British Journal of Radiology (2007) 80, 719-723
© 2007 British Institute of Radiology
doi: 10.1259/bjr/87219886
Developing a robust and efficient pathway for the referral and investigation of women with post-menopausal bleeding using a cut-off of 
4 mm for normal thickness
S C Williams, FRCR
1
C Lopez, FRCR
2
A Yoong, FRCPath
3 and
J M McHugo, FRCR
2
1 Department of Radiology, Good Hope Hospital, Rectory Road, Sutton Coldfield B75 7RR, Departments of 2 Radiology and 3 Histopathology, Birmingham Women's Health Care Trust, Metchley Park Road, Edgbaston, Birmingham B15 2TG, UK
Correspondence: Dr Josephine McHugo, Consultant Radiologist, Radiology Department, Birmingham Women's Hospital, Metchley Park Road, Edgbaston, Birmingham B15 2TG, UK. E-mail: jo.mchugo{at}bwhct.nhs.uk
 |
Abstract
|
|---|
The majority of women with post-menopausal bleeding (PMB) do not have endometrial cancer, and transvaginal ultrasound (TVUS) is accepted as the screening tool of choice to identify those at highest risk of malignancy. A new referral pathway was instigated in 2000, with patients only being referred to a gynaecologist following an abnormal TVUS result. An initial audit revealed a low positive predictive value for TVUS and a low incidence of detected malignancy. The cut-off value for a normal endometrium was increased from 3 mm to 4 mm in line with published data. This re7hyphen;audit evaluated the effectiveness of implemented changes and identified other areas for improvement. Of the 277 women referred during the study period, 193 had an abnormal or unseen endometrium and were subsequently seen by a gynaecologist. For patients without a histological diagnosis, clinical notes were reviewed and the hospital cancer database scrutinized for all endometrial cancers. Despite a 15.4% increase in referrals, only an additional 2.9% were assessed by gynaecologists. 14 cases of malignancy were identified with a mean endometrial thickness (ET) of 15.7 mm. Failed endometrial sampling was more prevalent with a minimally thickened endometrium. 80 patients with abnormal TVUS had no recorded histological diagnosis. Increasing the ET cut-off value has reduced unnecessary investigations in women at low risk of malignancy. No woman discharged back to her GP has been diagnosed with an endometrial malignancy within 1 year of initial referral. An algorithm has been proposed to further improve the investigation of women with PMB following an abnormal TVUS.
|
Introduction
|
|---|
Post-menopausal bleeding (PMB) is defined as bleeding following a 12 month period of amenorrhoea. Although no pathological cause is found in the majority of women presenting with this symptom, endometrial cancer is eventually diagnosed in approximately 10% undergoing curettage [1]. Surgical curettage and hysteroscopy are considered the diagnostic gold standards, but in recent years transvaginal ultrasonography (TVUS) measuring endometrial thickness (ET) has been accepted as an indispensable tool in the investigation of patients presenting with PMB [2].
An audit was originally completed in 2002 to assess the effectiveness of a new referral system for women with PMB that had been introduced 2 years previously. This pathway resulted in women having an initial screening TVUS and then those with a normal result (endometrial thickness of 
3 mm) being referred back to their GPs with advice to examine the lower genital tract. Only those with an unseen or abnormal endometrium are referred to a hospital gynaecologist. The audit revealed a low positive predictive value for the TVUS examination compared with published data, a low incidence of endometrial malignancy and a high percentage of women discharged without a histological diagnosis. Hormone replacement therapy (HRT) showed no effect for exclusion of cancer by thickness. An action plan, agreed by the members of the multidisciplinary team, involved changing the cut-off value for normal ET to 4 mm and devising an investigative algorithm. The purpose of this audit was to evaluate the outcome of these changes and assess the overall management of women with PMB at this institution.
|
Patients and methods
|
|---|
Over a 6 month period in 2004, 277 women were referred directly to the Radiology Department with a history of post-menopausal bleeding and were seen within 2 weeks. Initially, a transabdominal examination of the pelvis was performed using a 4 MHz curvilinear transducer of an Acuson Sequoia machine to obtain a global view of the pelvic organs. Patients were then asked to empty their bladder and give verbal consent to a TVUS, which was performed using a 5–7.5 MHz endovaginal-sector transducer. The examinations were conducted by a consultant radiologist, specialist registrars and senior radiographers specializing in gynaecological imaging. All practitioners rigorously followed a departmental protocol for defining normal and abnormal examinations (Table 1
).
