SEARCH

SEARCH BY CITATION

Keywords:

  • diagnosis;
  • osteopontin;
  • ovarian neoplasms;
  • ovary;
  • ultrasonography

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Aim

The aim of this study was to evaluate the role of the serum osteopontin (OPN) level as a biomarker for discriminating between malignant and benign ovarian tumors. Furthermore, comparisons with the diagnostic usefulness of the other tests were performed.

Methods

The study included 114 consecutive women with ovarian tumors (82 benign and 32 malignant) who were referred to our division.

Results

A cut-off level of 28.0 ng/mL for OPN showed a sensitivity of 71.87% and a specificity of 89.02%. The area under the receiver–operator curve (ROC) was 0.812. There were no differences in diagnostic utility between OPN and the other studied tests. OPN levels were lower in patients with endometriotic ovarian cysts than in those with other benign ovarian tumors (14.00 vs 19.50 ng/mL; P = 0.018). The difference between the median OPN level in patients with endometriotic cysts (14.0 ng/mL) and those with malignant tumors (40.85 ng/mL) was also statistically significant (P < 0.0001). The calculated OPN/CA-125 ratio was significantly different between patients with endometriotic cysts (median, 0.36; range, 0.05–2.89) and those with other benign tumors (median, 1.25; range, 0.05–5.70) (P = 0.0002). There was also a statistically significant difference in the median OPN/CA-125 ratio between patients with endometrial cysts (median, 0.36; range, 0.05–2.89) and those with malignant tumors (median, 0.12; range, 0.01–3.39) (P = 0.004).

Conclusion

The diagnostic utility of OPN is similar to that of ultrasonographic evaluation and CA-125 level assessment. Thus, OPN may be useful in differential diagnosis for less experienced ultrasonographers and is especially valuable for differential diagnosis of endometriotic cysts.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Ovarian cancer is the leading cause of death due to gynecological malignancies in the Western world.[1] The high mortality rate associated with ovarian cancer is strictly related to delayed diagnosis. Currently, as many as 70% of patients are diagnosed at International Federation of Gynecology and Obstetrics (FIGO) stage III and IV. By then, the disease is widespread and the prognosis is poor.[2] One of the crucial elements in the management of ovarian cancer is an appropriate referral to a gynecological oncology center because surgical treatment by a specialist significantly improves prognosis.[3] On the other hand, most benign tumors should be treated by a general gynecologist, and a laparoscopic approach may be indicated. Appropriate referral requires an accurate differential diagnosis of ovarian tumors, which remains an unsolved problem in gynecology. There are several methods for predicting malignancy in adnexal tumors including ultrasonographic evaluation via subjective assessment based on ‘pattern recognition’ or application of morphological indices and mathematical models.[4, 5] However, these methods, especially subjective assessment, require experience in ultrasonography; thus, easier methods such as the evaluation of biomarker levels may be useful for less experienced ultrasonographers.[6, 7]

Osteopontin (OPN) is a candidate biomarker for ovarian cancer. Physiologically, OPN is secreted by osteoblasts and the epithelial cells of multiple organs as well as by activated T lymphocytes, macrophages and leukocytes at the site of inflammation.[8, 9] However, this glycoprotein is strongly associated with ovarian cancer development and progression. In 2008, Song and co-workers showed that OPN promotes ovarian cancer growth in vitro and in vivo and increases the survival of ovarian cancer cells under certain stress conditions through activation of the PI3K/Akt/HIF-1α signaling pathway.[10] Furthermore, an increase in HIF-1α synthesis may facilitate angiogenesis in tumors.[11] Tilli et al. showed that OPN-c, an OPN splicing variant, contributed to the increased proliferation, migration and invasion of ovarian cancer cells.[12] Moreover, in ovarian cancer patients, increased OPN expression was shown to correlate with poor prognosis. In a 2007 study by Bao et al., increased OPN expression in metastatic lesions indicated decreased survival.[13] Additionally, OPN expression in ovarian cancer tissue samples was positively correlated with clinical stage, histological grade and lymph node metastasis.[14] Although several studies have focused on the role of OPN in ovarian cancer screening,[15] the utility of OPN for differentiating between malignant and benign ovarian tumors has not been sufficiently elucidated.

