Long-term outcome data for patients with hormone receptor-positive early breast cancer participating in the WSG PlanB trial after preselection by gene expression analysis: 10-year survival results from the WSG PlanB registry Authors: U. Nitz, M. Graeser, O. Gluz, S. Kümmel, M. Just, C. Jackisch, C. zu Eulenburg, M. Christgen, N. Harbeck, West German Study Group Published: 2025 ABSTRACT BackgroundThe PlanB registry evaluated long-term follow-up data for clinical candidates for chemotherapy enrolled in the PlanB trial in hormone receptor (HR)-positive early breast cancer (eBC) patients preselected using the 21-gene expression assay.Patients and methodsPlanB randomly assigned pT1-4c, pN+, pN0/high-risk human epidermal growth factor receptor 2 (HER2)-negative eBC patients to four cycles of epirubicin + cyclophoshamide followed by four cycles of docetaxel (EC-T arm) or to six cycles of docetaxel + cyclophosphamide (TC); patients with locally HR-positive tumors and recurrence score (RS) <12 received endocrine therapy (ET) only. Following the end of PlanB, patients with HR-positive tumors were included in a prospective, noninterventional PlanB registry. The primary objective was to compare invasive disease-free survival (iDFS); the secondary objectives were the comparisons of overall survival (OS), and distant disease-free survival (dDFS).ResultsThe registry included 699 patients (ET only: n = 119, TC: n = 298, EC-T: n = 289). In the TC and EC-T groups, respectively, 41% and 41% were postmenopausal; 63% and 55% were pN0; and 23% and 24% had RS >25. Ten-year survival rates (10.3 years median follow-up) in the TC and EC-T arms were, respectively, 90.8% [95% confidence interval (CI) 86.6% to 93.7%] and 92.1% (95% CI 88.2% to 94.7%, P = 0.546) for iDFS; 94.4% (95% CI 90.9% to 96.5%) and 93.2% (95% CI 89.5% to 95.6%, P = 0.789) for dDFS, and 96.8% (95% CI 94.0% to 98.3%) and 96.3% (95% CI 93.3% to 98.0%, P = 0.974) for OS. Overall, 10-year OS was 94.2% (95% CI 88.9% to 97.1%) in RS <12 (77.9% received ET-only), 96.8% (95% CI 94.4% to 98.2%) in RS 12-25, and 96.1% (95% CI 90.9% to 98.4%) in RS >25 group. RS was not predictive of the efficacy of EC-T versus TC.ConclusionsTen-year outcomes for TC and EC-T were similar and excellent; six cycles of TC is an effective option in HER2-negative eBC with pN0 or pN1 and intermediate-to-high-risk disease, according to gene expression analysis. Introduction Anthracycline-free regimens have been incorporated into the standard of care options for neoadjuvant and adjuvant therapy in selected patients with human epidermal growth factor receptor 2 (HER2)-negative early breast cancer (eBC).1 Compared with sequential anthracycline and taxane therapy, anthracycline-free regimens have a shorter duration and are associated with fewer acute grade 3 and 4 toxicities, as well as reduced long-term toxicity, particularly in relation to acute myeloid leukemia (AML).2,3 In a recent meta-analysis, proportional recurrence risk reduction for anthracycline-containing therapy was similar in hormone receptor (HR)-positive and HR-negative eBC.2 Nevertheless, evidence for HR-positive disease was mainly derived from clinical trials in which the chemotherapy indication was guided by clinical risk assessment. In the meantime, another generation of trials has emerged that investigated guiding the chemotherapy by gene expression assays, such as Oncotype DX or MammaPrint (e.g. TAILORx, NCT00310180; RxPONDER, NCT01272037; MINDACT, NCT00433589; and PlanB trials, NCT01049425).4, 5, 6, 7, 8 These trials consistently demonstrated that chemotherapy use could be limited to a relatively small but diverse group of HR-positive/HER2-negative eBC patients. This indicates that several eBC cases included in the meta-analyses were actually not optimal candidates for chemotherapy. The West German Study Group (WSG) PlanB trial is the first trial to test the anthracycline-free, docetaxel-containing regimen against a standard anthracycline followed by docetaxel therapy in a HER2-negative eBC population preselected using the 21-gene expression assay. Eligible patients were candidates for chemotherapy in case of clinical high risk and—together with their treating physician—could opt against chemotherapy if the recurrence score (RS) was <12. Together with the TAILORx and