Excellent Locoregional Control in Inflammatory Breast Cancer With a Personalized Radiation Therapy Approach

Practical Radiation Oncology (2019) xx, e1-e8

www.practicalradonc.org

Basic Original Report

Excellent Locoregional Control in Inflammatory Breast Cancer With a Personalized Radiation Therapy Approach Shane R. Stecklein MD, PhD a, Kelly J. Rosso MD b,c, Jenny Nuanjing BS, CMD a, Audree B. Tadros MD, MPH b,d, Anna Weiss MD b,e, Sarah M. DeSnyder MD b, Henry M. Kuerer MD, PhD b, Mediget Teshome MD, MPH b, Thomas A. Buchholz MD a,f, Michael C. Stauder MD a, Naoto T. Ueno MD, PhD g,h, Anthony Lucci MD b,h, Wendy A. Woodward MD, PhD a,h,* a

Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas; bDepartment of Breast Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas; cBanner MD Anderson Cancer Center, Gilbert, Arizona; dDepartment of Surgery, Memorial Sloan Kettering Cancer Center, New York, New York; eDepartment of Surgical Oncology, Brigham and Women’s Faulkner Breast Center and Dana-Farber Cancer Institute, Boston, Massachusetts; fScripps MD Anderson Cancer Center, La Jolla, California; gDepartment of Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas; and hMorgan Welch Inflammatory Breast Cancer Research Program and Clinic, The University of Texas MD Anderson Cancer Center, Houston, Texas

Received 11 February 2019; revised 15 May 2019; accepted 20 May 2019

Abstract Purpose: Inflammatory breast cancer (IBC) has been characterized by high locoregional recurrence (LRR) rates even after trimodality therapy. We recently reported excellent locoregional control among patients treated since formal dedication of an IBC-specific clinic and research program in 2006. Institutionally, a standard twice-daily (BID) dose escalation regimen for all patients with IBC was de-escalated in select cases in 2006 after review demonstrated that young age, incomplete response to neoadjuvant therapy, and positive margins identified subsets with maximal benefit from dose escalation. We report local control and toxicity rates specific to BID versus once-daily (QD) radiation therapy approaches. Methods and Materials: From a prospectively collected database, we identified 103 patients with nonmetastatic IBC who received trimodality therapy at our institution from 2007 to 2015. Descriptive statistics were used to describe the study cohort and compare retrospectively extracted

Supported in part by Cancer Center Support Grant CA016672 from the National Cancer Institute, National Institutes of Health, to The University of Texas MD Anderson Cancer Center. Disclosures: None. * Corresponding author. E-mail address: [email protected] (W.A. Woodward). https://doi.org/10.1016/j.prro.2019.05.011 1879-8500/Ó 2019 American Society for Radiation Oncology. Published by Elsevier Inc. All rights reserved.

e2

S.R. Stecklein et al

Practical Radiation Oncology: --- 2019

rates of radiation therapyeassociated toxicity. The actuarial rate of LRR-free survival was analyzed using the Kaplan-Meier method. Results: The median follow-up is 3.6 years. Thirty-nine patients (37.9%) received postmastectomy radiation therapy (PMRT) to the chest wall and undissected regional lymphatics in QD fractions (median dose, 50.0 Gy in 25 fractions [fx]; median boost dose, 10.0 Gy in 5 fx) and 64 patients (62.1%) received BID PMRT (median dose, 51.0 Gy in 34 fx; median boost dose, 15.0 Gy in 10 fx). Crude rates of toxicity were not different between patients treated with QD or BID PMRT. Two BID patients (3.1%) and no QD patients (0.0%) experienced LRR (P Z .53). The 3- and 5year LRR-free survival were 95.1% and 100.0% for BID and QD patients, respectively (P Z .25). Conclusions: Tailoring radiation therapy to clinical risk factors was associated with excellent locoregional control. De-escalation of PMRT from BID to QD was not clearly associated with reduced toxicity compared with BID, although retrospective data collection may limit this comparison. Ó 2019 American Society for Radiation Oncology. Published by Elsevier Inc. All rights reserved.

