Phase II Study of Proton Beam Radiation Therapy for Patients With Breast Cancer Requiring Regional Nodal Irradiation

Abstract

PURPOSE

To evaluate the safety and efficacy of proton beam radiation therapy (RT) for patients with breast cancer who require regional nodal irradiation.

METHODS

Patients with nonmetastatic breast cancer who required postoperative RT to the breast/chest wall and regional lymphatics and who were considered suboptimal candidates for conventional RT were eligible. The primary end point was the incidence of grade 3 or higher radiation pneumonitis (RP) or any grade 4 toxicity within 3 months of RT. Secondary end points were 5-year locoregional failure, overall survival, and acute and late toxicities per Common Terminology Criteria for Adverse Events (version 4.0). Strain echocardiography and cardiac biomarkers were obtained before and after RT to assess early cardiac changes.

RESULTS

Seventy patients completed RT between 2011 and 2016. Median follow-up was 55 months (range, 17 to 82 months). Of 69 evaluable patients, median age was 45 years (range, 24 to 70 years). Sixty-three patients (91%) had left-sided breast cancer, two had bilateral breast cancer, and five had right-sided breast cancer. Sixty-five (94%) had stage II to III breast cancer. Sixty-eight (99%) received systemic chemotherapy. Fifty (72%) underwent immediate reconstruction. Median dose to the chest wall/breast was 49.7 Gy (relative biological effectiveness) and to the internal mammary nodes, 48.8 Gy (relative biological effectiveness), which indicates comprehensive coverage. Among 62 surviving patients, the 5-year rates for locoregional failure and overall survival were 1.5% and 91%, respectively. One patient developed grade 2 RP, and none developed grade 3 RP. No grade 4 toxicities occurred. The unplanned surgical re-intervention rate at 5 years was 33%. No significant changes in echocardiography or cardiac biomarkers after RT were found.

CONCLUSION

Proton beam RT for breast cancer has low toxicity rates and similar rates of disease control compared with historical data of conventional RT. No early cardiac changes were observed, which paves the way for randomized studies to compare proton beam RT with standard RT.

INTRODUCTION

Regional nodal irradiation (RNI) is used for patients with both high-risk early-stage and locally advanced breast cancer to decrease the risk of locoregional failure (LRF) and improve disease-free survival. The recently published MA.20 and European Organisation for Research and Treatment of Cancer trials demonstrated a benefit in disease-free survival among patients with stage I to III breast cancer who received RNI versus whole-breast/chest wall irradiation alone.^1,2 In both studies, RNI included the first three intercostal spaces of the internal mammary nodes (IMNs). ASCO consensus guidelines for postmastectomy radiation recommend coverage of the IMNs among patients with stage II or greater breast cancer, and National Comprehensive Cancer Network guidelines suggest that when delivering RNI, inclusion of the IMNs should be strongly considered.^3,4 However, radiation therapy (RT) to the IMNs significantly increases cardiopulmonary RT exposure. Cardiac exposure to RT has been shown to increase the risk of major cardiac events with risk proportional to mean heart RT dose.^5,6 Recently, there has been significant public interest in the cardiovascular risks of breast cancer therapy and strategies to mitigate these risks.^7-9

Proton beam RT may spare cardiopulmonary radiation exposure significantly. The potential advantage of proton beam RT results from the physical property of the radiation beam, wherein the path of the beam has a finite range that permits for the absence of an exit dose, which prevents delivery of dose beyond the location of the target.

Multiple dosimetric planning studies that compared proton RT with photon RT have demonstrated superior delivery to targeted areas while sparing more of the heart and lungs.^10-12 However, prospective clinical data are lacking to support the safety and efficacy of proton RT for breast cancer.¹³ We report the first, to our knowledge, prospective trial of proton RT for breast cancer. These results were used as preliminary data for the currently accruing RadCOMP Consortium Trial (Pragmatic Randomized Trial of Proton Versus Photon Therapy for Patients with Nonmetastatic Breast Cancer: A Radiotherapy Comparative Effectiveness; ClinicalTrials.gov identifier: NCT02603341), a phase III randomized trial that is comparing conventional RT with proton beam RT for locally advanced breast cancer.

