2.1. Age as a Selection Criterion The concept of omission of RT for low-risk breast cancer dates to the early 1980s. At that time, the seminal NSABP B-06 trial had established the importance of RT in reducing the risk of an ipsilateral in-breast tumor recurrence (IBTR) after lumpectomy in women with node-positive or node-negative breast cancer [1]. Results from many clinical trials and EBCTCG meta-analysis established the benefit of adjuvant radiation therapy in reducing both IBTR and breast cancer deaths [2], as well as the benefit of endocrine therapy [3,4,5] in the treatment of early-stage HR+ breast cancer. The widely accepted recommendation for HR+ breast cancer includes five years of tamoxifen (TAM) [5] or an aromatase inhibitor (AIs) [4]. However, with the increasing use of mammography resulting in the detection of occult tumors, and data demonstrating the benefit of TAM in women with HR+ breast cancers, many began to question whether the benefit of RT observed in NSABP B-06 would apply to clinically low-risk patients. Key randomized controlled trials subsequently designed to evaluate clinical outcomes after the omission of RT in early-stage HR+ breast cancer are included in Table 1. The NSABP B-21 study opened in 1989 and randomized 1009 women with tumors ≤1 cm to TAM alone, RT, or RT + TAM [3]. There was a substantial risk of an IBTR after lumpectomy in patients with small tumors ≤1 cm despite TAM (16.5% at 8 years of follow-up versus 2.8% in the RT + TAM group). TAM alone was less effective than RT alone in reducing IBTRs. A Canadian trial on 769 women with pT1-2, N0 breast cancer from 1992–2000 randomized women to TAM alone or TAM + RT and reported an 8-year IBTR of 17.6% versus 3.5%, respectively [6]. In a planned subgroup analysis of T1, HR+ women, the 8-year IBTR was 15.2% versus 3.6%. In an unplanned subgroup analysis restricted to women with tumors ≤1 cm, HR+, and age ≥60, there was no difference in the rate of local relapse at 5 years (1.2% vs. 0%) or 8 years (3.6% vs. 0%) [7], suggesting that age could be used as a criterion to select patients for omission.In two subsequent modern randomized control trials, age was used as a criterion for omission after breast-conserving surgery (BCS) in patients with HR+ disease undergoing ET. In the CALGB 9343, 636 women ≥70 with clinical stage I, ER+ breast cancer, were randomized to TAM or RT + TAM [8]. At 10 years, the locoregional recurrence-free survival was 98% in the RT + TAM arm versus 90% in the TAM arm. This did not translate into differences in OS, DFS, or breast preservation. Among the patients who died, the causes of death were largely unrelated to breast cancer. In the PRIME II study, 1326 women ≥65 with node-negative, ER+ breast cancer were randomized to RT versus no RT [9]. All patients received ET. At 10 years, the IBTR rate was 0.9% versus 9.8% in the RT and no RT arms, respectively (p = 0.0008). However, there were no differences in regional recurrences, contralateral breast cancer, or distant metastases. These data suggest that the absolute risk of recurrence is lower when selecting patients of older age, with most dying of non-breast cancer-related causes. As a result, the National Comprehensive Cancer Networks Guidelines included adjuvant ET alone as a category one recommendation for women >70 years of age with early-stage HR+ breast cancer.

Of note, despite modest absolute reductions in IBTR in PRIME II and NSABP B-21, the hazard ratio for the benefit of RT was similar (0.12 in PRIME II versus 0.19 in NSABP B21) and validated a persistent benefit of RT even in these early-stage breast cancer populations. The omission of RT based solely on age and HR status alone may lack appropriate risk-stratification power. It remains critical to identify a patient population with “low enough” IBTR risk following lumpectomy, for whom adjuvant RT may not provide a convincingly meaningful benefit.

