The impact of Oncotype DX testing on adjuvant chemotherapy decision making in 1–3 node positive breast cancer

Abstract Background Oncotype DX testing has reduced the use of adjuvant chemotherapy in node‐negative early breast cancer but less is known about its impact in node positive patients. Aim This study aimed to investigate the impact of Oncotype DX gene assay testing on the decision to offer adjuvant chemotherapy in oestrogen positive, human epidermal growth factor receptor 2 negative, 1–3 lymph node positive patients. Methods Retrospective review of all node positive patients who underwent Oncotype DX testing at a single centre. Clinicopathological data, as well as estimated survival benefit data (from the PREDICT tool), was evaluated by a multidisciplinary group of surgeons and oncologists. Treatment decisions based on clinicopathological data were compared to recurrence scores (RS). A cut off RS > 30 was used to offer adjuvant chemotherapy. Results The 69 patients were identified, of which 9 (13%) had an RS > 30 and assigned a high‐genomic risk of recurrence. The 32 patients (46.4%) were offered adjuvant chemotherapy. Overall based on the use of the RS, the decision to offer adjuvant chemotherapy changed in 36% of patients, and ultimately 24 patients (34.7%) would have been spared chemotherapy. Conclusion Using clinicopathological data alone to make decisions regarding adjuvant chemotherapy in node positive breast cancer leads to overtreatment. Additional information on tumour biology as assessed by the Oncotype DX RS helps to select those patients who will benefit from adjuvant chemotherapy and spare patients from unnecessary chemotherapy.

clinicopathological features of the patient and tumour. More recently, gene expression profiling which captures tumour biology has been increasingly used to aid clinical decision-making.
The Oncotype DX test (Genomic Health, Redwood City, CA, USA) is a reverse transcriptase polymerase chain reaction based assay which measures the expression of a panel of 21 genes (16 cancerrelated, five reference) and generates a recurrence score (RS)-a value from 0 to 100. 4 This score reveals underlying tumour biology and also represents an individualised estimate of the risk of disease recurrence and prognosis.
Previous work with ER+, HER2-, node-negative patients has shown that the RS accurately predicts the benefit of adjuvant chemotherapy helping identify patients who would not benefit from the addition of chemotherapy to adjuvant endocrine therapy. In the landmark TAILORx trial, it was shown that ER+, HER2-, node-negative women with RS 0-15, and women above age 50 with RS 0-25 did not benefit from adjuvant chemotherapy with no difference in survival compared to endocrine therapy alone. 5 Multiple studies have shown the use of Oncotype DX testing in the node-negative setting has changed treatment decisions in up to 30% of patients. [6][7][8] More recently, attention has turned to the utility of the RS in node positive patients, to try to identify patients who would benefit from adjuvant chemotherapy. 9 Retrospective data from the SWOG-8814 trial found that only ER +, HER2-, node positive patients identified as high risk, which was defined as a RS >30, benefitted from the addition of adjuvant chemotherapy compared to tamoxifen alone. 10 There was no benefit from the addition of adjuvant chemotherapy in low risk patients (RS < 18) or for the intermediate (RS 18-30) group, although this trial was underpowered to detect a subtle difference.
The prospective WSG/planB trial showed that node positive patients with a RS < 11 had excellent (94.3%) 5-year disease free survival without chemotherapy. 11 Patients with intermediate-risk (defined as RS 12-25) underwent chemotherapy and showed a similar 5-year disease free survival of 94.3% suggesting that there may be some benefit of chemotherapy in this group. The ongoing RxPONDER trial is evaluating the role of adjuvant chemotherapy in node positive patients with a RS ≤25. 9 Patients will be randomised to hormonal therapy with or without adjuvant chemotherapy. The estimated completion date for the trial is 2022. 12 From these studies, it is clear that gene expression profiling has a role in guiding treatment selection of node positive patients to identify those patients who would not benefit from chemotherapy. 9,13 More evidence to support this is still needed.
Decisions about breast cancer treatment are made by a multidisciplinary team (MDT) including breast surgeons and oncologists. In addition to clinicopathological data, standard UK practice is to use, a prognostication tool such as PREDICT to aid decision making. 14   The simulated MDT decision to offer adjuvant chemotherapy was compared to the RS recommendation for chemotherapy. A RS of greater than 30 was used to select high-risk patients who would benefit from chemotherapy based on previous studies. 10 The two decisions were compared for each patient aiming to examine the impact of the Oncotype DX test in real life.

| Cost analysis
A cost analysis to assess the financial impact of Oncotype testing was performed using previously published sources. The current list price to the NHS for the Oncotype DX test is £2580. 16 The full cost of a course of chemotherapy, including drug costs, delivery costs and toxicity, is estimated to be between £3866.17 and £4863.70 depending on regimen. 16,17 The cost of performing the Oncotype DX test on all patients in the cohort was compared to the potential savings of omitting chemotherapy in patients with a low RS. Further subgroup cost analysis excluding patients with favourable PREDICT scores (<2% predicted survival benefit) was also performed.   Using the current NHS list price, the Oncotype DX test costs £2580, which reflects a total cost of £178 020 to test our cohort.

| RESULTS
From previous studies, the cost of a course of adjuvant chemotherapy in breast cancer is estimated to be between £3866.17 and £4863.70 depending on regimen. 16,17 In our study, 32 patients were recommended chemotherapy by the MDT, but only nine patients had an   21 Further studies will be needed to robustly define high and low-risk groups based on RS.
We found a moderately positive correlation between estimated additional 10-year overall survival benefit based on PREDICT and RS. To the best of our knowledge, this is the first to examine this relationship in node-positive patients. Previous studies in node-negative patients have shown a correlation between RS and low-risk patients, but discordance in high-risk patients. 22,23 This study is one of the very few showing real-life data on the effect of Oncotype test on adjuvant chemotherapy decisions in 1-3 node positive patients. The management of micrometastases or isolated tumour cells in the axilla remains contentious 24,25 and therefore were excluded from our study. In patients with four or more lymph nodes involved, most MDTs would recommend adjuvant chemotherapy without the need to resort to genetic assessment tests such as Oncotype. By excluding these two categories of patients, we avoid confounding factors into our data.
There are some limitations in this study. Firstly, a relatively small number of node positive patients were tested and the study was retrospective in nature. There is also selection bias, as the contempora- to help aid decisions about adjuvant chemotherapy. 28 As more robust prospective data supporting Oncotype testing in node positive patient emerges it will help inform the future revision of these guidelines. 13 Advances in modern oncology with the increasing use of targeted therapies and novel prognostic markers will ultimately result in more personalised approaches to the treatment of breast cancer patients.
Gene-based assays such as Oncotype DX provide additional valuable information on individual tumour biology and prognosis. Given the short and long term toxicity associated with adjuvant chemotherapy, it is imperative to identify patients who have a high risk of recurrence and therefore most likely to benefit, as well as identify low-risk patients in whom chemotherapy can be safely omitted. Our study shows that decision-making based solely on clinicopathological data may lead to overtreatment of node positive patients. However, prospective randomised trial data is still lacking. The results of ongoing trials, such as RxPonder, are eagerly anticipated to fully inform the role of Oncotype DX testing in node positive patients.

ACKNOWLEDGMENTS
This work was supported by the Norfolk & Norwich Hospitals Charity.

CONFLICT OF INTEREST
The authors have stated explicitly that there are no conflicts of interest in connection with this article.

AUTHOR CONTRIBUTIONS
All authors had full access to the data in the study and take responsi- writingreview and editing (equal).

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.

ETHICAL STATEMENT
Our study was registered as a clinical audit and did not require any formal ethical approval.