Could serologic and ultrasonografic indexes be useful for therapeutic decisions in patients with ovarian cancerP. Achimas-Cadariu, Al. Irimie, I. Neagoe, L. Achimas-Cadariu, R. Buiga
Original article, no. 3, 2009
* UMPh-Iuliu Hatieganu, Cluj-Napoca, Institute of Oncology-Prof. Dr. Ion Chiricuta
* Department of Surgery, Octavian Fodor-Emergency Hospital, Cluj-Napoca, Romania
* Institute of Oncology-Prof. Dr. Ion Chiricuta, Cluj-Napoca
Ovarian cancer is responsible for the majority of deaths related to malignant neoplasia of the female genital tract in industrialized countries. The disease is detected in advanced stages, fact that influences treatment choices and survival rates. (1) Ovarian cancer prognosis depends on clinical, morphological, biological and therapeutic variables. However, too little is currently known about the real influences of these variables and their interrelations. (2)
Tumor stage and grading, residual tumor after surgery, histological type and age at diagnosis are the most important prognostic parameters studied in ovarian cancer. (3,4,5) A limited number of studies communicate the prognostic significance of some clinical predictors as presence of affected lymph nodes or ascites. (6) Previous studies indicated that none of the studied biomarkers used individually was sufficient to diagnose the disease. The preoperative CA125 serum level seems to be an independent predictor of disease free interval and overall survival in patients with epithelial ovarian cancer. (3,7,8,9) The use of highly specific cell marker CD-34 aimed to find an association between the density of the microvascularization and disease-free survival and overall survival (10,11,12).
Primary surgery performed in specialized onco-gynecologi cal hospitals improves ovarian cancer prognosis as complete cytoreduction is usually attempted (9,13,14,15,16, 17,18,19,20).
Some prognostic indexes have been proposed but there is no evidence that disease free survival and overall survival in patients with ovarian cancer can be improved by means of serum CA-125 value, pelvic ultrasound or other procedures (21,22).
Nevertheless, all these factors fail to fully reveal the biology and aggressiveness of the tumor. The identification of predictors or prognostic indexes for the aggressiveness of ovarian tumors opens the path for targeted therapeutic decisions.
This study is based on the results and conclusions of previous research that the author of the current paper published in two Romanian medical journals (23,24) observing that: in epithelial ovarian carcinoma the time to recurrence and the overall survival depends on disease stage at first hospital admission, tumor grading and first therapeutic scheme; specific prognostic factors for time to recurrence are age at diagnosis and the degree of angiogenesis expressed by CD-34 values and vascular grading (highly sensitive for the occurrence of metastases). The overall survival is specifically predicted by disease free interval and the expression of the angiogenic marker Platelet Derived Endothelial Cell Growth Factor (PDECGF). Both CD-34 and PDECGF expression increase significantly in high-grade tumors.
The study aims to look for an association between prognostic factors (clinical, morphological, therapeutical and biological) of ovarian cancer and to quantify it; to establish prognostic indexes for different associations of predictors; to establish patterns that can explain the aggressiveness of some tumors (using non-invasive preoperative investigations); to rank the predictive value of tests used in ovarian tumor cases as well as to involve the above-mentioned issues in therapy decision.
Material and method
This prospective research comprised 124 cases from 574 consecutive patients admitted at the Institute of Oncology, Cluj-Napoca between 01.04.2003 and 01.04.2006 with pelvic or pelviabdominal tumors. All patients signed an informed consent. The study design has been assessed by the University of Medicine and Pharmacy “Iuliu Hatieganu” Cluj-Napoca, Romania (www.umfcluj.ro) Scientific Council and Ethical Commission under guidelines according to the Helsinki Declaration of 1975 on human experimentation.
To select the 124 cases the following exclusion criteria were used: primary tumor not recorded, no follow-up possibility during the study, uterine pathology or another pathology not related to the adnexa, multiple localizations (at least one extra-adnexial), lack of standardized treatment, presence of a second primary localization, admission to hospital in terminal stage, suspicion of ovarian metastases, unknown histopathology, lack of or inconclusive biological determinations.
This study was both observational, as it followed the naturally development of the disease and the influences of prognostic factors, and also analytical, as it compared the various studied groups.
