Contemporary management of breast cancer in Nigeria: insights from an institutional database

Abstract

High-quality data are needed to guide interventions aimed at improving breast cancer outcomes in Sub Saharan Africa. We present data from an institutional breast cancer database to create a framework for cancer policy and development in Nigeria. An institutional database was queried for consecutive patients diagnosed with breast cancer between January 2010 - December 2018. Sociodemographic, diagnostic, histopathologic, treatment and outcome variables were analyzed.

Of 607 patients, there were 597 females with a mean age of 49.8±12.2 years. Most patients presented with a palpable mass (97%) and advanced disease (80.2% ≥stage III). Immunohistochemistry was performed on 21.6% (131/607) of specimens. Forty-percent were estrogen receptor positive, 32.8% were positive for HER 2 and 43.5% were triple negative. Surgery was performed on 49.9% (303/607) of patients, while 72% received chemotherapy and 7.9% had radiotherapy. At a median follow-up period of 20.5 months, the overall survival was 43.6% (95% CI-37.7–49.5). Among patients with resectable disease, 18.8% (57/303) experienced a recurrence. Survival was significantly better for early stage disease (I and II) compared to late stage disease (III or IV) (78.6% vs 33.3%, p<0.001). Receipt of adjuvant radiotherapy after systemic chemotherapy was associated with improved survival in patients with locally advanced disease (68.5%, CI-46.3–86 vs. 51%, CI 38.6–61.9 p<0.001).

This large cohort highlights the dual burden of advanced disease and inadequate access to comprehensive breast cancer care in Nigeria. There is a significant potential for improving outcomes by promoting early diagnosis and facilitating access to multi-modality treatment.

Introduction

Breast cancer is the most common cause of cancer-related death in women globally¹. Unfortunately, there is a significant disparity in breast-cancer specific outcomes between high-income countries (HICs) and many low-middle income countries (LMICs) due to a variety of factors. In Nigeria, the incidence of breast cancer, which is currently 54.3/100,000, has risen dramatically over the last 10–20 years^2,3. This increase is projected to continue in the coming years^4–6. Phenotypically, evidence suggests a relatively young median age of diagnosis and higher proportion of triple-negative breast cancer (TNBC)⁹. These biological variables are confounded by delays in presentation that result in a significant burden of advanced stage disease^7,8,10. Understanding these factors and how they relate to poor outcomes is central to the design and implementation of effective interventional programs.

High-quality cancer registry data, which is lacking in many LMICs, is critical to understanding poor outcomes. The African Research Group for Oncology (ARGO) is a National Cancer Institute (NCI) recognized international cancer consortium based in Nigeria. It was established in 2013 in collaboration with Memorial Sloan-Kettering Cancer Center (MSKCC) in New York. The consortium developed a breast cancer database and biobank at a tertiary care academic hospital in Ile-Ife, Nigeria with the aim of providing high-quality data capable of informing meaningful interventions. This study presents the first full analysis of the ARGO, breast cancer database, including patterns of presentation, histopathology, treatment and outcomes. Based on this analysis, we present clinical and system-level domains for future intervention.

Methods

Development of the database

A prospective, electronic breast cancer database was launched at Obafemi Awolowo University Teaching Hospitals Complex (OAUTHC) in Ile Ife, Nigeria in 2016. Ethical approval from the Hospital Institutional Review Board was obtained and subsequently updated. Prior to this, patients with breast cancer were captured from the clinical record (2010–2016). These records were uploaded into the database and updated. Additional data points that were introduced after the launch of the database as well as missing records were obtained from retrospective review of the physical clinical record. Through telephone interview, patients or their next of kin were contacted to provide a follow-up update. All patients who were alive and reachable were offered regular follow-up every 3 months, which was incorporated into the prospective database. From 2016, patients were prospectively enrolled into the database with their details updated during every clinic visit as well as via telephone follow-up. In addition to clinical data, patients who presented from 2016 onward also had a part of the tumour specimen and serum stored in a −80° Celsius biobank for future analysis. Patients were consented for both clinical and biological data collection. The database is maintained by trained research assistants at OAUTHC and quality assurance is provided by support staff at MSKCC.

Sociodemographics and disease staging

The database includes information on sociodemographic variables, such as age, level of education and socioeconomic status. Socioeconomic status was defined by categorizing patients into 5 classes using occupation and level of education as the defining variables as described by Oyedeji in a study conducted in a Nigerian cohort¹¹. Those in classes 1 and 2 belonged to the upper class, class 3-middle class while classes 4 and 5 belonged to the low SE class. Risk factors for breast cancer including: date of menarche, supplemental hormone exposure and family history of malignancy; are also elicited. Clinical details such as tumour size and physical exam findings are recorded. Stage of disease at presentation is reported according to the American Joint Committee on Cancer. Metastatic work-up included a chest x-ray and abdominal ultrasound as per institutional protocol and international resource-stratified guidelines¹².

