Identification of high- and low-risk groups for ipsilateral recurrence within 10 years after breast-conserving surgery for ductal carcinoma in situ and personalized treatment: a retrospective study

Introduction

Breast cancer is one of the most common malignant tumors affecting women worldwide (1,2). According to statistics, in 2021, there were 297,790 new cases of breast cancer in women in the United States, accounting for 31% of all newly diagnosed cancers in women. There were 43,170 deaths in 2021, representing 15% of all cancer-related deaths. Over the past 20 years, the incidence of breast cancer has been steadily increasing at a rate of 0.5% per year, with the majority being early-stage breast cancer (3). According to the latest data released on February 1, 2024, by the International Agency for Research on Cancer (IARC) under the World Health Organization, approximately 2.3 million new cases of breast cancer were diagnosed worldwide in 2022. With 670,000 deaths, breast cancer surpassed gastric cancer to become the fourth leading cause of cancer-related deaths globally, far exceeding gynecological cancers such as uterine and ovarian cancers (4). In response to the growing breast cancer epidemic, public health departments worldwide are intensifying prevention and intervention efforts, including raising awareness of breast cancer through education, promoting cancer prevention, expanding screening and early diagnosis techniques, and improving early detection rates. Lifestyle improvements, such as a balanced diet, increased physical activity, and stress reduction, are also being encouraged to lower the risk of developing breast cancer.

Ductal carcinoma in situ (DCIS) is a non-invasive breast cancer characterized by abnormal ductal epithelial cells that have not yet invaded the myoepithelial cells covering the ductal basement membrane (5). Current common treatments for DCIS include mastectomy, breast-conserving surgery (BCS), radiotherapy, and endocrine therapy (ET). The National Comprehensive Cancer Network (NCCN) recommends the following options for local treatment of DCIS: mastectomy (level 2A evidence), BCS with radiotherapy (level 2A evidence), and BCS without radiotherapy (level 2B evidence) (6,7).

BCS, also known as lumpectomy or partial mastectomy, has become a widely accepted treatment option for patients with DCIS, allowing for the removal of the tumor while preserving breast tissue (2,8,9). Although BCS is considered a common treatment for DCIS due to its favorable long-term prognosis, some patients still experience ipsilateral breast tumor recurrence (IBTR) after surgery, presenting a significant challenge for both patients and clinicians (10,11). Tumor recurrence often necessitates further treatment, such as mastectomy, intensified radiotherapy, or systemic therapy (12-14). Therefore, it is crucial to accurately identify predictive factors for IBTR risk following BCS. These factors help guide treatment decisions, optimize follow-up strategies, and improve patient outcomes (3,15,16). Previous studies have identified various prognostic markers, including age, tumor size, histologic grade, hormone receptor (HR) status, surgical margins, and adjuvant therapies (17-19). However, comprehensive analysis of large, diverse patient populations is still needed to develop reliable predictive models and to create more individualized treatment plans. To achieve this, precise assessment of recurrence risk in patients is essential.

The Surveillance, Epidemiology, and End Results (SEER) Program database (20) of the U.S. National Cancer Institute is a large cancer-related database that includes information on cancer incidence, treatment, and outcomes from multiple institutions across the United States since 1973. It contains a substantial number of DCIS patients (21,22). In this study, we developed two nomograms based on extensive sample data from the SEER database to estimate the 5- and 10-year risks of IBTR in patients with DCIS treated with BCS. These nomogram models were then used to stratify DCIS patients after BCS according to their risk of IBTR, and targeted treatment plans were formulated for patients with different risk levels for clinical reference. We present this article in accordance with the TRIPOD reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-100/rc).

Methods

Data extraction

We searched and downloaded the medical records of female patients with DCIS from the SEER database (SEER Research Plus Data, 18 Registries, Nov 2020 Sub, 2000–2018), which covers cancer incidence and survival records for more than one-third of the U.S. population. This study retrospectively investigates the probability of ipsilateral second breast cancer within 10 years in patients who were first diagnosed with Tis breast cancer and underwent BCS between 2000 and 2008. We set a 10-year observation period to ensure that each patient had been observed for at least 10 years, with death treated as a censored event. The screening process was mainly divided into three steps: the first step was to select patients who experienced recurrence within 10 years, the second step was to select patients without recurrence within 10 years, and the third step was to combine the patient information from the first and second steps for further screening, as detailed in Figure 1. Researchers have obtained formal authorization and explicit consent from the SEER program to access and use these data, ensuring patient privacy is protected throughout the process This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Figure 1 The flow diagram of participant inclusion and exclusion. N stage, node stage; SEER, Surveillance, Epidemiology, and End Results; T-stage, tumor stage.

The clinicopathological characteristics of the patients included the year of diagnosis (2000–2008), primary site (upper outer quadrant, upper inner quadrant, lower outer quadrant, lower inner quadrant, central portion and nipple, other), age at diagnosis (20–44, 45–55, 56–80, >81 years), race (White, Black, Asian or Pacific Islander, other), histological type (intraductal carcinoma-NOS, cribriform, comedo, solid type, micropapillary, papillary adenocarcinoma, other), tumor size (0 < to ≤3, >3 cm, unknown/other), radiation therapy (yes, no/unknown), grade (grade I–II, grade III–IV, unknown), and HR status (negative/unknown, positive). The outcome variables were the presence or absence of recurrence and the time to recurrence.

