The outcome of endometrial carcinoma surveillance by ultrasound scan in women at risk of hereditary nonpolyposis colorectal carcinoma and familial colorectal carcinoma

Authors


Abstract

BACKGROUND

Endometrial carcinoma is the most common extracolonic malignancy associated with hereditary nonpolyposis colorectal carcinoma syndrome (HNPCC). The risk of endometrial carcinoma in HNPCC mutation carriers is approximately ten times that of the general population, and endometrial ultrasound surveillance to detect early cancer in asymptomatic individuals is recommended by the International Collaborative Group on HNPCC. There is little, if any, published data addressing the effectiveness of surveillance in HNPCC and familial colorectal carcinoma.

METHODS

The outcomes of endometrial carcinoma surveillance scans were collected from the St Mark's Hospital Imperial Cancer Research Fund Family Cancer Clinic in the UK and the Netherlands Foundation for the Detection of Hereditary Tumors. Two hundred ninety two women from HNPCC (171) or HNPCC-like (98) families between the ages of 25-65 years were offered pelvic ultrasound surveillance scans for a period of up to 13 years.

RESULTS

Results were available from 269 women. The study period included a total of 825.7 years of risk. Two cases of endometrial carcinoma were reported. Neither case was detected by surveillance scanning. Both cases presented at an early stage with symptoms and were subsequently cured.

CONCLUSIONS

Endometrial carcinoma surveillance in hereditary colorectal carcinoma may not offer obvious clinical benefits. Cancer 2002;94:1708–12. © 2002 American Cancer Society.

DOI 10.1002/cncr.10380

Hereditary nonpolyposis colon carcinoma (HNPCC, or Lynch syndrome) is an autosomal dominant syndrome caused by mutations in DNA mismatch repair genes which predispose individuals to carcinoma of the colon and a number of extracolonic carcinomas, including endometrial, small bowel, and urothelial carcinomas. Endometrial carcinoma has been identified as the most common extracolonic cancer in the syndrome.1, 2 The lifetime risk of endometrial carcinoma in female HNPCC gene carriers is approximately 40%; consequently surveillance has been suggested for asymptomatic cases.3, 4

In the general population, endometrial carcinoma is the most common gynecological cancer. Although it has a high incidence rate, it usually presents in the sixth and seventh decades of life with postmenopausal bleeding and has a favorable survival rate. In HNPCC, cases of endometrial carcinoma present at an earlier age, with the peak incidence in the fourth and fifth decades and almost all cases occurring by age 65 years. Death from endometrial carcinoma itself occurs in 12% of women with HNPCC affected with endometrial carcinoma.2

Although surveillance for endometrial carcinoma is recommended by the International Collaborative Group on HNPCC (ICG-HNPCC), there are no prospective studies that demonstrate that it is effective. The current recommendations are for annual or biennial transvaginal ultrasound scan (TVUS) in known mutation carriers or at-risk females from age 30-35 years.5 A randomized controlled trial of ultrasound surveillance versus no surveillance could be considered unethical in light of the reported high risk for endometrial carcinoma, but data on the outcome of the recommended surveillance strategy may be useful in assessing effectiveness, patient compliance, and cost implications and as a reference point for future collaborative research studies. Information on the outcome of endometrial carcinoma surveillance in women from HNPCC or HNPCC-like families was collected from the St Mark's Hospital Imperial Cancer Research Fund (ICRF) Family Cancer Clinic in the UK and the Netherlands Foundation for the Detection of Hereditary Tumors and is presented in this article.

PATIENTS AND METHODS

A total of 292 women from HNPCC or HNPCC-like families were offered annual or biennial endometrial carcinoma surveillance by ultrasound at either the ICRF Family Cancer Clinic at St Mark's Hospital, London, UK, or the Netherlands Foundation for the Detection of Hereditary Tumors.

St Mark's Hospital Patients

The ICRF Family Cancer Clinic was established in 1986. At that time, the Amsterdam Criteria (AC; at least three relatives with colorectal carcinoma [CRC], with one being a first degree relative of the other two; at least two successive generations affected; and at least one case of CRC diagnosed under the age of 50 years) had not yet been established, and endometrial carcinoma surveillance was offered to women from families whose pedigree suggested a dominantly inherited predisposition and who were at 50% risk of being mutation carriers. The families were subsequently reclassified as HNPCC (AC-positive) or HNPCC-like. One hundred eighty four women were recruited into the endometrial carcinoma surveillance program between March 1986 and March 1997. Recruitment was aimed at women between 25 and 65 years of age; 67 women were from families fulfilling the AC and 117 were from HNPCC-like families, defined as not fulfilling the AC because no cases of cancer occurred under the age of 50 years, or fulfilling the “modified” AC (ACII), with the inclusion of cases of endometrial carcinoma.6 Women were advised to arrange annual or biennial transvaginal pelvic ultrasound scans. If transvaginal ultrasound was not available, transabdominal pelvic scanning was recommended. The median age at the time of the first scan for the AC-positive group was 40 years (range, 24-64); for the HNPCC-like group the median age was 45 years (range, 20-71). The results of the scans were collected from the ICRF Family Cancer Clinic database, the hospital where the scan was carried out, and supplemented by information from each individual patient by means of a postal questionnaire (Fig. 1).