Histopathology

Histopathological data included pathological diagnosis, tumour grade (Nottingham grading system) and immunohistochemistry (IHC). Estrogen and progesterone receptor status reflect the Allred scoring system for receptor expression. Human epidermal growth factor receptor two (HER 2) positivity was defined by the presence of uniform intense membrane staining of more than 30% of invasive tumour cells (3+). Facilities for fluorescent in-situ hybridization were not available for patients with equivocal HER2 findings, they were therefore considered negative.

Treatment

The timing (neoadjuvant vs. adjuvant), regimen and number of cycles of chemotherapy was captured for each patient. Patients were classified as compliant if they received the recommended number of courses at appropriate intervals. Patients who had surgery had the type and extent of surgery documented.

Follow-up

Status at the time of last follow up was noted as ‘alive and well’, ‘alive with disease’, ‘deceased’ or ‘unknown’. Last follow up date was January 2020. Patients with unknown status were either classified as defaulters or lost to follow up. Patients who defaulted did not receive or complete their treatment after initial presentation. Lost to follow-up was used to classify patients who had received the recommended treatment but no longer attended clinic or were not reachable after consecutive 3 monthly calls for up to one year or more. Those who later became reachable during subsequent follow up calls had their designation changed depending on their status at that time (Dead, alive and well, alive with disease). Time intervals from identification of symptoms to presentation, treatment and the occurrence of events (recurrence, death) were noted. Disease-specific mortality was not captured; thus, death reflects ‘all cause’ mortality.

Data analysis

Descriptive statistics were used to capture the sociodemographic, histopathological and treatment characteristics of the cohort. Stage at diagnosis was compared between earlier years (2010–2014) and later years (2015–2018) of data capture. This was aimed at assessing the impact of recent interventions. Kaplan Meier methods were used to estimate survival patterns and cohorts were compared using the log-rank test. Overall survival (OS) was calculated from date of presentation until death; patients who defaulted or were lost to follow up were excluded from the survival analysis. Survival patterns were estimated based on stage at presentation. Patients with stage III disease which constitute a homogeneous cohort had their survival estimated based on the treatment received. Predictors of death were identified by multivariate analysis with only variables that were significant on univariate analysis included in the multivariate model. A p value ≤ 0.05 was considered significant for all statistical tests. All analyses were performed using SPSS version 25.

Results

Sociodemographic characteristics

There were 607 patients captured in the database between January 2010 and December 2018. There were 10 males (1.6%) and 597 females (98.4%) with a mean age of 49.8±12.2 years. The majority (422, 69.5%) were low income earners (Table 1). The mean age of menarche for the women was 15.2±2 years with a mean age at first birth of 25±5years and a mean parity of 3.8±1.9 children. The mean duration of breast feeding was 16.4±7 months. Amongst those whose menopausal status was known, 38.7% (186/481) were post-menopausal with a mean menopausal age of 49.2±5.8 years. Only 1 (0.2%) patient reported cigarette use while 15 (2.5%) patients reported a positive family history of breast cancer. The mean body mass index was 26.2±5.7 kg/m².

Table 1-.

Sociodemographic characteristics and risk factors for breast cancer

		N(607)	%
Age (years)	20 – 29	15	2.5
	30 – 39	111	18.3
	40 – 49	184	30.4
	50 – 59	152	25.1
	60 – 69	98	16.2
	>70	46	7.6
Socioeconomic status	Low	422	69.5
	Middle	177	29.2
	Upper	8	1.4
Level of education	None	53	8.7
	Primary	198	32.6
	Secondary	141	23.2
	Tertiary	210	34.6
	Unknown	5	0.8
Age at Menarche (years)	Mean±S.D	15.24±2.01
Menopausal Status (years)	Post-menopausal	253	41.7
	Pre-menopausal	228	37.6
	Unknown	116	19
	NA	10	1.6
Age at Menopause (years)	Mean±S.D	49.22±5.81
Use of OCP and HRT		92	15.4
Age at first birth (years)	Mean±S.D	25.02±4.95
Parity	Mean±S.D	3.83±1.89
Duration of Breastfeeding (months)	Mean±S.D	16.35±6.92
Smoking	Yes	1	0.2
	No	605	99.7
	Unknown	1	0.2
Family History of breast cancer	Yes	15	2.5
	No	588	96.9
	Unknown	4	0.7
BMI (Kg/M2)	Mean±S.D	26.2±5.7

