Introduction Adnexal masses are formations originating from the ovaries, fallopian tubes, cervix and uterus or other pelvic organs, from which ovarian cancer stands out due to its high mortality rate. Ovarian cancer is diagnosed in an advanced stage, so the five-year survival rate is low. Ovarian cancer is a heterogeneous disease, it differs in histological type, grade, molecular features, and consequently the treatment differs. Despite the implementation of treatment that includes surgical treatment, systemic application of chemotherapy, immunotherapy, targeted therapy and hormonal therapy, the prognosis for patients is devastating. For this reason, the diagnostic value of the tumor markers HE4, CA 125 and ROMA index, as well as the expression of tissue HE4 and Ki-67 antigen in the preoperative differentiation of ovarian tumors, was investigated. This study examined the possibility of preoperative stratification in subjects with a tumor mass in the pelvis using markers CA 125, HE4 and the multiparameter model ROMA index. Regression analysis was used to compare the preoperative values of CA 125, HE4 and ROMA index, especially for identifying malignancy. Postoperatively, ovarian cancer was confirmed by pathohistological analysis, as well as other pathohistological subtypes of ovarian tumors: benign, borderline-malignant, tumors that have grafts and tumors resembling lesions that served as a control group. The expression of the tissue marker HE4, which is not routinely determined on ovarian tumor tissue, but according to pathohistological tumor types, was tested. Antigen Ki-67 is an indicator of cell proliferation, so determination of antigen Ki-67 on ovarian tumor tissues served as an indicator of tumor growth. The concordance of tissue expression of HE4, antigen Ki-67 and serum values of HE4, CA 125 and ROMA index with the stage of disease extension, and according to the FIGO classification, should be proof of whether the disease can be recognized as early as possible. Additionally, the association of the mentioned parameters with the size and blood supply of the ovarian tumor was investigated. Due to the influence of the habit of smoking cigarettes, it will be examined whether the habit of smoking has an influence on the measured results, especially for HE4. Materials and methods The research was carried out in the Department of Medical Biochemistry in Oncology at the University Hospital for Tumors, Sestre milosrdnice University Hospital Center, in cooperation with the Department of Gynecological-Oncological Surgery of the Institute for Surgical Oncology at the University Hospital for Tumors and the Clinical Institute of Pathology "Ljudevit Jurak", Sestre milosrdnice University Hospital Center. The criterion for inclusion in the study was evidence of an adnexal mass in the pelvis of unknown etiology in the subjects and an indication of surgical treatment. Preoperatively, the blood of 159 subjects was sampled for the purpose of measuring the serum values of the tumor markers HE4 and CA 125. For each subject, the ROMA index was calculated after classification according to menopausal status, and according to formulas for premenopause and postmenopause. The obtained results were used to examine the diagnostic potential of all three parameters: CA 125, HE4 and ROMA index. Postoperatively, the test subjects were divided according to pathohistological types of tumors by looking at the pathohistological findings in which the tumor tissue is described as benign, borderline-malignant or malignant, and the histological type and degree of tumor differentiation determined, microinvasion, assessment of the status of lymph nodes, the presence of ascites, involvement of the tubes, omentum, the presence of residual tumor and metastases. Subsequently, the expression of tissue marker HE4 and antigen Ki-67 was determined by immunohistochemistry from archived samples of deparaffinized sections of tumor tissue, which served as an additional marker for the assessment of cell proliferation. Monoclonal antibodies specific for tissue HE4 and Ki-67 antigen were used for samples, 140 in total, for which material was available. Tumor grade and disease stage according to the FIGO classification (stage I-IV): ovarian carcinomas limited to the ovaries (FIGO stage I), ovarian carcinomas extending to the pelvic organs (FIGO stage II), ovarian carcinomas with peritoneal metastases (FIGO stage III) and ovarian carcinomas with distant metastases (FIGO stage IV) were compared with all serum and tissue parameters. Results In the first part of the results, using the ROC (Receiver Operating Characteristics) statistical