- Email: [email protected]
Global distribution pattern of histological subtypes of epithelial ovarian cancer: A database analysis and systematic review
Gynecologic Oncology 133 (2014) 147–154
Contents lists available at ScienceDirect
Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Global distribution pattern of histological subtypes of epithelial ovarian cancer: A database analysis and systematic review Pi-Lin Sung a,b,1, Yen-Hou Chang a,b,1, Kuan-Chong Chao a, Chi-Mu Chuang a,b,c,⁎, Task Force on Systematic Review and Meta-analysis of Ovarian Cancer a b c
Section of Gynecologic Oncology, Department of Obstetrics and Gynecology, Taipei Veterans General Hospital, Taipei, Taiwan Institute of Clinical Medicine, School of Medicine, National Yang-Ming University, Taipei, Taiwan Institute of Public Health, School of Medicine, National Yang-Ming University, Taiwan
H I G H L I G H T S • There existed significant variations of subtype distribution among countries and regions in the world. • Serous and endometrioid subtype showed less distribution variation, while larger differences were seen in mucinous and clear cell subtype. • A guide map for selecting countries or regions to implement clinical trials for epithelial ovarian cancer was provided.
a r t i c l e
i n f o
Article history: Received 26 November 2013 Accepted 11 February 2014 Available online 17 February 2014 Keywords: Epithelial ovarian cancer Subtype Targeted therapy Clinical trials Global
a b s t r a c t Background. Epithelial ovarian cancer is basically a heterogeneous disease with different chemosensitivity and distinct molecular alternations for each histological subtype. In order to assess whether the results of clinical trials can be extrapolated to a new country, it is critical to first examine whether the relative frequencies is homogenous across countries. Methods. Cancer registry database from a single institution in Taiwan combined with systematic review of the global literature on the relative frequencies of histological subtypes between 2003 and 2012 was provided. Results. Of 175 titles identified, 41 studies met inclusion/exclusion criteria. Globally, for each subtype, the median value of relative frequencies for serous subtype was 45.0%, with the Philippines (16.0%), Indonesia (22.7%), and Brazil (30.1%) as the three lowest countries and South Africa (68.0%), Greece (71.5%), and India (86.7%) as the three highest countries; for mucinous subtype, 11.4%, Italy (3.0%), Australia (3.4%), and Japan (5.4%) were the three lowest countries, while Indonesia (29.1%), Singapore (30.3%), and South Korea (38.6%) were the three highest countries; for endometrioid subtype, 12.6%, India (1.6%), Greece (5.7%), and Portugal (7.6%) were the three lowest countries, while Taiwan (24.8%), Egypt (25.0%), and Austria (25.5%) were the three highest countries; and for clear cell subtype, 5.3%, Pakistan (1.0%), Iran (2.0%), and Brazil (2.1%) were the three lowest countries while Thailand (16.0%), Taiwan (16.8%), and Spain (18.8%) were the three highest countries. Conclusions. Relative frequencies of subtypes were not homogenous across countries. This diversity may reflect the geographical and ethnic variations. Globally, epithelial ovarian cancer is a heterogeneous disease with a heterogeneous distribution pattern. © 2014 Elsevier Inc. All rights reserved.
Introduction Among the gynecological malignancies, ovarian cancer is the leading cause of mortality in developed countries with estimated 225,500 new cases and 140,200 deaths worldwide [1]. In the United States, it is
⁎ Corresponding author at: Section of Gynecologic Oncology, Department of Obstetrics and Gynecology, Taipei Veterans General Hospital, 7F, No.201, Shih-Pai road, Beitou District, Taipei, Taiwan, 11217. Fax: +886 2 723 2788. E-mail address: [email protected] (C.-M. Chuang). 1 First and second author contributed equally to the study.
http://dx.doi.org/10.1016/j.ygyno.2014.02.016 0090-8258/© 2014 Elsevier Inc. All rights reserved.
estimated that 22,240 women will be diagnosed with ovarian cancer in 2013 among whom 14,030 will die [2]. The majority of ovarian cancer is of epithelial origin. The major histological subtypes of epithelial ovarian cancer include: serous, mucinous, endometrioid, clear cell, undifferentiated and unclassified [3]. Each of these subtypes is genetically distinct with unique molecular pathogenesis and different susceptibility to chemotherapeutic agent. Nevertheless, the regulatory mechanisms underlying this heterogeneity remain poorly understood [4,5]. Currently, clinical trials do not differentiate these subtypes but treat them as a homogeneous group. As a consequence, the results are difficult to interpret as it is not clear whether
148
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
the findings are applicable to a specific subtype or to another country if the subtype distribution pattern is not consistent between countries. Generally, U.S.- or E.U.-based pivotal multicenter clinical trials for epithelial ovarian cancer seldom include clinical centers in Asia, Central and South America, and Africa, which are referred as the “new region” under the setting of bridging trials. Results of clinical trials are often generalized without further investigation, and the findings are used to support new drug applications in the new regions [6]. Due to the fact that epithelial ovarian cancer is a heterogeneous disease with four major subtypes, it is imperative for researchers and clinicians to know the distribution pattern of the histological subtypes and the potential differences between countries or populations when evaluating the applicability of trial results to a new region. In the current study, we aimed to conduct a global systematic review to assess the distribution pattern of subtypes of epithelial ovarian cancer. Relative frequencies of each subtype were calculated using data from cancer registries, controlled clinical trials, cohort studies, or studies of archives of surgical samples. Furthermore, by employing cluster analysis, we explored the aggregation patterns among the countries examined on the basis of their similarities in subtype distributions. Methods Database retrieval Electronic database from a total of 648 primary epithelial ovarian cancer was retrieved, which is a stably constructed registration system of consecutively treated patients of ovarian cancer, set up in Taipei Veteran General Hospital between January 2003 and December 2012. All information was collected under protocols approved by a hospital Institutional Review Board. Search strategy of systematic reviews The systematic review was undertaken in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) [7]. For included observational studies, Meta-analysis of Observational Studies in Epidemiology (MOOSE) was further followed [8]. A comprehensive computerized systemic review of published reports, including cancer registry database, randomized controlled trials, cohort studies, and studies of surgical specimens, was performed by searching the following databases: Medline, EMBASE, Cancerlit, Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials, ISI Web of Science, and Google Scholar. The key search terms included ‘epithelial ovarian cancer,’ ‘serous,’ ‘mucinous,’ ‘endometroid,’ ‘clear cell,’ in combination with the following terms: ‘cohort,’ ‘controlled trials,’ ‘database,’ ‘survey,’ ‘epidemiology,’ ‘registry,’ ‘specimens,’ and ‘surgical archives’. The search was limited to human studies published in English from January 1990 to December 2012. Inclusion and exclusion criteria Included were studies that reported the distributions of all subtypes (serous, mucinous, endometroid, clear cell, and others) of epithelial ovarian cancer. We excluded studies (i) with inadequate sample size, defined as less than 50 cases or surgical archives, (ii) missing cases in any one of four major subtypes, (iii) subtype-specific studies, (iv) animal xenograft studies using human cancer cell lines, and (v) abstracts, letters and posters where the full study was not published. Screening and data extraction The systematic search described above was completed by March, 2013. Two independent reviewers (P.L.S. and C.M.C.) assessed the potential relevance of all titles and abstracts identified from the electronic searches. Full articles were retrieved for further assessment when the
abstracts indicated that they might meet the inclusion criteria. Disagreements were resolved through discussion and consensus. A third reviewer (M.S.Y.) was consulted in case of persisting disagreement. The reviewed data were extracted and entered on to an ad hoc standardized data entry form by each reviewer. Data extracted for comparison included study of origin (continent/country), year of publication, research design, number of cases for each subtype, length of recruitment period, source of information. Minor subtype (e.g. transitional and squamous), undifferentiated and non-otherwise specified carcinoma, are categorized under “others” category. Relative frequencies of each subtype were calculated for each retrieved article. The definition of relative frequency points to the relative percentage of one subtype to the overall epithelial subtypes. As such, the sum of relative frequency of the four major subtypes may not be equal to 100% due to the presence of others (including squamous, transitional cell, and undifferentiated carcinoma). As a rule, we selected one representative study from each country or region for final analysis. For Europe, because several randomized trials are purely derived from European countries, therefore we decided to include these studies for the purpose of comparison. For the United States, because of existing Surveillance, Epidemiology and End Results (SEER) program, therefore we decided not to include randomized clinical trials. Instead, we present two SEER results for the purpose of comparison. In decreasing order, the priority of study selection is database analysis, followed by randomized controlled trials, then by observational studies, and finally by studies of surgical archives. Assessment of methodological quality The quality of randomized controlled trials was evaluated using validated Jadad scoring system which ranges from 0 (bad) to 5 (good). On the basis of Jadad scoring, we dichotomized the quality of reporting into poor (score b 3) or good (score ≥ 3) [9]. Quality of observational studies (e.g. database analysis, cohort study, and surgical archives) was scored according to the Newcastle–Ottawa Quality Assessment Scale which ranges from 1 (poor) to 9 (excellent) [10]. Because there are no descriptive anchors for this scale except the lowest and highest score, we decided to classify studies with a total score equal to or greater than 7 as high-quality studies. Statistical analysis Studies were grouped using the United Nations classification, which categorizes the world into 5 macrogeographical (continental) regions and 22 geographical subregions [11]. Agreement on the inclusion of studies was assessed using a kappa statistic. Relative frequency of each subtype for each included studies was presented by continent. The coefficient of variation (CV) was calculated by dividing the standard deviation by the mean measurement of relative frequency. Discrepancy of relative frequency between countries was assessed by chi-square test. Hierarchical cluster analysis was conducted to classify countries according to the distributions of all subtypes by calculating the distances according to the closeness of between-country distances. In this analysis, we used an agglomerative clustering procedure based on standardized Euclidean distances and the average linkage algorithm [12]. All countries were represented by their relative positions on a cluster tree (dendrogram) that shows the similarities and dissimilarities between countries. The merging of countries with similar features leads to the formation of a cluster, where the length of the branch indicates the degree of relation. Thus, countries with tightly related features appear closer together, while the degree of separation in the cluster tree increases with further dissimilarity. In order to calculate the distances between all variables in the analysis, they were further standardized by transforming the data to have a mean = 0 and a variance = 1. All statistical tests were two-sided with a significance level of p-value of 0.05
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
unless otherwise specified. Statistical calculations were performed using STATA 12.0 (College Station, TX, USA). Heatmap plot was created by GENE-E, available at the website http://www.broadinstitute.org/ cancer/software/GENE-E/.