Pattern of presentation, clinical, radiological and pathological details

The median duration of symptoms prior to presentation was 6 months (1–36 months). Mean tumour size at presentation was 10.5±7.8cm, with 135 (22.2%) patients presenting with ulcerated masses. Only 1.2% and 15.2% of woman presented with stage I and II disease, respectively. Sixty-percent (369/607) of women presented with stage III disease, while 20.3% (123/607) presented with stage IV disease. Stage at presentation was significantly associated with level of education with patients who had no education having the largest proportion of advanced cases (90.6%) compared to those with tertiary education (76.5%, p<0.001). Computerized axial tomography was used in the staging of five patients (0.8%). Immunohistochemistry was performed in 131 cases (21.6%). Amongst these, 53 (40.5%) were ER positive, 43(32.8%) were HER-2 positive and 57 (43.5%) were triple negative. In the recent cohort (2015–2018), a significantly higher number of women (p=0.03) presented with early-stage disease compared to those who presented earlier (2010–2014, Figure 1).

Figure 1:

Stage at presentation based on year of diagnosis

Treatment

Chemotherapy was administered to the majority of patients 426 (70.2%), but only 172 (40.4%) completed the recommended courses of chemotherapy at appropriate intervals. Of the 43 eligible patients based on HER-2 status, 6 (14%) received trastuzumab while 39 (73.6%) of the 53 eligible patients had hormonal therapy. Surgery was performed on 303 patients (50%), the majority of whom had modified radical mastectomy (279, 92%). Twenty patients had simple mastectomy (3.3%), while 1 patient had lumpectomy (0.2%). Forty-eight patients (7.9%) had radiotherapy administered (Table 2).

Table 2-.

Disease characteristics, treatment and outcomes

Variable		N(607)	%
Duration of symptoms (months)	Mean±S.D	10.10±13.94
Tumour Size (cm)	Mean±S.D	10.46±7.76
Breast Ulceration	Yes	135	22.2
	No	469	77.3
	Unknown	3	0.5
Laterality	Unilateral	576	94.9
	Bilateral	31	5.1
Group Staging	1	7	1.2
	2	92	15.2
	3	369	60.8
	4	123	20.3
	Missing	23	3.8
Immunohistochemistry ER Status	Positive	N(131) 53	% 40.5
	Negative	78	59.5
PR Status	Positive	42	32.1
	Negative	89	67.9
HER 2 Status	Positive	43	32.8
	Negative	88	67.1
Treatment received	Chemotherapy	426	70.2
	Surgery	303	49.9
	Radiotherapy	48	7.9
	Traztuzumab	5/43	11.6
	Hormonal therapy	39/53	73.6
Recurrence		N(303)	%
	Yes	57	18.8
	No	236	77.9
	Unknown	10	3.3
Site of Recurrence		N(57)	%
	Contralateral breast	6	10.5
	Loco-regional	40	70.6
	Distant	11	17.5
Survival		N(607)	%
	Dead	214	35.3
	Alive	266	43.8
	Unknown	127	20.9

Outcomes

Follow up data showed that 266 patients (43.8%) were alive at the time of last follow up, 214 (35.3%) were dead and 127 (20.9%) either defaulted before completing treatment or were lost to follow up. Patients lost to follow up were seen in the earlier part of the study between 2010–2016. Loss to follow up was however not significantly associated with any disease or demographic variable. Among those who had surgery, 18.8% (57/303) developed recurrent disease, of which 70.6% were loco-regional (Table 2). The median time to recurrence was 10 months.

Survival analysis was based on 472 patients with 135 patients were excluded). Of these, 21 were lost to follow up, 106 were defaulters and 8 had incomplete records. At a median follow-up of 20.5 months, the 5-year overall survival was 43.6% (95% CI-37.7–49.5). Inclusion of patients who were LFU in the analysis did not show a significant change in the 5-year survival 44.4% (CI-38.4–50.2).

Stage at presentation was significantly associated with survival (p<0.001, Figure 2). Patients presenting with early stage disease (stage I and II) had a cumulative 5-year survival of 78.6% (CI-64.6–87.5) while those with advanced disease (stages III and IV) had a 5-year survival of 33.3% (CI-23.7–42.4). Among patients with stage III disease, those who had surgery with chemotherapy and radiotherapy had significantly better 5-year survival (68.5%, CI-46.3–86) compared to those who had surgery with chemotherapy (51%, CI-38.6–61.9), surgery alone (45.9%, CI-11–75.7) or no surgery (23%, CI-11.8–37.1, p<0.001, Figure 3). On multivariate analysis, advanced stage disease was an independent predictor of mortality (Table 3).

