Refining chemotherapy decision-making in older adults: A prospective validation of a simplified two-level CARG score...

The intermediate-risk category of the Cancer and Aging Research Group (CARG) toxicity score (6-9) is heterogeneous and complicates treatment decisions in older adults receiving chemotherapy. The CARG score is a validated geriatric assessment tool that predicts chemotherapy-related toxicity using clinical, laboratory, and functional variables. However, the broad intermediate-risk group limits its clinical applicability. We aimed to validate a simplified binary classification (CARG <8 vs ≥8), termed the CARG-8 score, as a pragmatic tool for toxicity risk stratification.

This two-phase multicenter study included a retrospective cohort of 206 patients aged ≥65 years treated with cytotoxic chemotherapy between 2019 and 2023, and a prospective cohort of 175 patients enrolled between 2022 and 2025. Patients were categorized as low-risk (CARG <8) or high-risk (CARG ≥8). The cutoff of 8 was determined using receiver operating characteristic (ROC) curve analysis and Youden's J index to optimize prediction of grade 3+ toxicity. The primary endpoint was severe (3+) toxicity; logistic regression evaluated statistical associations. The effect of upfront chemotherapy dose reduction (<90% of standard) was also explored among high-risk patients.

In the retrospective cohort, high-risk patients experienced higher rates of grade 3+ toxicity (41.0% vs 24.5%; OR 2.14, P = 0.02), treatment delays (50% vs 27%; OR 2.66, P = 0.001), and dose reductions (38% vs 16%; OR 3.21, P < 0.001). Findings were confirmed in the prospective cohort, with greater severe toxicity (29% vs 14%; OR 2.58, P = 0.02) and delays (39% vs 14%; OR 4.03, P < 0.001). Among high-risk patients, upfront dose reduction markedly decreased severe toxicity both retrospectively (23% vs 71%; OR 0.13, P < 0.001) and prospectively (16% vs 75%; P < 0.001). Multivariable analysis identified CARG ≥8 as the only independent predictor of grade 3+ toxicity (OR 4.20, P = 0.01).

A CARG cutoff of 8 effectively stratifies toxicity risk and identifies older adults who may benefit from initial dose attenuation, supporting its use for risk-adapted chemotherapy in geriatric oncology.