Results Our analysis included a ten-year database analysis in a single referral center along with 40 published studies [13–52], selected on the basis of the criteria reported previously, from which subtype distributions were calculated. Of the total 41 studies included, 17 (41.5%) were database analysis, 6 (14.6%) were controlled clinical trials, 4 (9.8%) were cohort studies, and 14 (34.1%) were from molecular studies of surgical sample archives. Characteristics of included studies and their quality score are listed in Supplementary Table. The exclusion/inclusion procedure is described according to the QUOROM-guidelines (Quality of Reporting of Meta-analyses) (Fig. 1) [53]. For the United States, there were two SEER database studies with comparable period of patients' recruitment and case numbers, therefore we decided to present pooled estimates of both studies. For European countries, we included several pan-Europe clinical trials with the intention for comparison. Quality assessment revealed that Jadad scores of all included clinical trials met the criteria of high quality (score ≥ 3), whereas New Newcastle–Ottawa Quality
149
Assessment Scale for some observational studies did not meet the criteria of high quality (score ≥ 7). Globally, for each subtype, the median value of relative frequencies for serous subtype was 45.0%, with the Philippines (16.0%), Indonesia (22.7%), and Brazil (30.1%) as the three lowest countries and South Africa (68.0%), Greece (71.5%), and India (86.7%) as the three highest countries; for mucinous subtype, the median value was 11.4%, with the three lowest countries being Italy (3.0%), Australia (3.4%), and Japan (5.4%) and the three highest countries being Indonesia (29.1%), Singapore (30.3%), and South Korea (38.6%); for endometrioid subtype, the median value was 12.6%, with the three lowest countries being India (1.6%), Iran (4.9%), Greece (5.7%), and the three highest countries being Taiwan (24.8%), Egypt (25.0%), and Austria (25.5%); and for clear cell subtype, the median value was 5.3%, with the three lowest countries being Pakistan (1.0%), Iran (2.0%), and Germany (2.0%) and the three highest countries being Thailand (16.0%), Taiwan (16.8%), and Spain (18.8%). Due to significant inter-country variability of relative frequencies, we further calculated the coefficient of variation (CV) for each subtype. The CV for serous subtype was 29.8%, mucinous subtype 58.3%, endometrioid subtype 42.3%, and clear cell subtype 62.7%. These estimates indicated that serous subtype appeared to have more consistent global distribution, followed by endometrioid subtype, whereas mucinous and clear cell subtype varied significantly across countries (Figs. 2–5).
Fig. 1. Quorum flow diagram of the steps in the reviewing process.
150
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
Fig. 2. Geographical distribution of relative frequency of serous subtype. CV, coefficient of variation. Relative frequency is calculated by number of serous subtype / number of (four major subtypes + minor subtypes). USA SEER data were pooled estimates of references [17] and [18]. Filled circles denote statistically higher percentage than SEER, whereas empty circles denote statistically lower percentage.
Hierarchical cluster analysis revealed that distribution pattern of subtypes can be broadly divided into five clusters (Fig. 6). Cluster 1 is a large category, composed of the United States, Canada, and countries
of European Union. Taiwan is in cluster 2, with a close proximity to Spain. Cluster 3 is composed of a mixture of Asian, Africa, and European countries. Cluster 4 was composed of exclusively Asian countries, such
Fig. 3. Geographical distribution of relative frequency of mucinous subtype. CV, coefficient of variation. Relative frequency is calculated by number of mucinous subtype / number of (four major subtypes + minor subtypes). USA SEER data were pooled estimates of references [17] and [18]. Filled circles denote statistically higher percentage than SEER, whereas empty circles denote statistically lower percentage.
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
151
Fig. 4. Geographical distribution of relative frequency of endometrioid subtype. CV, coefficient of variation. Relative frequency is calculated by number of endometrioid subtype/number of (four major subtypes + minor subtypes). USA SEER data were pooled estimates of references [17] and [18]. Filled circles denote statistically higher percentage than SEER, whereas empty circles denote statistically lower percentage.
as Singapore, South Korea, Philippines, and Indonesia. Cluster 5 comprised only one country, India, due to its unique distribution pattern. The dendrogram shows that countries in which the major pharmaceutical
discoveries and development for ovarian cancer are located on the left side of the figure (e.g. United States, Switzerland, France, Germany, and United Kingdom in cluster 1), whereas the major countries for drug
Fig. 5. Geographical distribution of relative frequency of clear cell subtype. CV, coefficient of variation. Relative frequency is calculated by number of clear cell subtype number of (four major subtypes + minor subtypes). USA SEER data were pooled estimates of references [17] and [18]. Filled circles denote statistically higher percentage than SEER, whereas empty circles denote statistically lower percentage.
152
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
Fig. 6. Dendrogram of hierarchical cluster analysis of distribution pattern of subtypes among countries. The lengths of the horizontal lines in the dendrogram provide relative qualitative information about the linkage distance between various clusters.
