Analysis of Prognosis Factors of Breast Cancer Patients thru Socio Economic Profile: An Application of Binary Logistic Regression

Background: The idea of Prognosis factor is based on the variables that can be used to assess the chance of recovery from a disease. It may also be defined as the prior knowledge about any disease before treatment. Method: In this paper, selective prognostic factors (Age, Node and Tumour size) are analysed by logistic regression in patients who are suffering from Breast cancer based on data collected from the Cachar Cancer Hospital and Research Centre, Silchar, Assam, India. The purpose of the research is to analyse the effect of the prognosis factors on the remission of breast cancer; separately for economically weaker as well as well to do patients. Results: The study claims that there are 50.1 percent and 65.8 percent chance of remission of cancer for patients of age above 50 in case of breast cancer with below the poverty line and above the poverty line respectively. The present study has considered the cutoff value of 2 cm as the determining prognostic factor in relation to tumour size. Thus, the chance of remission from cancer is 22.3 percent and 34.7 percent for below poverty line and above poverty line patients, respectively, if tumour size is greater than 2 cm. It also endeavours to ascertain that there are 10.9 percent and 18.1 percent chance of remission of cancer, if the disease has metastasized to regional lymph nodes, for below poverty line and above poverty line patients respectively. Conclusion: There is a significant difference between the two poverty lines (APL and BPL) in terms of node and tumour size of breast cancer. The increasing sizes of tumour and node have lesser chance to follow-up as well as poorer survival and has a significant difference for patients belonging to both the poverty levels. The prognosis factors have the significant impact on the remission of breast cancer and depends on the socio-economic status of the patients due to the different standard of living, tendency of early diagnosis and the awareness level of cancer disease.


Background
Prognosis factors are defined as the variables that can be used to assess the chance of recovery from a disease. It is also considered as the prior knowledge about any disease before treatment.
The concept is applied to the cancer patients to get an idea of how cancer will affect the body and how it will respond to the therapies. It is difficult for the common people to take decisions about treatment methods due to lack of knowledge and different socio-economic backgrounds.
Thus, the prior knowledge will benefit the common people to participate in clinical and health policy decisions through prognostic and economic evaluation of cancer treatments. Although many prognostic factors have been identified over the last few decades, affecting the survival outcomes for breast cancer, there are some that have been proven to be of definite significance through various statistical methods. These includetumour size, nodal status, distant metastasis, histologic grade, histologic type, mitotic figure counts, hormone receptor status like Estrogen Receptor (ER) and Progesterone Receptor (PR) positivity and age of the patient (Russo.et.al, 1987). The present study will focus on tumour size, nodal status and age in relation to survival outcomes for breast cancer. Though these factors are already proven to be of significant importance in various studies, the present study will bring out their importance so as to guide future treatment decisions. In addition, the economic status of the patients has been evaluated as a possible demographic factor affecting choice of treatment and the resulting survival outcomes.
It is a matter of serious concern that new cases of the breast cancer are growing up day by day all over the world (Ali.et.al, 2011). Thus, the effect of the prognosis factors on cancer should be studied scientifically so that the patients can get an idea of economic and health policy decisions during the ongoing treatment processes. A comparative study was performed by Kroman.et.al, in the year 2000, among the treated vs. not treated through adjuvant cytotoxic treatment in case of breast cancer. The study revealed that there is a negative prognostic effect of young age in women diagnosed with low risk disease who did not receive adjuvant cytotoxic treatment (Kroman.et.al, 2000). Another study estimated the variations of Out-of-Field Dose, that are associated with radiotherapy, for the different parameters like field size and depth of cancer using the Markus Ionization Chamber Detector (Abdelaal.et.al, 2020). Again, in the year 2018, a study concluded that the effects of blood pooling have the impact on the levels of radioactivity measured in cancer tissues (Yavari K, 2018). A research was conducted in the recent year and concluded that Saliva officinalis can potentially prevent breast cancer (Zare H,2019). treatment is expensive and the income losses are significant (Nair.et.al, 2014). It seems logical that economically weaker patients are not in the favour of early diagnosis of cancer due to the high cost of the treatments, thus results in poor chance of survival. Taking treatment at early stage of cancer can relief the pain as well prevent the cancer from metastasis (Sun.et.al, 2017).
The patients are not often informed about the cause, nature and cost of therapies/treatments. Similarly, patients mostly do not have any information on the nature of prognostic factors and this lack of information may lead to the more advanced stages of cancer at diagnosis (Caplan L, 2014). The 'Stage' of the cancer is an important risk factor as due to longer time between the onset of cancer symptoms and the patient's presentation to health care, leading to later-stage diagnoses and therefore less eligibility for potentially curative treatment (Walter.et.al, 2015).
One study claims that the efforts to promote early detection should be continued in fighting with breast cancer as the primary prevention of breast cancer is still not available (Caplan L, 2014). Therefore, serious research is needed on the prognosis factors of breast cancer, which might help the patient/family members and the clinics/hospitals to take optimal decisions for cancer treatments. Although, different researchers have successfully brought up the importance of prognosis factors for different sites of cancer, but there is still a lack of studies on the relationship between the prognosis factors and the socio-economic status of the patients. Thus, keeping these points in mind, the present research is structured to get a conclusion of how the prognosis factors effect on the chance of remission of breast cancer; separately for economically weaker as well as economically well-off patients. The findings of the present research work will benefit the society for better treatment of cancer.