Statistical analysis

Patient and tumor characteristics were presented as percentages, and the data were analyzed using chi-squared tests or Fisher’s exact tests. A statistically significant difference was considered when P<0.05. All statistical analyses were performed using SPSS version 25.0. To evaluate the relative risk of each predictive factor on the outcomes, the dataset was randomly divided into training and validation cohorts at a 3:1 ratio. As shown in Table S1, there were no significant differences in baseline variables between the training and validation sets. Univariate and multivariate Cox regression analyses were subsequently conducted on the training set. Hazard ratios were calculated based on the Cox proportional hazards model, with adjustments made for primary site, age, race, histology, tumor size, radiation therapy, grade, and HR status. Cox proportional hazards regression analysis was used to compare breast cancer-specific survival (BCSS) between patients with recurrence and those without, as well as the proportion of recurrence-free patients in different risk groups. Competing risk analysis was used to estimate the cumulative incidence of competing events, expressed as DCIS recurrence and invasive recurrence. The Gray test was used to assess differences in IBTR rates across various subgroups.

Model development and validation

Nomograms based on regression models, calibration curves, and survival-related curves were generated using the RMS, Foreign, Survival, Cmprsk, and other software packages. A two-tailed P value <0.05 was considered statistically significant (*, P<0.05; **, P<0.01). Receiver operating characteristic (ROC) curves were used to describe the sensitivity and specificity of the constructed nomograms, while calibration plots were used to evaluate and validate their accuracy. The prediction model was validated using 500 bootstrap samples to prevent overfitting and obtain relatively unbiased estimates. Based on the risk scores of different tumor characteristics in the nomogram, we identified patients at extremely high risk, normal risk, and extremely low risk.

Results

Baseline characteristics of enrolled patients

As shown in Table 1, a total of 40,770 eligible patients were collected for this study, of which 2,614 experienced IBTR within 10 years, accounting for 6.41%. In the stratified analysis by year of diagnosis, the total number of cohort patients increased gradually from 2000 to 2008, and the number of IBTR patients also showed an upward trend, though the proportion fluctuated. The variable ‘Laterality’ did not show a statistically significant difference (P=0.065), whereas all other clinical predictive factors showed significant differences (P<0.01). The majority of patients had their primary lesion located in the upper outer quadrant (n=14,417, 35.36%). Compared to patients without recurrence, patients with recurrence showed significantly higher proportions in certain clinicopathological subgroups. For example, a higher proportion of IBTR was observed in the central portion and nipple region (8.46%). Additionally, the 18–44 and >81 years age groups (8.93%, 11.73%), Black patients (9.19%), patients with comedocarcinoma histology (7.22%), those with grade III–IV differentiation (6.74%), tumor size greater than 3 cm (6.74%), patients with hormone status classified as negative/unknown (6.89%), and patients who did not receive adjuvant radiotherapy (9.17%) had the highest proportions in their respective groups. Figure 2A shows a comparison of BCSS between IBTR and non-IBTR patients, with non-IBTR patients demonstrating higher survival rates. Competing risk analysis of the first recurrence probability indicated that the risk of invasive recurrence was higher than that of DCIS recurrence (Figure 2B).

Figure 2 BCSS and competing risks of recurrence among DCIS patients with or without IBTR. (A) Comparison of survival rates between patients with IBTR and those without IBTR. (B) Ten-year competing risk analysis of DCIS vs. invasive recurrence types among patients with IBTR. BCSS, breast cancer-specific survival; CI, confidence interval; DCIS, ductal carcinoma in situ; IBTR, ipsilateral breast tumor recurrence.

Univariate and multivariate Cox regression analysis

As shown in Table 2, univariate and multivariate Cox regression analyses were performed on all predictive factors in the training set of the overall population. Regarding the primary site, compared to the upper inner quadrant, the risk of ipsilateral recurrence did not significantly differ for the upper outer (P=0.962) and lower outer quadrants (P=0.057), while the central region had the highest risk of ipsilateral recurrence [hazard ratio =1.45, 95% confidence interval (CI): 1.21–1.75], followed by the lower inner quadrant (hazard ratio =1.33, 95% CI: 1.10–1.62). In terms of age, compared to the 56–80 years age group, the 18–44 years age group (hazard ratio =1.61, 95% CI: 1.45–1.80) and the 80+ years age group (hazard ratio =1.82, 95% CI: 1.51–2.19) had significantly higher risks of ipsilateral recurrence. The 45–55 years age group did not show statistically significant differences in the univariate analysis (P=0.122). In terms of race, Black patients had the highest risk of ipsilateral recurrence (hazard ratio =1.53, 95% CI: 1.36–1.71), followed by Asian or Pacific Islander patients (hazard ratio =1.16, 95% CI: 1.02–1.32). Regarding histological type, there was no statistical difference in recurrence risk for DCIS (NOS), solid type, micropapillary carcinoma, or papillary carcinoma (P=0.548), while carcinoma with necrosis had the highest recurrence risk (hazard ratio =1.15, 95% CI: 1.01–1.30). In terms of grade, high-grade patients had a significantly higher recurrence risk compared to low-grade patients (hazard ratio =1.19, 95% CI: 1.09–1.31). In terms of tumor size, tumors larger than 3 cm had the highest risk of ipsilateral recurrence (hazard ratio =1.45, 95% CI: 1.19–1.77). Regarding HR status, univariate regression analysis showed that HR-positive patients had a lower risk of ipsilateral recurrence compared to HR-negative patients (hazard ratio =0.84, 95% CI: 0.77–0.90). However, multivariate regression analysis did not show statistical significance (P=0.091). Regarding adjuvant radiotherapy, patients who received adjuvant radiotherapy had a lower risk of ipsilateral recurrence compared to those who did not receive radiotherapy (hazard ratio =0.51, 95% CI: 0.47–0.55).

IBTR prediction model

Among the initial 40,770 patients, 30,577 were allocated to the training cohort (75%), and 10,193 were allocated to the validation cohort (25%). A nomogram was developed based on the clinicopathological characteristics of the 30,577 patients, as shown in Figure 3A. The concordance index (C-index) for the 10-year ROC curve in the validation cohort was 0.682 (95% CI: 0.653–0.710), indicating strong agreement between the observed outcomes and predicted probabilities (see Figure 3B). The 10-year calibration curve in the validation cohort was close to the diagonal line for each sampling, showing no deviation from our expected predictions (see Figure 3C). Similarly, we identified 26,095 patients who received adjuvant radiotherapy and 14,675 patients who did not from the entire cohort of 40,770 patients, and constructed two subgroup-specific nomograms using the same methodology. The 10-year AUC value of the radiotherapy subgroup was 0.635 (95% CI: 0.595–0.675), and that of the non-radiotherapy subgroup was 0.632 (95% CI: 0.604–0.660) (as shown in Figure 3D-3I).