importation (i.e., China, South Korea, Singapore, Japan, and India in clusters 3–5) are on the right side of the dendrogram. Discussion Contrary to the common belief that the distribution pattern of histological subtypes of epithelial ovarian cancer is consistent across countries, instead, our systematic review revealed that there existed significant variations among countries and regions in the world. Among subtypes, fewer variations were observed in serous and endometrioid subtype, while larger differences were seen in mucinous and clear cell carcinoma. Among the four major subtypes, the latter two subtypes (mucinous and clear) are thought to be relative resistant to traditional chemotherapy [54]. It has been recognized that ovarian cancer is a heterogeneous disease with differential chemosensitivity and distinct molecular alternations for each subtype [4,54,55]. The large variations in subtype distribution patterns across countries and regions raise the issue of bridging study, which is designed to bridge a newly developed medication from the original region's (foreign) data in the original population to a new region [56]. The International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) had issued E5 document entitled “Ethnic Factors in the Acceptability of Foreign Clinical Data” in 1998. The ICH-E5 guideline provides a general framework for evaluating the potential impact of ethnic factors on the acceptability of foreign clinical data [57]. Differential distribution pattern in a specific country or region is one of the important ethnic factors which should be considered first when a new agent is to be registered to a new region for the treatment of epithelial ovarian cancer. To date, most landmark clinical trials of chemotherapeutic or molecular targeted agents for epithelial ovarian cancer were conducted in North America and/or Europe. When the approved agent is to be registered to a new region, pharmacokinetic bridging studies are usually requested by regulatory agencies of the new region, mainly countries in Asia, Africa, and Latin America, to extrapolate the trial results. However, the hierarchical cluster analysis in the current study
demonstrated that the countries with relative large market capability in new region (e.g. Japan, China, South Korea, Singapore, and India in cluster 3–5 in Fig. 6) showed significant different distribution patterns compared to countries in the United States, Canada, Switzerland, France, Germany, and United Kingdom (clusters 1–2), the major countries of drug development and export. Accordingly, the regulatory agency in the new region may consider conducting more complete clinical trials instead of merely implementing regular pharmacodynamic or pharmacokinetic bridging studies. Alternatively, considering the uncertainty of drug efficacy and safety due to the heterogeneous subtype distributions across countries, it is beneficial for the pharmaceutical industry to conduct multi-regional trials, designed for simultaneous drug development, submission, and approval around the world for epithelial ovarian cancer [58]. For the development of subtype-specific molecular targeted agents, our work provides a guiding map by which pharmaceutical industry can identify the candidate countries where high percentages of a specific subtype are observed to recruit patients more efficiently. The frequencies of both endometrioid and clear cell subtypes are relatively high in Spain and Taiwan. These two subtypes are closely linked to a distinct entity of endometriosis-associated ovarian carcinoma (EAOC) that is etiologically distinct from other ovarian carcinomas in several aspects [59]. Studies have documented that endometriosis is associated with approximately three-fold risk of endometrioid and clear cell subtype. Patients with EAOC had a lower stage of cancer, a distribution of histological subtypes that differs from the general population (i.e. endometrioid and clear cell cancer being the most common), predominantly lower grade endometriosis lesions, and significantly better overall survival as compared with other ovarian carcinomas [60,61]. Whether the populations in these two countries bear higher genetic polymorphisms associated with endometriosis-related ovarian cancer warrants further investigations. There are several limitations in the present study. First, there is a lack of adequate information from Africa and Latin America. Given this glaring lacking of available data, it is recommended that a global consortium for ovarian cancer registration be established to collect more update and
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
integrated resources for pharmaceutical development and academic research. Second, data from some countries were only available from surgical specimen archives in which selection bias may occur, thus raising some doubts whether the observed subtype distributions reflected the true values in the population. Third, the period of patients' recruitment was not consistent for all included studies. Some countries (e.g. Netherlands, Norway, and South Africa) presented the data that predated the routine use of monoclonal antibodies to differentiate subtype assignments, including high-grade endometrioid carcinoma versus high-grade serous carcinoma, undifferentiated histological types versus specifically differentiated histological subtype, and mixed histological subtypes versus pure histological subtype [62]. Fourth, data from some countries (South Africa, Greece, and Spain) did not provide details for “others” subtypes, thus the relative frequency estimates of four major subtypes may be inflated. Fifth, not all included studies mentioned standardized procedure of central pathology review, leaving the quality of subtype diagnosis a major concern. Sixth, data from some countries (Japan, Norway, and Greece) did not cover all stages, potentially making serous subtype over-presented. Seventh, many included papers predate the proposal of two-tiered grading system of serous carcinoma [63], precluding the presentation of data of low-grade versus high-grade serous subtype in our work. In summary, our analysis provides an overview of global distribution pattern of subtypes in epithelial ovarian cancer. By cluster analysis, we revealed that there existed considerable differences in distribution pattern of subtypes between most Asian countries and those in North American and Europe. These geographical and ethnic differences are by no means negligible, and they warrant further investigation. These figures may serve as a reference in the design and implementation of bridging studies when a medication is submitted for registration in a new region or country. Furthermore, this work provides a guide map for pharmaceutical industry to select pertinent countries or regions to implement clinical trials for epithelial ovarian cancer in the future. Finally, the glaring absence of studies from Africa and Latin America highlights the urgent need to construct a global network consortium that garners and integrates databases from countries around the world in a comprehensive and contemporary fashion. In addition to being a heterogeneous disease, epithelial ovarian cancer further demonstrates heterogeneous distribution patterns of subtypes. When designing clinical trials for epithelial ovarian cancer, it is imperative to take into account these heterogeneous distribution patterns. In order to accelerate drug development, submission, and approval around the world, implementing multi-regional trial at the initial design stage of new drug clinical trials for epithelial ovarian cancer is recommended. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2014.02.016. Conflict of interest statement The authors declared no conflict of interest with regard to any drugs or materials relevant to this paper.
Funding This work was supported both by Taiwan National Science Council NSC 101-2314-B-075-071 and Taipei Veterans General Hospital Grant V102C-077. References [1] Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:69–90. [2] Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2013;63:11–30. [3] Tavassoli FA, Devilee P. IARC WHO Classification of Tumours, No 4, Pathology and genetics of tumors of the breast and female genital organs. IARC Press; 2003. [4] McCluggage WG. My approach to and thoughts on the typing of ovarian carcinomas. J Clin Pathol 2008;61:152–63.
153
[5] Shih Ie M, Kurman RJ. Ovarian tumorigenesis: a proposed model based on morphological and molecular genetic analysis. Am J Pathol 2004;164:1511–8. [6] Shih WJ. Clinical trials for drug registrations in Asian-Pacific countries: proposal for a new paradigm from a statistical perspective. Control Clin Trials 2001;22:357–66. [7] Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009;151:264–9. [8] Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Metaanalysis of observational studies in epidemiology: a proposal for reporting. Metaanalysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283:2008–12. [9] Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJM, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:1–12. [10] Wells G, Shea B, Peterson J, Welch V, Losos M, Tugwell P. The Newcastle–Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. 3rd symposium on systematic reviews: beyond the basics; 2000. p. 3–5. [11] Division UNS. Standard country and area codes classifications (M49). Available from URL: http://unstats.un.org/unsd/methods/m49/m49regin.htm. [accessed 20 May, 2013]. [12] Agresti A. Categorical data analysis. John Wiley & Sons; 2002. [13] Alshenawy HA. Immunohistochemical expression of epidermal growth factor receptor, E-cadherin, and matrix metalloproteinase-9 in ovarian epithelial cancer and relation to patient deaths. Ann Diagn Pathol 2010;14:387–95. [14] Guidozzi F, Daponte A. Estrogen replacement therapy for ovarian carcinoma survivors: a randomized controlled trial. Cancer 1999;86:1013–8. [15] Paes MF, Daltoé RD, Madeira KP, Rezende LCD, Sirtoli GM, Herlinger AL, et al. A retrospective analysis of clinicopathological and prognostic characteristics of ovarian tumors in the State of Espirito Santo, Brazil. J Ovarian Res 2011;4:14–9. [16] Zhang J, Ugnat AM, Clarke K, Mao Y. Ovarian cancer histology-specific incidence trends in Canada 1969–1993: age-period-cohort analyses. Br J Cancer 1999;81:152–8. [17] Chan JK, Teoh D, Hu JM, Shin JY, Osann K, Kapp DS. Do clear cell ovarian carcinomas have poorer prognosis compared to other epithelial cell types? A study of 1411 clear cell ovarian cancers. Gynecol Oncol 2008;109:370–6. [18] Quirk JT, Natarajan N. Ovarian cancer incidence in the United States, 1992–1999. Gynecol Oncol 2005;97:519–23. [19] Yang X, Hou M, Yang K, Wang H, Peng Z, Cao Z, et al. Prognosis in epithelial ovarian cancer: clinical analysis of 287 pelvic and para-aortic lymphadenectomy. Chin Ger J Clin Oncol 2007;6:492–6. [20] Chitrathara K, Sheikh ZA, Vijaykumar DK, Kuriakose S, Anupama R, Nandeesh M. Is hysterectomy needed in ovarian cancer? Indian J Cancer 2011;48:471–6. [21] Koh SCL, Huak CY, Lutan D, Marpuang J, Ketut S, Budiana NG, et al. Combined panel of serum human tissue kallikreins and CA-125 for the detection of epithelial ovarian cancer. J Gynecol Oncol 2012;23:175–81. [22] Arab M, Khayamzadeh M, Hashemi M, Hosseini M, Tabatabaeefar M, Anbiaee R, et al. Crude and age-specific incidence rate patterns for histopathologic subtypes of ovarian cancer in Iran. Arch Iran Med 2010;13:203–8. [23] Katsumata N, Yasuda M, Takahashi F, Isonishi S, Jobo T, Aoki D, et al. Dose-dense paclitaxel once a week in combination with carboplatin every 3 weeks for advanced ovarian cancer: a phase 3, open-label, randomised controlled trial. Lancet 2009;374:1331–8. [24] Chung HH, Hwang SY, Jung KW, Won YJ, Shin HR, Kim JW, et al. Ovarian cancer incidence and survival in Korea: 1993–2002. Int J Gynecol Cancer 2007;17:595–600. [25] Thanikasalam K, Ho CM, Adeed N, Shahidan MN, Azizah WK. Pattern of ovarian tumours among Malaysian women at General Hospital, Kuala Lumpur. Med J Malaysia 1992;47:139–46. [26] Bhurgri Y, Shaheen Y, Kayani N, Nazir K, Ahmed R, Usman A, et al. Incidence, trends and morphology of ovarian cancer in Karachi (1995–2002). Asian Pac J Cancer Prev 2011;12:1567–71. [27] Redaniel MTM, Laudico A, Mirasol-Lumague MR, Gondos A, Uy GL, Toral JA, et al. Ovarian cancer survival population differences: a “high resolution study” comparing Philippine residents, and Filipino-Americans and Caucasians living in the US. BMC Cancer 2009;9:1–8. [28] Zhao C, Annamalai L, Guo C, Kothandaraman N, Koh SCL, Zhang H, et al. Circulating haptoglobin is an independent prognostic factor in the sera of patients with epithelial ovarian cancer. Neoplasia 2007;9:1–7. [29] Tangjitgamol S, Manusirivithaya S, Khunnarong J, Jesadapatarakul S, Tanwanich S. Expressions of estrogen and progesterone receptors in epithelial ovarian cancer: a clinicopathologic study. Int J Gynecol Cancer 2009;19:620–7. [30] Piccart MJ, Bertelsen K, James K, Cassidy J, Mangioni C, Simonsen E, et al. Randomized intergroup trial of cisplatin-paclitaxel versus cisplatin-cyclophosphamide in women with advanced epithelial ovarian cancer: three-year results. J Natl Cancer Inst 2000;92:699–708. [31] ICON2: randomised trial of single-agent carboplatin against three-drug combination of CAP (cyclophosphamide, doxorubicin, and cisplatin) in women with ovarian cancer. ICON Collaborators. International Collaborative Ovarian Neoplasm Study. Lancet 1998;352:1571–6. [32] International Collaborative Ovarian Neoplasm G. Paclitaxel plus carboplatin versus standard chemotherapy with either single-agent carboplatin or cyclophosphamide, doxorubicin, and cisplatin in women with ovarian cancer: the ICON3 randomised trial. Lancet 2002;360:505–15. [33] Perren TJ, Swart AM, Pfisterer J, Ledermann JA, Pujade-Lauraine E, Kristensen G, et al. A phase 3 trial of bevacizumab in ovarian cancer. N Engl J Med 2011;365:2484–96. [34] Gram IT, Lukanova A, Brill I, Braaten T, Lund E, Lundin E, et al. Cigarette smoking and risk of histological subtypes of epithelial ovarian cancer in the EPIC cohort study. Int J Cancer 2012;130:2204–10.
