Systematic Review of Breast Cancer Biology in Developing Countries (Part 2): Asian Subcontinent and South East Asia

There has been no systematic appraisal of ethnicity-based variations in breast cancer (BC) biology amongst women from developing countries. A qualitative systematic review was conducted of breast cancer size, stage, grade, histological type, extra-mammary involvement, hormone receptor status as well as patient demographics. This review includes patients from Africa, the Middle East, Eastern Europe, Mexico, the Caribbean and South America. BC in these regions present at an earlier age with large aggressive tumours. Distant metastases are frequently present at the time of diagnosis. African women have a higher frequency of triple negative tumours. Over half of Middle Eastern women have lymph node involvement at the time of diagnosis. Despite experiencing a lower incidence compared to the Ashkenazi Jewish population, Palestinian women have poorer five-year survival outcomes. The majority of women from Mexico and South America have stage two or three disease whilst over sixty percent of women from Eastern Europe have either stage one or stage two disease. The biological characteristics of BC in the Caribbean cannot be fully assessed due to a paucity of data from the region. BC amongst the developing world is characterised by an early peak age of onset with aggressive biological characteristics. Strategies that improve breast cancer awareness, address amenable risk factors and improve early detection are essential.


Introduction
There is a paucity of data regarding Breast cancer (BC) in the developing world. We have previously reported on the BC characteristics from women residing in Africa, the Middle East, Eastern Europe, Mexico, the Caribbean and South America [1]. Here we present the results for the Asian Subcontinent and South East Asia.

Methods
The methods and search strategy employed was the same as described in the first paper of this series [1].

Results
A flow diagram (QUROUM statement) outlining the process by which papers were selected for further evaluation is presented in the first manuscript of this series [1].
An overview of the biological features of BC within Pakistan, India and Sri Lanka is presented in Table 1. Pakistan has one of the highest reported incidence rates of BC in Asia with an ASR of 54 per 100,000 reported in Karachi South [2]. Bhurgri et al. found that 60% of newly diagnosed cases of BC occurred in those less than 50 years of age [2]. Data on tumour biology can be found in Table 1.

Hormone receptor status
Her-2 positivity was found in 25% and 31% of patients from Karachi and Rawalpindi respectively [3,6]. Azizun-Nisa et al. found ER and PR expression to be significantly lower in Her-2 positive tumours compared with Her-2 negative tumours (ER 84% vs. 70%; PR 92% vs. 78%) [3]. Analysis of 315 breast tumours with IDC found over expression of p53 in 55% of cases with p53 positivity associated with regional lymph node involvement and distant metastases [11].

Background
For the period 2001-2004, BC was the commonest cancer amongst females in many of India"s urban cities such as Mumbai and Delhi [12]. However, in certain rural communities the measured incidence is less than that of cervical cancer [12]. 3 [13]. Data on combined receptor profiles found 25% were ER+/PR+, 7% were ER+/PR−, 21% were ER−/PR+ and 47% were ER−PR− [13]. ER and PR immunoreactivity increased with advancing age and correlated with the presence of elastosis [13]. Higher grades of IDC were associated with reduced ER/PR positivity while the presence of necrosis and lymphovascular invasion was inversely related to hormone receptor status [13]. Twenty percent of patients were found to be Her-2 positive [13].  [16]. Of those with malignancy, 96% of patients were greater than 30 years of age [16]. Ductal carcinoma was reported in 97% (338/348) of cancer FNA specimens with mucinous carcinoma and medullary carcinoma accounting for one percent (3/348) each [16]. Furthermore, 52 cases suspicious of malignancy were evaluated later by histology and 63% (27/43) were found to be malignant [16].
Sharma et al. reported histology data for 23 cases of which 91% (21/23) were IDC with single cases of medullary and lobular carcinoma reported [17]. Stage three cancer was reported in 55% (12/22) of cases followed by stage two at 27% (6/22) and stage one and four disease at nine percent each (2/22) [17].

South East Asia
The tumour characteristic of BC amongst women from South East Asia has been outlined in Table 2.  (1981)(1982)(1983)(1984)(1985) to 16% (1996-2000) while the frequency of stage three cancer decreased from 25% to 17% over the same period [18]. Over the 20-year period, the rates of axillary lymph node involvement decreased from 52% to 46% while the rates of local recurrence/distant metastatic disease declined from 28% to 14% [18].

Tumour biology
In a study of 1028 patients younger than 50 years of age, Lin et al. found 92% of cases were IDC with 22% reported as grade one, 55% as grade two and 23% as grade three (Table 2) [20]. Patients older than 50 years had fewer grade one cancers (18% vs. 22%) and a greater frequency of grade three tumours (30% vs. 23%) [20]. No significant difference in tumour histological type, cancer stage or nodal involvement was observed between the two age groups [20].

