Predictors of locoregional outcome in patients receiving neoadjuvant therapy and postmastectomy radiation†
Presented as an oral presentation at the 51st Annual Meeting of the American Society for Radiation Oncology; San Diego, California; October 31 to November 4, 2010.
The objective of this study was to identify predictors of locoregional recurrence (LRR) after neoadjuvant therapy (NAT) and postmastectomy radiation (PMRT) in a cohort of patients with stage II through III breast cancer and to determine whether omission of the supraclavicular field had an impact on the risk of LRR.
The authors reviewed records from 464 patients who received NAT and PMRT from January 1999 to December 2009.
The median patient age was 50 years (range, 25-81 years). Clinical disease stage was stage II in 29% of patients, stage III in 71%, and inflammatory in 14%. Receptor status was estrogen receptor (ER)-positive in 54% of patients, progesterone receptor (PR)-positive in 39%, human epidermal growth factor receptor 2 (HER2)-positive in 24%, and negative for all 3 receptors (triple negative) in 32%. All patients received NAT and underwent mastectomy, and 19.6% had a complete pathologic response in the breast and axilla, 17.5% received radiation to the chest wall only, and 82.5% received radiation to the chest wall and the supraclavicular field; omission of the supraclavicular field was more common in patients with lower clinical and pathologic lymph node status. The median follow-up was 50.5 months, and the 5-year cumulative incidence of LRR was 6% (95% confidence interval, 3.9%-8.6%). Predictors of LRR were clinical stage III (P = .038), higher clinical lymph node status (P = .025), higher pathologic lymph node status (P = .003), the combination of clinically and pathologically positive lymph nodes (P < .001), inflammatory presentation (P = .037), negative ER status (P = .006), negative PR status (P = .015), triple-negative status (P < .001), and pathologic tumor size >2 cm (P = .045). On univariate analysis, omission of the supraclavicular field was not associated significantly with LRR (hazard ratio, 0.89; P = .833); however, on multivariate analyses, omission of the supraclavicular field was associated significantly with LRR (hazard ratio, 3.39; P = .024).
Presenting stage, receptor status, pathologic response to neoadjuvant therapy, and omission the supraclavicular field were identified as risk factors for LRR after neoadjuvant therapy and PMRT. Cancer 2013. © 2012 American Cancer Society.
Three major randomized trials1-3 and the landmark Early Breast Cancer Trialists' Collaborative Group meta-analysis4 have demonstrated that postmastectomy radiation (PMRT) reduces the risk of locoregional recurrence (LRR) and the improves survival of patients with breast cancer. On the basis of these and other studies, the role of PMRT traditionally has been determined by pathologic staging, with surgery as the first treatment modality.
However, the use of neoadjuvant therapy (NAT) has become increasingly common, particularly in locally advanced disease. Although it has not been demonstrated that the sequence of chemotherapy has an impact on survival, younger premenopausal patients may derive a greater benefit from NAT,5 and response to NAT is predictive of both local control and survival.6, 7 NAT has an impact on the extent of disease at the time of surgery and results in a pathologic complete response (pCR) in the breast and axilla in 5% to 40% of patients. The response to NAT is affected by many variables, including receptor status, chemotherapy type, and initial stage at presentation.5, 7-9 Although NAT has many advantages, its impact on surgical staging reduces the applicability of the traditional pathologic guidelines for PMRT, and clear guidelines for the use of PMRT after NAT have not been established. Retrospective series have demonstrated that the elimination of PMRT after NAT in high-risk patients, particularly those with stage III breast cancer, results in unacceptably high rates of recurrence, even in the setting of a pCR.10, 11 For this reason, the role of PMRT generally is determined by clinical staging before NAT without regard for the response to NAT. Risk factors for recurrence in this specific setting are not well established.
