Evaluating the Comprehensive Geriatric Assessment Model As an Assessment Tool to Predict Chemotoxicity in Elderly Patients with Lymphoma

Elderly patients with hematological malignancies including lymphoma are at greater risk of chemotherapy related adverse events with poorer outcomes compared to younger patients. Conventional assessment such as the Eastern Cooperative Oncology Group (ECOG) performance status scores and the Cancer and Aging Research Group (CARG) Chemotherapy Toxicity Tool may over-estimate the fitness of elderly patients in receiving anti-cancer treatment, hence exposing them to greater risks. This study aims to evaluate the Comprehensive Geriatric Assessment (CGA) model as a tool to identify frail elderly (Asian) patients with lymphoma who are at greater risk of chemo-toxicities, and compare it with ECOG and CARG assessment tools.

A prospective analysis was conducted on patients over 65 years of age with newly diagnosed lymphoma receiving chemotherapy. CGA scores were assessed prior to treatment, looking at the following geriatric domains: activities of daily living, falls history, hearing/visual impairment, nutritional assessment, cognition, polypharmacy and psychosocial support. ECOG performance status and CARG scores were also collected. Treatment related adverse events (TRAE) were recorded at each clinical encounter, and classified according to the Common Terminology Criteria for Adverse Events. The relationship between CGA scores, ECOG scores, CARG scores, Overall Survival (OS) and incidence of severe TRAEs (Grades 3 to 5) was evaluated.

104 patients were analyzed, with a median age of 75 years (range 65–93). Diffuse Large B Cell Lymphoma was the most common subtype (72 patients, 69.2%). 29 (27.9%) patients were identified as Fit, 61 (58.7%) as Pre-Frail, and 14 (13.4%) as Frail. Of note, 50% of patients identified as Frail on CGA were also categorized as ECOG 0, whilst 62.3% of Pre-Frail patients were categorized as ECOG 0. Similarly, 28.6% of patients identified as Frail were categorized as Low and Medium risk on CARG scoring, and 14.8% of patients identified as Pre-Frail were also categorized to be Low risk on CARG. An upfront chemotherapy dose reduction strategy was employed in 78.6% of CGA Frail patients, compared to 57.4% and 20.7% of CGA Pre-Frail and Fit patients respectively.

699 severe TRAEs were reported throughout the follow-up duration. Fit patients experienced an average of 4.97 severe TRAEs, whilst Pre-Frail and Frail patients experienced a higher average of 7.26 and 8.00 events respectively (P-value=0.165). Severe hematological TRAEs were more common (63.2%) compared to non-hematological TRAEs, with Fit patients experiencing an average of 3.27 hematological events compared to 4.61 and 4.71 events respectively for Pre-Frail and Frail patients (P-value=0.504). Fit patients had an average of 1.69 severe non-hematological TRAEs, compared to 2.65 and 3.28 events respectively for Pre-Frail and Frail patients (P=0.067). Patients with High Risk on CARG had similar trends of increased incidence of TRAEs and hematological TRAEs (10.7 / 6.26 events) compared to Medium (6.56 / 4.39 events) and Low Risk (6.96 / 4.59 events) patients.

Median OS of Frail patients was 274 days, while the median OS of Pre-Frail and Fit patients were not reached. Hazard ratios (HRs) for death of Frail patients and Pre-Frail patients compared to Fit patients were 5.72 (95% CI 1.72-19.10), and 1.23 (95% CI 0.39-3.92) respectively. Conversely, CARG scoring did not predict for OS, with HRs for death for Medium Risk patients and High Risk patients compared to Low Risk patients being 0.36 (CI 0.10-1.34) and 1.90 (0.68-5.33) respectively.

Conclusion

A significant proportion of Frail and Pre-Frail patients were initially identified as ECOG performance status 0, as well as Low to Medium risk under the CARG Toxicity Tool. Frail and Pre-Frail patients have a trend to higher incidence of severe TRAEs compared to Fit patients, however this was not statistically significant within the limits of this small study. Additionally, Frail patients demonstrate significantly lower overall survival rates compared to Pre-Frail and Fit patients. ECOG and CARG are widely used to predict the patient’s ability to tolerate chemotherapy, but our study shows that these tools may not accurately predict toxicity risk. CGA appears to be a better tool in predicting chemotoxicity and overall survival in elderly lymphoma patients and should be utilized instead.