Introduction Inflammatory breast cancer (IBC) is an uncommon highly aggressive form of breast cancer characterized by high rates of locoregional recurrence (LRR) and early metastatic dissemination. Approximately 70% of patients with IBC have locoregionally confined disease at diagnosis and undergo curative-intent trimodality therapy (neoadjuvant systemic therapy, modified radical mastectomy, and postmastectomy radiation therapy [PMRT]). Before this treatment paradigm, outcomes for IBC were dismal with LRR rates of 30% to 70% and 5-year survival rates as low as 6%.1,2 Previous studies have demonstrated that patients with IBC who receive trimodality therapy have significantly improved locoregional control and survival.3 Improvements in cytotoxic chemotherapy and incorporation of targeted agents are expected to increase pathologic complete response (pCR) rates and lead to improved locoregional control. Currently, only a few retrospective studies have described the association between pCR and risk of LRR in IBC.4-6 A crude weighted average of published institutional series spanning several decades puts expected LRR after PMRT in patients with IBC at approximately 20%.6 Various modifications to PMRT have been used to improve locoregional control in IBC. These have included more liberal use of tissue-equivalent bolus, overlapping field junctions, and frank escalation of nominal prescription dose with or without acceleration or hypofractionation.4,7,8 Researchers at our institution began to prescribe dose-escalated accelerated hyperfractionated radiation therapy in 1985 in an attempt to improve locoregional control in IBC. A previous analysis of our experience with once-daily (QD) versus twice-daily (BID) doseescalated radiation therapy showed that dose-escalated accelerated hyperfractionated radiation therapy selectively benefitted patients <45 years old; those with positive, close, or unknown margin status; and those who had less than a partial response to neoadjuvant systemic

therapy, defined as <50% reduction in the clinical skin changes and bidirectional tumor measurements using clinical and radiographic means.9 These results prompted a more personalized radiation therapy dose and fractionation approach for each patient based on these factors. Our recent report of patients treated with contemporary neoadjuvant systemic therapy, surgery, and tailored IBCspecific PMRT described excellent locoregional control with 2-year risk of LRR of 3% and 5-year overall survival of 69%.10 In the present study, we examined details of the radiation therapy, locoregional control among those who were offered de-escalation with QD PMRT, and, to the extent possible, acute and late toxicity between QD and BID PMRT approaches.

Methods and Materials Patient selection An institutional review boardeapproved retrospective review of all patients with IBC in a prospectively collected database of patients treated at The University of Texas MD Anderson Cancer Center from 2007 through 2015 was performed. Patients received a diagnosis of IBC from a multidisciplinary team per contemporary guidelines.11 After establishing a clinical diagnosis, confirmatory tissue biopsy and imaging were performed. Baseline imaging studies included bilateral mammography, breast and nodal ultrasound, and positron emission tomography/ computed tomography to evaluate for metastatic disease. Medical photography was used to document extent of disease before treatment. A total of 277 patients with IBC and complete data were identified during this period. Patients with stage IV disease at presentation or who developed distant metastatic disease during treatment and thus did not complete trimodality therapy, those with <60-day follow-up, and those who did not receive radiation therapy at our institution or only received palliative radiation therapy were excluded. A total of 103 patients

Practical Radiation Oncology: --- 2019

Personalized radiotherapy in IBC

e3

were included in the final analyses. None of these patients were included in our earlier publications regarding LRR with hyperfractionated radiation therapy.9,12

Treatment All patients were evaluated in a multidisciplinary dedicated IBC clinic by a breast medical oncologist, breast surgical oncologist, and breast radiation oncologist. Clinical response to neoadjuvant systemic therapy (complete response, partial response, minor response, or stable disease) was determined by imaging and defined in accordance with the WHO Response Classification.13 Breast and nodal pCR in this study was defined as no invasive carcinoma in breast specimen and lymph nodes, respectively. Patients with hormone receptorepositive tumors and triple-negative tumors received AC (cyclophosphamide, doxorubicin) followed by weekly paclitaxel. Carboplatin was added in certain cases of triple-negative tumors. All patients with HER2 amplification received targeted therapy (trastuzumab, pertuzumab, and/or lapatinib) in addition to a paclitaxel-based regimen in the neoadjuvant setting. All patients in this study underwent modified radical mastectomy with the intent of obtaining negative margins and complete level I and II axillary nodal dissection. Level III nodes were dissected in cases in which baseline axillary ultrasound identified suspicious level III nodes. Wide resection of skin was performed to optimize chances for negative margins. If primary closure was not possible, skin grafts or myocutaneous flap closures were used. Skin-sparing mastectomy or immediate reconstruction were not performed. Patients received PMRT to the chest wall, undissected axilla, supraclavicular nodes, and internal mammary nodes. Pretreatment medical photography and imaging were used to design radiation fields. Regional nodal irradiation including the supraclavicular and internal mammary lymph node basins was prescribed in all cases. It is our practice to contour the undissected axilla, supraclavicular fossa, and internal mammary chain in the first 3 intercostal spaces. The treatment planning goal was for the 90% isodose line to cover all nodal target volumes. In cases in which upfront regional nodes were positive and evident on cross sectional imaging, care was taken to review these and extend dose coverage as needed to provide margin on grossly involved but unresected regional nodal basins. Patients were treated with either QD radiation therapy (median dose, 50.0 Gy in 25 fractions; median boost dose, 10.0 Gy in 5 fractions) or accelerated hyperfractionated radiation therapy in BID fractions (median dose, 51.0 Gy in 34 fractions; median boost dose, 15.0 Gy in 10 fractions) depending on age, surgical margin status, and response to neoadjuvant systemic therapy.9 BID fractionation was recommended for all patients younger than 45 years; those with close, positive, or unknown surgical