Figure 3 Nomogram and related concordance assessment metrics. (A) Nomogram based on the training cohort of the overall population. (B) ROC curve of the test set within 10 years for validating the nomogram of the overall population. The concordance index (C-index) for the 10-year ROC curve in the validation cohort was 0.682 (95% CI: 0.653-0.710), indicating strong agreement between the observed outcomes and predicted probabilities. (C) Calibration curve for validating the nomogram of the overall population. The prediction model was validated using 500 bootstrap samples to prevent overfitting and obtain relatively unbiased estimates. The 10-year calibration curve in the validation cohort was close to the diagonal line for each sampling, showing no deviation from our expected predictions. (D) Nomogram developed from the training set of the radiotherapy population. (E) Ten-year ROC curve of the test set for validating the nomogram in the non-radiotherapy population. (F) Calibration curve for validating the nomogram in the non-radiotherapy population. (G) Nomogram developed from the training set of the non-radiotherapy population. (H) ROC curve of the test set for validating the nomogram of the radiotherapy population. (I) Calibration curve for validating the nomogram of the radiotherapy population. CI, confidence interval; HR, hormone receptor; IBTR, ipsilateral breast tumor recurrence; NOS, not otherwise specified; ROC, receiver operating characteristic.

Comparison of recurrence rates between patients with and without radiotherapy in the low-risk group

Based on the nomogram model constructed from the non-radiotherapy population (Figure 3B), a score of 44 corresponded to a 10-year recurrence rate of 5%. Therefore, patients with a score ≤44 were defined as the extremely low-risk group in this dataset. Furthermore, using the nomogram in Figure 3B, we identified patients with scores ≤44 from the overall population, both with and without adjuvant radiotherapy, and compared their recurrence rates. The results showed no significant difference in recurrence risk between the two groups (P=0.065, see Figure 4A).

Figure 4 Risk-stratified effects of adjuvant radiotherapy on ipsilateral recurrence in DCIS. (A) Comparison of IBTR rates in low-risk individuals: adjuvant radiotherapy vs. no adjuvant radiotherapy. (B) Comparison of IBTR risk between extremely high-risk and normal-risk patients receiving postoperative radiotherapy. CI, confidence interval; DCIS, ductal carcinoma in situ; IBTR, ipsilateral breast tumor recurrence.

Recurrence status in high-risk patients receiving adjuvant radiotherapy

Using the nomogram developed from the dataset of patients who received adjuvant radiotherapy (Figure 3C), it was observed that a score of 76 corresponded to a 10-year recurrence rate of 10%. Therefore, in this dataset, patients with a score greater than or equal to 76 were classified as the very high-risk group, while the remainder were classified as the normal-risk group. Compared to the very high-risk group, the recurrence risk was significantly reduced in the normal-risk group (P<0.001) (Figure 4B).

Comparison of IBTR rates in subgroups

In the comparison across different age groups, we observed the highest recurrence risk in patients aged 81 years and above (P<0.001), followed by those aged 18–44 years. Patients aged 45–55 and 56–80 years had significantly lower recurrence risks than those aged 18–44 years (P<0.001). A comparison across different primary sites showed that the recurrence risk for peripheral DCIS was significantly lower than that for central and nipple regions (P<0.001). The larger the tumor, the higher the IBTR rate (P<0.001). In the comparison based on different HR statuses, patients with HR-positive (HR+) tumors had a significantly lower recurrence risk than those with HR-negative (HR−) or unknown HR status (P<0.001). Comparison of different tumor grades showed that patients with grade III–IV tumors had a significantly higher recurrence risk than those with grade I–II tumors (P<0.001). Analysis of different racial groups indicated that Black patients had the highest recurrence risk (P<0.001). Patients who received adjuvant radiotherapy had a significantly lower recurrence risk than those who did not (P<0.001). A comparison based on different histological types revealed that patients with cribriform carcinoma had a significantly lower recurrence risk than those with comedocarcinoma, followed by patients with intraductal carcinoma (NOS) (P<0.001) (Figure 5).

Figure 5 Comparison of IBTR rates in subgroups. (A) In the comparison across different age groups, we observed the highest recurrence risk in patients aged 81 years and above (P<0.001), followed by those aged 18–44 years. Patients aged 45–55 and 56–80 years had significantly lower recurrence risks than those aged 18–44 years (P<0.001). (B) A comparison across different primary sites showed that the recurrence risk for peripheral DCIS was significantly lower than that for central and nipple regions (P<0.001). (C) The larger the tumor, the higher the IBTR rate (P<0.001). (D) In the comparison based on different HR statuses, patients with HR-positive (HR+) tumors had a significantly lower recurrence risk than those with HR-negative (HR−) or unknown HR status (P<0.001). (E) Comparison of different tumor grades showed that patients with grade III–IV tumors had a significantly higher recurrence risk than those with grade I–II tumors (P<0.001). (F) Analysis of different racial groups indicated that Black patients had the highest recurrence risk (P<0.001). (G) Patients who received adjuvant radiotherapy had a significantly lower recurrence risk than those who did not (P<0.001). (H) A comparison based on different histological types revealed that patients with cribriform carcinoma had a significantly lower recurrence risk than those with comedocarcinoma, followed by patients with intraductal carcinoma (P<0.001). ***, P<0.001. DCIS, ductal carcinoma in situ; HR, hormone receptor; IBTR, ipsilateral breast tumor recurrence.

