Abstract
Evaluation of: Oikonomopoulou K, Li L, Zheng Y et al. Prediction of ovarian cancer prognosis and response to chemotherapy by a serum-based multiparametric biomarker panel. Br. J. Cancer 99(7), 1103–1113 (2008).
The use of biomarkers may contribute to a better prediction of the clinical outcome and response to chemotherapy for patients with ovarian cancer. Previous results from the group in the reviewed study have shown an association between the protein expression in solid tumors of different kallikreins, B7-H4 and CA125 and ovarian cancer classification, prognosis and response to chemotherapy. Oikonomopoulou and colleagues have now investigated the role of these proteins in serum samples from 98 patients with ovarian cancer. The odds ratio and hazard ratio were calculated and time-dependent receiver-operating characteristic curves were utilized to evaluate the prognostic performance of the biomarkers. Internal validation of the data was performed based on resampling the original data. The results showed that elevated levels of various proteins were associated with worsened outcomes regarding clinical response, overall survival, progression-free survival and time from first chemotherapy to progression. However, the internal validation results did not confirm a correlation at the same level. These data demonstrate that a combination of different kallikreins, CA125, B7-H4 and spondin-2 may be potential biomarkers to predict survival and response to chemotherapy for patients with ovarian cancer. Further evaluation of the proteins is required, with validation on external sample sets focusing on the most important clinical questions, to be able to draw conclusions about their applicability as serum biomarkers.
Oikonomopoulou and colleagues performed a study of potential serum biomarkers. They hypothesized that a panel of proteins from the kallikrein (KLK) family and other proteins may provide new ways to predict response to chemotherapy, progression-free and overall survival, and time-to-progression of ovarian cancer patients. In this hypothesis-driven study, they used previous information from publications and their own work to select the ten biomarkers in the panel: CA125, KLK 5, -6, -7, -8, -10 and -11, as well as B7-H4, Reg-IV and spondin-2. This is an important way to further validate potential biomarkers and their role as predictive factors. However, the use of many different clinical end points weakens the results and conceals the most prominent way to use the biomarkers. It is important to reduce the number of end points in further studies to address the most relevant clinical questions.
The markers were quantified by sandwich-type ELISA, except for CA125 values, which were measured with a Roche Elecsys method, in serum samples from 98 patients diagnosed with epithelial ovarian cancer. The samples were well characterized in terms of known prognostic factors such as International Federation of Gynecology and Obstetrics stage and age, and these factors were adjusted in the statistical analysis. However, histopathology and residual tumor after primary optimal debulking surgery were not presented and adjusted for. The latter is especially important as prognosis is strongly related to optimal debulking before adjuvant or after neoadjuvant chemotherapy Citation[1]. The median age of the patients was 50 years (range: 22–77 years), and the patient group were younger than the average ovarian cancer patient, which may affect the results. Furthermore, the patients were treated with a combination of carboplatin–cyclophosphamide or Taxol®–carboplatin. The authors did not consider the differences in treatment in the analysis. When different treatments are used, it is preferable to check whether the two treatment groups can be merged before combining them into one group.
Among the patients, 58 (61%) had a good response, categorized as complete remission (CR; n = 29), partial remission (PR; n = 21) or stable disease (n = 8), and 37 had poor clinical response with progressive disease. Among the 50 CR and PR patients, 27 experienced disease progression, with a median time-to-progression of 6 months (range: 3–24 months). Patients were monitored for survival and disease progression with a median follow-up of 24 months (range: 2–36 months). A total of 45 patients died within the time limits of the study.
Oikonomopoulou and colleagues evaluated the clinical response characterized as good (i.e., CR, PR or stable disease) versus poor response, overall survival, progression-free survival and time from the first chemotherapy to progression among CR and PR patients. The odds ratio and hazard ratio were calculated and time-dependent receiver-operating characteristic curves were utilized to evaluate the prognostic performance of the biomarkers. The predictive accuracy was evaluated based on a cross-validation of the original data by calculating the overestimation of the sensitivities at given specificities. The results are summarized in .
The predicted accuracy need to be confirmed with internal validations and, ideally, on an independent set of samples Citation[2]. In the reviewed study, an external validation set was not available and Oikonomopoulou and colleagues, appropriately, performed an internal validation by calculating the overestimations. This enables the reader to estimate the implication of the results. Oikonomopoulou and colleagues reported an interesting association between the biomarkers and the patients’ survival and response to chemotherapy. However, the calculations of the overestimations indicate that their usefulness in clinical practice is low. Due to the heterogeneity of ovarian cancer, the potential of the present (and other) biomarkers may be missed. Further evaluations of the biomarkers may show that it is possible to better predict the response and survival for subset of patients.