Figure 1 Postmastectomy radiation therapy technique. (A) Typical electron boost fields encompassing the mastectomy scar, drain sites, and chest wall flaps. (B) Chest wall, regional nodal, and boost fields for patients with undissected or residual nodal disease after neoadjuvant systemic therapy. Nondivergent tangential photon fields (red field, left panel) are matched medially to en face appositional electron fields covering the internal mammary lymph nodes (blue and pink fields, left panel) and superiorly to an oblique photon supraclavicular field (yellow field, left panel). The orange inset in the supraclavicular field is a photon supplement used for this particular patient to ensure adequate coverage of the supraclavicular nodal volume. This technique is used if the primary supraclavicular field fails to provide adequate deep coverage. Chest wall boost fields are designed as described earlier (aqua and purple fields, right panel), and involved but undissected lymph node basins are boosted with electrons or photons (green, blue, and red fields, right panel). For this particular patient, the supraclavicular lymph nodes were boosted with an oblique photon field (blue field, right panel) and supplements were again used to improve deep coverage in 2 areas (red insets in blue field, right panel). The internal mammary nodal chain was boosted with an en face appositional electron field (green, right panel). For involved but undissected nodal volumes that completely respond to neoadjuvant systemic therapy, the cumulative target dose is 60 Gy. For undissected nodal volumes with residual disease after neoadjuvant systemic therapy the cumulative target dose is 66 Gy.

margins; and those with less robust responses to neoadjuvant systemic therapy. Although we did not use any quantitative residual disease burden metrics to make decisions regarding fractionation, our general approach, based on our previous experience, was to recommend QD radiation therapy for patients who achieved a pCR or who had only minimal residual breast disease and to recommend BID radiation therapy for patients with more than

e4

Practical Radiation Oncology: --- 2019

S.R. Stecklein et al

Table 1

Planning goals

Structure/volume Primary field targets Chest wall Axillary apex/undissected axilla Supraclavicular lymph nodes Internal mammary lymph nodes Boost field targets Chest wall Nodal boost volumes Organs at risk Heart Ipsilateral lung Spinal cord Esophagus Brachial plexus

Goal Ensure that the 98% isodose line does not break up Ensure coverage by the 90% isodose line Ensure coverage by the 90% isodose line Ensure coverage by the 90% isodose line Ensure coverage by the 90% isodose line Ensure coverage by the 90% isodose line Mean dose <5 Gy V20 <35% Dmax <45 Gy Dmax <45 Gy Dmax <54 Gy (primary fields) Composite* mean dose <60 Gy Composite* max dose <68 Gy

* Plan integrating the dose from primary fields and all boost fields.