METHODS

Patient Eligibility

Patients were enrolled in a prospective clinical trial approved by the institutional review board of the Dana-Farber/Harvard Cancer Center. Adult patients age 18 years or older with nonmetastatic breast cancer and who required postoperative RT inclusive of the IMNs were eligible if the treating physician determined that breast reconstruction prevented adequate target coverage or 5% or more of the heart would have received 20 Gy or more and/or the left anterior descending artery (LAD) would have received 20 Gy or more with conventional radiation.

The primary end point was the incidence of grade 3 or higher radiation pneumonitis (RP) or any grade 4 toxicity within 3 months after RT. This end point was selected because it is an established short-term safety end point for breast RT that also could verify the potential cardiopulmonary-sparing aspect of proton therapy. Secondary end points were 5-year rate of LRF and overall survival (OS), acute skin toxicity per Common Terminology Criteria for Adverse Events (version 4.0), and acute and late toxicity of breast reconstruction.

All patients were seen by a physician before RT; weekly during RT; and at 4 weeks, 8 weeks, and annually for 5 years after RT. Patients consented to photographs to document skin toxicity during treatment and at follow-up visits. In addition, patients underwent cardiac strain echocardiograms and testing for blood-based cardiac biomarkers (amino-terminal pro-B-type natriuretic peptide, ultrasensitive troponin I) before RT and at 4 and 8 weeks after RT to assess subclinical cardiac changes.

Radiation Treatment Planning and Technique

Computed tomography–based treatment planning was performed. No respiratory gating or deep inspiration breath-hold techniques were used. RT began within 3 to 19 weeks from surgery or chemotherapy. The Radiation Therapy Oncology Group breast contouring atlas was used for delineation of target structures and organs at risk (OARs).¹⁴ Target structures included the breast/chest wall (excluding ribs), IMNs, supraclavicular lymph nodes, and axilla. For patients with breast reconstruction, the prosthesis was included in the chest wall target. OARs included the heart, LAD for left-sided treatment, bilateral lungs, and esophagus.

The relative biological effectiveness (RBE) was set at 1.1 per institutional standard. Thus, the dose unit gray equivalent was proton dose in grays × an RBE of 1.1. The dose prescribed to the chest wall was 50.4 Gy (RBE) in 28 fractions per day of 1.8 Gy (RBE) 5 days per week. For patients with intact breasts, the whole breast was prescribed 45.0 Gy (RBE) in 1.8 Gy (RBE) per day over 25 fractions 5 days per week followed by a 14.4-Gy (RBE) fractionated boost to the lumpectomy cavity. The IMNs were prescribed between 45 and 50.4 Gy (RBE) in doses of 1.8 to 2.0 Gy (RBE) per day. In the case of residual or suspected residual tumor in the IMNs, an additional dose to gross disease was permitted. No patient received any component of photon or electron treatment. The planning algorithm was designed to achieve full coverage of all target structures while maximally sparing OARs. Normal tissue constraints included a maximum esophageal dose of 40 Gy (RBE), a mean heart dose less than 1.5 Gy (RBE), a maximum LAD dose of 10 Gy (RBE) a mean LAD dose of less than 2 Gy (RBE), and less than 20% of the ipsilateral lung volume receiving 20 Gy (RBE).

All treatments were delivered using either three-dimensional passively scattered proton (3D-conformal proton therapy [CPT]) or pencil beam scanning (PBS) proton therapy at the Massachusetts General Hospital. The transition from 3D-CPT to PBS occurred in 2013 and corresponded to an upgrade in technology at our institution. Daily imaging for localization was required.

