualization. Multiple sagittal and coronal images of the endometrium were obtained with an Antares (Siemens, Redmond, WA), an Elegra (Siemens), or a Voluson 730 (General Electric Medical Systems, Milwaukee, WI) scanner and 5-MHz or multifrequency transvaginal transducers. Per our protocol, antibiotic postprocedure prophylaxis (doxycycline 200 mg orally) was given if the patient was diabetic or immunocompromised.
Saline-infusion sonography endometrial sampling was performed at the end of the sonohysterography under direct transvaginal sonographic visualization after the catheter was removed. The biopsy was taken at the sight of the endometrial abnormality (thickened endometrium, polyp) using an endometrial sampling curette (Uterine Explora II Curette, Milex, Trumball, CT) or at a representative site in the uterine cavity if the endometrium had a normal appearance. The design of the curette is such that the aperture is in line with the flat surface of the curette handle so that directionality of the curette is known after insertion into the endometrial cavity. Only one board-certified ob-gyn physician/sonologist performed all the sonohysterography and saline-infusion sonography endometrial sampling procedures. Similarly, all saline-infusion sonography endometrial sampling specimens were reviewed by a single, board-certified pathologist who specializes in ob-gyn cytology/histology. This pathologist was blind to the patient’s previous endometrial biopsy.
The ability of saline-infusion sonography endometrial sampling compared with endometrial biopsy to obtain tissue for diagnosis was examined using McNemar’s test. The histopathologic diagnoses of the blind endometrial biopsy and saline-infusion sonography endometrial sampling then were correlated to final outcomes. Final outcomes constituted either a gold standard histological diagnosis obtained on hysteroscopy-dilation and curettage (D&C) or hysterectomy specimen or 1-year follow-up in patients without final pathology who had benign saline-infusion sonography endometrial sampling or sonohysterography diagnoses. A Kappa value for each technique was calculated for diagnostic accuracy, and, to evaluate agreement between methods, Kappa values were compared using weighted least squares for correlated Kappas.
Ninety-eight perimenopausal and postmenopausal women were recruited prospectively and consecutively during the 22-month enrollment period. Of these 98 consented patients, no saline-infusion sonography endometrial sampling specimens were obtained in seven patients owing to either withdrawal from the study before the saline-infusion sonography endometrial sampling (five of seven) or inability to perform sonohysterography secondary to cervical stenosis (two of seven). Saline-infusion sonography endometrial sampling specimens were insufficient in three patients. Therefore, 88 patients had diagnostic saline-infusion sonography endometrial sampling (Table 1). There were no complications incurred with saline-infusion sonography endometrial sampling.
Forty-five (51%) patients were perimenopausal (average age 48.5 years, range 40–53), and forty-three (49%) were postmenopausal (average age 59.7 years, range 48–85). The majority of our patients were Hispanic (47 of 88 or 53%), followed by African or African American (28 of 88 or 32%), white (11 of 88 or 12.5%), and Asian (two of 88 or 2%).
Saline-infusion sonography endometrial sampling specimens were obtained successfully in 88 patients. In contrast, only half of the patients had diagnostic endometrial biopsies (Table 1). Of the 47 nondiagnostic endometrial biopsies, 28 specimens (almost one third) were insufficient and 19 patients had no endometrial biopsy specimen at all. In comparing the ability of saline-infusion sonography endometrial sampling with endometrial biopsy to obtain tissue for diagnosis, saline-infusion sonography endometrial sampling provided a diagnosis 89% of the time (95% confidence interval [CI] 82–95%) compared with endometrial biopsy at 52% (95% CI 42–62%). This difference was statistically significant (P<.001).
Of the 88 saline-infusion sonography endometrial sampling specimens, we had final outcomes in 80 cases. Final outcomes constituted either a gold standard histological diagnosis obtained on hysteroscopy-D&C or hysterectomy specimen (46 patients) or 1-year follow-up in patients without final pathology who had benign saline-infusion sonography endometrial sampling or sonohysterography diagnoses (34 patients). Diagnoses were categorized as either 1) physiologic, which would include a normal, functional endometrium in a perimenopausal woman or atrophy in a postmenopausal woman, 2) benign polyp, or 3) hyperplasia/cancer. Of the 51 patients with diagnostic endometrial biopsies, we had final outcomes in 45 patients.
