Oncotype DX Score Research Articles

Abstract PD14-06: Does chemotherapy benefit patients with HR+/HER2- invasive lobular breast cancer?

Abstract Introduction Invasive lobular breast cancer (ILBC) is the second most common subtype of breast cancer. More evidence is emerging on the unique clinical features of ILBC. However, most systemic therapy guidelines for ILBC are derived from clinical trials with no specific focus on ILBC. Also, the benefit of chemotherapy for the treatment ILBC has been questioned. Using a large national registry, we aimed at comparing outcomes of various systemic therapy regimens in hormone receptor-positive (HR+) and HER2- ILBC. Methods We performed a year-stratified simple random sampling method to select 10% of ER+ or PR+ (HR+) and HER2- female patients with ILBC with AJCC stages I-III who underwent surgery from the National Cancer Database (NCDB) (2010-2017). Patients were grouped into four treatment cohorts: surgery only (S), chemotherapy alone (CT), endocrine therapy alone (ET), combined chemotherapy followed by endocrine therapy (CET). Overall survival (OS) among groups was first examined using the Kaplan-Meier method and Cox proportional hazard modeling. Focusing on ET and CET, propensity score (PS) match in logistic regression was performed to reduce selection biases due to demographical and clinical characteristics. OS between PS matched ET and CET pairs were evaluated again using the same methods. OS differences between ET and CET by lobular histology group and by disease stage were also examined. Results Approximately 90% of ILBC were HR+ and HER2+. N=17,789 patients were included in this analysis: 59.7% received ET, 26.2% CET, 2.3% CT and 11.9% S. 5-years OS was significantly better for patients who received ET (90.7%) or CET (90.4%) as compared to CT (79.4%) or S only (79.7%) (p<.001). Of interest, after adjusting for other demographic and clinical characteristics, 5-years OS was similar for both ET and CET (p=.064). Increased ET administration (52.6% to 67.4%, p for trend <.001) and decreased CET (30.4% to 20.2%, p for trend <0.001) were observed from 2010 to 2017. Among n=15,273 patients who underwent ET (n=10,616; 69.5%) or CET (n=4,657; 30.5%), both groups had similar histology findings (45.5% vs 47.7% lobular histology mixed with other features). However, significant demographic and clinical characteristics differences were identified between both cohorts. Patients in the CET group were younger (mean age 56.9±10.5 vs. 65.5±11.6) (p<.001), had a higher stage at diagnosis (34.8% stage III vs. 4.2%), more positive regional LN (67.5% vs. 13.7%), and received more radiation therapy (74.1% vs. 58.5%) as compared to those in the ET group. 5-years OS remained similar between ET and CET when adjusting for age, stage, LN involvement, histology, and radiation therapy (adjusted HR, 95% CI: 1.04, 0.89-1.20). Importantly, OS was similar between ET and ECT among patients with OncotypeDX (ODX) score < 26 (96.4% vs 97.5%) and ODX score ≥ 26 (5-year OS: 92.2% vs 91.4%). PS match yielded 3,002 pairs; 5-years OS remained similar between ET (92.6%) and CET (91.6%) (p=0.94). In addition, no significant difference was detected when stratifying by disease stage and histology. Conclusion In this large analysis , chemotherapy did not improve survival in patients with HR+ and HER2- lobular cancer treated with ET, regardless of stage or lymph node involvement, or ODX score. The majority of patients with early stage HR+/HER2- with ILBC in the US receive ET, and one-fourth receive CET. Research investigating de-escalated chemotherapy-free regimens for ILBC are warranted. Exploring biomarkers-based approach to identify subgroups who would benefit from chemotherapy would better define systemic treatment options for ILBC. Citation Format: Marita Yaghi, Iktej Jabbal, Nadeem Bilani, Maroun Bou Zerdan, Leah Elson, Hong Li, Diana Saravia, Elizabeth Stone, Zeina Nahleh. Does chemotherapy benefit patients with HR+/HER2- invasive lobular breast cancer? [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD14-06.

Abstract P1-04-05: Multiplexed immunofluorescence staining of intra-tumoral immune cell populations and associations with immunohistochemical, clinical, and pathologic variables in breast cancer

Abstract Background: Although immune checkpoint inhibitors (ICI) are most widely used in PD-L1 positive, metastatic triple-negative breast cancer (TNBC), the clinical utility of ICI in other tumor subtypes and patient populations is not well defined. In this study, we measured infiltration of multiple immune cell populations in primary and metastatic HR+ tumors to infer potential benefit from ICI. We investigated how risk stratifying features such as pathologic grade and Oncotype Dx scores affect immune cell infiltration across subtypes. Lastly, we compared immune cell infiltration in breast tumors from younger versus older women, as our pre-clinical studies demonstrated differential enrichment of immune response pathways by age and association with improved response to ICI. Methods: ImmunoProfile, a multiplexed immunofluorescence (IF) assay, was performed on FFPE tissue from 167 archived breast cancer tumors from the Dana-Farber Cancer Institute (HR+/HER2- n=124, HER2+ n=21, TNBC n=22). Unmatched samples were obtained from either primary breast (HR+/HER2- n=104, HER2+ n=13, TNBC n=15) or metastatic core biopsies. IF staining signal was measured for T cell markers CD8 and co-localized CD8/PD-1, T regulatory cell marker FOXP3, PD-L1 and PD1 expression on tumor cells as well as PD-L1 expression on inflammatory cells. Inflammatory, tumoral (TPS) and combined inflammatory and tumor (CPS) pathologic PD-L1 scores were calculated. PD-L1 staining was also assessed by IHC (E1L3N) staining and compared to multiplexed IF stains. IF staining signal (Vectra analyzed, log2+1 transformed) was analyzed against clinical and pathologic variables, including tumor subtype, primary versus metastatic status, pathologic grade, and Oncotype Dx score, using non-parametric statistical comparisons. Given the decline in adaptive immunity that occurs with normal physiological aging and our own group’s previous reports, we also explored whether immune cell staining signal associated with age. Results: TPS, CPS, and inflammatory PD-L1 scores strongly correlated between IHC and IF staining methods (Spearman r = 0.660, 0.683, 0.638 respectively, p < 2.2e-16). When HR+/HER2- (n=124), HER2+ (n=21) and TNBC (n=22) cases were analyzed separately, there was no statistical difference between unmatched primary and metastatic cases in any of the analyzed immune cell population stains for any subtype. When all primary cases were analyzed, TNBC showed higher CD8, CD8/PD1, PD1 (all p < 0.05), and FOXP3 staining as well as IF CPS, TPS and inflammatory PD-L1 scores (all p < 0.005) than HR+ cases. HR+/HER2- cases from women < age 65 (n=56) showed higher intra-tumoral CD8 (p=0.039), PD1 (p=0.043), and CD8/PD1 (p=0.022) than cases from women > age 65 (n=38). There was a statistically significant step-wise increase in IF staining of all immune cell populations with increasing pathologic grade (I, II, III) as well as Oncotype Dx scores (< 11, 11-25, >25) among primary HR+/HER2- cases (n=74). Conclusions: Strong correlations between PD-L1 scores based on IF and IHC methods suggest clinical utility for multiplexed IF immune cell staining in therapeutic planning. ImmunoProfile demonstrated the previously known pattern that primary TNBC contains higher immune cell infiltration than HR+ or HER2+ tumors. Women < age 65 with HR+ cancer showed higher levels of CD8 and PD1 stained populations, suggesting there may be greater utility of ICI in younger patients. Lastly, the association of higher pathologic grade and Oncotype Dx score with increasing immune cell infiltration supports exploration of ICI in biologically unfavorable primary HR+ cancers. Citation Format: Tess A. O'Meara, Tanya E. Keenan, Adrienne G. Waks, Kristen D. Felt, Bijaya Sharma, Scott Rodig, Melissa Hughes, Nancy U. Lin, Judith Agudo, Jennifer L. Guerriero, Sandra S. McAllister, Elizabeth Mittendorf, Sara Tolaney. Multiplexed immunofluorescence staining of intra-tumoral immune cell populations and associations with immunohistochemical, clinical, and pathologic variables in breast cancer [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P1-04-05.

