Epidemiological, clinical and diagnostic profile of breast cancer patients treated at Potchefstroom regional hospital, South Africa, 2012-2018: an open-cohort study

Introduction Breast cancer is the second most diagnosed cancer worldwide. We aimed to depict the diagnostic approach as well as the epidemiological and clinical profile of patients with breast cancer at Potchefstroom regional hospital, South Africa. Methods This descriptive open-cohort study included patients with primary invasive breast cancer, confirmed by histology results and treated at the hospital from 01 January 2012 to 31 December 2018. Data such as demographics, patient history, histology, breast clinical findings, physical mass description and diagnostic investigations were captured from hospital registries and patient files. Result One-hundred thirty-eight patients (mean age 56.2 (SD: 14.4) (95% CI 54.6-59.7) years) met inclusion criteria. Most patients were female (98.6%), from African (67.4%) or Caucasian (23.9%) descent. Findings included mostly left-sided breast involvement (51.8%), lesions in the upper-outer quadrant (43.1%), extensions to the skin (25.6%, N = 39), and tumour size of 2 ≤ 5 cm (49.3%), or > 5 cm (39.1%). Most patients (57.9%, N = 135) were categorised as BIRADS-5, with a ductal pattern (89.6%) (p < 0.01). Patients mostly presented in stages II to IV of disease (89.1%; p < 0.05). Late-stage (stages III-IV) at time of diagnosis (n = 84) was significantly associated with mass location (p = 0.006; Cramér's V = 0.280), tumour size (p < 0.001, Cramér's V = 0.239), and skin changes (p = 0.027, Cramér's V = 0.492). Conclusion Most patients consulted at a late-stage of the disease, indicating a need for the promotion of breast awareness campaigns, early detection, and timeous referral.


Introduction
Breast cancer is the second most diagnosed cancer accounting for an estimated 2.1 million cases (11.6%) of all diagnosed cancers worldwide, and a leading cause of cancer-related morbidity and mortality with more than 626 000 deaths in being home to cosmopolitan city centres, comfortable neighbourhoods and suburbs, but also to impoverished townships" [4]. The country is divided into nine provinces, which vary considerably in size. Gauteng, a highly urbanised region is the smallest, whereas the North West Province is the 6th largest in terms of size. The population of the North West Province resemble that of a developing country with approximately 60% of the population living in rural areas [5]. In lowerto middle-income countries, breast cancer patients often present late in their diagnostic stage with advanced disease [6][7][8]. The clinical stage at which patients present when diagnosed with breast cancer is a vital prognostic survival factor that may challenge the management thereof [9]. Cancer is usually detected during a screening mammogram, where patients with Breast, Imaging, Reporting and Data System (BIRADS) four and five will have a biopsy of the breast lesion done. BIRADS is a lexicon that forms part of a standardised mammography report whereby findings are put in a small number of well-defined categories that aim to aid quality assurance, communication, research, and improved patient care [10]. The screening mammogram and the greater breast cancer awareness of patients allow earlier diagnosis (stage I or II) [10]. The goals of the South African National Department of Health's breast cancer control and management policy [11] include the improvement of survival, decreased time to presentation, patient diagnosis and treatment initiation. It furthermore aims to decrease the stage at which treatment is started, thereby improving the quality of life in both survivorship and palliation, and to effectively monitor and evaluate the programme implementation and the impact of breast cancer interventions. Within South Africa itself, some disparities have been noted between regions and population groups in an early study [12].
With the South African National Department of Health's policy goals at hand and the attentiveness of disparities regarding access to healthcare, the need for local/regional profiling of breast cancer patients derived. The aim of the study was to depict the diagnostic approach as well as the epidemiological and clinical profile of breast cancer in Potchefstroom regional hospital, South Africa. The study is, therefore, a snapshot of the current status in a rural area and should pave the way for improvement.
investigation findings and patient's human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) status. The tool was converted to a Microsoft Excel® data capturing sheet. A 5% data re-entry method was followed, whereby 5% of the data were entered into a different dataset, where after the datasets were compared electronically. All discrepancies flagged as errors were resolved manually by comparing the electronic dataset to the data collection tool, using the patient number indicated on the tool/sheet.
If any discrepancies were found, another 5% of data were re-entered and the process repeated until no discrepancies were found. Data were also checked for any outliers. Age group was divided into 10year intervals. Population group was categorised as either African (urban and non-urban), Indian, Coloured, Caucasian and other as demarcated in the South African Demographic and Health Survey [5].
Statistical analysis: data were analysed using SPSS ® version 25.
Tests for normality (Q-Q plots) were used to determine data distribution. Continuous variables were expressed as means, standard deviations and 95% confidence intervals (CIs), and tested for heterogeneity of means by analysis of group variances using the student t-test. Categorical variables were expressed as counts and percentages. To test if single categorical variables follow a hypothesised population distribution, we performed the one-sample chi-square test, whereas Pearson's chi-square test and Fisher's exact test were used for the assessment of the relationship between categorical variables. A two-tailed p-value, where p < 0.05, was considered significant. The practical significance of associations was computed when p-values were significant. To determine the effect size of associations, we used the Cramér's V statistic. A Cramér's V ≥ 0.1 was deemed as a weak association, Cramér's V ≥ 0.3 was seen as a moderate association and Cramér's V ≥ 0.5 regarded as a large effect/association. To assess the size of the difference between means we used Cohen's d-value, with d ≥ 0.8 regarded as practically significant. To describe the association between the demographic, epidemiological and clinical profile of breast cancer patients in Potchefstroom Hospital, we grouped patients as either early stage (ES) (stages I and II) or late-stage (LS) (stages III and IV). Table 1 present the demographic characteristics of the study population. Table   2 and Table 3 show the findings of the clinical examination and physical mass description, and diagnostic investigation respectively.
Percentages in each category were calculated according to valid (nonmissing) cases as the denominator.
Ethical considerations: the Health Research Ethics Committee of the North-West University approved the study (NWU-00007-19-S1).
Approval was also obtained from the North West Provincial Department of Health and the Potchefstroom Hospital Patient's Safety group.