RxPONDER trials, PlanB reported 5-year data for different RS risk groups for patients with up to three involved lymph nodes.4,5,9,10 The successor WSG ADAPT-HR+/HER2− trial (NCT01779206) had the same inclusion criteria and guided chemotherapy in the intermediate-risk RS group through preoperative assessment of endocrine sensitivity.11,12 ADAPT-HR+/HER2− tested the hypothesis that patients at intermediate risk of recurrence with endocrine-sensitive disease have a comparable outcome to that of the low-risk RS group. In the context of these trials, we established a registry for HR-positive patients from the PlanB trial; here, we present the long-term follow-up data. Materials and methods Study design The PlanB registry is an open, prospective, multicenter, observational, noninterventional long-term follow-up registry for patients who participated in the PlanB trial. The phase III PlanB trial tested the noninferiority of an adjuvant anthracycline-free regimen: six cycles of docetaxel (T, 75 mg/m2) plus cyclophosphamide (C, 600 mg/m2) once every 3 weeks (q3w; TC arm), compared with an anthracycline-containing regimen: four cycles of epirubicin (E, 90 mg/m2) plus C (600 mg/m2) q3w followed by four cycles of T (100 mg/m2) q3w (EC-T arm). After including 274 patients, the trial was amended to recommend endocrine therapy (ET) alone (omitting chemotherapy, ET-only arm) in patients with locally HR-positive tumors and RS <12. Other treatments followed national guidelines at the time. Enrollment in the PlanB trial closed in 2011, and the PlanB registry began in 2018 to document the outcomes of patients with HR-positive disease. Trial participants were followed up for 5 years after the initial diagnosis, and patients included in the registry were observed for an additional 5 years, for a total of 10 years of follow-up. The study has been approved by the Ethics Committee of the North Rhine State Chamber of Physicians (Ethikkommission der Ärztekammer Nordrhein)—No.: 2017320. Study participants As previously described,8 the PlanB trial included female patients aged 18-75 years with histologically confirmed, unilateral primary invasive breast cancer, adequate surgical treatment, and no evidence of metastatic disease. Key inclusion criteria were HER2-negativity, pT1-4c, known HR status, pN+ or pN0 with at least one risk factor (pT2 or greater, grade 2-3, high urokinase-type plasminogen activator (uPA)/plasminogen activator inhibitor-1 (PAI-1), age <35 years, or HR-negativity), Eastern Cooperative Oncology Group performance status <2 or Karnofsky index ≥80%. Tumor and nodal status were assigned according to the American Joint Committee on Cancer (AJCC) Cancer Staging Manual, 7th edition, 2010.13 Patients with HR-positive disease were contacted by the recruiting center to obtain a separate signed informed consent for participation in the PlanB registry. Additionally, data from patients who had an event before the end of trial treatment were included in the registry. Endpoints The primary objective of the registry was to compare invasive disease-free survival (iDFS) between patients treated with adjuvant TC and those treated with EC-T. Secondary objectives included the comparison of overall survival (OS), distant disease-free survival (dDFS), local recurrence-free survival (LRFS), and the type of further anticancer treatment. iDFS was defined as the time interval from randomization until any invasive cancer event or death from any cause, whichever occurred first; dDFS was defined as the time interval from randomization until distant invasive relapse or death from any cause, whichever occurred first; OS was defined as the time interval from randomization until death from any cause. Additionally, survival was also evaluated in the ET-only arm. Statistical considerations Associations between two categorical variables were tested by chi-square or Fisher’s exact test. Differences in survival curves were tested using the Kaplan–Meier method and the log-rank test. The impacts of study treatment, and baseline variables on iDFS, dDFS, LRFS, and OS were estimated using Cox regression analysis and expressed as hazard ratios (HRs) and 95% confidence intervals (CIs). To test whether baseline variables interact with the chemotherapy treatment, an interaction between the baseline variable