Clinico-epidemiological profile of breast cancer patients and the retrospective application of Gail model 2: An Indian perspective

Abstract

AIM:

To study the clinical and epidemiological profile of patients of breast cancer presenting at our center at New Delhi, India and to evaluate the applicability of Gail model 2 as a means of measuring 5-year and lifetime risk in our already diagnosed cases of breast cancer.

METHODS:

This was a retrospective study conducted at Lady Hardinge Medical College Hospital in New Delhi, India, between January 2011 and July 2014. Two hundred and twenty two diagnosed cases of breast cancer were included. Information was collected retrospectively on a Performa from the medical record section and the Pathology department of the hospital.

The predicted five-year and lifetime risk was calculated using GM2 prediction model from the NCI’s breast cancer risk assessment tool website.

RESULTS AND CONCLUSIONS:

Breast cancer in India is a far more biologically aggressive disease than in the west with a widely different spectrum of presentation and behavior and late presentation in an advanced stage. The accepted risk factors routinely associated with breast cancer in western literature do not appear to be relevant in the Indian population. Accepted western models do not seem to apply in the Indian scenario.

Introduction

Breast cancer, the second most common cancer in the world is the most frequent cancer among women [1]. It is the most common cancer amongst women in USA accounting for 29% of newly diagnosed cancers [2]. India reports about 100,000 cases of breast cancer annually, with an estimated increased rate of 3% per year with an estimated 250,000 cases per annum by 2015 [3]. The death rates due to breast cancer are highest in India, surpassing China and USA and is a cause of concern [1].

Western studies show that breast cancer is increasingly being detected at an early stage, as mammographically suspicious lesions and presents most commonly in the fifth and sixth decades of life. The risk factors include early menarche, late menopause, nulliparity, lack of breast feeding, presence of previous breast biopsies showing atypia, and a family history of breast cancers. However, in India, locally advanced breast cancer (LABC) accounts for 50–70% of patients presenting for treatment [4]. In addition, the above risk factors have not been evaluated in detail in the Indian context.

Models like GAIL model (GM) 1 and 2, and the Clauss model incorporate some of these risk factors and have been used to assess the 5-year and lifetime risk of breast cancer for females. The GM is the most commonly used risk prediction model well studied and validated in various studies, mostly from the west [5,6].

This retrospective study was hence carried out to study the clinico-epidemiological profile of Indian patients of breast cancer presenting at our center, and to evaluate the applicability of Gail model 2 in our already diagnosed breast cancer cases.

Methods

This retrospective study, conducted at a university college hospital in northern India, between January 2011 and July 2014 included two hundred and twenty two diagnosed cases of breast cancer. Information was collected on a Performa from the medical record section and the Pathology department. Parameters studied included age and gender of the patient, ethnicity, presenting complaints (breast lump or nipple discharge), side and site of breast lump, age of menarche and menopause, age at first child birth, breast feeding, history of previous breast biopsy, TNM stage at presentation, histopathology, and the molecular subtyping. The predicted five-year and lifetime risk was calculated using GM2 prediction from the NCI’s breast cancer risk assessment tool website, accessed at http://www.cancer.gov/bcrisktool.

Observations

A total of 222 patients diagnosed with breast cancer were included of which, only one was male. The socio-demographic profile, clinical presentation and staging of the disease are shown in Table 1.

The majority of patients were in the age group of 35–54 years (68.91%, median age 45 years), with the youngest being 24 years old and oldest being 88 years old. 126 patients were from a rural background and 96 from an urban background, the ratio being 1.3:1.

The duration of presenting complaints ranged from 7 days to 24 months, with a median of 9 months. 52.73% of patients had complaints lasting more than 3 months.

210 (94.5%) patients presented with a breast lump, 143 (68.18%) on the right side, and 101 (45.45%) in the upper outer quadrant of the breast. One patient came with bilateral breast lumps. Four patients (4.53%) presented with a large fungating breast mass. Two (0.9%) patients presented with blood stained nipple discharge without any lump.

On evaluation of the stage and molecular profile, 64.73% patients were diagnosed in AJCC TNM (7th edition) stage III at presentation, and the most common molecular type was basal [46.41%, 103 patients] as shown in Table 1.

Histopathology was consistent with Infiltrating Ductal Carcinoma in 201 patients, Neuroendocrine carcinoma in five patients, apocrine carcinoma in three patients, papillary in four, Malignant Fibrous Histiocytoma/Squamous Cell Carcinoma/Squamous Cell Carcinoma in epidermoid cyst each in one patient each, LCIS/medullary carcinoma and DCIS each in two patients each respectively.

