A Nationwide Study of Breast Cancer Histopathology in Cameroon (Central Africa)

Background and objectives: Breast cancer is the commonest cancer in Africa. Like other countries in that continent, Cameroon has very week pathology and health information services. This study was conducted to determine the histopathologic features of breast cancer all over Cameroon. Methods: The study was retrospective and cross-sectional. Data were collected in the 10 regional hospitals and in national referral hospitals. We included complete files of patients diagnosed with breast cancer from January 1, 2008 to December 31, 2015. Doubloons were avoided for patients seen both at regional and national hospitals. Files with conflicting pathology results were excluded. Results: Of the 1666 files included, 180 (12.6%) with conflicting pathology results were excluded and1486 cases were analyzed. Mean age was 47.9 years (range 17-91 years).Ethnic groups from 3 regions out of 10 accounted for 78.1% of patients. 82.3% of cases were diagnosed in the two regions main towns of the country. Diagnosis was late, with 97.5% of patients already presenting symtpoms. Histopathological analysis was performed for 1371 (92.3%) of cases. Pathology specimens were mainly biopsies (79.4%). All male breast cancers (15 cases) and 87.7% of female cases were ductal carcinoma. Grade 2 of Scarff-Bloom-Richardson’s classification was the most frequent (51.1%) followed by Grade 3 (303%). Discussion: Ethnicity seems to have an influence on breast cancer distribution in Cameroon. Pathological diagnosis is mainly done on biopsies and ductal carcinoma is the most frequent type of breast cancer in the country. Pathology services should be evenly distributed around the country.


Introduction
Cameroon like other sub-Saharan African (SSA) countries is not prepared to cope with the worsening burden of non-communicable diseases (NCDs) that accounts for 31% (3% due to cancers) of total deaths in the country [1,2]. The country does not meet 6 of the 9 World Health Organization (WHO) targets of health system response to NCDs; this includes the absence of a national population-based cancer registry [2].
In its 2016-2027 health sector strategy Cameroon engaged itself to ensure availability of quality health information at all levels for evidence based decision-making [3]. Breast Cancer (BC) is the commonest cancer in Cameroon [4]. Data on cancers from the Cameroon health system are still incomplete while those from global health partners are obtained after extrapolation [4,5]. Data published by the Ministry of Health revealed that cancers were responsible for 4.5% of deaths in 2016 [5]. In comparison, during the same year the following conditions contributed to the death toll as follows: tuberculosis (2.1%), liver cirrhosis (2.4%), road traffic accidents (4.4%), cardiovascular diseases (11.6%) and neurological diseases (0.9%) [5].
In sub-Saharan Africa, only 10% of histopathology needs are met and this is a major barrier to comprehensive management of cancers [6]. In Cameroon there are 19 pathologists currently in practice for 22,179,707 inhabitants and there is no pathology-oriented clinician [5]. Several authors have described histopathology of BC in Cameroon, but the scope of their data was not national [7][8][9][10][11][12]. Our objective was to describe histopathologic features of BC all over Cameroon.

Study design
We carried out a retrospective cross-sectional study from December 1, 2017 to march 31, 2017. The study took place inreferral hospitals of all the ten administrative regions and in all national referral hospitals of Cameroon. Figure 1 shows the administrative map of Cameroon with regional capitals. Clearances were obtained from the institutional ethical committee of the Faculty of Medicine and Biomedical Sciences of the University of Yaounde 1 (Ethical clearance number 0113/UYI/ FMSB/vn issued on the 4th January 2017) and from the managers of all those health facilities. We included complete files of patients managed for breast cancer from January 1, 2008 to December 31, 2015. Files with conflicting pathology results were excluded (180 files).

Data collection and management
Data were collected under three subheadings: sociodemographics (age, sex, education, occupation, marital status, regions of origin and recruitment), clinical features (diagnostic circumstances and clinical involvement of axillary lymph nodes) and histopathologic features (type of specimens, histologic type, Scarff-Bloom-Richardson's grade and axillary lymph nodes involvement). Files of patients referred from regional to national referral hospitals were recruited only once (at regional level).
All data were collected on paper forms then computerized using the version 6.3.2 of the software Census and Survey Processing System. Descriptive statistics were computed with the version 23 of the Software Statistical Package for Social Sciences (SPSS ® ).

Results
For the 8 year-period under scrutiny, 1666 files were included (208.3 cases/year). Those with conflicting pathology results were excluded (180) and the remaining 1486 were analyzed.

Sociodemographic characteristics
Women accounted for 1470 (98.9%) of cases. Most patients were married (70.4%). An overwhelming majority of cases were found in the centre and littoral regions (43.6% and 38.7% respectively). Mean age (standard deviation) at diagnosis was 47.9 (13.3) years, with extremes of 17 and 91 years. Distribution of patients according to occupation was: government employee (11.6%), private corporations' employee (11.6%), the professions (7.4%) and jobless (69.4%). The distribution of cases by age, education and region of origin is presented in Table 1.