Table 2 compares endometrial biopsy with final outcomes. Endometrial biopsy overcalled physiologic lesions, categorizing 41 specimens as physiologic when final outcomes proved this diagnosis in only 26 cases, and grossly underperformed in the diagnosis of benign polyps, finding only one of 16 (6%). Endometrial biopsy underdiagnosed malignancies, including incorrectly categorizing a hyperplasia on final pathology as physiologic. The positive predictive value of endometrial biopsy was 61% (95% CI 45–76%), and the negative predictive value was 75% (95% CI 19–99%). The Kappa value (95% CI) was 0.18 (−0.03 to 0.39).
When comparing saline-infusion sonography endometrial sampling with the final outcomes as shown in Table 3, saline-infusion sonography endometrial sampling exhibited greater agreement than endometrial biopsy (Kappa 0.57, 95% CI 0.40–0.73). Saline-infusion sonography endometrial sampling correctly identified 20 of the 30 benign polyps and did not miss any of the four premalignant or malignant lesions. Ten saline-infusion sonography endometrial samples were categorized as hyperplasia/cancer, with only four of these confirmed on final histology (Table 4). Most of these lesions involved diffuse endometrial thickening or multiple masses or both (Fig. 1). There were six hyperplasias on saline-infusion sonography endometrial sampling that did not match the final pathology (Table 5). In contrast, these lesions were primarily localized with focal atypia or focal areas of thickening (Fig. 2). Saline-infusion sonography endometrial sampling had a positive predictive value of 80% (95% CI 66–91%) and a negative predictive value of 74% (95% CI 56–87%). Figure 3 compares the Kappa results of the techniques: endometrial biopsy (Kappa 0.18) did not perform as well as saline-infusion sonography endometrial sampling (Kappa 0.57). This difference was statistically significant (P<.005).
Multiple studies have demonstrated that the combination of blind endometrial biopsy and sonohysterography enhances the diagnostic accuracy of either procedure performed alone.12,13 Some have hypothesized that sampling the endometrium after sonohysterography and performing the biopsy under direct visualization of sono-guidance would improve the ability to obtain the tissue at the site of the endometrial abnormality.14–17
In our present study, endometrial biopsy proved inferior to saline-infusion sonography endometrial sampling in obtaining tissue for diagnosis (P<.001). Only half of the patients had diagnostic endometrial biopsy specimens. The diagnostic yield of saline-infusion sonography endometrial sampling was enhanced because, with this directed biopsy, we sampled the appropriate endometrial site as identified on sonohysterography associated with an increased risk of pathology. This focused sampling cannot be achieved with the traditional blind endometrial biopsy, which repeatedly has been shown to sample only a small portion of the endometrium and to miss many endometrial lesions.18
In addition, our study addresses the accuracy of sonohysterographically obtained biopsies as opposed to blind endometrial biopsy by direct comparison with final outcomes. This was not done in any of the previous studies.14–17
Our results demonstrate that endometrial biopsy performed poorly in diagnosing endometrial pathology in perimenopausal and postmenopausal women with abnormal uterine bleeding, particularly with focal lesions, missing 15 of the 16 endometrial polyps, which is similar to previous reports.2 It is more concerning, however, that endometrial biopsy misdiagnosed a premalignant lesion as physiologic. Because the Kappa value between endometrial biopsy and final outcome was only 0.18 with a 95% CI that included zero, the likelihood of the agreement between endometrial biopsy and the final outcome was no better than by chance alone.