In addition, the role of regional lymph node irradiation in patients with breast cancer is increasingly controversial. In the postmastectomy setting, adjuvant radiation has most commonly been defined as radiation of the chest wall and regional lymph nodes, including supraclavicular (SCV) and internal mammary (IMN) lymph nodes. Although treatment of the SCV field is generally well tolerated, it can result in skin/cosmetic toxicities and increased risk of brachial plexopathy, pneumonitis, lymphedema,12, 13 and cardiac/cerebrovascular mortality.14 Thus, identification of patients who can safely have this field omitted is desirable. However, to our knowledge, no studies to date have assessed omission of the SCV field in the setting of PMRT after NAT.
At our institution, use of the SCV field as a component of PMRT previously was not standardized, leading to variation in clinical practice over the last 10 years, with omission of the SCV field in selected patients. This variation in practice allows a meaningful, retrospective review of outcomes in PMRT patients who received PMRT to the chest wall with or without the SCV field. Our objectives were to identify predictors of LRR after NAT and PMRT, to assess the patterns of use of the SCV field in relation to clinical and pathologic lymph node status at our institution, and to determine whether omission of the SCV field in selected patients resulted in poorer survival or a higher risk of LRR.
MATERIALS AND METHODS
This retrospective review was approved by the University of Miami Institutional Review Board. A review of the medical records of patients with breast cancer who received PMRT between January 1999 and December 2009 at Jackson Memorial Hospital and at the University of Miami's Sylvester Comprehensive Cancer Center identified 464 patients who received NAT. We collected data on patient demographics, disease and treatment characteristics, and clinical outcomes.
Breast cancer staging was determined clinically based on physical examination at the time of diagnosis. Tumor size was determined by physical examination or by imaging studies if no palpable tumor was present. Patients did not undergo sentinel lymph node biopsy before they received NAT. Fine-needle aspiration of clinically suspicious axillary lymph nodes also was not standard during the treatment period assessed but was performed in selected patients, and clinically suspicious lymph nodes that were negative on fine-needle aspiration were staged as clinically negative (cN0).
Follow-up was calculated from the date of diagnosis. The date of progression was selected as the date of first event, including LRR, distant metastasis (DM), or death. Local failure was defined as tumor recurrence in the chest wall. Regional failure was defined as recurrence in the axilla, internal mammary lymph nodes, or supraclavicular fossa. Local failure and regional failure were defined together as LRR. Progression-free survival (PFS) was calculated from the date of diagnosis to the earliest occurrence of LRR, DM, or death from any cause. Overall survival (OS) calculated from the date of diagnosis to the date of death from any cause, with surviving patients censored at date of last contact. PFS and OS were estimated using the Kaplan-Meier method. A pCR to NAT was defined as the absence of invasive disease in the breast and axilla at the time of surgery.
The incidence of LRR, with or without synchronous distant failure (events <3 months apart), and the incidence of DM as first failure were estimated using the method of cumulative incidence described by Gray using the “cuminc” procedure in the R statistical package “cmprsk” (R Foundation for Statistical Computing, Vienna, Austria),15 with death as competing risk. The effects of potential prognostic factors were examined using the Gray test, which compares cumulative incidence curves, or the test of Fine and Gray, which is based on the competing risk Cox proportional hazards regression method implemented in the “crr” procedure in the “cmprsk” package. Statistical analyses were conducted using SAS software (version 9.2; SAS Institute Inc., Cary, NC) and R software (version 2.15.0; The R Development Core Team).
Patient and Disease Characteristics
Patient demographics and tumor characteristics are provided in Table 1. The median patient age at diagnosis was 50 years (range, 25-81 years), and 55% of patients were premenopausal or perimenopausal. Twenty-five percent of patients were black, 73% were white, 2% were Asian/other, and 59% were Hispanic. Clinical stage at presentation was as follows: 5.8% of patients had stage IIA disease, 23.5% had stage IIB disease, 39.5% had stage IIIA, 26.1% had stage IIIB disease, and 5% had stage IIIC disease. Inflammatory disease was observed in 14.2% of patients. Tumor histology was ductal in 81.9% of tumors, lobular in 7.8%, medullary in 3.4%, metaplastic in 2.8%, and other in 4.1%. Receptor status was as follows: 24% of patients were positive for human epidermal growth factor 2 (HER2), 54% were positive for estrogen receptor (ER), 39% were positive progesterone receptor (PR), and 32% were positive for all 3 receptors (triple negative).
