ESTIMATING MEANINGFUL CHANGE THRESHOLDS (MCTS) FOR EUROPEAN ORGANISATION FOR RESEARCH AND TREATMENT OF CANCER (EORTC) SCALES IN A PHASE 3 TRIAL FOR PATIENTS WITH LOCALLY RECURRENT INOPERABLE OR METASTATIC TRIPLE-NEGATIVE BREAST CANCER (MTNBC)
Author(s)
Nikki Ow, PhD1, Yanyan Zhu, MPH, PhD2, Rachel Lai, MPH3, Manjunatha Ankathatti Munegowda, MD4, Petra Vukovic, MD5, Kechen Zhao, MD6, Micah Maxwell, MD7, Karson Tse, MD7, Cristina Ivanescu, PhD8, Paolo Eusebi, MD9;
1Evinova, Oncology R&D, AstraZeneca, Mississauga, ON, Canada, 2Evinova, Oncology R&D, AstraZeneca, Waltham, MA, USA, 3Evinova, Oncology R&D, AstraZeneca, Gaithersburg, MD, USA, 4AstraZeneca, Mississauga, ON, Canada, 5AstraZeneca, Cambridge, United Kingdom, 6AstraZeneca, Wilmington, DE, USA, 7AstraZeneca, Gaithersburg, MD, USA, 8IQVIA, Patient-Centered Endpoints, Amsterdam-Zuidoost, Netherlands, 9IQVIA, Milan, Italy
1Evinova, Oncology R&D, AstraZeneca, Mississauga, ON, Canada, 2Evinova, Oncology R&D, AstraZeneca, Waltham, MA, USA, 3Evinova, Oncology R&D, AstraZeneca, Gaithersburg, MD, USA, 4AstraZeneca, Mississauga, ON, Canada, 5AstraZeneca, Cambridge, United Kingdom, 6AstraZeneca, Wilmington, DE, USA, 7AstraZeneca, Gaithersburg, MD, USA, 8IQVIA, Patient-Centered Endpoints, Amsterdam-Zuidoost, Netherlands, 9IQVIA, Milan, Italy
OBJECTIVES: To estimate the MCT for selected EORTC scales used in trial participants with first-line, locally recurrent, inoperable or mTNBC (NCT05374512).
METHODS: Pooled blinded data were analyzed to determine MCTs for within-participant deterioration from baseline to week 12 on selected EORTC IL146 and IL116 scales, using anchor-based methods and empirical cumulative distribution function curves. Distribution-based estimates (half-standard deviation [SD] at baseline and one standard error of measurement [SEM]) ensured that anchor-based estimates were above the natural variability of the change scores. A maximum of three anchors per scale were considered, with Patient Global Impression of Severity and Patient Global Impression of Change as primary anchors. Anchor appropriateness was assessed via polyserial correlations, with values ≥0.371 considered adequate.
RESULTS: 422 (Global Health Status [GHS]/Quality of Life [QoL]/Fatigue) and 423 (all other scales) participants completed EORTC IL146 at baseline, and 329 (GHS/QoL) or 338 (all other scales) at 12 weeks. For IL116, 428 participants completed scales at baseline and 340 (Arm symptoms) or 339 (Breast symptoms) at 12 weeks. For EORTC IL146, half-SD at baseline ranged from 9.59 (Cognitive Functioning) to 15.44 (Pain), and SEM from 5.60 (GHS/QoL) to 12.07 (Cognitive Functioning). For EORTC IL116, half-SD at baseline/SEM, was 11.98/11.53 (Arm) and 13.08/9.87 (Breast). All scales showed appropriate correlations with ≥1 anchors, except Cognitive Functioning (≥0.30 but <0.371). Distribution-based estimates were used for Cognitive Functioning, and Breast and Arm symptoms, given anchor estimates were lower. Following triangulation and considering minimum observable change values, within-participant MCTs for deterioration were 22.22 (Fatigue, Arm symptoms), 16.67 (Pain, Breast symptoms), −16.67 (Emotional, Cognitive, Social Functioning), −33.33 to −16.67 (GHS, QoL, Role Functioning), and −13.33 (Physical Functioning).
CONCLUSIONS: Interpretation of PRO endpoints is limited without pre-defined thresholds. These MCTs allow for better interpretation of the PRO endpoints in this trial and understanding of longitudinal change in PRO in clinical trials.
METHODS: Pooled blinded data were analyzed to determine MCTs for within-participant deterioration from baseline to week 12 on selected EORTC IL146 and IL116 scales, using anchor-based methods and empirical cumulative distribution function curves. Distribution-based estimates (half-standard deviation [SD] at baseline and one standard error of measurement [SEM]) ensured that anchor-based estimates were above the natural variability of the change scores. A maximum of three anchors per scale were considered, with Patient Global Impression of Severity and Patient Global Impression of Change as primary anchors. Anchor appropriateness was assessed via polyserial correlations, with values ≥0.371 considered adequate.
RESULTS: 422 (Global Health Status [GHS]/Quality of Life [QoL]/Fatigue) and 423 (all other scales) participants completed EORTC IL146 at baseline, and 329 (GHS/QoL) or 338 (all other scales) at 12 weeks. For IL116, 428 participants completed scales at baseline and 340 (Arm symptoms) or 339 (Breast symptoms) at 12 weeks. For EORTC IL146, half-SD at baseline ranged from 9.59 (Cognitive Functioning) to 15.44 (Pain), and SEM from 5.60 (GHS/QoL) to 12.07 (Cognitive Functioning). For EORTC IL116, half-SD at baseline/SEM, was 11.98/11.53 (Arm) and 13.08/9.87 (Breast). All scales showed appropriate correlations with ≥1 anchors, except Cognitive Functioning (≥0.30 but <0.371). Distribution-based estimates were used for Cognitive Functioning, and Breast and Arm symptoms, given anchor estimates were lower. Following triangulation and considering minimum observable change values, within-participant MCTs for deterioration were 22.22 (Fatigue, Arm symptoms), 16.67 (Pain, Breast symptoms), −16.67 (Emotional, Cognitive, Social Functioning), −33.33 to −16.67 (GHS, QoL, Role Functioning), and −13.33 (Physical Functioning).
CONCLUSIONS: Interpretation of PRO endpoints is limited without pre-defined thresholds. These MCTs allow for better interpretation of the PRO endpoints in this trial and understanding of longitudinal change in PRO in clinical trials.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
CO46
Topic
Clinical Outcomes
Disease
SDC: Cardiovascular Disorders (including MI, Stroke, Circulatory)