Discussion

To make recurrence prediction more universal and cost-effective for DCIS patients after BCS, we developed a nomogram using a large sample of data and multiple practical clinical predictive factors. This nomogram allows for a simple yet accurate assessment of the 10-year recurrence risk probability in these patients. It helps clinicians better evaluate recurrence risk and the need for further treatment. The concordance index for the validation set over 10 years was 0.682 (95% CI: 0.653–0.710), which aligns with the concordance index of the nomogram constructed by Memorial Sloan Kettering Cancer Center (concordance index of 0.70, with an internal validation of 0.69) (18). This indirectly supports the accuracy of the nomogram we developed. The 10-year calibration curve for the validation set was close to the diagonal line for each sampling point, with each calibration point near the 45-degree line, indicating good agreement between observed and predicted results.

Before constructing the nomogram, we focused on whether recurrence contributes to increased mortality. As shown in Figure 2A, the BCSS analysis demonstrated a statistically significant difference (P<0.001), indicating that patients with IBTR belong to the high-risk mortality group. Similarly, Wapnir et al. found that patients who developed IBTR—especially those with invasive recurrences—exhibited significantly higher breast cancer-specific mortality than patients without recurrence (23). In their analysis of long-term trial data, an invasive IBTR conferred approximately a 75% increase in the risk of breast cancer-related death (P<0.001), whereas a recurrence confined to DCIS did not adversely affect mortality. Notably, over half of the deaths following an invasive recurrence in Wapnir’s study were attributed to breast cancer. These findings reinforce that IBTR, particularly invasive IBTR, is a strong predictor of poorer BCSS, aligning with our observation that patients with IBTR belong to a high-risk mortality group.

To assess the aggressiveness of tumors after recurrence, we classified the second cancer event based on the T-stage information of the patients and performed a competing risk analysis. The results indicated a higher likelihood of invasive recurrence compared to recurrence as DCIS (Figure 2B). The EBCTCG meta-analysis on breast-conserving therapy for DCIS reported that approximately half of IBTR were invasive cancers, which may differ from the findings of our study (24,25). In our analysis, we observed that among DCIS patients, invasive recurrences predominated within 10 years after surgery, accounting for approximately 40% of all IBTRs, while in situ recurrences accounted for about 20%, resulting in a ratio of roughly 2:1. The remaining 40% of recurrences occurred after 10 years. Several factors may contribute to this discrepancy. First, we applied strict inclusion and exclusion criteria to the SEER dataset, excluding patients who did not undergo BCS, were diagnosed outside the 2000–2008 window, experienced IBTR within 6 months, died within 10 years, had contralateral breast cancer, lobular carcinoma, positive lymph nodes, or received chemotherapy. These criteria yielded a homogeneous cohort characterized by BCS and limited systemic treatment. As a result, early in situ recurrences—potentially due to margin involvement—were significantly reduced, and most recurrences observed were delayed and more likely to manifest as new invasive tumors (26).

Previously, the most common models for assessing or predicting the risk of local recurrence after BCS for DCIS included the University of Southern California/Van Nuys Prognostic Index (USC/VNPI) published in 1996, the nomogram constructed by Memorial Sloan Kettering Cancer Center, the Oncotype DX DCIS score, the NCCN prognostic index, and DCISionRT. The USC/VNPI, published in 1996, is based on patients’ age, tumor size and grade, and margin width, classifying patients into three risk groups and offering stratified treatment recommendations (17). This model provides some reference value for assessing the probability of recurrence, but it lacks reliable independent validation (27,28).

A retrospective analysis of 222 cases of DCIS detected by mammography and treated with surgical excision without radiotherapy revealed no significant difference in breast tumor recurrence among the low-, intermediate-, and high-risk USC/VNPI groups over five years. Therefore, the model has not been well validated in actual clinical practice (29). Subsequently, Memorial Sloan Kettering Cancer Center (MSKCC) developed a nomogram based on 1,681 consecutive patients with DCIS who underwent BCS. This nomogram includes 10 clinical and pathological variables (such as ET, radiotherapy, palpability, age, family history, margin status, number of excisions, grade, necrosis, and year of surgery) to predict the 5- and 10-year risks of ipsilateral breast recurrence (18). While the MSKCC model incorporates various clinicopathological factors to predict recurrence, some variables, such as the “year of surgery”, may limit its clinical generalizability in contemporary practice. Although this variable was appropriately included to account for evolving treatment patterns over time when the model was developed, it may not directly reflect tumor biology and is less applicable when predicting outcomes for patients treated under modern standardized protocols. Moreover, the score does not include tumor size or molecular markers [such as estrogen receptor (ER)/progesterone receptor (PR)], which are now recognized as important prognostic factors for DCIS (30,31). In addition to predictive models incorporating clinicopathological factors, there are currently models that include genetic predictors. The Oncotype DX DCIS score (32) is based on the expression of 12 genes, including 7 proliferation genes (Ki67, STK15, Survivin, CCNB1, and MYBL2, PR, and GSTM1), as well as 5 reference genes (ACTB, GAPDH, RPLPO, GUS, TFRC). It provides data on the risk of local recurrence and invasive local recurrence for DCIS patients and identifies which patients may benefit from radiotherapy (32). The DCISionRT model, developed by PreludeDx, determines the benefit of radiotherapy by analyzing a set of markers (HER2, PR, Ki-67, COX2, p16/INK4A, FOXA1, and SIAH2) and 4 clinicopathological features (such as age, tumor size, palpability, and margin status). However, it is important to note that both Oncotype DX DCIS and DCISionRT are proprietary assays with limited transparency. To date, no independent studies have published external validation metrics such as area under the ROC curve or calibration curves for these models in separate validation populations. This lack of published accuracy data raises concerns regarding their generalizability and cost-effectiveness in broader clinical practice (19).

