Journal of Cancer Research and Therapeutics, Vol. 3, No. 4, October-December, 2007, pp. 218-224
Postmastectomy radiation and survival in patients with breast cancer
Yadav BS, Sharma SC, Singh R, Singh G, Kumar V
Department of Radiotherapy, Postgraduate Institute of Medical Education and Research, Chandigarh - 160 012
Code Number: cr07054
AbstractPurpose: To analyze the impact of postmastectomy radiotherapy on locoregional control and overall survival in patients with carcinoma breast.
Materials and Methods: Between 1995 and 2000, 688 patients of carcinoma breast were analyzed. Out of these, 608 received postmastectomy radiotherapy and 80 patients were not given any radiation therapy. At a median follow-up of 67 months, the outcomes studied were locoregional recurrence (LRR), distant metastases, disease-free survival (DFS) and overall survival (OS) using univariate and multivariate analyses.
Results: The frequency of LRR with or without distant metastases was 8.5%, and distant metastases was seen in 18.7% of patients. On univariate analyses, factors affecting LRR were age <40 years (0.019), tumor stage ( P = 0.001 ), grade ( P = 0.027 ), pathological nodal status ( P < 0.0001 ), deep resection plane (0.041), ER/PR status ( P = 0.032 ) and postmastectomy radiation therapy (PMRT) ( P < 0.0001 ). DFS rate was 69% at 5 years. Factors affecting distant metastases were age <40 years (0.005), tumor stage ( P < 0.0001 ), grade ( P = 0.0007 ), pathological nodal status ( P < 0.0001 ), extra capsular extension (ECE) ( P = 0.002 ), hormonal therapy ( P < 0.0001 ) and PMRT ( P < 0.0001 ). The OS rate was 81% at 5 years. Factors affecting OS were tumor stage ( P < 0.0001 ), grade ( P = 0.0001 ), pathological nodal status ( P < 0.0001 ), ECE ( P = 0.002 ) ER/PR status ( P = 0.008 ), hormonal therapy ( P = 0.001 ) and PMRT ( P = 0.004 ). On multivariate analysis, factors affecting LRR were age ( P = 0.001 ), tumor stage ( P = 0.021 ), deep resection plane (0.003), ECE ( P = 0.022 ) and PMRT ( P = 0.047 ). Factors affecting distant metastases were menopause ( P = 0.044 ), grade ( P = 0.012 ), ECE ( P = 0.017 ) and PMRT ( P = 0.012 ). Factors affecting OS were menopausal status ( P = 0.017 ), tumor stage ( P = 0.029 ), pathological nodal status ( P = 0.011 ) and PMRT ( P = 0.002 ).
Conclusion: PMRT improves LRR as well as OS in patients with carcinoma breast. Other factors of prognostic importance were menopausal status, tumor stage and pathological nodal status.
Keywords: Breast cancer, postmastectomy radiation therapy, prognostic factors
Despite years of clinical research, there are still many unanswered questions regarding postmastectomy radiation therapy (PMRT). Trials conducted during 1960s and 1970s showed significant reduction in the rates of locoregional recurrence (LRR) but no improvement in overall survival (OS). , This is due to micrometastases that affect OS regardless of locoregional therapy. Thus, the use of adjuvant radiotherapy declined. However, randomized trials demonstrated that PMRT improves survival in women with node-positive breast cancer, who also received adjuvant systemic therapy. Hence, when systemic therapy is given for a micrometastatic disease, locoregional radiation may further improve OS by reducing the local disease burden, which is also considered to be the origin of distant metastases. Three randomized trials investigating PMRT in patients treated with systemic therapy demonstrated a reduction in LRR from approximately 30-10%. ,, The potential for LRR, in conjunction with systemic therapy, to reduce metastases and deaths from breast cancer has long been recognized. The meta-analysis published by Early Breast Cancer Trialists′ Collaborative Group (EBCTCG) showed that radiation treatment improved 20-year cause-specific survival rates by 13% ( P = 0.0001 ), despite the fact that most of the trials included in the analyses used obsolete radiation techniques and did not include systemic treatment. ,
All these data have been derived from countries where patients enjoy good general condition, treatment facilities and less financial problems as treatment expenses are borne by insurance companies. However, in developing countries like India, most of the patients present in advanced stage, poor general health and belong to low socioeconomic status and have to travel long distances for treatment and follow-up. Moreover, there is scarcity of randomized data and most of the data come from retrospective analysis. In our research, we analyzed postmastectomy patients who had received radiation along with systemic treatment for various prognostic factors.