Figure 2:

Survival based on stage at presentation

Figure 3:

Survival pattern for locally advanced breast cancer based on treatment received

Table 3:

Predictors of Recurrence and Death

	Univariate analysis	Multivariate analysis Odds Ratio (95% Confidence interval)	p-value
Predictors of death
Age	0.004
60+		1
40 – 59		0.68 (0.41–1.13)	0.13
< 40		1.27 (0.68–2.37)	0.46
Laterality	0.001
Unilateral		1
Bilateral		4.42 (1.34–14.57)	0.01
Unknown		3.22 (0.60–17.39)	0.17
Ulceration	0.001
No		1
Yes		2.00 (1.20–3.32)	0.01
Stage	0.001
2		1
3		2.54 (1.35–4.79)	0.01
4		11.33 (5.01–25.66)	0.01
Adjuvant Chemotherapy	0.001
No		1
Yes		1.41 (0.90–2.19)	0.13
Predictors of recurrence
Ulceration	0.01
No		1
Yes		2.96 (1.44–6.09)	0.01
Laterality	0.02
Unilateral		1
Bilateral		6.04 (1.29–28.24)	0.02
Stage	0.01
2		1
3		1.19 (0.48–2.95)	0.71
4		2.42 (0.73–8.09)	0.15
Neoadjuvant Chemotherapy	0.047
Yes		1
No		0.25 (0.52–1.20)	0.08

Discussion

Breast cancer is a becoming a major public health problem in many LMICs with rising incidence and persistently poor survival. Effective interventions to address this trend require high-quality data that is often limited by lack of resources and poor support systems¹³. The ARGO database is the product of an international collaboration, which is a good model for combating cancer disparities^{14, 15}. This began with a single institution, but currently involves multiple hospitals and with future expansion across the ARGO consortium, has the potential to generate nationally representative data.

This review comprehensively describes the pattern of presentation, clinicopathological profile and treatment outcomes for breast cancer care at an academic tertiary care referral hospital in Southwest Nigeria. Our study reports a 5-year overall survival of 43.6% for all stages at a median follow up period of 20 months. This is significantly higher than the 24% 5-year survival previously reported from a retrospective review (2005–2008) of 224 patients managed at a similar facility in Lagos, Nigeria¹⁶. Comparison with studies from other parts of sub-Saharan Africa (SSA) is quite challenging due to differences in methodology, survival end points and period of study. However, a study of over 2,500 patients managed between 2008 and 2015 from 14 population-based cancer registries in 12 countries from SSA (excluding Nigeria) provides one of the most comprehensive reviews of the African situation. It showed a wide variation in breast cancer outcomes based on the Human Development Index (HDI). Mauritius with a high HDI had a 5-year survival of 83%, close to figures from Europe and the USA while Uganda with a low HDI reported a five-year survival as low as 12%¹⁷. A review of 602 patients managed between 2009 and 2011 in Soweto South Africa also reported a 5-year terminal illness free survival of 61% while Ghana which shares similar demographic characteristics and resource capabilities with Nigeria reported a 5-year survival of 39%^18,19.

Stage at presentation is an important determinant of outcome. This is clearly depicted in this study with the significantly higher survival associated with early disease (78.6% for stage I/II vs. 33.3% for stage ≥ III). This highlights the possibility of achieving better outcomes with currently available resources, by promoting early diagnosis and treatment. It is quite encouraging to observe a small increase in the proportion of early breast cancer cases in the latter half of the cohort. However, sustained efforts are required to achieve more significant gains in survival. Current efforts by national and international agencies need to be scaled up, targeting the various levels of delay, both at the patient-mediated and health system levels. Awareness creation and provision of relevant breast cancer education are key interventions to address patient delay. To date, awareness of breast cancer screening in Nigeria remains poor and screening programs are not coordinated. In a recent survey of over a thousand women in a Nigerian community, only about 20% of the women had ever had a clinical breast examination, the majority not in the last one year²⁰. Only 2.3% of women reported ever having had a mammogram²¹.