Additionally, we conducted subgroup analyses for each predictive factor. The results indicated statistically significant differences in recurrence outcomes across the subgroups of each predictive factor (Figure 4), suggesting that each subgroup carries a certain weight in influencing outcomes. Therefore, we ranked the recurrence risk from high to low based on the following predictive factors: age (81+ > 18–44 > 45–55 > 56–80 years), race (Black > Asian or Pacific Islander > White), radiotherapy (none/unknown > yes), grade (III–IV > I–II), HR status (negative/unknown > positive), primary site (central & nipple > lower inner > lower outer > upper inner > upper outer), histologic type (ductal carcinoma > DCIS, NOS/papillary carcinoma/solid type/micropapillary carcinoma > cribriform carcinoma), tumor size (3+ > 0–3 cm).

Previous studies have primarily focused on identifying which types of DCIS patients can be exempt from radiation therapy after BCS (19). When the predicted 10-year IBTR risk is less than 5%, breast radiation therapy is generally recommended to be omitted (33-35). In fact, a subset of low-risk patients may have a low recurrence risk even without radiotherapy, suggesting the potential for treatment de-escalation. Currently, the literature supports using a 10-year recurrence rate of ≤5% as the threshold for the low-risk group (36), though some studies propose a threshold of ≤7.5% (37). Based on this, we adopted a 10-year recurrence rate of ≤5% as the cutoff to define the ultra-low-risk cohort. Using a nomogram built from population-based datasets, we performed a comparative analysis of IBTR survival rates in the ultra-low-risk cohort (10-year recurrence rate ≤5%) with and without adjuvant radiotherapy. The results showed no statistically significant difference in recurrence risk between patients who received adjuvant radiotherapy and those who did not (P=0.065, see Figure 4B). Previous research has shown that while adjuvant radiotherapy reduces recurrence rates, it does not significantly impact overall survival, indicating that treatment de-escalation could be considered for this ultra-low-risk population (23). Moreover, studies are currently exploring whether surgery can also be omitted for this group, in addition to radiotherapy, as evidenced by four ongoing large clinical trials [LORIS (38), COMET (39), LORD (40), LORETTA (41)] investigating active surveillance in DCIS patients. Due to recruitment challenges and slow progress, no results have been reported from these trials so far. Therefore, accurately predicting recurrence risk after BCS in DCIS patients is of great value, enabling individualized treatment and follow-up plans that achieve the optimal balance between benefit and minimizing toxicity and harm to the patient.

Additionally, using our nomogram model (Figure 3B), we identified a subset of high-risk DCIS patients (i.e., those who received radiotherapy with a 10-year recurrence rate >10%) (34,36). This group exhibited a statistically significant difference in recurrence risk compared to the normal-risk group after receiving BCS combined with radiotherapy (P<0.001). The increased recurrence risk suggests that despite adjuvant radiotherapy, these patients still face a high risk of recurrence, indicating the need for more intensive monitoring and additional adjuvant therapies.

In a meta-analysis of randomized trials conducted by the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), the 10-year IBTR rate among patients with DCIS treated with BCS alone was 28.1%. With the addition of adjuvant radiotherapy, this rate was reduced to 12.9%, corresponding to a relative risk reduction of over 50% (25). The EORTC study reached a similar conclusion, with 6-year follow-up data showing that IBTR recurrence rates were reduced by ≥50% in patients who received BCS + radiotherapy compared to those treated with BCS alone (42). These findings confirm the positive effect of radiotherapy following BCS in DCIS patients.

Of particular interest is the subset of patients who experience recurrence after BCS + radiotherapy, as they remain at a high risk of recurrence. This raises the question: Can we accurately identify this patient population and provide them with more advanced treatment? Studies have shown that adding ET such as tamoxifen to the BCS + radiotherapy regimen can further reduce recurrence risk (both ipsilateral and contralateral) in DCIS patients (43-45). The NSABP B-24 trial demonstrated that the addition of tamoxifen significantly reduced the cumulative incidence of breast events, including both ipsilateral and contralateral breast cancers. At 5 years, the event rate was 8.2% in the tamoxifen group compared to 13.4% in the placebo group (P=0.0009) (43). Tamoxifen was particularly effective in reducing both ipsilateral recurrence and contralateral new primary tumors, regardless of margin status or comedo necrosis. Similarly, long-term results from the UK/ANZ DCIS trial showed that tamoxifen significantly reduced all new breast events by 29% (hazard ratio =0.71, 95% CI: 0.58–0.88, P=0.002), with a notable reduction in ipsilateral DCIS recurrence (hazard ratio =0.70, P=0.03) and an even greater effect in reducing contralateral breast cancer (hazard ratio =0.44, P=0.005) (46). These findings support the use of tamoxifen as an effective adjuvant therapy to further reduce recurrence risk in DCIS patients treated with BCS and radiotherapy. However, in our study HR-negative status did not remain a significant independent predictor of recurrence on multivariate analysis (P=0.091). As we know, ER+ DCIS treated with ET would have a significantly lower IBTR rate (47). One possible explanation is confounding by other risk factors that co-occur with HR positivity. Notably, chi-squared analysis of our cohort (Table S2) showed that HR-positive DCIS cases were disproportionately represented among Black patients and those with tumors >3 cm. Black women with DCIS have been reported to experience higher local recurrence rates and worse outcomes compared to White patients (48,49), and large DCIS lesion size is a well-established independent risk factor for recurrence (50). The clustering of these adverse prognostic factors in the HR-positive subgroup could thus offset the usual protective effect of HR expression, explaining why HR status failed to reach significance in the adjusted model.

