Influence of Ethnicity on Survival of Breast Cancer Patients in Turkey

Breast cancer (BC) is a significant health problem among women worldwide. In western studies it has been reported that BC survival rates varies among different ethnic groups within the same national populations (Li et al., 2003; Spallek et al., 2012). Despite the overall improvements in BC survival for the last two decades, the difference in survival rates continues between ethnic groups (Wojcik et al., 1998). Several factors such as tumor biology and socioeconomic status (SES) are blamed for the survival disparity among ethnic groups. Breast cancer (BC) survival may differ in different regions depending on breast health awareness, and diagnostic and therapeutic backgrounds in Turkey (Ozmen, 2008). Western studies implicated that ethnicity is a predictor of survival in BC (McKenzie and Jeffreys, 2009). The impact of ethnicity on survival for BC has not yet been studied in 3 major ethnic groups (Turkish, Kurdish, and Arabic) living in Turkey. In the studies of Turkey and Western countries, the population of Turkey was all called as ‘‘Turkish’’, that is, the ethnicities of the women


Introduction
Breast cancer (BC) is a significant health problem among women worldwide. In western studies it has been reported that BC survival rates varies among different ethnic groups within the same national populations (Li et al., 2003;Spallek et al., 2012). Despite the overall improvements in BC survival for the last two decades, the difference in survival rates continues between ethnic groups (Wojcik et al., 1998). Several factors such as tumor biology and socioeconomic status (SES) are blamed for the survival disparity among ethnic groups.
Breast cancer (BC) survival may differ in different regions depending on breast health awareness, and diagnostic and therapeutic backgrounds in Turkey (Ozmen, 2008). Western studies implicated that ethnicity is a predictor of survival in BC (McKenzie and Jeffreys,

Influence of Ethnicity on Survival of Breast Cancer Patients in Turkey
Abdurahman Kuzhan 1 *, Mustafa Adli 2 , Hakan Buyukhatipoglu 3 included in these studies were not specified or taken into consideration.
Using Gaziantep University Oncology Hospital-based cohort of BC patients from Turkey, we investigated the impact of ethnicity on survival for BC, and the possible mechanisms in order to explain the survival disparities.

Materials and Methods
723 women BC patients whose first diagnosis times were ranged from 1986 to 2011 referred to the Oncology Hospital included in the study. The study was approved by Institutional review board of School of Medicine and written informed consent was obtained from all patients. Patients' ethnicity, age and stage at diagnosis, menopausal status, pathologic diagnosis, tumor size, histological grade, hormone receptor [(HR); estrogen receptor (ER), progesterone receptor (PR)] status, human epidermal growth factor receptor 2 (HER2) status, past treatments (surgery, chemotherapy, radiotherapy and hormone therapy), and survival times were recorded.
Staging was performed based on the American Joint Commission on Cancer 2010 staging system. Tumor grade was defined based on Bloom-Richardson criteria as I, II, III, and other/unknown (Harris et al., 2007). ER and PR status was recorded on the basis of immunohistochemistry (IHC) [positive (when >5% of tumor cells stained positive during IHC testing), negative or unknown]. HER2 status was recorded on the basis of IHC score (negative: 0 and 1+, positive: 3+). Specimens scored 2+ were further evaluated by fluorescence in situ hybridization (FISH) technique (copies≤2: negative, copies>2: positive).
Statistical analysis: Frequency of distribution was calculated for tumor characteristics, age and stage at diagnosis, menopausal status, and treatment history (surgery, chemotherapy, radiotherapy and hormone therapy) in ethnic groups. Chi-square test was used for comparison of the distributions of tumor characteristics and treatment types among ethnic groups. Kaplan-Meier analysis was used to estimate the overall survival and log-rank test was used to compare the survival curves. Analyses were performed by using Statistical Package for the Social Sciences (SPSS, version 18) software.

Results
Of the patients included 496 (68.7%) were Turkish, 189 (26.2%) were Kurdish, 37 (5.1%) were Arabic and 1 was Armenian. The only Armenian woman was not included in the statistical analysis. Median age for all groups was 47 (20-84). 650 patients had invasive ductal (90%), 32 had lobular (4.4%) and 40 had other carcinoma (5.5%). No significant difference was found in terms of histopathology among ethnic groups. However, invasive lobular carcinoma was not encountered in Arabic women. Median tumor size was 3.7 (0.5-7.2) cm for all groups. Tumor size in Kurdish women was larger than in Turkish patients (p=0.006). Kurdish BC patients were generally diagnosed at later stages (p=0.06); p value did not reach a significant level however there was a trend towards significance (Table 1). ER-positive tumor rates were 75%, 59.8% and 75.7% for Turkish, Kurdish and Arabic women, respectively (Table 1). ER and PR-positive tumor rates between Turkish and Arabic women were similar but they were lower in Kurdish patients (p<0.001). HER2positive tumor rates were similar among three ethnic groups (p>0.2). Triple-negative subtype (ER-/PR-/HER2-) rates were similar in Kurdish and Arabic women. Triplenegative subtype rates were lowest in Turkish patients (p=0.026) ( Table 1). Compared to Turkish and Arabic women with BC, Kurdish women were significantly less likely to receive hormonal therapy due to more HR-negative tumors (p=0.001). However there were no significant treatment differences (surgery, chemotherapy and radiation therapy) among ethnic groups (    Figure 1).