The main aim of this study was to evaluate the utility of the serum OPN level for differential diagnosis of ovarian tumors. Furthermore, biomarker level estimation was compared with several ultrasonographic tests and subjective assessments. Finally, we determined whether additional OPN or CA-125 assessment improves ultrasonography-based tumor evaluation.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

The study included 114 consecutive women who were referred to the Division of Gynecological Surgery, Poznan University of Medical Sciences, Poland, because of ovarian tumors between 2010 and 2012. An anamnesis was obtained from each patient. Data regarding each patient's age and menopausal status were collected. Postmenopausal status was defined as more than 12 months of amenorrhea or an age of more than 51 years in women who had undergone hysterectomy. All other women were considered premenopausal. Body mass index (BMI) was also calculated for each woman. The demographic characteristics of the studied patients are presented in Table 1.

Table 1. Demographic characteristics of the studied patients
 Benign ovarian tumorMalignant ovarian tumorP-value
Median age (range)39 years (18–74)55 years (31–80)P < 0.0001
Menopausal status (no. of patients)Premenopause, 64Premenopause, 11P < 0.0001
Postmenopause, 18Postmenopause, 21
Median body mass index (range)26 kg/m2 (18–38)22 kg/m2 (17–35)P < 0.0001

The tumors were classified according to World Health Organization criteria. Based on the results of histopathological examination, the women were divided into two groups: the benign ovarian tumor group (n = 82) and the malignant ovarian tumor group (n = 32). The histopathological findings obtained from both groups are summarized in Table 2. The tumor grade was defined using a 3-step scale. Regarding expectant management of ovarian tumors, the resolution of the tumor at a follow-up examination (usually after 3 months) was equivalent to the diagnosis of a functional cyst. Persistent tumors were managed operatively. Tumors with low malignant potential and metastatic adnexal tumors were classified as malignant ovarian tumors.

Table 2. Histopathological findings among the groups of patients studied
Benign ovarian tumor group (n = 82)
Serous cystadenoma10
Mucinous cystadenoma8
Endometriotic cysts35
Adult teratoma11
Functional ovarian cysts14
Paraovarian cysts1
Theca cell tumors2
Other (pedunculated leiomyoma)1
Malignant ovarian tumor group (n = 32)
Serous adenocarcinoma11
Serous borderline tumor2
Mucinous adenocarcinoma2
Mucinous borderline tumor2
Endometrioid adenocarcinoma3
Clear cell adenocarcinoma2
Undifferentiated carcinoma8
Other (ovarian lymphoma and hemangiopericytoma)2

Ultrasound examination was performed by one physician (R. M.) with experience in the field of gynecological ultrasonography. The tumors were evaluated according to the rules proposed by the International Ovarian Tumor Analysis Group in 2000.[16] In the case of bilateral tumors, the bigger and more complex tumor was considered for diagnosis. On the basis of ‘pattern recognition’ and subjective impression, the examiner (R. M.) classified all tumors as ‘benign’ or ‘malignant’.[4] The tumors were also classified according to the Gynecologic Imaging Reporting and Data System (GI-RADS) on the basis of definitions proposed by the authors in 2009.[17] Briefly, normal ovaries, which were not included in the study, were classified as GI-RADS 1. Tumors that were thought to be of functional origin were classified as GI-RADS 2, whereas neoplastic lesions that were likely benign (i.e. ovarian endometrioma, simple cysts) were classified as GI-RADS 3. Likely malignant tumors with one or two findings suggestive of malignancy were classified as GI-RADS 4, and tumors that were very likely to be malignant were classified as GI-RADS 5. Furthermore, the tumors were assessed using two morphological indices: the sonomorphological index (SM),[5] which was developed in our division in 2005, and the scoring system developed by Alcazar et al. in 2003,[18] which is also used in our division.