Statistical Analysis

The primary end point for evaluating safety was the rate of grade 3 or higher RP or any grade 4 toxicity within 3 months after RT by Common Terminology Criteria for Adverse Events (version 4.0). Compared with 3D-CPT, PBS planning allows for greater conformality of the radiation beam and controlled skin dose. Therefore, a two-stage design was used in each of the 3D-CPT and PBS cohorts to provide 81% power to detect an acceptable rate of 5% and reject a 15% rate at an α-level of 0.20. The early stopping rule was associated with 88% probability of proceeding from the first stage to full cohort accrual if the overall rate of grade 3 or higher RP or any grade 4 toxicity were truly only 5%.

Failure end points were calculated from start of RT. Wilcoxon rank sum test was used to compare ordinal grades of skin toxicity between proton techniques with an exact P value reported. Distant metastasis–free survival (DMFS) and OS were estimated by the Kaplan-Meier method, where DMFS failure was defined as the development of distant metastasis and death, whichever was earlier. Competing risks for LRF were distant metastasis and death, whereas any recurrence and death were considered as competing risks for surgical re-intervention. LRF and surgical re-intervention were estimated as the cumulative incidence function, and Gray’s test was used for comparison. Data analysis was performed using SAS 9.4 statistical software (SAS Institute, Cary, NC), and P values were based on a two-sided hypothesis.

For cardiac biomarkers, two-dimensional and Doppler echocardiography data were collected for standard and strain evaluation. Left ventricular ejection fraction was calculated from apical four- and two-chamber views using a modified Simpson biplane method. Measurement of serum-based amino-terminal pro-B-type natriuretic peptide and high-sensitivity cardiac troponin I was performed on a Siemens platform (Atellica IM high-sensitivity troponin I, Siemens Healthineers) using standard procedures.

RESULTS

Seventy patients were enrolled between August 2011 and November 2016. One patient was excluded from the analysis after withdrawing consent following her first radiation treatment, but she completed therapy as intended. Table 1 lists the characteristics of the patients eligible for analysis. All staging was performed using the American Joint Commission on Cancer seventh edition staging manual.¹⁵ Median age of the cohort was 45 years (range, 24 to 70 years). Fifty patients (72%) were younger than age 50 years. Sixty-three patients (91%) received RT to the left-side chest wall/breast, two (3%) received bilateral chest wall RT, and four (6%) received right-side chest wall/breast RT. Sixty-five patients (94%) had stage II or III breast cancer. Sixty-four patients (93%) underwent mastectomy. Sixty-eight patients (99%) received chemotherapy, of whom 39 (57%) received neoadjuvant chemotherapy and five (13%) experienced a pathologic complete response. Of 30 patients with involved lymph nodes who received chemotherapy, 13 (43%) converted to node negative. Forty-seven patients (68%) underwent axillary dissection, and the remainder underwent sentinel lymph node biopsy alone. Fifty-three (95%) of 56 patients with hormone receptor–positive breast cancer received endocrine therapy; the remaining three had minimal hormone receptor positivity and forewent endocrine therapy.

TABLE 1.

Patient Characteristics

Radiation Dose

Dosimetric data are listed in Table 2. Target coverage was achieved for all patients. Six patients with gross IMN disease were prescribed IMN doses that ranged from 56.0 to 64.4 Gy (RBE). The median of mean delivered dose to the IMN was 48.8 Gy (RBE), in keeping with intended dosing. With regard to normal structures, the mean heart dose was a median of 0.50 Gy (RBE), and the mean LAD dose was a median of 1.16 Gy (RBE).

TABLE 2.

Dosimetric Data

Toxicity

The incidence of grade 3 or higher RP in the cohort was 0%, and the incidence of any grade 4 to 5 toxicity was 0%. Table 3 lists grade 2 or higher treatment-related toxicities. One patient developed grade 2 RP, which was diagnosed on the basis of clinical symptoms alone 4 months after radiation and was treated successfully with oral corticosteroids. One patient developed a severe infection that involved her bilateral chest wall 4 months after RT to the left-side chest wall and required intravenous antibiotics.

TABLE 3.

