Analysis of breast cancer survival in a northeastern Brazilian state based on prognostic factors: A retrospective cohort study.

Breast cancer is a major health problem worldwide. Analysis of breast cancer epidemiology in emerging countries enables assessment of prognostic factors, cancer care quality, and the equity of resource distribution. We aimed to estimate the overall (OS) and cancer-specific survival (SS) of breast cancer patients in the northeastern Brazilian state of Sergipe to identify independent prognostic factors. We analyzed a cohort for the factors age at diagnosis, place of residence, time to treatment, staging, and molecular classification, using the Kaplan-Meier method, log-rank test, Pearson's chi-squared test and Cox regression model. The outcome was the vital status at the end of the study. Our analysis showed an OS probability of 0.72 and an SS probability of 0.75. In multivariate analysis, time to treatment within 60 days, stage IV, and triple-negative classification remained independent prognostic factors for both OS [unadjusted hazard ratio (HRp) 1.50 (1.21; 1.86), HRp 16.56 (8.35; 32.85), and HRp 2.73 (1.73; 4.29), respectively] and SS [HRp 1.43 (1.13; 1.81), HRp 20.53 (9.45; 44.56), and HRp 3.14 (1.88; 5.26), respectively]. Better survival was demonstrated for the following patients: those receiving their first treatment after 60 days, with an OS of 52.5 months (51.2; 53.8) and SS of 53.5 months (52.3; 54.7); stage I patients, with an OS of 58.8 months (57.7; 60.0) and SS of 59.2 months (58.1; 60.3); patients without nodal metastasis, with an OS of 54.2 months (53.0; 55.4) and SS of 55.6 months (54.5; 56.7); and patients with luminal A classification, with an OS of 56.8 months (55.0; 58.5) and SS of 57.8 months (56.2; 59.4). This study identified independent prognostic factors and that OS and SS were lower for patients from Sergipe than for patients in high-income areas. Therefore, determining the profiles of breast cancer patients in this population will inform specific cancer care.

Introduction

Breast cancer is a major public health problem because it has high incidence rates with consequent high morbidity and mortality. It is the second most common cancer in women after nonmelanoma skin cancer worldwide, with over two million new cases annually [1-3]. Despite early detection resulting in favorable prognosis, breast cancer is still the leading cause of cancer death among women, especially in economically deprived regions [1,4,5].

In some high-income countries, breast cancer incidence and mortality rates have steadily increased, while in others these rates have decreased [6]. However, in low- and middle-income countries, breast cancer incidence and mortality rates have consistently increased [3,7,8]. According to estimates from the Brazilian National Cancer Institute, breast cancer was the most common type of cancer in women (excluding nonmelanoma skin cancer) in 2020, with over 66,000 cases and a mean age-standardized rate (ASR) of 43.74 per 100,000 women. In Northeast Brazil, an economically deprived region, the mean ASR was 43.74 per 100,000. In Sergipe, a northeastern Brazilian state, the mean ASR was 44.27 per 100,000 [3]. The Brazilian Mortality Information System recorded 16,593 deaths from breast cancer in 2019, and the North and Northeast regions displayed the highest rates [9].

Cancer statistics can reveal the effectiveness of public health policies, the equity of resource distribution, and the impact of predictive factors on survival. Therefore, assessing breast cancer survival can help establish criteria for the objective evaluation of patient prognosis and can contribute to the improvement of cancer control strategies [10].

In Brazil, factors such as study design, calendar year, and region and population studied might explain differences in survival [11-13]. Several other factors, such as staging, age at diagnosis, time from diagnosis to treatment, race, histology, and socioeconomic status, may also play a role; however, their contributions are still uncertain [13-18]. It is also noteworthy that breast neoplasms of similar histological subtype can present different outcomes, which may be consistent with the molecular subtype [18].

Based on these assumptions and considering that survival studies in economically deprived regions are scarce, the present study aimed to identify independent prognostic factors for breast cancer survival in the northeastern Brazilian state of Sergipe and to assess how they influenced survival in the study population.

Materials and methods

We collected data from a retrospective hospital cohort of breast cancer patients treated in the main cancer facility in the state of Sergipe, Brazil from 2014 to 2010. Patients were followed up for at least 60 months.

We used the hospital-based cancer registry (HCR) database to retrieve information from women with invasive breast cancer diagnosed at either the facility or elsewhere. This facility is the largest referral center for the study population; therefore, a considerable number of advanced cases are registered for management. The HCR personnel input demographic, tumor, and treatment variables into the HCR information system, data obtained solely from medical records.

We selected breast cancer cases using the International Classification of Diseases, Oncology, 3rd edition (ICD-O-3), topographical codes C50.0 to C50.9, and morphological codes, 8050/3, 8211/3, 8480/3, 850_/3, 851_/3, 852_/3, 853_/3, and 854_/3.

To define the vital status and to collect information concerning the date and underlying cause of death, the HCR was used to search the Brazilian Mortality Information System of the Ministry of Health. To complement information concerning death, the HCR was used to access the following databases: 1) the National Deceased Registry (CNF Brazil), 2) the Federal Revenue Service of Brazil, 3) the Brazilian Electoral System, and 4) the National Health Registry.

The Authorization for Outpatient Procedures database was accessed for additional information concerning staging, molecular classification, hormone therapy, chemotherapy, and radiotherapy. The database of the Authorization Hospital Admissions of the Ministry of Health provided information on clinical and surgical admissions.