Table 1. Patient Demographics and Tumor Characteristics
|All patients||464||100|| || || |
|Race|| || ||Clinical stage, n = 463|| || |
|Asian and other||9||1.9||IIIA||183||39.5|
|Ethnicity|| || ||IIIB||121||26.1|
|Non-Hispanic||192||41.4||Clinical tumor classification|| || |
|Age at diagnosis, y|| || ||Tx||2||0.4|
|Mean ± SD||50.4 ± 9.9|| ||T4d||66||14.2|
|Median (range)||50 (25-81)|| ||Clinical lymph node classification|| || |
|Menopausal status|| || ||Nx|| || |
|Histology|| || ||N2||200||43.1|
|Lobular||36||7.8||Clinical tumor size, cm||20||4.3|
|Medullary||16||3.4||Mean ± SD|| || |
|Metaplastic||13||2.8||Median (range)||7.3 ± 4.1|| |
|Other||19||4.1|| ||6 (0.5-32)|| |
|Hormone receptor status|| || || || || |
|ER|| || ||HER2, n = 460|| || |
|Positive||249||53.7||Positive||112 (24.3)|| |
|PR, n = 454|| || ||TN, n = 463|| || |
|Positive||179||39.4||Negative for ER, PR, and HER2||149||32.2|
Treatment characteristics are summarized in Table 2. All patients received NAT. Forty-four percent of patients received a platinum salt/docetaxel/anthracycline as a component of their NAT regimen, 19.6% received anthracycline/taxane-based chemotherapy, 20% received trastuzumab-containing regimens, and 16.2% received other chemotherapy regimens. Among patients who had HER2-overexpressing tumors, 83% received trastuzumab-containing therapy. All underwent mastectomy, and 98% underwent axillary lymph node dissection. A median of 17 lymph nodes were removed in each patient. Three patients had positive surgical margins. All patients received radiation to the chest wall with or without regional lymph nodes, and the median total chest wall dose inclusive of boost was 60.4 grays (range, 46.8-74.4 grays). Radiation generally was delivered using medial and lateral photon tangents with a 1-cm bolus placed over the chest wall every other day; treatment of the internal mammary lymph nodes rarely was performed; and, if those lymph nodes were included, then the technique was most commonly deep tangents. It is difficult to determine the number of patients retrospectively who had the internal mammary lymph nodes included, but we estimate this was done in <5% of patients. Radiation to the SCV field was delivered by a standard anterior or anterior oblique field, prescribed to a depth of 3 to 5 cm. The majority of patients in this series received treatment using a 2-dimensional technique, with standard bony anatomic field boundaries for the SCV field, including the transverse process of the vertebral bodies medially, the match line inferiorly, lateral to the humeral head laterally, and above the humeral head superiorly. The match line typically was placed at the head of the clavicle.
Table 2. Treatment Characteristics
|Chemotherapy sequence|| || |
|Neoadjuvant and adjuvant||216||46.6|
|Neoadjuvant chemotherapy regimen|| || |
|Breast surgery|| || |
|Axillary surgery|| || |
|Sentinel LN biopsy||6||1.3|
|Axillary LN dissection||456||98.3|
|No. of positive LNs, n = 462|| || |
|Among 257 patients with positive LNs|| || |
|Mean ± SD||6.2 ± 6.2|| |
|Median (range)||4 (1-32)|| |
|Pathologic LN status|| || |
|Clinical and pathologic LN status|| || |
|pCR|| || |
|pCR in breast and axilla||91||19.6|
|pCR in breast||129||27.8|
|pCR in axilla||205||44.2|
|Pathologic tumor size in 335 patients with size >0, cm|
|Mean ± SD||3.2 ± 3.0|| |
|Median (range)||2.5 (0.1-25.0)|| |
|RT|| || |
|Adjuvant RT to CW only||81||17.5|
|Adjuvant RT to CW + SCV||383||82.5|
|Adjuvant RT to CW + SCV + PABa||28||6.1|
The majority of patients (82.5%) received radiation to the SCV field, whereas 17.5% had the SCV field omitted. Table 3 summarizes use of the SCV field in the entire cohort and by clinical and pathologic lymph node-stage combinations. The omission of the SCV field was significantly more common in lower clinical and pathologic lymph node classification: 29.8% of 205 patients who had pathologically negative lymph node status after NAT had the SCV field omitted, whereas 46.3% of 80 patients who had clinically and pathologically negative lymph node status had the SCV field omitted. Delivery of both chest wall and SCV radiation also was associated significantly with increasing clinical stage group, clinical T-classification, lack of pCR in the breast and axilla, and lack of pCR in the axilla.