analysis, the curve visualized and classified the properties of three parameters: CA 125, HE4 and ROMA index. A regression analysis of the obtained results was made, separately for the premenopausal and postmenopausal groups. In the premenopausal group, the area under the ROC curves was 84.6 % for HE4, 86.7 % for CA 125, and 84.6 % for the ROMA index. The threshold values obtained in the ROC analysis are 86.1 pmol/L for HE4, 40.7 U/mL for CA 125, and 21.9 % for the ROMA index. Diagnostic sensitivity was 85.7 % for HE4, 75 % for CA 125, and 85 % for the ROMA index. Diagnostic specificity was 80 % for HE4, 69.3 % for CA 125, and 62 % for the ROMA index. The positive predictive value was 67 % for HE4, 35.3 % for CA 125, and 67 % for the ROMA index. The negative predictive value was 92.4 % for HE4, 91.9 % for CA 125 and 95.2 % for the ROMA index. In the postmenopausal group, the area under the ROC curves was 92.8 % for HE4, 89.9 % for CA 125, and 92.7 % for the ROMA index. The threshold values obtained in the ROC analysis are 99.8 pmol/L for HE4, 45.8 U/mL for CA 125, and 38.4 % for the ROMA index. Diagnostic sensitivity was 91.2 % for HE4, 88.9 % for CA 125, and 91.7 % for ROMA index. Diagnostic specificity was 80.3 % for HE4, 77.2 % for CA 125, and 80.3 % for the ROMA index. The positive predictive value was 75.6 % for HE4, 77.5 % for CA 125, and 91.2 % for the ROMA index. The negative predictive value was 92.4 % for HE4, 91.9 % for CA 125 and 95.2 % for the ROMA index. The tumor marker HE4, applied to the test subjects in our study, has a higher diagnostic capacity than CA 125 in the differential diagnosis of ovarian tumors, mainly in premenopausal patients. Analysis of the distribution of test subjects according to the pathohistological types of the disease and according to the menopausal status of the test subjects showed a significantly higher proportion of malignant cysts (84.2 %) in postmenopausal subjects, while a statistically significantly higher proportion of TLL was demonstrated in premenopausal subjects (65.22 %), p = 0.001. The youngest subjects were those who had TLL with a median age of 49.0 (43.5- 53.5) years and were significantly younger than benign cysts (P = 0.003), borderline changes (P = 0.010), malignant cysts (P < 0.001) and metastasis (P < 0.001). In the group of malignant cysts, the most represented status was IIIC with 23 (60.5 %) according to FIGO, and as for grades, the most represented was grade 3 (62.9 % of subjects with malignant cysts). The analysis of tumor flow and volume in the groups according to the pathohistological type of the disease was not statistically significant (P = 0.080). By tumor volume, benign cysts were significantly smaller than malignant (P < 0.001) and metastases (P = 0.021). The value of HE4 in the serum is 13.10 pmol/L, and the highest is 7700.00 pmol/L. The subgroup with malignant cysts had the highest values of serum HE4 with a median of 302.7 pmol/L, which were significantly higher than all other subgroups (P < 0.001) including metastases (P = 0.010). The metastasis subgroup had significantly higher values than all other subgroups except malignant cysts with a median of 117.65 pmol/L. A statistically significant difference between serum HE4 concentrations (P < 0.01) was demonstrated between individual subgroups: benign cyst subgroup versus malignant cyst subgroup, benign cyst subgroup versus metastasis subgroup, borderline subgroup versus malignant cyst subgroup, borderline subgroup versus metastasis subgroup, TLL subgroup vs. subgroups of malignant cysts, subgroups of TLL vs. subgroups of metastases. The malignant cyst subgroup had the highest tissue HE4 values which were significantly higher than the benign cyst subgroup (P = 0.049), the TLL subgroup (P < 0.001), and the metastases subgroup (P = 0.030). Subgroup TLL had the lowest value of tissue HE4: it had no expressed tissue HE4 at all, which was significantly lower compared to all other subgroups (P < 0.001). Differences in measured CA 125 values between subgroups of pathohistological types: the highest values were recorded in the group of metastases with a median of 241.8 U/mL, which were significantly higher than all other subgroups except for malignant cysts, median of 199.15 U/mL (P = 0.964). Ki-67 antigen in ovarian tissue was the most expressed in the group of malignant cysts with significantly higher values compared to all other pathohistological types (P < 0.001) except metastases (P = 0.183). Given that the ROMA index is calculated from HE4 and CA 125 values, the differences in values are similar to those values. The highest ROMA index values were in malignant