Additionally, some studies suggest a benefit from HER2-targeted therapy. An analysis of data from the National Cancer Database (NCDB) of DCIS patients from 2004 to 2015 screened 1,927 biopsy-confirmed DCIS cases, of which 430 patients (22.3%) received anti-HER2 targeted therapy, while 1,497 patients (77.7%) did not. The results showed that patients who received targeted therapy had a higher 5-year OS compared to those who did not (97.7% vs. 95.8%, P=0.043), indicating a survival benefit from anti-HER2 therapy (51). This is further supported by a large NCDB-based study of 1,927 HER2-positive DCIS patients, which demonstrated that receipt of HER2-targeted therapy was associated with improved 5-year overall survival (97.7% vs. 95.8%, P=0.043) and remained an independent predictor of better outcomes in multivariate analysis (hazard ratio =0.348, P=0.046) (50). Besides, a 2017 JAMA Oncology study of 4,131 DCIS patients showed that adding a radiotherapy boost after whole-breast radiotherapy significantly reduced IBTR (hazard ratio =0.68, P=0.01), with a 15-year absolute benefit of 3.6%, particularly in patients with negative margins. Combined with endocrine and HER2-targeted therapies, radiotherapy boost offers a valuable strategy to further lower recurrence risk in DCIS patients with a life expectancy exceeding 10–15 years (52). In terms of surgical options, studies suggest that for this high-risk patient group, unilateral mastectomy or nipple-sparing mastectomy may be more recommended surgical approaches (53). Therefore, accurately identifying this high-risk population and exploring new, intensified treatment regimens is essential.

Finally, it is undeniable that there are certain limitations in this study. The SEER database does not provide information regarding surgical margins (54), Ki-67 expression, multifocality, ET (55), presence of tumor necrosis (56), BRCA1/2 mutations (57), comorbidities, family history (58), whether DCIS was diagnosed based on symptoms or discovered through mammography, and whether radiotherapy included a boost to the tumor bed. Many retrospective studies have reported that the recurrence rate of DCIS largely depends on the status of surgical margins, with wider margins being associated with a lower risk of local recurrence (59). Furthermore, the SEER database lacks descriptions of the multicentric nature of DCIS in histopathology. It is known that DCIS can involve the main duct and different levels of branching ducts, so single-center versus multicentric involvement may imply different recurrence risks (18). The inclusion of this critical information could help further refine our nomogram model and improve predictive accuracy. In the future, we plan to incorporate more potentially valuable clinicopathological indicators through domestic or international multicenter collaborations to optimize the predictive power and accuracy of our model.

Conclusions

To our knowledge, our study is the largest population-based cohort study to date. In this study, we developed a nomogram to predict the 10-year probability of IBTR in patients with DCIS after BCS. Furthermore, based on this nomogram and the corresponding IBTR analysis, we were able to classify patients with a 10-year recurrence rate greater than 10% as a high-risk group, indicating the need for intensified treatment. Conversely, patients with a recurrence rate below 5% were categorized as a very low-risk group, suggesting the potential for de-escalation of treatment. Overall, our study stratified DCIS patients’ recurrence risk after BCS based on readily accessible clinical and pathological factors. This stratification can provide precise adjuvant management and follow-up for these patients. Personalized treatment plans for DCIS can be strategized using our model, and this study lays the groundwork for future clinical trials on treatment de-escalation and escalation.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-100/rc