Table 3. Use of the Supraclavicular Field in the Entire Cohort and by Key Variables
|All patients||464||100||383||82.5||81||17.5|| |
|TN, n = 463|| || || || || || || |
|Negative for ER, PR, and HER2||149||100||121||81.2||28||18.8||.613|
|Other combinations||314||100||261||83.1||53||16.9|| |
|Clinical stage, n = 463|| || || || || || || |
|Clinical tumor classification|| || || || || || || |
|Clinical tumor size, cm|| || || || || || || |
|Mean ± SD||7.3 ± 4.1|| ||7.6 ± 4.3|| ||6.1 ± 2.8|| ||.0002|
|Median (range)||6 (0.5-32)|| ||7 (0.5-32)|| ||5 (1.5-15)|| || |
|Clinical LN status|| || || || || || || |
|No. of positive LNs, n = 462|| || || || || || || |
|Pathologic tumor size in 335 patients with tumors >0, cm|| || || || || || || |
|Mean ± SD||3.2 ± 3.09|| ||3.3 ± 3.1|| ||2.5 ± .9.0|| ||.014|
|Median (range)||2.5 (0.1-25)|| || || ||2.0 (0.1-9)|| || |
|Clinical and pathologic LN status|| || || || || || || |
|pCR in breast and axilla, n = 243|| || || || || || || |
|No pCR||373||100||315||84.5||58||15.5|| |
|pCR in breast, n = 243|| || || || || || || |
|No pCR||335||1000||282||84.2||53||15.8|| |
|pCR in axilla, n = 243|| || || || || || || |
|No pCR||259||100||239||92.3||20||7.7|| |
The median follow-up from the date of diagnosis was 45.9 months for all patients and 50.5 months for the patients who remained alive with no evidence of disease. In the whole cohort, the 4-year PFS rate was 76.5% (95% confidence interval, 71.9%-80.5%), and the OS rate was 86.3% (95% confidence interval, 82.3%-89.5%).
The cumulative 4-year incidence of LRR with or without synchronous DM as the first site of failure for the entire cohort was 5.8% (95% confidence interval, 3.8%-8.4%), which occurred in 25 patients. There were 25 LRRs as the first site of failure, including 12 isolated LRRs and 13 LRRs associated with synchronous distant failure (6 events were diagnosed simultaneously, 4 events were diagnosed <1 month apart, and 3 remaining events were diagnosed <3 months apart). The failure events are listed in Table 4. Among the total 25 LRRs, 16 were recurrences in the chest wall only, 3 were recurrences in the chest wall and regional lymph nodes, and 6 were regional lymph node recurrences only without chest wall failure. Among 9 regional failures, 6 were recurrences in the SCV field, and only 1 of those 6 recurrences developed in a patient who had the SCV field omitted. Among the 6 SCV failures, 4 occurred in the setting of inflammatory disease (the SCV field was treated in all 4 patients), 1 occurred in the setting of pathologic ypT3N2 disease (the SCV field was treated), and 1 occurred in a patient who had clinically positive lymph nodes but pathologically negative lymph nodes after NAT and had the SCV field omitted. Six patients had axillary failures, all 6 patients had received treatment to the SCV field, and 5 of these 6 patients had triple-negative disease.