The variables used were as defined on the standard tumor registry form [19]. For age at diagnosis, we employed age groups according to the hormonal phases (≤45 years, 45 to 54 years, 55 to 64 years, and ≥65 years). We selected other variables, such as place of residence (whether from the capital or countryside/outside), time from diagnosis to treatment, staging [20], histological type, molecular classification (defined as luminal A, luminal B, HER2/neu-enhanced, or triple-negative after immunohistochemical profiling and Fluorescence In Situ Hybridization (FISH) test whenever necessary), and vital status. The time to treatment was set at ≤60 days and >60 days to conform to Law No. 12 732/2012 [21]. Missing data were considered confounding variables that might influence survival estimates.

Statistical analysis

We used the Kaplan–Meier method to estimate the survival probability of the cohort and then calculated OS and SS for each time interval as the number of women surviving divided by the total number at risk.

To appraise differences among survival distributions, we applied the log-rank test and checked whether any factor would influence the time to event. The Bonferroni method provided compensation for the effect of multiple comparisons, provided correction for the log-rank test results, and assessed differences among several subgroups of variables to control significance levels by adjusting P values.

To evaluate the effect of multiple independent variables and the burden that some prognostic factors may impose upon the outcome, we resorted to Cox’s proportional risk model. The method required evaluation of the independent variables by a univariate analysis and then by a multivariate analysis, identifying hazard ratios, adjusted (HR) and unadjusted (HRp), and 95% confidence intervals. To test the proportional hazards assumptions, we employed the method based on scaled Schoenfield residuals.

Pearson’s chi-squared model was used to analyze the differences in proportions between the categorized variables to a 5% significance level. The backward selection method selected variables that would fit the tests. We used R Core Team 2020 to perform all the analyses.

Ethical considerations

The Research Ethics Committee of the Federal University of Sergipe approved this research. We conducted all methods in accordance with relevant guidelines and regulations. As patient databases remained anonymized, obtaining informed consent was not possible. Consequently, as specified in Resolution number 466, December 12, 2012, of the Ministry of Health of Brazil, the ethics committee granted exemption from the necessity for informed consent. In addition, all data remained confidential to be used exclusively for scientific purposes.

Results

We included 1,278 women with invasive breast cancer in this analysis. Of these, 966 were alive and 312 had died by the end of follow-up. Considering place of residence, 60.7% of the patients lived in the countryside. The median age at diagnosis was 55 years and patients were distributed similarly among the age groups. Invasive ductal carcinomas were the most frequent breast neoplasm subtype (90.1%). A high number of patients (47.6%) had their first treatment 60 days after diagnosis. Most patients were stage II (32.2%); however, this information was missing from the medical records of 22.9% of cases. Most of the patients did not have lymph node involvement (42.4%) but, again, the number of missing data points was high (32.4%). Most of the cases (30.6%) had their molecular status determined as luminal B. It should be noted that Ki67 was not stained for in 27.5% of cases, preventing determination as either luminal A or B; therefore, these cases were considered luminal X (Table 1).

Table 1

Descriptive statistics and mean overall and cancer-specific survival of the cohort during the study.

Characteristic	Findings N (%)	MOS (95% CI)	p-value	MSS (95% CI)	p-value
No. of patients    Deceased    Alive	1,278966 (75.6)312 (24.4)	50.5 (49.5; 51.5)		52.0 (51.1; 53.0)
Residence    Capital    Outside the capital	502 (39.3)776 (60.7)	51.4 (49.9; 52.9)49.9 (48.6; 51.2)		52.2 (50.7; 53.6)51.9 (50.7; 53.1)
Age groups    <45    45–55    56–65    >65    Median age (years)	356 (27.9)352 (27.5)290 (22.7)280 (21.9)55	50.2 (48.3; 52.1)50.8 (48.9; 52.6)51.0 (49.0; 53.0)49.9 (47.7; 52.1)	0.772	51.0 (49.2; 52.9)51.8 (50.0; 53.6)53.7 (51.9; 55.4)51.9 (49.8; 54.0)	0.364
Histology types    IDC    ILC    Special	1151 (90.1)54 (4.2)73 (5.7)	50.7 (49.6; 51.7)48.4 (43.8; 53.1)49.3 (44.9; 53.7)	0.123	52.2 (51.2; 53.2)50.3 (45.8; 54.8)50.4 (46.2; 54.7)	0.101
T to T (days)    ≤60    >60    Missing	537 (42.0)608 (47.6)133 (10.4)	48.0 (46.3; 49.6)52.5 (51.2; 53.8)51.2 (48.0; 54.4)	<0.001	50.1 (48.5; 51.7)53.5 (52.3; 54.7)52.7 (49.6; 55.8)	0.001
Staging (TNM)    I    II    III    IV    Missing	196 (15.3)412 (32.2)348 (27.2)29 (2.3)293 (22.9)	58.8 (57.7; 60.0)55.5 (54.3; 56.8)46.5 (44.5; 48.5)34.9 (27.1; 42.6)44.0 (41.5; 46.6)	<0.001	59.2 (58.1; 60.3)56.7 (55.5; 57.8)48.5 (46.5; 50.4)36.8 (29.0; 44.6)46.0 (43.5; 48.5)	<0.001
Lym node    Absent    Present    Missing	542 (42.4)322 (25.2)414 (32.4)	50.5 (48.5; 52.4)54.2 (53.0; 55.4)45.6 (43.5; 47.7)	<0.001	52.0 (50.1; 53.8)55.6 (54.5; 56.7)47.3 (45.2; 49.3)	<0.001
Mol Clas    luminal A    luminal B    luminal X    HER2 over    Triple-negative    Missing	140 (11.0)391 (30.6)352 (27.5)63 (4.9)201 (15.7)131 (10.3)	56.8 (55.0; 58.5)53.5 (52.1; 55.0)51.9 (50.2; 53.6)47.3 (42.3; 52.3)43.0 (40.0; 46.0)43.9 (39.9; 47.9)	<0.001	57.8 (56.2; 59.4)54.6 (53.2; 56.0)53.1 (51.5; 54.7)52.5 (48.2; 56.7)44.6 (41.6; 47.6)46.0 (42.1; 49.9)	<0.001

MOS: Mean overall survival; 95% CI: 95% confidence interval; MSS: Mean cancer-specific survival; Missing: Cancer registrar did not identify data in the medical records; T to T: Time to treatment; TNM: TNM Staging System, 7th Edition; Lym node: Lymph node involvement; Mol Clas: Molecular classification; luminal X: Estrogen and/or Progesterone receptor positivity, Ki67 not tested.

