The study was cross-sectional and analytical, with retrospective and prospective data collection at the oncology departments of the Yaoundé Gyneco-Obstetric and Pediatric Hospital (YGOPH), and the Yaoundé General Hospital (YGH). The study lasted 7 months. The period of study was 06 years, from January 2017 to December 2022.

The study population consisted of patients with breast cancer at the YGOPH and the YGH during the period of study. We consulted the files of these patients and included all cases (non-probabilistic sampling) meeting the inclusion criteria. We excluded patients whose files did not contain the diagnosis, and for whom latest news could not be obtained, because they did not allow survival calculations. We calculated the minimum sample size using the Lorentz formula, considering as the prevalence of the event studied (5-year survival) to be 62%, according to a previous study done in Cameroon in 2022 by Noa et al. [7] . This gave us a sample size of 362.

We used data from all YGOPH and YGH outpatient oncology registers from 2017 to 2022, to have a database of all breast cancer patients seen during this period. Secondly, we keenly searched the archives of the Gynecology-Obstetrics and Oncology departments for the medical files of these patients with breast cancer in order to collect data. We completed the missing information, and data relating to survival, which did not appear in the files, directly from the patients or their relatives through a telephone conversation. Data collected included sociodemographic/epidemiological, clinical, paraclinical, and therapeutic variables, and data relating to follow-up and survival.

We chose as the original date, the date of diagnosis of breast cancer and as the point date, the end date of the study, that is, May 6, 2023. The date of the latest news was the date on which we last heard from the patient. This date could be the date of death or the date of the last consultation, if the patient did not present the event studied (death), or if the subject was lost to follow-up, and the point date for those who were alive (dated May 6, 2023). The latest news was that the patient could be dead, alive, or lost to follow-up. The follow-up is the time that elapsed between the start date of the study and the end date. Participation time is the time between the original date and the latest news date. If the patient was lost to follow-up at the time point, she was right censored when determining the survival curve.

We obtained authorization from the YGH(N˚ 137-23/HGY/DG/DPM/APM-TR of February 13, 2023), and from the Institutional Research Ethics Committee for the Human Health (CIERSH) of YGOPH (N˚ 406/CIERSH/DM/2023 of March 3, 2023).We also had ethical clearance from the Institutional Ethics and Research Committee (CIER) of the Faculty of Medicine and Biomedical Sciences of the University of Yaoundé I (N˚ 0061/UYI/FMSB/VDRC/DAASR/CSD of February 6, 2023).

We used SPSS software version 23.0 (Statistical Package for Social Sciences) for analyses. Quantitative variables were expressed by the median followed by the range, and the mean followed by the standard deviation. Survival analysis was done with estimated survival curves using the Kaplan-Meier method. Survival curves were compared using the Log Rank test. Factors influencing survival were evaluated using the Cox model, through multivariate analysis. The statistical significance threshold was set at 0.05 and the confidence interval at 95%.

The ages of the 500 patients studied ranged from 22 to 83 years with a mean age of 47.19 ± 11.61 years. The most represented age group was 30 to 45 years old (45.8%) and the median was 46 years. More than half of the patients were single (51.4%) and had attained a higher education level (57.2%). The informal sector of the economy had a high representation (43.6%), while the majority resided in the urban setting (53.8%), and were Christians (94%). The highest presence, with respect to region of origin, was the West (37.4%), followed by the Center (27.8%) regions. We also recorded 29 participants from Chad and Gabon (5.8%).

The average consultation time (between the appearance of the first symptoms and the first consultation) was 10.26 ± 10.62 months with extremes of 1 month and 61 months. In our study population, only 3.8% were diagnosed during a screening campaign or routine visit. The remaining 96.2% were diagnosed following a consultation for symptoms. The most frequent reasons for consultation were the perception of a breast nodule or mass (79.9%) followed by breast pain (14.8%) and skin or nipple abnormalities (4.7%).