Results
Patient demographics: the demographic characteristics of the study population are described in Table 1 (Table 2). Table 3 depicts the findings from the analysis of diagnostic investigations.

Discussion
This is the first study describing the diagnostic approach, the epidemiological and clinical profile of breast cancer in Potchefstroom regional hospital, located in a predominantly rural area of South Africa.
We showed that most patients in our study consulted at late-stage of the disease, similar to recent studies conducted in peri-urban areas (residential space on the outskirt of large cities) of South Africa [8,13], and other sub-Saharan African countries [14][15][16]. In contrast, in developed countries such as the United States of America (USA), more than 60% of breast cancer patients present at stages I and II, and only 28% at stage III and III [17]. According to Black and Richmond [14], advanced disease at the time of diagnosis in sub-Sahara African regions may be ascribed to financial, logistical and sociocultural constraints hampering the implementation and sustainability of screening programs. Rambau et al. [16] are further of the opinion that this may also be the result of poor referral systems and natural aggressive biological behaviour of tumours.
Our population were mainly from African or Caucasian descent. The distribution of patients could be explained by the local demographics of Potchefstroom municipal area's population group as described in 2011 national statistics, indicating 70.8% as African, 7.1% as Coloured, 1.4% as Asian and 20.6% as Caucasian [5]. Research has, however, shown that the profile of breast cancer varies among different population groups. For example, Vorobiof et al. [18] and Walker et al . [19], determined that less than 22% of black women in peri-urban areas in South Africa consulted at Stage I and II of breast cancer, with the majority presenting at stage IV. Jedy-Agba et al. [20] identified similar trends among black women from the southern African region. In our study, however, although 63.4% and 63.6% of patients from African and Coloured descent, respectively, presented in stages III and IV compared to 51.5% of Caucasian patients, there was no association between population group and staging. Because of the limited sample size and missing data, associations between clinical findings and patient demographic characteristics could not be explored extensively.
The majority of patients in our study population were adults (mean age 56.2 ± 14.4 years (95% CI, 54.6-59.7), which is in agreement with findings from other studies conducted in large urban and periurban areas in South Africa [13,21]. Our population was older than the mean age of women in West Africa (mean age for presentation between 35 and 45 years) [22][23][24], Tanzania (44.7 years) [25], and the Democratic Republic of the Congo (mean of 47.8 ± 12.1 years) [26], although younger than the median age of patients with breast cancer in the USA which is 62 years [27]. This can be ascribed to a higher life expectancy in developed countries than in developing countries [28]. For example, based on population statistics, the highest proportion of the population (both in North West and South Africa) fell within the age category of 0-24 years (with proportions ranging from 9.3% to 11.3%) [5]. Out of the 138 patients, 1.5% (n = 2) were male. This percentage is close to the 1% found by other authors in South Africa [13], and other countries [29][30][31]. Risk factors are similar in both male and female patients; which include, among others, a family history of cancer (implying a genetic predisposition), radiation exposure history and hormonal imbalances [32]. Because of the small sample size and missing data on family history, it was not possible to contextually investigate this further in our study. Motloutsi et al. [  . This study presented with only one 34-year old woman with metastatic melanoma, treated for melanoma on the gluteal region