In contrast to endometrial biopsy, saline-infusion sonography endometrial sampling performed better when compared with final outcomes. Saline-infusion sonography endometrial sampling did not miss the premalignant or malignant lesions. By sampling the site of the endometrial abnormality, saline-infusion sonography endometrial sampling not only improved the ability to diagnose endometrial pathology but also improved the accuracy of diagnosing endometrial disease, including cancer. Unlike endometrial biopsy, nothing called physiologic or benign by saline-infusion sonography endometrial sampling was found to be hyperplasia or cancer. Our study not only confirms the accuracy of saline-infusion sonography endometrial sampling with global endometrial disease but also demonstrates its accuracy in diagnosing focal endometrial lesions.
Saline-infusion sonography endometrial sampling found 10 hyperplasia/cancers, but only four of these were confirmed on gold standard pathology specimens. The six hyperplasias on saline-infusion sonography endometrial sampling that did not match the final pathology were primarily localized with focal atypia or focal areas of thickening. Four of these lesions ranged in size from only 5 mm to 11 mm. This occurrence is not dissimilar to cervical biopsies that show dysplasia only to have the conization specimen reveal no residual disease. Secondly, all but two final pathology diagnoses were obtained by hysteroscopy-D&C, which is known to miss lesions.19 With regard to the two patients who underwent hysterectomy, one patient was treated with progesterone therapy for 9 months before surgery, so the atrophy on final histology is certainly not reflective of the endometrium at the time of saline-infusion sonography endometrial sampling. The other patient had a focal thickening seen on sonohysterography and saline-infusion sonography endometrial sampling which was positive for focal glandular atypia and also p53 protein and Ki-67 immunoperoxidase stains, which often are expressed in endometrial cancers. Subsequently, we do not believe that these six discordant saline-infusion sonography endometrial sampling specimens were false positives, but in fact that sampling may have diagnosed and treated these focal lesions simultaneously.
From a technical standpoint, the addition of the saline-infusion sonography endometrial sampling added little time to the procedure (approximately 5 minutes). No procedure had to be terminated prematurely because of patient discomfort. In fact, of the 59 patients who completed postprocedural surveys, the majority of patients (45 of 59, 76%) stated that they had less discomfort with saline-infusion sonography endometrial sampling than during the previous endometrial biopsy. Sampling under sono-guidance guaranteed that the Uterine Explora II curette was positioned properly in the endometrial cavity before sampling. For this same reason, saline-infusion sonography endometrial sampling would seem safer than the traditional blind biopsy because of the direct visualization of the biopsy instrument during the sampling; in fact, no complications were incurred. It was difficult to maintain adequate uterine cavity distension during saline-infusion sonography endometrial sampling. Because coaxial placement of the biopsy instrument proved difficult, especially with postmenopausal patients, the biopsy instrument was introduced after removal of the sonohysterography catheter. Subsequently, some of the uterine distension was lost. However, even with a uterus empty of saline, the biopsy instrument still was visible and could be directed to the area of abnormality as identified during the sonohysterography.
In summary, our findings lead us to the conclusion that saline-infusion sonography endometrial sampling is superior to endometrial biopsy in diagnosing endometrial pathology in perimenopausal and postmenopausal women. As a result, we believe that the blind endometrial biopsy may no longer be considered an appropriate initial diagnostic modality for evaluating these women with abnormal uterine bleeding. We suggest that the management of abnormal uterine bleeding should commence with sonographic evaluation of the endometrial cavity, with subsequent sonohysterography and directed saline-infusion sonography endometrial sampling for abnormal endometrial cavities.
1. Stovall TG, Photopulos GJ, Poston WM, Ling FW, Sandles LG. Pipelle endometrial sampling in patients with known endometrial carcinoma. Obstet Gynecol 1991;77:954–6.
2. Guido RS, Kanbour-Shakir A, Rulin MC, Christopherson WA. Pipelle endometrial sampling. Sensitivity in the detection of endometrial cancer. J Reprod Med 1995;40:553–5.
3. Goldstein SR, Nachtigall M, Snyder JR, Nachtigall L. Endometrial assessment by vaginal ultrasonography before endometrial sampling in patients with postmenopausal bleeding. Am J Obstet Gynecol 1990;163:119–23.
