Establishment of an appropriate screening test for ovarian cancer has long been sought. Symptoms that are associated with this malignancy are typically nonspecific and the association is often not recognized until the disease is advanced . Elevated serum CA125 levels have been detected in 50% and 92% of OC in early and late stages, respectively , similarly to the results observed for HE4 . Several studies have aimed to explore the link between serum VEGF levels and ovarian cancer [27, 28].
In this study, we investigated the diagnostics usefulness of VEGF and in combination with HE4 and CA125 in the patients with epithelial ovarian malignancies. Analysis of the usefulness of cytokine alone or in combination with other tumor markers, may improve the effectiveness of non-invasive diagnostics of this cancer . Furthermore, we made a comparison of the received results with the results of the tumor markers and with the control groups (benign ovarian lesions patients and healthy subjects). Additionally, we estimated the diagnostic utility of mentioned above parameters in correlation to the stage of ovarian cancer.
Our results show that VEGF concentrations as well as HE4 and CA125 levels in all groups of ovarian cancer patients were statistically significantly higher in comparison to the group of healthy subjects and patients with benign tumors (exception CA125 I stage). These data are very similar to the studies of other authors, who compared patients with ovarian [36–40] or breast cancer  with healthy volunteers however, there were differences in the number of patients. Our results are consistent with the results of Sedlakova et al.  and Li et al.  who observed increased levels of VEGF in the group of ovarian cancer patients compared to the benign ovarian tumor patients, although the tested groups with non-malignant lesions were much smaller (15 and 25 women respectively). Similar results were also received by Cheng and others  who observed higher concentrations but of VEGF-C in ovarian cancer (109 patients) in comparison to different benign gynecological lesions (76 subjects). In opposition to our observations there was no difference in the VEGF concentrations between OC patients and benign tumors group although only 7 serous malignant and 3 borderline tumors were included . Significantly higher serum levels of CA125 and HE4: (p = 0.005)  or (p < 0.0001) [38, 45] were found in OC patients than in those with benign diseases though there was various composition and menopausal status of the groups compared.
Moreover, it was observed that VEGF concentrations, similarly to HE4 and CA125, were statistically higher in every group (the analysis related to the stage of advancement of ovarian cancer), compared to the healthy controls. Similar results were confirmed by different authors , though they observed the highest concentrations of VEGF-C in more advanced stages and various sub-types of OC were enrolled in the study (serous-papillary, mucinous, endometrioid, clear cell). Interestingly, it was found that VEGF concentrations in serum significantly surpassed the control level in breast cancer patients (stages I-II) . Others found circulating concentrations of HE4 and CA125 significantly higher in patients with stage IA-IIB compared with healthy women, though more than four sub-types of ovarian cancer were included in the investigation and there was different the ethnical characteristics of the population selected .
Furthermore, significantly elevated levels of all parameters tested in the groups with early and late stage of ovarian cancer were observed compared to the cyst group. Our results are in agreement with different publications about our studied comparative tumor markers [36, 46, 47]. We were unable to confirm findings about VEGF in the papers published, since no reports on the subject are available.
The analysis of research results from literature revealed also almost identical to our observations significant differences when comparing patients in more advanced stage to metastasis-free early stage [36, 39, 40, 46], and is in opposition to another publication where a half smaller group of postmenopausal women was studied . It should be underlined also that in our investigations VEGF showed a significant difference in the concentrations between the two groups with more advanced stage of OC (III and IV – higher values for III stage). This may indicate the highest diagnostic usefulness of VEGF in locally advanced and metastases free stages of OC. One possible explanation for relatively low levels in patients with IV stage could be that VEGF is rapidly metabolized and excreted in case of presence many small, metastatic tumors. To our knowledge, little is known about the pharmacokinetic or pharmacodynamic characteristics of this cytokine. Undoubtedly, VEGF plays an essential role in the growth and metastases of ovarian cancer. Mattern and colleagues  showed the close correlation of VEFG expression with tumor cell proliferation. Vascular endothelial growth factor may also contribute to ovarian cancer metastasis by directly stimulating proliferation, survival, and/or migration of tumor cells . It has been validated by others that serum VEGF was an independent prognostic parameter in a large series of patients with all stages of OC .
To date no large studies had examined VEGF levels in healthy postmenopausal women with benign gynecological disorders. Cheng and colleagues  found no differences between benign ovarian diseases and healthy controls for serum levels of VEGF-C and CA125 (p > 0.05), though patients with cystis serous or cystis endometrioides were not enrolled in the group studied.
In opposition, we found in our previous study significant higher concentrations of CA125 in the group of 70 postmenopausal women with benign lesions of the ovary (cysts) . Our present observations about both tumor markers are in agreement with published evidence .
We showed the VEGF concentrations significantly different in the patients with studied sub-types of OC vs patients with cystis endometrioides, though we did not observe statistical differences in the concentrations of mentioned above cytokine between the two control groups: patients with cystis endometrioides and the healthy women group. On the basis of obtained results the VEGF could be presented as a useful diagnostic tool for the differentiation of patients with OC sub-type endometrioid epithelial, though we can not confirm these findings by other publications. Similarly to the discussed above results, HE4 also presented usefulness in the diagnostics of both investigated sub-types of OC. Moreover, patients with serous epithelial OC had significantly higher plasma levels of HE4 than patients with endometrioid epithelial OC. Patients with benign endometrioid epithelial ovarian tumors had lower HE4 and higher CA125 concentrations than patients with benign serous epithelial ovarian tumors. These data are in variance with previous observations [44, 52], and may be in part affected by the patients’ selection, though other studies have recently reported similar results . In agreement with the publications of Li et al.  and Czekierdowski et al.  we found no significant differences in VEGF levels among different pathological sub-types of ovarian carcinoma.
