Value of endometrial thickness measurement for diagnosing focal intrauterine pathology in women without abnormal uterine bleeding

Transcription

1 Ultrasound Obstet Gynecol 29; 33: Published online in Wiley InterScience ( DOI: 1.12/uog.6256 Value of endometrial thickness measurement for diagnosing focal intrauterine pathology in women without abnormal uterine bleeding E. DREISLER*, S. STAMPE SORENSEN*, P. H. IBSEN and G. LOSE* Departments of *Obstetrics and Gynecology and Pathology, Glostrup Hospital, University of Copenhagen, Glostrup, Denmark KEYWORDS: accuracy; endometrial polyps; hydrosonography; ROC curves; saline contrast sonohysterography; sensitivity; specificity; transvaginal sonography; uterine polyps ABSTRACT Objective To assess the diagnostic value of transvaginal sonographic (TVS) measurement of endometrial thickness for diagnosing focal intrauterine pathology in women without abnormal uterine bleeding (AUB). Methods A random selection from the Danish Civil Registration System was made: 166 women aged 2 74 years were invited to participate and 686 women were eligible and accepted inclusion (429 pre- and 257 postmenopausal). The women underwent TVS measurement of endometrial thickness and saline contrast sonohysterography (SCSH). Hysteroscopic resection with histopathology (gold standard) was performed when focal intrauterine pathology was suspected at SCSH. We excluded women with AUB (n = 237), failure of SCSH (n = 5), a scan that was not in the follicular phase (n = 11), hysteroscopy contraindicated (n = 2), and users of sequential hormone therapy (n = 9) or selective estrogen receptor modulators (n = 2). Thus, 375 women without AUB were included (217 pre- and 158 postmenopausal). Receiver operating characteristics (ROC) curves for endometrial thickness and focal lesion were analyzed. Results Focal intrauterine pathology was confirmed in 41 women (35 with polyps, five with submucosal myomas and one with polypoidal growing cancer). For premenopausal women, the area under the ROC curve (AUC) was.79 (95% CI,.68.89) and for postmenopausal women it was.84 (95% CI,.76.92). For premenopausal women, the best negative likelihood ratio (LR =.11) was obtained at an endometrial thickness of 5.2 mm, with a negative predictive value (NPV) of 99% and a positive predictive value (PPV) of 1%. For postmenopausal women the best LR (.8) was obtained at an endometrial thickness of 2.8 mm, with a NPV of 99% and a PPV of 26%. Conclusions In women without AUB, TVS measurement of endometrial thickness is a poor diagnostic test, but is apparently efficacious in excluding focal intrauterine pathology, especially in postmenopausal women. The 4 5-mm threshold conventionally used to exclude endometrial malignancy in women with postmenopausal bleeding is not transferable to women without AUB for excluding focal intrauterine pathology. Copyright 29 ISUOG. Published by John Wiley & Sons, Ltd. INTRODUCTION Transvaginal sonography (TVS) is a commonly used supplement to the general gynecological examination. Requirements from patients, easy access to equipment and legal issues are among the explanations for this. Incidental visualization of the endometrium is possible when diseases in the pelvic area are explored. Sonographic measurement of endometrial thickness is documented to be a valuable primary tool in evaluating patients with postmenopausal bleeding 1 3. In premenopausal women with abnormal uterine bleeding (AUB), analysis of the test performance of TVS and cut-off levels has produced divergent results 4 9. In women without AUB, when a thick endometrium is discovered there is a clinical dilemma as to which strategy should be chosen. The value of endometrial thickness measurement in these women has not been documented. This study aimed to assess the diagnostic value of TVS measurement of the endometrium for detection of focal intrauterine pathology in women without AUB. Correspondence to: Dr E. Dreisler, Department of Obstetrics and Gynecology, Glostrup Hospital, University of Copenhagen, Ndr. Ringvej, Dk-26 Glostrup, Denmark ( dreisler@dadlnet.dk) Accepted: 9 June 28 Copyright 29 ISUOG. Published by John Wiley & Sons, Ltd. ORIGINAL PAPER

