Elsevier

The Breast

Volume 46, August 2019, Pages 116-125
The Breast

Original article
Nomogram update based on TAILORx clinical trial results - Oncotype DX breast cancer recurrence score can be predicted using clinicopathologic data,☆☆

https://doi.org/10.1016/j.breast.2019.05.006Get rights and content

Highlights

  • A nomogram/calculator can predict Oncotype DX results without performing the test.

  • A nomogram/calculator uses clinicopathologic data from pathology reports.

  • An updated nomogram was built from 84,339 patients' data (National Cancer Data Base).

  • A nomogram correctly assigns 92.7% of cases to the Oncotype DX lowor high-risk category at calculated probability range.85%-100%.

  • An updated nomogram/calculator is available at https://utgsm.shinyapps.io/OncotypeDXCalculator.

Abstract

Objectives

Oncotype DX (ODX), 21-gene breast cancer (BC) assay, predicts risk of recurrence and benefits of addition of chemotherapy to hormonal therapy for early-stage BC. We previously published a nomogram/calculator that could predict ODX results without performing the test by using clinicopathologic characteristics of BC available from pathology reports. Patients with intermediate-risk (11–25) ODXRS (RS) were excluded from that nomogram. This update tests the predictive value of clinicopathologic variables for forecasting the ODXRS while including intermediate-risk-ODXRS patients and stratifying ODXRS based on recently published TAILORx clinical trial results (0–25 = low-risk, 26–100 = high-risk-ODXRS; intermediate-risk-ODXRS belongs to the low-risk category).

Material and methods

The nomogram was built on 65,754 ODX-tested ER+/HER2-/lymph-node-negative patients with 6–50 mm tumor, captured by the National Cancer Data Base (NCDB) from 2010 to 2014. Five clinicopathologic variables (age, tumor size, grade, progesterone-receptor status (PR) and BC-histologic type) were assessed with logistic regression to predict for a low-risk (0–25) or a high-risk (26–100) ODXRS. Results were validated on a separate 18,585 ODX-tested cohort from 2015.

Results

Grade and PR were the highest significant predictors of both low-risk and high-risk-ODXRS, followed by histologic type, tumor size and age. The Receiver Operator Characteristic (ROC) curve showed strong statistical model for both low-risk and high-risk-ODXRS prediction outcomes (c-index = 0.81).

Conclusions

An updated nomogram is now developed/validated on the entire population of ODX-tested patients (84,339) captured by the NCDB. The nomogram/calculator, available on-line at the UTMCK/Shiny website (https://utgsm.shinyapps.io/OncotypeDXCalculator/), will continue serving as a surrogate for BC patients for which ODX testing is not affordable, available or necessary.

Introduction

Oncotype DX (ODX), a commercially available 21-gene recurrence score breast cancer assay (Genomic Health, Redwood City, CA), provides prognostic and predictive breast cancer recurrence information for hormone-positive, HER2-negative, lymph-node negative patients [1,2]. It is endorsed by the National Comprehensive Cancer Network (NCCN) [3] and the American Society of Clinical Oncology (ASCO) [4]. It is used for staging purposes by the revised 8th edition (2018) of American Joint Commission on Cancer staging manual [5].

Initial results from the TAILORx clinical trial published in 2015 [6] prospectively validated a 21-gene expression assay in breast cancer. The study revealed that patients with low Oncotype DX recurrence score (ODXRS) from 0 to 10 can safely forego adjuvant chemotherapy, while adjuvant chemotherapy is recommended for high ODXRS (≥26). No adjuvant chemotherapy recommendations were previously available for intermediate (11–25) ODXRS patients until 6/3/2018, when the TAILORx clinical trial results on intermediate ODXRS were published [7]. According to the new data, patients with ODXRS 11–25 can now safely forego adjuvant chemotherapy, although some benefit of adjuvant chemotherapy was noted in patients younger than 50 years of age with ODXRS of 16–25. These new data allow categorization of ODXRS as a binary variable (0–25 as a low-risk ODXRS and 26–100 as a high-risk ODXRS), in which intermediate-risk score patients (11–25) are now included in the low-risk category.

According to Orucevic's et al. data, ODX is performed in approximately 1/3 of eligible early-stage breast cancer patients in the United States [8]. It is performed for less than 20% of patients in European countries [9]. ODX use in the United States is greatly influenced by race and socioeconomic status. Patients of Caucasian race and high socioeconomic status are tested significantly more often than patients of other races and low socioeconomic status [8,[10], [11], [12], [13]].

We previously developed and published [14] a user-friendly nomogram based on clinicopathologic characteristics of ODX tested patients captured by the National Cancer Data Base (NCDB) as a surrogate prediction model for the ODX assay. Intermediate-risk ODXRS patients (11–25) were excluded from that nomogram. The objective of this update is to test the predictive value of clinicopathologic variables for forecasting the new TAILORx binary ODXRS stratification, in which intermediate-risk ODXRS patients are included in the low-risk category (0–25 = low-risk). Five clinicopathologic variables readily available from pathology reports and established in clinical practice as prognostic and/or predictive were chosen for the development of the nomogram: age, tumor size, tumor grade, progesterone receptor status (PR) and the histologic type of breast cancer. Lymphovascular invasion was excluded in the current nomogram since lymphovascular invasion is now deemed as an optional data reporting element for invasive breast cancer by the College of American Pathologists’ guidelines [15].