We estimated the five-year survival probability of the study cohort as cancer-specific survival (SS) of 0.77 (95% CI 0.74; 0.80) and overall survival (OS) of 0.72 (95% CI 0.69; 0.75) (Figs 1 and 2).

Fig 1

Overall survival and mean overall survival of breast cancer patients by histology type, time to treatment, staging, lymph node involvement, and molecular classification.

Fig 2

Cancer-specific survival and mean cancer-specific survival of breast cancer patients by histology type, time to treatment, staging, lymph node involvement, and molecular classification.

Less favorable survival estimates were produced for patients who had their first treatment within 60 days, with an OS of 48 months (95% CI 46.3; 49,6) and SS of 50.1 months (95% CI 48.5; 51.7); for patients in stage IV, with an OS of 34.9 months (95% CI 27.1; 42.6) and SS of 36.8 months (95% CI 29.0; 44.6); for patients with lymph node metastasis, with an OS of 50.5 months (95% CI 48.5; 52.4); and for patients with triple-negative classification, with an OS of 43 months (95% CI 40.0; 46.0) and SS of 44.6 months (95% CI 41.6; 47.6). In contrast, better survival estimates were produced for patients who had their first treatment after 60 days, with an OS of 52.5 months (95% CI 51.2; 53.8) and SS of 53.5 months (95% CI 52.3; 54.7); for patients in stage I, with an OS of 58.8 months (95% CI 57.7; 60.0), and SS of 59.2 months (95% CI 58.1; 60.3); for patients without lymph node metastasis, with an OS of 54.2 months (95% CI 53.0; 55.4) and SS of 55.6 months (95% CI 54.5; 56.7); and for patients with luminal A classification, with an OS of 56.8 months (95% CI 55.0; 58.5) and SS of 57.8 months (95% CI 56.2; 59.4) (Table 1).

In the univariate (unadjusted) analysis, time to treatment within 60 days, stage IV and triple-negative molecular classification significantly impacted survival. In the multivariate analysis, these variables remained independent prognostic factors for OS [HRp 1.50 (95% CI 1.21; 1.86), HRp 16.56 (95% CI 8.35; 32.85), and HRp 2.73 (95% CI 1.73; 4.29), respectively] and SS [HRp 1.43 (95% CI 1.13; 1.81), HRp 20.53 (95% CI 9.45; 44.56), and HRp 3.14 (95% CI 1.88; 5.26), respectively] (Table 2). Even though the Schoenfield test rejects the hypothesis of hazard proportionality, S1 and S2 Figs show that hazards remain fairly constant throughout the follow-up period, except for the cancer-specific survival variables. Thus, to explain non proportionality, time-dependent variables were presented (Table 3).

Table 2

Effect of different prognostic factors on the survival probability of patients with breast cancer using the univariate (unadjusted) and multivariate (adjusted) cox regression model.

	OVERALL SURVIVAL			CANCER-SPECIFIC SURVIVAL
Variable	HR (95% CI)	HRp (95% CI)c	p-value	HR (95% CI)	HRp (95% CI)	p-value
Age group≤45≤40≤5546–5556–65>65	1.01.26 (0.98; 1.64)0.97 (0.80; 1.19)0.96 (0.73; 1.26)0.93 (0.70; 1.24)1.08 (0.82; 1.44)			1.01.39 (1.05; 1.83)1.16 (0.92; 1.45)0.94 (0.70; 1.25)0.75 (0.54; 1.04)0.93 (0.68; 1.26)
ResidenceOutside the capitalCapital	1.00.90 (0.73; 1.10)			1.01.02 (0.81; 1.28)
T to T>60≤60Missing	1.01.53 (1.24; 1.90)0.91 (0.62; 1.34)	1.50 (1.21; 1.86)0.64 (0.42; 0.97)	<0.0010.036	1.01.47 (1.16; 1.85)0.83 (0.53; 1.29)	1.43 (1.13; 1.81)0.57 (0.35; 0.91)	0.0030.019
StagingIIIIIIIVMissing	1.02.61 (1.47; 4.63)7.65 (4.42; 13.22)15.78 (7.97; 31.26)7.61 (4.38; 13.25)	2.41 (1.36; 4.28)6.97 (4.01; 12.09)16.56 (8.35; 32.85)6.86 (3.93; 11.97)	0.003<0.001<0.001<0.001	1.02.85 (1.46; 5.58)8.75 (4.60; 16.67)19.66 (9.07; 42.62)8.93 (4.66; 17.10)	2.62 (1.33; 5.13)7.98 (4.17; 15.27)20.53 (9.45; 44.56)7.99 (4.16; 15.37)	0.005<0.001<0.001<0.001
Lym NodeMissingNegativePositive	1.01.46 (1.12; 1.91)2.00 (1.57; 2.53)			1.01.57 (1.17; 2.11)2.25 (1.73; 2.93)
Mol ClasLuminal ALuminal BLuminal XHER2 overTriple-NegMissing	1.01..45 (0.93; 2.28)1.82 (1.17; 2.84)2.55 (1.44; 4.52)3.44 (2.19; 5.40)2.90 (1.78; 4.71)	1.27 (0.81; 1.99)1.25 (0.79; 1.95)1.95 (1.09; 3.49)2.73 (1.73; 4.29)2.82 (1.69; 4.71)	0.3070.3380.024<0.001<0.001	1.01.62 (0.97; 2.71)1.97 (1.18; 3.28)2.07 (1.03; 4.15)4.00 (2.40; 6.68)3.24 (1.86; 5.63)	1.01.40 (0.84; 2.34)1.33 (0.79; 2.22)1.52 (0.74; 3.11)3.14 (1.88; 5.26)3.18 (1.78; 5.66)	0.2000.2800.253<0.001<0.001
HistologyIDCILCSpecial	1.01.52 (1.00; 2.32)1.17 (0.77; 1.77)			1.01.54 (0.97; 2.45)1.31 (0.85; 2.03)

