Do 21-Gene Recurrence Score Influence Chemotherapy Decisions in T1bN0 Breast Cancer Patients?

Purpose: Hormone receptor (HR)-positive breast cancer patients with tumor size ≤1.0 cm and negative node have favorable outcomes. The 21-gene Recurrence Score (RS) could predict response to chemotherapy for HR+ breast cancer, but its role in T1bN0 disease is challenging. Methods: T1bN0 breast cancer patients diagnosed between January 2014 and June 2019 with RS results were included and categorized as Low- (RS < 18), Intermediate- (RS 18–30), or High-risk (RS > 30) groups. Univariate and multivariate analysis were used to assess factors associated with RS distribution and chemotherapy recommendation. Chemotherapy decisions change and patient adherence after 21-gene RS testing were also evaluated. Results: Among 237 patients with T1bN0 tumors, proportions of Low-, Intermediate-, and High-risk RS were 19.8, 63.3, and 16.9%, respectively. Multivariate analysis found that ER expression (P = 0.011), PR expression (P < 0.001), and Ki-67 index (P = 0.001) were independently associated with RS distribution. Adjuvant chemotherapy was recommended for 31.6% of patients, which was more frequently given to patients with higher tumor grade [Odds ratio (OR) = 2.99 for grade II, OR = 59.19 for grade III, P = 0.006], lymph vascular invasion (OR = 8.22, P = 0.032), Luminal-B subtype (OR = 5.68, P < 0.001), and Intermediate-to High-risk RS (OR = 10.01 for Intermediate-risk, OR = 192.42 for High-risk, P < 0.001). Chemotherapy decision change was found in 18.6% of patients, mainly in those with Intermediate- to High-risk RS tumor with the majority from no-chemotherapy to chemotherapy. The treatment compliance rate after the 21-gene RS testing with MDT was 95.4%. Conclusion: RS category was related to ER, PR, and Ki-67 expression, which was recognized as an independent factor of chemotherapy recommendation in T1bN0 breast cancer. The 21-gene RS testing would lead to a chemotherapy decision change rate of 18.6% as well as a high treatment adherence, which can be applied in T1bN0 patients.


INTRODUCTION
Breast cancer, the most common cancer diagnosed in women, is characterized by molecularly heterogeneous. The molecular subtype based on Estrogen receptor (ER), Progesterone receptor (PR), Human epidermal growth factor receptor-2 (HER2), and Ki-67 index could predict prognosis as well as response to treatment. The hormone receptor (HR)-positive breast cancer accounts for 60-75% of all cases (1,2), which always shows a favorable prognosis with adjuvant endocrine therapy and might avoid the cytotoxicity of chemotherapy.
In the past generation, several gene-based assays shed light on the genetic feature of breast cancer, among which the 21-gene RS testing is a reliable and widely used one (3). The 21-gene assay, presenting as a numerical variable ranging from 0 to 100 after calculating by a specific algorithm (4), is carried out by reverse transcription-PCR (RT-PCR) on fixed, paraffin-embedded tumor tissues with a panel composing of 16 cancer-related genes and 5 reference genes. The original study categorized RS < 18 as Lowrisk, RS 18-30 as Intermediate-risk, and RS > 30 as High-risk, in order to predict the distant recurrence rate of HR-positive, lymph-node negative patients treated with tamoxifen (4). Then, the treatment benefit of chemotherapy was certified in patients with High-risk RS in the NSABP B-20 cohort (5). The pivotal study TAILORx demonstrated that patients with RS < 25 derive a little survival benefit from adjuvant chemotherapy, especially for the elderly (6).
With the prevalence of common screening of breast, including clinical breast exams, mammograms, and ultrasounds, the incidence of small breast cancer has increased in the past few decades (7). Those patients always harbor promising clinical outcomes with 5-year disease-free survival (DFS) higher than 90% (8)(9)(10). This may give us a hint that small tumors share different biological features and may deserve treatment deescalation. Under this circumstance, choosing what kind of biomarkers or tools to guide therapeutic decision-making is well worth discussing. Notably, the TAILORx trial included tumors >1.0 cm or tumors between 0.5 and 1.0 cm with intermediate and/or high grade (11).
To our known, there were several studies focus on the survival outcome of patients with small tumors. However, the usage and influence of RS on small tumors were rarely researched, especially in tumors that sized from 0.5 to 1.0 cm (T1b). In the current study, we aim to evaluate the RS distribution, adjuvant chemotherapy decision, and therapeutic recommendation change due to the 21-gene RS testing in patients with T1b, HR-positive/HER2-negative lymph nodenegative breast cancer.

Study Population and Information
Clinical data including clinicopathological patterns, treatment decision and follow up information were extracted from Shanghai Jiao Tong University Breast Cancer Database (SJTU-BCDB). All patients were diagnosed between January 2014 and June 2019, at Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China. Patients were eligible if they met the following criteria: (1) HR-positive and HER2negative, primary invasive breast cancer, (2) underwent the 21gene RS testing, (3) the longest diameter of tumor was larger than 0.5 cm and no more than 1.0 cm. Exclusion criteria included: (1) patients with multifocal or multicenter tumor, (2) tumor was larger than 1.0 cm, (3) patients who had malignant breast lesions other than HR-positive and HER2-negative tumor, (4) male breast cancer, (5) pathologically confirmed lymph node-positive. The study was approved by the Ethical Committees of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine.

The 21-Gene RS Assay Testing
The 21-gene RS assay was performed on formalin-fixed, paraffinembedded tissue. RNA was extracted from two 10 µm unstained sections on hematoxylin and eosin-stained slides and was measured after ensuring the absence of DNA contamination. Gene-specific reverse transcription was performed followed by standardized quantitative RT-PCR reactions in 96 well plates using Applied Biosystems (Foster City, CA) 7500 Real-Time PCR System. The expression level of each cancer-related gene was normalized by 5 reference genes, and the 21-gene recurrence score was then calculated by a specific algorithm. RS was stratified as categorical variables with standard cutoffs (RS < 18 as Lowrisk, RS 18-30 as Intermediate-risk, and RS > 30 as High-risk) (4) and TAILORx cutoffs (RS < 11 as Low-risk, RS 11-25 as Intermediate-risk, and RS > 25 as High-risk) (6). The following text used the standard cutoffs in general unless otherwise noted.

Treatment Decision and Actual Usage of Chemotherapy
The Multidisciplinary Team (MDT), consisting of breast surgeons, oncologists, radiologists, and specialized breast nurses, would make the first-round MDT treatment decision after knowing the clinicopathological parameters of the surgical lesion but without 21-gene RS result (Pre-RS decision). And Post-RS decision would be determined after the results of the 21gene testing were presented, which will also be recorded as the final recommendation in the second-round MDT. The actual administration of the chemotherapy was confirmed by the follow-up information.

Statistical Analysis
Chi-square test (exact Fisher test if necessary) was used to evaluate the RS distribution and chemotherapy usage among patients with different clinicopathological characteristics. Multiple logistic regression models were used to generate adjusted odds ratios (ORs) with 95% confidence intervals (CIs) in order to assess factors associated with RS distribution and chemotherapy. Two-sided p < 0.05 were required for statistical significance. All statistical analyses were carried out by SPSS version 25.0.

Adjuvant Chemotherapy Decision in T1bN0 Patients
Patients' adherence to treatment decisions after the 21-gene RS testing was 95.4% (226 out of 237). Moreover, the treatment compliance rate was higher than 90% (range, 92.5-100.0%) in each subgroup when stratified by RS and tumor grade ( Table 5, Table S5).

Follow-Up and Disease Outcome
After a median follow-up period of 22.69 ± 16.17 months, only one patient with RS of 22 had disease recurrence. The patient was 36 years old, diagnosed with a 0.6 cm tumor in her left breast. The pathology showed it to be infiltrating ductal carcinoma with vascular invasion. Endocrine therapy alone had been used, and after 15 months, contralateral ductal carcinoma in situ was found.

DISCUSSION
In the current study, we included 237 HR+/HER2-, nodenegative breast cancer patients with T1b tumors who underwent the 21-gene RS testing. We found that ER expression, PR expression, and Ki-67 index were independently related to RS distribution. All patients were discussed in two rounds of MDT for their adjuvant treatments. Chemotherapy recommendation was given to 31.6% of T1bN0 patients, which was more frequent in those with higher grade, lymph vascular invasion, Luminal B subtype, and Intermediate-or High-risk RS tumors. Adjuvant chemotherapy decision was changed in 18.6% of patients after physicians knowing the 21-gene RS results, which was mainly from non-chemotherapy to chemotherapy and mostly occurred to patients with Intermediate-to High-risk RS, indicating that physician-intended recommendation would be influenced by the 21-gene RS testing. Moreover, we found the adherence rate to the post-RS decision was higher than 90.0% in these T1bN0 patients, which partly due to the MDT and the 21-gene RS testing. The popularity of breast screening and self-exam greatly contributed to the decrease in tumor size in the past few decades. Data from Surveillance, Epidemiology, and End Results (SEER) program showed that the proportion of small tumors had an increase of nearly 30% (7) from 1975 to 2012. This phenomenon merits concern and further investigation of the clinical feature and treatment pattern of small tumors. Several studies reported that tumor size is not an independent prognostic factor in T1a and T1b tumors (14)(15)(16)(17). In another hand, distinct clinical outcomes were observed in a Korean cohort of small tumors according to molecular subtypes (18), among which HR+/HER2-breast cancer accounted for 56.6% with the best prognosis. For those patients, endocrine therapy could reduce the risk of recurrence, whereas the benefit of adjuvant chemotherapy should be weighed against the treatment-related risk. The predictive value of RS on chemotherapy benefit was universally acknowledged, especially when the result of TAILORx was published. Thus, for all breast tumor ≥0.5 cm, the NCCN guideline strongly recommends the 21-gene assay. But there was no specific data about the clinical significance of 21-gene RS in T1bN0 population. So, we performed the current study in the T1bN0 patients to describe the biological characteristics of small tumors manifested by the RS, as well as to figure out whether the adjuvant chemotherapy decision and actual administration would be influenced by the 21-gene testing in this group of patients.
Regarding the distribution of RS, our results showed that Low-, Intermediate-, High-risk RS accounts for 19.8, 63.3, and 16.9% in the T1bN0 cohort. The distribution observed by our previous study regardless of the tumor size was 26.1, 49.3, and 24.6% in three groups, respectively (19). We noticed that the proportion of Low-or High-risk RS decreased and more tumors were categorized as Intermediate-risk. We postulated that the small tumor size (≤1 cm) in the current study contributes to the difference, since small tumors may be associated with relatively better biological behavior (20,21), as a reason for the lower proportion of High-risk. Meanwhile, patients with T1a tumors were excluded in this study, who may more likely to be genetically Low-risk. Pomponio et al. reported that the proportion of three RS groups account for 65.6, 29.9, and 4.5% in T1b tumors (22), and the NCDB data was 59.0, 33.4, 7.6%, respectively (23). The discordance between our results and theirs' may attribute to the genetic disparities between Chinese and western (24). Another possible reason may be due to the difference in enrollment criteria for the 21-gene RS testing, since we consecutively performed the assay on eligible patients, whereas other institutions may select patients by other clinical parameters.
Many pieces of research focused on clinicopathological factors associated with RS in order to find possible surrogates for the 21-gene testing, among which tumor grade and PR status were the most often discussed (20,25). We noticed that in T1bN0 tumors, ER expression, PR expression, and Ki-67 index were significantly associated with the RS category. The proportion of patients with high-grade tumors was 2.1, 3.3, and 10.0% in Low-, Intermediate-, and High-risk RS groups, whereas we didn't observe the influence of tumor grade on RS in multivariate analysis, which may ascribe to the limitation of tumor size ≤1.0 cm. With respect to the IHC assessment of ER, PR, and Ki-67, we used 50% as a cutoff for ER expression and found that more than 90% of patients had high ER expression. The cutoff for PR status (20%) and Ki-67 index (14%) was according to St. Gallen 2013 expert panel. We suspected that selecting different cutoff points for IHC results may lead to discordance in results. Meanwhile, we had to admitted that the IHC-based classification may be more feasible but imperfect when comparing with the gene-based subtyping of intrinsic subtype (26,27), and additional assays to clarify the biological diversity may also warrant investigation. The chemotherapy usage was significantly associated with tumor grade, lymph vascular invasion, Luminal subtype, and RS in our study. Chemotherapy usage rate was 4.3, 26.0, and 85.0% in Low-, Intermediate-, and High-risk RS groups, which correspond approximately to the literature (23). Several studies and guidelines highlight the superiority of the 21-gene testing over routine clinicopathological parameters (28,29). Meanwhile, the significance of tumor grade on chemotherapy usage, especially in tumors <1 cm, was also noted constantly (30,31). Among all the independent factors in our study, High-risk RS had the highest OR for chemotherapy usage, followed by grade III tumors. This result could reflect the significant importance of these two factors in T1bN0 patients when the treatment decision was made. Furthermore, Ignatov et al. observed that in T1a/b breast cancer, the effect of systemic therapy on survival could only be seen in Luminal-B like tumors but not in Luminal-A like ones (30). Our physicians may take this into consideration when determining the therapeutic recommendation, so Luminal-B like patients were more likely to be asked for chemotherapy.
In the practice of clinic, the alteration in chemotherapy recommendation occurred to ∼30% of cases after physicians knowing the RS (32,33), and the application of the multigene testing was accompanied by the decrease in chemotherapy usage (34,35). We focused on T1bN0 tumors and found that the decision change rate was 18.6%. Most of the changes (42 out of 44, 95.5%) were found in patients with Intermediate-risk (19 out of 150) or High-risk (23 out of 40) tumors, leading to an escalation in treatment pattern. Our results were consistent with the previous article, that adjuvant chemotherapy was given more frequently in T1bN0 patients who underwent the Oncotype in the present study, as high as 95.4%. The multi-gene assay may contribute to this since an adherence rate of 91% had also been reported by using the 70-gene signature (36). On the other hand, the two rounds of MDT in the decisionmaking process render the therapeutic schedule more reliable and acceptable. Clinical outcomes of T1bN0 tumors were constantly excellent with 5-year DFS over 95.0% (8,10,16,37). HR+ breast cancer was associated with ∼20-30% lower risk of death when compared with other molecular subtypes (17) and had the best prognosis in small tumors (18,38). Parise et al. found that 5-year Breast cancer-specific survival of HR+ T1b patients was 99.4% (39), which was concordant with the result published by the NCCN database (40). In our study, due to the small number of patients, only one case of contralateral DCIS was observed who received endocrine therapy alone.
Our study was designed to evaluate the role of 21-gene RS testing in T1bN0 patients which were outside of the TAILORx study, but there were several limitations. The retrospective design was the major weakness since the validation of the results may be negatively influenced by selection or information bias. We used multivariate analysis to narrow the confounding effect. Secondly, we used the IHC method for molecular subtyping instead of the intrinsic subtype identified by genebased assay. It was because the latter one is not feasible for large-scale clinical applications, though maybe more valid to decipher the heterogeneity of breast cancer. Thirdly, the small number of patients together with the short followup period leads to few recurrent events to investigate the survival outcome difference, thus further evaluation is worthy of consideration.

CONCLUSION
Our study included patients with HR+/HER2-, T1bN0 breast tumor and found that RS distribution was associated with ER expression, PR expression, and Ki-67 index, which was independently influenced adjuvant chemotherapy decision. The performance of the 21-gene RS testing would lead to an 18.6% change in therapeutic recommendation but rarely for low-grade tumors. Patient compliance to MDT suggestion was high after the 21-gene RS testing, which warrants further evaluation.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Ethical Committees of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
JY and XC contributed conception and design of the study. JW, OH, JH, ZL, WC, and YL organized the database. JY performed the statistical analysis and wrote the first draft of the manuscript. KS and XC contributed to manuscript revision and funding acquisition. All authors read and approved the submitted version.