Background
For the period 2003 to 2005, BC was the leading cancer amongst Korean females (15%) [30]. Ahn et al. reported on the chronological changes of Korean BC between 1996 and 2004 and found the number of registered cases had risen from 3801 to 9667 (154% rise) [21]. The median age at presentation remained constant at 47 years [21].

Tumour biology
Between 1996 and 2004, Ahn et al. found the frequency of stage zero cancer increased from four percent to 10% while stage one cancer increased from 20% to 36% [21] The frequency of cancer stages two to four declined from 76% to 55% [21]. The incidence of ductal carcinoma in situ increased from four percent to 10% over the same period [21]. Choi et al. compared the biological features of early onset BC (before 45 years) between women from Korea and the United States of America (USA) and found the former had a larger mean tumour size (28 mm vs. 21 mm) with no significant differences in histological type and number of lymph nodes involved [31].

Hormone receptor status
Son et al. reported ER and PR positivity in 57% and 51% of cases respectively [32]. Choi et al. reported that Korean women had greater expression of Her-2 receptors at 48% (28/59) compared to Caucasian women from the USA at 16% (9/57) [31] Expression of ER, PR, p53 and cyclin D1 was not statistically significant between the two groups [31].
Rhee et al. reported that 20% (136/683) of patients had triple negative breast tumours (negative for ER/PR/Her-2) and this was correlated with a younger age (<35 years), shorter relapse free survival and more aggressive clinicopathological characteristics as evidenced by higher rates of p53 and Ki67 expression, negative bcl-2 expression and greater positivity for epidermal growth factor receptor [33].

Background
BC is the second most common cancer diagnosed in Thai women [34]. For the period 1998-2000, the estimated incidence was 21 per 100 000 while the peak age-specific incidence was at 45 years [34]. Lertsanguansinchai et al. reviewed 399 patients with BC and reported a mean age of 50 years [35].

Tumour biology
IDC was reported between 76% and 91% of cases across nine Thai registries, followed by lobular carcinoma [34]. In the city of Lampang, cancer stage was reported as local, regional and distant in 29%, 50% and 10% of cases respectively [34]. In Bangkok, the distributions of cancer stage was known in 54% of cases and of these, 20% were reported as having local involvement while 28% and six percent had regional and distant involvement respectively [34].

Kuala Lumpur
Tumour biology Malay women have lower incidence rates of BC compared to Chinese and Indian women, however they present with larger tumours with more advanced stage disease [44]. Ong et al. reported on 385 cases of BC without distant metastasis of which 61% of patients were Chinese [24]. The mean tumour diameter was 37 mm with grade two tumours reported in 53% (150/284) of cases, while 36% (103/284) were grade three and 11% (31/284) were grade one [24]. Lymphovascular involvement was noted in 33% (71/214) of cases [24]. When patients with stage four disease were also considered Malay patients were statistically more likely to present at a younger age, have larger tumours (>50 mm), higher disease stage (Stages three or four) with a greater frequency of lymph node involvement [24].
Hisham et al. reported on 774 new cases with the median age among the three ethnic groups of 50 years [43]. Furthermore, 50% to 60% of subjects had clinical stage three or four disease of which only 5% were detected through mammography screening [43]. A mean tumour size of 54 mm was reported (range 1-200 mm). The same authors had reported on a further 752 new cases and subsequently found a mean tumour size of 42 mm with 30-40% of cases with clinical stage three or four disease [43].  [46]. The age-standardized incidence was 6.9 per 100 000 [46].