Abstract P1-13-06: The influence of the oncotype DX breast recurrence score on persistence to endocrine therapy in patients with low-risk ER/PR+ invasive breast cancer

Abstract Intro: The Oncotype DX breast recurrence score provides newly diagnosed patients with low-risk ER/PR+ invasive breast cancer with an estimation of their 10-year risk of distant recurrence if compliant with five years of endocrine therapy. Despite the well-established benefits of endocrine therapy, to include a decrease in local and distant disease recurrence, many patients will discontinue endocrine therapy prior to five years, defined as non-persistence. We hypothesized that the individual perception of risk of recurrence conveyed by the Oncotype DX breast recurrence score directly influences persistence with endocrine therapy and patients with higher Oncotype DX breast recurrence scores may have higher rates of persistence when compared to a similar cohort with lower recurrence scores. Further, we hypothesized that disease recurrence in women with low-risk (defined as less than 26) ER/PR+ breast cancer is due primarily to non-persistence with endocrine therapy. Methods: Our study included 198 patients with low-risk ER/PR+ breast cancer treated at a comprehensive cancer center between April 2011 and April 2014 with sufficient-follow-up data of at least 5 years. Persistence was determined by chart review and defined as a minimum duration of 55 months on endocrine therapy. Multivariable ANOVA and Fisher Chi Square Exact test were used to compare the recurrence score, 10-year risk of distant recurrence, and rates of recurrence between persistent vs non-persistent patients. Results: The average recurrence score of our cohort was 14.74 with an average 10-year risk of distant recurrence of 9.8 percent with five years of endocrine therapy. The average recurrence score was 15.24 in the persistent group vs. 13.18 for the non-persistent group (p=0.029). The average 10-year risk of distance recurrence was 10.1% in persistent patients vs 9% in non-persistent patients (p=0.019). The recurrence rate (both local and distant) was 4% in persistent patients vs 16% in non-persistent patients at 5- year follow (P=<0.01). Discussion: Non-persistence to endocrine therapy represents a major unmet need in breast cancer care. The Oncotype DX breast recurrence score significantly influences persistence to endocrine therapy as our retrospective review demonstrated that patients with low-risk ER/PR+ locally invasive breast cancer with higher Oncotype DX scores are more likely to complete the recommended 5-years of endocrine therapy. Further, despite having a lower average Oncotype DX score overall, the non-persistent group had a higher rate of recurrence when compared to the persistent group. This finding further substantiates the overall effectiveness of endocrine therapy in reducing the risk of future recurrence in patients with low-risk ER/PR+ breast cancer. Mindful framing of Oncotype DX test results by providers represents a potential avenue to increase persistence through the modulation of individual risk perception and may ultimately lead to decreased rates of disease recurrence in patients with low-risk ER/PR+ invasive breast cancer. Citation Format: Jonathan Pirruccello, Elaine Kuhn, Mary Chamberlin. The influence of the oncotype DX breast recurrence score on persistence to endocrine therapy in patients with low-risk ER/PR+ invasive breast cancer [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P1-13-06.

Abstract P3-18-04: Evaluating de-escalation of breast radiation (DEBRA) following lumpectomy for breast conservative treatment of stage 1, hr+, HER2-, RS ≤18 breast cancer: NRG-BR007 a phase III trial