Objective of the study
The paper is designed keeping in view the following objective:  To study the effect of prognosis factors on the chance of remission of breast cancer; separately for economically weaker and economically well-off patients.

Methods
The data used for the present study is secondary in nature collected from the Cachar Cancer Hospital and Research Centre, Silchar, Assam, India. The dataset is stratified on the basis of the economic condition of the patients (BPLbelow poverty line and APLabove poverty line).
The classification is as the patients included in this study belong to different economic backgrounds. Thus, the patients, who have the BPL card 1 are considered as Below Poverty Line category and those without the card are considered to be an Above Poverty Line category. The following table gives an overview of the dataset. In this study, data were collected from a total of 400 female patients, half of which belongs to BPL group and the other half to APL group. There seems to be the difference between the two poverty lines in terms of size of the tumour as well as node status and thus z test for two sample proportion has been performed. It is appropriate for the comparison between the groups in proportion (Kleinbaum and Kupper, 1978) and the test statistics can be computed as,  Again, the difference between the two poverty lines (APL and BPL) has been tested in terms of the sizes of the tumour, the node and the follow-up period of the breast cancer patients. We are interested to observe the progressing sizes of nodal and tumour have any impact on the followup period of the breast cancer patients in both the groups (APL and BPL) separately. Thus, the multiple regression technique is applied and the models can be written as, Before including the independent variables (Node size and Tumour size) in the regression models, one important point should keep in mind that some independent variables that can be included in the models may play a redundant role, which could direct effect on the models and thus cannot be considered as reliable. Thus, Variance Inflation Factor (VIF) technique is applied through Multicolinearity Analysis.
The z test statistic is also computed to test the statistical difference between the regression equations in terms of node and tumour size and it should be noted the 'test of normality' must be performed before the z test (Kleinbaum and Kupper, 1978). The statistics can be computed as, = the difference of standard error between the regression coefficients associated with node size of APL group and BPL group.  In the present study, it will be convenient if we write our binary logistic regression model as follows: The following flowchart gives a better understanding of the Binary Logistic Regression model.

Results
At the outset, z test for two sample proportion has been performed to test the difference between the two poverty lines (APL and BPL) in terms of tumour size and node status of breast cancer.
The following table provides the results of z test for proportion of both the poverty lines separately.   We compute z test statistic to test the significant difference between the regression equations in terms of node and tumour size and it should be noted the 'test of normality' must be performed before the z test. The following table provides the result of test of normality.

Breast Cancer
Node size 1.056 Tumour size 1.031   We have obtained from Table 7 that in case of breast cancer with below poverty line category, there is a 50.1 percent chance of remission from the cancer if the patient age is above 50, which is better in comparison to that in patient age below 50. Again, if the cancer has already metastasized to axillary lymph nodes, the chance of remission from the cancer is only 10.9 percent, which is worse as compared to the cancer that has not spread to the lymph nodes. The patients with tumour size greater than 2 cm have a 22.3 percent chance of remission from cancer after the treatment, which is worse in comparison to the patients with the size of tumours less than 2 cm. Similarly, from Table 8, it is obtained that there is 65.8 percent more chance of remission from the cancer if the patient age is above 50 in the comparison to the patient of age below 50. Again, if the cancer has already metastasized to axillary lymph nodes, the chance of remission from the cancer is only 18.1 percent as compared to the cancer that has not metastasized to regional lymph nodes. The patients with tumour size greater than 2 cm have a 34.7 percent chance of remission of cancer in the comparison to the patients with the size of tumours less than 2 cm in the case of breast cancer with above poverty line category.

