Earlier brain tumor detection: Assessing the impact on survival, recurrence, inpatient length of stay and neurological deficit.

Abstract

e23268 Background: Brain tumors commonly present in an emergency setting, but the majority of patients diagnosed with a brain tumor will have had one or more primary care visits in the preceding months. This suggests opportunities for earlier diagnosis. Early economic evaluation of a brain cancer liquid biopsy identified an evidence gap in quantifying the impact of earlier/later diagnosis on patient outcomes. The aim of this study is to estimate the association between two indicators of earlier/later diagnosis, tumor size and the length of diagnostic interval, and the patient outcomes of cancer-specific survival, overall survival, new or worsened neurological deficit, cancer recurrence and inpatient length of stay in 12 months following diagnosis. Methods: A retrospective cohort study utilising a database of 1400 patients from Southeast Scotland, UK, diagnosed with brain tumors between 2010 and 2020. Diagnostic interval was defined as time from first presentation to radiological diagnosis. Tumor size was defined as the recorded maximum diameter at time of diagnosis. Hazard ratios for cancer-specific and overall mortality, per mm increase in tumor size or per day increase in diagnostic interval, were calculated. Odds ratios for new or worsened neurological deficit and for tumor recurrence were similarly calculated. Linear regression was used to estimate effects on total inpatient bed days. Regression was used for all outcomes to adjust estimates for possible confounding factors. Alternative diagnostic window definitions were explored in sensitivity analysis. Analysis was repeated stratified analysis by brain tumor subtype was used to explore generalisability across subtypes. Results: There were 1196 eligible patients that were included in the study. The mean tumor size was 40.6mm (SD: 17.5mm). Each 1mm increase in tumor size increases mortality risk by approximately 1%, increases the expected inpatient days by 0.15 days, and the risk of new or worsened neurological deficit by 2%. Diagnostic interval results (in preparation) will be presented at the meeting. Conclusions: Based on previous estimates of brain tumor growth rates, a tumor with a 40mm diameter at diagnosis (mean diameter in this patient population) would be expected to be 28mm if diagnosed 1 month earlier. Applying the effect size estimates reported above would translate into clinically important differences in outcomes. This highlights the improvement in patient outcomes that could be expected if early intervention is made possible by new diagnostic tools.