and the therapy arm was included in a Cox model that contained only the chemotherapy arms. Multivariate Cox regression models employed a backstep algorithm, keeping only variables with P < 0.1 in the model. P values <0.05 were considered statistically significant. No adjustments for multiple testing were made due to the exploratory character of the analyses. All statistical data analyses were carried out with the Stata software (version 16.0, StataCorp LLC). Results Patient characteristics Between February 2018 and November 2021, 699 patients were included in the PlanB registry. Of 2116 patients who completed chemotherapy treatment per protocol in the PlanB trial, 298 patients were from the TC arm, 288 from the EC-T arm, and 113 from the ET-only group. Most of the remaining patients could not be reached because they were lost to follow-up after the trial ended, or they chose not to give separate informed consent to participate in the registry. Moreover, although 23 deaths have been documented, it is likely that the actual number of deceased patients is higher, as many individuals were lost during follow-up. The characteristics of included patients were well-balanced between the chemotherapy arms (Table 1). Of patients in the TC arm, 58.7% had a high clinical risk, and 61.1% in the EC-T arm had a high clinical risk (tumor diameter >3 cm and grade 1, or tumor diameter >2 cm and grade 2, or tumor diameter >1 cm and grade 3).14 In the TC and EC-T arms, 67.8%-69.8% had RS 12-25 and 22.1%-22.9% had RS >25, respectively.Table 1Baseline patient characteristicsCharacteristicTC (n = 298)EC-T (n = 288)ET only (n = 113)Overall (N = 699)P valueAge at registration in years, median (IQR)54 (48-62)54 (48-62)55 (49-64)54 (48-63)0.16aAge category0.95b ≤50 years105 (35.2)104 (36.1)39 (34.5)248 (35.5) >50 years193 (64.8)184 (63.9)74 (65.5)451 (64.5)Menopausal status0.81b Premenopausal112 (37.6)106 (36.8)38 (33.6)256 (36.6) Postmenopausal159 (53.4)154 (53.5)63 (55.8)376 (53.8) Unknown/unclear27 (9.1)28 (9.7)12 (10.6)67 (9.6)pT status0.76b 1162 (54.4)164 (56.9)67 (59.3)393 (56.2) 2127 (42.6)113 (39.2)41 (36.3)281 (40.2) 3-49 (3.0)11 (3.8)5 (4.4)25 (3.6)pN status0.002c 0159 (53.4)158 (54.9)77 (68.1)394 (56.4) 1115 (38.6)115 (39.9)36 (31.9)266 (38.1) 2-324 (8.1)15 (5.2)0 (0.0)39 (5.6)Histological grade (local)<0.001b 1-2229 (76.8)208 (72.2)103 (91.2)540 (77.3) 369 (23.2)80 (27.8)10 (8.8)159 (22.7)Clinical riskd0.088b Low114 (38.3)115 (39.9)56 (49.6)285 (40.8) High182 (61.1)169 (58.7)55 (48.7)406 (58.1) NA2 (0.7)4 (1.4)2 (1.8)8 (1.1)ER status (central)0.012c Negative16 (5.4)8 (2.8)0 (0)24 (3.4) Positive263 (88.3)254 (88.2)109 (96.5)626 (89.6) NA19 (6.4)26 (9.0)4 (3.5)49 (7.0)PR status (central)<0.001c Negative54 (18.1)53 (18.4)4 (3.5)111 (15.9) Positive219 (73.5)209 (72.6)105 (92.9)533 (76.3) NA25 (8.4)26 (9.0)4 (3.5)55 (7.9)Recurrence score<0.001b <1221 (7.0)11 (3.8)113 (100.0)145 (20.7) 12-25202 (67.8)201 (69.8)0 (0)403 (57.7) >2566 (22.1)66 (22.9)0 (0)132 (18.9) NA9 (3.0)10 (3.5)0 (0)19 (2.7)Recurrence score, median (IQR)19 (15-21)19 (15-25)9 (6-10)17 (12-23)<0.001aKi-670.47b 0-10128 (43.0)120 (41.7)54 (47.8)302 (43.2) 11-2082 (27.5)73 (25.3)33 (29.2)188 (26.9) >2086 (28.9)91 (31.6)24 (21.2)201 (28.8) NA2 (0.7)4 (1.4)2 (1.8)8 (1.1)Ki-67 baseline0.053a Median (IQR)12 (5.5-21)13 (5.5-22.5)11 (6-18)12 (6-21) NA2 (0.7)4 (1.4)2 (1.8)8 (1.1)HR+ and HER2− and Ki-67 <20% (luminal-A-like)0.06b No91 (30.5)91 (31.6)24 (21.2)206 (29.5) Yes188 (63.1)174 (60.4)85 (75.2)447 (63.9) NA19 (6.4)23 (8.0)8 (7.1)46 (6.6)Values are given as n (%) unless indicated otherwise. P values calculated for nonmissing categories.EC-T, epirubicin + cyclophosphamide followed by docetaxel; ER, estrogen receptor; ET, endocrine therapy; HER2, human epidermal growth factor receptor 2; HR, hormone receptor; IQR, interquartile range; NA, not applicable; PR, progesterone receptor; TC, docetaxel + cyclophosphamide.aF-test.bPearson’s chi-square test.cFisher’s exact test.dLow clinical risk: tumor diameter ≤3 cm and grade 1, or tumor diameter ≤2 cm and grade 2, or tumor diameter ≤1 cm and grade 3; high clinical risk: tumor diameter >3 cm and grade 1, or tumor diameter >2 cm and grade 2, or tumor diameter >1 cm and grade 3. Table 1Baseline patient characteristicsCharacteristicTC (n = 298)EC-T (n = 288)ET only (n = 