Peer Review File: Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-100/prf

Funding: The work was supported by the National Natural Science Foundation of China (Nos. 82272855, 81972453, 81972597), Natural Science Foundation of Zhejiang Province (Nos. LR22H160011, LY19H160055, LY19H160059, LY18H160005, and LY20H160026), and Medical and Health Science and Technology of Zhejiang Province (Youth Talent Program) Project (No. 2021RC016). The work was also sponsored by Zheng Shu Medical Elite Scholarship Fund.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-100/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Woolston C. Breast cancer. Nature 2015;527:S101. [Crossref] [PubMed]
2. Sakorafas GH, Farley DR, Peros G. Recent advances and current controversies in the management of DCIS of the breast. Cancer Treat Rev 2008;34:483-97. [Crossref] [PubMed]
3. Salvatorelli L, Puzzo L, Vecchio GM, et al. Ductal carcinoma in situ of the breast: an update with emphasis on radiological and morphological features as predictive prognostic factors. Cancers (Basel) 2020;12:609. [Crossref] [PubMed]
4. IARC [Internet]. International Agency for Research on Cancer. Available online: https://www.iarc.who.int/featured-news/global-cancer-burden-growing-amidst-mounting-need-for-services/
5. Fu F, Gilmore RC, Jacobs LK. Ductal carcinoma in situ. Surg Clin North Am 2018;98:725-45. [Crossref] [PubMed]
6. Malhotra RK, Indrayan A. A simple nomogram for sample size for estimating sensitivity and specificity of medical tests. Indian J Ophthalmol 2010;58:519-22. [Crossref] [PubMed]
7. Fisher MJ, Marshall AP, Mitchell M. Testing differences in proportions. Aust Crit Care 2011;24:133-8. [Crossref] [PubMed]
8. Barrio AV, Van Zee KJ. Controversies in the treatment of ductal carcinoma in situ. Annu Rev Med 2017;68:197-211. [Crossref] [PubMed]
9. Shah C, Wobb J, Manyam B, et al. Management of ductal carcinoma in situ of the breast: a review. JAMA Oncol 2016;2:1083-8. [Crossref] [PubMed]
10. Rezai M, Kraemer S, Kimmig R, et al. Breast conservative surgery and local recurrence. Breast 2015;24:S100-7. [Crossref] [PubMed]
11. Ekatah GE, Turnbull AK, Arthur LM, et al. Margin width and local recurrence after breast conserving surgery for ductal carcinoma in situ. Eur J Surg Oncol 2017;43:2029-35. [Crossref] [PubMed]
12. Greenberg CC, Habel LA, Hughes ME, et al. Characterization and treatment of local recurrence following breast conservation for ductal carcinoma in situ. Ann Surg Oncol 2014;21:3766-73. [Crossref] [PubMed]
13. Jordan RM, Oxenberg J. Breast cancer conservation therapy. In: Treasure Island (FL); 2024.
14. Lazzeroni M, Puntoni M, Provinciali N, et al. Estimating the magnitude of clinical benefit of systemic therapy in patients with DCIS or pre-invasive disease of the breast. Breast 2019;48:S39-43. [Crossref] [PubMed]
15. Williams KE, Barnes NLP, Cramer A, et al. Molecular phenotypes of DCIS predict overall and invasive recurrence. Ann Oncol 2015;26:1019-25. [Crossref] [PubMed]
16. Benson JR, Wishart GC. Predictors of recurrence for ductal carcinoma in situ after breast-conserving surgery. Lancet Oncol 2013;14:e348-57. [Crossref] [PubMed]
17. Silverstein MJ, Lagios MD, Craig PH, et al. A prognostic index for ductal carcinoma in situ of the breast. Cancer 1996;77:2267-74. [Crossref] [PubMed]
18. Rudloff U, Jacks LM, Goldberg JI, et al. Nomogram for predicting the risk of local recurrence after breast-conserving surgery for ductal carcinoma in situ. J Clin Oncol 2010;28:3762-9. [Crossref] [PubMed]
19. Shah C, Bremer T, Cox C, et al. The clinical utility of DCISionRT(®) on radiation therapy decision making in patients with ductal carcinoma in situ following breast-conserving surgery. Ann Surg Oncol 2021;28:5974-84. [Crossref] [PubMed]
20. The Surveillance, Epidemiology, and End Results (SEER) Program database [Internet]. Available online: https://seer.cancer.gov/explorer/
21. Hankey BF, Ries LA, Edwards BK. The surveillance, epidemiology, and end results program: a national resource. Cancer Epidemiol Biomarkers Prev 1999;8:1117-21.
22. Hayat MJ, Howlader N, Reichman ME, et al. Cancer statistics, trends, and multiple primary cancer analyses from the Surveillance, Epidemiology, and End Results (SEER) Program. Oncologist 2007;12:20-37. [Crossref] [PubMed]
23. Wapnir IL, Dignam JJ, Fisher B, et al. Long-term outcomes of invasive ipsilateral breast tumor recurrences after lumpectomy in NSABP B-17 and B-24 randomized clinical trials for DCIS. J Natl Cancer Inst 2011;103:478-88. [Crossref] [PubMed]
24. Stuart KE, Houssami N, Taylor R, et al. Long-term outcomes of ductal carcinoma in situ of the breast: a systematic review, meta-analysis and meta-regression analysis. BMC Cancer 2015;15:890. [Crossref] [PubMed]
25. Early Breast Cancer Trialists' Collaborative Group (EBCTCG). Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials. Lancet 2011;378:1707-16. [Crossref] [PubMed]
26. Byng D, Retèl VP, Schaapveld M, et al. Treating (low-risk) DCIS patients: What can we learn from real-world cancer registry evidence? Breast Cancer Res Treat 2021;187:187-96. [Crossref] [PubMed]
27. Silverstein MJ, Poller DN, Waisman JR, et al. Prognostic classification of breast ductal carcinoma-in-situ. Lancet 1995;345:1154-7. [Crossref] [PubMed]
28. Silverstein MJ, Barth A, Poller DN, et al. Ten-year results comparing mastectomy to excision and radiation therapy for ductal carcinoma in situ of the breast. Eur J Cancer 1995;31A:1425-7. [Crossref] [PubMed]
29. MacAusland SG, Hepel JT, Chong FK, et al. An attempt to independently verify the utility of the Van Nuys Prognostic Index for ductal carcinoma in situ. Cancer 2007;110:2648-53. [Crossref] [PubMed]
30. Mazouni C, Delaloge S, Rimareix F, et al. Nomogram for risk of relapse after breast-conserving surgery in ductal carcinoma in situ. J Clin Oncol 2011;29:e44-author reply e45-6. [Crossref] [PubMed]
31. Ballehaninna UK, Chamberlain RS. Inclusion of tumor biology molecular markers to improve the ductal carcinoma in situ ipsilateral breast tumor recurrence nomogram predictability. J Clin Oncol 2011;29:e97-8; author reply e99. [Crossref] [PubMed]