minimal residual breast disease or any residual nodal disease. Eleven patients were recommended BID fractionation (5 for age <45 years, 5 for less than partial response to neoadjuvant systemic therapy, and 1 for less than partial response to neoadjuvant systemic therapy and close surgical margin) but ultimately received QD treatment owing to patient preference. Tissue-equivalent bolus (3 mm) was used on the chest wall, including any portion overlying the infraclavicular fossa in the supraclavicular field, for all cases. For BID fractionation, bolus was prescribed with each fraction for the first 5 days, then QD for the second 5 days, then as needed to achieve brisk skin erythema. For QD fractionation, bolus was prescribed with every other fraction for the first 10 days, then as needed to achieve brisk erythema. Field junctions were double-treated with a 3 mm overlap to ensure that all skin in the field received adequate dose. The chest wall was boosted in all cases. Chest wall boost volumes were drawn to cover the surgical bed including the mastectomy flaps, drain sites, and scar with adequate dosimetric margin (Fig 1A). Regional nodes were boosted in any patient with regional nodal involvement at presentation that was outside of the regions targeted by lymph node dissection (Fig 1B). For residual nodal disease after neoadjuvant systemic therapy, the intent was to boost the involved but undissected nodal basin(s) to 66 Gy. In the case of a radiographic complete

response to neoadjuvant systemic therapy, the undissected nodal basin(s) were boosted to 60 Gy. Planning goals for targets and organs at risk are listed in Table 1.

Statistical analysis All patients were analyzed according to treatment regimen received. Fisher exact test was used to compare descriptive statistics between groups. Actuarial LRR was calculated from the end date of PMRT using the KaplanMeier method. Toxicity was retrospectively collected from radiation treatment summaries and follow-up notes. Where appropriate and possible, toxicity was retrospectively scored according to Common Terminology Criteria for Adverse Events version 4.0 criteria. All statistical tests are 2-tailed. Analyses were performed in SPSS (IBM Corp., Armonk, NY, version 24).

Results Patient and treatment characteristics A total of 103 patients with stage IIIB (55.3%) or IIIC (44.7%) disease were analyzed. The median follow-up was 3.6 years (range, 0.8-8.6 years). The majority of patients were white (79.6%) and peri- or postmenopausal (63.1%). Surgical margins <2 mm were uncommon (3.9%), and no patients had a positive margin. Breast and nodal pCR were observed in 26.2% and 37.6% of patients, respectively. Demographic, tumor, and treatment characteristics for all patients and by PMRT regimen are detailed in Table 2. As expected, patients in the BID treatment group had more aggressive disease characteristics, including more positive metastatic lymph nodes (P < .0001) and higher rates of lymphovascular space invasion (P Z .002), and lower rates of pathologic response in both the breast (P < .0001) and lymph nodes (P < .0001).

Locoregional recurrence There have been 2 LRRs in the BID PMRT patients and no LRRs in the QD PMRT patients during the period of observation. One patient had isolated LRR in the axilla and 1 patient had LRR in the axilla and supraclavicular lymph nodes. The 4-year probability of LRR was 4.9% for BID PMRT patients and 0.0% for QD PMRT patients (Fig 2A; P Z .26). It was not possible to examine predictors of LRR given the small number of events.

Overall survival The median overall survival (OS) was 6.4 years for BID PMRT patients and was not reached for QD PMRT patients. Five-year overall survival was 55.9% for BID

Practical Radiation Oncology: --- 2019 Table 2

e5

Personalized radiotherapy in IBC

Patient and treatment characteristics

Characteristics Age (median), y Race White, non-Hispanic Hispanic Black Asian Other Menopausal status Premenopausal Peri- or postmenopausal Clinical N stage cN0 cN1 cN2 cN3 Clinical stage IIIB IIIC Tumor subtype HRþ/HER2e HRþ/HER2þ HRe/HER2þ HRe/HER2e Days to surgery (median) Lymphovascular space invasion Negative Positive Surgical margins Negative Close (<2 mm) Positive (tumor on ink) No. of nodes removed (median) Breast pCR No Yes Nodal pCR* No Yes No. of positive nodes (median) Days to radiation therapy (median) Adjuvant chemotherapy No Yes Adjuvant hormone therapyy No Yes

All patients (N Z 103)

QD PMRT (n Z 39)

BID PMRT (n Z 64)

52 (22-78)

52 (39-78)

52 (22-78)

82 12 5 1 3

32 4 2 0 1

50 8 3 1 2

(79.6) (11.7) (4.9) (1.0) (2.9)

(82.1) (10.3) (5.1) (0.0) (2.6)

P Value .29 .98

(78.1) (12.5) (4.7) (1.6) (3.1) .09

38 (36.9) 65 (63.1)

10 (25.6) 29 (74.4)

28 (43.8) 36 (56.3)

2 47 7 47

0 17 3 19

2 30 4 28

.83 (1.9) (45.6) (6.8) (45.6)

