Barbara Fowble, MD

ABSTRACT: With the increased use of doxorubicin-basedchemotherapy, chemoendocrine therapy, and high-dose chemotherapy with autologousbone marrow transplantation or peripheral blood progenitor-cell reinfusion,the role of postmastectomy radiation in the treatment of stages II-IIIbreast cancer has been challenged. Despite these therapies, 20% to 30%of patients with four or more positive nodes, primary tumor size 5 cm ,or more or pectoral fascia involvement will develop an isolated locoregionalrecurrence. Postmastectomy radiation decreases the incidence of locoregionalrecurrence to 10% or less in these high-risk patients, and modificationsin technique can minimize long-term cardiac mortality. A recent meta-analysishas demonstrated a decrease in breast cancer mortality of approximately10% with postmastectomy radiation, which is similar to the effect of adjuvantsystemic therapy in axillary node-positive patients. Future studies areneeded to further define the subset of patients who will benefit from postmastectomyradiation and to more precisely quantitate this benefit. [ONCOLOGY 11(2):213-237,1997]

The rationale for postmastectomy adjuvant radiation is the preventionof recurrent disease in the chest wall and regional nodes following surgery.Theoretically, clinically occult disease at the operative site or in regionalnodes could serve as a source for the development of distant metastases.Therefore, eradication of this disease may have the potential for decreasingthe incidence of distant metastases and improving disease-free and overallsurvival. Although the relationship between locoregional control and distantmetastases has generated considerable debate, there is increasing evidencethat all patients with breast cancer do not have systemic disease fromthe inception,[1,2] and that there may be a causal relationship betweenlocal control and distant metastases.[3,4] 

Using a competing risk approach, Arriagada et al[4] recently analyzedthe results of the Stockholm trial, which randomized 960 women to premastectomyor postmastectomy radiation vs observation. This analysis recognized thecompeting risks for a first event, including locoregional failure, distantmetastases, contralateral breast cancer, non-breast cancer, second malignancy,and death due to other causes. The addition of radiation significantlydecreased the 15-year cumulative incidence of local failure as a firstevent (6% vs 26%; P less .0001) or any event (11% vs 36%; P less .0001).Radiation did not significantly reduce the incidence of distant metastasesas a first site of failure (29% vs 23%; P = .48), but it did significantlydecrease the overall 15-year cumulative incidence of distant metastases(37% vs 45%; P = .04). This benefit, however, was limited to the axillarynode-positive patients (54% cumulative incidence with radiation vs 72%without; P = .01). 

Following locoregional recurrence, 66% of the observation group developeddistant metastases, as compared with 82% of the radiation group. Therefore,radiation decreased the total incidence of distant metastases by decreasingthe incidence of locoregional recurrence as the first site of failure andthe subsequent distant metastases associated with such a recurrence. Theauthors estimated a 2% and 4% sur-vival benefit with postmastectomy radiationat 10 years and 15 years, respectively, for axillary node-negative patients.For axillary node-positive patients, the survival benefits were 6% at 10years and 9% at 15 years. 

An additional argument in support of postmastectomy radiation is theprevention of the morbidity associated with uncontrolled locoregional disease.Delaying radiation until the time of re-currence ultimately results inlong-term control in only 50% to 60% of patients.[5-11] 

As a result of the increased use of doxorubicin-based chemotherapy,che- moendocrine therapy, and high-dose chemotherapy supported by autologousbone marrow transplantation or peripheral blood progenitor-cell reinfusion,the role of postmastectomy radiation in the treatment of stages II-IIIdisease has been challenged. This article will survey the long historyof trials evaluating postmastectomy radiation--both the earlier studies(then) and more recent trials (now)--to determine whether or not this challengeis justified. 

Some of the earliest prospective randomized trials of breast cancertreatment evaluated the role of postmastectomy radiation. These trialscan be divided into two groups: (1) those evaluating the role of radiationfollowing radical or modified radical mastectomy,[12-20] and (2) thoseevaluating the role of radiation following simple mastectomy.[21-26] Theseearly trials have been criticized for the lack of a consistent randomizationprocess and inadequate stratification,[12-14] unequal patient distributionamong the treatment arms,[14] the use of confounding ad-juvant therapy,[14,19]and inadequate radiation in terms of the total dose[12-14,19,27,28] usedand/or the regions treated.[12-14,19] 

Despite these criticisms and variations in patient eligibility, theresults achieved were similar.[29] All but one of the trials[26] demonstrateda decrease in the incidence of locoregional recurrence with radiation.Moreover, radiation resulted in a decrease in the incidence of distantmetastases in five trials [15,16,19,21,22,26] and a significant improvementin disease-free survival in four trials.[15,16,19,24,26] None of the trialsshowed a significant overall survival benefit for radiation. 

A combined analysis of the long-term results of the Oslo and Stockholmtrials found no significant overall or disease-free survival benefit ofradiation therapy in axillary node-negative patients. However, radiationsignificantly improved the 10-year disease-free and distant disease-freesurvival rates in axillary node-positive patients. 

