Breast cancer is the most prevalent malignancy among Korean women. Its incidence has been steadily rising since 2000, and in 2021, the number of newly diagnosed cases exceeded 30,000. Notably, 44.4% of patients presented with stage Ⅰ disease at diagnosis, and hormone receptor (HR)–positive, human epidermal growth factor receptor 2 (HER2)–negative tumors constituted 69.1% of the newly diagnosed cases [1].

Following the National Surgical Adjuvant Breast and Bowel Project B-06 trial, radiotherapy (RT) after breast-conserving surgery (BCS) has been established as a standard treatment for early breast cancer, significantly improving local control and survival rates [2]. However, subsequent evidence suggested that certain subgroups may derive limited benefit from RT [3-8]. Given the proven efficacy of adjuvant endocrine therapy in reducing recurrence risk [3,9] along with concerns regarding the side effects of irradiation, studies have explored the feasibility of RT omission, and two large randomized trials have validated RT omission in carefully selected breast cancer patients [10-13]. Consequently, several guidelines [14-16] acknowledge the possibility of omitting RT in specific cases: (1) women aged ≥ 70 years with clinical node-negative, T1, HR-positive/HER2-negative tumors; and (2) women aged ≥ 65 years with pathological node-negative, T ≤ 3 cm, HR-positive/HER2-negative tumors treated with BCS and a minimum of 5 years of adjuvant endocrine therapy. Despite these recommendations, real-world data supporting this approach in Korean patients remain limited.

Therefore, we conducted a retrospective analysis to compare treatment outcomes for low-risk, early-stage breast cancer between patients who received postoperative RT and those who did not at a single institution in Korea. 

Between January 2010 and December 2020, a total of 24,582 patients underwent breast cancer surgery at Asan Medical Center. Among them, 13,967 patients who underwent BCS without neoadjuvant chemotherapy were identified. Of these, 7,134 patients were diagnosed with stage ⅠA disease, and 1,144 patients met the study’s definition of low-risk, early-stage breast cancer. After excluding patients with missing RT data and those with less than 6 months of follow-up, a final cohort of 742 patients was included in the analysis. Among them, 707 patients received postoperative RT (PORT group), while 35 patients did not receive adjuvant RT (RT omission group) (Fig. 1). The majority of patients (98.4%) had unilateral breast cancer, and most patients (98.2%) underwent sentinel lymph node biopsy without axillary dissection. The median tumor size was 1.1 cm (range, 0.1 to 2.0), with ductal carcinoma being the most common histologic subtype. Baseline characteristics were generally similar between the two groups; however, patients in the RT omission group were older (median, 76 years; range, 68 to 85) than those in the PORT group (median, 62 years; range, 55 to 87). Additionally, ductal carcinoma was less common in the RT omission group (80.0%) than in the PORT group (92.8%) (Table 1).

In the PORT group, the majority of patients (98.6%) received whole breast irradiation (WBI), while 10 patients underwent partial breast irradiation (PBI) to the tumor bed. Three patients (0.5%) received elective nodal irradiation at a dose of 45 Gy in 25 fractions, which encompassed the supraclavicular fossa and axillary apex. For those receiving WBI, a total dose of 41.6–50.4 Gy (median, 50.0 Gy) was administered in 16–28 fractions (median, 25). Tumor bed boost was delivered at a dose of 9–16 Gy (median, 10 Gy) in 3–8 fractions (median, 4) for patients with negative surgical margins, while those with positive surgical margins received a higher dose of either 12.5 Gy in 5 fractions or 15.0 Gy in 6 fractions. Among patients who underwent PBI, 70.0% received a total of 38 Gy in 10 fractions (Table 2). In the RT omission group, most patients did not receive treatment due to advanced age (68.6%) or comorbidities (14.3%), and six patients declined to undergo RT.

Median follow-up was 60.0 months (interquartile range [IQR], 43.0 to 86.0) in the PORT group and 52.0 months (IQR, 28.0 to 64.0) in the RT omission group (p = 0.003). Disease recurrence occurred in 10 patients (1.4%) in the PORT group and one patient (2.9%) in the RT omission group, corresponding to 5-year cumulative incidence rates of 1.28% (95% confidence interval, 0.32% to 2.24%) and 0%, respectively. This difference was not statistically significant (Gray’s test, p = 0.305) (Fig. 2A). No variable demonstrated a statistically significant association with the cumulative incidence of disease recurrence in univariate analysis, and multivariate analysis was not performed due to the limited number of recurrence events (Table 3).