Data collection followed the “representative sample” system, which allows more freedom in elaborating indicators. The groups included in the study matched representative criteria for the target population.
The variables studied were: anthropometric and social, clinical, serological ( preoperative CA 125; levels of various cytokines and vascular growth factors - FAST Quant Human Th1/Th2 was used to evaluate the following cytokines: INTERLEUKINS=IL (IL-1b, IL-2, IL-4, IL-5, IL-6, IL-10, IL-13), Tumor Necrosing Factor a (TNFa), Interferon g (IFNg) and FAST Quant Human Angiogenesis in order to evaluate several factors involved in the angiogenesis process: Platelet Derived Growth Factor (PDGF–BB), Vascular Endothelial Growth Factor (VEGF), Fibroblastic Growth Factor (FGF-b), Keratinocotic Growth Factor (KGF), Angiogenin and Angiopoetin-2), ultrasound, biological (identification of CD-34, expressed by the Intratumoral Microvascular Density, and the expression of Platelet Derived Endothelial Cell Growth Factor (PDECGF)), temporal, the treatment received by the patient: 1 for primary optimal citoreductive surgery (residual disease under 10 mm) followed by adjuvant chemotherapy, 2 for suboptimal surgery (residual disease over 10 mm) followed by adjuvant chemotherapy, 3 for optimal interval surgery (performed after neoadjuvant chemotherapy with residual disease under 10 mm), 4 for suboptimal interval surgery (performed after neoadjuvant chemotherapy with residual disease over 10 mm), 5 for surgery in cases of benign tumors.
An electronic database was created. The fields correspond to the variables studied; additional new prognostic indexes were introduced in the database by the combination and interaction of independent prognostic factors to obtain more predictive outcomes. Only the fields that represent original contributions of the present study are presented as follows:
- Prognostic Score established postoperatively: the sum of FIGO level (1 for FIGO I and II or 3 for FIGO III and IV) and tumor grade (1, 2 or 3) –minimum value = 2, maximum value = 6.
- Vascular Score established postoperatively: the sum of CD-23 level and PDECGF level – minimum value =2, maximum value = 8.
- CD-34 levels calculated according to descriptive statistics (as in Table 1 below): 1 (between minimum and 25%, respectively values from 1 to 101); 2 (between 25% and average, respectively values from 102 to 132); 3 (between average and 75%, respectively values from 133 to 176) 4 (between 75% and maximum, respectively values from 177 to 428).
- Platelet Derived Endothelial Cell Growth Factor levels calculated according to descriptive statistics (as in Table 1 below): 1 (between minimum and 25%, respectively values from 1 to 14); 2 (between 25% and average, respectively values from 15 to 26); 3 (between average and 75%, respectively values from 27 to 41) 4 (between 75% and maximum, respectively values from 42 to 78).
- General Score established postoperatively: the product of Prognostic Score and Vascular Score – minimum value = 4, maximum value = 8.
- Preoperative Vascular Score: the product of serum value of Angiogenin) and serum value of Interleukin 6 established preoperatively.
Two statistical software, Epiinfoä and Statistica 6.1 were used to calculate the following: epidemiological and statistical indices, statistical comparisons (ANOVA, a Parametric Test for Inequality of Population Means, Mann-Whitney/Wilcoxon Two-Sample Test for Inequal Population Variances and Chi-square 2=-tailed p), linear and multiple regressions, multivariate Cox proportional hazard analysis, survival rates compared by the log rank test.
For each category of data processing, (these were numerous and therefore unable to comply with space allocated by the editor), only the results that are most representative for discussions and conclusions are presented.
Table 1 shows the descriptive statistics indicators of quantitative variables for the general group of 124 patients included in the study.
The histopathological distribution of the tumors studied was as follows: Malignant=88 (71.0%); Benign=36 (29.0%). The 88 malignant tumors have been diagnosed mainly in advanced FIFO stages – 72(81,8%) in stages III and IV showing high grading of tumors – 41(47,8) grade 3.
