DETERMINATION OF MEANINGFUL WITHIN-PATIENT CHANGE THRESHOLDS (MWPC) FOR THE FUNCTIONAL ASSESSMENT OF CANCER THERAPY-GASTRIC (FACT-GA) FROM STAR-221 TRIAL
Author(s)
Mira Patel, PhD1, Alfred SACKEYFIO, MSc2, Avni Shah, MBA3, Christopher Pelligra, MSc4, SAEID SHAHRAZ, PhD1, Shien Guo, MHA, PhD5.
1Gilead Sciences, Inc., Foster City, CA, USA, 2Gilead Sciences, Inc., Stockley Park, United Kingdom, 3Arcus Biosciences, Hayward, CA, USA, 4PPD Evidera Patient-Centered Research, Thermo Fisher Scientific, Atlanta, GA, USA, 5PPD Evidera Patient-Centered Research, Thermo Fisher Scientific, Waltham, WA, USA.
1Gilead Sciences, Inc., Foster City, CA, USA, 2Gilead Sciences, Inc., Stockley Park, United Kingdom, 3Arcus Biosciences, Hayward, CA, USA, 4PPD Evidera Patient-Centered Research, Thermo Fisher Scientific, Atlanta, GA, USA, 5PPD Evidera Patient-Centered Research, Thermo Fisher Scientific, Waltham, WA, USA.
OBJECTIVES: The FACT-Ga is a 46-item, multi-domain patient-reported outcome (PRO) instrument developed to assess health-related quality of life in patients with gastric cancer (GC). The study aimed to determine MWPC thresholds for worsening and improvement in the FACT-Ga subscale, trial outcome index (TOI), and total score (TS).
METHODS: Analyses were conducted using the intent-to-treat population from a blinded interim dataset of the phase 3 STAR-221 trial (NCT05568095) in patients with previously untreated locally advanced unresectable or metastatic GC, gastroesophageal junction cancer (GEJC) or esophageal adenocarcinoma (EAC). Patients completed the FACT-Ga and anchor measures on Day 1 of each treatment cycle. Correlations between the changes in the FACT-Ga and candidate anchors were examined (|r|>0.3 considered acceptable). Anchor-based methods were used to estimate MWPC thresholds from mean and median FACT-Ga score changes within anchor-defined response categories. Distribution-based estimates, including standard error of measurement (SEM) and one-half standard deviation (1/2SD), and empirical cumulative distribution function (eCDF) curves were used to support threshold selection.
RESULTS: A total of 781 patients were included. Selected anchors, including the Patient Global Impression of Severity (PGIS) scale and EQ-5D-5L pain/discomfort and usual activities domains, demonstrated moderate correlations with changes in FACT-Ga scores (r=-0.38 to -0.54). For all FACT-Ga scales, selected anchor-based thresholds exceeded corresponding SEM and 1/2SD estimates. Proposed MWPC thresholds for worsening and improvement, respectively, were ≤-7.0 and ≥12.0 for FACT-Ga subscale (on a score range of 0 - 76), ≤-12.0 and ≥15.0 for the TOI (0 - 132), and ≤-14.0 and ≥16.0 for the TS (0 - 184). eCDF curves generally demonstrated clear separation between anchor-defined response groups.
CONCLUSIONS: The analyses established interpretable MWPC thresholds for the FACT-Ga subscale, TOI, and TS in first-line advanced GC/GEJC/EAC. These thresholds support interpretation of change in FACT-Ga scores and may facilitate evaluation of responder-based and time-to-event PRO endpoints in future GC clinical trials.
METHODS: Analyses were conducted using the intent-to-treat population from a blinded interim dataset of the phase 3 STAR-221 trial (NCT05568095) in patients with previously untreated locally advanced unresectable or metastatic GC, gastroesophageal junction cancer (GEJC) or esophageal adenocarcinoma (EAC). Patients completed the FACT-Ga and anchor measures on Day 1 of each treatment cycle. Correlations between the changes in the FACT-Ga and candidate anchors were examined (|r|>0.3 considered acceptable). Anchor-based methods were used to estimate MWPC thresholds from mean and median FACT-Ga score changes within anchor-defined response categories. Distribution-based estimates, including standard error of measurement (SEM) and one-half standard deviation (1/2SD), and empirical cumulative distribution function (eCDF) curves were used to support threshold selection.
RESULTS: A total of 781 patients were included. Selected anchors, including the Patient Global Impression of Severity (PGIS) scale and EQ-5D-5L pain/discomfort and usual activities domains, demonstrated moderate correlations with changes in FACT-Ga scores (r=-0.38 to -0.54). For all FACT-Ga scales, selected anchor-based thresholds exceeded corresponding SEM and 1/2SD estimates. Proposed MWPC thresholds for worsening and improvement, respectively, were ≤-7.0 and ≥12.0 for FACT-Ga subscale (on a score range of 0 - 76), ≤-12.0 and ≥15.0 for the TOI (0 - 132), and ≤-14.0 and ≥16.0 for the TS (0 - 184). eCDF curves generally demonstrated clear separation between anchor-defined response groups.
CONCLUSIONS: The analyses established interpretable MWPC thresholds for the FACT-Ga subscale, TOI, and TS in first-line advanced GC/GEJC/EAC. These thresholds support interpretation of change in FACT-Ga scores and may facilitate evaluation of responder-based and time-to-event PRO endpoints in future GC clinical trials.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
PCR243
Topic
Methodological & Statistical Research, Patient-Centered Research
Topic Subcategory
Patient-reported Outcomes & Quality of Life Outcomes
Disease
Oncology