--A meta-analysis of the mortality resultsfrom eight randomized trials evaluating postmastectomy radiation withoutadjuvant chemotherapy was first published in 1987 by Cuzick et al.[30]The analysis included results of the Manchester I, II and Regional trials,the National Surgical Adjuvant Breast and Bowel Project (NSABP) B04 trial,the Stockholm trial, the Oslo I and II trials, the Cancer Research Campaigntrial, the Edinburgh I trial, and an unpublished trial from Heidelbergthat compared radical mastectomy alone to radical mastectomy plus radiation.In this analysis, radiation therapy afforded no survival benefit for thefirst 10 years of follow-up. However, in patients who survived more than10 years, radiation significantly increased mortality. Data were unavailableon the specific causes of death. 

Subsequent reports from the individual trials[19,20,31-33] indicatedthat the increased mortality was related to cardiovascular disease. Patientswith left-sided tumors and those who received orthovoltage or cobalt-60irradiation of the regional nodes were at greater risk for death due tocardiovascular disease. 

The publication of this meta-analysis coincided with the increased useof doxorubicin-based chemotherapy regimens for axillary node-positive patientsand the initiation of clinical trials for adjuvant therapy in axillarynode-negative women. As a result, interest in and recommendations for postmastectomyradiation diminished. 

--In 1994, Cuzick et al[34] publishedan update of the 1987 meta-analysis with specific causes of death for allof the trials except NSABP B04. The previously reported significant increasein mortality from all causes in the 10-year survivors was no longer evident.An increase in cardiac deaths was observed in patients who received radiation,but this was offset by a decrease in the number of breast cancer-relateddeaths, especially in the more recent trials. After 10 years, no significantdifferences in mortality were seen among patients who received radiationfollowing a simple mastectomy. After 15 years, cardiac-related causes ofdeath were increased in patients who received radiation following a radicalor modified radical mastectomy. The increased mortality was most apparentin women who were 50 years or more of age and clinically node-negative.

Radiation technique was implicated in these results. Trials that employedorthovoltage irradiation[20,31] or cobalt-60 irradiation of the internalmammary nodes[19,33] demonstrated an increased incidence of cardiac deaths.This meta-analysis suggests that with improvements in radiotherapy technique,and hence, the avoidance of cardiac morbidity and mortality, postmastectomyradiation may have the potential for significantly improving survival dueto a decrease in breast cancer-related deaths. 

--In 1995, the Early Breast Cancer Trialists'Collaborative Group published a second meta-analysis of prospective randomizedtrials of radiation and surgery for operable breast cancer.[66] This analysisincluded 36 trials initiated before 1985, 32 of which evaluated the roleof postmastectomy radiation with or without chemotherapy. Data on causeof death were available from 28 trials. 

Overall mortality was not increased in the patients who received radiation.Rates of 10-year survival were 63% in axillary node-negative patients whoreceived radiation vs 62.8% in those who did not. The corresponding numberswere 47.7% and 46.3% for axillary node-positive patients. Radiation resultedin a decrease in the number of breast cancer deaths but an increase inthe number of non-breast cancer deaths, which, for the first 2 decadesafter treatment, was more evident in women 60 years or more of age. Radiationreduced the rate of local recurrence by 67%. The authors were unable toassess the relationship between improved local control and distant metastaseswith the available data. 

Therefore, two meta-analyses have demonstrated the ability of postmastectomyradiation to decrease the risk of locoregional recurrence and breast cancer-relateddeaths and emphasize the importance of minimizing long-term morbidity withappropriate techniques. 

Risk Factors for Locoregional Recurrence After Mastectomy and NoSystemic Therapy

In surgical series in which adjuvant systemic therapy has not been used,the risk of a local (chest wall) or regional (axilla, supraclavicular,or internal mammary node) recurrence following radical or modified radicalmastectomy has been related to primary tumor size and axillary nodal status.[14,35-47]Patients with primary tumors 5 cm or more in diameter and those with fouror more positive axillary nodes have a 25% to 30% incidence (high risk)of locoregional recurrence.[14,35-39,41,46] In contrast, patients withprimary tumors 2 cm or less or negative axillary nodes have a 5% to 10%incidence (low risk) of locoregional recurrence. Patients who have primarytumors 2 to less than 5 cm in diameter or one to three positive axillarynodes have a 10% to 15% incidence (intermediate risk) of locoregional recurrence.[14,35,365,39,41,46]Additional factors that correlate with an increased risk of locoregionalrecurrence following mastectomy include young age ( 40 years or less),[35,48,49]negative estrogen receptors,[50,51] high histologic grade,[44,52,53] andpectoral fascia involvement.[35,54] 

Data are limited regarding the correlation between mastectomy resectionmargin status and the risk of a chest-wall recurrence. Axillary node-negativepatients with primary tumors 5 cm or less and a close margin ( 4 mm orless) have a 6% incidence of locoregional recurrence.[40] Mentzer et al[55]were unable to correlate the incidence of locoregional recurrence withmargin status in 100 women with stage II breast cancer. Unfortunately,margin status was assessed grossly in 67% of these patients. None of thefour patients with a positive margin who did not receive radiation experienceda locoregional recurrence. Locoregional recurrences developed in 3 of 25patients with a margin 2 mm, or less 5 of 45 patients with a margin 5 mm,or less and 1 of 33 patients with a margin more than 5 mm. 