Treatment-Related Toxicities

The incidence of overall acute grade 2 or 3 skin toxicity was 86%, with no statistical difference between 3D-CPT and PBS (P = .094) in the distribution of grade 1 (4% v 20%), 2 (91% v 78%), and 3 (4% v 2%) toxicity. However, at 1-year follow-up, the crude incidence of grade 1 telangiectasia was 16%, and all patients with telangiectasia were treated with 3D-CPT (Appendix Fig A1, online only). Because 3D-CPT delivers a higher skin dose than PBS, this finding is not surprising, but it does highlight a difference in potential late toxicity. Grade 2 fatigue occurred in 24 patients (35%). Five patients (7%) had acute grade 2 dysphagia as a result of radiation esophagitis. No patients experienced grade 2 or higher lymphedema.

Although no grade 2 or higher rib fractures occurred in the study, five patients (7%) experienced grade 1 rib fracture, with a median time to fracture of 15.9 months. With proton RT, the RBE dose, or the effective delivered dose, increases at the beam’s distal edge. For proton breast treatment, as the dose is pulled away from the lungs, the distal edge of the field may be in the ribs, which might correlate with rib facture. For the purposes of this study, rib fracture was identified either through patient symptoms (eg, pain) and confirmed with imaging or through detection as an incidental finding on imaging obtained for another reason. For all patients with rib fracture, the location was compared with the treatment plan, and any fracture located within or in proximity to the treatment field was considered related to radiation.

Cardiac Outcomes

Across the entire cohort of patients, no significant decreases in global longitudinal strain or left ventricular ejection fraction and no significant increases in circulating biomarkers were observed¹⁶ (Awadalla et al, manuscript in preparation). The final analyses of the strain portion of the echocardiograms to assess for subclinical cardiac strain changes for the entire cohort, along with circulating cardiac biomarkers, will be reported separately.

Disease Control

Median follow-up of the cohort was 55 months (range, 17 to 82 months) among 62 surviving patients. Of 69 patients who completed therapy on protocol, one with a p53 mutation experienced an in-field recurrence, an LRF in the treated axilla (Appendix Fig A2, online only), and an in-field sarcoma (crude LRF rate, 1.4%). Five-year estimated LRF was 1.5% (95% CI, 0.1% to 7.0%; Fig 1A). There were eight distant relapses (crude distant failure rate, 12%). The 5-year estimated DMFS was 86% (95% CI, 75% to 93%), and the 5-year estimated OS was 91% (95% CI, 81% to 96%; Fig 1B).

FIG 1.

(A) Locoregional failure (LRF) and (B) overall survival (OS) after proton beam radiation therapy.

Breast Reconstruction, Complications, and Cosmesis

Of 64 patients who underwent mastectomy, 53 (83%) pursued reconstruction. Table 4 lists the details of breast reconstruction outcomes. Thirty-nine patients underwent mastectomy with immediate implant-based reconstruction, one patient experienced implant loss before RT, and 12 (32%) of 38 experienced complications after RT. Three of 12 experienced implant loss, although one of the three was not believed to be attributable to RT; the patient experienced a systemic infection after dental extraction. Nine of the 12 patients required revisions for asymmetry or contracture. Eleven patients underwent tissue expander-to-implant exchange before RT. Of those, one patient experienced implant loss before RT and eventually underwent a successful reconstruction, and three (27%) required revision for asymmetry or contracture after RT. Finally, 14 patients did not undergo reconstruction before RT. Of those, three attempted reconstruction after treatment, all with success, including one patient with flap failure who required a second surgery. Therefore, among the entire cohort of 53 patients who attempted reconstruction, 15 (28%) experienced an RT-related complication, and two (4%) experienced reconstructive loss attributable to RT. Among the 48 patients at risk for pre-RT implant loss, the actuarial rate of any unplanned surgical re-intervention was 33% at 5 years. A cosmetic outcome at 5 years is shown in Appendix Fig A3 (online only).

TABLE 4.