There were significant differences in several prognostic factors for LRR (Table 5). On univariate analysis, predictors of LRR were higher clinical lymph node status (P = .025), positive pathologic lymph node status (P = .003), the combination of clinically and pathologically positive lymph node disease (P < .001), clinical stage III versus stage II (P = .038), inflammatory disease ay presentation (P = .037), negative ER status (P = .006), negative PR status (P = .015), triple-negative status (P < .001), and residual pathologic tumor size >2 cm (P = .045). Lack of a pCR in the breast and axilla was not predictive of LRR (P = .150), but lack of a pCR in the axilla alone was strongly predictive of LRR (P = .003). None of the 3 patients who had positive surgical margins have developed an LRR or distant failure.
Table 4. Failure Events
|LRR|| || |
|Type of LRR|| || |
|CW only||16 (SCV omitted in 3)||3.4|
|SCV only||3 (SCV omitted in 1)||0.6|
|AX and SCV||1||0.2|
|CW and AX and SCV||2||0.4|
|CW and AX||1||0.2|
|Distant failure|| || |
|Yes||87 (SCV omitted in 10)||18.8|
|First failure|| || |
|Simultaneous LRF & distant||13||2.8|
|Progression|| || |
|Vital status|| || |
Table 5. Forty-Eight–Month Cumulative Incidence of Locoregional Recurrence With or Without Distant Metastasis and of Isolated Distant Metastasis as First Failure by Selected Prognostic Factors
|All patients||5.8 (3.8-8.4)||—||15.2 (11.8-19.0)||—|
|ER|| || || || |
|Positive||2.9 (1.2-5.8)||.006||15.4 (10.7-20.8)||0.843|
|Negative||9.5 (5.8-14.1)|| ||14.8 (10.1-20.4)|| |
|PR|| || || || |
|Positive||2.2 (0.6-5.9)||.015||14.0 (8.8-20.5)||0.531|
|Negative||8.4 (5.4-12.3)|| ||15.0 (10.8-19.9)|| |
|HER2|| || || || |
|Positive||2.0 (0.4-6.5)||.054||18.6 (11.2-27.5)||0.573|
|Negative||7.1 (4.6-10.4)|| ||14.3 (10.5-18.6)|| |
|TN status|| || || || |
|Not negative for ER, PR, and HER2||2.6 (1.2-5.1)||<.001||16.0 (11.7-20.9)||.998|
|Negative for ER, PR, and HER2||12.8 (7.8-19.2)|| ||13.5 (8.2-19.9)|| |
|Clinical stage|| || || || |
|II||2.6 (0.7-6.9)||.038||8.5 (4.3-16.0)||.009|
|III||7.2 (4.6-10.6)|| ||18.2 (13.7-23.2)|| |
|Clinical tumor classification|| || || || |
|T1-T3||4.9 (2.8-7.9)||.172||11.7 (8.1-15.9)||< .001|
|T4||8.3 (4.2-14.1)|| ||23.9 (16.4-32.3)|| |
|Not T4d||4.9 (2.9-7.5)||.037||14.0 (10.4-18.1)||.038|
|T4d, inflammatory||12.1 (5.2-22.0)|| ||22.3 (12.5-34.0)|| |
|Clinical LN status|| || || || |
|N0, Nx||1.8 (0.3-5.7)||.025||9.7 (5.2-15.8)||.032|
|N1||6.7 (3.7-11.1)|| ||14.2 (9.4- 20.1)|| |
|N2-N3||9.2 (4.6-15.6)|| ||24.5 (15.8- 34.2)|| |
|N0,Nx||1.8 (0.3-5.7)||.010||9.7 (5.2-15.8)||.068|
|N1-N3||7.7 (4.9-11.1)|| ||17.8 (13.3- 22.7)|| |
|No. of positive LNs at surgery|| || || || |
|0: pN0||2.5 (0.8-6.0)||.005||8.6 (5.0-13.5)||< .001|
|1-3||6.3 (2.8-11.8)|| ||16.2 (9.6-24.4)|| |
|≥4||10.4 (5.8-16.5)|| ||23.9 (16.5-32.2)|| |
|Pcr in axilla|| || || || |
|Yes||2.5 (0.8-6.0)||.003||8.7 (5.0-13.5)||< .001|
|No||8.5 (5.3-12.5)|| ||20.3 (15.1-26.0)|| |
|pCR in breast|| || || || |
|Yes||3.1 (0.8-8.1)||.076||9.0 (4.1-16.4)||.003|
|No||6.9 (4.4-10.1)|| ||17.6 (13.4-22.2)|| |
|pCR in breast and axilla|| || || || |
|Yes||3.2 (0.6-10.1)||.150||4.2 (1.1-10.7)||.001|
|No||6.5 (4.2-9.5)|| ||17.7 (13.6-22.2)|| |
|Clinical and pathologic LN status|| || || || |
|cNx/cN0→pN0||1.7 (0.1-8.0)||.001||7.8 (2.8, 16.1)||.001|
|cNx/cN0→pN+||1.8 (0.1-8.5)|| ||12.0 (5.2-21.8)|| |
|cN+→pN0||3.0 (0.8-7.8)|| ||8.2 (4.5-15.0)|| |
|cN+→pN+||10.7 (6.7-15.8)|| ||23.2 (16.8-30.2)|| |
|RT fields|| || || || |
|CW only||5.8 (1.8-13.3)||.832||7.7 (2.7-16.0)||.083|
|CW + SCV||5.8 (3.7-8.7)|| ||16.7 (12.9-21.2)|| |
|Pathologic tumor size, cm|| || || || |
|≤2||3.9 (1.9, 7.0)||.045||11.8 (7.9-16.6)||.003|
|>2||8.7 (5.1-13.5)|| ||20.0 (14.1-26.5)|| |