The most common site for a locoregional recurrence is the chest wall,either as a single site or a component of multiple sites.[5,6,10,11,56]Approximately 40% of the chest-wall recurrences involve the mastectomyscar.[52,57] The second most common site is the supraclavicular region.[5,6,10,11,56]Axillary recurrences are uncommon in the dissected axilla, especially when10 or more axillary nodes are removed.[58] The presence of extracapsularextension does not increase the risk of an axillary recurrence,[59-62]provided that an adequate dissection has been performed. Internal mammarynode recurrences represent less than 10% of ll locoregional recurrences.[5,6,10,11,56]

The majority of locoregional recurrences postmastectomy occur within5 years of surgery.[6,11,52,57,63,64] Incidence peaks in the second year.[52]Montague[65] reported that 34% of locoregional recurrences following mastectomyoccurred after 5 years and 10%, after 10 years. 

Impact of Adjuvant Chemotherapy on Locoregional Recurrence

New developments in adjuvant systemic therapy in the 1980s and '90schallenged the role of postmastectomy radiation: In the 1980s, the benefitof adjuvant chemotherapy in both axillary node-positive and node-negativepatients was demonstrated[66] and increased emphasis was placed on doxorubicin-basedregimens and combined chemoendocrine therapy. In the 1990s, dose-intensiveregimens with autologous bone marrow transplantation or peripheral bloodprogenitor-cell reinfusion became prominent. It was presumed that adjuvantsystemic therapy would eradicate micrometastatic disease and prevent locoregionalrecurrence with the same degree of efficacy. Adjuvant chemotherapy and/orendocrine therapy would, therefore, obviate the need for postmastectomyradiation. 

Identifying Patients at Risk of Treatment Failure

--Identificationof patients at a significant risk of locoregional recurrence despite adjuvantsystemic therapy is difficult. Few studies have specifically addressedthe issue of factors that predict an increased risk of locoregional recurrencein post- mastectomy patients who receive adjuvant chemotherapy.[46,47,51,54,67-69]Often, relapse-free and overall survival are reported without identifyingthe sites of first failure. When reported, patterns of first failure arefrequently mentioned as crude numbers. Median follow-up is often short.

In two studies,[47,69] 30% to 40% of locoregional recurrences in patientsreceiving adjuvant chemotherapy occurred after 3 years. Pisansky et al[69]reported 3- and 8-year cumulative incidences of isolated locoregional recurrenceof 12% and 20%, respectively, in node-positive patients entered into twoprospective randomized trials evaluating cyclosphosphamide (Cytoxan, Neosar),fluorouracil, and prednisone, with or without radiation. These observationssuggest that the risk of a locoregional recurrence will be underestimat-ed in series without long-term follow-up. In the Eastern Cooperative OncologyGroup (ECOG) analysis of locoregional recurrence in node-positive patientsreceiving chemotherapy results were reported at 3 years only.[54] 

Additional problems include the failure to separate patterns of relapsein patients undergoing conservative surgery plus radiation from those undergoingmastectomy. Factors that predict for an increased risk of locoregionalrecurrence in these two groups are not always the same.[11] 

The following discussion will review the results of prospective randomizedtrials in which the impact of adjuvant chemotherapy and/or endocrine therapyon locoregional recurrence in patients who have undergone mastectomy andaxillary dissection has been reported. For the tables, the incidence ofisolated locoregional recurrence (ie, without simultaneous distant metastases)is presented since this represents the pattern of failure most likely tobe altered by adjuvant radiation. 

Trials Comparing Chemotherapy to Observation

--Table1 and Table 2 summarize the resultsof randomized trials comparing adjuvant chemotherapy to observation inaxillary node-negative and node-positive women. In node-negative patients,the incidence of an isolated locoregional recurrence is approximately 10%,with median follow-up ranging from 4.5 to 10 years. Chemotherapy appearsto result in a modest decrease in the incidence of locoregional recurrencein these low-risk patients. 

The results for node-positive patients do not demonstrate a consistentpattern. Some studies show minimal or no decrease in the incidence of locoregionalrecurrence with adjuvant chemotherapy,[54,70,73] while others [72,74-77]demonstrate a decrease in the incidence of locoregional recurrence. Thedifferences do not appear to be related to the period of follow-up, sincetwo[70,71,73] of the three long-term studies failed to demonstrate a beneficialeffect, while one[74] did. 

The variation in results also is not explained by menopausal status.Two[54,72,74] of the postmenopausal studies found no difference in locoregionalrecurrence with the addition of chemotherapy, while two others[76,77] showeda difference. 

Both of the studies[74,75] that used melphalan (Alkeran) demonstrateda reduction in the incidence of locoregional recurrence, whereas three[54,70,71,73]of the six studies employing CMF (cyclophosphamide, methotrexate, fluorouracil)or a variation thereof failed to demonstrate a significant decrease. Inthe Guy's/Manchester trial,[72] CMF decreased locoregional recurrence inpremenopausal but not postmenopausal women. 

Approximately 60% of the patients entered into these trials would beconsidered at intermediate risk for a locoregional recurrence, ie, theyhad one to three positive nodes or primary tumor size less than 5 cm. Overall,adjuvant chemotherapy resulted in a 15% incidence of isolated locoregionalrecurrence. For patients not receiving adjuvant chemotherapy, the incidenceof locoregional recurrence ranged from 14% to 44%. Isolated locoregionalrecurrences comprised 30% to 40% of all recurrences in these trials. 