From the five subdivision of the general group, in accordance to the first applied therapy, the greatest part of the patients underwent primary citoreductive surgery of the ovarian tumor as following: 1 (primary optimal citoreductive surgery followed by adjuvant chemotherapy = 54 (43.4%), 2 (suboptimal surgery followed by adjuvant chemotherapy = 24 (19.1%), 3 (optimal interval surgery) = 7 (5.8%), 4 (suboptimal interval surgery) = 3 (2.6%), 5 (surgery in cases of benign tumors) = 36 (29,1)%.
The recurrence of disease occurred in 26 (21%) of patients from the general group (recurrence or metastases) within the 36 months of the research.
The significant differences for values of well known predictors for ovarian cancer in the two groups (Malignant and Benign) were determined, in order to confirm the relevance of the sample. Afterward, the same procedures were applied to less used parameters, in order to identify and select the prognostic factors that can be considered predictors (independently or in association). Below you will find a few examples for the parameters whose average value presented a significant or highly significant difference in the subgroups of benign and malignant ovarian tumors.
- Known predictors: “PDECGF” averages: P value= 0.0073; “CA 125” averages: P-value = 0.0027; “Il6” averages: P-value = 0.0071.
- Potential ultrasonographic predictors: “Tumor size” averages: P-value = 0.0001; “Multiloculare solid masses insight the tumor”: P-value = 0.0001; “Presence of ascites”: P-value = 0.0009; “Bilaterality of the ovarian masses”: P-value = 0.0002. The malignancy risk for the “Presence of ascites” variable, determined ultrasonographically, was calculated: Risk Ratio (RR) =3.5648 (1.7228-7.3765) (-95% Confidence Interval).
- Potential serologic predictors: “Angiogenin”: P-value = 0.0022; “Angiopoetin 2”: P-value = 0.0019; “PDGF”: P-value = 0.0049; “VEGF”: P-value = 0.0012.
The significance of statistical difference was calculated in order to highlight the influence of various group parameters (qualitative) on possible predictors for the evolution or aggressiveness of ovarian cancer. The statistically significant or highly significant results are presented below:
· The significant difference of the “Disease-free interval” variable averages was calculated according to the place of origin (urban/rural): P value =0.0020.
· The significant difference of the “CA 125 Value” averages was calculated according to FIGO level (I, II=low FIGO level); III, IV= high FIGO level): P value =0.0163.
· The significant difference of the “PDECGF Level” averages was calculated according to FIGO stage (I, II= low FIGO level; III, IV= high FIGO level): P value =0.0077.
· The significant difference of the “Peak Systolic Velocity” variable averages determined using Doppler was calculated according to the return of the disease (1=recurrence or metastasis; 0= disease-free patient): P value =0.0317.
· The significant difference of the “CD-34” variable averages was calculated according to the “Presence of ascites” determined ultrasonographically: P value =0.0022.
· The significant difference of the “PDECGF” variable averages was calculated according to the “Presence of ascites” determined ultrasonographically: P value =0.0013.
· The significant difference of the “CA125” variable averages was calculated according to the “Presence of ascites” determined ultrasonographically: P value =0.0090
The Log-Rank and Wilcoxon survival indices were calculated for the 88 patients with malignant ovarian tumors). Kaplan Mayer survival curves were also drawn. The difference among curves and its significance, as well as the predictivity degree of various parameters, compared with the “Disease-free interval” temporal variable, were calculated. (Fig. 1)
When calculating univariate regression coefficients for the “Histopathology: Malignant/Benign” dependent
variable, the following studied parameters can be considered group predictive variables: therapeutic protocol, disease-free interval, CA 125 value, resistive index, pulsatility index, Peak Systolic Velocity, tumor size, multiloculare solid masses inside the tumor, bilaterality of the ovarian masses, and prognostic score.
The predictors of the disease-free interval, considered as a dependent variable, were the following: the patient’s place of origin, age at onset of disease, therapeutic protocol, PDECGF value, multiloculare solid masses inside the tumor, and diastolic notch. Only the last two parameters registered the required value for being considered independent predictors.