The combined ET of the two layers was measured (excluding any fluid) in the longitudinal plane and classified as normal if 4 mm and regular in outline with no focal abnormality. Patients with a thin, regular endometrium were reassured and referred directly back to their GP. The report issued to the GP emphasized the importance of examining the lower genital tract and advised rereferral should the patient's symptoms recur. All patients with an abnormal examination and those in whom the endometrium could not be adequately visualized were referred directly to the oncology clinic for a booked or same day appointment.
Histological results for the cohort of 193 patients referred for gynaecological opinion were searched for using the TELEPATH computer system. Searches were conducted using the patient's name, date of birth and hospital identification number separately. Patients had undergone Pipelle aspiration, hysteroscopy, dilatation and curettage or hysterectomy, but for 80 patients for whom no histological diagnosis had been recorded, the clinical notes were requested.
The hospital cancer register was searched for a minimum of 12 months following initial referral (maximum period of 18 months) to identify any patients initially discharged who subsequently were found to have endometrial cancer.
|
Results
|
|---|
The 277 patients seen during the 6 month study period represent a 15.4% increase in referrals compared with the same period 2 years previously. The original audit had revealed a low positive predictive value for TVUS examinations classified as abnormal, which was addressed by increasing the threshold for normality in endometrial thickness to 4 mm (previously 3 mm). This was reflected in the increase in the proportion of normal examinations from 22.5% to 30.3% seen in the current audit. By redefining the criteria for normality, only a 2.9% increase in patient referrals to gynaecology outpatients was observed, despite the 15.4% increase in GP referrals to this institution. 27 patients who would have previously been referred to the gynaecologists (ET = 4 mm) avoided further investigations. A table of comparative results from both audits is provided (Table 2).
Seven patients did not undergo a TVUS, representing 2.5% of the total patient population. Analysis of this group showed that one was 98 years old, one had learning difficulties and was felt unable to give informed consent, two were virgo intacta and one had a large cystocele. In two cases, the reason for not performing a TV examination was not recorded.
15 women were found to have undergone a previous hysterectomy and this investigative pathway was not optimal for them. In five patients an IUCD was present and the endometrial thickness could not be accurately assessed. In addition to the above, the endometrium could not be adequately visualized in a further six patients. All of these women were referred for clinical assessment.
167 patients had a thickened endometrium (mean 8.6 mm, range 5–28 mm) and a histological diagnosis was available in only 54.5% of these cases. In this subgroup there were 14 cancers, 24 cases of endometrial polyps and 6 patients with endometrial hyperplasia. Seven patients were found to have endocervical polyps (two in association with endometrial polyps). A summary of histological results is depicted in Table 3. As expected, the proportion of malignant samples increased with increasing endometrial thickness. The mean ET associated with malignancy was 15.7 mm with a range of 7–26 mm.
All 11 patients on tamoxifen treatment had a thickened endometrium (mean ET = 11.6 mm, range 7–22 mm). Five were found to have endometrial polyps; in two cases the endometrium demonstrated a hormonal response; one was classified as normal; and no histology was available for the remaining three patients.
Of the group of 80 patients for whom no histological diagnosis was recorded, 56 clinical records were available for review. Pipelle aspiration had been attempted in 34. In 10 patients the operator was unable to pass the Pipelle through the cervix and, in the remaining 24, no tissue or inadequate tissue was aspirated. A majority of these patients had ET of 5–7 mm. Eight patients had been directly referred for hysteroscopy. In 11 patients, no attempt at obtaining a histological sample had been made. The remaining three patients had declined further investigation. A summary of the outcomes of these 56 patients is depicted in Figure 1.
View larger version (14K):
[in this window]
[in a new window]
|
Figure 1. Management of patients without a recorded histological diagnosis on TELEPATH.(Ix = investigation, H, D+C = hysteroscopy, dilatation and curettage, US = ultrasound examination, FU = follow-up, NAD = no abnormality detected, HRT = hormone replacement therapy, DNA = did not attend).
|
|
The subgroup of women in whom no histological sampling was attempted included two patients with ET of 5 mm who were reassured and discharged without follow-up. Four patients had repeat TVUS examinations, one of which demonstrated an increase in ET. One patient on HRT was classified as breakthrough bleeding and another woman as being perimenopausal. Two patients with a stenosed cervical os and ET of 6 mm and 7 mm were discharged with no further follow-up.