Reconstruction Outcomes

DISCUSSION

This study represents the first prospective trial of proton beam RT for breast cancer to our knowledge. In a cohort of 69 women who received proton beam RT with a median follow-up of 55 months, both toxicity and disease control rates compared favorably with historic data of conventional RT.^1,2 The growth in the number of proton centers over the past decade, combined with a heightened awareness of cardiac morbidity, has led to a marked increase in the number of patients with breast cancer who receive proton beam RT, both on and off a clinical trial.¹⁷ However, no published clinical trials have evaluated the safety and efficacy of this complex treatment to date.¹⁸ The phase III RadCOMP Consortium Trial is a large-scale, multicenter, randomized clinical trial that will assess whether proton therapy decreases the risk of cardiovascular morbidity and mortality. The preliminary data from our study served as the foundation for the rationale and treatment technique of RadCOMP, and we look forward to its findings.

Until then, this trial contributes several valuable pieces of information to the understanding of proton beam RT for breast cancer. First, because the main rationale for proton therapy is to spare cardiopulmonary radiation exposure, this study successfully demonstrates the feasibility of delivering homogenous, high doses of radiation to maximize cancer control while sparing the heart and lung. The doses achieved to both target structures and normal tissues compare favorably with doses achievable with even the most advanced conventional radiation techniques. In particular, the mean heart and LAD doses of 0.50 Gy (RBE) and 1.16 Gy (RBE), respectively, are so modest despite comprehensive coverage to the nearby IMNs that they cannot be replicated with other RT techniques. For example, mean heart doses for patients who receive conventional left-sided RT inclusive of the IMNs in the modern era are routinely between 2.5 and 6 Gy, and mean doses delivered to the LAD can be between 10 and 30 Gy.^16,19,20

In addition, all patients in the study tolerated treatment with minimal toxicity. This provides needed evidence that proton beam RT confers similar acute and subacute adverse effects as conventional RT. Specifically, RP was rare, and cosmetic complications were similar to modern reports of patients who receive conventional RT.^21-25 Most importantly, disease control rates in this high-risk patient population were consistent with those of patients who receive conventional radiation, with only one patient experiencing an LRF over the study period.^1,2,24,25 Of note, this was observed in patients for whom the chest wall target was contoured in a more conservative manner than the conventional radiation contouring guidelines currently in use, in which the entire chest wall inclusive of the ribs is targeted. That control rates remained high using a smaller target and more-precise radiation delivery with protons suggests that there may be the option to exclude the ribs as part of the chest wall contour in patients after mastectomy, unless there is suspicion for direct chest wall involvement.

Although it was not a predetermined study end point, five patients (7%) experienced rib fractures, which is slightly higher than in historical photon series.^21,22 Rib fracture is a theoretical concern with protons given that the end of range of the proton beam approximates the ribs. Although our sample size is too small to determine whether the risk of fracture is higher with protons, it warrants continued study.

Of note, skin toxicity was a concern when this trial was conceived because 3D-CPT necessitates the delivery of the full prescription dose to the skin. These patients overall did well with respect to skin toxicity. However, PBS protons, the modality used for the majority of patients in this trial and most readily used for patients currently, allow for modification of skin dose to the desired dose level, something not feasible with standard RT. We observed borderline improvements in acute toxicity in the PBS proton group, but additional modifications to our approach may allow for additional improvement in toxicity and cosmetic outcomes.

Our observation of no changes in cardiac biomarkers or strain echocardiography as a result of RT was reassuring. This stands in contrast to prior studies of conventional RT that showed an increase in serum troponin and a decrement in strain echocardiography after RT.^26-28 In addition to the RadCOMP Consortium Trial, study is under way, as a shared pursuit between our institution and the MD Anderson Cancer Center, to compare protons and conventional radiation in a randomized fashion as it relates to early cardiac changes, including strain echocardiography, as well as to quality-of-life parameters, including cosmesis, shoulder function, lymphedema, and pulmonary changes.