154
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
[35] Stadlmann S, Pollheimer J, Moser PL, Raggi A, Amberger A, Margreiter R, et al. Cytokine-regulated expression of collagenase-2 (MMP-8) is involved in the progression of ovarian cancer. Eur J Cancer 2003;39:2499–505. [36] Sedlakova I, Vavrova J, Tosner J, Hanousek L. Lysophosphatidic acid in patients with ovarian cancer. Ceska Gynekol 2010;3:41–6. [37] Soegaard M, Jensen A, Høgdall E, Christensen L, Høgdall C, Blaakær J, et al. Different risk factor profiles for mucinous and nonmucinous ovarian cancer: results from the Danish MALOVA study. Cancer Epidemiol Biomarkers Prev 2007;16:1160–6. [38] Kumpulainen S, Sankila R, Leminen A, Kuoppala T, Komulainen M, Puistola U, et al. The effect of hospital operative volume, residual tumor and first-line chemotherapy on survival of ovarian cancer — a prospective nation-wide study in Finland. Gynecol Oncol 2009;115:199–203. [39] Hamidou Z, Causeret S, Dabakuyo TS, Gentil J, Arnould L, Roignot P, et al. Populationbased study of ovarian cancer in Cote d'Or: prognostic factors and trends in relative survival rates over the last 20 years. BMC Cancer 2010;10:1–8. [40] Tianhui C, Lina J, Adam G, Katharina E, Bernd H, Alexander K, et al. Survival of ovarian cancer patients in Germany in the early 21st century: a period analysis by age, histology, laterality, and stage. Eur J Cancer Prev 2013;22:59–67. [41] Bamias A, Psaltopoulou T, Sotiropoulou M, Haidopoulos D, Lianos E, Bournakis E, et al. Mucinous but not clear cell histology is associated with inferior survival in patients with advanced stage ovarian carcinoma treated with platinum-paclitaxel chemotherapy. Cancer 2010;116:1462–8. [42] Ireland NCR. Cancer trends — cancers of the ovary, 14; 2012. p. 1–4. [43] Pignata S, Scambia G, Ferrandina G, Savarese A, Sorio R, Breda E, et al. Carboplatin plus paclitaxel versus carboplatin plus pegylated liposomal doxorubicin as firstline treatment for patients with ovarian cancer: the MITO-2 randomized phase III trial. J Clin Oncol 2011;29:3628–35. [44] Schouten LJ, Goldbohm RA, van den Brandt PA. Height, weight, weight change, and ovarian cancer risk in the Netherlands cohort study on diet and cancer. Am J Epidemiol 2003;157:424–33. [45] Wang Y, Helland A, Holm R, Kristensen GB, Borresen-Dale AL. PIK3CA mutations in advanced ovarian carcinomas. Hum Mutat 2005;25:1–5. [46] Dobrzycka B, Terlikowski SJ, Kinalski M, Kowalczuk O, Niklinska W, Chyczewski L. Circulating free DNA and p53 antibodies in plasma of patients with ovarian epithelial cancers. Ann Oncol 2011;22:1133–40. [47] Granstrom C, Sundquist J, Hemminki K. Population attributable fractions for ovarian cancer in Swedish women by morphological type. Br J Cancer 2008;98:199–205. [48] Levi F, Randimbison L, Blanc-Moya R, La Vecchia C. Second neoplasms after invasive and borderline ovarian cancer. Eur J Cancer Prev 2009;18:216–9.
[49] de la Torre FJ, Garcia A, Gil-Moreno A, Planaguma J, Reventos J, Cajal SR, et al. Apoptosis in epithelial ovarian tumors: Prognostic significance of clinical and histopathologic factors and its association with the immunohistochemical expression of apoptotic regulatory proteins (p53, bcl-2 and bax). Eur J Obstet Gynecol Reprod Biol 2007;130:121–8. [50] Faleiro-Rodrigues C, Macedo-Pinto I, Pereira D, Lopes CS. Prognostic value of E-cadherin immunoexpression in patients with primary ovarian carcinomas. Ann Oncol 2004;15:1535–42. [51] Beral V, Million Women Study C, Bull D, Green J, Reeves G. Ovarian cancer and hormone replacement therapy in the Million Women Study. Lancet 2007;369:1703–10. [52] Olsen CM, Nagle CM, Whiteman DC, Purdie DM, Green AC, Webb PM, et al. Body size and risk of epithelial ovarian and related cancers: a population-based case–control study. Int J Cancer 2008;123:450–6. [53] Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of reporting of meta-analyses. Lancet 1999;354:1896–900. [54] McCluggage WG. Morphological subtypes of ovarian carcinoma: a review with emphasis on new developments and pathogenesis. Pathology 2011;43:420–32. [55] Banerjee S, Kaye SB. New strategies in the treatment of ovarian cancer: current clinical perspectives and future potential. Clin Cancer Res 2013;19:961–8. [56] Gupta SK. Implications of ICH-E5: assessment of drug's sensitivity to ethnic factors and necessity of a bridging study for global drug development. Perspect Clin Res 2011;2:121–3. [57] ICH E5(R1): ethnic factors in the acceptability of foreign clinical data; Feb 1998. [58] Tsong Y, Chang WJ, Dong X, Tsou HH. Assessment of regional treatment effect in a multiregional clinical trial. J Biopharm Stat 2012;22:1019–36. [59] Worley MJ, Welch WR, Berkowitz RS, Ng SW. Endometriosis-associated ovarian cancer: a review of pathogenesis. Int J Mol Sci 2013;14:5367–79. [60] Pearce CL, Templeman C, Rossing MA, Lee A, Near AM, Webb PM, et al. Association between endometriosis and risk of histological subtypes of ovarian cancer: a pooled analysis of case–control studies. Lancet Oncol 2012;13:385–94. [61] Gayther SA, Song H, Ramus SJ, Kjaer SK, Whittemore AS, Quaye L, et al. Tagging single nucleotide polymorphisms in cell cycle control genes and susceptibility to invasive epithelial ovarian cancer. Cancer Res 2007;67:3027–35. [62] McCluggage WG, Young RH. Immunohistochemistry as a diagnostic aid in the evaluation of ovarian tumors. Semin Diagn Pathol 2005;22:3–32. [63] Malpica A, Deavers MT, Lu K, Bodurka DC, Atkinson EN, Gershenson DM, et al. Grading ovarian serous carcinoma using a two-tier system. Am J Surg Pathol 2004;28:496–504.