On univariate analysis, omission of the SCV field was not associated with the risk of LRR (Gray test: P = .832 [Table 5]; Gray and Fine test: HR, 0.89; P =. 0.833 [Table 6]). On multivariate analyses, omission of the SCV field was associated significantly with the risk of LRR. The HR was 3.39 (P = .024) in a model that was adjusted for positive pN status (HR, 10.23; P < .0001), triple-negative status (HR, 8.50; P < .0001), clinical stage III disease (2.56; P = .108), and inflammatory presentation (HR, 1.41; P = .460).
Table 6. Effect of Key Prognostic Factors on the Risk of Locoregional Recurrence With or Without Distant Metastasis and on the Risk of Isolated Distant Metastasis as First Failure
|LRR with or without DM|| || |
|Univariate model|| || |
|RT to CW only vs RT to CW + SCV||0.89 (0.31-2.58)||.833|
|Multivariate model|| || |
|RT to CW only vs RT to CW + SCV||3.39 (1.17-9.82)||.024|
|pN + vs pN0: No pCR vs pCR in axilla||10.23 (3.19-32.78)||< .0001|
|TN vs other||8.50 (3.48-20.79)||< .0001|
|Stage III vs II||2.56 (0.81-8.07)||.108|
|T4d vs other||1.41 (0.56-3.55)||.460|
|Isolated DM|| || |
|Univariate model|| || |
|RT to CW only vs RT to CW + SCV||0.51 (0.24-1.11)||.089|
|Multivariate model|| || |
|RT to CW only vs RT to CW + SCV||0.69 (0.31-1.54)||.360|
|No pCR vs pCR in breast and axilla||5.15 (1.59-16.66)||.006|
|TN vs other||1.15 (0.66- 2.01)||.620|
|Stage III vs II||1.87 (0.96-3.67)||.067|
|T4d vs other||1.35 (0.74-2.47)||.320|
The cumulative 4-year incidence of isolated DM as the first site of failure for the entire cohort was 15.2% (95% confidence interval, 11.8%-19%), occurring in 68 patients. Overall, there were 87 distant failures, including 13 failures that were synchronous with LRR, 6 failures after LRR, 68 patients who had isolated DM as the first failure, and a total of 68 deaths. There were significant differences in the cumulative incidence of isolated DM by several prognostic factors (Table 5). In univariate analysis, predictors of DM as the first failure included higher clinical stage (P = .009), inflammatory disease at presentation (P = .038), positive pathologic lymph node status (P < .001), the combination of clinically and pathologically positive lymph node disease (P = .001), lack of a pCR in the breast and axilla clinical (P = .001), and residual pathologic tumor size >2 cm (P = .003). Also in univariate analysis, omission of the SCV field was associated marginally with the risk of isolated DM as the first failure (Gray test: P = .083 [Table 5]; HR, 0.51; Gray and Fine test: P = .089) (Table 6). However, omission of the SCV field lost its significance for predicting isolated DM (HR, 0.69; P = .360) in a model that was adjusted for lack of a pCR in the breast and axilla (HR, 5.15; P = .006), triple-negative status (HR, 1.15; P = .62), clinical stage III disease (HR, 1.87; P = .067), and inflammatory presentation (HR, 1.35; P = .32). Omission of the SCV field was not associated with PFS (P = .078) or OS (P = .462) on univariate analysis.
In this large retrospective series of patients with stage II and III breast cancer who uniformly received NAT, underwent mastectomy, and received PMRT, we identified multiple risk factors for LRR, including negative ER status, negative PR status, triple-negative disease, advanced clinical stage, and poor response to NAT. On univariate analysis, omission of the SCV field was not associated with an increased risk of LRR. However, omission of the SCV field became significantly associated with an increased risk of LRR in multivariate analyses with adjustment for other established risk factors, including triple-negative status, pCR in axilla, advanced stage at presentation, and inflammatory disease.