The majority of kallikrein mRNA and/or proteins have been reported as upregulated in ovarian carcinoma tissues, serum samples from ovarian cancer patients and cancer cell lines. Kallikreins are also associated with more aggressive forms of ovarian cancers. Of these, KLK6 seems to be the most prominent candidate as a biomarker Citation[3]. Zheng and colleagues used solid tumors and studied the protein expression of kallikreins, B7-H4, and CA125 in multiparametric combinations with clinical variables. They found a significant relationship with ovarian cancer classification, prognosis and response to chemotherapy Citation[4]. The work by Oikonomopoulou and colleagues further develop this hypothesis by using serum samples instead of solid tumors. This step is necessary if one considers the set of biomarkers as a screening test for diagnosis of ovarian cancer. However, since optimal debulking surgery is the main first-line treatment for ovarian cancer, tumor tissue material will always be available for prognostic evaluation. Nevertheless, the study provides important data to the field of kallikreins and ovarian cancer, and the additional data regarding kallikreins could enable the development of new biomarkers. However, it is less likely that this set of proteins will stand if ever tested in an external validation, since this particular biomarker panel was found to be of low clinical value.
Expert commentary
The use of biomarkers may contribute to a better prediction of the clinical outcome for patients with ovarian cancer and may facilitate selecting the most optimal and individual treatment. Cytogenetic changes and differences in gene and protein expressions may be used to better predict patients’ survival and response to treatment. Technological advances, such as large scale analysis of tumor DNA, RNA and proteins, are new and rapidly evolving possibilities for the detection of potential biomarkers. Several attempts to find markers in serum samples have been performed Citation[5]. A search for ‘serum markers ovarian cancer and prognosis’ in Pubmed resulted in almost 2000 citations. The current challenge lies in selecting significant results from this great quantity of information and then further validating this data. Several promising biomarkers have been found that, later on, were shown to be of no significance. The accumulation of potential biomarkers requires further validation of the results to avoid continued growth in the amount of inadequate information. Consequently, to reduce the number of potential markers for analysis, it is important that studies with positive as well as negative validation results are published.
Five-year view
The exit screening for serum biomarkers accumulates a great quantity of information, and the search of accurate serum biomarkers has only just begun. When the search has reached its saturation point, it is important to focus on the most relevant markers for the development of new therapeutic strategies that can improve the patients’ situation. The accumulating data of biomarkers, together with increasing knowledge regarding methods of validating the predictive accuracy and clinical usefulness of different biomarker settings, will most definitely aid the development of new therapeutic strategies against ovarian cancer Citation[6].
Table 1. Summary of significant findings.
Key issues
• Oikonomopoulou and colleagues examined the expression of ten novel biomarkers in 98 serum samples from patients with epithelial ovarian cancer. The markers had been investigated previously in relation to different malignancies, including ovarian cancer.
• The biomarkers were analyzed in relation to prediction of response to chemotherapy, progression free and overall survival, and time to progression for ovarian cancer patients.
• The results showed that elevated levels of different biomarkers were associated with worse outcome in different tests. CA125, kallikreins 6 and 10, and B7-H4 were significant predictors in all comparisons. Receiver-operating characteristic curve analysis of different combinations of CA125, kallikreins 5, 7, 8 and 10, B7-H4, and Spondin-2 created area-under-curve values of 0.77–0.82, with corrected area-under-curve values of 0.62–0.65 .
• The results suggest that a combination of different kallikreins, CA125, B7–H4 and Spondin-2 may predict overall and progression-free survival. It may also be able to predict the patient’s response to chemotherapy and whether the patients will progress within 1 year or not.
• Further investigations are necessary, with validation on external sample sets focusing on the most important clinical questions, to establish the proteins applicability as biomarkers for ovarian cancer prognosis and response to chemotherapy.
Financial & competing interests disclosure
The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.
References
- Vergote I, van Gorp T, Amant F et al. Timing of debulking surgery in advanced ovarian cancer. Int. J. Gynecol. Cancer18(Suppl. 1), 11–19 (2008).
- Lusa L, McShane LM, Radmacher MD et al. Appropriateness of some resampling-based inference procedures for assessing performance of prognostic classifiers derived from microarray data. Stat. Med.26(5), 1102–1113 (2007).
- Paliouras M, Borgono C, Diamandis EP. Human tissue kallikreins: the cancer biomarker family. Cancer Lett.249(1), 61–79 (2007).
- Zheng Y, Katsaros D, Shan SJ et al. A multiparametric panel for ovarian cancer diagnosis, prognosis, and response to chemotherapy. Clin. Cancer Res.13(23), 6984–6992 (2007).
- Gagnon A, Ye B. Discovery and application of protein biomarkers for ovarian cancer. Curr. Opin. Obstet. Gynecol.20(1), 9–13 (2008).
- Henry NL, Hayes DF. Uses and abuses of tumor markers in the diagnosis, monitoring, and treatment of primary and metastatic breast cancer. Oncologist11(6), 541–552 (2006).