Figure 1.

Questionnaire sent to St Mark's Hospital patients.

Netherlands Registry Patients

One hundred eight women from 38 HNPCC families were entered into the endometrial carcinoma surveillance program between January 1, 1994, and September 1, 1999. Twenty five families were AC-positive and three families were ACII-positive. The remaining 10 families did not fulfill either the AC or the ACII, but pedigrees were suggestive of HNPCC, and HNPCC mutations were present. The median age at first scan was 42 years (range, 23-68 years). Women were advised to have a transvaginal ultrasound scan annually or biennially from age 30-35 years.

RESULTS

Data were available from 269 women, 171 from HNPCC families and 98 from HNPCC-like families. In total, 222 women had at least one scan, 159 from HNPCC and 63 from HNPCC-like families. A total of 825.7 years of risk was calculated from the time of enrollment until the end of the study period, hysterectomy, or time of death. During the study, 522 scans were carried out, resulting in a ratio of 0.63 for years scanned/years of risk. No cases of endometrial carcinoma were detected by surveillance ultrasound, but two interval cases of endometrial carcinoma did occur, one in the St Mark's Hospital group and one in the Netherlands Foundation group.

Case 1

The patient, a member of an HNPCC family, was initially seen in the ICRF Family Cancer Clinic in 1987 shortly after being treated for synchronous Duke Stage A and B colorectal carcinomas; at that time she was 51 years of age. She was enrolled in the pelvic cancer surveillance program in 1992, and had a normal transabdominal ultrasound scan in April 1992. She developed symptoms of postmenopausal bleeding early in 1994 and, after investigation, was found to have Stage 1 endometrial adenocarcinoma and underwent a total hysterectomy and bilateral salpingo-oophrectomy in July 1994. At the time of writing, she remains well and continues with regular colonoscopic surveillance. Mutation testing has since revealed a mutation in hMLH1.

Case 2

The patient is a known hMLH1 mutation carrier from an AC-positive family. Routine surveillance TVUS in December 1994 revealed no abnormalities. Over the following months she developed almost continual menstrual bleeding, and in May 1995, aged 46 years, a Stage 1 endometrial carcinoma was diagnosed. She underwent a total hysterectomy and remains well at the time of writing.

Three cases of endometrial carcinoma were brought to our attention at St Mark's Hospital in at-risk women from HNPCC families who had never been enrolled in the pelvic surveillance program. All three cases presented with symptoms at an early stage, underwent surgery, and remain well at the time of writing.

St Mark's Hospital Questionnaires

Of the questionnaires sent out, 86% (159 out of 184) were returned. Of the respondents, 79% had had at least one scan during the study period, 81% from AC-positive and 77% from HNPCC-like families. Every effort was made to contact nonresponders via family members and general practitioners. Eight individuals were reported to have died before the end of the study period, though none of these deaths were from endometrial carcinoma.

DISCUSSION

Surveillance for endometrial carcinoma in at-risk women or known HNPCC mutation carriers has been recommended by the ICG-HNPCC because of the high incidence of endometrial carcinoma in mutation carriers. In contrast, screening of the general population for endometrial carcinoma is not advocated as the disease usually presents early, survival rates are favorable, and there is a lack of evidence that screening is effective.7

We examined the outcome of endometrial carcinoma surveillance by ultrasound in 269 women from HNPCC and HNPCC-like families. Women were enrolled into the surveillance program on the basis of their family history (at 50% risk of inheriting a dominant mutation) and an unknown gene status. No cases of endometrial carcinoma were detected through surveillance in 825.7 patient years of followup. Two interval cases of endometrial carcinoma did present symptomatically at an early stage in women who were subsequently identified as mutation carriers and were successfully treated.

The current results suggest that annual or biennial pelvic ultrasound may not be an effective method to detect early endometrial carcinoma in this group of women, many of whom are premenopausal. The current results do confirm an increased risk of endometrial carcinoma in HNPCC women, as no cases would be expected in a matched population-risk group over this period. Between two and four cases of endometrial carcinoma were expected based on an annual risk of 1% in HNPCC mutation carriers and 0.5% in at-risk females. These results offer some reassurance, as there were no deaths from endometrial carcinoma during the study.

