Objective
The purpose of this study was to evaluate the performance of preoperative serum levels of human epididymis protein 4 (HE4) and cancer antigen 125 (CA125) in the prediction of the presence of metastases in endometrial carcinoma.
Study Design
Preoperative sera were collected from 98 women with a diagnosis of endometrial carcinoma. The concentrations of HE4 and CA125 were assessed by enzyme-linked immunosorbent assay and correlated with the results of the final histopathologic report.
Results
Fourteen patients had metastases (≥stage IIIA, International Federation of Gynecology and Obstetrics 2009 classification). The serum concentrations of HE4 and CA125 were higher in the group with metastases than in the group without metastases (median [interquartile range], 148.6 pmol/L [71.6–219.1 pmol/L] vs 77.2 pmol/L [52.9–99.3 pmol/L]; P = .001; and 20.0 U/mL [10.1–70.8 U/mL] vs 4.3 U/mL [2.9–10.4 U/mL]; P < .001, respectively). By a multivariate analysis, the combination of HE4 and CA125 (a risk score algorithm) was the only predictive factor for the presence of metastases (odds ratio, 21.562; 95% confidence interval, 5.472–84.963; P < .001), and the grade was the predictor for a deep (≥50%) myometrial invasion by the tumor (odds ratio, 2.005; 95% confidence interval, 1.123–3.581; P = .019). The sensitivity, specificity, positive predictive value, and negative predictive value for the combination of the markers to predict the presence of metastases were 71.4%, 89.5%, 55.6%, and 94.4%, respectively.
Conclusion
A combination of preoperative HE4 and CA125 seems to be a better predictor of metastatic disease than either 1 alone in endometrial carcinoma.
Endometrial carcinoma is the most common gynecologic malignant tumor in the developed countries. Endometrial carcinoma has generally a good prognosis, mainly because only a minority (25%) of the patients has a metastatic disease at presentation. The treatment of endometrial carcinoma is surgical, including a hysterectomy, bilateral salpingo-oophorectomy, peritoneal fluid sampling, and a pelvic and periaortic lymphadenectomy. However, the need for a routine lymphadenectomy recently has been debated in the case of the low-risk or stage IA grade 1-2 disease.
Serum markers are available for the diagnosis and follow-up of several cancers. Cancer antigen 125 (CA125) that is commonly used in ovarian cancer has been investigated for endometrial carcinoma also. A cutoff limit that would define normal and pathologic serum levels like the one for ovarian cancer or 35 U/mL has not yet been defined. However, an elevated CA125 level may be associated with a metastasized disease, and serial CA125 measurements can also be used during the treatment and follow-up examination of patients.
Human epididymis protein 4 (HE4) originally was isolated from the human epididymis but is also expressed in other tissues of the body. The HE4 protein contains 2 whey acidic protein domains and a core of 4 disulphides. Since its introduction, the biomarker capability of HE4 has been studied in various malignant tumors, which includes gastric, breast, and lung cancer. In endometrial carcinoma, the serum concentration of HE4 has been shown to correlate with the depth of myometrial invasion and the stage of the disease. Currently, the analysis of the concentration of HE4 in serum is used in parallel with CA125 to detect ovarian cancer, especially in premenopausal women. HE4 has also been shown to help in the discrimination of benign endometriotic cysts from malignant ovarian tumors.
The aim of this study was to investigate the performance of the preoperative serum HE4 and CA125 levels in the prediction an advanced stage and/or a deep (≥50%) myometrial invasion in endometrial carcinoma.
Materials and Methods
Ninety-eight consecutive women with a newly diagnosed endometrial carcinoma who were treated at Tampere University Hospital between September 2007 and October 2009 were enrolled in this prospective observational study. All patients gave written informed consent; the study was approved by the Ethics Committee of the Pirkanmaa Hospital District. All patients were scheduled for a hysterectomy, bilateral salpingo-oophorectomy, pelvic lymphadenectomy, and peritoneal fluid sampling. A paraaortic lymphadenectomy and an infracolic omentectomy were performed when indicated. Preoperative serum samples were collected at the preoperative outpatient office visit or on the day preceding the operation and stored at –70°C until analyzed.