2 TVS and focal lesions 345 METHODS Subjects The initial random selection (using a random number generator), from the Danish Civil registry system, of 166 women aged 2 74 years has been described elsewhere 1. Of these, 686 women, with a median age of 45 years, were eligible and accepted inclusion (429 pre- and 257 postmenopausal). Their symptoms were evaluated with a validated questionnaire 1, which included questions on medical history, menstrual bleeding patterns and use of hormones, including oral contraceptives and tamoxifen. The information obtained from the questionnaire was verified at the clinical examination for those included. We excluded the 237 women who had AUB, as well as those treated with medication that increases endometrial thickness 11,12 (sequential hormone therapy (n = 9), tamoxifen (n = 1) and raloxifen (n = 1)), those in whom SCSH failed (n = 5) and those in whom hysteroscopy was contraindicated (n = 2). Among premenopausal women, those not scanned in the follicular phase (menstrual cycle day 5 12) were excluded (n = 11). Thus, 375 women underwent full evaluation and were included in the final analysis (217 premenopausal and 158 postmenopausal). All participants gave written consent according to Helsinki declarations and The Scientific Ethics Committee of the County of Copenhagen (KA283) and the Danish Data Protection Agency approved the study. Sonography Ultrasound examinations were performed using a Hitachi EUB 6 ultrasound system (Hitachi Medical Corporation, Tokyo, Japan) with a MHz transvaginal probe by two researchers (E.D. and S.S.S.); E.D. is a research fellow who had had 2 years practice in gynecology before the examinations began and S.S.S. is a gynecologist with more than 2 years experience of ultrasound and hysteroscopic surgery. To optimize the quality and achieve consistent interpretation of the ultrasound examinations, both investigators were present at the first 1 examinations and in difficult cases. During TVS, the thickest part of the anteroposterior bilayer endometrial thickness was measured in the sagittal plane. Echogenicity of the endometrium was noted, but no tentative diagnoses were made. Digital images were transferred to a computerized database. For SCSH, a speculum was inserted into the vagina and the cervix was washed with antiseptic solution. Uterine cavities were visualized by inserting isotonic saline through a 1-cm, 8-charrière baby feeding tube (Unomedical, Hundested, Denmark) connected to a 2-mL syringe. In women with a narrow cervical orifice, a 5-charrière baby feeding tube was used. Generally, a tenaculum was not used. Uterine cavities were examined in a longitudinal plane from right to left uterine cornu and in a transverse plane from fundus to cervix. All examinations were saved as videoclips. A negative SCSH was defined as a uterine cavity with a well-defined endometrium and no echodense foci exceeding 5 mm in maximum diameter. Polyps were suspected when smooth-margined echogenic protrusions with fairly homogeneous texture were seen at SCSH. Submucosal myomas were suspected on visualization of a solid round structure of mixed echogenicity protruding into the uterine cavity and covered by endometrium on the surface. In cases with pedunculated myomas and myomas with intramural extension of < 5%, hysteroscopy was performed, with final diagnosis being made at histopathology. When SCSH was inconclusive, the procedure was repeated (in 21 cases). Complications of SCSH occurred in six women (1.6%): three had a vasovagal attack during the procedure, and two had lower abdominal pain and one had foul-smelling discharge afterwards. No cases of pelvic inflammatory disease occurred. Hysteroscopic resection and analysis of specimens In women with focal intrauterine