2.2. Combining Age and Biology Does age remain an important criterion for RT omission once we account for tumor subtype? In a retrospective study of 1434 patients who underwent breast-conserving therapy from 1997–2006, the crude risk of locoregional recurrence (LRR) varied drastically by tumor subtype [10]. Patients with luminal A (ER+/PR+/HER2-) and luminal HER2+ breast cancer had a 1.5% and 1.0% risk for local recurrence at a median follow-up of 85 and 82 months, respectively. In patients with the luminal B subtype, the risk rose to 4% at a median follow-up of 96.3 months. In patients with HER2+ and triple-negative subtypes, the risk rose to 10.9% and 8.8% at a median follow-up of 83.8 and 61.2 months, respectively. In young women, the luminal B and HER2 subtypes were associated with high rates of LR after BCS and RT, though increasing age was associated with decreased risk of LR independent of the breast cancer subtype. In the Canadian study discussed above, investigators developed a low-risk clinical model of patients with T1, grade 1 or 2 breast cancer and ≥60 years, and reported a 10-year IBTR rate of 4.6% compared to 13.7% for the high-risk clinical group [11]. Combining the low-risk model with the luminal A subtype, defined by IHC and proliferative index criteria, resulted in a 10-year IBTR rate of 1.3% with TAM compared to 5% with RT + TAM. On multivariable analysis, treatment with RT + TAM versus TAM alone, the intrinsic subtype and clinical risk group were significantly associated with IBTR. The LUMINA trial is a prospective multicenter cohort study of patients ≥age 55 with grade 1–2, T1N0 breast cancer status, post-lumpectomy with ≥1 mm margins and luminal A subtype (ER ≥ 1%, PR >20%, HER2 negative and ki-67 ≤13.25%). All patients are treated with adjuvant endocrine therapy (ET), and RT is omitted following BCS. Preliminary results from 501 patients were reported at the ASCO 2022 annual meeting. The median age was 67 years, with 88% of patients Moving beyond IHC, gene signatures based on multigene profiling assays are increasingly being used as predictive and prognostic markers for breast cancer recurrence and have been incorporated into ongoing clinical trials for the omission of RT (Table 2). The PAM50 assay is a standardized 50-gene set for intrinsic subtype classification used for risk stratification. This assay is currently being investigated for the biological selection of low-risk breast cancers in the PRECISION and EXPERT trials. The Oncotype DX is a 21-gene classifier containing the mRNA quantification of ER, PR, HER2, and tumor proliferation that provides a recurrence score (RS) to quantify the risk of distant recurrence. In a study investigating the association between RS and LRR in node-negative and ER+ breast cancer from two NSABP trials, NSABP B-14 (TAM treated: n = 895, placebo-treated: n = 355) and NSABP B-20 (n = 424), RS was found to be a significant predictor of LRR in TAM treated, placebo treated and chemotherapy + TAM treated patients. Of note, only 2% of patients enrolled in the NSABP 20 were over the age of 70 years, and no patients over the age of 70 years were enrolled in NSABP B14. On multivariable analysis, RS, age, and type of treatment were independent predictors of LRR. DEBRA is a phase III trial that randomizes patients between 50 to 70 years of age, with HR+ early-stage breast cancer, and with an Oncotype DX RS ≤18 to adjuvant RT versus observation.

We await results from these prospective clinical trials, which combine age and biology to optimize the selection of patients who may not benefit from adjuvant radiation.

