Unraveling the web of frailty: A novel approach to frailty using a CGA-derived spider plot.

Dana Cavanaugh,Sarah K Holt,Erin Petersen,Samia Jannat,Jonathan L. Wright,John L. Gore,George Schade,Sarah P. Psutka

DOI: https://doi.org/10.1200/jco.2024.42.4_suppl.579

IF: 45.3

2024-01-31

Journal of Clinical Oncology

Abstract:579 Background: Frailty in bladder cancer predicts increased complications and mortality. Measuring frailty is challenging because relying on one frailty instrument risks oversimplifying patients’ multidimensional vulnerability profiles. Current guidelines recommend using a comprehensive geriatric assessment (CGA) to quantify frailty across domains of physical function, mental health/cognition, nutrition, and multimorbidity in older adults. However, the CGA’s extensive data can be hard to synthesize and integrate into busy clinical practice. We propose a novel CGA-derived Frailty Spider Plot, synthesizing CGA-identified vulnerabilities into a clinically useful frailty profile. Methods: Urothelial cancer patients, prospectively enrolled between 9/2020 and 7/2021 in a multidisciplinary bladder cancer clinic, completed a CGA incorporating validated assessments of functional status, multimorbidity, nutrition, cognition, and mental health, augmented with CT-derived assessments of muscle mass and adiposity. Spearman Correlation Coefficients were used to quantify relationships between frailty tools and domains. Novel Frailty Spider plots were then created to visualize individuals’ distinct frailty phenotypes. The CGA instruments, grouped by domain, were the spokes of the plot, with each node determined by validated thresholds. Outer ring values indicated greater frailty. Results: The cohort included 67 patients (median age 71, 16.4% female), most with muscle-invasive bladder cancer (77.6%; cN+: 20.9%, M1: 7.6%). The CGA identified 31.4% vulnerable-to-moderately frail, 23.9% at risk for falls, 13.4% with mild depression, 7.5% with moderate-severe depression, 3% with mild/moderate dementia, 34% at risk for malnutrition, and 6% malnourished. Frailty measures were generally weakly correlated (rho <|0.5|). Individuals’ specific risk profiles were plotted on the novel Frailty spider plot which visually distinguished patients according to their frailty phenotypes. For instance, a patient with a high risk Mini-nutritional Assessment score, hypoalbuminema, and sarcopenic skeletal muscle mass, is distinguished with a Spider plot dominated by nutritional frailty spokes. Conclusions: In this prospective observational cohort, a CGA identified key vulnerabilities beyond a standard risk assessment. Frailty domains provide unique risk insights beyond relying on simple frail/not frail nomenclature. Notably, even within a domain, instruments had low correlations, contributing distinct insights into a patient’s vulnerability profile. Our novel Spider plot approach visually synthesizes this data in a single depiction of the predominant components of frailty that can be targeted by personalized prehabilitation interventions. Future work will refine risk assessment instruments to best predict clinical outcomes, creating parsimonious Frailty spider plots for clinical use.

oncology