HR: Unadjusted hazard ratio; HRp: Adjusted hazard ratio; 95%CI: 95% confidence interval; T to T: Time to treatment; Missing: Cancer registrar did not identify data in medical record; Lym Node: Lymph node involvement; Mol Clas: Molecular classification; HER2 over: HER2 overexpression; Triple-Neg: Triple-negative: IDC: Invasive ductal carcinoma; ILC: Invasive lobular carcinoma.

Table 3

Effect of time-dependent variables on the survival probability of patients with breast cancer using the univariate (unadjusted) and multivariate (adjusted) cox regression model.

Variable	HR (95% CI)	HRp (95% CI)	p-valor
Age group≤40	1,26 (0,98–1,64)
≤55	0,97 (0,80–1,19)
46–55	0,96 (0,73–1,26)
56–65	0,93 (0,70–1,24)
>65	1,08 (0,82–1,44)
ResidenceCapital	0,90 (0,73–1,10)
Treated	1,36 (0,94–1,98)
T to T≤60	1,53 (1,24–1,90)	1,50 (1,21–1,86)	<0,001
Missing	0,91 (0,62–1,34)	0,60 (0,40–0,92)	0,018
StagingII	2,61 (1,47–4,63)	2,53 (1,43–4,49)	0,001
III	7,65 (4,42–13,22)	7,25 (4,19–12,55)	<0,001
IV	15,78 (7,97–31,26)	16,90 (8,52–33,50)	<0,001
Missing	7,61 (4,38–13,25)	7,15 (4,10–12,45)	<0,001
Lym NodePositive	1,46 (1,12–1,91)
Negative	2,00 (1,57–2,53)
Mol ClasLuminal B	1,45 (0,93–2,28)
Luminal X	1,82 (1,17–2,84)
HER2 over	2,55 (1,44–4,52)
Triplo-Neg	3,44 (2,19–5,40)
Missing	2,90 (1,78–4,71)
HistologyILC	1,52 (1,00–2,32)
Non Special	1,17 (0,77–1,77)

HR: Unadjusted hazard ratio; HRp: Adjusted hazard ratio; 95%CI: 95% confidence interval; T to T: Time to treatment; Missing: Cancer registrar did not identify data in medical record; Lym Node: Lymph node involvement; Mol Clas: Molecular.

Discussion

The present study demonstrated that time to treatment, staging, and molecular classification of HER2 significantly impacted OS and SS in the univariate (unadjusted) analysis. In the multivariate (adjusted) analysis, time to treatment after 60 days, stage IV, and triple-negative classification remained independent prognostic factors. The survival estimates observed in the study were lower than those found in some affluent areas of Brazil [22-24], as well as in high-income countries [25,26] and China (89.4%) [27].

Patients in this study under the age of 40 years had a lower OS and SS than older patients. Some studies report that patients under 40 years of age usually present unfavorable prognostic characteristics, usually associated with advanced staging, HER2 overexpression and nodal metastasis [28]. Nixon et al. (1994) reported that women under 35 years of age had poor tumor differentiation, lymphatic involvement, necrosis, and estrogen receptor negativity; consequently, they had more recurrences and distant metastases [29]. In contrast, older women present less aggressive features but have several comorbidities that, when associated with advanced stages, might contribute negatively to survival [13], although this was not shown in our data.

An intriguing finding was that time to treatment ≤ 60 days can be considered as a prognostic factor. While reanalyzing our data, we assumed that this was because of factors such as younger age and more advanced and triple-negative tumors; however, it remained an independent prognostic factor after multivariate analysis. It is possible that improvement in health care plays a role in this finding. The difficulty in accessing diagnosis and treatment in low- and middle-income countries has to be overcome because better cancer survival is a consequence of early diagnosis and timely treatment [30,31].

Advanced staging remained an independent prognostic factor after multivariate analysis, and it became quite clear that both OS and SS decreased as staging progressed. Bulky tumors directly interfere with the quality of life of patients with breast cancer. Conversely, patients presenting at early stages undergo less aggressive modalities of treatment and face a lower risk of death [32]. In the present study, stage IV indicated an increased risk of death, as also determined by Höfelman et al. (2014) [32] and Fayer (2014) [33].

We were cautious in our analysis because of the high percentage of missing information, mainly in staging, which might have influenced the results. The group with missing data on staging was presented as an independent factor in the multivariate analysis, as also performed by Basílio (2011) [28] and Brito, Portela and Vasconcellos (2009) [34]. Ayala et al. (2019) warned that failure to register this information, especially in the Breast Cancer Information System (SISMAMA), would compromise data monitoring [13].