Discussion
Impact of the ethnicity on survival rates was reported in previous studies among BC patients (Greend and Pai 2008;Bhoo-Pathy et al., 2012;Spallek et al., 2012;Abdullah et al., 2013). This is the first study from Turkey investigating the influence of ethnicity on survival in women BC population. In the present study, we found Kurdish women with BC had worse survival rates than Turkish and Arabic women, although all had health insurance and equal access to care.
Tumor biology may be influenced by ethnicity, because of the higher frequency of poor prognostic factors such as high grade, HER2 over-expression, HR-negative and triple-negative status in certain ethnic groups (Amend et al., 2006;Telli et al., 2011). Amend et al. (2006) reported that Europan-American women with BC are more likely to be diagnosed with high-grade, ER-negative, PR-negative and triple-negative tumors compared to African-American women. A study from South East Asia showed that Malay and Indian women with BC were more likely to present with unfavorable tumor characteristics (ER-negative, PRnegative or less differentiate tumors) than Chinese women (Bhoo-Pathy et al., 2012).
Our previous study (Kuzhan et al., 2013) and the present study demonstrated that ER and PR status were different among Turkish, Arabic and Kurdish women with BC, however, tumor grade and HER2 status were similar. ER-negative tumor type was higher in Kurdish patients than in Turkish and Arabic women. In addition, Kurdish and Arabic patients had higher rates of ER-/PRnegative and triple-negative tumors than Turkish women. According to our study study results overall survival rates were better in Turkish and Arabic patients than Kurdish patients. Worst survival of Kurdish patients might be explained by more HR-negative and triple-negative tumor rates seen in this ethnicity.
Socioeconomic status (SES) and cultural values are reported to have a relationship with survival disparities seen among ethnic groups. Ethnicity has often been thought to represent SES (i.e. education status, household income, and insurance, etc.) in western literatures (Brawley, 2002). In western populations, BC patients with low SES were correlated with younger age, larger tumor, more advanced stage, higher tumor grade and higher proportion of HR-negative tumors (Li et al., 2002;Rosenberg et al., 2005;Andaya et al., 2012). Patients with low SES were also associated with fewer mammographic screening, unequal access to treatment and poorer treatment adherence (Bradley et al., 2002;Schootman et al., 2003;Hersman et al., 2005). Finally, in western studies, low SES is reported to be associated with more aggressive tumor characteristics and a worse clinical outcome. Unfortunately in Turkey there is no study data whether ethnicity and SES is correlated. However, Adli et al. (2010) reported that in contrast to western studies, there was no significant difference among ethnic groups in Turkey with respect to time to diagnosis, mammographic screening, and access to health care. In addition, our prevous study showed that tumor grade, age and stage at diagnosis were similar among ethnic groups (Kuzhan et al., 2012). However, the present study revealed that Kurdish women with BC were associated with larger tumor sizes compared to Turkish and Arabic women and more advanced stage than Turkish women.
Lifestyle factors, i.e. diet and age at first childbirth, are increasingly being recognized as important prognostic factors of BC (Kroman et al., 1998;McKenzie and Jeffreys, 2009). Due to differences in religious and cultural practices, lifestyle profiles may differ between the ethnic groups in the same population. For example, Chinese women have fewer children, breast feed for shorter period and have their first child later compared to the Malay and Indian. The incidence of BC was highest among Chinese but they normally present at early stages with lesser tumor and consequently, their survival is the higher compared to other ethnic groups in Malaysia (Yip et al., 2006;Yusuf et al., 2013). Muslim ethnic groups migrated from Turkey are at a younger age at their first childbirth and less likely to have alcohol consumption compared to local-born non-Muslims. These factors may explain the differences in incidence and mortality between immigrants from Turkey and local-borns (Zeeb and Razum, 2003;Spallek et al., 2012).
Treatment response to hormone therapy and chemotherapy may be affected by genetic differences. Differences in activity of CYP450 group of enzymes which metabolize antihormonal drugs and higher  (Lal et al., 2007;Fan et al., 2008;Ma et al., 2010). Hence, differences in the effectiveness of some anticancer drugs among certain ethnic groups may partly explain the survival disparities.
We had detailed information about tumor characteristics, prognostic and predictive determinants, and treatment records, however we did not have clear evidence about the causes of deaths whether they related to BC or not. In addition we did not have adequate data for social factors and response to treatment ratios.
In conclusion, Kurdish women with BC in this study are worse survival than Turkish and Arabic women, despite equal access to health care and similar treatment patterns. The underlying reasons for the survival disparity are unclear and additional epidemiological and genetic studies are required.