(0.0) (43.6) (7.7) (48.7)

(3.1) (46.9) (6.3) (43.8) .55

57 (55.3) 46 (44.7)

20 (51.3) 19 (48.7)

37 (57.8) 27 (42.2) .07

40 19 26 18 216

(38.8) (18.4) (25.2) (17.5) (140-389)

10 7 15 7 222

(25.6) (17.9) (38.5) (17.9) (156-389)

30 12 11 11 214

(46.9) (18.8) (17.2) (17.2) (140-296)

48 (46.6) 55 (53.4)

26 (66.7) 13 (33.3)

22 (34.4) 42 (65.6)

99 4 0 20

38 1 0 22

61 3 0 20

.96 .002

1.00 (96.1) (3.9) (0.0) (4-88)

(97.4) (2.6) (0.0) (5-88)

(95.3) (4.7) (0.0) (4-50)

76 (73.8) 27 (26.2)

20 (51.3) 19 (48.7)

56 (87.5) 8 (12.5)

63 38 3 39

15 24 1 37

48 14 5 41

.96 <.0001 <.0001

(62.4) (37.6) (0-29) (16-114)

(38.5) (61.5) (0-29) (17-59)

(77.4) (22.6) (0-23) (16-114)

75 (72.8) 28 (27.2)

24 (61.5) 15 (38.5)

51 (79.7) 13 (20.3)

8 (13.6) 51 (86.4)

4 (23.5) 13 (76.5)

4 (9.5) 38 (90.5)

<.0001 .63 .07

.21

Abbreviations: BID Z twice daily; pCR Z pathologic complete response; QD Z once daily. * In clinically node-positive patients before neoadjuvant systemic therapy. y In HR þ patients.

PMRT patients and 84.0% for QD PMRT patients (Fig 2B; P Z .04). The inferior survival in the BID PMRT patients is attributed to more aggressive disease characteristics and poorer responses to neoadjuvant systemic therapy.

Toxicity Radiation dermatitis and infection All patients developed brisk erythema and variable degrees of dry or moist desquamation. Because of the

Practical Radiation Oncology: --- 2019

S.R. Stecklein et al

B Locoregional Recurrence (%)

A 100

100

QD PMRT BID PMRT

75 50 25 0

P=0.26 0

2

4

6

8

10

Overall Survival (%)

e6

75 50 25 0

QD PMRT

0

2

Time (Years) 39 64

22 32

14 19

10 5

P=0.04

BID PMRT 4

6

8

10

2 2

0 0

Time (Years) 0 2

0 0

39 64

30 44

18 20

12 8

Figure 2 Kaplan-Meier curves for locoregional recurrence (A) and overall survival (B). Numbers below graphs denote number remaining at risk at each interval.

inherent limitations in objectively scoring retrospective descriptions of toxicity, we are unable to quantify or grade dermatitis with acceptable reliability. In our experience, patients treated with BID fractionation tend to have more intense acute dermatitis than those treated with QD fractionation, although the dermatitis is clinically manageable in nearly all cases. Treatment-limiting severe radiation dermatitis was only observed in 1 patient. She had extensive residual breast and nodal disease after neoadjuvant systemic therapy that would normally warrant BID fractionation. However, a decision was made to treat with concurrent capecitabine, and she was treated with QD PMRT with a planned boost to 66 Gy. Because of severe dermatitis in the sixth week of treatment, the final 3 fractions were omitted and her radiation therapy was discontinued after 60 Gy. Chest wall infection was observed in 1 QD PMRT patient and 2 BID PMRT patients (2.6% in QD PMRT vs 3.1% in BID PMRT; P Z 1.00). Symptomatic lymphedema and fibrosis Posttreatment lymphedema leading to recorded intervention including sleeve use, manual lymphatic drainage, or lymphovascular bypass surgery was observed in 10 QD PMRT patients and 8 BID PMRT patients (25.6% in QD PMRT vs 12.5% in BID PMRT; P Z .11). A single case of symptomatic fibrosis was observed in a QD PMRT patient (2.6%). Pneumonitis and brachial plexopathy Symptomatic pneumonitis was observed in 1 QD PMRT patient and 5 BID PMRT patients (2.6% for QD PMRT vs 7.8% for BID PMRT; P Z .40). For BID patients who developed pneumonitis, the mean ipsilateral lung dose was 17.49  1.97 Gy, and the ipsilateral lung V20 was 35.2%  4.6%, exclusive of lung contributions from the chest wall boost because these were clinically planned during the period of study. No cases of brachial