The different survival probability estimates for different molecular classifications might require different considerations. The different survival probabilities between luminal A and luminal X might be caused by the portion of HER2-overexpressing tumors that were not detected. In addition, missing data (approximately 10%) were shown to be an independent prognostic factor in multivariate analysis, also denoting a confounding factor for survival. Some studies report that a lack of this information may be associated with difficulty in accessing adequate diagnosis and treatment and may be correlated with social status [33,34]. Apart from that, triple-negative classification was a clear independent prognostic factor. Al-Thoubaity (2020) reported that HER2 overexpression and triple-negative features were the most frequently observed; they were associated with an early surge in young women, usually harboring bulky tumors and lymph node metastases [35].

In the present study, the most common histology type was invasive ductal carcinoma, while invasive lobular carcinoma comprised only a small part of the cohort, which was similar to the findings of Basílio (2011) [28]. In our study, lobular carcinomas indicated worse prognosis, which was also observed by others [36-41].

The independent prognostic factors identified in our multivariate analyses were in agreement with hospital-based studies [33,42]. Among these, lymph node metastasis only impacted OS and SS in univariate analysis. Some studies estimated that it indicated an increased risk of death of four to eight times [30,43].

Some limitations of this research should be considered, such as the use of retrospective secondary data from medical records without controls and a lack of standardization of the information in pathological reports. They might have interfered with the accuracy of the presented results.

Despite these limitations, we have determined prognostic factors and estimated the survival probabilities of cancer patients in a northeastern Brazilian state. The outcomes observed indicate the need to improve the cancer care system in this region. The data obtained support the implementation of targeted strategies to improve breast cancer survival irrespective of socioeconomic and cultural background, with the ultimate aim of healthcare equity.

Conclusions

In summary, our results indicate that independent prognostic factors, such as time to treatment ≤ 60 days, advanced stage, and triple-negative molecular classification, significantly impact OS and SS. In addition, a lack of information, such as staging and molecular classification, may compromise survival analysis and, consequently, jeopardize cancer care actions.

Estimating OS and cancer-specific survival provided a better understanding of the profile of breast cancer patients treated in the state of Sergipe. This emphasizes the need for specific health policies to improve access to cancer facilities for early diagnosis and timely treatment.

The proportional hazards assumptions for the cox model, considering overall survival estimates.

(TIFF)

Click here for additional data file.

The proportional hazards assumptions for the cox model, considering specific survival estimates.

(TIFF)

Click here for additional data file.

18 Mar 2021

PONE-D-20-36356

Breast cancer survival analysis based on prognostic fators in a state of Northeastern Brazil: a retrospective cohort study

PLOS ONE

Dear Dr. Marques,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Your manuscript has been assessed by two external reviewers, who have raised a number of concerns about the methodology and statistical analyses used. Please ensure that these are addressed in full as part of your revisions.

Please submit your revised manuscript by Apr 30 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Joseph Donlan

Senior Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In your ethics statement in the Methods section and in the online submission form, please provide additional information about the data used in your retrospective study. Specifically, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

3. Please include the date(s) on which you accessed the databases or records to obtain the retrospective data used in your study.

4. We note that the grant information you provided in the ‘Funding Information’ and ‘Financial Disclosure’ sections do not match.

When you resubmit, please ensure that you provide the correct grant numbers for the awards you received for your study in the ‘Funding Information’ section.

5. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

6. Please upload a copy of Supporting Information Files which you refer to in your text on page 19.

7. Thank you for stating in your Funding Statement:

"CAL - This research was conducted with the partial support of a Research Development Grant from the Fundação de Apoio à Pesquisa e à Inovação Tecnológica do Estado de Sergipe - FAPITEC/SE Protocol: 019.203.00961/2018-2. (https://fapitec.se.gov.br/)"

Please provide an amended statement that declares *all* the funding or sources of support (whether external or internal to your organization) received during this study, as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now.  Please also include the statement “There was no additional external funding received for this study.” in your updated Funding Statement.

Please include your amended Funding Statement within your cover letter. We will change the online submission form on your behalf.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: No

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: 1. The fact that the material method is more comprehensive in the abstract, such as the number of patients.

2. Is there contrast between the time between diagnosis and 1st treatment and the result? Shouldn't it be negative to survival as the time between diagnosis and 1st treatment is prolonged? should be discussed with the literature.

3. In the analysis, what is the ratio of unknowns in luminal without ki-67 % information, stage, lymph node, molecular classification. (As far as it is understood, these rates are over 10%.). Isn't it necessary to analyze these unknown data without consideration?

4. Texts in figures are not readable.

Reviewer #2: This would be interesting to see this manuscript published however this is significant and substantial work required before it meets publication standards.

• The Short title does not seem appropriate for the study presented – the aim was to identify prognostic factors. It might be more appropriate to have a Short title, “Breast cancer survival analysis to identify prognostic factors: a retrospective cohort study.

• As this is a single centre study, are referral patterns likely to have influenced these results?

• What is the likely impact of missing data?

• Missing data (for each respective variable) should be reported in the manuscript.

• A secondary aim was to consider the survival according to time to treatment. Surprisingly, women who were treated within 60 days had poorer survival than those treated after 60days.

• Age specific HR reported in Table 3 are confusing. Does the <=55 group also include the <=40 group? What is the reference category here?

o Ideally the reference category for each variable should be included in the Table.

o Are the HR reported (as opposed to the HRp) univariate or multivariate? The results should be clearly explained so there is no doubt for the reader in terms of what they are reading. Presumably the HRp results are the outcome of the backward selection process?

Methods

• Description of the Hospital Registry of Cancer should be included. How are these data collected? Do they include all breast cancer diagnoses and admissions? What variables are collected? Are these medical records, registry data, administrative data?

• Brief and do not seem to describe all methods used.

Statistical Analysis

• No mention of Bonferroni correction used in Table 1

• For the Cox PH models, were the assumptions of this model tested? And how?

Results

• A descriptive table (Table 1) should be provided that provides descriptive statistics of the cohort analysed.