Regarding locoregional recurrence, two (0.3%) and five (0.7%) patients in the PORT group experienced local and regional recurrences, respectively. No locoregional recurrence was observed in the RT omission group. Among those with local recurrence, one patient underwent salvage mastectomy and remained disease-free, while the other did not receive further treatment due to terminal illness. Among the five patients with regional recurrence, two underwent salvage axillary dissection, one received axillary dissection followed by RT to the supraclavicular and axillary lymph nodes, one received RT alone, and one did not receive further treatment due to pre-existing comorbidities.

For survival outcomes, 18 patients (2.5%) in the PORT group and three patients (8.6%) in the RT omission group died. Breast cancer-specific death occurred in four patients (0.6%) in the PORT group and none in the RT omission group. While the 5-year OS rates differed significantly between the two groups (98.5% vs. 92.7%, p = 0.007), the difference in BCSS was not statistically significant (99.8% vs. 100%, p = 0.737) (Fig. 2B, 2C).

Propensity score matching was conducted to minimize selection bias and potential confounding effects, yielding 70 matched patients in the PORT group and 32 in the RT omission group. After matching, baseline characteristics were generally well balanced – most covariates showed standardized mean differences of ≤0.2, and the largest remaining imbalances observed in patient age and histologic subtype were small and clinically negligible (Table 4). Median follow-up remained longer in the PORT group than in the RT omission group (63.5 months [IQR, 44.3 to 83.5] vs. 43.0 months [IQR, 27.3 to 68.3 ]; p = 0.006). Disease recurrence occurred in only one patient in each cohort, at 66 months in the PORT group and 100 months in the RT omission group. No locoregional recurrence was observed in both groups. A competing-risk analysis showed no significant difference in the cumulative incidence of disease recurrence between groups (Gray’s test, p = 0.468) (Fig. 3A), and no significant differences were observed in OS or BCSS between the two groups after matching (p = 0.502 and p = 0.617, respectively) (Fig. 3B, 3C). 

Our study demonstrated two principal findings. First, there was no statistically significant difference in overall disease recurrence between the groups. Second, locoregional recurrence rates were low across both cohorts, with no instances of locoregional recurrence observed among the 35 patients in the RT omission group.

Although the long-term follow-up results of the Cancer and Leukemia Group B 9343 trial demonstrated no survival benefit from adjuvant RT following BCS in women aged 70 years and older with stage Ⅰ estrogen receptor-positive tumors, there was a significant difference in terms of locoregional recurrence (10-year locoregional recurrence-free survival rate, 98% in the tamoxifen + RT group vs. 90% in the tamoxifen only group) [11]. The following PRIME Ⅱ trial reported comparable results, with a 10-year local recurrence rate of 0.9% in the RT group vs. 9.5% in the no-RT group [13] (Supplementary Table S1). Consequently, local recurrence remains a major concern for patients undergoing BCS. Although salvage treatments such as repeat lumpectomy or mastectomy are available, the psychological distress associated with recurrence and the necessity for additional interventions should not be underestimated. In this context, the omission of adjuvant RT becomes a challenging clinical decision, particularly when considering the potential tenfold increase in the risk of locoregional recurrence despite no observed difference in survival. This issue is especially pertinent in Korea, where cancer treatment costs are largely covered by the National Health Insurance Service and access to healthcare is exceptionally convenient, thereby minimizing the burden on patients. Our findings, particularly the absence of locoregional recurrence in the RT omission group, suggest that the omission of adjuvant RT in carefully selected patients may not adversely affect oncologic outcomes. This finding may be attributed to several factors. First, the stringent patient selection criteria applied in our study—such as smaller tumor size, favorable histologic subtypes, lower histologic and nuclear grades, and a predefined Ki-67 threshold—may have contributed to the low recurrence rates. In addition, the routine use of preoperative breast MRI may have enabled more accurate staging and facilitated the exclusion of hidden malignancies. Of note, no locoregional recurrences were observed even among patients in the RT omission group with positive margins, and further studies are warranted to clarify the clinical implications of margin status in patients with low-risk, early-stage breast cancer.