Discussion
This review has shown that females with BC in the developing world have an early peak age at presentation with large and aggressive tumours. Distant metastases are frequently present at the time of diagnosis. Over half the women from the Asian subcontinent have lymph node metastases at first presentation with most cases having either grade two or three tumours. The rise in BC incidence amongst South East Asian countries exceeds that of the western world, with women from Malaysia and Philippines frequently presenting with stage three or four cancer. African women have a higher frequency of triple negative tumours. Over half of Middle Eastern women have lymph node involvement at the time of diagnosis. Despite experiencing a lower incidence compared to the Ashkenazi Jewish population, Palestinian women have poorer five-year survival outcomes. The majority of women from Mexico and South America have stage two or three disease whilst over sixty percent of women from Eastern Europe have either stage one or stage two diseases.
BC has long been considered a disease predominantly affecting affluent nations with the highest incidence rates reported in North America and Western Europe [48,49]. This is thought to be, at least partly, related to a greater prevalence of BC risk factors and the detection of early stage cancer through breast screening programmes [49,50]. Conversely, a lower incidence pattern is reported in developing nations such as those in Asia and Africa [48,49]. This is probably partially related to an under-reporting of cases due to the absence of cancer registries as well as a lower prevalence of risk factors for the development of BC [49,50]. Despite this the majority of BC deaths are reported in developing countries which have a higher mortality-to-incidence ratio [37,48]. For example the incidence to mortality ratio in South East Asia is 0.46 compared to 0.19 in North America [49].
In 2002, there were approximately 1.38 million new cases of BC worldwide and by 2020 this is expected to escalate to 1.7 million [48,51]. Furthermore, a projected 50% increase in BC mortality is anticipated worldwide, however this value is expected to be higher (58%) in developing nations [50]. Increasing life expectancy as well as changes in BC risk factor profile amongst developing nations are thought to contribute to the overall rise in BC incidence in the future. [50]. In urban Shanghai, the incidence of BC has increased by 50% [52]. In Mexico, BC is the leading cause of female cancer death whilst amongst the urban centres of India, it has surpassed cervical cancer as the leading female cancer [12,53]. The impact of "westernisation" whereby a previously unexposed population are exposed to changes in risk profile (cohort effect) may partially be responsible for the increased BC risk observed [54]. Evidence to support this originates from migration studies which show a higher risk of BC in women who had migrated from Japan to Northern America [55]. Furthermore, data from the Shanghai BC study have identified similar hormonal and reproductive risk factors for the development of BC to that observed in the developed world which included earlier age at menarche, nulliparity later age at first live pregnancy and menopause and a lack of breastfeeding to be prevalent amongst affected cases [56,57]. Similar patterns have also been observed in Malaysia and other parts of the developing world [58,59].
The advanced nature of BC in the developing world has largely been attributed to the delays in seeking medical attention [60,61]. The reasons for this are multi factorial and include a lack of breast screening services combined with socioeconomic, cultural and political factors that underpin a propensity for women to present with advanced cancer [61,62]. Thongsuksai et al. had reported that 25% of BC patients from Thailand had waited 12 weeks from the recognition of symptoms before seeking medical advice, with similar patterns of behaviour observed amongst women from Iran (43%), Columbia (20%) and Peru (67%) [63][64][65][66]. Fears amongst patients and their families were identified as barriers to preventing timely access to early detection methods in Mexico [67]. Amongst Iranian women, a lack of knowledge regarding the necessity of such visits, fear, negligence, lack of access to physicians, and poverty were cited as the main reasons for delay [64]. The lack of BC education and the presence of other competing causes of morbidity and mortality means the community knowledge of BC is relatively limited [68]. Cultural factors are also influential, with affected cases only presenting to medical services once homeopathic and alternate therapies have been exhausted [25]. For example, Ajekigbe et al. identified preference for prayer houses or spiritual healing homes in 14% of cases with delayed presentation [69]. Errico et al. have shown that some women describe a sense of guilt of bringing "bad genes" into the family and are made to feel isolated by their communities over fears of "spreading" their illness [70]. Women also carry fears over BC treatment. Ajekigbe et al. found that amongst Nigerian women, the most common reason for delayed presentation (45%) was fear of mastectomy with similar reasons identified by women from Pakistan [69,71].
The quality of treatment for BC in the developing world is highly variable. Agarawal et al. has described reoperation rates of up to 40% in certain sites in India following suboptimal surgery [68]. Furthermore, imperfections in pathological reporting of tumours have also been reported [72]. Inadequate tissue fixation has important consequences on the interpretation of hormone receptor status and hence treatment with hormonal type therapies [39]. In addition the ability to offer breast conserving surgery in developing countries is limited by the advanced nature of disease combined with a lack of radiotherapy services. A study from Delhi found only 11% of patients underwent breast conserving surgery due to the lack of radiotherapy services [73]. El Saghir et al. reported a total of 84 radiation therapy centres amongst all Arab countries combined compared to 1875 in the USA, despite the population numbers for the two areas being equivalent (approx 300 million) [74]. Finally, the financial costs also influence a women"s decision to proceed with treatment as the costs of clinical visits and treatments have to be self funded.