plexopathy were noted in QD PMRT patients and 2 cases were reported in BID PMRT patients (3.1%). Both of these occurred with a maximum plexus dose of <66 Gy and presented as weakness and numbness in the ipsilateral extremity. The first case occurred in a patient with a leftsided cT4dN2aM0 ypT4dN2aM0 cancer whose left brachial plexus received a maximum point dose of 65.11 Gy. She noted onset of dermatomal left arm numbness approximately 1 year after completing radiation therapy. Magnetic resonance imaging showed increased T2 signal surrounding the left brachial plexus, but no infiltration or mass. At her last follow-up, approximately 7 years after completing radiation therapy, her numbness is markedly improved and she has no functional limitations. The second case of plexopathy occurred in a patient with a right-sided cT4dN3aM0 ypT0N0M0 cancer whose right brachial plexus received a maximum point dose of 57.95 Gy. The patient noted occasional mild right arm pain and numbness starting approximately 9 months after completing radiation therapy, with acute and marked worsening of right arm pain and numbness after undergoing a trans rectus abdominus muscle flap reconstruction with synchronous lymphovascular bypass 11 months after PMRT. Magnetic resonance imaging showed slight thickening and edema of the right brachial plexus, potentially consistent with post radiation therapy damage. At approximately 4.5 years after radiation therapy, the patient continues to have severe right-sided plexopathy and was referred for hyperbaric oxygen therapy and ongoing physical therapy.

Discussion We demonstrated that tailoring radiation therapy to clinical risk factors was associated with excellent locoregional control in patients with IBC treated with neoadjuvant systemic therapy and modified radical

Practical Radiation Oncology: --- 2019

mastectomy. LRR of IBC is highly morbid, and patients with cutaneous or soft-tissue recurrence of IBC have few treatment options. Accordingly, locoregional control is critically important in both the curative and palliative management of IBC, and strategies to simultaneously prevent locoregional disease recurrence and minimize treatment-related morbidity are imperative. Rates of LRR in IBC have declined over time with improvements in neoadjuvant systemic therapy, operative techniques, and IBC-specific modifications to PMRT, including liberal use of tissue-equivalent bolus, overlapping field junctions, and modifications to nominal prescription dose or fractionation.4,7,8 We recently reported outcomes for a cohort of 114 patients with IBC treated with neoadjuvant systemic therapy, modified radical mastectomy, and postmastectomy radiation therapy.10 We observed a 4-year LRR probability of 5.7%, which is lower than in older studies of trimodality therapy in IBC.14,15 Our previous analysis showed that radiation therapy intensification selectively benefitted patients <45 years old; those with positive, close, or unknown margin status; and those who had less than a partial response to neoadjuvant systemic therapy. We have largely used these criteria to determine which patients receive QD and BID PMRT.9 Our present analysis on an independent cohort of patients demonstrates that tailored deintensification of radiation therapy based on these clinical and pathologic parameters is not associated with inferior locoregional control. Tailoring PMRT based on patient and disease characteristics can be used both to increase the likelihood of disease control and to reduce treatment-related toxicity. In the present study we observed excellent locoregional control in both QD and BID PMRT patients, suggesting that our stratification based on patient age, surgical margin status, and response to neoadjuvant systemic therapy is appropriate. Because of the low number of LRR events in our study we are unable to identify additional clinical or pathologic features that may improve our ability to define patients who are best served by dose escalation or deescalation. Severe treatment-related toxicity was uncommon with both QD and BID PMRT, and we did not observe markedly less radiation therapyeassociated acute or late toxicity with the less intensive QD regimen. The retrospective nature of our study and lack of objective toxicity scoring may limit our ability to identify differences in treatment-related adverse effects and highlight the merit of prospective data collection. Although our median followup of 3.6 years is relatively short, our previous report with a median follow-up of >5 years demonstrated that nearly all observed locoregional failures occurred within the first 3 years.9 We therefore expect that the observed rates of locoregional control are likely to be durable with longer follow-up. Logistical considerations may also be important in determining whether to treat patients with QD or BID

