The impact of age and nodal status on variations in oncotype DX testing and adjuvant treatment

Oncotype DX (ODX) recurrence score (RS) is a validated tool to guide the use of adjuvant chemotherapy (AC) in hormone receptor+/HER2- breast cancer. In this analysis, we examine (1) characteristics associated with ODX testing and (2) the association between ODX RS and receipt of AC across age and nodal status. Women with HR+/HER2–, early-stage (T1-2, N0-1) breast cancers from 2010–2017 in the National Cancer Database were included. 530,125 met inclusion and 255,971 received ODX testing. Older women were less likely to receive testing; however, nodal positivity increased use of testing. High ODX RS was associated with increased mortality, though the association was not consistent across age and was most strongly associated with mortality among younger, node-negative women. Older women with high ODX RS, regardless of nodal status, were less likely to receive AC. Clinicians may be employing ODX RS to support treatment decisions against the receipt of AC.


INTRODUCTION
Unprecedented advances in science and technology, such as multigene expression analysis of tumors, have rapidly changed treatment paradigms for many cancers. The oncotype DX (ODX) recurrence score (RS) (Exact Sciences., Madison, WI) is a validated prognostic and predictive tool to estimate patient risk of distant breast-cancer recurrence and guide the use of adjuvant chemotherapy in hormone receptor-positive (HR+) human epidermal growth factor receptor-2 negative (HER2-), earlystage breast cancer 1,2 . While the first results of RxPONDER prospective trial (NCT01272037) assessing ODX RS use in nodepositive (N1) disease were reported in December 2020, a small prospective study 3 and multiple retrospective studies have reinforced its utility in this setting [4][5][6][7][8][9][10] . Many clinicians have already been using the test in this cohort to guide decisions on adjuvant therapy 11 . ODX RS enables personalized treatment approaches and has reduced the use of chemotherapy, which is associated with both negative short-and long-term sequela that can severely affect quality of life 12,13 .
Cancer treatment for the aging population is challenging, requiring clinicians to critically weigh the delicate balance of competing morbidities and quality of life. Prospective randomized control trials have shown that axillary staging and radiation therapy do not improve overall or disease-specific survival in select older women with early-stage breast cancer [14][15][16][17][18] . Tumor biology of older breast-cancer patients is distinct, and deimplementation of practices such as adjuvant radiation and sentinel node biopsy in low-risk patients has been slow [19][20][21][22] . The National Comprehensive Cancer Network and American Society of Clinical Oncology both recommend the use of oncotype DX RS testing in HR+/HER2-, lymph node-negative or micro-metastatic tumors, however, this recommendation is currently not stratified by age or menopausal status 23,24 .
The intent of this study was to (1) investigate the prevalence of oncotype DX RS testing in women diagnosed with HR+/HER2breast cancer and its association with mortality across age groups and nodal status, and (2) examine whether testing was associated with adjuvant chemotherapy use across both age and nodal status. We hypothesized that the predictive value of ODX RS may vary with age and nodal status and that testing and its impact on adjuvant therapy decisions may not be consistent across age groups.

RESULTS
Overall, 530,125 women were diagnosed with early-stage, HR+/HER2-breast cancer were included and 255,971 (48%) of these women received an ODX RS test.
Oncotype DX recurrence score (ODX RS) testing Demographic and clinical characteristics, stratified by testing status, are reported in Table 1. Women who received an oncotype DX RS test were on average younger, had fewer comorbidities, and lower pathologic T and N stages. Of those women who received an ODX RS test, the majority had private insurance (60% versus 40%). Most women (57%) in the untested cohort had Medicare/Medicaid. In the cohort that received an ODX RS test, node-positive women represented a smaller percentage of the total cohort compared to in the untested cohort (16% versus 20%).
Oncotype DX recurrence score and adjuvant chemotherapy Of the women who had oncotype DX RS testing (n = 255,971), 248,576 women had complete information including numerical ODX RS. Median oncotype score was similar across age groups and nodal positivity status, with a median score ranging from 15-18 ( Fig. 2A). Distribution of high, intermediate and low ODX recurrence scores across age subgroups and nodal status are depicted in Fig. 2B. High ODX RS was most prevalent in women <40 and low ODX RS was most prevalent in women ≥70, regardless of nodal status.
Receipt of chemotherapy, stratified by age cohort and oncotype DX RS, is shown in Table 2. After adjustment for Charlson score, insurance status, tumor grade, histologic subtype and tumor stage, ODX RS had the smallest impact on adjuvant chemotherapy use among younger women (<40 years old), with high ODX RS, compared to low ODX RS, associated with a fourfold increase in receipt of adjuvant chemotherapy. Among women 40-69 years old and ≥70 years old, high ODX RS, compared to low, was associated with substantially higher prevalence of adjuvant Sub-analysis was performed on receipt of adjuvant chemotherapy in women who received a high (≥26) ODX RS (Supplemental Table II). Of all women who received a high ODX RS, younger women were more likely to receive adjuvant chemotherapy than their older counterparts, (88.8% and 77.2% versus 51.3% in the <40, 40-69, and ≥70-year-old cohorts, respectively). Nodal positivity resulted in increased use of adjuvant chemotherapy across all age subgroups. However, among women over 70, adjuvant chemotherapy was used in 55% of node-positive and 51% of node-negative women with high ODX RS.

DISCUSSION
Our study highlights the current landscape of oncotype DX RS testing; that is, older women are less likely to receive an ODX RS test than their younger counterparts, but when RS testing is used in in women ≥70 years of age, it is more often performed in those with nodal positivity. High ODX RS was associated with higher allcause mortality across nodal status. However, this relationship was not consistent across age and high ODX RS was most strongly associated with mortality among younger, node-negative women. Older women with high ODX RS, regardless of nodal status, were less likely to receive adjuvant chemotherapy.
Our data is consistent with previous studies [25][26][27][28][29] demonstrating decreased use of ODX RS with increasing age. Similar to previous reports 30 , we found that median ODX RS was similar across age and nodal status. In a study of 610,350 tumor specimens from node-negative, micro-metastatic and nodepositive patients, Bello et al. demonstrated the independent nature of tumor biology and anatomic stage by showing similar distributions of ODX scores across nodal status. A second study showed a similar distribution of ODX RS across premenopausal and post-menopausal cohorts 24 . Despite similar median scores across age groups, older, node-positive women have a higher prevalence of low oncotype scores, very low rates of high oncotype scores, and high ODX scores are more frequent in the very young (<40) 8,31 . Our data similarly demonstrated a higher prevalence of low oncotype DX RS in women ≥70 years of age despite similar median scores to younger patients.
As found in prior studies 32,33 , younger age and nodal positivity were associated with increased use of adjuvant chemotherapy. Past literature has found 70-88% of women in receipt of a high ODX RS receive adjuvant chemotherapy [32][33][34][35] . Similarly, our analysis found receipt of AC following a high (≥26) ODX recurrence score to be 88.8% and 77.2% in women ≤40 years and 40-69 years, respectively. When stratified by <50 or ≥50 years Fig. 1 Trends in oncotype DX recurrence score (ODX RS) testing over time. A overall and stratified by B age group C nodal status, and D age and nodal status, among women with HR+/HER2-, early-stage breast cancer.
of age, Williams et al. found no difference in chemotherapy recommendation based on ODX RS 36 . However, in our cohort, among women over 70 years of age, only about half with high ODX RS (node negative or node positive) received adjuvant chemotherapy. This suggests that treatment decisions begin to differ not only with menopausal status (using age 50 as a surrogate), but instead among those with perceived competing risks for mortality. The recent presentation of the RxPONDER data confirms the safety of chemotherapy omission in postmenopausal women with 1-3 positive nodes and ODX RS <25 37 . While various reasons could explain the differential outcomes in pre-and post-menopausal women (including the ovarian suppressive effect of cytotoxic chemotherapy), these data may also reflect different tumor biology across age groups. Although our results describe the current landscape of the use and implications of ODX RS testing, further study examining the impact of these outcomes following the release of RxPONDER will be of interest.
The overall health benefit of chemotherapy must be considered in terms of decreased risk of relapse and improved survival 36,38 . Important drivers of this decision include estimation of life expectancy, chemotherapy-derived toxicity and the possibility of functional decline 39 . The decisions to recommend use ODX testing and use adjuvant chemotherapy, when indicated, are multifaceted and require insight into patient comorbidity and quality of living. Functional age, rather than chronological age, should be considered. Notably, a recent questionnaire-based prospective study studying the impact of chemotherapy on global health status in patients over 70 years of age found the negative impact of chemotherapy on quality life was statistically significant at 6 months, however, chemotherapy was no longer impactful on quality of life by 18 months 13 . Multiple validated tools have been proposed to aid physicians in objectively measuring the medical and psychosocial domains of the geriatric patient 40 . As observed in our data, oncotype DX RS testing is currently being used in the elderly breast-cancer population, however, a high ODX RS did not consistently result in receipt of chemotherapy in this cohort. Clinicians should consider comprehensive geriatric assessments before making treatment decisions so that costly procedures and tests, that may also have potential risks, can be avoided.
In our data, high ODX RS was most strongly associated with mortality among younger, node-negative women and this association weakened as age increased. The relationship between high oncotype DX RS and mortality persists across all ages and nodal status. We suspect these findings may be a result of both varied biology across age groups and a reflection of the bias of who clinicians elect to order the test for. Despite these differences, receipt of chemotherapy was associated with reduced mortality for women with high ODX RS in all age groups. Retrospective Fig. 2 Median Oncotype DX recurrence scores stratified by age and nodal status. The A median oncotype DX recurrence score (ODX RS) and B categorized ODX RS, stratified by age group and nodal status among women with HR+/HER2-, early-stage breast cancer. Patients who had missing staging information were excluded from analysis.