The majority of women had had more than 02 pregnancies (57.4%), and were multiparous (47.4%). Sixteen percent (16%) of women used a modern method of contraception with a predominance of combined oral contraceptives. The average age of first period was 14 ± 1.70 years with extremes of 9 years and 18 years. The most common age group for first pregnancy was 18 to 30 years (74.5%), the average being 22 years. Forty-one percent (41%) of women had reached menopause at the time of diagnosis. In our study population, 17.4% had comorbidities, the most common being hypertension (62.1%). Eight-point six percent of patients had a past history of breast disease, mainly fibrocystic mastopathy. The notion of a history of cancer in a family member was found in 39 patients (7.9%), mainly first degree, and 57 patients had familial breast cancers (11.5%). The nodule was located in the left breast in 58.6%, preferentially in the QSE (58%). The majority of patients were classified as stage 3 at the time of diagnosis (47.6%).

Infiltrating ductal carcinoma was the most represented histological type (84.7%), while grade 2 of the SBR classification, was the most represented histo-prognostic category (40.2%). Immunohistochemistry was performed by 34.20% of patients and the most represented molecular subtype was triple negative (41.8%) followed by Luminal A (30%). On imaging, the ACR 4 finding was the most common (59.6%) (Table 1).

Surgery was performed in 45% of cases and the majority of operated patients received radical treatment (88.4%). Chemotherapy was administered in 79.4% of cases; the most frequently used chemotherapy regimen was regimen D (1st line FAC/AC, 2nd and 3rd line not received) in 35.5% of cases. Hormone therapy was prescribed in 34.2% of patients while 14.6%received radiotherapy. Of all 500 patients retained, 17.2% received no treatment. In patients, whose assessment was known, partial response was observed in 42.1%, and complete response in 35.4% (Table 2).

Regimen A: 1^st line (FAC/AC), 2^nd line (Taxol/Taxotere), 3^rd line (Capecitabine);

Regimen B: 1^st line (FAC/AC), 2^nd line (Carboplatine + Taxol), 3^rd line: (Capecitabine);

Regimen C: 1^st line (FAC/AC), 2^nd line: (Taxol/Taxotere/Carboplatine + Taxol), 3^rd line: not received;

Regimen D: 1^st line (FAC/AC), 2^nd line: not received/ not known, 3^rd line: not received;

Unknown: treatment not initiated or undocumented.

The survival of patients with breast cancer at 1, 2, 3, 4, and 5 years was, respectively, 90.6%, 83.1%, 74.2%, 69.8%, and 59.2%. The slope of the survival curve reveals that mortality was higher in the first 40 months (the first three and a half years), before stabilizing and remaining constant after the fifth year. The median survival time was not reached. However, the first quartile (Q1) was 36.00 months (3 years). However, the average survival time of our study sample was 56.91 months (4.74 years). The minimum survival time was 0.73 months (22 days), while the maximum survival time was 76 months or 6.33 years (Figure 1).

Identification of factors associated with the death of patients with breast cancer was sought through univariate analysis among epidemiological factors (age, level of education, profession, marital status, religion), clinical factors (reasons for consultations, clinical signs, tumor location, clinical stage), paraclinical characteristics (histological type, differentiation grade, hormonal receptors, Her-2) and therapeutic variables (type of surgery, chemotherapy, radiotherapy, hormonal therapy). Multivariate analysis of the determinants of survival of patients with breast cancer, according to the Cox model, made it possible to identify independent factors associated with a reduction in survival. These include breast ulceration, presence of bilateral tumor, and clinical stage 3 disease. Meanwhile, the ACR3 classification on imaging was associated with improved patient survival.

Table 3 reveals that patients with ulcerated cancer are three times more likely

Table 3. Multivariate analysisof factors associated with survival following the Cox model.

Variables

Death

Adjusted P value

Adjusted HR

(95% CI)

Yes

No

Married

59 (11.8)

197 (39.4)

0.365

1.57 ( 0.91 - 2.47 )

Diabetes

7 (1.4)

8 (1.6)

0.129

3.74 ( 0.70 - 10.58)

Breast pain

22 (4.5)

51 (10.3)

0.071

1.95 (0.99 - 3.41)

Vegetative ulceration

8 (1.6)

9 (1.8)

0.002

3.23 (1.51 - 6.89)

Bilateral breast lesion

5 (1.0)

4 (0.8)

0.001

9.152 (3.52 - 23.85)

Stage 3

58 (11.6)

180 (36.0)

0.010

1.72 (1.14 - 2.56)