Section snippets

Patients and pathology variables selection

Nomograms that predict for a low-risk or a high-risk ODXRS (outcomes of interest) were constructed based on the methods published by Iasonos et al. [16]. Low-risk and a high-risk ODXRS cut-off values were based on the recently published TAILORx clinical trial results [7] (0–25 = low-risk ODXRS and 26–100 = high-risk ODXRS).

Some benefit of adjuvant chemotherapy was identified in the TAILORx trial [7] in patients younger than 50 years of age with ODXRS of 16–25. Therefore, we also evaluated our

Clinicopathologic characteristics of patients in training and validation cohorts used for building of nomograms/calculator

The training cohort consisted of 65,754 ODX tested ER+/HER2-/lymph-node-negative patients with 6–50 mm tumor size, captured by the NCDB from 2010 to 2014. An external validation cohort consisted of 18,585 ODX tested patients captured by NCDB in 2015. Descriptive clinicopathologic characteristics of these patients used for building of nomograms were presented in Table 1.

Nomogram/calculator model development

In multivariate logistic regression analysis, all five variables, tumor size, tumor grade, PR status, histologic tumor type,

Discussion

ODX continues to be the most commonly utilized breast cancer genomic assay in the United States based on the data gathered by the NCDB in the time period from 2010 to 2015. From 434,592 ER+/HER2-negative/lymph node-negative female breast cancer patients, 167,589 (38.5%) had genomic tests performed, with ODX being the most frequently utilized test (95.8%). While performance of breast cancer genomic assays gives personalized prognostic and predictive information for individual breast cancer

Summary

An updated nomogram is now developed and validated based on the entire population of ODX tested patients (84,339) captured by the NCDB from 2010 to 2015. The updated nomogram's c-index of 0.81 is considered a strong statistical model for both low-risk and high-risk ODX recurrence score prediction outcomes. The model overall correctly assigns 86.8% of cases to the low- or high-risk ODX category. Furthermore, the overall accuracy of our prediction model rises to 92.7% at calculated probability

Authors’ declaration of interest

None.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

For our study, formal consent is not required. The NCDB de-identifies data that include the names of patients and institutions. Because of this, criteria of 45 CFR 46.102 d research are met, and Institutional Review Board approval is

Funding

This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

Data availability statement

The data that support the findings of this study are available from the National Cancer Data Base, but restrictions apply to the availability of these data. These data were used under permission for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the National Cancer Data Base.

Authors’ contributions

Conception and design: Amila Orucevic, M.D., Ph.D.

Collection and assembly of data: Amila Orucevic, M.D., Ph.D., Megan King*, BS, Robert E. Heidel, Ph.D., Alison P. McNabb, MS.

*Note: Ms McNeil legally changed her last name now to King.

Data analysis and interpretation: All authors

Manuscript writing: All authors.

Final approval of manuscript: All authors.

Acknowledgement

Authors would like to thank Mr. Ryan King, BS in Mechanical Engineering, MS in Advanced Analytics, Analytics Professional/Sr. Statistician at General Motors, for his expertise and advice during development of the working version of updated calculator using R and Shiny application.

References (39)

  • AJCC cancer Staging Manual

    (2016)
  • J.A. Sparano et al.

    Prospective validation of a 21-gene expression assay in breast cancer

    N Engl J Med

    (2015)
  • J.A. Sparano et al.

    Adjuvant chemotherapy guided by a 21-gene expression assay in breast cancer

    N Engl J Med

    (2018)
  • A. Orucevic et al.

    Utilization and impact of 21-gene recurrence score assay for breast cancer in clinical practice across the United States: Lessons learned from the 2010 to 2012 National Cancer Data Base analysis

    Breast Canc Res Treat

    (2016)
  • M.A. Dinan et al.

    Initial trends in the use of the 21-gene recurrence score assay for patients with breast cancer in the medicare population, 2005-2009

    JAMA Oncology

    (2015)
  • A.A. Guth et al.

    Utilization of oncotype DX in an inner city population: Race or place?

    Int J Breast Cancer

    (2013)
  • M.C. Roberts et al.

    Racial variation in the uptake of oncotype DX testing for early-stage breast cancer

    J Clin Oncol

    (2016)
  • M.J. Lund et al.

    21-Gene recurrence scores: Racial differences in testing, scores, treatment, and outcome

    Cancer

    (2012)
  • A. Orucevic et al.

    Oncotype DX breast cancer recurrence score can be predicted with a novel nomogram using clinicopathologic data

    Breast Canc Res Treat

    (2017)
  • Cited by (0)

    Note: Part of this study was presented at the San Antonio Breast Cancer Symposium in the poster session (P2-08-09) on December 6, 2018.

    ☆☆

    Disclaimers: The National Cancer Data Base (NCDB) is a joint project of the Commission on Cancer (CoC) of the American College of Surgeons and the American Cancer Society. The CoC's NCDB and the hospitals participating in the CoC NCDB are the source of the de-identified data used herein; they have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors.

    View full text