Avoiding unnecessary radiation offers multiple advantages, such as reducing healthcare costs and patient inconvenience, while eliminating RT-related toxicities that may compromise cosmesis, diminish quality of life, or lead to serious health complications. The primary challenge lies in accurately identifying the very low-risk patients who can safely forgo RT while still achieving excellent outcomes with minimal physical, psychological, and financial burdens. Traditionally, clinicopathologic factors such as patient age, tumor size, and tumor grade have been utilized to predict the risk of local recurrence following BCS. Over the past decade, growing evidence has highlighted the molecular heterogeneity of early-stage breast cancer, emphasizing the need for personalized treatment strategies. Several clinical trials have sought to refine patient selection for RT omission based on molecular markers, even among younger patients. The LUMINA trial, which included patients with a Ki-67 index of ≤13.25%, reported a 5-year local recurrence rate of only 2.3% [21]. Similarly, the IDEA study used the Oncotype DX Breast Recurrence Score to guide decisions regarding RT omission (Supplementary Table S1) [22]. Other trials, such as PRECISION (NCT02653755) and PRIMETIME (ISCTRN41579286), have integrated clinical and biological factors, including the Prosigna PAM50 assay or the IHC4+Clinical biomarker, to further refine risk stratification. Additionally, two ongoing randomized trials—EXPERT (NCT02889874) and DEBRA (NCT04852887)—are investigating the role of endocrine therapy alone in patients selected through genomic profiling. Collectively, if these studies confirm that biomarkers can reliably predict local recurrence risk, they could significantly advance more personalized approaches to RT omission.

The limitations of our study highlight the need for a larger, nationwide study to validate our results and provide more generalizable evidence. First, its retrospective, single-center design introduces inherent selection biases. Second, the non-randomized treatment allocation resulted in a significant imbalance in patient numbers and baseline characteristics between groups. While this reflects the real-world clinical landscape, where RT omission remains infrequent and, when implemented, is typically reserved for patients older than those outlined in current guidelines [23], we attempted to mitigate these imbalances through propensity score matching. Additionally, the follow-up time differed between the two groups, with the PORT group having a longer follow-up duration both before and after matching. Although time-to-event analyses such as Kaplan-Meier estimation and Fine-Gray competing-risk models were used to adjust for this difference, the shorter follow-up in the RT omission group may still have limited the detection of late recurrences. This temporal imbalance should be taken into account when interpreting our findings. Moreover, the oncologic safety observed after RT omission must be interpreted in light of the universal and sustained use of adjuvant endocrine therapy. In our study, the planned five-year course of tamoxifen or aromatase inhibitor was initiated by all patients and completed by the vast majority, likely underpinning the absence of locoregional recurrence despite RT omission. However, real-world adherence to endocrine therapy is suboptimal, with 5-year persistence rates of only 60%–80% in HR-positive survivors [24]. Because discontinuation has been associated with higher locoregional recurrence, clinicians should adopt a cautious, individualized approach when considering RT omission for patients in whom long-term endocrine therapy is uncertain. Finally, while most patients (98.6%) in the PORT group of our study underwent WBI, PBI has increasingly been adopted in selected early-stage breast cancer patients [25-28]. The ongoing EUROPA trial (NCT04134598) reflects this evolving paradigm by comparing single-modality RT, primarily short-course PBI, with endocrine therapy alone in women aged ≥ 70 years with T1N0, luminal A-like tumors. Interim results suggest better quality of life and fewer adverse events in the RT arm (Supplementary Table S1) [29]. While our study explored the feasibility of RT omission in well-selected low-risk patients receiving standard endocrine therapy, future research should broaden the comparison to include PBI alone, with comprehensive evaluation not only of oncologic outcomes but also of patient convenience and cost-effectiveness. Collectively, these combined efforts may help inform more individualized treatment strategies.

In conclusion, our study suggests that postmenopausal women with T1N0, grade 1–2, HR-positive, HER2-negative tumors with a Ki-67 index of ≤20% who received adjuvant endocrine therapy exhibit comparable disease recurrence and survival rates, regardless of their post-BCS RT status. These findings support the feasibility of RT omission in carefully selected patients receiving endocrine therapy. However, further studies by the Korean Radiation Oncology Group are necessary to confirm these preliminary results and refine the selection criteria for RT omission in low-risk, early-stage breast cancer.