a few years prior to the breast mass presentation. Lymphoma of the breast, either as a primary or secondary location, is rare, representing 0.04% to 0.5% of malignant breast tumours [40]. It may also originate from lymphatic tissue present within the breast, adjacent to ducts and lobules, or from intra-mammary lymph nodes [41]. The single patient reported in our distribution was a 24year old HIV-positive woman with non-Hodgkin's lymphoma of the breast. This is an AIDS-defining cancer and considering the high prevalence of HIV in South Africa, healthcare professionals should always keep it as a differential diagnosis in young patients who are HIV-positive and presenting with breast tumours. Although less than 30% of our population presented with other clinical findings including skin changes, ulcerations, and nipple retractions, these are signs of advanced disease, as per the American Joint Committee on Cancer's (AJCC) Tumor, Node, Metastasis (TNM) breast staging classification and diagnostic features [42]. The latter findings will have an impact on the management and the survival of the patient. The three patients with Paget's disease of the breast, defined by Karakas as a [43] "rare type of cancer of the nipple-areola complex that is often associated with an underlying in situ or invasive carcinoma", all presented with nipple discharges and peri-areolar eczema-like lesions. These findings are the commonest findings which make even the diagnosis of Paget's disease very suggestive during clinical examination [44,45] These findings echoed Lee et al. [46] stating that BIRADS-3 is an evolving assessment category. As much as it reduces the number of unnecessary biopsies, it allows the person reading the report to keep a high index of suspicion for detections of early-stage breast cancer.
Michaels et al. [47] emphasised this by stating that there is significant inter-observer variability in the assessment of mammographic BIRADS-3 findings. Clinicians should, therefore, interpret the BIRADS finding in conjunction with other relevant clinical information.
The HIV-pandemic in South Africa prompted the study to include patients' HIV status. It is important to mention that about a third of the study population was HIV-positive. This percentage is higher than the 19.7% among Soweto women with breast cancer found by Cubasch et al. [48], and the 5.3% found in Cape Town, South Africa, by Langenhoven et al. [21]. This disparity is probably linked to the prevalence of HIV infection in the general population of these three provinces in South Africa, respectively 22.7% (North West), 17.6% (Gauteng) and 12.6% (Western Cape) [49]. The relationship between HIV/AIDS and breast cancer is still a field of extensive research worldwide. No study has established a direct correlation between the two conditions, but it appears as if these patients are younger with advanced stages of the disease. In the current era of highly active antiretroviral therapy, most of the HIV-positive patients, however, do not experience more complications due to breast cancer than those who are HIV-negative [50]. Finally, our study population comprised of patients from the South African public health sector. This limits the external validity of the study in that results cannot be generalised to the total South African population.

Conclusion
This study depicts the epidemiological and clinical profile as well as the diagnostic approach of breast cancer in Potchefstroom regional hospital, South Africa. Our study reiterated findings that, irrespective of the geographical status in South Africa, the majority of patients with breast cancer consulted late in both rural and peri-urban areas of developing countries, at stage III or stage IV of the disease,  Advanced stages of breast cancer at the time of diagnosis was independent of gender, age group, family cancer history, prior hormone therapy, HIV status, population group, side of involvement, tumour local invasion, and histology type.