Triple-negative receptor status is emerging as a clear negative prognostic indicator for LRR as well as PFS and OS. An analysis of the Danish postmastectomy trials demonstrated that the locoregional control benefit of PMRT was less in patients with triple-negative tumors,16 and our own series demonstrated a higher risk of LRR after PMRT in patients with triple-negative tumors.17 It remains unclear whether triple-negative breast cancer is a risk factor specifically for lymph node recurrence. Studies suggest that lymph node-positive disease at presentation is more common in patients with triple-negative breast cancer, and a retrospective series revealed that patients with triple-negative tumors had a higher risk of isolated regional lymph node failure in the setting of early stage breast cancer and breast-conserving therapy.18 In our series, of 9 regional failures, 7 occurred in patients with triple-negative disease. Because of the small number of regional events, we were unable to assess whether triple-negative breast cancer was a significant predictor of regional failure in particular, but the data overall do suggest that this is an entity with a high risk of recurrence in all forms, and omission of the SCV field in this patient subset should be avoided. However, because it appears that PMRT is less effective in the setting of triple-negative disease, it is possible that comprehensive PMRT alone is not sufficient. It is noteworthy that, in the setting of treatment with trastuzumab for the majority of patients with HER2-overexpressing tumors, HER2 overexpression did not emerge as a risk factor for LRR.
The role of regional lymph node irradiation in breast cancer has become increasingly controversial. The recently published American College of Surgeons Oncology Group (ACOSOG) Z0011 study assessed the role of completion axillary dissection in patients with very early stage, low-risk breast cancer who had clinically negative lymph nodes and 1 or 2 positive sentinel lymph nodes; those patients received adjuvant radiation to the whole breast using supine tangent fields (generally inclusive of a substantial component of the level I/II axillary lymph nodes), but not to the high axillary or supraclavicular lymph nodes. In that study, there was no difference in locoregional outcome between patients who underwent completion axillary dissection and those who did not, and the locoregional failure rates were very low despite omission of SCV radiation.19 A recent publication from France evaluated the locoregional impact of omission of the SCV field in patients who were lymph node-negative at the time they underwent breast-conserving surgery after NAT, regardless of the initial clinical lymph node status. Those investigators observed no difference in locoregional outcome or death with and without the receipt of regional lymph node radiation.20 Conversely, the National Cancer Institute of Canada (NCIC) MA.20 study, which randomized patients who had more advanced disease than the patients in ACOSOG Z0011 (the majority of had 1-3 positive lymph nodes) to receive whole breast radiation with or without regional lymph node irradiation, demonstrated a decreased risk of LRR and a trend toward improvement in OS with the addition of regional lymph node irradiation.13 It is critical to note that the ACOSOG Z0011 and NCIC MA.20 studies are not directly applicable to the patient population reviewed in our current study, because our patients had more locally advanced disease than the patients in those 2 studies, and they received NAT: Positive lymph node status after NAT has clearly emerged as a more serious prognostic indicator than positive lymph node status at the time of up-front surgery.