pathology exceeding 5 mm in diameter at SCSH, hysteroscopic resection was performed using a Versapoint bipolar (Johnson and Johnson, New Brunswick, NJ, USA) or an Olympus monopolar (Olympus, Hamburg, Germany) resectoscope with the patient under either local or general anesthesia. The hysteroscopist was aware of the findings at SCSH. Nine women had very small lesions of < 5 mm. The location and number of intrauterine projections were recorded. Focal lesions were resected in one piece or in several large pieces able to pass through the cervical canal. One woman developed a false passage at hysteroscopy, but no other complications occurred. The same pathologist, who was blinded to the endometrial measurement, examined all histopathological specimens and characterized the findings according to the World Health Organization s criteria 13. The gold standard was the histology report. Statistical analysis Continuous data are reported as median, range and 2.5 th and 97.5 th percentiles. The chi-square test was used for univariate analysis, and the Mann Whitney U-test for numerical, independent data. P <.5 was considered statistically significant. Receiver operating characteristics (ROC) curves were prepared (plot of sensitivity vs. 1 specificity) and the areas under the curves (AUC) estimated. AUC = 1 indicates a perfect test, AUC >.9 indicates high accuracy and AUC between.7 and.9 indicates moderate accuracy 14., specificity, positive predictive value (PPV), negative predictive value (NPV), and likelihood ratio (LR) were calculated. A LR is independent of the prevalence of the condition studied and provides information on post-test probability. To determine a cut-off level for endometrial thickness for excluding focal intrauterine pathology, the negative LRs (LR ) were evaluated. Positive LRs (LR+) were also evaluated. In general, LR <.1 orlr+ > 1 generate

3 346 Dreisler et al. large changes from pretest probability, LR =.1.2or LR+ =2 5 generate moderate changes, and LR >.2 or LR+ < 2 generate small changes in probability 15. Statistical analysis was performed using the SAS software package (Release 9.1; SAS Institute Inc., Cary, NC, USA). RESULTS The 217 premenopausal women included in the study were aged (median, 38) years and the 158 postmenopausal women were (median, 59) years. Eighty of the premenopausal women used oral contraceptives, and 15 of the postmenopausal women used continuous hormone therapy. Focal intrauterine pathology was confirmed in 7.4% (16/217) of premenopausal and (15.8%) (25/158) of postmenopaual women, and included 35 cases of polyps, five of submucosal myomas and one of polypoidal growing endometrial cancer. SCSH was false positive in two cases, in which normal endometrium was diagnosed at histopathological examination of the resected specimen. Endometrial thicknesses are given in Table 1. Endometrial thickness was affected by the vertical diameter of polyps in broad-based polyps, and by the horizontal diameter of stalky polyps and polyps arising from the fundal area. The contribution of polyp size to endometrial thickness is illustrated in Figure 1., specificity, PPV, NPV and LRs at different endometrial thickness cut-off values for pre- and postmenopausal women are given in Tables 2 and 3, respectively, and the ROC curves are shown in Figure 2. For premenopausal women, the AUC was.79 (95% CI,.68.89) and for postmenopausal women it was.84 (95% CI,.76.92). In postmenopausal women, when a maximum endometrial thickness of 5 mm was selected as the cut-off for a positive test, the nosographic sensitivity was 56%, and the nosographic specificity was 88%. The PPV was 47%, the NPV was 91%, and the LR was.5. If the aim is to exclude a focal intrauterine lesion, one must examine the LR. In these women the best LR (.8 (95% CI,.1.57)) was obtained at an endometrial thickness of 2.8 mm; at this limit the NPV of endometrial Table 1 Endometrial thickness in 375 women without abnormal uterine bleeding Group n