In the EBCTCG meta-analysis summarizing studies beginning before 2000, the rate of any first recurrence with BCS alone was 35% compared to 19.3% with BCS and RT, translating to a 10-year gain of 15.7% and a relative risk of 0.52 from RT [2]. In more modern series, the absolute local recurrence and the absolute benefit of whole breast RT are lower, at around 1.1–2% [9,12]. Therefore, improving the radiation delivery and toxicity has become increasingly important to maintain a favorable therapeutic ratio. Partial breast irradiation (PBI) involves focused radiotherapy delivery to the lumpectomy cavity rather than the whole breast. The rationale is that residual microscopic disease tends to lie within 2 cm of the initial tumor [13,14], and most local recurrences occur at or near the lumpectomy bed [15]. The UK IMPORT LOW study, which accrued patients between 2007 and 2010, evaluated outcomes after whole breast RT with 40 Gy in 15 fractions, reduced dose whole breast RT with 36 Gy in 15 fractions and 40 Gy in 15 fractions to the lumpectomy area, and partial breast RT with 40 Gy in 15 fractions [12]. The risk of local relapse at 6 years was 1.1%, 0.2%, and 0.5%, respectively. The toxicity of the two experimental arms with reduced dose and partial breast RT was lower, indicating that reducing RT fields does not compromise efficacy, nor does it increase side effects.Accelerated RT involves higher doses delivered over a shorter interval of time. The rationale is that a shorter RT course with a higher dose per fraction can achieve the same therapeutic effect as a longer treatment course with a lower dose per fraction. A one-week schedule for whole breast or chest wall RT was evaluated in the UK FAST FORWARD study and found not to be inferior to standard fractionation without compromising normal tissue effects [16,17]. Accelerated partial breast irradiation (APBI) combines field reduction and acceleration of treatment time. APBI has been evaluated in phase III trials using percutaneous approaches, including multicatheter interstitial brachytherapy (MIB), balloon catheter intracavitary brachytherapy, electronic radiotherapy, and non-percutaneous approaches, including three-dimensional conformal external beam radiotherapy (3D-CRT), or intensity modulated radiation therapy (IMRT), proton beams or stereotactic PBI (Table 3). The first randomized controlled trial of APBI came from the Hungarian National Institute of Oncology for patients with T1, N0-1 mi, >40 years of age with negative surgical margins. Investigators reported 10-year IBTR rates of 5.9% and 5.1% in the APBI and WBI arms, respectively, using HDR MIB or an electron beam [18]. Subsequent clinical trials have demonstrated even lower IBTR rates of 2.5–4%, with mature follow-up at 8–10 years [19,20,21,22,23]. The Florence trial showed equivalent outcomes for WBI (50 Gy in 25 fractions with tumor bed boost) vs. APBI (30 Gy in five every-other-day fractions) using IMRT. There was a significantly lower rate of acute and late toxicity and improved cosmetic outcomes with APBI than with WBI at a median follow-up of 10 years [23]. The NSABP B-39/RTOG 0413 trial is the largest prospective study to date, with 4216 patients randomized to WBI (50 Gy in 25 fractions) vs. APBI via either multicatheter brachytherapy (34 Gy in 10 fractions BID), intracavity brachytherapy (MammoSite 34 Gy in 10 fractions BID), or 3D conformal radiation (3D-CRT) (38.5 Gy in 10 fractions BID). The 10-year cumulative incidence of IBTR was 4.6% in the APBI group versus 3.9% in the WBI group, hence not meeting the criteria for equivalence. However, the absolute difference in IBTR was Table 3).Regarding toxicity and cosmesis, results have been mixed with a potential confounding effect from the radiation components of the trials. For one, the interfraction recovery kinetics impact tissue repair for late fibrosis is estimated to have a half-life of repair of 4.4 hours. The RAPID trial from Canada demonstrated non-inferiority using 3D-CRT (38.5 Gy in 10 fractions BID) with an interfraction interval of 6 hours [22]. This study found no difference in IBTR but an increase in moderate late toxicity and adverse cosmesis with APBI compared to hypofractionated WBI (42.5 Gy in 16 fractions). These findings suggest that perhaps 6 hours may have been insufficient for complete normal tissue repair. Furthermore, compared to the Florence trial in which 100% of patients were treated with IMRT, only 10% underwent IMRT in the RAPID trial, and the homogeneity and ipsilateral breast volume constraints were less strict.A retrospective study of 345 patients treated at New York University (NYU) confirmed the low rates of toxicity and poor cosmesis with a five-fraction regimen [24]. In this study, 94% of patients were treated prone, with 32% treated every other day and 68% on consecutive days with 6 Gy × 5. The ipsilateral breast was constrained to V50% n = 199, median follow-up 2.8 years) was 92.5% and 89.4%, respectively. There were low rates of telangiectasia (4.5% grade 1 and 1.5% grade 2), fibrosis (17.6% grade 1 and 3.0% grade 2), and retraction/atrophy (24.1% grade 1, 2.5% grade 2, and 0.5% grade 3). Overall, a once-per-day EBRT-APBI regimen with a sufficient interval between fractions appears to have a significantly less adverse effect on cosmesis. A notable finding of the NYU retrospective review was that when comparing patients with IBTR to those without IBTR, a higher proportion of those who did not receive endocrine therapy experienced IBTR (67% vs 27%, p = 0.048). The NYU S14-0136 trial of prone APBI is a randomized controlled non-inferiority trial to compare radiation fibrosis with five (6 Gy × 5) versus three fractions (8 Gy × 3). The primary endpoint of the study is the rate of fibrosis at 2 years. The study closed accrual in June 2020, and we await the maturation of the data for the primary endpoint.

The side effect profile of APBI over WBI in the modern era will likely continue to improve as we modify delivery schedules, volumes treated, and treatment techniques. The data indicate that APBI not only reduces the biological toxicity of radiotherapy but also reduces the financial toxicity, based on Medicare fee-for-service global reimbursement patterns. Future directions to continue improving the therapeutic ratio include increasing the use of conformal treatments with IMRT, the use of even smaller volumes with preoperative treatments, and additional research regarding dose-volume parameters that allow even safer delivery of treatment with regard to long-term toxicity.