Blood samples were collected within 5 days before surgical treatment or on the day of admission in the case of women who were managed expectantly. Blood samples were centrifuged, and sera were stored at −82°C. The OPN concentration was measured in the obtained sera using enzyme-linked immunosorbent assay kits (catalog no. DOST00; R&D Systems Minneapolis, MN, USA). Serum CA-125 levels were assessed using the immunoenzymatic test (ST AIA-Pack OVCA Tosoh, Kyoto, Japan, used in our hospital's laboratory).

Finally, statistical analyses were performed using MedCalc ver. 11 and Cytel Studio StatXact 9. The area under the receiver–operator curve (ROC) was determined to assess the utility of OPN for differential diagnosis of ovarian tumors. Moreover, net reclassification improvement (NRI) was used to estimate the improvement related to the addition of biomarker level evaluation to ultrasonographic examination. NRI was calculated using the formula published in 2008.[19]

This study was approved by the local research ethics committee.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

The median serum OPN level in the ovarian cancer group (40.85 ng/mL; range, 8.0–126.5) was significantly higher than that in the benign ovarian tumor group (16.63 ng/mL; range, 4.82–113.08) (P < 0.0001). Using the ROC, we determined that the best cut-off value for OPN was 28.0 ng/mL. Similarly, the median CA-125 level was higher in the malignant tumor group (396.4 IU/mL; range, 7.39–3657.0) than in the benign ovarian tumor group (22.34 IU/mL; range, 4.57–393.9) (P < 0.0001). The ROC was also used to determine that the best cut-off value for CA-125 was 100 IU/mL. A comparison of the areas under the ROC (AUROC) for OPN (0.833; 95% confidence interval [CI], 0.748–0.899) and CA-125 (0.874; 95% CI, 0.794–0.931) indicated no significant difference between these two markers with regard to their ability to distinguish between malignant and benign ovarian tumors (P = 0.536). Both ROC are presented in Figure 1(a). We noted a statistically significant positive correlation between OPN and CA-125 levels (Spearman's rank correlation coefficient R = 0.385; P = 0.048). In the group of premenopausal women, the AUROC for OPN and CA-125 were 0.737 and 0.855, respectively, and the difference between these values was not significant (P = 0.391, DeLong test). Similarly, in the group of postmenopausal women, there were no differences between the AUROC for OPN and CA-125 (0.844 and 0.893, respectively; P = 0.567).

figure

Figure 1. Areas under the receiver–operator curves for osteopontin (OPN) and CA-125 discriminating patients with benign and malignant tumors (a), for OPN discriminating patients with benign tumors and patients with endometriotic cysts (b) and patients with malignant tumors and patients with endometriotic cysts (c).

Download figure to PowerPoint

Table 3 shows the sensitivity, specificity, predictive value and AUROC for each of the ultrasonographic tests and biomarker assessments performed. The best results were obtained using subjective ultrasonographic assessment; however, there were no statistically significant differences in the AUROC between all the studied tests.

Table 3. Sensitivity, specificity, predictive value and AUROC for each test used in the differential diagnosis of patients with benign and patients with malignant ovarian tumors
TestCut-offSens.Spec.PPVNPVAUROC
  1. AUROC, area under the receiver–operator curve; GI-RADS, Gynecologic Imaging Reporting and Data System; NPV, negative predictive value; PPV, positive predictive value; Sens., sensitivity; SM, sonomorphological index; Spec., specificity.