• Kaplan Meier estimates of 5 year survival should be accompanied by confidence intervals to indicate the uncertainty around the estimates.

• Table 1 is very difficult to read and interpret. I would suggest at a minimum that survival days be converted to months or years. Are these median survival estimates?? I would expect to see a table (or two, for each survival outcomes) that reported % 5 yr survival and confidence intervals for each covariate.

• Unclear the purpose of Table 2

• Table titles and figure title should be meaningful and complete, indicating if estimates are adjusted or unadjusted.

• The results of the lymph node status (Table 3) are surprising – I would expect better survival for those who are lymph node negative. Could the authors please confirm these results are correct? Similarly for time to treatment, I would expect women treated earlier to typically have better survival.

Discussion

•

• Ln 242-244: Should this indicate that the study has reported prognostic factors for breast cancer AND overall survival? It is unclear what the reference to time between diagnosis and first treatment refer to – these are not outcomes that have been presented.

• All discussion around the survival outcomes should clearly differentiate between overall survival and breast cancer specific survival, as both these outcomes have been reported.

• Ln253-255: Unclear what results the correlation with time to treatment refers to.?

• LN262-266: Discusses that comorbidity levels and advanced stage contributes to poorer survival by older women however, the results presented in this manuscript indicate that poorer women have poorer breast cancer specific survival, and equivalent overall survival to older women. The authors should discuss and explain this finding.

• Ln270-273: The authors speculate that poor survival in those treated within 60 days may be due to advanced stage and/or triple negative tumours. But if the multivariate (if they are adjusted) results indicate that time to treatment is still a predictor when adjusted for stage and triple negatives, then this does not fully explain why survival is poorer when treated within 60 days.

• LN283 – Does standardise analysis refer to multivariate analysis? The latter would be the preferred term. Standardisation is something very different.

• There is a great focus in the Discussion and Results on the univariate analyses, which are not as important in a study that is looking to identify independent predictors of survival. It would be recommended to focus mainly on the multivariate results in response to the research question.

• Ln322-327: Discussion of missing data is appropriate however, the extent of which has not been quantified in results. It is important to do so.

• Ln 335 indicates that results were stratified, I have seen no evidence of this.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

1 Jun 2021

Response to Reviewers

Reviewer #1:

1. The fact that the material method is more comprehensive in the abstract, such as the number of patients.

R. We revised and completed material and methods to comply with this observation

2. Is there contrast between the time between diagnosis and 1st treatment and the result? Shouldn't it be negative to survival as the time between diagnosis and 1st treatment is prolonged? should be discussed with the literature.

R. We inserted the following to discuss this point: “An intriguing feature resulting from our estimates was time to treatment ≤ 60 days to be taken as a prognostic factor. After reanalyzing our data, we assumed it was due to factors such as younger age and more advanced and triple-negative tumors; however, it remained an independent prognostic factor after multivariate analysis. Whether improvement in health care would play any role in modifying it can be questioned. The difficulty in accessing diagnosis and treatment in low- and middle-income countries has to be overcome [30] because better cancer survival is a consequence of early diagnosis and timely treatment [31].”

3. In the analysis, what is the ratio of unknowns in luminal without ki-67 % information, stage, lymph node, molecular classification. (As far as it is understood, these rates are over 10%.). Isn't it necessary to analyze these unknown data without consideration?

R. We added this piece in discussion: “We are cautious in analyzing the high percentage of missing information, mainly in staging, which might have influenced the results. The group with missing data on staging was presented as an independent factor in the multivariate analysis, as Basílio (2011) [28] and Brito, Portela and Vasconcellos (2009) [45] confirmed. Ayala et al. (2019) warned that failure to register this information, especially in the Breast Cancer Information System (SISMAMA), would compromise data monitoring [13].”

4. Texts in figures are not readable

R. We edited figures to a different format.

Reviewer #2: This would be interesting to see this manuscript published however this is significant and substantial work required before it meets publication standards.

R. We had the manuscript edited by Nature Research Editing Service (certificate below)

• The Short title does not seem appropriate for the study presented – the aim was to identify prognostic factors. It might be more appropriate to have a Short title, “Breast cancer survival analysis to identify prognostic factors: a retrospective cohort study.

R. We agreed and changed it.

• As this is a single centre study, are referral patterns likely to have influenced these results?

R. We think so. In fact, we added “We used the hospital-based cancer registry (HCR) database to retrieve information from women with invasive BC diagnosed at either the facility or elsewhere. Since it is the largest referral center for the study population, a considerable number of advanced cases are registered for management” in Materials and Methods.

• What is the likely impact of missing data?

R. We added this piece in discussion: “We are cautious in analyzing the high percentage of missing information, mainly in staging, which might have influenced the results. The group with missing data on staging was presented as an independent factor in the multivariate analysis, as Basílio (2011) [28] and Brito, Portela and Vasconcellos (2009) [45] confirmed. Ayala et al. (2019) warned that failure to register this information, especially in the Breast Cancer Information System (SISMAMA), would compromise data monitoring [13].”

• Missing data (for each respective variable) should be reported in the manuscript.

R. That was reported in results and in Table 1

• A secondary aim was to consider the survival according to time to treatment. Surprisingly, women who were treated within 60 days had poorer survival than those treated after 60days.

R. We inserted the following to discuss this point: “An intriguing feature resulting from our estimates was time to treatment ≤ 60 days to be taken as a prognostic factor. After reanalyzing our data, we assumed it was due to factors such as younger age and more advanced and triple-negative tumors; however, it remained an independent prognostic factor after multivariate analysis. Whether improvement in health care would play any role in modifying it can be questioned. The difficulty in accessing diagnosis and treatment in low- and middle-income countries has to be overcome [30] because better cancer survival is a consequence of early diagnosis and timely treatment [31].”