Clinical features
Diagnosis was made under the following circumstances: consultation for symptoms (all men and 97.5% of women), ad-hoc screening campaign (0.8% of women) and breast self-examination (1.7% of women). Axillary lymph nodes were clinically involved in 3 (18.8%) men and 872 (59.3%) women. pathological diagnoses, 1089 (79.4%) were made on biopsies and 282 (20.6%) on resection specimens. Male BCs were all (15) ductal invasive carcinomas and had the following Scarff-Bloom-Richardson distribution: 2 of grade 1; 8 of grade 2 and 1 of grade 3 (1 result with regard to grade not available). Table 2 shows the distribution of female BCs by histologic types and Scarff-Bloom-Richardson's grades.

Sociodemographic and clinical features
The male/female breast cancer ratio in our series (1.1%) is close to that of 1% observed elsewhere than in Africa thougha meta-analysis found a much higher figure (4.2%) in Africa [13].
Over the 8-year period (2008-2015) under scrutiny we identified 1666 cases of BC; this contrasts with the 5-year prevalence in 2012 estimated at 8082 cases all over the Cameroon by the WHO after extrapolation [4]. Though our study systematically reviewed BCs in all the regions of Cameroon, it did not go beyond central and intermediate levels of the health system thus failing to include cases that remained peripheral (level of health district and inside the community). Another plausible explanation is that many cases (taken into account in WHO population extrapolations) never attended modern health facilities as it has been reported for other cancers in sub-Saharan Africa [1,[14][15][16][17].
Several factors explain why Yaounde and Douala accounted for more than 80% of cases. Firstly, the hospitals we included in those towns serve as referral facilities for the center and littoral regions which together account for 33.7% of the country population, both cities accounting for 24.7% [18]. Secondly, as the political and economic capital cities, those towns are characterized by higher proportions of educated and employed citizens more prone to attend modern hospitals. Thirdly, with 17 out of the 19 pathologists currently in practice in the country working in these two towns, access to pathology services is easier than in other regions.
It is quite striking that 78.1% of patients belonged to ethnic groups from three regions (Center, West and Littoral). It is known that incidence and mortality from BC depends on race and ethnicity but such atremendous disparity in Cameroon has not yet been reported [19,20]. Specific studies should be carried out to confirm and better characterized these finding in view of informing BC control strategies.
Mean age at diagnosis of BC in our series is within the range of previously reported figures (33-46 years) in Cameroon and neighboring countries [1,[7][8][9][10][11][12][21][22][23]. Like other researchers we found that diagnosis was very late (97.5% of cases were discovered following symptoms including clinical involvement of axillary lymph nodes). That result was foreseeable given the absence of screening (neither systematic nor opportunistic). Reasons for lateness in seeking modern cancer care in sub-Saharan are known and were out of the scope our study but high joblessness and illiteracy rates among patients (22.4% and 69.4% respectively) may have contributed [1,[15][16][17]. Late diagnosis and barriers to optimal holistic management of overt cases of BCs explain the ever depicted poor survival of patients in Cameroon [7][8][9][10][11][12].The tremendous scarcity of cancer-oriented clinicians and public health specialists that has been acknowledged by the ministry of health should to be tackled urgently in view of reversing the fast rising incidence of breast and other cancer in Cameroon [5].It is almost certain that such a human resource shall conceive and start a much more effective nationwide cancer control policy.

Histologic features
More than nine tenths of diagnosis was made by pathology services and archives were usable at the intermediate level of health system. Despite numerous challenges in implementing pathology services at national level in Cameroon, this can be considered as a promising step to be considered by policymakers when drawing national strategies for cancer control. Indeed6 of the 10 regions don't have a pathologist, meaning that there is already an effective mechanism for dispatching of specimens and results. In comparison, only 37.5% of BCs in Niger are confirmed by pathologists and there is only one pathology unit in the Central African Republic [22,23].
A major function of pathology in the management of BC is to establish the prognosis and guide surgical and adjuvant therapies (6). Unfortunately the histological prognosis has been based only on histological type and the Scarff-Bloom-Richardson's grade. There is a great need for an interdisciplinary strategy to foster a much more detailed histopathological analysis of cancer specimens. A wide range of specialists should intervene (surgical oncologists, medical oncologists, radiation oncologist and pathologists) in the frame of routine tumor boards at institutional levels. This will generate request and relevant use of more histoprognostic details. At the national level, the ministry of health and other policymakers must intervene by allocating the necessary funds to equip pathology units with the necessary technology and (further) train pathologists.

Conclusion
The novelty in this study is the nationwide coverage which is an (initial) answer to the information gap on characteristics of BCs in Cameroon. The retrospective collection of data has certainly limited the scope but prospective collection of such data all over the country is the role of a national cancer registry. Another aspect to explore is how pathology results are used in managing BC patients in Cameroon.