Osteopontin28.0 ng/mL71.8789.0271.9089.000.812
CA-125100.0 IU/mL75.8693.3381.5090.900.886
CA-12535.0 IU/mL86.2172.0046.3092.000.886
Subjective assessment 96.8785.3772.1098.600.896
GI-RADS490.6285.3770.7095.900.880
SM871.8787.8069.7088.900.844
Alcazar's scoring system487.5074.3957.1093.800.846

The calculated NRI revealed that neither OPN nor CA-125 could improve the diagnostic utility of the selected ultrasonographic test. Furthermore, the additional evaluation of CA-125 levels with a cut-off level of 35 IU/mL significantly worsened the diagnostic performance of subjective assessment and the GI-RADS. A summary of the NRI calculation results is presented in Table 4.

Table 4. Results of the NRI calculation, which determines the added utility of biomarker level assessment to the ultrasonographic tests performed
TestNRIP-value
  1. Different cut-offs of CA-125 presented in parentheses. Alcazar, Alcazar et al.'s scoring system; GI-RADS, Gynecologic Imaging Reporting and Data System; NRI, net reclassification improvement; OPN, osteopontin; SM, sonomorphological index; SUB, subjective ultrasonographic assessment.

SUB + OPN−0.1770.100
SUB + CA-125 (35)−0.2450.006
SUB + CA-125 (100)−0.1010.23
GI-RADS + OPN−0.1650.164
GI-RADS + CA-125 (35)−0.2380.03
GI-RADS + CA-125 (100)−0.080.398
SM + OPN0.020.86
SM + CA-125 (35)−0.070.592
SM + CA-125 (100)0.050.686
Alcazar + OPN−0.0040.971
Alcazar + CA-125 (35)−0.090.428
Alcazar + CA-125 (100)0.1280.282
CA-125 + OPN−0.0980.383

The median OPN level assessed in the sera obtained from patients with early-stage (FIGO stage I and II) ovarian neoplasms (32.5 ng/mL; range, 8.0–63.35) was significantly lower than that in patients with advanced (FIGO III and IV) ovarian malignancies (68.0 ng/mL; range, 19.0–126.5) (P = 0.001). The mean serum OPN level in patients with serous ovarian carcinomas was 60.45 ng/mL (standard deviation [SD] ± 27.67 ng/mL), whereas, in the group with non-serous malignant tumors, the mean level was 40.73 ng/mL (SD ± 27.76 ng/mL). The difference between these groups was nearly significant (P = 0.057). The median OPN level in patients with mucinous ovarian carcinomas was 30.13 ng/mL (range, 12.87–63.35), which was not significantly different from the level in patients with non-mucinous ovarian malignant tumors (43.97 ng/mL; range, 8.0–126.5) (P = 0.254). The median OPN level was significantly lower in the sera of women suffering from endometriotic cysts (14.0 ng/mL; range, 4.82–106.0) than in women with other benign ovarian tumors (19.50 ng/mL; range, 6.00–113.0) (P = 0.018). On the contrary, the median CA-125 levels were significantly higher in the group with endometriotic cysts (36.84 ng/mL; range, 8.58–217.00) than in the group with other benign ovarian tumors (14.93 ng/mL; range, 4.57–393.90) (P = 0.0009). We compared the serum OPN level in patients with endometriotic cysts (14.0 ng/mL) with that in patients with malignant tumors (40.85 ng/mL) and found that the difference was statistically significant (P < 0.0001). We calculated the best cut-off value discriminating patients with benign tumors and patients with endometriotic cysts (19.5 ng/mL) and also patients with malignant tumors and patients with endometriotic cysts (25.2 ng/mL). Both ROC curves are presented in Figure 1(b,c). Table 5 shows the sensitivity, specificity, predictive value and AUROC for OPN in discrimination patients with endometriotic cysts and patients with benign or malignant ovarian tumors. We also found that the difference in the OPN/CA-125 ratio between patients with endometriotic cysts (median, 0.36; range, 0.05–2.89) and those with other benign tumors (median, 1.25; range, 0.05–5.70) was statistically significant (P = 0.0002). There was also a statistically significant difference in the median OPN/CA-125 ratio between patients with endometrial cysts (0.36; range, 0.05–2.89) and those with malignant tumors (0.12; range, 0.01–3.39) (P = 0.004).

Table 5. Sensitivity, specificity, predictive value and AUROC for OPN in discriminating patients with benign tumors and patients with endometriotic cysts and also patients with malignant tumors and patients with endometriotic cysts
OPN testCut-offSens.Spec.PPVNPVAUROC
  1. AUROC, area under the receiver–operator curve; NPV, negative predictive value; OPN, osteopontin; PPV, positive predictive value; Sens., sensitivity; Spec., specificity.

Benign vs endometriotic19.5 ng/mL78.951.052.078.10.655
Malignant vs endometriotic25.2 ng/mL78.190.989.381.10.871

We found no differences in the OPN levels of patients with different grades of ovarian malignant tumors (P = 0.30). The mean OPN levels in patients with grade 1, 2 and 3 ovarian malignancies were 33.95 ± 17.46, 53.04 ± 33.37 and 53.84 ± 30.59 ng/mL, respectively.

In the group with malignant ovarian tumors, the OPN levels were not correlated with patient age (Pearson R = 0.268; P = 0.144), tumor size (Pearson R = –0.06; P = 0.762) or BMI (Pearson R = 0.271; P = 0.147). Similarly, in the group with benign ovarian tumors, there was no relationship between OPN levels and patient age (Spearman's R = 0.162; P = 0.145), tumor size (Spearman's R = –0.078; P = 0.507) or BMI (Spearman's R = 0.152; P = 0.183).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

In the present study, we found no differences between subjective ultrasonographic assessment, the GI-RADS,[17] the SM,[5] the scoring system proposed by Alcazar et al.,[18] or CA-125 and OPN level evaluation with regard to the accuracy for predicting ovarian malignancy. Although all of these methods are able to distinguish between malignant and benign ovarian tumors, each of them has specific advantages and disadvantages. Subjective assessment and GI-RADS are highly sensitive and specific, but an experienced sonographer is needed. Morphological index systems are more objective, but scoring rules must be strictly adhered to and significant differences are commonly found in external validations of prognostic models when compared to original studies. Biomarkers are the most objective, and they can be useful for clinicians who are less experienced in ultrasonography; however, their predictive value is relatively low.[5]

Because biomarker assessment is the most objective method, it is suitable for less experienced ultrasonographers. CA-125 is the most popular and widely used ovarian cancer biomarker. However, its utility as a single test for differential diagnosis of ovarian tumors has been questioned.[20, 21] Specifically, an assessment of CA-125 levels is not accurate for the diagnosis of mucinous ovarian adenocarcinomas, and CA-125 levels are frequently elevated in several benign conditions (including endometrial cysts) that affect premenopausal women, thus giving false-negative results.[21, 22] OPN is a novel ovarian cancer biomarker that is expressed in 100% of ovarian cancers that lack CA-125 expression at the tissue level.[23] In a 2002 study by Kim et al.,[24] OPN had a specificity of 80.4% and sensitivities of 80.4% and 85.4% for detecting early-stage (I/II) and late-stage (III/IV) ovarian cancers, respectively. However, in a 2008 study by Moore et al., OPN achieved a sensitivity of 19.6% and a specificity of 90% for the differentiation of malignant and non-malignant pelvic masses.[25] Our results are more similar to those of Kim et al.[24] According to a study by Tilli et al., OPN-c, an osteopontin-c splicing isoform, is especially involved in ovarian cancer progression. The authors showed elevated expression of OPN-c mRNA in ovarian carcinoma and ovarian borderline tumor samples with a contrasting lack of expression in benign ovarian tumors and normal ovarian tissue.[12] Thus, further studies evaluating the utility of secreted specific OPN isoforms in ovarian cancer screening and differential diagnosis are needed.

In a 2002 study, Kim et al.[24] observed no differences in OPN levels between different histological types of ovarian cancer. Similar results were obtained in our study; however, the differences between serous and non-serous ovarian cancer were close to the level of significance. We found no differences in OPN levels between mucinous and non-mucinous ovarian cancers; thus, OPN may be superior to CA-125 for the diagnosis of mucinous ovarian cancer. Furthermore, contrary to the findings for CA-125, we found significantly lower OPN levels in the sera of patients with endometriotic cysts than in the sera of patients with other benign ovarian tumors. In the majority of cases, ovarian endometriotic cysts are accurately diagnosed using subjective ultrasonography.[26] However, for less experienced ultrasonographers, OPN may be more useful than CA-125 in patients with endometriotic cysts because these cysts are a common cause of false-positive results in CA-125 evaluation.[27]

Subjective ultrasonographic assessment performed by an experienced sonographer was shown to be superior to mathematical predictive models and CA-125 evaluation.[4, 28, 29] In the present study, subjective ultrasonographic assessment achieved the highest AUROC; however, the differences between the various tests examined were not statistically significant. Subjective assessment is effective only when it is performed by an experienced clinician;[7] mathematical models or morphological scales are used more often by less experienced ultrasonographers.[4] In the present study, we employed two morphological scales that are routinely used in our clinic: (1) the SM developed in our department;[5] and (2) a scoring system proposed by Alcazar et al.[18] that also employs Doppler blood flow examination. In the original study, the SM showed the highest AUROC among the four indices that were examined.[5] This index was developed based on data from 686 patients, and we believe that it is both easy to learn and comprehensive in its analysis of tumor morphology. The scoring system proposed by Alcazar et al.[18] is much simpler than the SM, and in this study, we confirmed its utility for the differential diagnosis of ovarian tumors. In our opinion, the main limitation of Alcazar et al.'s scoring system is the high rate of false-positive results in the evaluation of pedunculated leiomyomas and fibrothecomas, which are completely solid and frequently have a central blood flow. These two features are sufficient to achieve the cut-off level of Alcazar et al.'s scoring system and give a false-positive result.

The GI-RADS was developed in 2009 to simplify communication between the sonologist and the clinician/gynecologist.[17] In the second published study based on the GI-RADS, the authors assumed that tumors scored as GI-RADS 4 and 5 should be referred to a gynecological oncologist.[30] Thus, when the cut-off level was set at GI-RADS 4 and above, the system achieved a sensitivity of 99.1% and a specificity of 85.9% for predicting malignancy. These results were similar to our findings (Table 3); however, the sensitivity was lower in our study. The GI-RADS is a type of subjective assessment, and in our opinion, tumors classified as GI-RADS 4 encompass tumors that are difficult to assess subjectively. In a 2011 study by Valentin et al.,[31] only 7% of masses were difficult to classify; however, these tumors represent the greatest challenge in the differential diagnosis of adnexal masses.

In the second part of our study, we found that neither OPN nor CA-125 could improve the diagnostic utility of morphological scales, the GI-RADS or subjective ultrasonographic assessments. Furthermore, we showed that after subjective assessment or use of GI-RADS, additional CA-125 evaluation with the cut-off level set at 35 IU/mL significantly worsened the predictive accuracy of ultrasonography. Similar results were obtained in 2009 by Valentin et al.[32] These results indicate that biomarker assessment should preferentially be used as a first-line test to triage patients rather than as a second-line test to verify ultrasonographic assessment.

We have shown that the diagnostic performance of OPN assessment is similar to that of subjective ultrasonographic assessment, morphological scales and CA-125 assessment for the differential diagnosis of ovarian tumors. OPN assessment is valuable for the differential diagnosis of ovarian endometriotic cysts when elevated CA-125 gives a false-positive result. In addition to the evaluation of CA-125, evaluation of the OPN level may be useful for less experienced ultrasonographers.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

This research was supported by a grant from Poznan University of Medical Sciences.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References
  • 1
    Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA CancerJ Clin 2010; 60: 277300.
  • 2
    Heintz AP, Odicino F, Maisonneuve P et al. Carcinoma of the ovary. FIGO 26th Annual Report on the Results of Treatment in Gynecological Cancer. Int J Gynaecol Obstet 2006; 95 (Suppl 1): S161S192.
  • 3
    du Bois A, Rochon J, Pfisterer J, Hoskins WJ. Variations in institutional infrastructure, physician specialization and experience, and outcome in ovarian cancer: a systematic review. Gynecol Oncol 2009; 112: 422436.
  • 4
    Valentin L, Hagen B, Tingulstad S, Eik-Nes S. Comparison of ‘pattern recognition’ and logistic regression models for discrimination between benign and malignant pelvic masses: a prospective cross validation. Ultrasound Obstet Gynecol 2001; 18: 357365.
  • 5
    Szpurek D, Moszynski R, Zietkowiak W, Spaczynski M, Sajdak S. An ultrasonographic morphological index for prediction of ovarian tumor malignancy. Eur J Gynaecol Oncol 2005; 26: 5154.
  • 6
    Franchi D, Maria Teresa S, Sara B, Davide R, Mario S. Ultrasound imaging compared to a multivariate predictive algorithm combining HE4 and CA 125 (ROMA) in the pre-operative assessment of adnexal masses. Ultrasound Obstet Gynecol 2011; 38 (Suppl 1): 39.
  • 7
    Timmerman D, Schwarzler P, Collins WP et al. Subjective assessment of adnexal masses with the use of ultrasonography: an analysis of interobserver variability and experience. Ultrasound Obstet Gynecol 1999; 13: 1116.
  • 8
    Patarca R, Freeman GJ, Singh RP et al. Structural and functional studies of the early T lymphocyte activation 1 (Eta-1) gene. Definition of a novel T cell-dependent response associated with genetic resistance to bacterial infection. J Exp Med 1989; 170: 145161.
  • 9
    Rodrigues LR, Teixeira JA, Schmitt FL, Paulsson M, Lindmark-Mansson H. The role of osteopontin in tumor progression and metastasis in breast cancer. Cancer Epidemiol Biomarkers Prev 2007; 16: 10871097.
  • 10
    Song G, Cai QF, Mao YB, Ming YL, Bao SD, Ouyang GL. Osteopontin promotes ovarian cancer progression and cell survival and increases HIF-1alpha expression through the PI3-K/Akt pathway. Cancer Sci 2008; 99: 19011907.
  • 11
    Bryant CS, Munkarah AR, Kumar S et al. Reduction of hypoxia-induced angiogenesis in ovarian cancer cells by inhibition of HIF-1 alpha gene expression. Arch Gynecol Obstet 2010; 282: 677683.
  • 12
    Tilli TM, Franco VF, Robbs BK et al. Osteopontin-c splicing isoform contributes to ovarian cancer progression. Mol Cancer Res 2011; 9: 280293.
  • 13
    Bao LH, Sakaguchi H, Fujimoto J, Tamaya T. Osteopontin in metastatic lesions as a prognostic marker in ovarian cancers. J Biomed Sci 2007; 14: 373381.
  • 14
    Zhang LL, Shao SL, Wu Y. Expressions of osteopontin and B7-H4 in epithelial ovarian neoplasm and their significance. Chin J Cancer 2010; 29: 2529.
  • 15
    Visintin I, Feng Z, Longton G et al. Diagnostic markers for early detection of ovarian cancer. Clin Cancer Res 2008; 14: 10651072.
