Accuracy of preoperative histology and macroscopic assessment of cervical involvement in endometrial carcinoma

Prediction of cervical involvement in endometrial cancer by hysteroscopy

OBJECTIVE

This study evaluated the ability of hysteroscopy to preoperatively predict cervical involvement in endometrial cancer.

METHODS

The records of 110 surgically staged consecutive endometrial cancer patients treated at our institution from 1997 to 2003 were retrospectively analyzed. Data on demographics, preoperative staging procedures, surgical pathology reports, and adjuvant treatments were retrieved.

RESULTS

Fourteen (12.7%) patients had cervical involvement (stage II) according to the surgical pathology report, of whom nine (8.1%) had stage IIA and five (4.6%) had stage IIB. Clinical evaluation by speculoscopy and palpation had failed to reveal any indication of cervical involvement. Preoperative diagnostic hysteroscopy procedures were included. None of the hysteroscopy procedures revealed any suspicious lesion in the cervical canal.

CONCLUSION

Hysteroscopy and clinical examination fail to adequately predict cervical involvement in endometrial carcinoma. Fractional D&C appears to be the best method until a more effective alternative becomes available.

Endometrial endometrioid adenocarcinoma with a deceptive pattern of spread to the uterine cervix: a manifestation of stage IIb endometrial carcinoma liable to be misinterpreted as an independent carcinoma or a benign lesion

The prognosis of endometrial endometrioid adenocarcinoma is determined in part by stage; endocervical stromal involvement (stage IIB) imparts a worsened prognosis. We describe a deceptive pattern of stage IIB disease that mimics a primary endocervical glandular proliferation and may lead to understaging of endometrial endometrioid adenocarcinoma. Fifteen cases of endometrial endometrioid adenocarcinoma with a peculiar pattern of cervical involvement were identified from our consultation files. All cases were referred in consultation because of doubt about the nature of the cervical process and its relation to the corpus tumor; in a few instances, the cervical proliferation was considered possibly benign and in one case was misinterpreted as mesonephric hyperplasia. The patients ranged from 49 to 84 years in age (mean age 64.9 years). There was usually a grossly evident endometrial tumor. The cervix was unremarkable grossly in at least 11 patients. The cervical tumors were composed of variably shaped, often tubular glands with little or no stromal response and mainly invaded as widely spaced glands that often appeared deceptively benign. In 14 cases luminal secretions, mainly eosinophilic, were identified, often leading to consideration of a mesonephric lesion. Ten of the endometrial tumors were grade 1, four grade 2, and one grade 3. One was noninvasive, nine superficially invasive, and five deeply invasive. In four cases myoinvasion had, at least in part, a diffusely infiltrative pattern. The tumors in the cervix showed no in situ component and no definite surface involvement. Continuity with the corpus tumor could be demonstrated in 12 cases. Ten of the cervical tumors invaded more deeply than the endometrial tumor, four invaded to a similar depth, and only one was more superficial than its endometrial counterpart. The cervical and corpus tumors had a similar immunoprofile in nine cases: all were vimentin positive, eight estrogen positive and one negative, four carcinoembryonic antigen negative, and five with focal apical or rare cytoplasmic staining. This immunoprofile in conjunction with routine morphologic similarity between the two tumors and the usual documented continuity between them indicate that the cervical process represents spread from the endometrial endometrioid adenocarcinoma. It is important for both therapeutic and prognostic reasons that the cervical abnormality is not misinterpreted as a benign or malignant primary endocervical glandular process.

Value of dynamic MR imaging in assessing endometrial carcinoma involvement of the cervix

OBJECTIVE

The purpose of this study was to evaluate endometrial carcinoma involvement of the cervix using dynamic MR imaging compared with T2-weighted and contrast-enhanced T1-weighted MR imaging.

SUBJECTS AND METHODS

In 42 patients with endometrial carcinoma, T2-weighted MR imaging using rapid acquisition with relaxation enhancement, dynamic MR imaging using gradient-echo sequences, and contrast-enhanced T1-weighted MR imaging using spin-echo sequences were performed before treatment. We evaluated patterns of enhancement in the cervix and tumor. In 39 of the 42 patients who underwent surgical treatment, we compared MR imaging findings with histologic results concerning cervical involvement.

RESULTS

Enhancement of the cervical epithelium was greater than that of the tumor and cervical stroma on dynamic MR imaging in most patients. In assessing cervical involvement, the accuracy of T2-weighted, dynamic, and contrast-enhanced T1-weighted MR imaging was 85%, 95%, and 90%, respectively; no statistically significant difference was observed. False-positive cases on T2-weighted MR imaging were correctly identified as having no cervical involvement on dynamic MR imaging using the finding of continuous enhancement of the cervical epithelium. We found this finding to be reliable in assessing tumor involvement of the cervix.

CONCLUSION

We believe that, in combination with T2-weighted MR imaging sequences, dynamic MR imaging is useful in assessing endometrial carcinoma involvement of the cervix.