113)Overall (N = 699)P valueAge at registration in years, median (IQR)54 (48-62)54 (48-62)55 (49-64)54 (48-63)0.16aAge category0.95b ≤50 years105 (35.2)104 (36.1)39 (34.5)248 (35.5) >50 years193 (64.8)184 (63.9)74 (65.5)451 (64.5)Menopausal status0.81b Premenopausal112 (37.6)106 (36.8)38 (33.6)256 (36.6) Postmenopausal159 (53.4)154 (53.5)63 (55.8)376 (53.8) Unknown/unclear27 (9.1)28 (9.7)12 (10.6)67 (9.6)pT status0.76b 1162 (54.4)164 (56.9)67 (59.3)393 (56.2) 2127 (42.6)113 (39.2)41 (36.3)281 (40.2) 3-49 (3.0)11 (3.8)5 (4.4)25 (3.6)pN status0.002c 0159 (53.4)158 (54.9)77 (68.1)394 (56.4) 1115 (38.6)115 (39.9)36 (31.9)266 (38.1) 2-324 (8.1)15 (5.2)0 (0.0)39 (5.6)Histological grade (local)<0.001b 1-2229 (76.8)208 (72.2)103 (91.2)540 (77.3) 369 (23.2)80 (27.8)10 (8.8)159 (22.7)Clinical riskd0.088b Low114 (38.3)115 (39.9)56 (49.6)285 (40.8) High182 (61.1)169 (58.7)55 (48.7)406 (58.1) NA2 (0.7)4 (1.4)2 (1.8)8 (1.1)ER status (central)0.012c Negative16 (5.4)8 (2.8)0 (0)24 (3.4) Positive263 (88.3)254 (88.2)109 (96.5)626 (89.6) NA19 (6.4)26 (9.0)4 (3.5)49 (7.0)PR status (central)<0.001c Negative54 (18.1)53 (18.4)4 (3.5)111 (15.9) Positive219 (73.5)209 (72.6)105 (92.9)533 (76.3) NA25 (8.4)26 (9.0)4 (3.5)55 (7.9)Recurrence score<0.001b <1221 (7.0)11 (3.8)113 (100.0)145 (20.7) 12-25202 (67.8)201 (69.8)0 (0)403 (57.7) >2566 (22.1)66 (22.9)0 (0)132 (18.9) NA9 (3.0)10 (3.5)0 (0)19 (2.7)Recurrence score, median (IQR)19 (15-21)19 (15-25)9 (6-10)17 (12-23)<0.001aKi-670.47b 0-10128 (43.0)120 (41.7)54 (47.8)302 (43.2) 11-2082 (27.5)73 (25.3)33 (29.2)188 (26.9) >2086 (28.9)91 (31.6)24 (21.2)201 (28.8) NA2 (0.7)4 (1.4)2 (1.8)8 (1.1)Ki-67 baseline0.053a Median (IQR)12 (5.5-21)13 (5.5-22.5)11 (6-18)12 (6-21) NA2 (0.7)4 (1.4)2 (1.8)8 (1.1)HR+ and HER2− and Ki-67 <20% (luminal-A-like)0.06b No91 (30.5)91 (31.6)24 (21.2)206 (29.5) Yes188 (63.1)174 (60.4)85 (75.2)447 (63.9) NA19 (6.4)23 (8.0)8 (7.1)46 (6.6)Values are given as n (%) unless indicated otherwise. P values calculated for nonmissing categories.EC-T, epirubicin + cyclophosphamide followed by docetaxel; ER, estrogen receptor; ET, endocrine therapy; HER2, human epidermal growth factor receptor 2; HR, hormone receptor; IQR, interquartile range; NA, not applicable; PR, progesterone receptor; TC, docetaxel + cyclophosphamide.aF-test.bPearson’s chi-square test.cFisher’s exact test.dLow clinical risk: tumor diameter ≤3 cm and grade 1, or tumor diameter ≤2 cm and grade 2, or tumor diameter ≤1 cm and grade 3; high clinical risk: tumor diameter >3 cm and grade 1, or tumor diameter >2 cm and grade 2, or tumor diameter >1 cm and grade 3. Baseline patient characteristics Values are given as n (%) unless indicated otherwise. P values calculated for nonmissing categories. EC-T, epirubicin + cyclophosphamide followed by docetaxel; ER, estrogen receptor; ET, endocrine therapy; HER2, human epidermal growth factor receptor 2; HR, hormone receptor; IQR, interquartile range; NA, not applicable; PR, progesterone receptor; TC, docetaxel + cyclophosphamide. F-test. Pearson’s chi-square test. Fisher’s exact test. Low clinical risk: tumor diameter ≤3 cm and grade 1, or tumor diameter ≤2 cm and grade 2, or tumor diameter ≤1 cm and grade 3; high clinical risk: tumor diameter >3 cm and grade 1, or tumor diameter >2 cm and grade 2, or tumor diameter >1 cm and grade 3. Compared with TC and EC-T arms, patients in the ET-only arm had less frequent lymph node involvement (31.9% versus 46.6% and 45.1%; P = 0.002), fewer grade 3 tumors (8.8% versus 23.2% and 27.8%; P < 0.001), and lower median RS [9, interquartile range (IQR) 6-10, versus 19, IQR 15-21 and 19, IQR 15-25; P < 0.001). Furthermore, 96.5% and 92.9% of tumors in the ET-only group were positive for estrogen receptor or progesterone receptor, compared with the chemotherapy-containing arms (88.2%-88.3%, P < 0.001, and 72.6%-73.5%, P < 0.001, respectively). All patients in the ET-only arm had RS < 12, and 48.7% of them had a high clinical risk. Baseline characteristics of patients included in the PlanB registry and the PlanB trial differed significantly with respect to age, nodal status, grade, clinical risk, estrogen and progesterone receptor status, RS, and Ki-67 levels (Supplementary Table S1, available at https://doi.org/10.1016/j.esmoop.2025.105891). Outcome analysis After 10.3 years of median