• Age specific HR reported in Table 3 are confusing. Does the <=55 group also include the <=40 group? What is the reference category here?

o Ideally the reference category for each variable should be included in the Table.

o Are the HR reported (as opposed to the HRp) univariate or multivariate? The results should be clearly explained so there is no doubt for the reader in terms of what they are reading. Presumably the HRp results are the outcome of the backward selection process?

R. We rewrote the results and also the tables to better explained the findings

Methods

• Description of the Hospital Registry of Cancer should be included. How are these data collected? Do they include all breast cancer diagnoses and admissions? What variables are collected? Are these medical records, registry data, administrative data?

R. That was inserted: “We used the hospital-based cancer registry (HCR) database to retrieve information from women with invasive BC diagnosed at either the facility or elsewhere. Since it is the largest referral center for the study population, a considerable number of advanced cases are registered for management. Thus, the HCR personnel input demographic, tumor, and treatment variables into the HCR information system, data which were obtained solely from medical records.”

• Brief and do not seem to describe all methods used.

R. We inserted the following: “We selected BC cases using the International Classification of Diseases, Oncology, 3rd edition (ICD-O-3); topographical codes C50.0 to C50.9; and morphological codes 8050/3, 8211/3, 8480/3, 850_/3, 851_/3, 852_/3, 853_/3, and 854_/3.

To define the vital status and collect information as the date and underlying cause of death, we searched the Brazilian Mortality Information System (SIM) of the Ministry of Health. To complement information on death, we accessed the following databases: 1) National Deceased Registry (CNF Brazil), 2) Federal Revenue Service of Brazil, 3) Brazilian Electoral System, and 4) National Health Registry (NHR).

We also accessed the Authorization for Outpatient Procedures database for additional information concerning staging, molecular classification, hormone therapy, chemotherapy, and radiotherapy. The database of the Authorization Hospital Admissions (AHA) of the Ministry of Health provided us with information on clinical and surgical admissions.”

Statistical Analysis

• No mention of Bonferroni correction used in Table 1

R. We inserted the following: “The Bonferroni method provided compensation for the effect of multiple comparisons, provided correction for the log-rank test results, and assessed differences among several subgroups of variables to control significance levels by adjusting P values.”

• For the Cox PH models, were the assumptions of this model tested? And how?

R. We rewrote the information about the Cox model to try to comply with that: ” To evaluate the effect of multiple independent variables and the burden that some prognostic factor may impose upon the outcome, we resorted to Cox's proportional risk model. First, the method evaluated the independent variables as univariate analysis and then multivariate analysis, identifying hazard ratios, adjusted (HR) and unadjusted (HRp), and 95% confidence intervals.

Pearson's chi-squared model provided analysis of the differences in proportions between the categorized variables to a 5% significance level. The backward selection method selected variables that would fit the tests.”

Results

• A descriptive table (Table 1) should be provided that provides descriptive statistics of the cohort analysed.

R. Table 1 now brings the descriptive statistics of the cohort study

• Kaplan Meier estimates of 5 year survival should be accompanied by confidence intervals to indicate the uncertainty around the estimates.

R. We present confidence intervals.

• Table 1 is very difficult to read and interpret. I would suggest at a minimum that survival days be converted to months or years. Are these median survival estimates?? I would expect to see a table (or two, for each survival outcomes) that reported % 5 yr survival and confidence intervals for each covariate.

R. We redesigned the tables, resulting in different Table 1 and 2. I hope that will meet the concerns presented.

• Unclear the purpose of Table 2

R. Ok, tables were redesigned

• Table titles and figure title should be meaningful and complete, indicating if estimates are adjusted or unadjusted.

R. We redesigned them and hope that will do.

• The results of the lymph node status (Table 3) are surprising – I would expect better survival for those who are lymph node negative. Could the authors please confirm these results are correct? Similarly for time to treatment, I would expect women treated earlier to typically have better survival.

R. Sorry, that was a mistake, as the survival curves for that variable show. That was corrected.

Discussion

•

• Ln 242-244: Should this indicate that the study has reported prognostic factors for breast cancer AND overall survival? It is unclear what the reference to time between diagnosis and first treatment refer to – these are not outcomes that have been presented.

R. That was rewritten to “The present study demonstrated that time to treatment, staging, and molecular classification HER2 significantly impacted OS and SS in the univariate (unadjusted) analysis. In the multivariate (adjusted) analysis, time to treatment after 60 days, stage IV, and triple-negative classification remained independent prognostic factors.”

• All discussion around the survival outcomes should clearly differentiate between overall survival and breast cancer specific survival, as both these outcomes have been reported.

R. Discussion was rewritten and that was done.

• Ln253-255: Unclear what results the correlation with time to treatment refers to.?

R. We agreed with that and rewrote to “Patients under the age of 40 years had a lower OS and SS in the study. Some studies report that patients under 40 years of age usually present unfavorable prognostic characteristics, usually associated with advanced staging, HER-2 overexpression and nodal metastasis [28]. Nixon et al. (1994) reported that women under 35 years of age had poor tumor differentiation, lymphatic involvement, necrosis, and estrogen receptor negativity; consequently, they had more recurrences and distant metastases [29].”

• LN262-266: Discusses that comorbidity levels and advanced stage contributes to poorer survival by older women however, the results presented in this manuscript indicate that poorer women have poorer breast cancer specific survival, and equivalent overall survival to older women. The authors should discuss and explain this finding.

R. We agreed that was confusing and changed to “In contrast, older women present less aggressive features but have several comorbidities that, when associated with advanced stages, might contribute negatively to survival [13], although this was not shown in our data.”