Materials and Methods
Patient, tumor and treatment characteristics
The records of patients of carcinoma breast registered in the Department of Radiotherapy between January 1995 and December 2000 were reviewed. Patients registered after mastectomies were retrospectively analyzed. Prognostic factors for locoregional failure, distant failure, disease-free survival (DFS) and OS were determined using univariate and multivariate analyses. A total of 688 patients with carcinoma of the breast were analyzed. Out of these, 608 received PMRT. The patients′characteristics were as shown in [Table - 1]. Pathologic diagnosis was obtained from fine needle aspiration cytology or core biopsy. Ninety-four percent of the tumors were infiltrating ductal carcinoma, 4% were lobular carcinoma and 2% were medullary carcinoma. Total mastectomy with axillary clearance was done in 50% patients, simple mastectomy was done in 30% patients and 20% of the patients underwent modified radical mastectomy with axillary clearance.
Postmastectomy radiation therapy was delivered to 88.5% patients with a dose 35Gy/40Gy/15#/3 weeks. We followed Manchester shorter fractionation schedule  35Gy to chest wall and 40Gy to axilla and supraclavicular region. Axillary node-positive patients and some patients with axillary lymph node-negative T3 tumors, or when axillary status was unknown or there was incomplete axillary dissection, were given radiation to axilla and supraclavicular region. Anterior photon field was used to deliver radiation to supraclavicular, infraclavicular, axillary and internal mammary nodes. Two tangential opposed fields were used to irradiate the chest wall. The borders for chest wall radiotherapy were the anterior midline (medial), the mid-axillary line (laterally), the inframammary fold (inferior) and the bottom of the head of the clavicle (superior). Supraclavicular, infraclavicular and high axillary lymph nodes were treated with an anterior photon field; the inferior portion of this field was matched to the superior edge of the tangent fields. The head of the humerus was also shielded from radiation beam. Internal mammary nodes were irradiated with a separate 12 x 6 cm field. Dose was 40Gy/15#/3 weeks. Treatment was given using 60 Co units. Doses were prescribed at the midpoint of the central axis.
Two adjuvant chemotherapy regimes used were FAC (5-Fu 600 mg/m 2 , adriamycin 50 mg/m 2 and cyclophosphamide 600 mg/m 2 ) in 102 (14%) and CMF (cyclophosphamide 600 mg/m 2 , methotrexate 40 mg/m 2 and 5-Fu 600 mg/m 2 ) in 278 (41%) patients. Tamoxifen was given to 401 (58%) patients irrespective of ER/PR status; dose was 20 mg daily for 5 years. Patients were followed at regular intervals and further investigated only if they had any symptoms or evidence of recurrent or metastatic disease.
End points and statistical evaluation
Locoregional control and survival curves were generated using the Kaplan-Meier method. Locoregional control was defined as any recurrence in the skin or soft tissue over chest wall or a recurrence in the regional lymphatic sites (axilla, internal mammary, infraclavicular and supraclavicular). Univariate analysis was done using the Log rank test. Multivariate analysis was done using the Cox proportional hazards model for age, menopausal status, tumor stage, histological grade, nodal status, deep resection plane, ECE, LVI, ER/PR status, radiotherapy given, chemotherapy and hormones. The outcomes studied were LRR, distant failure, DFS and OS using univariate and multivariate analyses.
All statistical tests were two-tailed and differences were considered to be statistically significant if P ≤ 0.05. Statistical analysis was performed using SPSS software version 12.0 (Statistical Package of Social Science, Chicago).
Patient, tumor and treatment characteristics
Between January 1995 and December 2000, 1497 patients with breast cancer were registered. Out of these, 688 patients underwent mastectomy. Patient characteristics are as listed in [Table - 1]. The median age at presentation was 46 years (range 20-75 years). At a median follow-up of 67 months, 507 (74%) patients were alive without any evidence of disease, 104 (15%) were alive with disease and 77 (11%) were dead. The frequency of LRR with or without distant metastases was 8.5% and distant metastases was seen in 18.7% of patients. The OS rate was 81% at 5 years [Figure - 1]. The recurrence rate in the skin or soft tissue over ipsilateral chest wall or a recurrence in the regional lymphatic sites or distant sites at any time during follow-up was documented in 174 patients (18.7%), for a DFS rate of 69% at 5 years [Figure - 2]. LRR was seen in 59 patients and distant metastases in 115 patients. The site of recurrence was chest wall in 36 patients, regional lymph nodes in 23 patients and both local recurrence and distant metastases was seen in 14 patients. With 59 locoregional recurrences with or without distant metastases, the 5-year local control was 94.4%. The most common site of distant metastasis was bones (5.8%), followed by lungs (5.5%) and liver (2%).