follow-up (in patients alive at the time of follow-up), no significant differences in iDFS, dDFS, or OS were observed between the chemotherapy arms (Figure 1). Estimated 10-year survival rates in the TC and EC-T arms were 90.8% (95% CI 86.6% to 93.7%) and 92.1% (95% CI 88.2% to 94.7%, P = 0.546) for iDFS (HR 0.85, 95% CI 0.49-1.45, P = 0.55); 94.4% (95% CI 90.9% to 96.5%) and 93.2% (95% CI 89.5% to 95.6%, P = 0.789) for dDFS (HR 1.1, 95% CI 0.6-2.0, P = 0.79); and 96.8% (95% CI 94.0% to 98.3%) and 96.3% (95% CI 93.3% to 98.0%, P = 0.974) for OS (HR 1.0, 95% CI 0.5-2.3, P = 0.97). Local recurrences occurred in seven patients in the TC arm and seven patients in the EC-T arm.Figure 1Analysis of invasive disease-free survival (A), distant disease-free survival (B) and overall survival (C) in the docetaxel + cyclophosphamide (TC) and epirubicin + cyclophosphamide followed by docetaxel (EC-T) arms. Analysis of invasive disease-free survival (A), distant disease-free survival (B) and overall survival (C) in the docetaxel + cyclophosphamide (TC) and epirubicin + cyclophosphamide followed by docetaxel (EC-T) arms. Supplementary Table S2, available at https://doi.org/10.1016/j.esmoop.2025.105891 shows 10-year survival rates in chemotherapy arms according to nodal status. Univariate analysis found that higher pN status, high Ki-67 levels, and larger tumor size were negatively associated with iDFS, dDFS, and OS (Supplementary Table S3, available at https://doi.org/10.1016/j.esmoop.2025.105891). In multivariable analysis, pN2-3 status (versus pN0) was associated with iDFS (HR 7.4, 95% CI 3.5-15.4, P < 0.0001), dDFS (HR 11.1, 95% CI 4.8-25.6, P < 0.0001), and OS (HR 7.9, 95% CI 2.6-23.4, P = 0.0002). Figure 2 presents forest plots illustrating the impact of TC versus EC-T on survival according to several key variables, including age, RS, nodal status, grade, and luminal-A-like subtype (defined as HR-positive and HER2-negative and Ki-67 <20%). Anthracycline-free chemotherapy was comparable to anthracycline-containing chemotherapy in terms of iDFS, dDFS, and OS in all evaluated subgroups. No significant predictors of benefit from anthracycline use were identified for patients with a poorer prognosis according to any of the analyzed factors.Figure 2Forest plots showing the impact of docetaxel + cyclophosphamide (TC) versus epirubicin + cyclophosphamide followed by docetaxel (EC-T) on invasive disease-free survival (DFS) (A), distant disease-free survival (dDFS) (B), and overall survival (OS) (C) within subgroups. No significant subgroup-by-treatment interactions were found; all P-interaction values were >0.05. Hazard ratios (HRs) with 95% confidence intervals (CIs) are shown. RS, recurrence score. aOnly two deaths occurred in the TC arm, and therefore no comparison was possible. Forest plots showing the impact of docetaxel + cyclophosphamide (TC) versus epirubicin + cyclophosphamide followed by docetaxel (EC-T) on invasive disease-free survival (DFS) (A), distant disease-free survival (dDFS) (B), and overall survival (OS) (C) within subgroups. No significant subgroup-by-treatment interactions were found; all P-interaction values were >0.05. Hazard ratios (HRs) with 95% confidence intervals (CIs) are shown. RS, recurrence score. aOnly two deaths occurred in the TC arm, and therefore no comparison was possible. In the RS <12 group, with 113 out of 145 patients (77.9%) receiving ET only and 61 (42.1%) with node-positive eBC, 10-year iDFS, dDFS, and OS rates were 88.6% (95% CI 82.0% to 92.9%), 91.5% (95% CI 85.5% to 95.1%), and 94.2% (95% CI 88.9% to 97.1%), respectively (Figure 3). In the RS 12-25 and RS >25 groups, all patients had received chemotherapy; 175 (43.8%) out of 400, and 53 (40.2%) out of 132 patients had node-positive disease, respectively. Ten-year iDFS, dDFS, and OS rates were 93.4% (95% CI 90.4% to 95.5%), 95.2% (95% CI 92.5% to 97.1%), and 96.8% (95% CI 94.4% to 98.2%), respectively, in the RS 12-25 group, and 87.6% (95% CI 80.5% to 92.2%), 90.7% (95% CI 84.2% to 94.7%), and 96.1% (95% CI 90.9% to 98.4%) in the RS >25 group, respectively.Figure 3Analysis of invasive disease-free survival (A), distant disease-free survival (B) and overall survival (C) according to the RS group in pooled patients from the endocrine therapy (ET)-only, docetaxel + cyclophosphamide (TC), and epirubicin + cyclophosphamide followed by docetaxel (EC-T) arms. Analysis of invasive disease-free survival (A), distant disease-free survival (B) and overall survival (C) according to the RS group in pooled patients from the endocrine therapy (ET)-only, docetaxel + cyclophosphamide (TC), and epirubicin + cyclophosphamide followed by docetaxel (EC-T) arms. In the ET-only arm, median follow-up was 10.1 years; estimated 10-year survival rates were 88.9% (95% CI 81.2% to 93.5%) for iDFS, 92.6% (95% CI 85.7% to 96.2%) for dDFS, and 94.5% (95% CI 88.1% to 97.5%) for OS. Two patients had a local recurrence. Ten-year survival rates in patients in the ET-only arm aged ≤50 years and >50 years, respectively, were 94.8% (95% CI 80.8% to 98.7%) and 85.8% (95% CI 75.1% to 92.2%, P = 0.052) for iDFS, 100% and 88.7% (95% CI 78.6% to 94.2%, P = 0.014) for dDFS, and 100% and 91.6% (95% CI 82.2% to 96.1%, P = 0.054) for OS. Ten-year survival rates in the pN0 and pN1 groups were 93.5% (95% CI 85.0% to 97.2%) and 79.3% (95% CI 61.1% to 90.0%, P = 0.052) for iDFS, 96.1% (95% CI 88.4% to 98.7%) and 85.2% (95% CI 67.9% to 93.6%, P = 0.047) for dDFS, and 97.3% (95% CI 98.6% to 99.3%) and 88.6% (95% CI 72.4% to 95.6%, P = 0.014) for OS. Discussion We present the long-term outcomes from a follow-up registry of the first trial, designed to compare anthracycline-free and standard anthracycline-containing adjuvant therapy in HR-positive eBC patients selected by the 21-gene assay. After a 10-year follow-up, no differences in iDFS, dDFS, and OS were observed between the trial arms in the entire population analyzed, as well as in subgroups stratified by key clinical and biological characteristics. This indicates that anthracycline-containing and anthracycline-free regimens are similarly effective, even after long-term follow-up. Specifically, our results strongly suggest that a significant portion of patients with pN0 disease and high risk by gene expression analysis, and those with pN1 eBC with intermediate to high risk by gene expression testing, could be spared the toxicities associated with anthracycline use. So far, several analyses have examined the impact of anthracyclines on specific patient populations; however, most could not identify any group that would benefit from omitting anthracyclines. Data from the PlanB trial were part of a pooled analysis alongside the SUCCESS C trial data.3 Both trials had an identical six-cycle TC experimental arm and a standard arm involving sequential anthracycline and T (either four cycles of EC followed by four cycles of T or three cycles of fluorouracil, epirubicin, and cyclophosphamide followed by three cycles of T). Approximately 10% of the 5924 analyzed patients had more than three involved lymph nodes. In these clinically high-risk patients, DFS was significantly worse when treated with TC compared with anthracycline-containing regimens (HR 1.48, 95% CI 1.04-2.13, P = 0.032).3 This supports 5-year data from the ABC trials demonstrating the inferiority of TC in terms of DFS but not OS in a population with a higher baseline risk than that in the PlanB trial or pooled PlanB and SUCCESS C trials. Furthermore, pooled analysis of the PlanB and SUCCESS C trials demonstrated the superiority of anthracycline-containing versus anthracycline-free chemotherapy (HR 2.07, 95% CI 1.20-3.57, P = 0.008) in the relatively few cases of lobular eBC (12.3%), a finding that may inform decision making for these patients. A recent meta-analysis of anthracycline-containing and taxane-containing regimens was conducted by the Early Breast Cancer Trialists’ Collaborative Group in more than 100 000 women from 86 randomized trials.2 In that meta-analysis, 18 103 women across 15 trials received taxane therapy with or without anthracyclines. Anthracycline-containing regimens were associated with 14% lower recurrence rates compared with anthracycline-free therapy. Deaths without recurrence were not increased in the anthracycline-treated group; however, one additional AML case was reported per 700 patients. Overall, the regimens involving both anthracycline and taxane conferred the greatest benefit, whereas no significant differences were reported for the comparison of six cycles of TC versus sequential anthracycline and taxane. In the anthracycline