This study has several limitations. It was performed at a single institution, and patients were selected on the basis of physician judgment and insurance approval, which introduce the possibility of selection bias. In addition, the 5-year follow-up limits our ability to comment on long-term disease control and cardiac events, although the 5-year end points are promising, and the cardiac surrogate data are reassuring. Finally, although the study provides some insight with regard to the differences between proton and conventional RT, additional randomized data are needed to highlight differences in toxicity.

In summary, in our prospective trial of women with locally advanced breast cancer who required treatment of the IMNs, proton beam RT was safe and effective. Future research will provide needed information about the potential long-term normal tissue–sparing benefits of this complex treatment modality compared with conventional radiation.

APPENDIX

FIG A1.

Cosmetic outcome approximately 5.5 years after three-dimensional passively scattered protons to the chest wall after implant reconstruction. Telangiectasia are noted in the inferior chest wall below the implant.

FIG A2.

Positron emission tomography scan fused to radiation treatment planning computed tomography scans with doses shown demonstrate an in-field regional nodal failure for the patient in the study with a locoregional failure.

FIG A3.

Cosmetic outcome 5 years after pencil beam scanning proton radiation. Photographs of the breast at baseline before radiation therapy and at 5 years are shown.

AUTHOR CONTRIBUTIONS

Conception and design: Beow Y. Yeap, Michele A. Gadd, Michelle Specht, Alice Ho, Jonathan Passeri, Tomas G. Neilan, Hsiao-Ming Lu, Shannon M. MacDonald

Administrative support: Rachel B. Jimenez, Shea Hickey

Provision of study material or patients: Rachel B. Jimenez, Shea Hickey, Steven J. Isakoff, Barbara L. Smith, Alphonse G. Taghian

Collection and assembly of data: Rachel B. Jimenez, Shea Hickey, Nicolas DePauw, Estelle Batin, Michele A. Gadd, Steven J. Isakoff, Barbara L. Smith, Eric C. Liao, Hsiao-Ming Lu, Shannon M. MacDonald

Data analysis and interpretation: Rachel B. Jimenez, Shea Hickey, Nicolas DePauw, Beow Y. Yeap, Estelle Batin, Steven J. Isakoff, Barbara L. Smith, Amy S. Colwell, James L. Januzzi, Tomas G. Neilan, Alphonse G. Taghian, Shannon M. MacDonald

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Phase II Study of Proton Beam Radiation Therapy for Patients With Breast Cancer Requiring Regional Nodal Irradiation

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/site/ifc.

Rachel B. Jimenez

Employment: Biogen (I)

Research Funding: Focal Therapeutics

Beow Y. Yeap

Consulting or Advisory Role: Abcodia (I)

Steven J. Isakoff

Consulting or Advisory Role: AbbVie, Genentech, Roche, Myriad Genetics, Hengrui Therapeutics, Puma Biotechnology, Immunomedics

Research Funding: Genentech (Inst), PharmaMar (Inst), AbbVie (Inst), OncoPep (Inst), Merck (Inst), AstraZeneca (Inst), MedImmune (Inst)

Eric C. Liao

Honoraria: Allergan, Musculoskeletal Transplant Foundation

Consulting or Advisory Role: Allergan, Musculoskeletal Transplant Foundation

Research Funding: Musculoskeletal Transplant Foundation (Inst)

Travel, Accommodations, Expenses: Allergan

Amy S. Colwell

Honoraria: Allergan

Consulting or Advisory Role: Allergan

Travel, Accommodations, Expenses: Allergan

Alice Ho

Research Funding: Merck (Inst), Tesaro (Inst)

James L. Januzzi

Research Funding: Roche (Inst), Abbott Laboratories (Inst), Siemens Healthcare Diagnostics (Inst), Prevencio (Inst)

Tomas G. Neilan

Consulting or Advisory Role: Bristol-Myers Squibb, Parexel, Intrinsic Imaging, Aprea AB

Alphonse G. Taghian

Honoraria: UpToDate

Consulting or Advisory Role: Vision RT

Hsiao-Ming Lu

Employment: Mevion Medical Systems (I)

No other potential conflicts of interest were reported.