--Table3 presents the incidence of isolated locoregional recurrence in a numberof prospective randomized trials that evaluated various chemotherapeuticregimens in axillary node-positive patients. The incidence of isolatedlocoregional recurrence after mastectomy appears to be minimally influencedby multiagent chemotherapy when compared to a single agent, long vs shorterduration of therapy, modifications of CMF, the use of doxorubicin-basedregimens, or sequential vs alternating regimens. 

In the Ludwig V trial[78] a single course of perioperative chemotherapywas associated with an increased risk of locoregional recurrence. In theNSABP B11 trial,[79] the addition of doxorubicin to melphalan and fluorouracildecreased the incidence of iso-lated locoregional recurrence in es- trogen-receptornegative, axillary node-positive patients. However, similar findings werenot observed in the NSABP B12[79] trial, in which estrogen receptor-positive,axillary node-positive patients received tamoxifen (Nolvadex) in additionto chemotherapy. Three additional trials have also failed to demonstratea significant decrease in the incidence of postmastectomy locoregionalrecurrence with doxorubicin- based regimens.[73,80,81] Also, the use ofsequential or alternating non-cross-resistant regimens has not significantlylowered the incidence of locoregional recurrence.[82-85] 

Little data are available on the incidence of postmastectomy locoregionalrecurrence in patients receiving high-dose or dose-intensive chemotherapy,with or without autologous bone marrow transplantation or peripheral bloodprogenitor-cell reinfusion.[86-92] In the Cancer and Leukemia Group G (CALGB)trial,[88] low-dose, low-intensity CAF (cyclophosphamide, 

Adriamycin, and fluorouracil) resulted in a higher incidence of locoregionalrecurrence. However, dose-intensive CAF was not associated with a lowerincidence of locoregional recurrence when compared to the same total doseof CAF but given in a moderately dose-intensive regimen. 

Peters et al[92] reported that three of nine patients with 10 or morepositive nodes who received standard-dose CAF followed by high-dose chemotherapy(cyclophosphamide, cisplatin [Platinol], and carmustine [BCNU]) and autologousbone marrow support developed a locoregional recurrence as the first siteof failure. This observation has prompted the routine use of postmastectomyradiation in patients with 10 or more positive nodes in the two ongoingprospective randomized trials evaluating high-dose chemotherapy with autologousbone marrow transplantation (CALGB 9082 and Intergroup 0121). 

--For the trials listed in Table2 and Table 3, the proportion offailures that were isolated locoregional was approximately 30% to 40%;the majority of failures occurred at distant sites. Goldhirsch et al,[93]in an analysis of the Ludwig I-V trials, reported that adjuvant systemicchemotherapy and/or tamoxifen resulted in a greater reduction in the incidenceof locoregional recurrence, as opposed to bone or visceral metastases,as first site of failure. The authors suggested that current regimens havea greater impact on locoregional recurrence and a diminished effect ondistant failure. 

In the CALGB trial[88] low-dose, low-intensity CAF was associated withmore locoregional than distant failures. In the Milan trial of sequentialvs alternating non-cross-resistant regimens in patients with four or morepositive nodes, the more effective regimen (doxorubicin followed by CMF)resulted in an almost equal incidence of locoregional and distant failures(14% vs 17%). For the alternating regimen, the incidence of locoregionalfailure was 17% vs 32% for distant failure. Therefore, in this trial themore effective regimen had a greater impact on distant failure than onlocoregional recurrence. 

The early experience of Peters et al[92] with high-dose therapy suggestssimilar findings in the patients who did not receive postmastectomy radiation.Two series [54,69] have identified patients with primary tumors 5 cm ormore in size and four to seven positive nodes who received adjuvant chemotherapyas having a greater risk for locoregional recurrence than distant metastases.For the majority of node-positive patients, distant failures predominate.However, it is important to identify the subgroup of patients who are atgreatest risk for postmastectomy locoregional recurrence despite adjuvantchemotherapy and to quantitate this risk in order to better define thepotential benefit of radiation. 

Risk Factors for Locoregional Recurrence After Mastectomy Plus Chemotherapy

Several retrospective and prospective randomized trials have specificallyaddressed the issue of the risk of locoregional recurrence following mastectomyin patients receiving adjuvant chemotherapy. Factors analyzed for theirability to predict for a locoregional recurrence include age, menopausalstatus, obesity, primary tumor size and location, histology, histologicgrade, presence or absence of lymphatic invasion, pectoral fascia involvement,prominent lymphocytic infiltration or gross multicentric disease, numberof positive axillary nodes, type of surgery, interval from surgery to chemotherapy,dose intensity, and, more recently, DNA content and S-phase. The significantfindings of these studies are summarized in Table4. 

The two most important factors appear to be the number of positive nodesand primary tumor size. Table 5 presentsthe incidence of isolated locoregional recurrence related to these twofactors in patients receiving adjuvant chemotherapy. The incidence of isolatedlocoregional recurrence ranges from 8% to 36% for patients with four ormore positive nodes and 18% to 71% for primary tumors 5 cm or more, dependingon the period of follow-up. 