Median Range Endometrial thickness (mm) 2.5 th centile 97.5 th centile No focal lesion Premenopausal* Postmenopausal With focal lesion Premenopausal* Postmenopausal *Menstrual cycle day Postmenopausal women using sequential hormone therapy were excluded. Endometrial thickness (mm) Polyp size (mm) 15 2 Figure 1 Scatterplot showing largest anteroposterior endometrial thickness at transvaginal ultrasound and contribution from polyp size. Polyps were measured at saline contrast sonohysterography. thickness measurement for excluding focal intrauterine pathology was 98.5% (95% CI, ), implying that 98.5% of women with an endometrial thickness measurement below 2.8 mm had no focal lesion. If the aim is to detect focal lesions, one must examine the LR+. The best LR+ (12.41 (95% CI, ) was obtained at an endometrial thickness of mm, with a PPV of 7.%, i.e. the likelihood of having a focal lesion was 7% if the endometrial thickness was above 8 mm. For premenopausal women, the best LR (.11 (95% CI,.2.74)) was obtained at endometrial thickness Table 2 Diagnostic performance with regard to focal intrauterine pathology of transvaginal ultrasound measurement of endometrial thickness in 217 premenopausal women without abnormal uterine bleeding ET cut-off (mm) PPV NPV Specificity LR+ LR (7. 18.) 98.8 ( ) 93.8 ( ) 41.8 ( ) 1.6 ( ).15 (.2 1.) ( ) 99.1 ( ) 93.8 ( ) 53.7 ( ) 2.3 ( ).12 (.2.78) ( ) 99.1 ( ) 93.8 ( ) 56.2 ( ) 2.14 ( ).11 (.2.74) ( ) 95.8 ( ) 62.5 ( ) 68.7 ( ) 2. ( ).55 ( ) ( ) 95.8 ( ) 56.3 ( ) 8.1 ( ) 2.83 ( ).55 (.31.96) ( ) 95.1 ( ) 43.8 ( ) 88.6 ( ) 3.82 ( ).64 (.41.98) ( ) 94.6 ( ) 31.3 ( ) 95.5 ( ) 6.98 ( ).72 (.52 1.) ( ) 94.1 ( ) 25. ( ) 95.5 ( ) 5.58 ( ).79 ( ) ET, endometrial thickness; LR+, positive likelihood ratio; LR, negative likelihood ratio; NPV, negative predictive value; PPV, positive predictive value.

4 TVS and focal lesions 347 Table 3 Diagnostic performance with regard to focal intrauterine pathology of transvaginal ultrasound measurement of endometrial thickness in 158 postmenopausal women without postmenopausal bleeding ET cut-off (mm) PPV NPV Specificity LR+ LR ( ) 98.5 ( ) 96. ( ) 48.1 ( ) 1.85 ( ).8 (.1.57) ( ) 97.4 ( ) 92. ( ) 55.6 ( ) 2.7 ( ).14 (.4.54) ( ) 94.5 ( ) 76. ( ) 78.2 ( ) 3.49 ( ).31 (.16.62) ( ) 91.4 ( ) 56. ( ) 88. ( ) 4.66 ( ).5 (.32.78) ( ) 9.5 ( ) 48. ( ) 93.2 ( ) 7.9 ( ).56 (.38.82) ( ) 88.7 ( ) 36. ( ) 94.7 ( ) 6.84 ( ).68 (.5.91) ( ) 87.8 ( ) 28. ( ) 97.7 ( ) ( ).74 (.57.94) ( ) 87.8 ( ) 28. ( ) 97.7 ( ) ( ).74 (.57.94) ( ) 87.2 ( ) 24. ( ) 97.7 ( ) 1.64 ( ).78 (.62.97) ET, endometrial thickness; LR+, positive likelihood ratio; LR, negative likelihood ratio; NPV, negative predictive value; PPV, positive predictive value Specificity Figure 2 Receiver operating characteristics curve for the relationship between transvaginal sonographic endometrial thickness and detection of intrauterine focal pathology in 217 premenopausal ( ) and 158 postmenopausal (...) women. Area under curve (AUC) for premenopausal women =.79 (95% CI,.68.89) and AUC for postmenopausal women =.84 (95% CI,.76.92). of 5.2 mm, with a NPV of 99.1%, indicating that focal lesion can be excluded with a likelihood of 99% when endometrial thickness is below this threshold. The best LR+ (6.98 (95% CI, )) was obtained at an endometrial thickness of 8.9 mm, with a PPV of 35.7%. When women with polyps < 5 mm in any diameter were excluded from analyses, the results did not differ markedly. DISCUSSION To the best of our knowledge, this is the largest general population studied so far in the evaluation by TVS and SCSH of women without AUB. With respect to the diagnostic performance of sonographic endometrial thickness measurement for excluding focal intrauterine pathology, the test was generally of moderate accuracy: AUCs were.79 and.84 for pre- and postmenopausal women, respectively. Our data indicate that measurement of endometrial thickness by TVS alone does have some value, particularly when endometrial thickness is < 5.2 