K. Iles et al.
analyses have also demonstrated that the receipt of chemotherapy results in improved overall survival not only in older women, but also in women with multiple comorbidities 41,42 . However, in a Surveillance, Epidemiology and Endports Registry (SEER) study, Zhou and colleagues 43 showed that while ODX RS was a strong predictor of breast-cancer-specific mortality, the addition of adjuvant chemotherapy did not decrease the risk of breastcancer-specific mortality in a competing risk model. In this context, our data again highlights the importance of assessing competing risks and patient preferences when choosing to use and apply results of the ODX RS test. Our analysis is not without limitations. The NCDB lacks cancerspecific survival data. While this has implications for the older patients in this investigation, due to their competing comorbid conditions, we believe the data is still informative in helping us understand the value of oncotype in older and node-positive women. We acknowledge that our findings may not fully describe the spectrum of ODX RS in relation to nodal status and all-cause mortality given the implications of selection bias. To better contextualize our results, we conducted analyses to understand the characteristics of who received testing compared to who did not. In spite of these limitations, we believe our analysis likely reflects real world practice of how clinicians are currently using this tool.
We have shown that oncotype DX testing is used less frequently in node-negative women over 70 years of age compared to younger women, and more commonly in nodepositive women ≥70 years. We demonstrate that high ODX score is most predictive of survival in the youngest patients. Similarly, and not surprisingly, adjuvant chemotherapy in women with high ODX RS of all ages and nodal statuses, is associated with reduced mortality, however, it is given only about half the time in women ≥70 with a high ODX RS. As we have elected to omit invasive procedures and treatments in this patient population, we recommend thoughtful use of ODX RS, taking into consideration life expectancy and patient values.

Study design and patient population
This retrospective cohort study used data from the National Cancer Database (NCDB), a joint database project of the Commission on Cancer (CoC) and the American College of Surgeons capturing incident cancer cases. Established in 1989, this nationwide, facility-based registry captures 70% of newly diagnosed cancer cases each year. This registry captures valuable patient and cancer characteristics in addition to therapeutics avenues in a de-identified fashion. Institutional Review Board approval was not required as the NCDB is a de-identified, publicly available database.
The two primary outcomes of interest were adjuvant chemotherapy initiation within 9 months of definitive diagnosis and 5-year and allcause mortality. Women were followed until loss-to-follow-up, death, or December 2017.

Statistical analysis
Descriptive statistics were used to characterize pertinent demographics and clinical characteristics, including race, Charlson-Deyo comorbidity score, insurance type, residence type, residential income, clinical stage, tumor grade, histologic subtype, in addition to pathologic T and N stages. Multivariable log-binomial regression was used to assess differences in the prevalence of ODX RS use across patient and cancer characteristics. An additional model with interaction terms was used to assess whether the association between testing and nodal status was consistent across age. Multivariable log-binomial regression was also used to assess whether ODX RS score was predictive of adjuvant chemotherapy use. Women who underwent ODX RS testing but were missing an associated numerical score were excluded from these analyses (n = 7395). Separate models with interaction terms between ODX RS score and age group were also run among node-negative and node-positive women. Models were adjusted for relevant covariates based on outcome of interest and are indicated accordingly within each analysis.
Kaplan-Meier curves were then used to examine survival in women with high ODX RS across age and nodal status by receipt of chemotherapy. Fig. 4 Interaction between age group and nodal positivty and its association with mortality. Interaction between age group and nodal positivity on 5-year, all-cause mortality, among women with HR+/HER2-, early-stage breast cancer who had an oncotype DX recurrence score (ODX RS) test. Models were adjusted for Charlson-Deyo score, histologic subtype, pathologic T stage, tumor grade, and insurance status.