Pisansky et al[69] reported that 40% of locoregional recurrences occurredafter 3 years. The 8-year cumulative incidence of locoregional recurrencewas 20% for all node-positive patients receiving cyclophosphamide, fluorouracil,and prednisone, with or without tamoxifen, in two prospective randomizedtrials. The 3- and 8-year cumulative incidences of locoregional recurrencewere 8% and 20%, respectively, for patients with one to three positivenodes; 14% and 18%, respectively, for those with four to seven positivenodes; and 22% and 29%, respectively, for those with more than seven positivenodes. For T1 tumors, the corresponding figures were 8% and 19%; for T2tumors, 13% and 20%; and for T3a tumors, 71% and 71%. 

Salvage of Locoregional Recurrence in Patients Receiving Chemotherapy

Few data are available regarding the ability to salvage an isolatedlocoregional recurrence following mastectomy in patients who have receivedadjuvant chemotherapy. Danoff et al[56] reported distant metastases in56% of patients who received adjuvant chemotherapy and subsequently developedan isolated locoregional recurrence. Tennrall-Nittby et al[94] describedoutcome in 1,153 stage II patients entered into the South Sweden BreastCancer Trial. Premenopausal women were randomized to radiation, cyclophosphamide,or both. Salvage therapy was successful in 11 (48%) of 23 premenopausalpatients who developed an isolated locoregional recurrence following chemotherapy.

Goldhirsch et al[95] reported a 5-year survival rate of 53% in the 54premenopausal or perimenopausal women who developed local failure followingCMF with or without oophorectomy. In contrast, the 5-year survival ratewas 5% in the 59 patients who developed a regional failure. These datasuggest that, at best, salvage therapy may prove effective in 50% of womenwith an isolated locoregional recurrence following mastectomy and adjuvantchemotherapy. 

Effect of Tamoxifen on Locoregional Recurrence

Several studies have suggested that tamoxifen may be more effectivein preventing locoregional recurrence than distant metastases.[76,96-98]The question arises as to whether tamoxifen alone or in combination withchemotherapy can decrease the incidence of postmastectomy locoregionalrecurrence and obviate the need for radiation. The incidence of isolatedlocoregional recurrence reported from prospective randomized trials evaluatingchemoendocrine therapy in axillary node-positive patients is presentedin Table 6. The addition of oophorectomyor tamoxifen to chemotherapy in premenopausal women has not significantlydecreased the incidence of locoregional failures or altered the proportionof failures that are locoregional. 

In the Gynecologic Adjuvant Breast Group study,[99] CMF was more effectivethan tamoxifen in preventing locoregional recurrence in women with estrogen-receptor-positivetumors and one to three positive axillary nodes (4% vs 18%). In women 49years old or less who had four or more positive nodes or estrogen receptor-negativetumors, the addition of tamoxifen to doxorubicin and cyclophosphamide resultedin a modest decrease in the incidence of locore- gional recurrence (from21% to 17%). Similarly, in postmenopausal women, the addition of tamoxifento chemotherapy[54,77,100-102] or the addition of chemotherapy to tamoxiifen[76,100,102-105]modestly lowered the incidence of locoregional recurrence. However, ina recent update[106,107] of the National Cancer Institute of Canada ClinicalTrials Group study of 706 postmenopausal, node-positive, estrogen or progesteronereceptor-positive women, there was no significant difference between tamoxifenalone and CMF plus tamoxifen with regard to locoregional progression-freesurvival. 

In the Gynecologic Adjuvant Group study,[99] tamoxifen was more effectivethan CMF in preventing locoregional recurrence in women 50 years of ageor more with one to three positive nodes and estrogen receptor-positivetumors (7% vs 16%). In women with four or more positive nodes, the additionof tamoxifen to doxorubicin and cyclophosphamide decreased the incidenceof locoregional recurrence from 27% to 17%. The NSABP B14[83] trial forestrogen receptor-positive, axillary node-negative women demonstrated adecrease in the incidence of local recurrence with tamoxifen when comparedto observation alone. These studies suggest that the maximum benefit interms of locoregional control is achieved with the combination of chemotherapyand tamoxifen. 

Is Postmastectomy Radiation Better Than Chemotherapy in PreventingLocoregional Recurrence?

Two prospective randomized trials have provided evidence that postmastectomyradiation is more effective than chemotherapy in preventing locoregionalrecurrence.[15,108,109] In the Stockholm Breast Cancer Group trial,[15]premenopausal and postmenopausal women with positive axillary nodes orprimary tumor size more than 3 cm were randomized to postmastectomy radiationor CMF. Rates of locoregional recurrence were 12% in the radiation groupand 22% in the chemotherapy group (P = .001). The median follow-up was6.5 years. 

In the Glasgow randomized trial comparing postmastectomy radiation,CMF, and radiation plus CMF, the locoregional recurrence rate was 31% forpatients receiving CMF, as compared with 16% for those receiving radiation.

Postmastectomy Chemotherapy or Tamoxifen vs Systemic Therapy PlusRadiation

--A total of 13 prospective randomizedtrials have evaluated the role of postmastectomy radiation and adjuvantchemotherapy in patients with positive axillary nodes. Three of these trialsincluded only patients with operable T3 tumors,[110-113] and five[108,114-117]reported results for patients with four or more positive nodes. The accrualperiod, number of patients, chemotherapy regimen, and percentage of patientswith four or more positive nodes or T3 tumors are shown in Table7. In seven of these trials, the majority of patients would be consideredat intermediate risk for a locoregional recurrence following mastectomy,ie, they had one to three positive nodes. 