Predictors of locoregional recurrence after mastectomy
On univariate analyses, factors affecting LRR were age < 40 years ( P = 0.019 ), tumor stage ( P = 0.001 ), grade ( P = 0.027 ), pathological nodal status ( P < 0.0001 ), deep resection plane ( P = 0.041 ), ER/PR status ( P = 0.032 ) and PMRT ( P < 0.0001 ). Strongest association was seen with the use of radiotherapy ( P < 0.0001 ) as shown in [Table - 2]. Of the various prognostic factors studied, PMRT was associated with all adverse outcomes. LRR in those patients who received PMRT was 5.26% as compared to 31.26% in those who had not received PMRT. Factors affecting distant metastases were age < 40 years ( P = 0.005 ), tumor stage ( P < 0.0001 ), grade ( P = 0.0007 ), pathological nodal status ( P < 0.0001 ), ECE ( P = 0.002 ), hormonal therapy ( P < 0.0001 ) and PMRT ( P < 0.0001 ). Distant metastases in those patients who received PMRT were 15.3% as compared to 26.25% in those who had not received PMRT. Factors affecting OS were tumor stage ( P < 0.0001 ), pathological nodal status ( P < 0.0001 ), grade ( P < 0.0001 ), ECE ( P = 0.002 ), ER/PR status ( P = 0.007 ), PMRT ( P < 0.0001 ) and adjuvant hormonal therapy ( P < 0.0001 ); here too PMRT emerged as the strongest factor affecting OS [Table - 2]. On multivariate analysis, factors affecting LRR were younger age ( P < 0.001 ), tumor stage ( P = 0.030 ), DRP ( P = 0.003 ), ECE ( P = 0.047 ) and PMRT ( P = 0.018 ), and for distant relapse were menopausal status ( P = 0.044 ), grade ( P = 0.012 ), ECE ( P = 0.017 ) and PMRT ( P = 0.012 ). Factors affecting OS were menopausal status ( P = 0.018 ), tumor stage ( P = 0.043 ), pathological nodal status ( P = 0.011 ) and PMRT ( P = 0.002 ).
Multivariate analysis was done for all points using Cox proportional hazard model [Table - 3]. For LRR, younger age [hazard ratio (HR) 4.16; 95% confidence interval (CI), 1.884-9.196], tumor stage (HR 0.353; CI 0.138-0.9), DRP (HR 1.474; CI 1.915-10.979) and PMRT (HR 2.060; CI 1.107-3.833) were significant independent adverse factors. For distant metastases, premenopausal status (HR 2.692; CI 1.029-7.044), grade (HR 4.158; CI 1.366-12.650), ECE (HR 3.572; CI 1.257-10.151) and PMRT (HR 4.203; CI 1.348-13.102) were the adverse independent prognostic factors. The only factor affecting RFS was LVI (HR 2.778; CI 1.079-7.154). For OS, premenopausal status (HR 2.619; CI 1.191-5.759), T3/T4 tumor stage (HR 0.3; CI 0.098-0.882), pathological nodal status (HR 0.308; CI 0.124-0.763) and PMRT (HR 4.78; CI 1.694-9.815) were significant independent adverse prognostic factors.
In this analysis, it was also seen that about 64% of all relapses occurred within 2 years, 74% within 3 years, 86% within 5 years and 95% within 10 years. So, 10 years′ follow-up is adequate for relapses to occur, and only 5% of the overall relapses occurred after that. For LRR, 68% of recurrences occur within 2 years, 75% within 3 years, 88% within 5 years and 94% at 10 years. So, 10 years′ follow-up is adequate for LRR to occur, and only 6% of the LRR occurred after that.
WHO acute grade III skin toxicity was seen in 7.1% of patients after PMRT. Acute pneumonitis was seen in 3% of patients and grade-1 and -2 oesophagitis were seen in 10% of patients. Mild arm edema developed in 8.7% of patients after PMRT. Cancer in the opposite breast developed in 2.7% patients; other second cancers seen were ovarian and cervix in four patients and basal cell carcinoma, multiple myeloma and carcinoma of the vallecula each in one patient. The overall incidence of second cancers was 4.36%. None of our patients developed any cardiac complication or radiation-induced malignancy.