On risk factor evaluation, none of the patients were nullipara and only 4 (1.81%) patients had not breast fed their children for more than twelve months. Age at menarche ranged from 11–18 years, with a mean of 14 years and 90 (40.54%) patients had menarche before 12 years of age. 160 (72%) patients were pre-menopausal. 10 patients (4.50%) had a positive family history of breast cancer in single first degree relative, and 6 patients (2.72%) had a first full term pregnancy after the age of 30 years. Risk factors such as alcohol consumption and hormone replacement therapy were not present and none had any personal history of cancer or high dose radiation therapy to the chest in the past. Tobacco consumption was present in 110 patients in chewable form and 10 patients by smoking (Table 2).

On applying GM 2 to our already diagnosed breast cancer patients (Table 3), the estimated 5-year and lifetime risk calculated put 92% patients in a low 5-year risk and 86% patients in a low lifetime risk group for breast cancer. The risk could not be calculated for 8% patients because they fell in age group less than 35 years or more than 85 years for which the risk cannot be calculated using the presently validated model.

Discussion

An estimated 1.67 million new breast cancer cases were diagnosed in 2012 (25% of all cancers). It is the most common cancer in women with slightly more cases in less developed (883,000 cases) than in more developed (794,000) regions representing about 12% of all new cancer cases and 25% of all cancers in women [1].

In 2012, for the United States, 232,714 women were newly detected and 43,909 women died of breast cancer and for India, 144,937 women were newly detected and 70,218 women died of breast cancer with deaths more than any other country in the world (second: China – 47,984 deaths and third: US – 43,909 deaths [1]).

The estimated rate of breast cancer in India is 80 new cases per 100,000 population per year, and in Delhi, it is 146 new cases per 100,000 population per year [3]. Breast cancer is thus a major concern in India and its incidence is showing a rising trend.

India–USA comparison of demographics and presentation of disease

Breast cancer is predominantly a disease of females with female gender being the strongest non modifiable risk factor. The incidence of male breast carcinoma in our study was 0.45%, lower than 1.4% reported in the literature [7].

The majority of our patients were in the age group of 35–54 years (68.91%, median age 45 years) with youngest patient being 24 years and oldest being 88 years. Similar results were also observed by Suresh et al. in their study from Delhi in 2013 [8]. However, this is in contrast to reports from the US SEER data which show that 68.80% female breast cancer patients are in 45–74 years age group, with a median age of 61 years [9]. This significant difference in age distribution is shown in Fig. 1. The disease presentation more than a decade-and-a-half earlier than in the west is particularly alarming as early age at presentation of breast cancer is an independent adverse prognostic factor [12].

Premenopausal women formed 72% of the cohort in our study. A previous study from Delhi by Raina et al. in 2005 showed that these form around 50% of the cohort [10]. This highlights the increasing incidence of breast cancer in premenopausal woman in India. 126 patients in our study were from a rural background and 96 from an urban background, the ratio being 1.3:1. Other reports from India as well as United States show a higher incidence in the urban population [7].

The duration of presenting complaints ranged from 7 days to 24 months, with a median of 9 months. Breast lump was the presenting complaint in 210 patients (94.5%). Only two (0.9%) patients presented with blood stained nipple discharge without any lump. This is in contrast to developed nations where, due to the increased awareness amongst females regarding breast self examination and availability of screening programs, a shift toward the diagnosis of clinically occult and non palpable lesions has been noted [4]. This is reflected in the US NCDB data wherein 20% are in situ lesions and 39% are stage I disease at presentation in United States [11], whereas 64.73% of our cases presented in stage III. The stage wise comparison between our data and US data is shown in Fig. 2. Other reports from India have also shown that a majority of the patients (65.9%) present with LABC [4], which is a relatively uncommon presentation (5–20%) in economically developed countries [12]. The long duration of symptoms before diagnosis (median duration 9 months), and the advanced stage (stage III) at presentation is worrisome as it reflects social inhibitions, lack of health education and awareness, absence of regular breast self examination and screening modalities, and a reluctance to subject oneself to a clinical breast examination.

Our study, as well as other reports from India and the western world show that IDC is the most commonly encountered histopathology. Ductal carcinoma in situ (DCIS) accounts for over 20% of the breast carcinoma cases in the western world. However, in developing countries, a very low incidence of DCIS is seen. Primary neuro endocrine carcinoma of breast, malignant fibrous histiocytoma and squamous cell carcinoma are rare histological variants. In our studies these combined to less than 3% [13–15].