mm in premenopausal women and < 2.8 mm in postmenopausal women; at these thresholds the best LR and NPVs (about 99%) were obtained. PPVs were consistently low (1 36%) in premenopausal women, indicating that measurement of endometrial thickness alone is not a useful test for diagnosing focal intrauterine pathology in these women, even when they are examined during the follicular phase of the menstrual cycle. For postmenopausal women, the test performed somewhat better; the best PPV was 7% at an endometrial thickness 8 mm. In asymptomatic women (without AUB), focal intrauterine pathologies are most frequently benign endometrial polyps 1, with no need for therapy. However, focal intrauterine lesions may in very rare cases harbor endometrial cancer. Consequently, the detection of a thick endometrium in an asymptomatic woman raises a clinical dilemma as to which strategy should be chosen. Due to the high false-positive rate of endometrial thickness measurement on TVS, there would be a substantial risk of overtreatment, with consequences such as patient anxiety, potential complications from surgical intervention and wasted resources for the healthcare system. This is in agreement with the finding that screening for endometrial cancer is not cost effective 16,17. Furthermore, endometrial cancer usually presents with uterine bleeding at an early and non-advanced stage. Thus, in most cases the clinical strategy should be restrictive; however, in cases with obvious risk factors for endometrial cancer (eg. old age, obesity, long-term treatment with unopposed estrogens) a more active strategy is warranted. The 4 5-mm limits generally used 1 3 for excluding endometrial malignancy in symptomatic postmenopausal women were not transferable to women without AUB for

5 348 Dreisler et al. exclusion of focal lesion. In line with a previous study of symptomatic postmenopausal women 18, we obtained the lower optimal endometrial thickness cut-off value of 2.8 mm; however, at this value the PPV declined as the NPV increased, leaving the clinical problem as to what to do when the endometrium exceeds the cut-off value. In asymptomatic postmenopausal women, benign focal lesions will be the main finding; in these women, the prevalence of endometrial polyps is 12% and the risk of endometrial cancer is low 1. In our series, the best LR+ (12.41) for focal intrauterine lesion was obtained at an endometrial thickness of 8 mm. Based on a theoretical cohort, a cut-off value of 11 mm for endometrial biopsy has been suggested for asymptomatic postmenopausal women 19. However, the optimal strategy in asymptomatic women with a finding of a thick endometrium still needs to be established, and the combination of TVS/SCSH with evaluation of endometrial texture, echogenicity and Doppler findings may be useful to exclude malignancy. TVS endometrial thickness in cases with polyps was correlated positively with polyp size measured at SCSH (Figure 1). Some polyps seemed to be compressed in the uterus, and after installation of saline, the polyp distended. Therefore, the polyp s size in some cases exceeded the endometrial thickness. We intended to optimize the study design by timing the ultrasound examination according to the menstrual cycle and excluding from analysis women treated with sequential hormone therapy. However, we did not sample the endometrium in all women, leaving undiagnosed possible hyperplasias and cancers with diffuse growth; this may have affected our results. Failure of SCSH or lack of hysteroscopy led to the exclusion of 52 (mainly postmenopausal) women; this is a source of verification bias. Finally, there were nine women with lesions < 5mm in diameter at SCSH; these were judged to be normal according to our protocol, but no gold standard was available for these women. In conclusion, in women without AUB, the TVS measurement of endometrial thickness was useful for exclusion of focal intrauterine pathology. However, when a thick endometrium