and taxane group, the proportional risk reductions in recurrence were similar between estrogen receptor-positive and -negative eBC and did not differ by age, nodal status, grade, or tumor size.2 The limited ability of clinical characteristics alone to identify patients who could safely omit anthracyclines prompted investigations using gene expression analyses to stratify patients according to tumor biology and sensitivity to taxane-based treatment. Exploratory analyses from the RxPONDER and TAILORx trials provided interesting results on the efficacy of different chemotherapy regimens in populations preselected by RS. Results from the RxPONDER trial in patients with N1 disease and RS ≤25 suggested that the use of anthracyclines plus taxane may improve 5-year iDFS in premenopausal patients, whereas the impact of cyclophosphamide plus taxane on the postmenopausal group was not significant.5 Furthermore, the TAILORx trial demonstrated that patients with N0 disease and RS >25 had comparable 5-year iDFS and OS rates following therapy with taxane plus cyclophosphamide (88.1% and 95.8%) and anthracycline with (88.6% and 97.2%) or without taxane (87.4% and 96.7%, respectively).4 The lack of anthracycline impact on iDFS and OS was sustained in the 9-year analysis (taxane plus cyclophosphamide: 78.6% and 89.3%; and anthracycline with: 77.9% and 90.9%; or without taxane: 73.4% and 88.3%, respectively). These findings support the hypothesis that N0 patients with a higher baseline RS benefit from lower toxicity of anthracycline-free regimens; however, these results should be interpreted with caution as follow-up data are limited.4 As previously reported,8,10 survival analysis at 3 and 5 years in the PlanB trial demonstrated a substantial prognostic impact of RS without any interaction with chemotherapy type. Although we observed some imbalances in the demographic data between the PlanB trial and PlanB registry populations, the long-term survival data confirm the initially observed prognostic categories. Three-year dDFS rates reported from the PlanB trial were 97.4% in the RS <12, 97.8% in the RS 12-25, and 91.9% in the RS >25 group.8 The corresponding 10-year dDFS rates from the PlanB registry are now 91.5% in RS <12, 95.2% in RS 12-25, and 90.7% in the RS >25 group, with chemotherapy used in 22.1%, 100%, and 100% of patients, respectively, and node-positive disease in 42.1%, 43.8%, and 40.2% of patients, respectively. Compared with the PlanB registry, 9-year dDFS rates in TAILORx were numerically higher in the RS <11 group receiving ET only (96.8%), comparable in the RS 11-25 group receiving chemoendocrine therapy (95.0%), and lower in the RS >25 group receiving chemoendocrine therapy (86.8%); the corresponding 12-year dDFS rates were 93.2%, 92.8%, and 84.8%, respectively.15,16 However, these indirect comparisons should be interpreted in light of differences between the studies. For example, more than one-third of PlanB patients in each RS group had nodal involvement, while a quarter received chemotherapy in the RS <12 group; in contrast, all patients in TAILORx were node-negative, while none in the RS <11 group received chemotherapy. In the high-risk RS groups, the TAILORx trial furthermore reports a substantial number of patients at long-term risk in the RS >25 group, despite the use of chemoendocrine therapy. The PlanB registry demonstrated 10-year iDFS rates of 87.6% in the RS >25 population. The substantial remaining risk in this RS group, which is not adequately addressed by standard chemotherapy and endocrine therapy, may require alternative approaches, such as CDK4/6 inhibitors, as in the ongoing ADAPTlate trial (NCT04565054). PlanB was one of the first trials to prospectively omit chemotherapy in patients with RS <12 and up to three involved lymph nodes. Although patient allocation to ET-only treatment was not randomized, the PlanB registry demonstrated promising long-term survival data in this low-risk RS group. While the 10-year dDFS rate at 92.6% is slightly lower than in TAILORx, potentially due to the inclusion of node-positive patients, our findings support the evidence of excellent dDFS rates in N0 patients with RS <11 receiving only ET. The same is true for the RS <12 group in the ADAPT-HR+/HER2− trial, where a comparable proportion of patients had N1 disease. Collectively, available data indicate that chemotherapy has no substantial impact on survival in these subgroups with low risk, as determined by gene expression analysis. iDFS and dDFS data from the TAILORx trial show that the risk of locoregional recurrence was even more significant than the risk associated with distant recurrence.15 In the PlanB registry, we report only two local recurrences in the low-risk ET-only arm (RS <12) and seven local recurrences in each chemotherapy arm. Lastly, patients in the intermediate-risk group have excellent outcomes with chemotherapy. Data are less consistent when specifically examining the younger age groups and those at higher risk, as defined by N1 disease. The RxPONDER trial in pN1 patients with RS 0-25 did not demonstrate better survival after the addition of chemotherapy to ET at a 5-year analysis in the overall population.5 However, the benefit from chemotherapy was significant in the premenopausal group. Subgroup analysis in the RS 0-10 groups demonstrated that 5-year iDFS rates in the chemotherapy followed by ET and in the ET-only arms were comparable in postmenopausal women (95.8% and 92.7%, respectively), whereas premenopausal women appeared to derive an iDFS benefit from chemotherapy (96.6% versus 92.4%).5 In the MINDACT trial, women aged ≤50 years who were at high clinical risk and low risk according to the MammaPrint gene expression assay derived a small but significant dDFS benefit from chemotherapy at 8-year analysis.7 The same, albeit weaker, signal is observed in the TAILORx pN0 population, showing a chemotherapy benefit in women aged ≤50 years with RS 16-20 and RS 21-25 (absolute difference in 5-year iDFS: 9% and 6.3%, respectively).15 Interesting results in this context are reported from the (neo)adjuvant ADAPT-HR+/HER2− trial, which combined the risk assessment by the 21-gene assay with testing of the early proliferation response to 3-week preoperative ET (Ki-67 ≤10%).11,12 Five-year survival differences related to age >50 years or pN0/1 were not seen in the preselected RS 12-25/endocrine responder population. Therefore, the ADAPT-HR+/HER2− trial confirmed the hypothesis that chemotherapy can be safely omitted in patients with three or more involved lymph nodes and an intermediate RS, who were preselected based on the endocrine response. Endocrine responsiveness and the endocrine effect of chemotherapy, especially in the perimenopausal population, seem to be important hidden variables affecting outcomes. Therefore, therapy guided by both baseline RS and endocrine response could be a feasible approach to optimally select the patients for chemotherapy-free, de-escalated approaches. Limitations Our study has several limitations. Firstly, since our registry was a follow-up study requiring additional informed consent, we could not ensure enrollment of all patients who had participated in the PlanB trial. In addition to the fact that a number of patients could not be contacted, there is a potential selection bias that may affect the results. Secondly, the PlanB trial, and by extension, the PlanB registry, was not sufficiently powered to detect and quantify small differences between the subgroups. Thirdly, patient allocation to the ET-only arm was not randomized, and, therefore, definitive conclusions on the benefit of treatment cannot be drawn. Fourthly, although we did not detect benefits from anthracycline use in most of the analyzed subgroups, we cannot dismiss the possibility that a particular subgroup derives substantial benefit from anthracycline-containing regimens. Conclusions The long-term follow-up data from the PlanB registry align with previous findings and suggest that anthracycline-containing and anthracycline-free adjuvant TC therapy may offer a comparable efficacy in HR-positive eBC patients with intermediate to high recurrence risk and up to three positive lymph nodes. Moreover, favorable long-term outcomes were observed with chemotherapy-free ET treatment in patients with low-risk disease. Based on the sustained survival benefit, our results contribute to the growing body of evidence suggesting overtreatment in a clinically meaningful group of patients with intermediate RS and strongly support current efforts toward the development of de-escalated approaches.