cysts and were significantly higher than all other subgroups (P < 0.001). In the high-risk group, malignant cysts were the most common (45.3 %), while in the low-risk group those were benign cysts (63.1 %; P < 0.001). Also, in the group with high risk, the most represented was FIGO classification IIIC – 23 (67.6 %) respondents. The agreement of the clinical significance of the expression and serum concentration of HE4, Ki-67 and ROMA index with the FIGO classification, as well as the assessment of the clinical significance of the serum concentration of HE4 by comparison with the expression of HE4 in the tissue with the stage of the histological tumor grade, does not significantly correlate with any of the mentioned clinical values, while higher FIGO classification significantly positively correlates with serum HE4 (rho = 0.475, P = 0.004), CA 125 (rho = 0.530, P = 0.001) and ROMA index (rho = 0.546, P = 0.001). Considering the strength of the correlation coefficients, the strongest significant correlation is with the ROMA index itself, which points to clinical importance. Ki-67 did not significantly correlate with grade, FIGO classification, or ROMA index. Also, tissue HE4 did not significantly correlate with grade and FIGO classification. The research also examined the influence of smoking on the frequency and type of tumors. There is no proven connection between the pathohistological types of the disease in relation to the smoking status and the number of cigarettes smoked per day. Conclusion The HE4 marker improves the diagnostic specificity of CA 125 in distinguishing between benign and malignant pathology, allowing valuable surgical anticipation in case of a possible malignancy result. The regression analysis of the obtained measurement results showed differences for the premenopausal and postmenopausal groups. The limit values in our research are higher than declared by the manufacturer, and have proven to be optimal. Marker HE4 has the best diagnostic specificity in the diagnosis of ovarian tumors in premenopausal patients. Diagnostic sensitivity is equal for HE4 and ROMA index. The ROMA index is the marker that has the best diagnostic ability to recognize malignancy in postmenopausal patients, which proves the highest diagnostic sensitivity compared to CA 125 and HE4. This also affects the presence of malignant cysts in the high-risk group, so that HE4 and the ROMA index surpass CA 125 in the differential diagnosis of adnexal masses, and represent useful tools in the diagnosis of ovarian cancer. Malignant cysts had the highest values of serum HE4, which were significantly higher than all other examined subgroups, and the highest values of serum CA 125 were recorded in the group of metastases. Given that the ROMA index is calculated from HE4 and CA 125 values, the differences in values are similar to those values. Given the existence of statistically significant differences in the measured values of the HE4 marker in the serum, confirmation of its hypothetical value as a specific marker for ovarian cancer was obtained. This conclusion is in accordance with scientific evidence based on research on similar populations of patients. Tissue HE4 was expressed in all ovarian tumor subtypes, except in the control group, which was surprising. It can be concluded that the obtained results did not confirm the literature data stating that tissue HE4 as a marker is also expressed in healthy ovarian tissue, which represents the control group. Expression of the HE4 marker in all examined subtypes proves the growth of cells regardless of whether they are malignant or benign. Nevertheless, malignant cysts had the highest values of tissue HE4, so it can be concluded that determination of tissue HE4 for ovarian cancers is useful. The concordance of the clinical significance of the expression and serum concentration of HE4, Ki-67 and ROMA index with the FIGO classification, as well as the assessment of the clinical significance of the serum concentration of HE4 by comparison with the expression of HE4 in the tissue with the stage of the disease, differed by individual parameters. The strongest correlation was demonstrated for the ROMA index, which indicates its clinical importance. Ki-67 did not significantly correlate with grade, FIGO classification, or ROMA index. Also, tissue HE4 did not significantly correlate with grade and FIGO classification. The grade of the tumor does not significantly correlate with any of the mentioned clinical values. The conclusion is that the ROMA index is the most useful, given that it is a mathematical calculation based on the measured values of HE4 and CA 125.