Contents lists available at ScienceDirect
Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Global distribution pattern of histological subtypes of epithelial ovarian cancer: A database analysis and systematic review Pi-Lin Sung a,b,1, Yen-Hou Chang a,b,1, Kuan-Chong Chao a, Chi-Mu Chuang a,b,c,⁎, Task Force on Systematic Review and Meta-analysis of Ovarian Cancer a b c
Section of Gynecologic Oncology, Department of Obstetrics and Gynecology, Taipei Veterans General Hospital, Taipei, Taiwan Institute of Clinical Medicine, School of Medicine, National Yang-Ming University, Taipei, Taiwan Institute of Public Health, School of Medicine, National Yang-Ming University, Taiwan
H I G H L I G H T S • There existed significant variations of subtype distribution among countries and regions in the world. • Serous and endometrioid subtype showed less distribution variation, while larger differences were seen in mucinous and clear cell subtype. • A guide map for selecting countries or regions to implement clinical trials for epithelial ovarian cancer was provided.
a r t i c l e
i n f o
Article history: Received 26 November 2013 Accepted 11 February 2014 Available online 17 February 2014 Keywords: Epithelial ovarian cancer Subtype Targeted therapy Clinical trials Global
a b s t r a c t Background. Epithelial ovarian cancer is basically a heterogeneous disease with different chemosensitivity and distinct molecular alternations for each histological subtype. In order to assess whether the results of clinical trials can be extrapolated to a new country, it is critical to first examine whether the relative frequencies is homogenous across countries. Methods. Cancer registry database from a single institution in Taiwan combined with systematic review of the global literature on the relative frequencies of histological subtypes between 2003 and 2012 was provided. Results. Of 175 titles identified, 41 studies met inclusion/exclusion criteria. Globally, for each subtype, the median value of relative frequencies for serous subtype was 45.0%, with the Philippines (16.0%), Indonesia (22.7%), and Brazil (30.1%) as the three lowest countries and South Africa (68.0%), Greece (71.5%), and India (86.7%) as the three highest countries; for mucinous subtype, 11.4%, Italy (3.0%), Australia (3.4%), and Japan (5.4%) were the three lowest countries, while Indonesia (29.1%), Singapore (30.3%), and South Korea (38.6%) were the three highest countries; for endometrioid subtype, 12.6%, India (1.6%), Greece (5.7%), and Portugal (7.6%) were the three lowest countries, while Taiwan (24.8%), Egypt (25.0%), and Austria (25.5%) were the three highest countries; and for clear cell subtype, 5.3%, Pakistan (1.0%), Iran (2.0%), and Brazil (2.1%) were the three lowest countries while Thailand (16.0%), Taiwan (16.8%), and Spain (18.8%) were the three highest countries. Conclusions. Relative frequencies of subtypes were not homogenous across countries. This diversity may reflect the geographical and ethnic variations. Globally, epithelial ovarian cancer is a heterogeneous disease with a heterogeneous distribution pattern. © 2014 Elsevier Inc. All rights reserved.
Introduction Among the gynecological malignancies, ovarian cancer is the leading cause of mortality in developed countries with estimated 225,500 new cases and 140,200 deaths worldwide [1]. In the United States, it is
⁎ Corresponding author at: Section of Gynecologic Oncology, Department of Obstetrics and Gynecology, Taipei Veterans General Hospital, 7F, No.201, Shih-Pai road, Beitou District, Taipei, Taiwan, 11217. Fax: +886 2 723 2788. E-mail address: [email protected] (C.-M. Chuang). 1 First and second author contributed equally to the study.
http://dx.doi.org/10.1016/j.ygyno.2014.02.016 0090-8258/© 2014 Elsevier Inc. All rights reserved.
estimated that 22,240 women will be diagnosed with ovarian cancer in 2013 among whom 14,030 will die [2]. The majority of ovarian cancer is of epithelial origin. The major histological subtypes of epithelial ovarian cancer include: serous, mucinous, endometrioid, clear cell, undifferentiated and unclassified [3]. Each of these subtypes is genetically distinct with unique molecular pathogenesis and different susceptibility to chemotherapeutic agent. Nevertheless, the regulatory mechanisms underlying this heterogeneity remain poorly understood [4,5]. Currently, clinical trials do not differentiate these subtypes but treat them as a homogeneous group. As a consequence, the results are difficult to interpret as it is not clear whether
148
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
the findings are applicable to a specific subtype or to another country if the subtype distribution pattern is not consistent between countries. Generally, U.S.- or E.U.-based pivotal multicenter clinical trials for epithelial ovarian cancer seldom include clinical centers in Asia, Central and South America, and Africa, which are referred as the “new region” under the setting of bridging trials. Results of clinical trials are often generalized without further investigation, and the findings are used to support new drug applications in the new regions [6]. Due to the fact that epithelial ovarian cancer is a heterogeneous disease with four major subtypes, it is imperative for researchers and clinicians to know the distribution pattern of the histological subtypes and the potential differences between countries or populations when evaluating the applicability of trial results to a new region. In the current study, we aimed to conduct a global systematic review to assess the distribution pattern of subtypes of epithelial ovarian cancer. Relative frequencies of each subtype were calculated using data from cancer registries, controlled clinical trials, cohort studies, or studies of archives of surgical samples. Furthermore, by employing cluster analysis, we explored the aggregation patterns among the countries examined on the basis of their similarities in subtype distributions. Methods Database retrieval Electronic database from a total of 648 primary epithelial ovarian cancer was retrieved, which is a stably constructed registration system of consecutively treated patients of ovarian cancer, set up in Taipei Veteran General Hospital between January 2003 and December 2012. All information was collected under protocols approved by a hospital Institutional Review Board. Search strategy of systematic reviews The systematic review was undertaken in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) [7]. For included observational studies, Meta-analysis of Observational Studies in Epidemiology (MOOSE) was further followed [8]. A comprehensive computerized systemic review of published reports, including cancer registry database, randomized controlled trials, cohort studies, and studies of surgical specimens, was performed by searching the following databases: Medline, EMBASE, Cancerlit, Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials, ISI Web of Science, and Google Scholar. The key search terms included ‘epithelial ovarian cancer,’ ‘serous,’ ‘mucinous,’ ‘endometroid,’ ‘clear cell,’ in combination with the following terms: ‘cohort,’ ‘controlled trials,’ ‘database,’ ‘survey,’ ‘epidemiology,’ ‘registry,’ ‘specimens,’ and ‘surgical archives’. The search was limited to human studies published in English from January 1990 to December 2012. Inclusion and exclusion criteria Included were studies that reported the distributions of all subtypes (serous, mucinous, endometroid, clear cell, and others) of epithelial ovarian cancer. We excluded studies (i) with inadequate sample size, defined as less than 50 cases or surgical archives, (ii) missing cases in any one of four major subtypes, (iii) subtype-specific studies, (iv) animal xenograft studies using human cancer cell lines, and (v) abstracts, letters and posters where the full study was not published. Screening and data extraction The systematic search described above was completed by March, 2013. Two independent reviewers (P.L.S. and C.M.C.) assessed the potential relevance of all titles and abstracts identified from the electronic searches. Full articles were retrieved for further assessment when the
abstracts indicated that they might meet the inclusion criteria. Disagreements were resolved through discussion and consensus. A third reviewer (M.S.Y.) was consulted in case of persisting disagreement. The reviewed data were extracted and entered on to an ad hoc standardized data entry form by each reviewer. Data extracted for comparison included study of origin (continent/country), year of publication, research design, number of cases for each subtype, length of recruitment period, source of information. Minor subtype (e.g. transitional and squamous), undifferentiated and non-otherwise specified carcinoma, are categorized under “others” category. Relative frequencies of each subtype were calculated for each retrieved article. The definition of relative frequency points to the relative percentage of one subtype to the overall epithelial subtypes. As such, the sum of relative frequency of the four major subtypes may not be equal to 100% due to the presence of others (including squamous, transitional cell, and undifferentiated carcinoma). As a rule, we selected one representative study from each country or region for final analysis. For Europe, because several randomized trials are purely derived from European countries, therefore we decided to include these studies for the purpose of comparison. For the United States, because of existing Surveillance, Epidemiology and End Results (SEER) program, therefore we decided not to include randomized clinical trials. Instead, we present two SEER results for the purpose of comparison. In decreasing order, the priority of study selection is database analysis, followed by randomized controlled trials, then by observational studies, and finally by studies of surgical archives. Assessment of methodological quality The quality of randomized controlled trials was evaluated using validated Jadad scoring system which ranges from 0 (bad) to 5 (good). On the basis of Jadad scoring, we dichotomized the quality of reporting into poor (score b 3) or good (score ≥ 3) [9]. Quality of observational studies (e.g. database analysis, cohort study, and surgical archives) was scored according to the Newcastle–Ottawa Quality Assessment Scale which ranges from 1 (poor) to 9 (excellent) [10]. Because there are no descriptive anchors for this scale except the lowest and highest score, we decided to classify studies with a total score equal to or greater than 7 as high-quality studies. Statistical analysis Studies were grouped using the United Nations classification, which categorizes the world into 5 macrogeographical (continental) regions and 22 geographical subregions [11]. Agreement on the inclusion of studies was assessed using a kappa statistic. Relative frequency of each subtype for each included studies was presented by continent. The coefficient of variation (CV) was calculated by dividing the standard deviation by the mean measurement of relative frequency. Discrepancy of relative frequency between countries was assessed by chi-square test. Hierarchical cluster analysis was conducted to classify countries according to the distributions of all subtypes by calculating the distances according to the closeness of between-country distances. In this analysis, we used an agglomerative clustering procedure based on standardized Euclidean distances and the average linkage algorithm [12]. All countries were represented by their relative positions on a cluster tree (dendrogram) that shows the similarities and dissimilarities between countries. The merging of countries with similar features leads to the formation of a cluster, where the length of the branch indicates the degree of relation. Thus, countries with tightly related features appear closer together, while the degree of separation in the cluster tree increases with further dissimilarity. In order to calculate the distances between all variables in the analysis, they were further standardized by transforming the data to have a mean = 0 and a variance = 1. All statistical tests were two-sided with a significance level of p-value of 0.05
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
unless otherwise specified. Statistical calculations were performed using STATA 12.0 (College Station, TX, USA). Heatmap plot was created by GENE-E, available at the website http://www.broadinstitute.org/ cancer/software/GENE-E/.