In order to reinforce the prognostic value of the cytokines for their inclusion in the group of early (preoperative), non-aggressive predictors, the difference of their average value was studied in connection with other ovarian cancer prognostic factors (independent variables). The significant difference of the “Angiogenin” variable averages was calculated according to the return of the disease: P-value = 0.041
To evaluate the discriminative power of proposed predictor indexes, called SCORES in this study, their correlation was tested versus different recognized prognostic factors (clinical and serological). (Table 2)
The Vascular Score values were also associated with disease recurrence (p=0.0009), and multilocular solid masses inside the tumor (P=0.0043); General Score values were associated with positive values of Ca125 (>35U/ml) (p=0.0067), with Central Blood Flow (P=0.0030) and with presence of Diastolic Notch (p=0.0200)
The association among potential preoperative and post-operative predictors had to be established for patients selected to undergo the same therapeutic protocol in order to formulate an early prediction on the aggressiveness of malignant ovarian tumors. The group of patients who underwent primary optimal surgical cytoreduction was chosen for this purpose since it was the largest of the groups of patients with malignant tumors.
The significance of the difference of the studied predictors in patients with and without recurrence of disease and the predictive cut off values for recurrence or metastasis was tested. The values bearing statistical significance are presented below (the cut off values were considered the maximum level of the group with lower average). (Table 3)
The main aim of the research – to establish associations between clinical, morphological, therapeutic and biological evolution and prognostic factors – was reached. Indices and scores that highlight the evolution of certain stages and histopathological types of ovarian cancer were elaborated. The main objective was also realized as the association studied was quantified by elaborating indexes able to predict the evolution and aggressiveness related to the angiogenic component in ovarian carcinoma. The number of cases in the study was the main limitation of the research. This restriction was due to the limited time allotted for this research and the number of cases referred to the Institute of Oncology, Cluj-Napoca.
The high proportion of malignant tumors (71%) of all cases included in the study is obviously explained by a pre-selection of the patients referred to the Institute of Oncology Cluj. The advanced stages at the time of diagnosis (81.8% in FIGO III and IV stages) can be attributed to the long asymptomatic period of the disease in the context of absence of a national early detection strategy.
The evaluation of prognostic factors by descriptive statistical indices presented in the beginning of the subchapter “Results” allows the sample to be placed within the value ranges of the studied parameters and certifies the relevance of the patient group included in the study, through the discriminatory capacity of prognostic factors for the malignancy of ovarian tumors. The same statistical comparisons confirmed the statistically significant differences by means of the malignant/benign groups for the potential ultrasonographic and serologic predictors (little used in current oncological practice).
Consecutively, the qualitative (grouping) factors, that could influence known or supposed evolution or aggressiveness predictors of ovarian cancer, were assessed. Possible associations between clinical, serologic, biological and ultrasonographic factors (e.g. CA125 value and tumor stage, PDECGF value and tumor stage or CD34, PDECGF and CAS125 values and the presence of ascites) could have been shown.
The relevance for the evolution of ovarian cancer, illustrated by the disease-free interval of (known or supposed) variables included in the study, was tested; only the FIGO study (Log rank P=0.023) could be considered an individual predictor. Previous studies have also shown the predictive significance of FIGO for global survival (5, 7, 8). The limited duration of the prospective study (36 months) did not allow valid estimates of global survival in the present study.
The calculation of univariate regression for group predictors of malignancy showed, along with the first therapeutic protocol applied, the disease-free interval and the positivity of the CA125 level, also presented in other studies (4,11,18,19), as well as other factors: 2D ultrasonographic aspects and Doppler indices.
In addition to already known predictors of the disease-free interval, such as age at onset of disease (3) or first therapeutic protocol applied (9), this study identified 2D and Doppler ultrasonographic predictors (diastolic notch and multilocular solid masses inside the tumor can be considered independent predictors in our study).
After testing the discriminatory capacity of some angiogenic cytokines (growth factors and interleukins) for malignancy in patients with ovarian tumors, the interrelation of highly significant factors in the evolution of ovarian cancer was tested. Highly significant values were only found for the recurrence of the disease.
In search of high predictions, resulting from the association or interaction of known prognostic factors, which were investigated by other researchers as well (6), we verified the relevance of three indices.
The preoperative vascular index (calculated as product of the preoperatively determined values of angiogenin and interleukin 2) is not highly correlated with Doppler indices but is highly correlated with the postoperative vascular index (the sum of CD-34 and PDECGF levels determined immunohistochemically from the paraffin blocks of the operative pieces): r2=0.65. The high level of the correlation shows that both indices similarly reflect the implication of angiogenic factors in the evolution of ovarian cancer. Thus, by using serological tests and by establishing a threshold of the preoperative vascular score value for the making of diagnosis, the therapeutic decision can be refined, by associating antiangiogenic therapy to neoadjuvant or adjuvant therapy.