Analysis of the clinical notes of the 56 women revealed that 21 (37.5%) eventually underwent the gold-standard procedure of hysteroscopy, dilatation and curettage. However, 24 women with an endometrial thickness of 5 mm or greater were discharged without an adequate histological diagnosis or undergoing hysteroscopy (five cases because of patient non-compliance).
No patients with an initial TVUS classed as "normal" were diagnosed with an endometrial cancer within 12 months of initial referral (this time being chosen to complete the audit).
|
Discussion
|
|---|
There is persistent debate regarding the most appropriate methods for investigating and managing PMB. TVUS is generally accepted as an appropriate first-line investigation to identify patients presenting with post-menopausal bleeding who are at higher risk of endometrial carcinoma. It is a relatively non-invasive procedure with good negative predictive value if properly performed [3]. The risk of malignancy increases as the endometrial thickness increases. It is accepted that in a majority of patients no organic cause for the bleeding will be found. Bleeding in HRT users is less likely to be the result of a gynaecological malignancy.
The cut-off point for defining a normal endometrium has to be kept to a sufficiently low level to maintain sensitivity, but specificity will also be low when one ideally wants to minimize more invasive diagnostic procedures. A prospective study by Malinova and Pehlivanov [4] using TVUS to correlate ET with histopathology showed a sensitivity of 100% and specificity of 64% when a 6 mm cut off was used. All endometrial carcinomas in this group were found in patients with ET 
7 mm, which led the authors to suggest that it was reasonable not to perform dilatation and curettage if the ET 6 mm. However, advanced endometrial carcinoma has been known to occur in cases without appreciable endometrial thickening on ultrasound [5]. Several studies have used cut-off values of either 4 mm or 5 mm and a meta-analysis published in 2002 showed that a negative test result reduced the post-test probability of carcinoma from 14% to 1.2% with a 4 mm cut off and 2.3% at 5mm (a less than sixfold reduction) [6]. The value of TVUS is to identify those patients who will require further investigation and reduce the use of costly, invasive investigations in those with a very low probability of malignant disease.
Colour Doppler is no longer believed to clarify whether a thickened endometrium is the result of benign pathology or a malignancy. Sheth et al [7] conducted a study of 45 post-menopausal women with ET 8 mm on TVUS. All underwent colour duplex Doppler endovaginal ultrasound of the endometrium. Pulsatility and resistive indices were recorded from arterial waveforms generated in areas of increased vascularity. They concluded that low-impedance arterial flow was seen in various diseases with significant overlap between benign and malignant conditions.
Malignancy has been estimated to occur in 10% of those undergoing curettage [8] but our malignancy rate was 7.2% in patients with a thickened endometrium. Assuming a similar patient population, this suggests that we may still be missing malignancies and that our investigative pathway is not suitably robust. A prospective study to identify those women who are re-referred with PMB or subsequently present with endometrial cancer would clarify this.
The Pipelle endometrial sampler is a relatively cheap, safe and simple device that can be used in the outpatient setting. However, it is a blind procedure and focal endometrial abnormality cannot be targeted. Success can be improved when this technique is combined with TVUS. A retrospective study by Farrell et al [9] found that, of 141 patients with a Pipelle biopsy reported as "insufficient for diagnostic purposes", 20% were found to have uterine pathology after secondary investigation. Uterine malignancy was eventually diagnosed in 3% of this group. They concluded that an inadequate or negative Pipelle sample should not be considered reassuring but an indication for further investigation since serious underlying uterine pathology may be present.
Endometrial sampling is less successful in women with a thin endometrium and this is reflected in the relatively high percentages of women having no histological diagnosis following Pipelle sampling (69% in the 5 mm group, 48% in the 6–10 mm group).