However, as techniques for radiation delivery improve and advances are made in systemic therapies, the omission of radiation to “elective” fields has become more common in disease sites, like the lung, pancreas, and others, to maximize the therapeutic ratio of treatment. The toxicity of regional lymph node irradiation, such as skin/cosmetic toxicities; the risk of brachial plexopathy, pneumonitis, lymphedema12, 13; and the risk of cardiac/cerebrovascular mortality are quite low. Nonetheless, the omission of regional lymph node irradiation is desirable if it does not offer patients a significant benefit. However, our series suggests that treatment of regional lymph nodes may have an impact on locoregional outcome. We were not able to identify a subset of patients who received NAT and underwent mastectomy and could safely received treatment to the chest wall only with omission of the SCV field. Although the rate of LRR was low at 5.8% in the entire cohort, and there was a very low LRR rate (1.9%) in patients who had pathologically negative lymph nodes after NAT, our multivariate analysis demonstrated that omission of the SCV field had a negative impact on the risk of LRR in our cohort. The potential benefit of regional lymph node radiation is a reduction in the distant failure rate, in addition to LRR, as demonstrated in the NCIC MA.20 study.13 Regional lymph node radiation in our series was not associated with distant failure and did not have an impact on PFS, possibly because of the small numbers of events.
Our study included only patients who received PMRT; thus, we were not able to assess the overall benefit of PMRT in this patient population. Although guidelines for the use of PMRT after up-front surgery are relatively well established, guidelines after NAT are less clear. Currently, common indications for PMRT after NAT include a clinical tumor size >5 cm, clinically positive lymph nodes, and/or residual lymph node disease after NAT; generally, all patients with clinical stage III disease who receive NAT are offered PMRT, and patients with stage II disease usually are offered PMRT, depending on other risk factors. The low rate of LRR in patients who achieved an axillary pCR in our series, as well as that observed in other series,11 suggests the possibility that PMRT may be avoided altogether in patients who have a pCR to NAT, and trials designed to assess possibility this are currently being developed by cooperative groups.
It is important to note that, during the period covered in this review, breast cancer staging was determined clinically, and pathologic assessment of axillary lymph nodes before chemotherapy was not routinely performed. At our center, we do not perform sentinel lymph node biopsy before NAT. Since 2009, we have adopted ultrasound staging with fine-needle aspiration or core biopsy of clinically or radiographically suspicious lymph node(s) before the initiation of NAT. Although this approach has a lower sensitivity than sentinel lymph node biopsy, ranging from 28% to 56%,21, 22 it does not have an impact on subsequent surgical staging and does not delay treatment. In addition, regional lymph node recurrence is likely underestimated in this series, because dedicated imaging is required to identify clinically occult recurrences and was not routinely performed in our patients.
This series is limited by its reliance on retrospective data and small patient numbers in each category of lymph node status and SCV field use combinations. We view the current data as a preliminary step toward well designed, prospective studies that will assess the tailoring of radiation field selection to patients individually, with hormone and HER2 status combinations as well as lymph node response to NAT as factors in clinical decision-making pathways, to determine whether any particular subset of patients receiving NAT can safely have the SCV field omitted from their PMRT treatment.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.