Results Our analysis included a ten-year database analysis in a single referral center along with 40 published studies [13–52], selected on the basis of the criteria reported previously, from which subtype distributions were calculated. Of the total 41 studies included, 17 (41.5%) were database analysis, 6 (14.6%) were controlled clinical trials, 4 (9.8%) were cohort studies, and 14 (34.1%) were from molecular studies of surgical sample archives. Characteristics of included studies and their quality score are listed in Supplementary Table. The exclusion/inclusion procedure is described according to the QUOROM-guidelines (Quality of Reporting of Meta-analyses) (Fig. 1) [53]. For the United States, there were two SEER database studies with comparable period of patients' recruitment and case numbers, therefore we decided to present pooled estimates of both studies. For European countries, we included several pan-Europe clinical trials with the intention for comparison. Quality assessment revealed that Jadad scores of all included clinical trials met the criteria of high quality (score ≥ 3), whereas New Newcastle–Ottawa Quality
149
Assessment Scale for some observational studies did not meet the criteria of high quality (score ≥ 7). Globally, for each subtype, the median value of relative frequencies for serous subtype was 45.0%, with the Philippines (16.0%), Indonesia (22.7%), and Brazil (30.1%) as the three lowest countries and South Africa (68.0%), Greece (71.5%), and India (86.7%) as the three highest countries; for mucinous subtype, the median value was 11.4%, with the three lowest countries being Italy (3.0%), Australia (3.4%), and Japan (5.4%) and the three highest countries being Indonesia (29.1%), Singapore (30.3%), and South Korea (38.6%); for endometrioid subtype, the median value was 12.6%, with the three lowest countries being India (1.6%), Iran (4.9%), Greece (5.7%), and the three highest countries being Taiwan (24.8%), Egypt (25.0%), and Austria (25.5%); and for clear cell subtype, the median value was 5.3%, with the three lowest countries being Pakistan (1.0%), Iran (2.0%), and Germany (2.0%) and the three highest countries being Thailand (16.0%), Taiwan (16.8%), and Spain (18.8%). Due to significant inter-country variability of relative frequencies, we further calculated the coefficient of variation (CV) for each subtype. The CV for serous subtype was 29.8%, mucinous subtype 58.3%, endometrioid subtype 42.3%, and clear cell subtype 62.7%. These estimates indicated that serous subtype appeared to have more consistent global distribution, followed by endometrioid subtype, whereas mucinous and clear cell subtype varied significantly across countries (Figs. 2–5).
Fig. 1. Quorum flow diagram of the steps in the reviewing process.
150
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
Fig. 2. Geographical distribution of relative frequency of serous subtype. CV, coefficient of variation. Relative frequency is calculated by number of serous subtype / number of (four major subtypes + minor subtypes). USA SEER data were pooled estimates of references [17] and [18]. Filled circles denote statistically higher percentage than SEER, whereas empty circles denote statistically lower percentage.
Hierarchical cluster analysis revealed that distribution pattern of subtypes can be broadly divided into five clusters (Fig. 6). Cluster 1 is a large category, composed of the United States, Canada, and countries
of European Union. Taiwan is in cluster 2, with a close proximity to Spain. Cluster 3 is composed of a mixture of Asian, Africa, and European countries. Cluster 4 was composed of exclusively Asian countries, such
Fig. 3. Geographical distribution of relative frequency of mucinous subtype. CV, coefficient of variation. Relative frequency is calculated by number of mucinous subtype / number of (four major subtypes + minor subtypes). USA SEER data were pooled estimates of references [17] and [18]. Filled circles denote statistically higher percentage than SEER, whereas empty circles denote statistically lower percentage.
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
151
Fig. 4. Geographical distribution of relative frequency of endometrioid subtype. CV, coefficient of variation. Relative frequency is calculated by number of endometrioid subtype/number of (four major subtypes + minor subtypes). USA SEER data were pooled estimates of references [17] and [18]. Filled circles denote statistically higher percentage than SEER, whereas empty circles denote statistically lower percentage.
as Singapore, South Korea, Philippines, and Indonesia. Cluster 5 comprised only one country, India, due to its unique distribution pattern. The dendrogram shows that countries in which the major pharmaceutical
discoveries and development for ovarian cancer are located on the left side of the figure (e.g. United States, Switzerland, France, Germany, and United Kingdom in cluster 1), whereas the major countries for drug
Fig. 5. Geographical distribution of relative frequency of clear cell subtype. CV, coefficient of variation. Relative frequency is calculated by number of clear cell subtype number of (four major subtypes + minor subtypes). USA SEER data were pooled estimates of references [17] and [18]. Filled circles denote statistically higher percentage than SEER, whereas empty circles denote statistically lower percentage.
152
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
Fig. 6. Dendrogram of hierarchical cluster analysis of distribution pattern of subtypes among countries. The lengths of the horizontal lines in the dendrogram provide relative qualitative information about the linkage distance between various clusters.