The HE4 and CA125 concentrations in serum samples were measured by sandwich enzyme-linked immunosorbent assay (ELISA) kits (Fujirebio Diagnostics Inc, Malvern, PA; and Abnova GmbH, Heidelberg, Germany, respectively). All measurements were performed at room temperature according to the manufacturers’ instructions blinded to the results of the histopathologic report. The plates were read 2 and 5 minutes after the administration of the stop solution at the wavelength of 405 nm for HE4 and 450 nm for CA125, respectively. The minimum detection limits for HE4 and CA125 were 15 pmol/L and 5 U/mL, respectively. The coefficient of variance was 6.9% for HE4 and 13.6% for CA125, respectively.
The results of the ELISA analysis were correlated with the final histopathologic report. The patients were staged originally according to the International Federation of Gynecology and Obstetrics 1988 guidelines; however, for the purpose of this study, they were restaged according to the new 2009 guidelines.
The statistical analysis was performed with SPSS software (version 18.0; SPSS, Chicago, IL) and MedCalc software (version 12.0; MedCalc Software, Mariakerke, Belgium). The distribution of the continuous variables was assessed with the use of the Kolmogorov-Smirnov test. The comparison of the groups was performed with the use of the Mann-Whitney U test and the Kruskal-Wallis test when appropriate. The Spearman’s rank correlation test was used to evaluate correlations between the continuous variables. To compare the predictive performance of the markers, a receiver operating characteristics (ROC) analysis was accomplished. The variables that were found to be statistically significant in the univariate analysis were compared in a multivariate regression analysis by a definition of the dependent variables as metastasis or nonmetastasis and the presence or absence of a deep (≥50%) myometrial invasion. A probability value of < .05 was considered to be statistically significant. All tests were 2-sided.
Results
The mean age of the patients was 66.8 ± 8.8 years (range, 33–87 years). The demographics of the patients and the histopathologic characteristics of the tumors are presented in Table 1 . Three patients had a non–uterine cancer in the final histopathologic report, although the preoperative diagnosis based on the endometrial biopsy or curettage specimen had been endometrial carcinoma. Of these cases, 1 was a cervical cancer; 1 was a Fallopian tube cancer, and 1 was an endometrioid ovarian cancer. These 3 patients were excluded from the study. In 5 operations, the surgeon refrained from performing a lymphadenectomy because of the obesity or comorbidity of the patient. One of these patients had metastases in the ovaries. The other 4 women were excluded from the statistical analysis when the variables were correlated with respect to the presence of metastases.
| Variable | Measure |
|---|---|
| Age, y a | 66.8 ± 8.8 (33–87) |
| Weight, kg a | 80.6 ± 16.3 (52–130) |
| Body mass index, kg/m² a | 30.3 ± 6.0 (20.3–46.1) |
| Premenopausal, n | 3 |
| Postmenopausal, n | 92 |
| Histologic condition, n | |
| Endometrioid | 86 |
| Serous | 4 |
| Clear cell | 1 |
| Mixed | 2 |
| Carcinosarcoma | 2 |
| Grade, n | |
| 1 | 40 |
| 2 | 24 |
| 3 | 31 |
| Stage, n b | |
| IA | 47 |
| IB | 28 |
| II | 6 |
| IIIA | 6 c |
| IIIB | 0 |
| IIIC | 6 |
| IVA | 0 |
| IVB | 2 |
| Myometrial invasion, n | |
| <50% | 50 |
| ≥50% | 45 |
| Personal history of malignancy, d n | 6 |
a Data are presented as mean ± SD (range)
b International Federation of Gynecology and Obstetrics 2009 classification
The median concentrations of HE4 and CA125 in serum are presented in Tables 2 and 3 . Both markers exhibited a correlation with metastases and a deep myometrial invasion. An association with the patient’s age and body mass index (BMI) was observed for the concentration of HE4, whereas no similar association was seen for CA125. A positive correlation was found between the levels of HE4 and CA125 ( Table 4 ). The HE4 levels furthermore correlated with the histologic grade (median, 69.8 [interquartile range, 51.3–101.9] vs 77.2 [interquartile range, 52.7–95.3] vs 99.2 [interquartile range, 64.7–172.3] for grades 1, 2 and 3, respectively; P = .012, Kruskal-Wallis test). However, when the analysis was repeated and the metastatic cases were excluded, no statistically significant correlation was found.