This analysis provides evaluation of the impact of PMRT for breast cancer in developing world where general condition and physique of the patient is totally different from that in the developed countries. The treatment setup and facilities are inadequate. Ours is a tertiary level institute situated in northern part of India where many patients are referred to from outside. Patients are operated at different centers, and adequacy of surgery, including level of axillary dissection, varies depending on the surgeon′s expertise. Many a times the surgical and histopathological details are incomplete, leading to deficient data. Equally, women included in the analyses were young and old, premenopausal and postmenopausal and had contrast life styles (very low smoking rates, minimal cardiac toxicity, lower socioeconomic strata, rural population, low fat intake, many children, prolonged breastfeeding habits, etc.). Chemotherapy used in most of the patients was CMF (90%) and may be in suboptimal doses after dose reduction due to toxicity, particularly hematological and gastrointestinal (vomiting and mucositis), low body surface area and bulky disease. In this research, we have tried to bring forward our institutional data from retrospective analyses for various prognostic factors in postmastectomy patients with breast cancer.
On multivariate analysis, factors affecting local recurrence were younger age (< 40 years), T3T4 tumor, DRP, ECE and PMRT. Younger age has been reported to be associated with increased incidence of LRR in many studies, particularly in patients with conservative breast surgery.  It could be due to biologic aggressiveness of the tumor in young age related to the hormonal levels and other factors such as higher grade and larger tumor size at presentation. Advanced clinical stage (T3T4) at presentation was an independent predictor for LRR in our analysis. Positive DRP was also associated with increased LRR. LRR in patients with positive DRP was 5.87% as compared to 1.7% in patients with negative DRP.
Our results indicated that ECE is a marker of aggressive disease with high propensity for LRR. ECE is important from the point of view that it represents the bulk of the disease; once the disease has extended outside the capsule it can cause local recurrence as well as throw distant metastases easily through vascular and lymphatic channels. Local recurrence can be due to spillage while removing nodes because nodal dissection does not mean the disease outside the capsule is also removed. ECE seems to be a more reliable prognostic factor than nodal positivity because nodal disease is removed by surgery, or if microscopic disease is still present in a node after surgery, it can be taken care of by radiation. ECE will be present only when nodes are involved and it signifies more bulk of disease as compared to only nodal positivity. Hence, patients with ECE should be considered candidates for axillary and supraclavicular field irradiation. LRR in patients with ECE was 8% as compared to 4% in those who did not have. Our results are comparable to those from other institutions, ,,, which report a 4-7% risk of axillary recurrence after surgery and adjuvant therapy for patients with ECE.
Two recently published retrospective series explored various factors of prognostic importance for LRR in more than a 1000 women. ,, In a study by Recht et al.,  they found pathological nodal status, tumor size and ER/PR status affecting LRR on multivariate analysis. Katz et al. , performed a multivariate Cox regression analysis for various prognostic factors and found that tumor size> 5 cm, presence of four or more nodes, close or positive surgical margins or clinically or grossly pathological multicentric disease and presence of LVI, was an independent predictor of LRR. It is difficult to compare the frequencies of LRR among different studies because the selection and treatment of patients vary as well as definition of LRR. In some studies, only local recurrences (i.e. chest wall) are scored as LRR; supraclavicular nodes are sometimes counted as distant failures and sometimes as LRR. In our study as well as those of Recht et al.  and Katz et al.,  relapses in the chest wall and in the axilla, supraclavicular, infraclavicular fossae and internal mammary nodes were taken as LRR. These can affect the statistics. Consistent with most previous reports, our study reveals that the chest wall and supraclavicular fossae were the most common sites for LRR.
Impact on survival
Factors affecting the OS were menopausal status, tumor stage, grade, ECE, ER/PR status, hormonal therapy and use of PMRT. OS at 5 years was 92% in patients with T1T2 tumors as compared to 68% in T3T4 tumors; OS at 5 years was 55% when ECE was present as compared to 74% when it was absent. DFS at 5 years was 84% in patients treated with PMRT as compared to 49% in those without PMRT, and OS at 5 years was 89% vs. 53% in the two groups, respectively [Figure - 3]. The significant difference seen in OS between the groups treated with and without PMRT indicates that second-line treatment cannot compensate for inadequate primary therapy even with high-dose combination chemotherapy. Analysis at 20 years from the British Columbia randomized trial has also shown significant benefit of PMRT for DFS and OS. 