ACR3

3 (0.6)

44 (8.8)

0.005

0.19 (0.59 - 5.98)

CIS carcinoma

0

29 (5.8)

0.955

0.79 ( 0.76 - 1.83)

Luminal A

4 (0.8)

47 (9.4)

0.078

0.32 (0.11 - 1.92)

to die early (aHR = 3.23; p = 0.002). Mortality is nine times higher in breast cancer patients with bilateral involvement (aHR = 9.2; p<0.001). Mortality is twice as high (1.72 times) in the group presenting clinical stage 3, in a statistically significant manner (aHR = 1.72; p = 0.010). On the other hand, the survival of patients with breast cancer is significantly improved for those whose imaging results classified their lesion as ACR 3 (aHR = 0.19; p = 0.005).

The survival of patients with breast cancer at 1, 2, 3, 4 and 5 years was respectively 90.6%, 83.1%, 74.2%, 69.8, and 59.2%. Mortality was higher in the first three years. The median survival in our study population was not reached. However, the first quartile was Q1 = 36 months (or 3 years). This means that the death rate in our population did not reach 50% during the follow-up period. However, 25% of patients died before 3 years. These results are similar to those observed in Cameroon by Noa et al. and Mapoko et al. The study conducted by Noa et al. in 2022 [7] showed 2, 3 and 5 year survival of 79.4%, 73.3% and 62% of patients operated on for breast cancer. In 2023, Mapoko et al. had found respectively at 3 and 5 years 65.11% and 58.6% [8] . In Africa, in general, and sub-Saharan Africa in particular, low 5-year survival rates close to 50% have been reported. South Africa in 2018 reported a 5-year survival of 61%. On the other hand, this rate is twice higher than that found by Kemfang et al in 2015 in Cameroon, in a study carried out over the period from 1995 to 2007, where 5-year survival was 30% [9] . This difference could be explained by the fact that our study focused on patients whose follow-up is recent from 2017 to 2022, and suggests that therapeutic advances have improved survival. However, our 5-year survival is still low compared to those observed in Japan and Sweden, where survival rates of around 80% are reported for women with breast cancer [28] . The 5-year survival was 73% and 85% in African American and White American women respectively [29] . In 2016 in France, the 5-year survival rate for breast cancer was 88% [30] . These variabilities in survival observed in Africa and elsewhere could be explained by the fact that in African countries, more than half of the patients had advanced stages at diagnosis, as in our study population. Furthermore, financial, geographical and therapeutic accessibility are limited, and the use of alternative therapies and problems with therapeutic compliance are significant [17] .

After multivariate analysis, the independent factors associated with a reduction in survival were breast ulceration (aHR = 3.23; p = 0.002), bilateral location of the tumor (aHR = 9.2; p < 0.001) and clinical stage 3 (aHR = 1.72; p = 0.010) while ACR3 classification on imaging (aHR = 0.19; p = 0.005) improves patient survival. However, many other factors were not associated with survival despite the fact that they were shown to be in many other studies. These include age, histological type, receptor expression, hormonal and Her-2 receptors, and treatment modalities.

· Breast ulceration: Breast ulceration (aHR = 3.23; p = 0.002) reduced survival. This is explained by the fact that skin involvement in breast cancer classifies the lesion as stage T4b. This underscores why the ulcerated state is considered an indication of advanced disease.

· Bilateral location of the lesion: It was associated with poor survival (aHR = 9.2; p < 0.001). As observed in a study conducted in China in 2021 by Haaron et al, the bilaterality of the breast lesion (p = 0.01) was a significant prognostic parameter reducing the overall survival of breast cancer [31] .

· Clinical stage 3: It was associated with reduced survival (aHR = 1.72; p = 0.010). These results are similar to those observed in Tunisia in 2021 where clinical stage 3 (p=0.016) was associated with a reduction in survival [25] .

· The ACR 3 classification: Patients whose imaging classification ACR3 (aHR = 0.19; p = 0.005) had better survival. This could be explained by the fact that they were going to consult early and that the diagnosis could have been made early during the monitoring of the ACR3 lesion and treatment accordingly started earlier.

These were mainly due to missing data, either due to incomplete or unfound files, or to patient loss to follow-up.