is present, the diagnostic value of this measurement is limited. The 4 5-mm threshold used conventionally in symptomatic postmenopausal women is not transferable to women without AUB for excluding focal intrauterine lesion. REFERENCES 1. Karlsson B, Granberg S, Wikland M, Ylostalo P, Torvid K, Marsal K, Valentin L. Transvaginal ultrasonography of the endometrium in women with postmenopausal bleeding a Nordic multicenter study. Am J Obstet Gynecol 1995; 172: Smith-Bindman R, Kerlikowske K, Feldstein VA, Subak L, Scheidler J, Segal M, Brand R, Grady D. Endovaginal ultrasound to exclude endometrial cancer and other endometrial abnormalities. JAMA 1998; 28: Gupta JK, Chien PF, Voit D, Clark TJ, Khan KS. Ultrasonographic endometrial thickness for diagnosing endometrial pathology in women with postmenopausal bleeding: a metaanalysis. Acta Obstet Gynecol Scand 22; 81: Dijkhuizen FP, Brolmann HA, Potters AE, Bongers MY, Heinz AP. The accuracy of transvaginal ultrasonography in the diagnosis of endometrial abnormalities. Obstet Gynecol 1996; 87: Vercellini P, Cortesi I, Oldani S, Moschetta M, De Giorgi O, Crosignani PG. The role of transvaginal ultrasonography and outpatient diagnostic hysteroscopy in the evaluation of patients with menorrhagia. Hum Reprod 1997; 12: Goldstein SR, Zeltser I, Horan CK, Snyder JR, Schwartz LB. Ultrasonography-based triage for perimenopausal patients with abnormal uterine bleeding. Am J Obstet Gynecol 1997; 177: Schwarzler P, Concin H, Bosch H, Berlinger A, Wohlgenannt K, Collins WP, Bourne TH. An evaluation of sonohysterography and diagnostic hysteroscopy for the assessment of intrauterine pathology. Ultrasound Obstet Gynecol 1998; 11: Dueholm M, Jensen ML, Laursen H, Kracht P. Can the endometrial thickness as measured by transvaginal sonography be used to exclude polyps or hyperplasia in pre-menopausal patients with abnormal uterine bleeding? Acta Obstet Gynecol Scand 21; 8: Getpook C, Wattanakumtornkul S. Endometrial thickness screening in premenopausal women with abnormal uterine bleeding. J Obstet Gynaecol Res 26; 32: Dreisler E, Stampe Sorensen S, Ibsen PH, Lose G. Prevalence of endometrial polyps and abnormal uterine bleeding in a Danish population aged 2 74 years. Ultrasound Obstet Gynecol 29; 33: Affinito P, Palomba S, Pellicano M, Sorrentino C, Di Carlo C, Morgera R, Arienzo MP, Nappi C. Ultrasonographic measurement of endometrial thickness during hormonal replacement therapy in postmenopausal women. Ultrasound Obstet Gynecol 1998; 11: Van den Bosch T, Van Schoubroeck D, Ameye L, De Brabanter J, van Huffel S, Timmerman D. Ultrasound assessment of endometrial thickness and endometrial polyps in women on hormonal replacement therapy. Am J Obstet Gynecol 23; 188: World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of the Breast and Female Genital Organs. Tavassoli FA, Devilee P (eds). IARC Press: Lyon, 23; Swets JA. Measuring the accuracy of diagnostic systems. Science 1988; 24: Jaeschke R, Guyatt GH, Sackett DL. Users guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The Evidence-Based Medicine Working Group. JAMA 1994; 271: Gull B, Karlsson B, Milsom I, Wikland M, Granberg S. Transvaginal sonography of the endometrium in a representative sample of postmenopausal women. Ultrasound Obstet Gynecol 1996; 7: Gerber B, Krause A, Muller H, Reimer T, Kulz T, Kundt G, Friese K. Ultrasonographic detection of asymptomatic endometrial cancer in postmenopausal patients offers no prognostic advantage over symptomatic disease discovered by uterine bleeding. Eur J Cancer 21; 37: Van den Bosch T, Vandendael A, Van Schoubroeck D, Wranz PA, Lombard CJ. Combining vaginal ultrasonography and office endometrial sampling in the diagnosis of endometrial disease in postmenopausal women. Obstet Gynecol 1995; 85: Smith-Bindman R, Weiss E, Feldstein V. How thick is too thick? When endometrial thickness should prompt biopsy in postmenopausal women without vaginal bleeding. Ultrasound Obstet Gynecol 24; 24: