Association of quality of life (QoL) with mortality in patients with adenoid cystic carcinoma using an internationally validated QoL questionnaire (EQ-5D-5L): A new baseline.

6097 Background: An evaluation of impact on quality of life (QoL) is required for regulatory approval and reimbursement of new drug therapies. Almost all studies of new therapies in adenoid cystic carcinoma (ACC) are non-randomised with no comparator QoL measure. To support Health Technology Assessments and the evaluation of the clinical utility of these therapies, we sought to determine the QoL measures in a cohort of ACC patients and to assess for change over time and associations with clinical or prognostic factors. Methods: An internationally validated QoL questionnaire (EQ-5D-5L) was completed by ACC patients referred to an experimental medicine centre between 2019 and 2024, who provided written informed consent to analysis. EQ-5D value scores (EQV) were calculated from each questionnaire using the EuroQol England value set. A Cox proportional hazards model was built with EQV as a time dependent variable, with a variant using penalised smoothing splines to assess linearity. A base non-linear mixed effects (nlme) model was constructed between time and EQV. The relationship between EQV and other predictors (NOTCH1 alteration status, age at diagnosis, sex, local and/or metastatic recurrence, and primary site of disease) were tested against the base model using variant models built for each predictor. Both analyses used R Statistical Software (v4.3.2) with survival and nlme packages in Rstudio (v2023.12.0.0). Results: Between 2019 and 2023, 563 EQ-5D-5L questionnaires were completed by 161 patients with ACC (median 2 responses/patient, range 1 to 20). NOTCH1 alteration was seen in 8% (13/161), median age at diagnosis 49 years (range 17 to 81), 40% male, 36% (58/161) had a major salivary gland as primary site, and 14% had local recurrence, 46% metastatic recurrence and 26% both for 86% total with recurrent disease (139/161). For all 161 patients, median EQV was 0.81 (range -0.22 to 1.0); an EQV of 1 is ideal QoL and 0 is QoL as bad as death. The median EQV for the clinical subgroups were 0.83 (NOTCH1, IQR 0.74-0.87), 0.84 (no recurrence, IQR 0.75-0.92), 0.80 (local recurrence only, IQR 0.70-0.86), 0.81 (distant metastasis only, IQR 0.70-0.92), 0.83 (distant and local, IQR 0.75-0.92), 0.84 (major primary site, IQR 0.75-0.92), 0.81 (other primary site, IQR 0.70-0.89), and no statistical differences were identified between these groups. However, a decrease in EQV from 1 to 0 was associated with an eightfold increase in risk of death in the total population (HR= 0.118, 95% CI 0.057 to 0.244, p= <0.001). There was a significant non-linear relationship between EQV and survival over time (p=0.011), with a greater marginal effect on survival for EQVs >0.81. Conclusions: For patients with ACC, a worse QoL as measured by EQ-5D-5L was associated with a significantly increased risk of death. NOTCH1 and disease recurrence were not associated with a worse QoL.