• Ln270-273: The authors speculate that poor survival in those treated within 60 days may be due to advanced stage and/or triple negative tumours. But if the multivariate (if they are adjusted) results indicate that time to treatment is still a predictor when adjusted for stage and triple negatives, then this does not fully explain why survival is poorer when treated within 60 days.

R. We looked into our date and noticed that the calculation was correct; then, we rewrote to “An intriguing feature resulting from our estimates was time to treatment ≤ 60 days to be taken as a prognostic factor. After reanalyzing our data, we assumed it was due to factors such as younger age and more advanced and triple-negative tumors; however, it remained an independent prognostic factor after multivariate analysis. Whether improvement in health care would play any role in modifying it can be questioned. The difficulty in accessing diagnosis and treatment in low- and middle-income countries has to be overcome [30] because better cancer survival is a consequence of early diagnosis and timely treatment [31].”

• LN283 – Does standardise analysis refer to multivariate analysis? The latter would be the preferred term. Standardisation is something very different.

R. We corrected that.

• There is a great focus in the Discussion and Results on the univariate analyses, which are not as important in a study that is looking to identify independent predictors of survival. It would be recommended to focus mainly on the multivariate results in response to the research question.

R. We rewrote discussion to comply with that.

• Ln322-327: Discussion of missing data is appropriate however, the extent of which has not been quantified in results. It is important to do so.

R. We quantified in results and added the lines in discussion:

“We are cautious in analyzing the high percentage of missing information, mainly in staging, which might have influenced the results. The group with missing data on staging was presented as an independent factor in the multivariate analysis, as Basílio (2011) [28] and Brito, Portela and Vasconcellos (2009) [45] confirmed. Ayala et al. (2019) warned that failure to register this information, especially in the Breast Cancer Information System (SISMAMA), would compromise data monitoring [13]”;

“In addition, missing data (approximately 10%) were shown to be an independent prognostic factor in multivariate analysis – in fact, also denoting a confounding factor for survival. Some studies report that a lack of this information may be associated with difficulty in accessing adequate diagnosis and treatment and may be correlated with social status [33,35]”;

“Some limitations of this research should be considered, such as the use of retrospective secondary data without controls for completeness of information as found in medical records and a lack of standardization of the information in pathological reports. They might have interfered with the accuracy of the presented results.”

• Ln 335 indicates that results were stratified, I have seen no evidence of this.

R. Sorry, that was a mistake. We replaced for “Despite these limitations, it is important to note that the present study estimated the survival probabilities of cancer patients in a northeastern Brazilian state, separated by prognostic factors”.

We thank the reviewers for consideration taken with our manuscript.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

1 Jul 2021

PONE-D-20-36356R1

Breast cancer survival analysis based on prognostic factors in a northeastern Brazilian state: A retrospective cohort study

PLOS ONE

Dear Dr. Marques,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please provide additional description and discussion of  testing of the proportional hazards assumption for the Cox (proportional hazards) model. Please also proof-read and correct English grammatical errors.

Please submit your revised manuscript by Aug 15 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Nancy Lan Guo, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

Reviewer #2: Thank you for addressing my earlier comments. These have been done adequately although the testing of the proportional hazards assumption for the cox (proportional hazards) model needs to be undertaken and/or described.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

2 Sep 2021

We had the final manuscript edited by Jeremy Allen, PhD, from Edanz (www.edanz.com/ac) and added acknowledgment.

Reviewer #1:

R. We thank you for all the comments addressed previously.

Reviewer #2: Thank you for addressing my earlier comments. These have been done adequately although the testing of the proportional hazards assumptions for the cox (proportional hazards) model needs to be undertaken and/or described.

R. We have described the proportional hazards assumption for the cox model and added S3 and S4 Figs in Supporting information. We also added the comments “To test the proportional hazards assumptions, we employed the method based on scaled Schoenfield residuals”, lines 120-121; and “Even though the Schoenfield test rejects the hypothesis of hazard proportionality, probably because of test sensitivity caused by sample size, Figs S3 and S4 show that hazards remain fairly constant throughout the follow-up period, lines 178-180.

We hope we have answered adequately and thank you for the comments.

14 Sep 2021

24 Nov 2021

Response to Reviewers

We had the final manuscript edited by Jeremy Allen, PhD, from Edanz (www.edanz.com/ac) and added acknowledgment.

Reviewer #1:

R. We thank you for all the comments addressed previously.

Reviewer #2: I am not convinced that the test of slope of schoenfeld residuals is too sensitive given the sample size (which isn't that large). Looking at the graphs of residuals, it appears that the top left graph may reflect non proportionality. Ideally, the authors would include a time dependent variable for that characterisitc. If this interaction term is signficant, that would be provide more evidence of non proportionality, and conveniently, address that at the same time.

R. After revision with the statistician, we agree with the reviewer’s opinion. So, we are including the lines “Even though the Schoenfield test rejects the hypothesis of hazard proportionality, Figs S3 and S4 show that hazards remain fairly constant throughout the follow-up period, except for the for the cancer-specific survival variables. Thus, to explain non proportionality, time-dependent variables were presented (Table 3)”; and also, are adding Table 3; as can be seen in the revised manuscript with track changes.

Submitted filename: Response to Reviewers - PONE-D-20-36356R3.docx

Click here for additional data file.

17 Jan 2022

Analysis of breast cancer survival in a northeastern Brazilian state based on prognostic factors: A retrospective cohort study

PONE-D-20-36356R3

Dear Dr. Manuscript,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Nancy Lan Guo, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: Yes: Dr Elizabeth Buckley

25 Jan 2022

PONE-D-20-36356R3

Analysis of breast cancer survival in a northeastern Brazilian state based on prognostic factors: A retrospective cohort study

Dear Dr. Marques:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Nancy Lan Guo

Academic Editor

PLOS ONE