The postoperative vascular index is highly correlated with preoperatively determined Doppler indices, thus reflecting the angiogenic component of tumor evolution. Calculation of the postoperative vascular index (which is extremely expensive and laborious) becomes redundant in the case of highly accurate serological, ultrasonographic and Dopper determinations.
Prognostic score is not correlated with Doppler indices or angiogenic cytokine values, which might suggest that it does not influence the angiogenic component of the disease evolution.
Table 2 shows that no index considered above is significantly correlated with the duration of the disease-free interval.
In order to assess tumor aggressiveness in the patients included in the study, the association between predictors and a parameter relevant for disease evolution (recurrence) was monitored in a group, as homogeneous as possible for this parameter (the group of patients who followed the same initial protocol: optimal primary surgical cytoreduction). In addition to CD34 and CA125 values, only a few factors illustrating angiogenic activity (Doppler indices and cytokines with angiogenic action) presented significant differences in the means of the groups with and without recurrence of the disease. Comparing the means of the same parameters for the group of patients with recurrence versus the group showing metastasis, an even smaller number of predictors proved to be relevant (Doppler resistivity index and interleukin Il 6 values). Even if these values were calculated for a relatively small group (54 patients), the results could stimulate, by a subsequent verification of accuracy, the performance of extensive studies (possibly trials) that might generate standards in this field.
1. The values of serum (cytokines, CA-125, etc), tissular (CD-34 and PDECGF) and ultrasonographic (the ultrasonographic risk score and Doppler indicators) predictors are highly correlated; therefore their value is indicative of the evolution and prognosis of ovarian cancer (for well-established stages).
2. The creation of indexes resulting from the studied predictors can indicate the aggressiveness of the tumor in his angiogenic component and support the decision of treatment selected.
Opportunities generated by the research
The conclusions of the research, question the use of classical TNM staging (FIGO for ovarian tumor cases) in the selection of therapy. This new pattern of clinical thinking suggests that when staged, histopathology, grading and therapy are similar and the aggressiveness of the disease may vary in ovarian cancer patients with the same biological parameters, thus requiring differential therapeutic approaches within the function of the angiogenic component of the tumor evolution.
1. Bartlett, JMS - Ovarian cancer: methods and protocols. 2000 Humana Press New Jersey.
2. von Georgi, R., Franke, F.E., Munstedt, K. - Influence of surgery, postoperative therapy and tumor characteristics on patient prognosis in advanced ovarian carcinomas. Eur. J. Obstet. Gynecol. Reprod. Biol., 2003, 111:189.
3. Paramasivam, S., Tripcony, L., Crandon, A., Quinn, M., Hammond, I., Marsden, D., Proietto, A., Davy, M., Carter, J., Nicklin, J., Perrin, L., Obermair, A. - Prognostic importance of preoperative CA-125 in International Federation of Gynecology and Obstetrics stage I epithelial ovarian cancer: an Australian multicenter study. J. Clin. Oncol., 2005, 23:5862.
4. Obermair, A., Fuller, A., Lopez-Varela, E., van Gorp, T., Vergote, I., Eaton, L., Fowler, J., Quinn, M., Hammond, I., Marsden, D., Proietto, A., Carter, J., Davy, M., Tripcony, L., Abu-Rustum, N. - A new prognostic model for FIGO stage 1 epithelial ovarian cancer. Gynecol. Oncol., 2007, 104:607.
5. Massi, D., Susini, T., Savino, L., Boddi, V., Amunni, G., Colafranceschi, M. - Epithelial ovarian tumors in the reproductive age group: age is not an independent prognostic factor. Cancer, 1996, 77:1131.
6. Kosary, C.L. - FIGO stage, histology, histologic grade, age and race as prognostic factors in determining survival for cancers of the female gynecological system: an analysis of 1973-87 SEER cases of cancers of the endometrium, cervix, ovary, vulva, and vagina. Semin. Surg. Oncol., 1994, 10:31.