This analysis also demonstrates that PMRT reduces the rates of local and distant relapses and chances of death from breast cancer as compared to other trials where radiotherapy was not given. In an analysis of 2016 patients treated with mastectomy and chemotherapy on Eastern Cooperative Group Trials, Recht et al. reported 10-year LRR of 28.7% for patients with four or more lymph nodes and 12.9% for patients with one to three lymph nodes.  Similarly, Katz et al. showed a 10-year LRR of 25-34% for patients with four or more lymph nodes and 13% 10-year LRR for patients with one to three lymph nodes.  In our analysis, 10-year projected LRR was 15% for patients with four or more lymph nodes and 6% for patients with one to three lymph nodes, that is 10% absolute improvement in the 10-year LRR for patients with four or more lymph nodes and 7% for patients with one to three lymph nodes. PMRT also improved OS. This is contrary to the EBCTCG meta-analysis,  where radiotherapy was found to decrease the number of local recurrences significantly, but was found to have no impact on OS. However, a re-analysis of the 2000 EBCTCG overview showed a statistically significant reduction in all-cause mortality associated with radiation therapy among lymph node-positive patients and among trials started after 1975. 
Our results are consistent with those of the Danish pre- and postmenopausal trials, , where survival benefit of radiation therapy was confirmed in patients treated with adjuvant chemotherapy and with the meta-analysis of all adjuvant trials where radiation therapy given in conjunction with chemotherapy was compared with the same chemotherapy alone, confirming statistically significant reduction of overall mortality associated with radiation therapy.  There was no long-term mortality or significant morbidity caused by PMRT or its late side effects. It was also seen that addition of chemotherapy to PMRT favors decrease in distant metastases but does not improve OS.
Impact on distant metastases
The long-term reduction in the rate of distant metastases associated with radiation therapy seen in this analysis showed that locoregional microscopic disease that survives after mastectomy and chemotherapy is the cause of subsequent distant metastases and locoregional recurrences. Distant relapse rate was 15% in patients who were given radiation as compared to 26% in those who were not. It means that radiation therapy can effectively eradicate the source of metastases in 30-40% of patients who would otherwise be at risk of distant metastases. When patients with pN1 nodes were compared with pN2 and pN3 nodes, LRR was similar (3-4%) but distant relapse rates were 13%, 23% and 20%, respectively, which means PMRT reduces LRR along with distant metastases in these patients too. The residual disease after mastectomy, which may be microscopic, can still become resistant to chemotherapy and can be eradicated with locoregional radiotherapy leading to better outcome.
Shorter fractionation schedules of 35Gy/40Gy/15#/3 weeks to chest wall and drainage area respectively have several advantages. It is as effective, without added morbidity, as compared to longer treatment schedules. A shorter fractionation schedule minimizes the time the patient needs to spend away from home, family and job and also minimizes costs associated with travel, lodging and time away from work. Considering the high incidence of breast cancer among females, shorter fractionation schedule would also ease the demand on radiation treatment sources, which are already lacking in our country. It will also save the health care budget in the long run and reduce the long-waiting list of patients for radiotherapy. A shorter radiation fractionation schedule can also be delivered in the middle of a chemotherapy course without any delay between chemotherapy cycles. These advantages are clinically proved in randomized a trial against conventional fractionation.  This trial showed equal survival, local control, toxicity and cosmetic outcomes at 5 years in the two arms. Radiation therapy to supraclavicular fossae/axillary apex lead to 8.7% of mild lymphedema as compared to that of 24% reported by Halverson et al.  after conservative breast surgery.
ConclusionIn this multivariate analysis of a large cohort of Indian women treated with mastectomy and radiotherapy, the prognostic significance of various factors was confirmed. PMRT emerged as the most important prognostic factors for all outcomes. Other factors were younger age, clinically advanced tumor stage, pathological nodal stage and ECE. PMRT still plays an important role in women with breast cancer. Although benefits of adjuvant radiotherapy may be measured by both locoregional and systemic recurrences, the reduction of systemic recurrences associated with radiation therapy is of substantially higher importance, because systemic recurrence is a surrogate for eventual breast cancer mortality. Our results also indicate that in situations where residual disease remains, adjuvant chemotherapy alone in high-risk breast cancer patients is suboptimal, and that the addition of locoregional radiation therapy is important to achieve highest cure rate. ECE should be considered an indication for axillary and supraclavicular fossae irradiation because it is a hallmark for local as well as distant relapse. Furthermore, our results were obtained with shorter fractionation schedule of radiation therapy that may be clinically equal to radiation therapy schedules of 25 or more fractions presently used with important cost saving. Long-term toxicity of radiation therapy is minimal if delivered accurately.
Copyright 2007 - Journal of Cancer Research and Therapeutics
The following images related to this document are available:
Photo images[cr07054t2.jpg] [cr07054t3.jpg] [cr07054t1.jpg] [cr07054f3.jpg] [cr07054f2.jpg] [cr07054f1.jpg]