Personalized radiotherapy in IBC

e7

fractionation. Although our typical BID regimen results in a shorter overall treatment time (22 treatment days for BID compared with 30 treatment days for QD), BID treatments require a minimum 6-hour interfraction interval and can thus be very disruptive for patients’ daily routines. The effect of fractionation on patient-specific factors such as transportation, employment, and child care may need to be considered when determining fractionation. Although we strongly favor BID fractionation for patients <45 years old; those with positive, close, or unknown margin status; and those who had less than a partial response to neoadjuvant systemic therapy, there are alternative methods to intensify treatment that remain compatible with QD fractionation.4,7,8 All patients in this study were treated at our highvolume institution by specialized breast medical, surgical, and radiation oncologists with expertise in the management of IBC. Patients in this cohort were also healthy enough to tolerate aggressive trimodality therapy and had operable disease at presentation or after neoadjuvant systemic therapy and were thus offered curative-intent treatment. Our findings may not similarly apply in settings that are less familiar with IBC-specific trimodality therapy or in patients with more aggressive disease.

Conclusions Despite excellent locoregional control in our cohort, many patients developed metastatic disease and died of IBC. Improvements in systemic therapy are desperately needed, considering distant failure remains the major driver of IBC-associated mortality. As rates of distant metastasis improve, durable locoregional control and mitigation of locoregional therapyeassociated toxicity will remain important determinants of long-term cure and survivorship.

References 1. Droulias CA, Sewell CW, McSweeney MB, Powell RW. Inflammatory carcinoma of the breast: A correlation of clinical, radiologic and pathogic findings. Ann Surg. 1976;184:217-222. 2. Perez CA, Fields JN. Role of radiation therapy for locally advanced and inflammatory carcinoma of the breast. Oncology (Williston Park). 1987;1:81-94. 3. Fields JN, Kuske RR, Perez CA, Fineberg BB, Bartlett N. Prognostic factors in inflammatory breast cancer. Univariate and multivariate analysis. Cancer. 1989;63:1225-1232. 4. Brown L, Harmsen W, Blanchard M, et al. Once-daily radiation therapy for inflammatory breast cancer. Int J Radiat Oncol Biol Phys. 2014;89:997-1003. 5. Panades M, Olivotto IA, Speers CH, et al. Evolving treatment strategies for inflammatory breast cancer: a population-based survival analysis. J Clin Oncol. 2005;23:1941-1950. 6. Woodward WA. Postmastectomy radiation therapy for inflammatory breast cancer: Is more better? Int J Radiat Oncol Biol Phys. 2014;89: 1004-1005.

e8

S.R. Stecklein et al

7. Damast S, Ho AY, Montgomery L, et al. Locoregional outcomes of inflammatory breast cancer patients treated with standard fractionation radiation and daily skin bolus in the taxane era. Int J Radiat Oncol Biol Phys. 2010;77:1105-1112. 8. Liao Z, Strom EA, Buzdar AU, et al. Locoregional irradiation for inflammatory breast cancer: Effectiveness of dose escalation in decreasing recurrence. Int J Radiat Oncol Biol Phys. 2000;47: 1191-1200. 9. Bristol IJ, Woodward WA, Strom EA, et al. Locoregional treatment outcomes after multimodality management of inflammatory breast cancer. Int J Radiat Oncol Biol Phys. 2008;72:474-484. 10. Rosso KJ, Tadros AB, Weiss A, et al. Improved locoregional control in a contemporary cohort of nonmetastatic inflammatory breast cancer patients undergoing surgery. Ann Surg Oncol. 2017;24:29812988.

Practical Radiation Oncology: --- 2019 11. Dawood S, Merajver SD, Viens P, et al. International expert panel on inflammatory breast cancer: Consensus statement for standardized diagnosis and treatment. Ann Oncol. 2011;22:515-523. 12. Li J, Gonzalez-Angulo AM, Allen PK, et al. Triple-negative subtype predicts poor overall survival and high locoregional relapse in inflammatory breast cancer. Oncologist. 2011;16:1675-1683. 13. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer. 1981;47:207-214. 14. Greenbaum MP, Strom EA, Allen PK, et al. Low locoregional recurrence rates in patients treated after 2000 with doxorubicin based chemotherapy, modified radical mastectomy, and postmastectomy radiation. Radiother Oncol. 2010;95:312-316. 15. Warren LE, Guo H, Regan MM, et al. Inflammatory breast cancer: Patterns of failure and the case for aggressive locoregional management. Ann Surg Oncol. 2015;22:2483-2491.