7. Sevelda, P., Schemper, M., Spona, J. - CA 125 as an independent prognostic factor for survival in patients with epithelial ovarian cancer. Am. J. Obstet. Gynecol., 1989, 161:1213.
8. Geisler, J.P., Miller, G.A., Lee, T.H., Harwood, R.M., Wiemann, M.C., Geisler, H.E. - Relationship of preoperative serum CA-125 to survival in epithelial ovarian carcinoma. J. Reprod. Med., 1996, 41:140.
9. van Dalen A, Favier J, Baumgartner L, Hasholzner U, De Bruijn, H. - Prognostic significance of CA 125 and TPS levels after chemotherapy in ovarian cancer patients. Anticancer Res., 1999, 19:2523.
10. Pepper, M.S. - Manipulating angiogenesis: from basic science to the bedside. Rterioscler. Thromb. Vasc. Biol., 1997, 17:605.
11. Sharma, A., Bernacki, R.J. - Ovarian cancer patients with high CA-125 but not symptoms – should antiangiogenic treatments be considered? Oncol. Res., 1997, 9:53.
12. Hollingsworth, H.C., Kohn, E.C., Steinberg, S.M., Rothenberg, M.L., Merino, M.J. - Tumor angiogenesis in advanced stage ovarian carcinoma. Am. J. Pathol., 1995, 147:33.
13. Tingulstad, S., Skjeldestad, F.E., Hagen, B. - The effect of centralization of primary surgery on survival in ovarian cancer patients. Obstet. Gynecol., 2003, 102:499.
14. Li, A.J., Karlan, B.Y. - Surgical advances in the treatment of ovarian cancer. Hematol. Oncol. Clin. North Am., 2003, 17:945.
15. Berman, M.L. - Future directions in the surgical management of ovarian cancer. Gynecol. Oncol., 2003, 90:S33.
16. Eisenkop, S.M., Spirtos, M., Friedman, L., Lin, W.C., Pisani, A.L., Perticucci, S. - Relative influences of tumor volume before surgery and the cytoreductive outcome on survival for patients with advanced ovarian cancer. Gynecol. Oncol., 2003, 90:390.
17. Kecmanovic, D.M., Pavlov, M.J., Kovacevic, P.A., Ceranic, M.S., Stamenkovic, A.B. - Cytoreductive surgery for ovarian cancer. Eur. J. Surg. Oncol., 2003, 29:315.
18. Sato, N., Akahira, J., Kitamura, T., Aoki, M., Yoshinaga, K., Okamura, C., Utsunomiya, H., Shimizu, T., Niikura, H., Ito, K., Okamura, K., Yaegashi, N. - Prognostic significance of second-look laparotomy for surgically confirmed early-stage epithelial ovarian cancer: a multicenter retrospective study. Int. J. Clin. Oncol., 2003, 8:40.
19. Morice, P., Dubernard, G., Rey, A., Atallah, D., Pautier, P., Pomel, C., Lhomme, C., Duvillard, P., Castaigne, D. - Results of interval debulking surgery compared with primary debulking surgery in advanced stage ovarian cancer. J. Am. Coll. Surg., 2003, 197:955.
20. Engelen, M.J., Kos, H.E., Willemse, P.H., Aalders, J.G., de Vries, E.G., Schaapveld, M. - Surgery by consultant gynecologic oncologists improves survival in patients with ovarian carcinoma. Cancer, 2006, 106:589.
21. Lie, D. - USPSTF Recommends Against Routine Screening for Ovarian Cancer. Ann. Fam. Med., 2004, 2:260.
22. Bourne, T.H., Campbell, S., Reynolds, C. - Transvaginal color flow imaging: A possible new screening technique for ovarian cancer. Br. Med. J., 1989, 299:1367.
23. Achimas Cadariu, L., Lazãr, A., Irimie, M., Mureşan, D., Eniu. - Evolution and prognostic fators in ovarian cancer. Applied Medical Informatics, 2005, 16:34.
24. Achimaş Cadariu, P., Lazãr, L., Irimie, Al., Neagoe, Ioana, Buiga, R. - Predictori angiogenici biologici în cancerul ovarian. Radioterapie şi Oncologie Medicalã, 2006, 12:59.