In recent years breast cancer has been classified based on its molecular characteristics by gene expression profiling. A similar classification using immunohistochemistry has been identified [16]. Breast cancers which do not express oestrogen, progesterone, or HER-2 neu receptors are triple negative breast carcinomas, found to be extremely aggressive with a poor prognosis [17]. As shown in Fig. 3, our patients predominantly exhibit a triple negative molecular phenotype [46.41%, 103 patients], in comparison to US SEER data which shows 40% of US patients to be Luminal A phenotype [ER+/PR+/Her 2−] with only 25% of US patients being of the triple negative phenotype [9]. The triple negative phenotype of our patients points to an aggressive disease with a poor prognosis.

Gail model 2 and risk factor evaluation in our patients.

Models have been developed to estimate the risk of breast cancer in western literature. Mitchell Gail developed a model in 1989 to assess the risk of breast cancer risk based on the results from the BCDDP – a large screening study that included 284,780 women who had been undergoing annual mammographic examination and included factors of age at menarche, number of breast biopsies, age at first live birth and number of first degree relatives with breast cancer and applied to an age group of 35–74 years [5,6]. Later, it was modified (GM2) by involving age of patient at presentation, race/ethnicity and age of patient at presentation given maximum risk factor. BRCA positivity, previous biopsy suggestive of LCIS/DCIS, history of breast cancer, history of radiation therapy for Hodgkin’s, are all exclusion for using this model. Also, the age group was expanded to 35–85 years. Other models have also been formulated, but most countries in the West use the Gail Model 2. A patient is said to have low 5 year risk and lifetime risk when the calculated value is less than 1.66% and between 10% and 15% respectively and high 5 year risk and lifetime risk when the calculated value is more than 1.66% and 15% respectively.

The risk factors, their associated relative risk as per the US SEER data, and their relevance in our data is shown in Table 3 which clearly shows that apart from age at menarche, none of the other risk factors of Gail model 2 are significant in our data on already diagnosed cancer patients and the other risk factors implicated in western literature don’t feature in our data. This is alarming as it shows is that the risk factors that we routinely study to be associated with breast cancer does not appear to be risk factors in our data. The fact is more clearly shown in Fig. 4, when the Gail model 2 is retrospectively applied to our already diagnosed breast cancer patients, the estimated 5 year and lifetime risk calculations show a ‘low 5-year risk’ and a ‘low lifetime risk’ for 92% and 86% of the patients respectively! Also, the risk could not be calculated for 8% of our patients for the reason stated above.

Breast cancer is a serious health problem in India. It is more common amongst low and middle socioeconomic groups. It presents more than 15 years earlier in India (median age 45 years) than in the west. The long duration of symptoms (median duration 9 months) and the late stage at presentation (stage III) reflects social inhibitions, lack of health education and awareness, absence of regular breast self examination and screening modalities, and a reluctance to subject oneself to clinical breast examination in contrast to the west where the majority of patients are either in situ lesions or Stage I disease. An overwhelming number of our patients have a poor prognosis triple negative molecular status (46.41%), compared to the USA where 40% of patients are of the good prognosis Luminal A phenotype [ER+, PR+, Her2neu−]. This indicates that a large number of our patients (for reasons not clear) have aggressive, poor prognostic biologic behavior.

Risk factors routinely associated with breast cancer in western literature do not appear to be relevant in the Indian population. In our study, only 7.7% of patients were over 65 years of age, none were nulliparous, only 2.72% had their first full term pregnancy after 30 years of age, only 1.8% of our patients had not breast fed for >1 year, only 15.31% patients have had previous breast biopsies, and only 4.5% had a positive family history in a first degree relative. None had a history of alcohol consumption, personal history of breast cancer, high dose radiation to chest, or hormone replacement therapy. The only probable significant factor was early menarche (<12 years of age) in 40.54% of patients. Application of Gail model 2 to our already diagnosed breast cancer patients categorized 92% as having ‘low 5-year risk’, and 86% having a ‘low lifetime risk’!! The Gail model 2 thus does not appear to be useful in our population.

Conclusions

Breast cancer in India is thus a far more biologically aggressive disease than in the west with a widely different spectrum of presentation and behavior. Compounding this is a late presentation in an advanced stage. This calls for aggressive public health awareness and education to emphasize regular breast self examination after the age of 30, and to consult a doctor immediately for any breast complaints. Further research is required to study the biologic behavior and risk factors in the Indian scenario, and to formulate risk assessment models and prevention strategies in our setup.