4. Emanuel MH, Verdel MJ, Wamsteker K, Lammes FB. A prospective comparison of transvaginal ultrasonography and diagnostic hysteroscopy in the evaluation of patients with abnormal uterine bleeding: clinical implications. Am J Obstet Gynecol 1995;172:547–52.
5. Bronz L, Suter T, Rusca T. The value of transvaginal sonography with and without saline instillation in the diagnosis of uterine pathology in pre- and postmenopausal women with abnormal bleeding or suspect sonographic findings. Ultrasound Obstet Gynecol 1997;9:53–8.
6. Granberg S, Wikland M, Karlsson B, Norstrom A, Friberg LG. Endometrial thickness as measured by endovaginal ultrasonography for identifying endometrial abnormality. Am J Obstet Gynecol 1991;164:47–52.
7. Sheth S, Hamper UM, Kurman RJ. Thickened endometrium in the postmenopausal woman: sonographic-pathologic correlation. Radiology 1993;187:135–9.
8. Dueholm M, Forman A, Jensen ML, Laursen H, Kracht P. Transvaginal sonography combined with saline contrast sonohysterography in evaluating the uterine cavity in premenopausal patients with abnormal uterine bleeding. Ultrasound Obstet Gynecol 2001;18:54–61.
9. Widrich T, Bradley LD, Mitchinson AR, Collins RL. Comparison of saline infusion sonography with office hysteroscopy for the evaluation of the endometrium. Am J Obstet Gynecol 1996;174:1327–34.
10. Williams CD, Marshburn PB. A prospective study of transvaginal hydrosonography in the evaluation of abnormal uterine bleeding. Am J Obstet Gynecol 1998;179:292–8.
11. Epstein E, Ramirez A, Skoog L, Valentin L. Transvaginal sonography, saline contrast and hysteroscopy for the investigation of women with postmenopausal bleeding and endometrium ≥5 mm. Ultrasound Obstet Gynecol 2001;18:157–62.
12. Dubinsky TJ, Parvey HR, Gormaz G, Curtis M, Maklad N. Transvaginal hysterosonography: comparison with biopsy in the evaluation of postmenopausal bleeding. J Ultrasound Med 1995;14:887–93.
13. O’Connell LP, Fries MH, Zeringue E, Brehm W. Triage of abnormal postmenopausal bleeding: a comparison of endometrial biopsy and transvaginal sonohysterography versus fractional curettage with hysteroscopy. Am J Obstet Gynecol 1998;178:956–61.
14. Moschos E, Gibbs M, Twickler DM. Comparison of saline infusion sonography-guided endometrial sampling and endometrial biopsy in the diagnosis of endometrial pathology. Poster presentation P18.16. 16th World Congress of Ultrasound in Obstetrics and Gynecology, London (UK): September 2006.
15. Dubinsky TJ, Reed S, Mao C, Waitches GM, Hoffer EK. Hysterosonographically guided endometrial biopsy: technical feasibility. AJR Am J Roentgenol 2000;174:1589–91.
16. Bernard JP, Metzger U, Camatte S, Robin F, Vilde F, Taurelle R, et al. Comparison of three catheters for endometrial sampling during sonohysterography: results of a preliminary study. J Obstet Gynaecol 2002;22:84–5.
17. Wei AY, Schink JC, Pritts EA, Olive DL, Linheim SR. Saline contrast sonohysterography and directed extraction, resection and biopsy of intrauterine pathology using a Uterine Explora Curette. Ultrasound Obstet Gynecol 2006;27:202–5.
18. Clark TJ, Mann CH, Shah N, Khan KS, Song F, Gupta JK. Accuracy of outpatient endometrial biopsy in the diagnosis of endometrial cancer: a systematic quantitative review. BJOG 2002;109:313–21.
19. Epstein E, Ramirez A, Shoog L, Valentin L. Dilatation and curettage fails to detect most focal lesions in the uterine cavity in women with postmenopausal bleeding. Acta Obstet Gynecol Scand 2001;80:1131–6.
Figure. No caption available.