Geriatric assessment-based therapy for older patients with diffuse large B-cell lymphoma: results of a two-decade clinical experience

Treating older patients with diffuse large B-cell lymphoma (DLBCL) is challenging due to heterogeneity in fitness and tolerance to chemotherapy with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP). Thus, we previously established the Age, Comorbidity, and Albumin (ACA) index for this population, a geriatric assessment comprising four categories. However, its role in treatment decision-making remains unclear. We conducted a retrospective study in patients ≥ 65 years old with DLBCL who were treated with R-CHOP between 2015 and 2022 (post-ACA cohort), evaluating initial dose intensity (IDI), relative dose intensity (RDI), and 2-year overall survival (OS). During this period, dose intensities of cytotoxic drugs were adjusted by the multidisciplinary oncology team according to the ACA index categories (Excellent, Good, Moderate, and Poor). Then, we compared these outcomes with a historical cohort of patients receiving R-CHOP between 2001 and 2012 (pre-ACA cohort), where the dose intensities were determined by each physician’s discretion. Propensity score matching (PSM), adjusted for the International Prognostic Index and the ACA index, was used to balance baseline characteristics between the cohorts. In the post-ACA cohort comprising 135 patients with a median age of 76 years (range: 65–91 years), the median IDI, median RDI, and 2-year OS according to the aforementioned ACA categories were 97%, 80%, 70%, and 70% (p < 0.001); 93%, 78%, 62%, and 59% (p < 0.001); and 96%, 89%, 66%, and 51% (p < 0.001), respectively. After PSM, 98 patients from the pre- and post-ACA cohorts were paired. Despite well-balanced patient demographics, granulocyte-colony stimulating factor prophylaxis was more commonly used in the post-ACA group (100% vs. 49%, p < 0.001). Consequently, the post-ACA group had higher median IDI and RDI (80% vs. 68%, p < 0.001, and 77% vs. 47%, p < 0.001, respectively) and better 2-year OS (81% vs. 66%, p = 0.014). The ACA index can serve as a feasible standard-of-care framework for individualized dose intensity adjustment in older patients with DLBCL. Although this strategy organizes treatment decision-making, vulnerable patients still exhibit significantly inferior outcomes, highlighting the need for further therapeutic optimization and novel strategies.