importation (i.e., China, South Korea, Singapore, Japan, and India in clusters 3–5) are on the right side of the dendrogram. Discussion Contrary to the common belief that the distribution pattern of histological subtypes of epithelial ovarian cancer is consistent across countries, instead, our systematic review revealed that there existed significant variations among countries and regions in the world. Among subtypes, fewer variations were observed in serous and endometrioid subtype, while larger differences were seen in mucinous and clear cell carcinoma. Among the four major subtypes, the latter two subtypes (mucinous and clear) are thought to be relative resistant to traditional chemotherapy [54]. It has been recognized that ovarian cancer is a heterogeneous disease with differential chemosensitivity and distinct molecular alternations for each subtype [4,54,55]. The large variations in subtype distribution patterns across countries and regions raise the issue of bridging study, which is designed to bridge a newly developed medication from the original region's (foreign) data in the original population to a new region [56]. The International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) had issued E5 document entitled “Ethnic Factors in the Acceptability of Foreign Clinical Data” in 1998. The ICH-E5 guideline provides a general framework for evaluating the potential impact of ethnic factors on the acceptability of foreign clinical data [57]. Differential distribution pattern in a specific country or region is one of the important ethnic factors which should be considered first when a new agent is to be registered to a new region for the treatment of epithelial ovarian cancer. To date, most landmark clinical trials of chemotherapeutic or molecular targeted agents for epithelial ovarian cancer were conducted in North America and/or Europe. When the approved agent is to be registered to a new region, pharmacokinetic bridging studies are usually requested by regulatory agencies of the new region, mainly countries in Asia, Africa, and Latin America, to extrapolate the trial results. However, the hierarchical cluster analysis in the current study
demonstrated that the countries with relative large market capability in new region (e.g. Japan, China, South Korea, Singapore, and India in cluster 3–5 in Fig. 6) showed significant different distribution patterns compared to countries in the United States, Canada, Switzerland, France, Germany, and United Kingdom (clusters 1–2), the major countries of drug development and export. Accordingly, the regulatory agency in the new region may consider conducting more complete clinical trials instead of merely implementing regular pharmacodynamic or pharmacokinetic bridging studies. Alternatively, considering the uncertainty of drug efficacy and safety due to the heterogeneous subtype distributions across countries, it is beneficial for the pharmaceutical industry to conduct multi-regional trials, designed for simultaneous drug development, submission, and approval around the world for epithelial ovarian cancer [58]. For the development of subtype-specific molecular targeted agents, our work provides a guiding map by which pharmaceutical industry can identify the candidate countries where high percentages of a specific subtype are observed to recruit patients more efficiently. The frequencies of both endometrioid and clear cell subtypes are relatively high in Spain and Taiwan. These two subtypes are closely linked to a distinct entity of endometriosis-associated ovarian carcinoma (EAOC) that is etiologically distinct from other ovarian carcinomas in several aspects [59]. Studies have documented that endometriosis is associated with approximately three-fold risk of endometrioid and clear cell subtype. Patients with EAOC had a lower stage of cancer, a distribution of histological subtypes that differs from the general population (i.e. endometrioid and clear cell cancer being the most common), predominantly lower grade endometriosis lesions, and significantly better overall survival as compared with other ovarian carcinomas [60,61]. Whether the populations in these two countries bear higher genetic polymorphisms associated with endometriosis-related ovarian cancer warrants further investigations. There are several limitations in the present study. First, there is a lack of adequate information from Africa and Latin America. Given this glaring lacking of available data, it is recommended that a global consortium for ovarian cancer registration be established to collect more update and
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
integrated resources for pharmaceutical development and academic research. Second, data from some countries were only available from surgical specimen archives in which selection bias may occur, thus raising some doubts whether the observed subtype distributions reflected the true values in the population. Third, the period of patients' recruitment was not consistent for all included studies. Some countries (e.g. Netherlands, Norway, and South Africa) presented the data that predated the routine use of monoclonal antibodies to differentiate subtype assignments, including high-grade endometrioid carcinoma versus high-grade serous carcinoma, undifferentiated histological types versus specifically differentiated histological subtype, and mixed histological subtypes versus pure histological subtype [62]. Fourth, data from some countries (South Africa, Greece, and Spain) did not provide details for “others” subtypes, thus the relative frequency estimates of four major subtypes may be inflated. Fifth, not all included studies mentioned standardized procedure of central pathology review, leaving the quality of subtype diagnosis a major concern. Sixth, data from some countries (Japan, Norway, and Greece) did not cover all stages, potentially making serous subtype over-presented. Seventh, many included papers predate the proposal of two-tiered grading system of serous carcinoma [63], precluding the presentation of data of low-grade versus high-grade serous subtype in our work. In summary, our analysis provides an overview of global distribution pattern of subtypes in epithelial ovarian cancer. By cluster analysis, we revealed that there existed considerable differences in distribution pattern of subtypes between most Asian countries and those in North American and Europe. These geographical and ethnic differences are by no means negligible, and they warrant further investigation. These figures may serve as a reference in the design and implementation of bridging studies when a medication is submitted for registration in a new region or country. Furthermore, this work provides a guide map for pharmaceutical industry to select pertinent countries or regions to implement clinical trials for epithelial ovarian cancer in the future. Finally, the glaring absence of studies from Africa and Latin America highlights the urgent need to construct a global network consortium that garners and integrates databases from countries around the world in a comprehensive and contemporary fashion. In addition to being a heterogeneous disease, epithelial ovarian cancer further demonstrates heterogeneous distribution patterns of subtypes. When designing clinical trials for epithelial ovarian cancer, it is imperative to take into account these heterogeneous distribution patterns. In order to accelerate drug development, submission, and approval around the world, implementing multi-regional trial at the initial design stage of new drug clinical trials for epithelial ovarian cancer is recommended. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2014.02.016. Conflict of interest statement The authors declared no conflict of interest with regard to any drugs or materials relevant to this paper.
Funding This work was supported both by Taiwan National Science Council NSC 101-2314-B-075-071 and Taipei Veterans General Hospital Grant V102C-077. References [1] Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:69–90. [2] Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2013;63:11–30. [3] Tavassoli FA, Devilee P. IARC WHO Classification of Tumours, No 4, Pathology and genetics of tumors of the breast and female genital organs. IARC Press; 2003. [4] McCluggage WG. My approach to and thoughts on the typing of ovarian carcinomas. J Clin Pathol 2008;61:152–63.
153
[5] Shih Ie M, Kurman RJ. Ovarian tumorigenesis: a proposed model based on morphological and molecular genetic analysis. Am J Pathol 2004;164:1511–8. [6] Shih WJ. Clinical trials for drug registrations in Asian-Pacific countries: proposal for a new paradigm from a statistical perspective. Control Clin Trials 2001;22:357–66. [7] Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009;151:264–9. [8] Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Metaanalysis of observational studies in epidemiology: a proposal for reporting. Metaanalysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283:2008–12. [9] Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJM, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:1–12. [10] Wells G, Shea B, Peterson J, Welch V, Losos M, Tugwell P. The Newcastle–Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. 3rd symposium on systematic reviews: beyond the basics; 2000. p. 3–5. [11] Division UNS. Standard country and area codes classifications (M49). Available from URL: http://unstats.un.org/unsd/methods/m49/m49regin.htm. [accessed 20 May, 2013]. [12] Agresti A. Categorical data analysis. John Wiley & Sons; 2002. [13] Alshenawy HA. Immunohistochemical expression of epidermal growth factor receptor, E-cadherin, and matrix metalloproteinase-9 in ovarian epithelial cancer and relation to patient deaths. Ann Diagn Pathol 2010;14:387–95. [14] Guidozzi F, Daponte A. Estrogen replacement therapy for ovarian carcinoma survivors: a randomized controlled trial. Cancer 1999;86:1013–8. [15] Paes MF, Daltoé RD, Madeira KP, Rezende LCD, Sirtoli GM, Herlinger AL, et al. A retrospective analysis of clinicopathological and prognostic characteristics of ovarian tumors in the State of Espirito Santo, Brazil. J Ovarian Res 2011;4:14–9. [16] Zhang J, Ugnat AM, Clarke K, Mao Y. Ovarian cancer histology-specific incidence trends in Canada 1969–1993: age-period-cohort analyses. Br J Cancer 1999;81:152–8. [17] Chan JK, Teoh D, Hu JM, Shin JY, Osann K, Kapp DS. Do clear cell ovarian carcinomas have poorer prognosis compared to other epithelial cell types? A study of 1411 clear cell ovarian cancers. Gynecol Oncol 2008;109:370–6. [18] Quirk JT, Natarajan N. Ovarian cancer incidence in the United States, 1992–1999. Gynecol Oncol 2005;97:519–23. [19] Yang X, Hou M, Yang K, Wang H, Peng Z, Cao Z, et al. Prognosis in epithelial ovarian cancer: clinical analysis of 287 pelvic and para-aortic lymphadenectomy. Chin Ger J Clin Oncol 2007;6:492–6. [20] Chitrathara K, Sheikh ZA, Vijaykumar DK, Kuriakose S, Anupama R, Nandeesh M. Is hysterectomy needed in ovarian cancer? Indian J Cancer 2011;48:471–6. [21] Koh SCL, Huak CY, Lutan D, Marpuang J, Ketut S, Budiana NG, et al. Combined panel of serum human tissue kallikreins and CA-125 for the detection of epithelial ovarian cancer. J Gynecol Oncol 2012;23:175–81. [22] Arab M, Khayamzadeh M, Hashemi M, Hosseini M, Tabatabaeefar M, Anbiaee R, et al. Crude and age-specific incidence rate patterns for histopathologic subtypes of ovarian cancer in Iran. Arch Iran Med 2010;13:203–8. [23] Katsumata N, Yasuda M, Takahashi F, Isonishi S, Jobo T, Aoki D, et al. Dose-dense paclitaxel once a week in combination with carboplatin every 3 weeks for advanced ovarian cancer: a phase 3, open-label, randomised controlled trial. Lancet 2009;374:1331–8. [24] Chung HH, Hwang SY, Jung KW, Won YJ, Shin HR, Kim JW, et al. Ovarian cancer incidence and survival in Korea: 1993–2002. Int J Gynecol Cancer 2007;17:595–600. [25] Thanikasalam K, Ho CM, Adeed N, Shahidan MN, Azizah WK. Pattern of ovarian tumours among Malaysian women at General Hospital, Kuala Lumpur. Med J Malaysia 1992;47:139–46. [26] Bhurgri Y, Shaheen Y, Kayani N, Nazir K, Ahmed R, Usman A, et al. Incidence, trends and morphology of ovarian cancer in Karachi (1995–2002). Asian Pac J Cancer Prev 2011;12:1567–71. [27] Redaniel MTM, Laudico A, Mirasol-Lumague MR, Gondos A, Uy GL, Toral JA, et al. Ovarian cancer survival population differences: a “high resolution study” comparing Philippine residents, and Filipino-Americans and Caucasians living in the US. BMC Cancer 2009;9:1–8. [28] Zhao C, Annamalai L, Guo C, Kothandaraman N, Koh SCL, Zhang H, et al. Circulating haptoglobin is an independent prognostic factor in the sera of patients with epithelial ovarian cancer. Neoplasia 2007;9:1–7. [29] Tangjitgamol S, Manusirivithaya S, Khunnarong J, Jesadapatarakul S, Tanwanich S. Expressions of estrogen and progesterone receptors in epithelial ovarian cancer: a clinicopathologic study. Int J Gynecol Cancer 2009;19:620–7. [30] Piccart MJ, Bertelsen K, James K, Cassidy J, Mangioni C, Simonsen E, et al. Randomized intergroup trial of cisplatin-paclitaxel versus cisplatin-cyclophosphamide in women with advanced epithelial ovarian cancer: three-year results. J Natl Cancer Inst 2000;92:699–708. [31] ICON2: randomised trial of single-agent carboplatin against three-drug combination of CAP (cyclophosphamide, doxorubicin, and cisplatin) in women with ovarian cancer. ICON Collaborators. International Collaborative Ovarian Neoplasm Study. Lancet 1998;352:1571–6. [32] International Collaborative Ovarian Neoplasm G. Paclitaxel plus carboplatin versus standard chemotherapy with either single-agent carboplatin or cyclophosphamide, doxorubicin, and cisplatin in women with ovarian cancer: the ICON3 randomised trial. Lancet 2002;360:505–15. [33] Perren TJ, Swart AM, Pfisterer J, Ledermann JA, Pujade-Lauraine E, Kristensen G, et al. A phase 3 trial of bevacizumab in ovarian cancer. N Engl J Med 2011;365:2484–96. [34] Gram IT, Lukanova A, Brill I, Braaten T, Lund E, Lundin E, et al. Cigarette smoking and risk of histological subtypes of epithelial ovarian cancer in the EPIC cohort study. Int J Cancer 2012;130:2204–10.