As population-based screening is impractical in most resource limited settings, interventions should be aimed toward timely presentation and earlier diagnosis of clinically apparent disease²². Screening directed at high-risk populations is one potential strategy. However, findings from this study suggest that some of the well validated risk factors described among Caucasian populations may not be as applicable to patients in Nigeria. Our cohort was enriched for women with a relatively late age of menarche, early age at first birth, multiparity, long duration of breast feeding and infrequent use of hormonal contraceptives or hormone replacement therapies. The development and validation of locally relevant breast cancer risk models will help to identify a high-risk population that may allow for more targeted screening. A recent study evaluating a Nigerian-based breast cancer risk model seems to hold some promise and further validation may confirm its usefulness²³. The ARGO consortium is currently piloting a screening program among high-risk Nigerian women based on a positive family history with the use of a novel, hand-held, low technology device. If performance metrics are favorable, this may be scaled up for adoption at a regional or national level. More resources and research efforts should therefore be directed towards awareness creation, community mobilization and patient navigation.

A critical component of early diagnosis and treatment is access. Health care spending in Nigeria is largely out-of-pocket, with only about 5% of the population covered by the National Health Insurance Scheme²⁴. The majority of patients in this study were of lower socio-economic status, which may have contributed to delayed presentation and poor compliance with treatment. The burden of out-of-pocket expenditures for modern breast cancer diagnosis is also captured by the paucity of individuals with locally advanced disease who received cross-sectional imaging as part of their staging; despite the availability of the imaging modality at OAUTHC. This certainly limits the accuracy of staging and consequently affects the choice of treatment. This study also shows how few Nigerian patients are benefiting from the significant advancements in targeted therapy (e.g. HER-2 targeted therapy) which has significantly improved outcomes in HICs.

A significant barrier to effective multi-modality treatment of breast cancer in Nigeria is the poor access to radiotherapy. Despite the majority of patients presenting with locally advanced disease, <10% received adjuvant radiotherapy. Our study evaluated the impact of comprehensive care on outcomes by comparing survival among stage III patients based on the treatment received. The addition of adjuvant radiotherapy after surgery and chemotherapy was associated with significantly better survival and is congruent with a similar series from Lagos Nigeria²⁵. Although this study did not explore reasons for the omission of radiotherapy, lack of physical access to radiotherapy facilities and financial limitation are known barriers to radiotherapy in Nigeria²⁶. Given the significant cost of radiotherapy and other treatment modalities, access to care may be impossible without subsidies from governments and donor agencies. This should be considered an important area of intervention.

There are some limitations of this study. Unfortunately, about 20% of patients defaulted or were lost to follow-up. However, this is a significant improvement over historical series and represents one of the largest complete cohorts of breast cancer patients in SSA^{19, 25}. The survival analysis reported in this study was limited to overall survival. Disease specific survival could not be determined because the cause of death was unavailable for the majority of patients. We were also unable to report on outcomes based on receptor subtype (e.g. ER positive vs. TNBC) due to the limited number of patients with IHC data. However, such a comprehensive review will be feasible in the near future given the routine performance of IHC on all breast cancer specimens over the past two years within the ARGO network of hospitals. Staging of patients in this study was limited to x-rays and ultrasound in almost all cases; this has the potential of affecting the accuracy of staging, particularly patients with metastatic disease who might have been under-staged.

Conclusion

This study clearly depicts the dual burden of advanced breast cancer and inadequate access to comprehensive breast cancer care in a Nigerian cohort. It shows how better outcomes can be achieved given early presentation and adequate treatment despite obvious limitations. By highlighting the burden of advanced stage disease, poor access to modern diagnostics and targeted therapy, we identify multiple areas for further research and development. Perhaps most importantly, the data presented on long-term outcomes of breast cancer management in Nigeria provides a reference for measuring the impact of current and future interventions.

What’s new?

Despite being the leading cause of cancer deaths in Nigeria, there is limited data on long term outcomes of breast cancer. This study provides data from a prospective institutional database on the current management of breast cancer in Nigeria, highlighting the facts, the gaps and possible interventions. Survival estimates provided in this study are rare in most Nigerian breast cancer series. The impact of stage at presentation and treatment on survival were also assessed, highlighting how much is achievable with current resources given early presentation and access to care. This study provides a reference for current and future interventions in Nigeria.

List of abbreviations

ARGO

African Research Group for Oncology

ER

Estrogen Receptor

HDI

Human Development Index

HER

Human Epidermal growth factor Receptor

HIC

High Income Country

IHC

Immuno-Histo-Chemistry

LMIC

Low Middle Income Country

MSKCC

Memorial Sloan-Kettering Cancer Centre

NCI

National Cancer Institute

OAUTHC

Obafemi Awolowo University Teaching Hospitals Complex

OS

Overall survival

PR

Progesterone Receptor

TNBC

Triple Negative Breast Cancer

DATA AVAILABILITY STATEMENT

The data used in this study can be made available upon reasonable request.