Abstract Roughly 50% of newly diagnosed breast cancer is stage 1, the majority being ER/PR positive, HER2- negative. Genomic assays such as Oncotype DX® have identified patients with reduced distant metastasis and lack of chemotherapy benefit, allowing patients to avoid excess toxicity. These genomic assays have been shown to be prognostic for local-regional recurrence (LRR). The de-escalation of therapy is of interest to patients, providers, and payers. Low risk, as identified by both the use of Oncotype and Mammaprint® is associated with low LRR after lumpectomy and breast radiotherapy. TRIAL DESIGN: In the DEBRA trial, we hypothesized that breast-conserving surgery (BCS) alone is non-inferior to BCS plus radiotherapy for in-breast cancer control and breast preservation in women intending appropriate endocrine therapy for stage 1 (ER and/or PR-positive, HER2-negative, and Oncotype DX Recurrence Score [RS] low) breast cancer. Stratification is by age (<60; ≥60), tumor size (≤1 cm; >1-2 cm), and RS <11, RS 11-18. Patients are randomized to either breast radiotherapy (RT) plus endocrine therapy (arm 1) or to observation and endocrine therapy (arm 2). Arm 1 therapy is post-lumpectomy breast RT using standard methods (hypo- or conventional-fractionated whole breast irradiation with or without boost, accelerated partial breast irradiation) and at least 5 years of endocrine therapy (tamoxifen or aromatase inhibitor). In arm 2, at least 5 years of endocrine therapy (tamoxifen or aromatase inhibitor) will be given. The specific regimen of endocrine therapy in both arms is at the treating physician’s discretion. ELIGIBLITY: Patients who are stage 1: pT1 (2 cm), pN0, age ≥50 to <70 years, status post (s/p) lumpectomy with negative margins (no ink on tumor ), s/p axillary nodal staging (SNB or AND), ER and/or PR positive by ASCO/CAP, HER2-negative by ASCO/CAP, and Oncotype DX RS of ≤18 on diagnostic core biopsy or resected specimen. ENDPOINTS: Primary: In-breast recurrence (IBR). Secondary: Breast conservation rate, invasive in-breast recurrence (IIBR), relapse free interval (RFI), distant disease-free survival (DDFS), overall survival (OS), patient-reported breast pain, patient-reported worry about recurrence, and adherence to endocrine therapy. STATISTICS: We assume a clinically acceptable difference in IBR of 4% at 10 years to judge omission of RT as non-inferior (10-year event-free survival for RT group is 95.6% versus 91.6% for the omission of RT group). To be able to detect non-inferiority with 80% power and a one sided α=0.025, and assuming that there would be a ramp-up in accrual in the first two years of the study (leveling off in Years 3-5), 1,670 (835 per arm) patients are required to be randomized. This conservatively assumes loss to follow-up will be 1% per year. Some of the T1a patients accrued to this study will have oncotype DX scores >18, making them ineligible for the study. An extra step in the accrual process will require us to register 1,714 patients to ensure our final randomized cohort is 1,670 patients. Accrual: Screen 1,714 to randomize 1,670 into the study. Contact information: Protocol: CTSU member website: https://www.ctsu.org. NRG Oncology Pgh Clinical Coordinating Dpt: 1-800-477-7227 or ccd@nsabp.org. Support: U10CA180868, U10CA180822. NCT04852887. Citation Format: Julia White, Stewart J Anderson, Eleanor ER Harris, Eleftherios P Mamounas, Daniel G Stover, Patricia A Ganz, Reshma Jagsi, Reena S Cecchini, Carmen Bergom, Valerie Theberge, Mahmoud El-Tamer, Rich C Zellars, Dean A Shumway, Guang-Pei Chen, Thomas B Julian, Norman Wolmark. Evaluating de-escalation of breast radiation (DEBRA) following lumpectomy for breast conservative treatment of stage 1, hr+, HER2-, RS ≤18 breast cancer: NRG-BR007 a phase III trial [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P3-18-04.

Abstract P1-08-28: Real world use of OncotypeDx testing in the management of breast cancer. The North East England experience

Abstract OncotypeDx testing is now routinely used to help decide management of hormone positive HER2 negative breast cancer. We have reviewed adjuvant treatment and long-term outcomes of breast cancer patients in those who had a genomic OncotypeDx study performed by the oncology team based at the Northern Centre for Cancer Care (NCCC) in Newcastle Upon Tyne over a 10 year period between 2011-2021. Guidance on the use of chemotherapy according to age and recurrence score changed from July 2018, which has been factored in to our review. We also reviewed the other clinicopathological features that could affect outcomes, including tumour size, grade, nodal status , Nottingham Prognostic index and the presence or absence of in situ disease and lymphovascular invasion. The notes of 908 patients who underwent a total of 974 oncotype Dx scores were reviewed to assess:1.Were the management decisions in line with the guidance on recurrence scores in use at that time?2.What were the outcomes for these patients (we have assessed disease free and overall survival for those who have completed 3 years or more of follow up)?3.For patients who relapsed, would a different decision have been made if current guidelines are applied?. 508 tumours had a low recurrence score (18 or less) and would typically be advised that chemotherapy would not be of benefit. 198 tumours scored in the intermediate group and a decision about chemotherapy currently would be made on the basis of their age. 268 tumours had a high recurrence score of 25 or greater and current advice would advocate chemotherapy in this group. The cut off figures for oncotype score groups have changed with previous ‘intermediate’ recurrence score figures from 18-30, and the ‘high’ score category above 30. 232 tumours of the 974 were node positive (80 with micrometastases, 148 with 1-3 nodes, and 4 had more than 4 involved nodes). In the 255 records evaluated thus far for patient outcome, only 22 relapses have occurred, with 9 local recurrences and 13 distant recurrences. Only two patients in the low recurrence score group evaluated so far (n=147) received chemotherapy and there were 6 patients with recurrences in this group; 4 were local and 3 distant. In the high recurrence score group (n=99 )85 patients received chemotherapy and 14 patients did not, of whom 5 were due to patient choice. Seven patients have so far relapsed (2 local & 5 distant recurrences). In the intermediate group 15 patients received chemotherapy of the total group evaluated to date(n=73). Six patients in the intermediate group have relapsed (2 local & 4 distant recurrences). Under modern guidelines, only two further patients in this group would have been offered chemotherapy. 25 of the patients in the intermediate group were 55 years or above and 4 of these received chemotherapy - under current UK guidance these patients would avoid chemotherapy. Whilst the aim is to share results from the complete dataset in our poster, early results from this study confirm the benefit of OncotypeDx testing in the real world management of patients with breast cancer. AcknowledgementsMark Verrill, NicCresti, Daniella Lee, Radha Todd, Helen Turnbull, Kathryn Wright, NajibahMahtab, Ed Hallam, Sarah Reynia Citation Format: Abigail Gault, Jayshree Veeratterapillay, Wendy Taylor. Real world use of OncotypeDx testing in the management of breast cancer. The North East England experience [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P1-08-28.

Abstract P1-02-12: Genomic testing on breast cancer core biopsy material is feasible and provides useful treatment information

Abstract Background: Genomic profiling assays for invasive breast cancer provide useful predictive and prognostic information and are performed most commonly on surgical resection specimens. Obtaining the molecular profile at the time of initial core needle biopsy is ideal because it could provide information that could significantly alter preoperative decision-making and avoid unnecessary treatments. In January 2020, the breast center began a 3 month pilot program that would reflexively send core biopsy material on all newly diagnosed patients for genomic testing using the MammaPrint (MPT) and BluePrint (BPT) assays regardless of receptor status. The goals were to determine feasibility and examine the impact to patient care. Methods: Breast core biopsy tissue was triaged by an attending pathologist as soon as slides were available. If invasive carcinoma with at least 3 mm in linear extent of tumor with 20% cellularity was identified, materials were immediately sent to Agendia for MPT/BPT testing even before ER/PR/HER2 testing was available to the pathologist. Clinicopathologic information, turn-around time (TAT) and adequacy data were tracked. The impact was discussed regularly at breast tumor board to determine if the results led to altered decision-making that could reduce unnecessary interventions and time to initial treatment. Results: 445 core biopsy specimens were sent for genomic testing. Of those 442, (97%) yielded genomic results with an average TAT from biopsy to genomic result of 11 calendar days. MPT identified 233 (53%) cases that were low risk and 209 (47%) cases that were high risk. BPT showed that 60 (14%) cases were Basal, 18 (4%) cases were HER2, and 364 (82%) cases were Luminal. Further analysis of the Luminal subgroup demonstrated that 154 (42%) were Luminal A low risk, 78 (21%) were Luminal A Ultralow risk and 132 (36%) were Luminal B High Risk. Analysis by race demonstrated a significantly higher percentage of high risk tumors in African-American women including a higher percentage of basal cancers (26%) as compared to Caucasian women (10%). Of note, 51 patients in the pilot study had additional OncotypeDX (ODX) testing on subsequent surgical resection. There was fair correlation between the assays with the majority of the low risk MPT having low risk ODX scores (<=25) and the high risk MPT having high risk ODX scores (>25), though MPT identified more patients as being high risk. The COVID-19 pandemic altered plans to assess time to treat and treatment interventions as initially intended. However, the knowledge of genomic result enhanced the ability to triage patients, allowing those patients with low risk tumors to begin endocrine therapy and delay surgery. Use of preoperative hormone therapy was considered more often in place of neoadjuvant chemotherapy in patients with low genomic risk ER+/HER2- patients. Because of the continued feedback indicating positive impact, the pilot ultimately was extended to 11 months to allow time for formal implementation. Conclusions: Genomic testing using the MPT/BPT assay on core biopsy samples with at least 3 mm yields results 97% of the time with an average TAT of 11 days from biopsy date to result. The genomic information at the time of initial diagnosis impacted patient care most notably in the ER+ setting. The results led to the immediate implementation of direct reflex testing of all newly diagnosed ER+ and HER2- or IHC 2+ cancers by pathology after the initial pilot phase. Citation Format: Hannah Gilmore, Philip Bomeisl, Alberto Montero. Genomic testing on breast cancer core biopsy material is feasible and provides useful treatment information [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P1-02-12.

Utility of Oncotype DX score in clinical management for T1 estrogen receptor positive, HER2 negative, and lymph node negative breast cancer.

The management of estrogen receptor positive (ER+)/HER2- and lymph node (LN) negative breast cancers can be influenced by Oncotype DX recurrence score (RS) in the USA. However, the benefit of RS in T1 tumors (≤ 1cm) is not clear. We retrieved 199 T1 ER+/HER2-/LN- breast cancer diagnosed between 1993 and 2016 that had undergone RS testing. The median follow-up time was 51months. We examined the disease-free survival (DFS) and distant metastasis and their association with RS and other clinicopathologic features. Of the 199 cases, 40 were T1a (≤ 0.5cm) and 159 were T1b (> 0.5cm to 1cm) tumors. In the 40 T1a tumors, 11 would benefit from chemotherapy by the TAILORx study results. Of these T1a tumors, 36 were Nottingham grade 1/2, 3 were grade 3, and 1 was microinvasive carcinoma; 2 (5%) had local recurrence and 1 (2.5%) had distant metastasis to the bone. The only patient with T1a tumor (Nottingham grade 3, RS = 42) and distant metastasis to bone had received adjuvant chemotherapy. In the 159 T1b tumors, 25 would benefit chemotherapy by the TAILORx results. Of the T1b tumors, 149 were Nottingham grade 1/2 and 10 were grade 3. Nine (5.7%) had local recurrence and 2 (1.3%) had distant metastasis to bone and mediastinum, respectively. The two T1b tumors with distant metastasis had a RS 20 and Nottingham grade 2, and RS 27 and Nottingham grade 3, respectively. Both patients received adjuvant chemotherapy. In multivariate analysis of the entire cohort (T1a and T1b tumors), Nottingham tumor grade and receiving chemotherapy were significantly associated with DFS. In univariate analysis of the entire cohort, Nottingham tumor grade, receiving adjuvant chemotherapy, and RS were significantly associated with distant metastasis. This study demonstrates that the metastatic rate of T1a and T1b ER+/HER2-/LN- breast cancer is very low. Patients with low grade (1 or 2), T1a ER+/HER2-/LN- breast cancer may not need RS for treatment decision-making; however, in patients with high-grade T1a or T1b ER+/HER2-/LN- breast cancer, RS analysis should be strongly considered.

Factors Associated With the Decision to Decline Chemotherapy in Patients With Early-stage, ER+/HER2- Breast Cancer and High-risk Scoring on Genomic Assays

IntroductionThe rate of refusal of chemotherapy ranges from 3% to 19%, but varies widely by patient profile and treatment setting. Using a large national registry, we explore factors significantly associated with the decision to decline chemotherapy in patients with early-stage, HR+/HER2- breast cancer (BC) despite high risk scoring on multigene sequencing analysis for OncotypeDX (ODX) or MammaPrint (MP), in which the survival benefit of chemotherapy is clear. Patients and MethodsPatients with HR+/HER2- BC and high risk scoring on ODX (score >26) or MP were selected from the National Cancer Database (2004-2017). Only those who refused to get chemotherapy despite their physician's recommendations were included. Univariate frequency and proportion statistics were used to describe the patient cohort. Bivariate Chi-square analysis evaluated the association between refusal of recommended chemotherapy and sociodemographic characteristics. Significant variables (P < .05) were included in a multivariable logistic regression model. ResultsN = 43,533 patients were included (88.7% ODX, 11.3% MP). A total of n = 4415 (10.1%) patients declined chemotherapy despite recommendation by the patient's primary oncologist. Age >70 (OR: 3.46, 95% CI: 2.96-4.04, P < .001), black race (OR: 1.20, 95% CI: 1.07-1.36, P = .01), non-private insurance, lobular carcinoma histology (OR: 1.21, 95% CI: 1.09-1.35, P < .001), and tumor grade of I significantly predicted chemotherapy decline. ConclusionIdentifying and addressing many of the factors that contribute to under-treatment in minorities is to be key to reducing cancer disparity and improving equity in cancer care and outcome.

Cost-effectiveness analysis of Oncotype DX from a Brazilian private medicine perspective: a GBECAM multicenter retrospective study.

Oncotype DX (ODX) is a validated assay for the prediction of risk of recurrence and benefit of chemotherapy (CT) in both node negative (N0) and 1-3 positive nodes (N1), hormone receptor positive (HR+), human epidermal growth factor receptor 2-negative (HER2-) early breast cancer (eBC). Due to limited access to genomic assays in Brazil, treatment decisions remain largely driven by traditional clinicopathologic risk factors. ODX has been reported to be cost-effective in different health system, but limited data are available considering the reality of middle-income countries such as Brazil. We aim to evaluate the cost-effectiveness of ODX across strata of clinical risk groups using data from a dataset of patients from Brazilian institutions. Clinicopathologic and ODX information were analyzed for patients with T1-T3, N0-N1, HR+/HER2- eBC who had an ODX performed between 2005 and 2020. Projections of CT indication by clinicopathologic criteria were based on binary clinical risk categorization based on the Adjuvant! Algorithm. The ODX score was correlated with the indication of CT according to TAILORx and RxPONDER data. Two decision-tree models were developed. In the first model, low and high clinical risk patients were included while in the second, only high clinical risk patients were included. The cost for ODX and CT was based on the Brazilian private medicine perspective. In all, 645 patients were analyzed; 411 patients (63.7%) had low clinical risk and 234 patients (36.3%) had high clinical risk disease. The ODX indicated low (<11), intermediate (11-25), and high (>25) risk in 119 (18.4%), 415 (64.3%), and 111 (17.2%) patients, respectively. Among 645 patients analyzed in the first model, ODX was effective (5.6% reduction in CT indication) though with an incremental cost of United States Dollar (US$) 2288.87 per patient. Among 234 patients analyzed in the second model (high clinical risk only), ODX led to a 57.7% reduction in CT indication and reduced costs by US$ 4350.66 per patient. Our study suggests that ODX is cost-saving for patients with high clinical risk HR+/HER2- eBC and cost-attractive for the overall population in the Brazilian private medicine perspective. Its incorporation into routine practice should be strongly considered by healthcare providers.

362 Comparison of PREDICT Score to Oncotype Dx Score in Early Invasive Breast Cancer

Abstract Aim The PREDICT score is used to help define decisions about benefits of chemotherapy in early invasive breast cancer. With a PREDICT score of 5% or more chemotherapy should be considered, 3% to 4% is a borderline decision and less than 3% chemotherapy is not pursued. We set out to compare the PREDICT score to Oncotype Dx recurrence score (RS), which predicts chemotherapy benefit based on 21-gene breast cancer assay. Method Data for patient and tumour characteristics, and treatment choice was collected from electronic patient records of 98 patients who had an RS score done between February 2017-September 2020. This data was then compared to the patients’ PREDICT scores. Results 98 patients were analysed with 2 excluded due to having nodal macro-metastases. Mean patient age was 53 years with an average tumour size of 27mm. 31/96 (32%) patients had a PREDICT score ≥5%, with 21/31 (68%) having a low RS score (&amp;lt;26). 57/96 (59%) patients had a PREDICT score of 3-4%, with 40/57 (70%) having a low RS score. 8/96 (8%) patients had a PREDICT score of &amp;lt; 3%, with 7/8 (88%) having a low RS score. Additionally, 7/8 (88%) patients in this group were pre-menopausal. Conclusions 68% of patients in the ≥5% PREDICT group had a low RS score and so have avoided chemotherapy. In the 3-4% group, approximately 1 in 3 tested were shown to have a chemotherapy benefit. In the &amp;lt;3% group most patients were pre-menopausal, and chemotherapy was safely avoided. A bigger dataset and comparison to other health boards is required.

Oncologists’ perspectives on omission of sentinel lymph node biopsy in women &gt;70 years with early-stage hormone receptor-positive breast cancer.

48 Background: Trial data show that omission of surgical axillary staging does not affect overall survival in women &gt;70 with cT1N0 hormone receptor-positive (HR+) breast cancer, and the Society of Surgical Oncology’s Choosing Wisely recommendations advise against routine use of sentinel lymph node biopsy (SLNB) in patients with early-stage HR+ cancers. Despite this, almost 80% of women eligible for omission still undergo SLNB. We sought to explore oncologists’ perspectives of omission of SLNB in this patient population. Methods: We conducted an exploratory qualitative study using semi-structured telephone interviews with surgical, medical, and radiation breast oncologists throughout North America from 3/2020 to 1/2021. Purposive snowball sampling ensured a range of practice types. Interviews were transcribed and a team trained in qualitative analysis undertook thematic analysis guided by grounded theory to identify emergent themes. Results: Participants included sixteen surgical, six medical, and seven radiation oncologists (55% female) (Table). Overall, while oncologists in all fields expressed acceptance regarding SLNB omission in certain women &gt;70 with cT1N0 HR+ disease, many viewed it as a complex choice based on patient comorbidities, chronologic age, patient preferences, and disease factors. Although patients’ physiologic age and life expectancy were also important decisional factors, almost all participants assessed these subjectively despite knowing that validated tools existed. Most surgeons perceived the data backing the Choosing Wisely recommendation as weak, although knowledge of specific supporting studies was low. While all participants agreed that SLNB omission does not affect survival, several radiation oncologists expressed anxiety about resultant increased regional recurrence risk. In the absence of known nodal status, medical and radiation oncologists stated they were more likely to order additional imaging, rely on OncotypeDX scores to make systemic therapy decisions, add high tangents, and be reluctant to offer partial breast irradiation. Conclusions: While surgeons are aware of the Choosing Wisely recommendation, high SLNB rates in patients eligible for omission may be driven by perceptions of the quality of the supporting data and differing ideas regarding appropriate candidacy for omission. There are downstream effects of SLNB omission on medical and radiation oncology treatment decision making and surgeons should engage in multidisciplinary discussion prior to surgery.[Table: see text]

Mammography-based radiomics for predicting the risk of breast cancer recurrence: a multicenter study.

This study aimed to establish a mammography-based radiomics model for predicting the risk of estrogen receptor (ER)-positive, lymph node (LN)-negative invasive breast cancer recurrence based on Oncotype DX and validated it by using multicenter data. A total of 304 potentially eligible patients with pre-operative mammography images and available Oncotype DX score were retrospectively enrolled from two hospitals. The patients were grouped as training set (168 patients), internal test set (72 patients), and external test set (64 patients). Radiomics features were extracted from the mammography images of each patient. Spearman correlation analysis, analysis of variance, and least absolute shrinkage and selection operator regression were performed to reduce the redundant features in the training set, and the least absolute shrinkage and selection operator algorithm was used to construct the radiomics signature based on selected features. Multivariate logistic regression was utilized to construct classification models that included radiomics signature and clinical risk factors to predict low vs intermediate and high recurrence risk of ER-positive, LN-negative invasive breast cancer in the training set. The models were evaluated with the receiver operating characteristic curve in the training set. The internal and external test sets were used to confirm the discriminatory power of the models. The clinical usefulness was evaluated by using decision curve analysis. The radiomics signature consisting of three radiomics features achieved favorable prediction performance. The multivariate logistic regression model including radiomics signature and clinical risk factors (tumor grade and HER 2) showed good performance with areas under the curve of 0.92 (95% confidence interval [CI] 0.86 to 0.97), 0.88 (95% CI 0.75 to 1.00), and 0.84 (95% CI 0.69 to 0.99) in the training, internal and external test sets, respectively. The DCA indicated that when the threshold probability is ranges from 0.1 to 1.0, the radiomics model adds more net benefit than the "treat all" or "treat none" scheme in internal and external test sets. As a non-invasive pre-operative prediction tool, the mammography-based radiomics model incorporating radiomics and clinical factors show favorable predictive performance for predicting the risk of ER-positive, LN-negative invasive breast cancer recurrence based on Oncotype DX. The mammography-based radiomics model incorporating radiomics and clinical factors shows favorable predictive performance for predicting the risk of ER-positive, LN-negative invasive breast cancer recurrence.

149P Can high Ki67 predict distant recurrence in early stage breast cancer with low oncotype Dx score?

Oncotype Dx recurrence score provides prognostic and predictive information for recurrence in early stage breast cancer. Patients with low oncotype Dx scores (=<25) are less likely to recur and often adjuvant chemotherapy is not necessary. However, a small minority of patients with low Oncotype Dx scores will experience recurrence. In this retrospective study, we examined whether high Ki67 could predict the distant recurrence in early stage breast cancer with low oncotype Dx scores.

Deep learning from HE slides predicts the clinical benefit from adjuvant chemotherapy in hormone receptor-positive breast cancer patients

We hypothesized that a deep-learning algorithm using HE images might be capable of predicting the benefits of adjuvant chemotherapy in cancer patients. HE slides were retrospectively collected from 1343 de-identified breast cancer patients at the Samsung Medical Center and used to develop the Lunit SCOPE algorithm. Lunit SCOPE was trained to predict the recurrence using the 21-gene assay (Oncotype DX) and histological parameters. The risk prediction model predicted the Oncotype DX score > 25 and the recurrence survival of the prognosis validation cohort and TCGA cohorts. The most important predictive variable was the mitotic cells in the cancer epithelium. Of the 363 patients who did not receive adjuvant therapy, 104 predicted high risk had a significantly lower survival rate. The top-300 genes highly correlated with the predicted risk were enriched for cell cycle, nuclear division, and cell division. From the Oncotype DX genes, the predicted risk was positively correlated with proliferation-associated genes and negatively correlated with prognostic genes from the estrogen category. An integrative analysis using Lunit SCOPE predicted the risk of cancer recurrence and the early-stage hormone receptor-positive breast cancer patients who would benefit from adjuvant chemotherapy.

Abstract 896: Comparison of tumor profiles of second cancers among breast cancer survivors to first cancers in the SEER registries: A nationwide study

Abstract Background: In the US, the number of breast cancer (BC) survivors is expected to increase to 5 million by 2030. Although it is well established that BC survivors are at increased risk of second primary cancer (SPC), little is known about the characteristics of these tumors. Methods: We conducted a retrospective cohort study using the Surveillance, Epidemiology, and End Results (SEER) 18 database to compare the tumor characteristics of first primary cancer (FPC) to SPC among women diagnosed with BC between 2000-2016. For this analysis, we focused on breast, lung, colorectal, uterine, and cervical cancer. Odds ratio (OR) and 95% confidence interval (CI) for tumor characteristics comparing SPC to FPC were estimated from logistic or multinomial logistic regression, adjusting for race/ethnicity, year of diagnosis, and age at diagnosis. Tumor characteristics included summary tumor stage, T stage, tumor grade, hormone receptor (HR) status of BC, molecular subtype of BC, Oncotype DX score of BC for a subset (N=48,826), and histologic subtype of lung cancer. Results: The study population included 14,392 second and 586,788 first BC, 5,202 second and 246,543 first lung cancers, 3,525 second and 190,640 first colorectal cancers, 2,517 second and 125,818 first uterine cancers, as well as 230 second and 42,601 first cervical cancers. Women with SPC were older than women with FPC. Compared to FPC, SPC were more likely to be diagnosed at local stage for breast (OR: 1.30, 95%CI: 1.25-1.34), lung (OR: 1.57, 95%CI: 1.47-1.68), colorectal (OR: 1.20, 95%CI: 1.12-1.28), and cervical cancer (OR: 1.59, 95%CI: 1.19-2.12), but not for uterine cancer (OR: 1.07, 95%CI: 0.98-1.17). Second breast (OR: 1.26, 95%CI: 1.21-1.31) and uterine (OR: 1.83, 95%CI: 1.65-2.03) cancers were more likely to present at poorly or undifferentiated tumor grade after adjusting for tumor stage. No difference in tumor grade was found for lung, colorectal, and cervical cancer. Second BC were more likely to be of hormone receptor (HR) negative (OR: 1.64, 95%CI: 1.57-1.71), human epidermal growth factor receptor 2 (HER2) enriched (OR:1.30, 95%CI: 1.16-1.45) and triple negative (OR: 1.66, 95%: 1.55-1.77) subtype, and having Oncotype DX score above 26 (OR: 1.73, 95%CI: 1.47-2.04). Second lung cancers were more likely to be of small cell histology compared to first lung cancers (OR =1.15, 95%CI: 1.06-1.25). Conclusion: Screen-detectable second cancers were diagnosed at an earlier stage compared with first cancers. Further, second breast, uterine, and lung cancer had a more aggressive tumor profile compared to women with a similar type of first cancer suggestive of underlying biological differences between the two groups. It is plausible that more intensive screening strategies akin to screening recommendations for women at high risk of BC may be needed for survivors at risk for certain second cancers. Citation Format: Zhengyi Deng, Miranda R. Jones, Kala Visvanathan. Comparison of tumor profiles of second cancers among breast cancer survivors to first cancers in the SEER registries: A nationwide study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 896.

Abstract 2695: Identification of an age-related breast cancer gene expression signature with strong prognostic value

Abstract Background and Objective: Breast cancer (BC) diagnosed at ages &amp;lt;40 years is known to present more aggressive tumor phenotypes and poorer clinical outcome compared to older BC patients. To gain a better understanding of the possible age-related expression difference in BC promoting genes, we aimed to identify transcriptional alterations supporting cancer progression in BC of the young. Methods: We studied mRNA gene expression data from METABRIC discovery (n=997), and validation cohort (n=995), and explored the transcriptional patterns associated with BC of the young. We investigated differentially expressed genes (DEGs) between primary BC for patients below and above 40 years at time of diagnosis and explored gene sets (by Gene Set Enrichment Analysis) enriched in BC of the young. Protein-protein interaction (PPI) networks were explored by the STRING database and visualized in Cytoscape software. Hub genes were identified by MCODE and CytoHubba plugins. Results: Among genes differentially expressed between BC of the young and older, 91 DEGs were commonly (for the two cohorts) upregulated and 99 commonly downregulated in the young (Fold change ≥1.5/≤ -1.5). Oncogenic- and proliferation associated signatures were enriched in the young BC, including signatures reflecting KRAS, MTOR and MYC. Proliferation and cell cycle processes were the dominating enriched gene ontology categories (FDR &amp;lt;10%). Six hub genes presenting highest PPI network connectivity were identified, described to be associated with the cell cycle and cell proliferation. A signature score was made by summarizing the hub genes expression values. High signature score was significantly associated with high tumor size, high histologic grade, lymph node metastasis, ER negativity, basal-like, and HER2 enriched subtypes (P &amp;lt;0.001). The signature score showed strong correlation with the proliferation signatures OncotypeDx and PCNA (ρ=0.90-0.96, P &amp;lt;0.001). Our signature score associated with reduced cancer specific survival (P &amp;lt;0.001), also when adjusted for tumor size, histologic grade, and lymph node status in multivariate analyses in both cohorts (HR: 1.074-1.081, 95% CI: 1.046-1.109, P ≤0.003). The signature score maintained independent, significant association with survival when adjusting for the traditional clinico-pathologic variables in OncotypeDx-low tumors in discovery (HR: 1.077, 95% CI: 1.034-1.122) and validation cohort (HR: 1.154, 95% CI: 1.065-1.251), both P &amp;lt;0.001. Conclusion: By genome-wide gene expression analyses, the current study provides new insights into age-related gene expression alterations in breast cancer. We demonstrate enrichment of oncogenic signaling and evidence of higher tumor cell proliferation in young BC patients, and identify a gene expression signature reflecting tumor proliferation, aggressive tumor features and reduced survival, also in subsets of low OncotypeDx score. Citation Format: Lise Martine Ingebriktsen, Lars Andreas Akslen, Elisabeth Wik. Identification of an age-related breast cancer gene expression signature with strong prognostic value [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2695.

OncotypeDx scores according to race/ethnicity and age in a diverse cohort of breast cancer patients.

e12512 Background: Breast cancer incidence and mortality rate differ across racial/ethnic populations in the US, but little is known about the relationship between Oncotype DX scores and race/ethnicity. Oncotype DX scores are used in early stage, hormone positive breast cancers to estimate the likelihood of breast cancer recurrence and benefit from receiving chemotherapy. Most available literature assesses disparities between race and receipt of an Oncotype DX score, with only some assessing distribution of scores by race. Populations previously studied mainly focus on Black compared to White women, with very few including Asian and Native Hawaiian/Pacific Islander subpopulations. When included, they are often grouped together as Asian/Pacific Islander. Current studies suggest there is a similar distribution of scores among Black and White women, but some studies report that Black women may be more likely to have high-risk scores. Methods: We examined 476 unique breast cancer cases in the Hawaii Pacific Health system diagnosed in 2018-2020. We used univariable and multivariable analyses on all of those cases that received an Oncotype DX score to determine correlation to age and race. Results: In 328 breast cancer cases that received an Oncotype DX score, age of diagnosis ranged from 29 to 84 with race/ethnicity including Japanese (n = 90, 27%), White (n = 78, 24%), Filipino (n = 54, 17%), Native Hawaiian (n = 52, 16%), Chinese (n = 24, 7%), and other (n = 30, 9%) populations. Cases with age at diagnosis between 60-69 (n = 118, 36%) were found to have a mean Oncotype DX score of 13.91 (CI 12.19-15.62, p = 0.04) which was statistically lower than other ages groups on both univariable and multivariable analysis. We found no other significant relationships between Oncotype DX score and race or age at diagnosis on our analyses. Conclusions: These findings contribute more information about Oncotype DX scores within Asian and Native Hawaiian populations to the available literature.

Impact of oncotype Dx score on treatment and long-term outcomes.

e12518 Background: Oncotype dx is a 21 gene breast cancer assay that helps predict benefit of chemotherapy in early-stage hormone receptor positive (HR+), HER2 negative, 0 to 3 node positive breast cancer as reported by TAILORx and RxPonder trials. Both trials found no benefit of adjuvant therapy in mid-range scores of 11-25 with the exception of some benefit in age under 50 or premenopausal status, no benefit if score was 0-10, and benefit of chemotherapy if &gt;25. Methods: All breast cancer patients between 2019-2020 were chart checked to evaluate oncotype dx scores. Patient characteristics were analyzed between oncotype dx scores of 0-10, 11-25, and &gt;25 using X square tests. 2000-2020 institutional cancer registry data was analyzed to understand recurrence and mortality in patients with oncotype dx testing. Results: 173 patients had oncotype dx testing between 2019-2020, and most were HR+, HER2-, and had 0-3 lymph nodes. There was no difference in race or ethnicity between the groups of oncotype dx scores (Table), and average age at diagnosis for all three groups was 58 years. More patients with lower oncotype dx score had hormone therapy, but this was not significant (0-10 at 93.6%, 11-25 at 90.8%, and &gt;25 at 82.1%, p=0.26). Patients with a higher oncotype score were significantly more likely to have chemotherapy (0-10 at 0%, 11-25 at 17.3%, and &gt;25 at 82.1%, p&lt;0.01). When looking at long term outcomes, the mortality rate was highest in the 11-25 score group at 9.5%, but was not significantly higher than 0-10 at 4.7% and &gt;25 at 2.1% (p=0.15). The recurrence was highest in the 0-10 group at 5.4%, but not significant compared to the 11-25 group at 4.1% and &gt;25 group at 4.2% (p=0.09). Conclusions: Race, ethnicity, and age does not impact predisposition to high oncotype dx score, which is interesting since literature shows minority women have higher mortality rates. This could be because oncotype dx looks at early-stage breast cancers and many minority women have later stages at diagnosis. There was no difference in which group received hormone therapy and there was significantly higher rate of chemotherapy administration in patients with scores &gt;25, which is consistent with prospective trial recommendations. When evaluating long term outcomes between oncotype dx groups over 20 years, there was no significant difference in recurrence of disease or mortality indicating that oncotype dx scoring accurately predicted overall benefit of chemotherapy.[Table: see text]

Comparing practice patterns and outcomes for early-stage hormone receptor-positive and node-positive breast cancer patients with high-intermediate-risk Oncotype Dx scores in the National Cancer Database.

e12520 Background: Oncotype Dx (ODX) is a commercial diagnostic test primarily used to predict the likely benefit from chemotherapy in ER+, HER2-, and node negative breast cancer. The prognostic value (recurrence risk) has also been demonstrated to apply to early stage lymph node positive (LN+) disease in a number of retrospective and prospective studies. The ongoing RxPONDER trial aims to clarify the predictive value of RS in LN+ population. In light of the initial results, we analyzed the practice patterns and outcomes for HR+/Her2 -/node positive breast cancer patients receiving ODX testing in the years from 2010-2017 with RS 14-25 in a retrospective observational study of the NCDB. Methods: Women with HR+/Her2 -/node positive breast cancer receiving ODX testing from 2010-2017 were identified in the NCDB using TAILORx and RxPONDER patients’ inclusion criteria: ages 18-75, 6-50mm invasive tumors, N1, M0, ER+/HER2 -. The impact of ODX results in the high-intermediate range (14-25) and other clinico-pathologic variables on the receipt of chemotherapy were compared. Additionally, we examined the impact of chemotherapy on overall survival (OS). Frequencies, Kaplain-Meier analysis, and changepoint analysis using the Contal and O’Quigley method were utilized. Results: There were 109,652 T1-2 and N1 patients of whom 32,506 (29.6%) received ODX testing. 13,461 (41.4%%) women had scores in the high-intermediate (14-25) range. The majority tended to have only 1 LN involved (1LN: 77.2%, 2LNs: 17.5%, 3LNs: 5.3%), had a mean age of 57.8y, were Caucasian (86.4%), and were preferentially tested at academic or comprehensive community cancer programs (79.2%). 6,610 (49.3%) patients were recommended chemotherapy, the median ODX score for all women who were recommended chemotherapy was 20 compared to 17 for those whom chemotherapy was not recommended. 5,068 (76.7%) women had documentation of receiving chemotherapy which correlated with improved OS regardless of age. Conclusions: In the group of women with HR+/Her2 -/node positive breast cancer, clinicians appear to utilize ODX testing in less than one-third of patients, possibly finding RS to be most useful in guiding adjuvant therapy recommendations when only 1LN is involved. Both the recommendation and receipt of chemotherapy correlated linearly with increasing RS, as expected based on the current NCCN guideline recommendations. We identified an OS benefit when chemotherapy was administered, regardless of patient age. Long-term follow-up in the RxPONDER trial will likely continue to clarify the predictive value of RS &lt; 25 in the ER+/HER2-/node positive breast cancer population.

Oncotype Dx score associated to pathological features to better prognosticate the risk of recurrence for breast cancer patients.

e12562 Background: Oncotype Dx (ODX) prognosticates the risk of recurrence and predicts the benefit of adjuvant chemotherapy in estrogen-receptor-positive breast cancer (BC). However, it puts aside important clinical features. Our objective is to develop a model that uses routine clinical parameters associated with Oncotype Dx score to estimate the recurrence rate in our population. Methods: We retrospectively reviewed ODX and pathology reports from 371 early BC patients diagnosticated between 2014 and 2016 in a specialized cancer center, with a 5 year follow up. Variables were selected first with a univariate cox regression, then the statistically significant variables underwent a multivariate cox regression. The β coefficients from selected variables were used to create the prediction formula. Results: Among the tested variables, ki67, estrogen receptor (ER) and ODx score were significantly associated with Recurrence. The linear predictor is: (0.1331 x Ki67) + (0.4232 x ER)+(0.2112xODx score). The threshold score for this equation was set on 33 (median value) to discriminate low and high estimated risks for recurrence. The correct classification rate (CCR) for the training and validation sets was 76% and 68%, respectively. Conclusions: It is possible to optimize the Oncotype Dx to better determine a recurrence score with selected pathological features.