154
P.-L. Sung et al. / Gynecologic Oncology 133 (2014) 147–154
[35] Stadlmann S, Pollheimer J, Moser PL, Raggi A, Amberger A, Margreiter R, et al. Cytokine-regulated expression of collagenase-2 (MMP-8) is involved in the progression of ovarian cancer. Eur J Cancer 2003;39:2499–505. [36] Sedlakova I, Vavrova J, Tosner J, Hanousek L. Lysophosphatidic acid in patients with ovarian cancer. Ceska Gynekol 2010;3:41–6. [37] Soegaard M, Jensen A, Høgdall E, Christensen L, Høgdall C, Blaakær J, et al. Different risk factor profiles for mucinous and nonmucinous ovarian cancer: results from the Danish MALOVA study. Cancer Epidemiol Biomarkers Prev 2007;16:1160–6. [38] Kumpulainen S, Sankila R, Leminen A, Kuoppala T, Komulainen M, Puistola U, et al. The effect of hospital operative volume, residual tumor and first-line chemotherapy on survival of ovarian cancer — a prospective nation-wide study in Finland. Gynecol Oncol 2009;115:199–203. [39] Hamidou Z, Causeret S, Dabakuyo TS, Gentil J, Arnould L, Roignot P, et al. Populationbased study of ovarian cancer in Cote d'Or: prognostic factors and trends in relative survival rates over the last 20 years. BMC Cancer 2010;10:1–8. [40] Tianhui C, Lina J, Adam G, Katharina E, Bernd H, Alexander K, et al. Survival of ovarian cancer patients in Germany in the early 21st century: a period analysis by age, histology, laterality, and stage. Eur J Cancer Prev 2013;22:59–67. [41] Bamias A, Psaltopoulou T, Sotiropoulou M, Haidopoulos D, Lianos E, Bournakis E, et al. Mucinous but not clear cell histology is associated with inferior survival in patients with advanced stage ovarian carcinoma treated with platinum-paclitaxel chemotherapy. Cancer 2010;116:1462–8. [42] Ireland NCR. Cancer trends — cancers of the ovary, 14; 2012. p. 1–4. [43] Pignata S, Scambia G, Ferrandina G, Savarese A, Sorio R, Breda E, et al. Carboplatin plus paclitaxel versus carboplatin plus pegylated liposomal doxorubicin as firstline treatment for patients with ovarian cancer: the MITO-2 randomized phase III trial. J Clin Oncol 2011;29:3628–35. [44] Schouten LJ, Goldbohm RA, van den Brandt PA. Height, weight, weight change, and ovarian cancer risk in the Netherlands cohort study on diet and cancer. Am J Epidemiol 2003;157:424–33. [45] Wang Y, Helland A, Holm R, Kristensen GB, Borresen-Dale AL. PIK3CA mutations in advanced ovarian carcinomas. Hum Mutat 2005;25:1–5. [46] Dobrzycka B, Terlikowski SJ, Kinalski M, Kowalczuk O, Niklinska W, Chyczewski L. Circulating free DNA and p53 antibodies in plasma of patients with ovarian epithelial cancers. Ann Oncol 2011;22:1133–40. [47] Granstrom C, Sundquist J, Hemminki K. Population attributable fractions for ovarian cancer in Swedish women by morphological type. Br J Cancer 2008;98:199–205. [48] Levi F, Randimbison L, Blanc-Moya R, La Vecchia C. Second neoplasms after invasive and borderline ovarian cancer. Eur J Cancer Prev 2009;18:216–9.
[49] de la Torre FJ, Garcia A, Gil-Moreno A, Planaguma J, Reventos J, Cajal SR, et al. Apoptosis in epithelial ovarian tumors: Prognostic significance of clinical and histopathologic factors and its association with the immunohistochemical expression of apoptotic regulatory proteins (p53, bcl-2 and bax). Eur J Obstet Gynecol Reprod Biol 2007;130:121–8. [50] Faleiro-Rodrigues C, Macedo-Pinto I, Pereira D, Lopes CS. Prognostic value of E-cadherin immunoexpression in patients with primary ovarian carcinomas. Ann Oncol 2004;15:1535–42. [51] Beral V, Million Women Study C, Bull D, Green J, Reeves G. Ovarian cancer and hormone replacement therapy in the Million Women Study. Lancet 2007;369:1703–10. [52] Olsen CM, Nagle CM, Whiteman DC, Purdie DM, Green AC, Webb PM, et al. Body size and risk of epithelial ovarian and related cancers: a population-based case–control study. Int J Cancer 2008;123:450–6. [53] Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of reporting of meta-analyses. Lancet 1999;354:1896–900. [54] McCluggage WG. Morphological subtypes of ovarian carcinoma: a review with emphasis on new developments and pathogenesis. Pathology 2011;43:420–32. [55] Banerjee S, Kaye SB. New strategies in the treatment of ovarian cancer: current clinical perspectives and future potential. Clin Cancer Res 2013;19:961–8. [56] Gupta SK. Implications of ICH-E5: assessment of drug's sensitivity to ethnic factors and necessity of a bridging study for global drug development. Perspect Clin Res 2011;2:121–3. [57] ICH E5(R1): ethnic factors in the acceptability of foreign clinical data; Feb 1998. [58] Tsong Y, Chang WJ, Dong X, Tsou HH. Assessment of regional treatment effect in a multiregional clinical trial. J Biopharm Stat 2012;22:1019–36. [59] Worley MJ, Welch WR, Berkowitz RS, Ng SW. Endometriosis-associated ovarian cancer: a review of pathogenesis. Int J Mol Sci 2013;14:5367–79. [60] Pearce CL, Templeman C, Rossing MA, Lee A, Near AM, Webb PM, et al. Association between endometriosis and risk of histological subtypes of ovarian cancer: a pooled analysis of case–control studies. Lancet Oncol 2012;13:385–94. [61] Gayther SA, Song H, Ramus SJ, Kjaer SK, Whittemore AS, Quaye L, et al. Tagging single nucleotide polymorphisms in cell cycle control genes and susceptibility to invasive epithelial ovarian cancer. Cancer Res 2007;67:3027–35. [62] McCluggage WG, Young RH. Immunohistochemistry as a diagnostic aid in the evaluation of ovarian tumors. Semin Diagn Pathol 2005;22:3–32. [63] Malpica A, Deavers MT, Lu K, Bodurka DC, Atkinson EN, Gershenson DM, et al. Grading ovarian serous carcinoma using a two-tier system. Am J Surg Pathol 2004;28:496–504.