The correlation between examined markers levels was estimated by Spearman’s rank correlation test, because the value of each marker was not in a parametric distribution. Most of the measured values tended to be increased for every marker and in opposition to the publications of Tan and colleagues  we found a positive correlation between preoperative VEGF and CA125 levels, similarly to the other publication (r = 0.44; p = 0.0015) , though Manher et al.  found these parameters inversely correlated (r = −0.460; p = 0.01). Furthermore, we found positive correlation between these two mentioned above parameters also in control group (cysts) in opposition to the results of others (r = 0.10; p = 0.28)  and between VEGF and HE4 in healthy individuals group but no publications could confirm our findings. In accordance to demonstrated in our study results some authors showed positive correlation between CA125 and HE4 levels among patients with ovarian carcinoma (r = 0.54 p < 0.0001) [37, 55].
The ability to detect early cancers would definitely improve patient prognosis. It was demonstrated that the diagnostic sensitivity increased with the progress of cancer disease and was the highest for CA125 (from 40% in stage I up to 84% in stage IV of the disease). Our results are in agreement with the literature . On the other side, we could not demonstrate a better sensitivity than CA125 for ovarian cancer by alone VEGF testing, neither in early stages nor in advanced stages (though SE of this cytokine reached equal or even higher values to HE4 in particular stages). This could be explained by other authors’ investigations that levels of detectable VEGF in serum may depend on the binding to vascular endothelium or soluble receptors and/or degradation. In addition, VEGF levels in serum may also depend on the biosynthesis and release of VEGF from blood mononuclear cells and activated platelets . It is interesting to point out that we found out a maximum increase in the diagnostic sensitivity for the combination of all parameters tested to 84% in stage I, even to 96% in stages II-IV, and to 94-95% in every histological sub-type of OC. Results of this study suggest that combining VEGF with CA125 and HE4 measurements might enable improved early detection of this cancer as compared with use of either marker alone or of both comparative tumor markers. This observation is in accordance to our previous papers, in which we evaluated diagnostic criteria of selected cytokines and commonly used tumor markers in various gynecological malignancies [51, 57, 58]. Several studies have found sensitivity to be greater than in either marker alone: VEGF and CEA (carcinoembrionic antigen) in colorectal  or HE4 and CA125 in ovarian cancer patients [5, 45]. Diagnostic specificity (SP) was very high and reached the value of 94% for VEGF and HE4, and 92% for CA125. We found similar results in the existing literature [37, 39, 44].
The positive predictive value of the parameters tested were high in the total group of OC (94%). The negative predictive value shows the ability to exclude malignant disease and was the highest for CA125 (58%) in the total group of OC, though in current study the combination of both comparative tumor markers or with VEGF had unquestionable higher value (95%). These observations are in agreement with those recently reported by Molina et al.  and in opposition to the results of other authors who observed higher values of PV-NR for CA 125 or HE4 alone (92% and 72%; respectively) in Asian women , or distinctly higher values of these diagnostic criteria for HE4 than CA125 though for comparison of epithelial OC vs various benign lesions of the ovary in premenopausal women . We were unable to compare our findings about VEGF by the papers published, since no reports on the subject are available. The results of the current study support our previous findings regarding another cytokine - M-CSF in this malignancy .
The area under the ROC curve (AUC) indicates the clinical usefulness of a tumor marker. Our results show that CA125 had the highest diagnostic power from the parameters tested in the total group of OC patients (AUC = 0.9276) and its value was negligibly higher than that of HE4 (AUC = 0.8321). Interestingly, VEGF had the largest AUC (0.7984) in the patients with I stage, and comparable to HE4 in patients with II and III stage. Diagnostic accuracy was increased when three parameters were included resulting in the highest diagnostic power in every group studied regarding to the stage and histological sub-type of epithelial OC. Furthermore, we observed that the areas under the curve for VEGF and comparative markers were statistically significantly larger compared to AUC = 0,5 - borderline of diagnostic usefulness of the test. Test data published by Candido dos Reis and colleagues  indicated area under the ROC curve – 0.99 for VEGF in diagnosis of epithelial ovarian tumors, though the cyst fluid was used as a material for the assay. In contrary to the current results Cheng et al.  found the AUC of VEGF-C larger than CA125 in different screening groups: OC versus benign ovarian diseases and healthy controls (0.826; 0.760; respectively), OC vs healthy individuals (0.862; 0.853) and finally OC vs benign ovarian diseases (0.802; 0.681). In a few previous studies the AUC values of CA125 and HE4 for differentiating ovarian cancer were 0.82-0.95 and 0.85-0.96, respectively [37, 38, 45–47, 62] and was significantly higher for the combination of mentioned above markers  which were similar to our findings. Slight differences of the results between the studies might be caused by differences in the number of patients and the histological types or disease stages of ovarian cancer enrolled in each study.