Wake Forest University School of Medicine Winston Salem, NC, United States
R. T. Hughes1, N. Razavian1, and M. Farris2; 1Department of Radiation Oncology, Wake Forest University School of Medicine, Winston Salem, NC, 2Department of Radiation Oncology, Wake Forest University School of Medicine, Winston-Salem, NC
Purpose/Objective(s): To determine whether an electronic frailty index (eFI) calculated at time of consultation is associated with survival and radiotherapy (RT) outcomes. Materials/
Methods: Patients eligible for analysis received RT within our comprehensive cancer center network between August 2019 to July 2022 and had eFI calculated at the time of consultation. An eFI, developed and previously reported by investigators at our institution, constructs a score using electronic health record (EHR)-based patient factors such as vital signs, smoking status, comorbid diagnosis codes, medications, laboratory values, and functional information from at least 2 ambulatory visits over the preceding 2 years. The eFI is calculated automatically in the EHR; scores range between 0 to 1. Patients were categorized as fit (eFI =0.10), pre-frail (>0.10 to 0.21), and frail (>0.21). If available, additional comorbidity indices also calculated in the corresponding encounter (Charlson Comorbidity Index [CCI] and Elixhauser Comorbidity Index [ECI]) were extracted. Hospitalizations during or within 6 months of RT completion were recorded. Overall survival (OS) from the date of consultation was estimated using the Kaplan Meier method. A P-value <0.05 was considered statistically significant. Results: In total, 399 patients were analyzed. Median age was 73 years; 52% were male, 83% white and 15% black. The most common primary malignancies were lung (27%), breast (21%), and genitourinary (19%). Median eFI score was 0.14 (range, 0-0.52); 30%, 49%, and 22% were categorized as fit, pre-frail, and frail, respectively. CCI and ECI metrics were available for 269 (67%) and 271 (68%) patients, respectively. eFI was significantly correlated with CCI (Spearman’s ?=0.51, p<0.01) and ECI (?=0.35, p<0.01). Three-year OS estimates for fit, pre-frail, and frail patients were 69.4%, 50.7%, and 44.0%, respectively (log-rank p<0.01). Of the 252 patients treated with RT, intent was curative in 184 (73%) and palliative in 68 (27%). In patients treated with curative RT, median prescribed dose was 54 Gy (range, 20-79.2) and median number of fractions was 25 (3-44). eFI category was associated with prescribed number of fractions (fit: 28, pre-frail: 25, frail: 24; p=0.02) and time on treatment (fit: 38 days, pre-frail: 32, frail: 23; p=0.02). Increasing frailty as determined by the eFI was associated with higher rates of incomplete treatment (fit: 0%, pre-frail: 4%, frail: 10%; p=0.05). The hospitalization rate during or within 6 months of RT was 11.5%. There was no association between eFI and hospitalization for patients treated with curative or palliative RT. Conclusion: The eFI calculated at the radiation oncology consultation is associated with OS in all patients and RT completion in patients treated with curative intent RT. This frailty metric, that is automatically calculated by the EHR at the point of contact, may assist radiation oncologists in prognostic counseling and fractionation decisions.