32. Solin LJ, Gray R, Baehner FL, et al. A multigene expression assay to predict local recurrence risk for ductal carcinoma in situ of the breast. J Natl Cancer Inst 2013;105:701-10. [Crossref] [PubMed]
33. Stenmark Tullberg A, Sjöström M, Niméus E, et al. Integrating Tumor-Intrinsic and Immunologic Factors to Identify Immunogenic Breast Cancers from a Low-Risk Cohort: Results from the Randomized SweBCG91RT Trial. Clin Cancer Res 2023;29:1783-93. [Crossref] [PubMed]
34. Rakovitch E, Parpia S, Koch A, et al. DUCHESS: an evaluation of the ductal carcinoma in situ score for decisions on radiotherapy in patients with low/intermediate-risk DCIS. Breast Cancer Res Treat 2021;188:133-9. [Crossref] [PubMed]
35. Pilewskie M, Olcese C, Patil S, et al. Women with low-risk DCIS eligible for the LORIS trial after complete surgical excision: how low is their risk after standard therapy? Ann Surg Oncol 2016;23:4253-61. [Crossref] [PubMed]
36. Paszat L, Sutradhar R, Zhou L, et al. Including the ductal carcinoma-in-situ (DCIS) score in the development of a multivariable prediction model for recurrence after excision of DCIS. Clin Breast Cancer 2019;19:35-46. [Crossref] [PubMed]
37. Rakovitch E, Nofech-Mozes S, Hanna W, et al. Multigene expression assay and benefit of radiotherapy after breast conservation in ductal carcinoma in situ. J Natl Cancer Inst 2017;109:djw256. [Crossref] [PubMed]
38. Kanbayashi C, Thompson AM, Hwang ESS, et al. The international collaboration of active surveillance trials for low-risk DCIS (LORIS, LORD, COMET, LORETTA). J Clin Oncol 2019;37:TPS603.
39. Elshof LE, Tryfonidis K, Slaets L, et al. Feasibility of a prospective, randomised, open-label, international multicentre, phase III, non-inferiority trial to assess the safety of active surveillance for low risk ductal carcinoma in situ - The LORD study. Eur J Cancer 2015;51:1497-510. [Crossref] [PubMed]
40. Francis A, Bartlett J, Billingham L, et al. Abstract OT1-03-01: The UK LORIS trial: Randomizing patients with low or low intermediate grade ductal carcinoma in situ (DCIS) to surgery or active monitoring. Cancer Res 2017;77:OT1-03-01-OT1-03–01.
41. Kanbayashi C, Iwata H. Current approach and future perspective for ductal carcinoma in situ of the breast. Jpn J Clin Oncol 2017;47:671-7. [Crossref] [PubMed]
42. Julien JP, Bijker N, Fentiman IS, et al. Radiotherapy in breast-conserving treatment for ductal carcinoma in situ: first results of the EORTC randomised phase III trial 10853. EORTC Breast Cancer Cooperative Group and EORTC Radiotherapy Group. Lancet 2000;355:528-33. [Crossref] [PubMed]
43. Fisher B, Dignam J, Wolmark N, et al. Tamoxifen in treatment of intraductal breast cancer: National Surgical Adjuvant Breast and Bowel Project B-24 randomised controlled trial. Lancet 1999;353:1993-2000. [Crossref] [PubMed]
44. Margolese RG, Cecchini RS, Julian TB, et al. Anastrozole versus tamoxifen in postmenopausal women with ductal carcinoma in situ undergoing lumpectomy plus radiotherapy (NSABP B-35): a randomised, double-blind, phase 3 clinical trial. Lancet 2016;387:849-56. [Crossref] [PubMed]
45. Forbes JF, Sestak I, Howell A, et al. Anastrozole versus tamoxifen for the prevention of locoregional and contralateral breast cancer in postmenopausal women with locally excised ductal carcinoma in situ (IBIS-II DCIS): a double-blind, randomised controlled trial. Lancet 2016;387:866-73. [Crossref] [PubMed]
46. Cuzick J, Sestak I, Pinder SE, et al. Effect of tamoxifen and radiotherapy in women with locally excised ductal carcinoma in situ: long-term results from the UK/ANZ DCIS trial. Lancet Oncol 2011;12:21-9. [Crossref] [PubMed]
47. Allred DC, Anderson SJ, Paik S, et al. Adjuvant tamoxifen reduces subsequent breast cancer in women with estrogen receptor-positive ductal carcinoma in situ: a study based on NSABP protocol B-24. J Clin Oncol 2012;30:1268-73. [Crossref] [PubMed]
48. Liu Y, Colditz GA, Gehlert S, et al. Racial disparities in risk of second breast tumors after ductal carcinoma in situ. Breast Cancer Res Treat 2014;148:163-73. [Crossref] [PubMed]
49. Madubata CC, Liu Y, Goodman MS, et al. Comparing treatment and outcomes of ductal carcinoma in situ among women in Missouri by race. Breast Cancer Res Treat 2016;160:563-72. [Crossref] [PubMed]
50. Collins LC, Achacoso N, Haque R, et al. Risk factors for non-invasive and invasive local recurrence in patients with ductal carcinoma in situ. Breast Cancer Res Treat 2013;139:453-60. [Crossref] [PubMed]
51. Lewis GD, Haque W, Farach A, et al. The impact of HER2-directed targeted therapy on HER2-positive DCIS of the breast. Rep Pract Oncol Radiother 2021;26:179-87. [Crossref] [PubMed]
52. Moran MS, Zhao Y, Ma S, et al. Association of Radiotherapy Boost for Ductal Carcinoma In Situ With Local Control After Whole-Breast Radiotherapy. JAMA Oncol 2017;3:1060-8. [Crossref] [PubMed]
53. Curigliano G, Burstein HJ, Winer EP, et al. De-escalating and escalating treatments for early-stage breast cancer: the St. Gallen International Expert Consensus Conference on the Primary Therapy of Early Breast Cancer 2017. Ann Oncol 2017;28:1700-12. [Crossref] [PubMed]
54. Solin LJ, Gray R, Hughes LL, et al. Surgical Excision Without Radiation for Ductal Carcinoma in Situ of the Breast: 12-Year Results From the ECOG-ACRIN E5194 Study. J Clin Oncol 2015;33:3938-44. [Crossref] [PubMed]
55. Shen J, Hunt KK, Mirza NQ, et al. Predictors of systemic recurrence and disease-specific survival after ipsilateral breast tumor recurrence. Cancer 2005;104:479-90. [Crossref] [PubMed]
56. Fisher ER, Dignam J, Tan-Chiu E, et al. Pathologic findings from the National Surgical Adjuvant Breast Project (NSABP) eight-year update of Protocol B-17: intraductal carcinoma. Cancer 1999;86:429-38. [Crossref] [PubMed]
57. Garcia-Etienne CA, Barile M, Gentilini OD, et al. Breast-conserving surgery in BRCA1/2 mutation carriers: are we approaching an answer? Ann Surg Oncol 2009;16:3380-7. [Crossref] [PubMed]
58. Seynaeve C, Verhoog LC, van de Bosch LM, et al. Ipsilateral breast tumour recurrence in hereditary breast cancer following breast-conserving therapy. Eur J Cancer 2004;40:1150-8. [Crossref] [PubMed]
59. Pilewskie M, Morrow M. Margins in breast cancer: How much is enough? Cancer 2018;124:1335-41. [Crossref] [PubMed]