  • 16
    Timmerman D, Valentin L, Bourne TH, Collins WP, Verrelst H, Vergote I. Terms, definitions and measurements to describe the sonographic features of adnexal tumors: a consensus opinion from the International Ovarian Tumor Analysis (IOTA) Group. Ultrasound Obstet Gynecol 2000; 16: 500505.
  • 17
    Amor F, Vaccaro H, Alcazar JL, Leon M, Craig JM, Martinez J. Gynecologic imaging reporting and data system: a new proposal for classifying adnexal masses on the basis of sonographic findings. J Ultrasound Med 2009; 28: 285291.
  • 18
    Alcazar JL, Merce LT, Laparte C, Jurado M, Lopez-Garcia G. A new scoring system to differentiate benign from malignant adnexal masses. Am J Obstet Gynecol 2003; 188: 685692.
  • 19
    Pencina MJ, D'Agostino RB, Sr, D'Agostino RB, Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med 2008; 27: 157172. discussion 207-12.
  • 20
    Jacobs I, Bast RC, Jr. The CA 125 tumour-associated antigen: a review of the literature. Hum Reprod 1989; 4: 112.
  • 21
    Malkasian GD, Jr, Knapp RC, Lavin PT et al. Preoperative evaluation of serum CA 125 levels in premenopausal and postmenopausal patients with pelvic masses: discrimination of benign from malignant disease. Am J Obstet Gynecol 1988; 159: 341346.
  • 22
    Zygmunt A, Madry R, Markowska J, Fischer Z. Estimation of the usefulness of neoplastic markers TPS and CA 125 in diagnosis and monitoring of ovarian cancer. Eur J Gynaecol Oncol 1999; 20: 298301.
  • 23
    Rosen DG, Wang L, Atkinson JN et al. Potential markers that complement expression of CA125 in epithelial ovarian cancer. Gynecol Oncol 2005; 99: 267277.
  • 24
    Kim JH, Skates SJ, Uede T et al. Osteopontin as a potential diagnostic biomarker for ovarian cancer. JAMA 2002; 287: 16711679.
  • 25
    Moore RG, Brown AK, Miller MC et al. The use of multiple novel tumor biomarkers for the detection of ovarian carcinoma in patients with a pelvic mass. Gynecol Oncol 2008; 108: 402408.
  • 26
    Valentin L. Pattern recognition of pelvic masses by gray-scale ultrasound imaging: the contribution of Doppler ultrasound. Ultrasound Obstet Gynecol 1999; 14: 338347.
  • 27
    Moore RG, Miller MC, Steinhoff MM et al. Serum HE4 levels are less frequently elevated than CA125 in women with benign gynecologic disorders. Am J Obstet Gynecol 2012; 206: 351 e1351 e8.
  • 28
    Van Holsbeke C, Van Calster B, Testa AC et al. Prospective internal validation of mathematical models to predict malignancy in adnexal masses: results from the international ovarian tumor analysis study. Clin Cancer Res 2009; 15: 684691.
  • 29
    Van Calster B, Timmerman D, Bourne T et al. Discrimination between benign and malignant adnexal masses by specialist ultrasound examination versus serum CA-125. J Natl Cancer Inst 2007; 99: 17061714.
  • 30
    Amor F, Alcazar JL, Vaccaro H, Leon M, Iturra A. GI-RADS reporting system for ultrasound evaluation of adnexal masses in clinical practice: a prospective multicenter study. Ultrasound Obstet Gynecol 2011; 38: 450455.
  • 31
    Valentin L, Ameye L, Savelli L et al. Adnexal masses difficult to classify as benign or malignant using subjective assessment of gray-scale and Doppler ultrasound findings: logistic regression models do not help. Ultrasound Obstet Gynecol 2011; 38: 456465.
  • 32
    Valentin L, Jurkovic D, Van Calster B et al. Adding a single CA 125 measurement to ultrasound imaging performed by an experienced examiner does not improve preoperative discrimination between benign and malignant adnexal masses. Ultrasound Obstet Gynecol 2009; 34: 345354.