| Grade | Metastases | |||||||
|---|---|---|---|---|---|---|---|---|
| n a | Human epididymis protein 4 (pmol/L) b | Cancer antigen 125 (U/mL) b | ||||||
| No | Yes | No | Yes | P value c | No | Yes | P value c | |
| 1 | 34 | 4 | 64.1 (48.9–99.0) | 128.8 (73.3–415.7) | .052 | 4.3 (2.2–7.0) | 25.3 (8.6–115.4) | .023 |
| 2 | 21 | 3 | 77.2 (52.7–92.6) | 148.3 (N/A) | .310 | 4.2 (3.2–8.3) | 28.9 (N/A) | .023 |
| 3 | 22 | 7 | 91.9 (63.4–130.3) | 174.9 (74.4–250.1) | .067 | 4.8 (2.7–12.4) | 14.4 (4.5–57.6) | .083 |
| All | 77 | 14 | 77.2 (52.9–99.3) | 148.6 (71.6–219.1) | .001 | 4.3 (2.9–10.4) | 20.0 (10.1–70.8) | < .001 |
a Four patients with unknown lymph node status were excluded.
b Data are presented as median (interquartile range).
| Grade | Deep invasion | |||||||
|---|---|---|---|---|---|---|---|---|
| n | Human epididymis protein 4 (pmol/L) a | Cancer antigen 125 (U/mL) a | ||||||
| No | Yes | No | Yes | P value b | No | Yes | P value b | |
| 1 | 26 | 14 | 63.5 (45.7–93.7) | 83.4 (61.5–145.6) | .039 | 3.4 (1.8–6.1) | 10.8 (4.7–23.8) | .001 |
| 2 | 15 | 9 | 75.2 (52.7–83.4) | 89.6 (47.3–169.9) | .290 | 4.3 (3.2–10.2) | 4.2 (3.2–40.5) | .446 |
| 3 | 9 | 22 | 70.3 (55.9–91.9) | 122.6 (81.6–207.7) | .009 | 3.2 (2.3–4.0) | 11.2 (4.7–20.5) | .002 |
| All | 50 | 45 | 68.3 (50.5–90.1) | 111.7 (63.3–164.7) | < .001 | 3.6 (2.2–5.4) | 10.9 (4.1–25.3) | < .001 |
a Data are presented as median (interquartile range).
| Variable | Human epididymis protein 4 | Cancer antigen 125 | ||
|---|---|---|---|---|
| r | P value a | r | P value a | |
| Age | 0.406 | < .001 | 0.078 | .452 |
| Body mass index | 0.406 | < .001 | −0.058 | .614 |
| Human epididymis protein 4 | 0.404 | < .001 | ||
| Cancer antigen 125 | 0.404 | < .001 | ||
With cutoff limits of 70 pmol/L and 35 U/mL for HE4 and CA125, respectively, the sensitivity, specificity, positive predictive value, and negative predictive value for the presence of metastases were calculated for both markers. The performance of the combination of the markers was evaluated by calculation of a risk score for the presence of metastases with the use of the algorithm for postmenopausal women that had been described by Moore et al. A risk score of >27.7 was considered positive. Because 2 of the 3 premenopausal patients were analogous to the postmenopausal group by their age (49 and 51 years, respectively), they were included in the risk score analysis. The youngest of the study participants (age, 33 years) was excluded ( Table 5 ). Nineteen women had a risk score of >27.7. All of them had a deep myometrial invasion, and 10 women also had a metastatic disease. A ROC analysis for the risk score algorithm produced an area under the curve of 0.824 (95% confidence interval [CI], 0.689–0.959) for the presence of metastases. When this was compared with the areas under the curve of HE4 and CA125 as single markers (0.763 [95% CI, 0.610–0.917] and 0.802 [95% CI 0.672–0.932]), no statistical difference was found (risk score vs HE4; risk score vs CA125; HE4 vs CA125; P = .170; P = .424; P = .545, respectively).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
