Session: 902. Health Services Research—Malignant Diseases: Poster III
Hematology Disease Topics & Pathways:
Diseases, Leukemia, Follicular Lymphoma, AML, Adult, CLL, Lymphoma (any), CML, Mantle Cell Lymphoma, Elderly, DLBCL, Non-Hodgkin Lymphoma, MDS, B-Cell Lymphoma, Study Population, Lymphoid Malignancies, Clinically relevant, Myeloid Malignancies, Quality Improvement
The comprehensive geriatric assessment (CGA) in older patients with cancer is the gold standard to identify robust, frail or poor prognosis patients according Balducci classification. In Spain, a new proposal of a specific Geriatric Assessment in Hematology (GAH) scale has been designed and validated in patients with hematologic malignancies such as MDS/AML, multiple myeloma and CLL. The GAH scale has not been explored in patients with lymphoma. In this study, we have analyzed the utility of using the GAH scales in patients with hematologic malignancies, mostly lymphoma patients.
Patients and methods.
From March 2016 and September 2017, patients with hematologic malignancies were prospectively referred to the Geriatric Oncology clinic after a frailty screening test using G8 scale and with score <14 points. All patients were assessed with CIRS-G and GAH scales performed by the oncology nurses and a comprehensive geriatric assessment performed by the geriatrician.
Of the 96 patients referred aged 70 years or over, 41 were males (42.7%) and 55 females (57.3%), the median age was 79 years (range, 70-89), and with the diagnosis of lymphoma in 53 patients (55.2%), multiple myeloma in 23 patients (24.0%), CLL in 13 patients (13.6%), MDS/AML in 5 patients (5.2%) and CML in 2 patients (2.0%). Seventy-five patients (78.1%) had good performance status with ECOG score 0-1. Regarding frailty, 20 patients (20.8%) had a score of 15 points or over at G8 scale and 76 patients (79.2%) were identified as frail because of a score of 14 points or below. Regarding comorbidities, the median CIRS-G score was 9 (range, 4-20).
After the GAH scale assessment, the median number of domains affected in robust patients was 2 (1-4) and in frail patients was 4 (3-5) (p=0.0001). In the ROC curve, with an AUC of 0.7595 and a likelyhood ratio of 9, the cut-off in this series was 2 domains with impairment, with a sentivity of 13.79% and a specificity of 92.5% (p= 0.0003). Using a correlation factor for each domain, the mean score at GAH scale in robust patients was 26 points and in frail patients was 42.5 points (p=0.0038). In the ROC curve, with an area under the curve of 0.7026 and a likelihood ratio of 2.04, the cut-off value to identify robust vs frail patients was 33 points in the GAH scale, with a sensitivity of 77.5% and a specificity of 62.07% (p=0.0043).
Analyzing the eight domains explored in the GAH scale, robust patients according CGA had less risk of polypharmacy of 31.25% vs 81.48% in frail patients (OR 0.1033, 95% CI 0.0472-0.2541) (p<0.0001), less gate speed/FAC impairment of 16.66% vs 81.48% (OR 0.04545, 95% CI 0.0183-0.1313) (p<0.0001), less ADL impairment 37.5% vs 85.19% (OR 0.1043, 95% CI 0.0398-0.2684) (p<0.0001), less mood impairment in 4.17% vs 40.74% in frail patients (OR 0.06324, 95% CI 0.01421-0.2498) (p<0.0001), less mental health impairments in 2.08% vs 22.22% in frail patients (OR 0.0744, 95% CI 0.0068-0.4531) (p=0.0023), less comorbidities in 2.08% vs 42.59% (OR 0.0286, 95% CI 0.0027-0.1817) (p<0.0001), less malnutrition in 10.42% vs 37.04% (OR 0.1977, 95% CI 0.0759-0.5495) (p=0.0024), and less poor self-reported well-being in 6.25% vs 66.67% (OR 0.0333, 95% CI 0.0101-0.1187) (p<0.0001). The median overall survival for patients with 3 or less domains impaired was not reached vs 90.77 months in those patients with 4-8 domains impaired (Log-rank test, p=0.0003), with HR (Log-rank) of 0.11 (95% CI, 0.04474-0.2846).
Mean G8 score were similar between robust (11.68) and frail (11.04) patients (p=n.s.) among all patients with score below 14 points. Robust patients had less comorbidities according to CIRS-G scale, with a median of 9 vs 11 points (p=0.0001). There was correlation between CIRS-G and ECOG with G8 score, not found in previous studies. There is a correlation between the brief comorbidity assessment in the GAH scale with CIRS-G score. Among patients identified as not having comorbidities, the median CIRS-G score was 9 vs 13.5 among patients with comorbidities according the GAH scale (p<0.0001).
The GAH scale is a valid tool for patients with hematologic malignancies, including patients with lymphoma, in order to classify patients according frailty phenotype. All domains explored in GAH scale were impaired with higher frequency in frail patients. Robust patients had less comorbidities and better performance status. The brief comorbidities assessment in the GAH scale correlates well with the CIRS-G.
Disclosures: No relevant conflicts of interest to declare.
See more of: Oral and Poster Abstracts