The results of the trials are summarized in Table8 and Table 9. All of the trialsexcept one[114] demonstrated a decrease in the incidence of locoregionalrecurrence with the addition of radiation. In the lone negative study,conducted at the University of Arizona Cancer Center,[114] randomizationto radiation therapy was not strict and low doses of radiation were frequentlyemployed. 

In only 1 of the remaining 12 trials was the decreased incidence oflocoregional recurrence with radiation associated with a decreased incidenceof distant metastases as the site of first failure. In the British Columbiatrial,[118] the distant disease-free survival rate at 12 years was 58%in the patients who received radiation vs 40% in those who did not (P =.007). The total incidence of distant metastases was not reported in thesetrials, and therefore, the impact of radiation on the decrease in distantmetastases occurring after a local failure is not evident.[4] Five of thetrials demonstrated a significant improvement in relapse-free survivalwith the addition of radiation [108,110,111,116,117,119,120] and two trialsreported a significant overall survival benefit.[110,111,119,120] 

The absence of a survival or relapse-freesurvival benefit in the majority of these trials is not unexpected. Intwo of the trials, radiotherapy was inadequate because of the low doseused[114] or regions treated.[121,122] As previously stated, the majorityof patients entered into these trials were not at high risk for a locoregionalrecurrence. 

The magnitude of the survival benefit from postmastectomy radiationrelated to the risk of locoregional recurrence can be calculated usingthe formula proposed by Suit and Westgate.[123] In patients with one tothree positive nodes receiving adjuvant chemotherapy, the incidence oflocoregional recurrence is approximately 10% (Table5). The addition of radiation decreases this incidence to approximately5% or less.[115] Therefore, the benefit in terms of locoregional recurrenceis only 5%. Approximately 20% to 30% of these patients will develop distantmetastases without a local recurrence,[54,69] and therefore, preventionof such a recurrence in these patients would not improve survival. Themaximum potential survival benefit (assuming no salvage of a locoregionalrecurrence) in patients with one to three positive nodes would be 3% to4% (5% - [1% - 2%]). None of the randomized trials evaluating postmastectomyradiation in this group of patients has had adequate numbers to demonstratea significant survival benefit of this magnitude. 

In patients with four or more positive nodes, the incidence of locoregionalrecurrence without radiation is 20% to 25% (Table5). The addition of radiation decreases this incidence to 10%.[110]A locoregional recurrence benefit of 15% translates into a survival benefitof 9% to 10% since 30% to 40% of these patients will develop distant metastaseswithout local failure.[54,69] Assuming a 5-year survival of 50% in thesepatients and a minimum follow-up of 3 years, a randomized trial evaluatingpostmastectomy radiation in this subset of patients would require 328 patientsin each arm to demonstrate a significant survival benefit of 10% (alpha= .05; beta = .2). 

Thus, it is evident from Table 7 thatinadequate patient numbers are responsible, in part, for the nonsignificantresults. For the five trials [108,114-117] reporting results for patientswith four or more positive nodes, the absolute overall survival benefitwith radiation has ranged from 6% to 12%.[108,116,117] In two of the trials,[114,115]patients who received radiation had a decrease in survival (absolute decrement,4%). 

For patients with T3 tumors, the benefit with radiation in terms oflocoregional recurrence is approximately 20% (Table5); this would translate into a survival benefit of 14% since 30% ofthese patients who achieved local control would develop distant metastases.[54,69]Table 9 presents the results of the randomizedtrials for operable T3 breast cancer. The absolute survival benefit rangesfrom 2% to 25%. In only one trial was this difference significant.[110,111]

Therefore, the results of the prospective randomized trials evaluatingpostmastectomy adjuvant chemotherapy, with or without radiation, have limitedclinical applicability due to the inclusion of patients at low or intermediaterisk for locoregional recurrence, for whom radiation would be expectedto have minimal benefit and inadequate patient numbers to demonstrate asignificant benefit for the high-risk patients. 

--The results of two prospective randomizedtrials[94,104,124,125] evaluating postmastectomy tamoxifen, with or withoutradiation, in axillary node-positive patients are summarized in Table10. The addition of radiation to tamoxifen decreased the incidenceof locoregional recurrence in both trials and significantly improved relapse-freesurvival in one trial[104,124,125] but did not afford an overall survivalbenefit. 

Three other trials[43,94,101,126] have evaluated the addition of tamoxifento radiation in axillary node-positive patients. In the South Sweden BreastCancer trial,[94] 719 postmenopausal women were randomized to radiationalone, radiation plus tamoxifen, or tamoxifen alone. With a median follow-upof 8 years, the incidences of locoregional recurrence in the three groupswere 8%, 6%, and 18%, respectively. There were no significant differencesamong the groups in terms of survival. 

In the Stockholm trial, [101] 427 postmenopausal women with positiveaxillary nodes or primary tumor size more than 3 cm were randomized toradiation alone, radiation plus tamoxifen, CMF alone, or CMF plus tamoxifen.With a mean follow-up of 6.5 years, locoregional recurrence rates were13% in patients who received radiation and 12% in those treated with radiationplus tamoxifen. At 10 years, overall and relapse-free survival rates were62% and 57%, respectively, in the patients treated with radiation; 65%and 63%, respectively, in those given radiation and tamoxifen; 50% and31%, respectively, in the CMF-treated patients; and 52% and 47%, in thosereceiving CMF and tamoxifen. 