Discussion
Preclusion of cancer is one of the most significant public health challenges of the 21 st century (Ali et. al, 2011). Further, new cases of breast cancer show an ever increasing incidence all over the world. Several prognostic factors have been identified which affect the outcomes of disease and treatment. Thus, the present study is mainly focused on the chance of remission of cancer in relation to three prognostic factors viz. Age, Node and Tumour size for the patients that are suffering from breast cancer.
The age of the patient is a well-defined prognosis factor for local recurrence. It is well established that patient age greater than 35 or 40 is associated with an increased frequency of local recurrence due to presence of various adverse pathologic features, such as lymph vascular invasion, grade 3 histology, absence of Estrogen Receptor (ER) and Progesterone Receptor (PR), presence of HER2 and presence of extensive intra-ductal component (Kollias.et.al, 1997).
Our study finds that there are 50.1 percent and 65.8 percent chance of remission of cancer for patients of age above 50 in case of breast cancer with below the poverty line and above the poverty line respectively. Thus, in agreement with previous studies, we can consider the higher age as a good prognostic factor and younger age as a poor prognostic factor in breast cancer.
Tumour size is considered to be the best measure of tumour behaviour in breast cancer. Patients with a primary tumour size of less than 1 cm exhibit a frequency of only 10 percent to 20 percent of nodal metastasis, such that the 10-year disease-free survival rate is about 90 percent (Carter.et.al, 1989). Since the American Joint Committee on Cancer (AJCC) has described tumour size less than or equal to 2 cm as T1, the present study has focused on this cutoff value of 2 cm as the determining prognostic factor in relation to tumour size. The chance of remission from cancer is 22.3 percent and 34.7 percent for BPL and APL patients, respectively, if tumour size is greater than 2 cm.
Axillary lymph node status has been described as the second most important prognostic factor in relation to disease-free survival, as well as overall survival in breast cancer. 70 percent of node positive patients are likely to develop a recurrence compared to only 20 percent to 30 percent of node-negative patients (Veronesi.et.al, 1993). Patients with 4 or more numbers of involved nodes have a worse prognosis when compared to those with less than 4 nodes (Fisher.et.al, 1993). It is found from our study that there are 10.9 percent and 18.1 percent chance of remission of cancer, if the disease has metastasized to regional lymph nodes, with below the poverty line and above poverty line respectively. Thus, we conclude that the parameters viz. Tumour size and Node status, which measure the stage of the cancer patients is significant prognostic factors that help in predicting tumour behaviour and survival outcomes.
Looking at the global scenario and studying the cancer incidence related spatial data, economic conditions of the patients are also a matter of serious concern (Nair.et.al, 2014).

Conclusion
The present study has successfully brought out the relationship between specific prognostic factors and survival outcomes in breast cancer with respect to the different socio-economic status of the patients. There is a significant difference between the two poverty lines (APL and BPL) in terms of node and tumour size of breast cancer. The increasing sizes of tumour and node have lesser chance to follow-up as well as poorer survival and has a significant difference for both the poverty lines. The application of logistic regression model to the selected prognostic factors has revealed that a tumour size greater than 2 cm denotes a dismal prognosis, as does the presence of axillary nodal involvement. Also, patient's age less than 50 years is associated with a worse prognosis and poor overall survival. The prognosis factors have the significant impact on the remission of breast cancer and it vary from socio-economic status of the patients due to the different standard of living, tendency of early diagnosis and the awareness level of cancer disease.
We expect that the present study will pave ways for further study on the topics and would be beneficial for the researchers in the field. Further research would be challenging to study the health financing methods of different socio-economic groups and suggest policies related to tailor made insurance and expenditure management for the cancer patients. Additional challenge might be to decompose the cost component (like doctors' fees, therapy related cost, etc.) according to the incidence of cancer and study the pattern over time. The Institutional Ethics Committee (IEC) took the decision that the project falls under the exempted category as no participation of Human volunteer is directly involved and data will be collected from secondary sources.

List of Abbreviations
The administration of the hospital (Cachar Cancer Hospital and Research Centre) had granted the permission to access and use the medical records for this particular study. The ethical committee decision report and permission letter of the hospital have been attached in the supplementary materials section through online process.

Consent for publication
"Not applicable"

Availability of data and material
All data generated or analysed during this study are included in this published article [and its supplementary information files].