Abstract PO1-14-05: A comparison of chemotherapy recommendations by NPI, Predict, and Oncotype DX testing in UK women with early node positive breast cancer

Abstract

Abstract Introduction: Gene expression testing is expensive and, unless it changes the chemotherapy (CT) decision, unnecessary. Both Nottingham Prognostic Index (NPI) and Predict Breast Cancer (Predict) are free to use and based on clinico-pathological data. Each gives a guide to the prognosis. Many clinical teams in the UK have used these indices to guide treatment and more recently to help decide who should have access to gene expression testing. NPI scores of ≤2.4 have a cancer-specific ten-year survival of 96%, and >2.4 to ≤3.4, >3.4 to ≤5.4 and >5.4 have survivals of 93%, 78% and 44% respectively. UK oncologists use it as a guide in recommending adjuvant treatment. CT is not recommended for patients with NPI ≤3.4, is discussed as on option for NPI >3.4 to ≤5.4 and is recommended for NPI >5.4. Similarly, Predict can be used to guide CT decisions as follows: < 3% benefit, CT not recommended; 3-5% CT discussed as a possible option; >5% CT recommended. The OncotypeDX® Breast Recurrence Score® test has been assessed recently in the RxPONDER trial which concluded that postmenopausal women with a Recurrence Score (RS) result < 26 showed no benefit from the addition of CT. This study compares the clinical utility of three indices (NPI, Predict, and RS result) by assessing the distribution of NPI and Predict across RS result, evaluates changes in CT recommendations based on risk group, and by estimating the likelihood of CT recommendation. Methods: We recently conducted a prospective UK decision impact, decision conflict and economic analysis of Oncotype DX® test in early node positive, hormone receptor positive and HER2 negative breast cancer involving 664 women. Using these data, we calculated the NPI and the Predict score. Descriptive statistics, logistic regression and McNemar’s tests were used to assess the associations between NPI, Predict, RS result and change in CT decisions for all patients (n=664) and in the cohort of post-menopausal patients recruited after RxPONDER reported (n=176). Results: Table 1 shows similar stratification for all three indices within the low-risk groups, but not for the high-risk groups. Similar results were observed in the post-menopausal, post-RxPONDER cohort. The results of the CT decision change in the post-menopausal, post-RxPONDER patients by NPI and Predict demonstrate the utility of RS results across all NPI and Predict risk groups (Table 2). There was a stronger association between RS result (26-100 vs 0-25) and CT recommendation (OR152.8, 95% CI 53.9-433.2, p< 0.001) as compared to the association between either NPI or Predict (OR 8.7, 95% CI 3.8-19.8, p< 0.001 and OR 7.2, 95% CI 2.8-18.3, p=0.003 respectively), suggesting that UK Oncologists are predominantly using RS to guide their CT decisions. Conclusions: Either basing CT decisions or selecting which patients should undergo OncotypeDX testing by pre-screening with either NPI or Predict risks misclassifing many patients who either need potentially lifesaving CT or who can safely avoid chemotherapy altogether. Table 1. Predict and NPI risk groups by RS result for all patients and for the post-menopausal, post-RxPONDER cohort. Table 2. CT recommendation pre-and post-assay by risk group for post-menopausal, post-RxPONDER patients. Citation Format: Simon Holt, Paige Innis, Priyadharshini Sai-Giridhar, Satish Seerapu. A comparison of chemotherapy recommendations by NPI, Predict, and Oncotype DX testing in UK women with early node positive breast cancer [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-14-05.