In the Danish Breast Cancer Cooperative Group study 77C,[43,126] 1,716postmenopausal women (29% of whom had four or more positive nodes and 24%of whom had T3 tumors) were randomized to radiation alone or radiationplus tamoxifen. The 9-year actuarial locoregional recurrence rate was 17%with radiation and 12% with radiation and tamoxifen. 

These studies suggest that the addition of tamoxifen to radiation affordsminimal benefit in terms of locoregional recurrence. 

Postmastectomy Radiation in High-Risk Patients Receiving AdjuvantChemotherapy

Several retrospective series have evaluated postmastectomy radiationin patients at significant risk for locoregional recurrence who were receivingadjuvant chemotherapy.[90,91,127-129] Fowble et al[127] reported 4-yearoverall and disease-free survival rates of 67% and 47%, respectively, in63 women who received postmastectomy radiation and chemotherapy (CMF withor without prednisone, or CAF). The median number of positive nodes waseight, and 43% of the patients had primary tumors 5 cm or more. The 4-yearactuarial isolated locoregional recurrence rate was 5%. 

In a study by Uematsu et al[128] the crude locoregional recurrence ratewas 4% in 49 patients with four to nine positive nodes who received postmastectomyradiation and chemotherapy (CMF or CAF). The 10-year actuarial relapse-freesurvival was 40%. In patients with 10 or more positive nodes, the crudelocoregional recurrence rate was 3% and the 10-year actuarial relapse-freesurvival was 10%. 

Ung et al[129] reported a crude locoregional recurrence rate of 12.5%in 64 patients (70% of whom had 10 or more positive nodes) who receivedpostmastectomy radiation and chemotherapy (CMF or AC [Adriamycin and cyclophosphamide]),with or without tamoxifen. The 10-year actuarial rates of freedom fromdistant metastases and overall survival were 35% and 48%, respectively.For patients with 10 or more positive nodes, the 5-year actuarial rateof freedom from distant metastases was 42%. 

Buzdar et al[130] reported a 5% local failure rate with postmastectomyradiation in patients with 10 or more positive nodes entered into fourprospective trials evaluating doxorubicin-containing adjuvant therapy.Marks et al[90] described the results of postmastectomy radiation in 40patients who received standard CAF followed by high-dose chemotherapy (cyclophosphamide,cisplatin, carmustine) with autologous bone marrow transplant support.In the 34 patients who completed radiation therapy, the 2-year disease-freesurvival rate was 87%. A single patient developed an isolated chest wallrecurrence. 

At present, the indications for postmastectomy radiation in patientsreceiving adjuvant systemic therapy include primary tumor size 5 cm ormore, four or more positive axillary nodes, and pectoral fascia involvement.There are minimal data regarding locoregional recurrence rates in patientswith a positive deep margin.[40,55] However, in clinical practice thishas often been an indication for postmastectomy radiation. 

The presence of extracapsular nodal extension is not an indication foraxillary irradiation, provided that an adequate axillary dissection hasbeen performed. Four studies[59,60-62] have failed to demonstrate an increasedrisk of axillary recurrence in patients with extracapsular extension whodid not receive axillary radiation. However, in the British Columbia trial,[116]postmastectomy radiation in patients with extensive extracapsular spreadresulted in a significant improvement in 10-year actuarial overall anddisease-free survival. The presence of gross multicentric disease alsois not an indication for postmastectomy radiation in the absence of a primarytumor size 5 cm or more or four or more positive nodes.[131] A single trialsupports the role of postmastectomy radiation in T3 axillary node-negativebreast cancer.[111] 

The most common sites for postmastectomy locoregional recurrence inpatients receiving adjuvant chemotherapy are the chest wall and supraclavicularregion.[43,54,56,105] Axillary node sampling in the Danish Breast CancerCooperative Group trials 82b and 82c indicated that 45% and 40% of thelocoregional recurrences occur in the axilla.[105] 

Internal mammary node recurrences are infrequent,[54] and treatmentwith radiation has not resulted in a significant decrease in this patternof failure[90,132] but has resulted in an increased incidence of complications,including symptomatic pneumonitis,[90] as well as the previously reportedincreased mortality from cardiovascular disease.[19,33] In the retrospectiveseries described by Fowble et al,[127] Uematsu et al,[128] and Ung et al,[129]no specific attempt was made to routinely treat the internal mammary nodesin high-risk patients receiving postmastectomy radiation and chemotherapy.In two[128,129] of these three series, there were no internal mammary noderecurrences. Fowble et al[127] reported a single internal mammary noderecurrence in 63 high-risk patients. 

Based on the results of all these studies, postmastectomy radiationfor axillary node-positive patients should include the chest wall and supraclavicularnodes. Because of the morbidity associated with internal mammary node irradiationand the fact that these nodes are an infrequent site of locoregional recurrence,routine irradiation of this region is not indicated. Irradiation of themid- and lateral axilla with a posterior axillary field has not resultedin a decreased incidence of axillary recurrence even in patients with extracapsularextension, provided that an adequate axillary dissection has been performed.[132]

Sequencing of Postmastectomy Radiation and Chemotherapy

The optimal sequencing of postmastectomy radiation and chemotherapyremains to be determined. 