The low peak age of incidence in the developing world may be at least partially explained by a lower overall life expectancy observed in developing countries [75]. Gukas et al. reported that only 5% of the Nigerian population were greater than 60 years of age compared to 21% of the British population [76]. Furthermore, the age standardised rate of BC was not higher amongst the Nigerian population. In fact, for both populations the age standardised rate was greater for women above the age of 50 [76]. In addition, it is well known that BC amongst premenopausal women tends to display poorer tumour characteristics [77,78]. Thus the earlier mean peak age at presentation may partially account for the aggressive pattern of cancer observed in much of the developing world.
Although increased parity has typically been associated with reduced risk of BC, Palmer et al. described the "dual effect of pregnancy" and found African American women who had more than four children prior to the age of 45 years had a greater risk of BC and this was associated with a protective effect in women greater than 45 [79]. The role of increased parity (>3 live births) and increased BC risk amongst younger African women (but not white American women) was also noted by Hall et al., although the results were deemed not significant [80]. Okobia et al. found increased parity (>4) carried an overall protective effect in a case control study of 250 Nigerian women with BC where the mean age of cases and controls was 46 and 47 respectively [81]. In contrast, Adebamowo et al. found a significant difference in the mean number of pregnancies in affected cases compared to controls also amongst Nigerian women [59].
A number of studies from Africa and the Middle East report a greater frequency of triple negative tumours. This has previously been related to poor tissue preservation [82]. Uy et al. reported a 10% increase in the number of hormone receptor positive tumours at the Philippines General Hospital after specific tissue fixation procedures were implemented [39]. Adawambabo et al. found no difference between ER/PR receptor status and age of Nigerian BC patients, although Awadelkarim et al. found a mean difference of 12 years between Sudanese and Italian cases positive for ER [60,82]. The Californian BC study showed premenopausal African American women had a greater frequency of basal like tumours (triple negative tumours) compared to non-African American women of any age [83]. Thus the lower frequency of hormone positive receptors reported amongst women living in Africa or other developing countries may reflect a combination of poor tissue fixation, patient age and possible differences in biology.
The aim of this study was to summarise the biological features of BC form developing countries. Due to the amount of detail already presented in this series we did not compare these findings to first world countries. For the same reason we did not elaborate on the management and treatment of BC in developing countries as discussion on this topic can be found elsewhere [50]. There were however some limitations to this report. A number of studies were based on the findings from single institutions within each country"s largest city. Thus the BC characteristics of the country were inferred based on the results from a single study which reflect cases that, on average, have the highest standard of living. This is also a limitation of some National Cancer registries as the true BC incidence, mortality and biological characteristics from many of the smaller towns and villages remain largely unknown. As such we have incorporated the findings from World Health Organisation GLOBOCAN database which provides an estimate of BC incidence and mortality around the world including developing countries. One of the limitations that hindered valid comparisons between developing counties related to a lack of standardised reporting of data of BC characteristics. For example, some centres reported on tumour size and cancer stage based on clinical rather than pathological values while not all centres had reported on hormone receptor status. As a consequence some of the outcomes reported in this review were incomplete due to the variability in the data recorded. There are a limited number of studies which have assessed the contribution of genetic mutations to the development of BC within the context of the developing world. The lack of specialised laboratories, genetic reporting services and personnel amongst the developing world renders it difficult to elicit whether true genetic differences exist within ethnic groups. The role of socioeconomic factors is associated with poor prognosis and was not accounted for in this study. Finally, we were only able to review studies and abstracts written in English.
The global BC health initiative group have devised guidelines aimed at risk factor modification, early diagnosis, treatment and management of BC for developing countries [50]. With an increasing lifespan amongst individuals living in the developing world coupled with changes in risk factor profiles such as obesity and child bearing practises, BC incidence rates will continue to climb and a shifting trend from pre to post menopausal BC may occur [50,54]. Mammography remains the gold standard for early detection of BC [84]. However, the cost of establishing and maintaining such a programme may not be feasible in much of the developing world. Furthermore, mammography screening amongst pre menopausal women and the associated risks of high false positive tests would impose difficulties amongst developing countries. While alternative screening approaches such as self breast examinations have not been shown to reduce mortality in two randomised trials, the use of clinically trained professionals in performing clinical breast examinations in Egypt and India has shown encouraging results [85][86][87][88][89]. The importance of educating women of the symptoms, risk factors as well as addressing perceptions/misconceptions in a culturally sensitive manner remains essential [90].

Conclusions
BC in the developing world is characterised by an early peak age of onset with aggressive biological characteristics. There are a limited number of detailed studies in this field with significant variability of outcomes reported. Further research into the molecular and biological features as well as genetic determinants into BC development is certainly warranted. Importantly, public health initiatives that improve BC awareness address amenable risk factors and allow for the early detection of BC will be essential in addressing the outcome inequalities that currently exist.