Although in the majority of clinical centers, women will be enrolled in surveillance programs on the basis of their family histories, results of further studies are awaited to determine whether ultrasound surveillance targeted at known mutation carriers will be more effective. A number of genes responsible for HNPCC have been identified. Mutations in two genes, hMLH1 and hMSH2, account for most cases, but mutations are also seen in hMSH6, hPMS1, and hPMS2. As mutation detection becomes more common, there is increasing evidence for genotype-phenotype correlations in HNPCC. hMSH2 females and those carrying hMSH6 mutations seem to be at greater risk for endometrial carcinoma, and surveillance may be yet more effective if targeted at this group.8, 9

There are a number of potential tools available for endometrial carcinoma surveillance. Transvaginal ultrasound scanning is the method of surveillance recommended by the ICG-HNPCC. It is a relatively noninvasive test that is readily available and inexpensive and allows examination of the endometrial cavity and the taking of measurements of the thickness of the endometrial lining. It is the initial examination of choice in the investigation of postmenopausal vaginal bleeding and, in this group, has been shown to be a highly sensitive but not very specific test. The addition of doppler color flow imaging may have the potential to increase the specificity for cancer diagnoses.10, 11 To our knowledge, the effectiveness of ultrasound surveillance for endometrial carcinoma in an asymptomatic population group has only been addressed in postmenopausal women receiving hormone replacement therapy or placebos.12 Langer et al. compared transvaginal ultrasonography with endometrial biopsy and found that the false positive rate of ultrasonography was unacceptably high, suggesting that it is not a practical screening procedure in this group.12

Other surveillance methods include endometrial cytologic sampling and endometrial biopsy. The former may be performed in an outpatient setting but has been shown to have a poor positive predictive value for early cancer detection in asymptomatic women.13 Endometrial biopsy is not an appropriate tool for screening/surveillance as it is uncomfortable, and may be technically impossible without dilatation and anesthetic in a significant number of women. It is unclear whether the results of existing trials are applicable to the HNPCC surveillance group, who are genetically predisposed to endometrial carcinoma and a significant number of whom will be premenopausal and not on hormone therapy.

In the St Mark's Hospital group, data from the questionnaire shows that the majority of women took up surveillance if it was advised, and the interval of one or two years was acceptable. No specific study of the women who did not arrange for scans was undertaken. On average, women were having scans at least biennially, as scans took place during 61% of the total years. The ICRF Family Cancer Clinic is a tertiary referral center, and individuals who are registered come from all parts of the UK. After their initial consultations they may choose to have the recommended surveillance carried out at their local hospitals, and the patients and/or local institutions are relied upon to forward any results to be entered into the database. Postal surveys are sent regularly to individuals to update the database, and, for the purpose of this study, a questionnaire regarding pelvic ultrasound surveillance was included. In patients in whom the questionnaire was the sole source of information regarding the ultrasound result, selection bias as well as the possibility of recall bias is acknowledged. Furthermore, there may have been cases of endometrial carcinoma in the St Mark's Hospital group that we have not accounted for in nonresponders. Every effort was made to contact individuals via family members and general practitioners of those patients for whom information was lacking, particularly to establish whether they had died from endometrial carcinoma. Inevitably some patients will be lost to follow up.

All cases of endometrial carcinoma reported, whether included in surveillance or not, are alive and well at a median of 5 years since diagnosis (range, 4-23 years) at the time of writing, which supports previously published data that endometrial carcinoma is a rare cause of death in HNPCC.2 Favorable survival rates for treated endometrial carcinoma should be given some consideration when giving advice to women about prophylactic hysterectomy. Information about ovarian pathology was not requested specifically in the current study, but one obvious advantage of pelvic ultrasound is the opportunity to visualize the ovaries to look for cancer.

In the current article, we have presented prospective data on the outcome of ultrasound surveillance in women from dominant pedigree colorectal carcinoma families who are at-risk for HNPCC. A randomized controlled trial of ultrasound surveillance versus no surveillance for endometrial carcinoma in at-risk HNPCC women would be considered unethical in light of the reported high risk for endometrial carcinoma. It is hoped that the current study may act as a basis for future collaborative studies and open a dialogue as to the effectiveness of ultrasound surveillance for endometrial carcinoma in HNPCC. The results raise the question as to whether alternative methods of surveillance for endometrial carcinoma in HNPCC should be assessed or, as there is a low mortality, symptomatic intermenstrual and postmenopausal bleeding should simply be investigated promptly.

Table 1. Results of Ultrasound Scan
GroupHNPCC n = 171HNPCC-like n = 98Total n = 269
  1. HNPCC: hereditory nonpolyposis colorectal carcinoma syndrome.

Years of risk525.7300825.7
Years scanned347175522
Years scanned/years risk.66.58.63
Cases of endometrial carcinoma202
Endometrial carcinoma detected by surveillance000
Deaths from endometrial carcinoma000

Acknowledgements

The authors thank H. Boonstra, Ph.D., J. de Graaff, Ph.D., R. H. M. Verheyen, Ph.D., John Shepherd, M.D., and Mrs. Maggie Stevens for their assistance.

Ancillary