--Several series[133-135] havefound an increased incidence of locoregional recurrence in patients whoreceived radiation following the completion of chemotherapy. Buchholz etal[133] reported a 5-year actuarial locoregional recurrence rate of 23%in 35 patients who received postmastectomy radiation more than 6 monthsafter mastectomy, as compared with a rate of 5% in 19 patients who receivedradiation within 6 months, (P = .192). Doxorubicin-based regimens weremore commonly used in the delayed-radiation group, and a greater percentageof patients in the early-radiation group had a close or positive marginor four or more positive nodes. In the study of Buzdar et al,[134] the10-year actuarial locoregional recurrence rate was 16% in patients whoreceived chemotherapy first vs 8% in patients who received postmastectomyradiation first. 

Locoregional recurrences during chemotherapy are infrequent. Stefaniket al[68] observed a single postmastectomy locoregional recurrence duringchemotherapy. Griem et al[115] reported two locoregional recurrences duringchemotherapy in 100 women with four or more positive axillary nodes. From70% to 80% of postmastectomy locoregional recurrences occur after the completionof chemotherapy.[68,115] 

--There are limited data regarding the impactof the sequencing of postmastectomy radiation and chemotherapy on survival.Two prospective randomized trials[117,136] demonstrated no adverse effecton survival when the initiation of chemotherapy is delayed until afterthe completion of postmastectomy radiation. In the Oncofrance[136] trialof AVCF (Adriamycin, vincristine, cyclophosphamide, and fluorouracil) vsCMF for axillary node-positive breast cancer, 55% of the patients receivedradiation. No difference in outcome was seen in the patients who receivedradiation first. 

In the Southeastern Cancer Group study,[117] the 10-year actuarial overallsurvival was 46% in patients who received radiation first followed by sixcycles of CMF, as compared with 34% in patients given six cycles of CMFwithout radiation. Patients who received radiation had an automatic 25%dose reduction for the first two cycles of chemotherapy. 

In a randomized trial reported by Lara Jimenez et al,[137] a 10-yeardisease-free survival of 57% was seen in patients who received postmastectomyradiation after an initial three cycles of CMF and followed by an additionalthree cycles of CMF; this was significantly better (P = .05) than the 10-yeardisease-free survival rates observed in patients treated with radiationfirst followed by six cycles of CMF (41%) or those given chemotherapy first(CMF for six cycles) followed by radiation (46%). A total of 248 axillarynode-positive patients were analyzed in this trial. 

Two retrospective series have reported a decreased survival when theinitiation of radiation is delayed.[133,135] Buchholz et al[133] reporteda 5-year actuarial survival rate of 78% in patients who received postmastectomyradiation within 6 months of chemotherapy, as compared with 73% in thosewho received it after 6 months (P = .32). In contrast, Brufman et al[138]retrospectively reviewed 87 patients who received postmastectomy radiationand CMF. The 7-year relapse-free survival rate was 77% in patients whoreceived concurrent chemotherapy, 77% for patients treated with chemotherapyfirst, and 42% in those given radiation first. 

At present, the most common method for integrating postmastectomy radiationand chemotherapy in high-risk axillary node-positive patients is to deliverradiation at the completion of chemotherapy. In the CALGB and intergrouprandomized trials evaluating dose-intensive chemotherapy with autologousbone marrow transplantation, patients receive radiation after the completionof standard chemotherapy or dose-intensive chemotherapy. 

Complications of Combined-Modality Therapy

The late effects of combined-modality therapy include secondary leukemiaand second cancers after adjuvant chemotherapy and/or radiation and cardiaceffects after adjuvant chemotherapy and/or radiation. These late complicationshave been described in two recent review articles.[139,140] This discussionwill focus on complications reported by recent series evaluating postmastectomyradiation and chemotherapy. 

--In two retrospective series of postmastectomy radiationand adjuvant chemotherapy, the incidence of severe arm edema ranged from9% to 13%.[127,129] In general, patients who developed this complicationreceived treatment to the chest wall and supraclavicular nodes. Also, aposterior axillary boost field was frequently employed. 

The radiation factor that appears to contribute most to the developmentof arm edema is the use of the supraclavicular field.[132] The additionof the posterior axillary field does not increase this risk.[132] The mostimportant non-radiation factors that increase the risk of subsequent armedema are infection in the ipsilateral arm and obesity or significant weightgain after breast cancer treatment.[141] 

--Symptomatic pneumonitis has been reportedin 5% of patients receiving postmastectomy radiation and standard chemotherapy.[127]Marks et al[90] noted a 16% incidence of symptomatic pneumonitis in 32patients who received dose-intensive chemotherapy (cyclophosphamide, cisplatin,and carmustine) with autologous bone marrow support plus postmastectomyradiation that included the internal mammary nodes. None of the six patientswho did not receive radiation to the internal mammary nodes experiencedsymptomatic pneumonitis. 

Articles

With the increased use of doxorubicin-based chemotherapy, chemoendocrine therapy, and high-dose chemotherapy with autologous bone marrow transplantation or peripheral blood progenitor-cell reinfusion, the role

Barbara Fowble, MD

Articles

Postmastectomy Radiation: Then and Now