Within this hospital, currently no universal algorithm exists for the investigation of women referred with post-menopausal bleeding and a positive ultrasound. This is graphically illustrated by the variety of secondary investigations in use and methods of follow-up in the 56 patients with "insufficient" or no Pipelle sample (Figure 1
). Other studies have similarly found a great diversity in the management of women with PMB [10]. We suggest that repeating the Pipelle aspiration only serves to delay other investigations that are likely to be more conclusive. Hysteroscopy in combination with dilatation and curettage is accepted as the "gold standard" and should be offered to all patients with thickened endometrium and an inconclusive Pipelle biopsy. In those patients in whom a more "wait and see" approach is appropriate, they should be followed up with repeat TVUS and undergo hysteroscopy if they have a further episode of bleeding or the ultrasound findings progress. TVUS follow-up at 4 months and a year from initial presentation is based on current practice but the value of re-screening at these time intervals to detect significant pathology should be an area for future audit. A proposed algorithm is presented in Figure 2. Patients with re-bleeding following a normal TVUS are referred directly to a gynaecologist as they represent a subgroup of patients with a higher likelihood of significant pathology. Patients with a normal endometrial thickness but with an abnormal TVUS examination (defined in Table 1) were also seen by a gynaecologist. None of these patients was diagnosed with an endometrial carcinoma within the subsequent 12 months.
This study has several limitations. Follow-up of patients was limited to 12 months. At the time of the audit, the hospital was the main regional centre for gynaecological malignancy. An assumption was made that the majority of women with recurrent symptoms would be referred back to this institution. However, data from women presenting to other hospitals within the region or moving out of region would not have been obtained and future audits should address this with reference to the regional cancer registry and follow-up from GPs.
The referral mechanism for women with PMB is efficient, with only those requiring further investigation seen by gynaecologists. We have successfully increased the percentage of endometrial malignancies diagnosed, reduced unnecessary investigations in women with a low risk of endometrial cancer and shielded the gynaecology outpatient department from increased GP referrals. The audit has identified further areas for improvement addressed by the investigation algorithm.
Received for publication April 4, 2006.
Revision received September 20, 2006.
Accepted for publication November 27, 2006.
|
References
|
|---|
- MacKenzie IZ, Bibby JG. Critical assessment of dilatation and curettage in 1029 women. Lancet 1978;ii:566–8.
- Gull B, Carlsson MD, Karlsson B, Ylöstalo P, Milsom I, Granberg S. Transvaginal ultrasonography of the endometrium in women with postmenopausal bleeding: Is it always necessary to perform an endometrial biopsy? Am J Obstet Gynecol 2000;182:509–15.[CrossRef][Medline]
- Gull B, Karlsson B, Milsom I, Granberg S. Can ultrasound replace dilatation and curettage? A longitudinal evaluation of postmenopausal bleeding and transvaginal sonographic measurement of the endometrium as predictors of endometrial cancer. Am J Obstet Gynecol 2003;188:401–8.[CrossRef][Medline]
- Malinova M, Pehlivanov B. Transvaginal sonography and endometrial thickness in patients with postmenopausal uterine bleeding. Eur J Obstet Gynecol Reprod Biol 1995;58:161–5.
- Buyuk E, Durumusoglu F, Erenus M, Karakoc B. Endometrial disease diagnosed by transvaginal ultrasound and dilatation and curettage. Acta Obstet Gynecol Scand 1999;78:419–22.[CrossRef][Medline]
- Gupta JK, Chien PF, Voit D, Clark J, Khan KS. Ultrasonographic endometrial thickness for diagnosing endometrial pathology in women with post-menopausal bleeding: a meta-analysis. Acta Obstet Gynecol Scand 2002;81:799–816.[CrossRef][Medline]
- Sheth S, Hamper UM, McCollum ME, Caskey CI, Rosenshein NB, Kurman RJ. Endometrial blood flow analysis in post-menopausal women: can it help differentiate benign from malignant causes of endometrial thickening? Radiology 1995;195:661–5.[Abstract/Free Full Text]
- Grimes DA. Diagnostic dilatation and curettage: a reappraisal. Am J Obstet Gynecol 1982;142:1–6.[Medline]
- Farrel T, Jones N, Owen P, Baird A. The significance of an "insufficient" pipelle sample in the investigation of post-menopausal bleeding. Acta Obstet Gynaecol Scand 1999;78:810–12.[CrossRef][Medline]
- Epstein E. Management of postmenopausal bleeding in Sweden: a need for increased use of hydrosonography and hysteroscopy. Acta Obstet Gynaecol Scand 2004;83:89–95.[CrossRef][Medline]
This article has been cited by other articles:
|
|
|
|
 
K. Canfell, Y. J. Kang, M. Clements, A. M. Moa, and V. Beral
Normal endometrial cells in cervical cytology: systematic review of prevalence and relation to significant endometrial pathology
J Med Screen,
December 1, 2008;
15(4):
188 - 198.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
