The charlson comorbidity index as a predictor of survival in peripheral T-cell lymphoma: A multicenter analysis from the gifu hematology study group

Abstract Background: Peripheral T-cell lymphoma (PTCL) is one of the most aggressive subtypes of non-Hodgkin's lymphoma, with a 5-year survival rate less than 40%. Commonly used prognostic models, such as the International Prognostic Index (IPI) and the Prognostic Index for T-cell lymphoma (PIT), often fail to capture the impact of non-oncological comorbidities. This is a significant limitation in Japan, where almost half of PTCL patients are aged 70 years or older. The Charlson Comorbidity Index (CCI) is a validated tool for quantifying the baseline comorbidity burden; however, its prognostic value in PTCL remains unclear. A multicenter, real-world study to assess the impact of the CCI on treatment decisions and survival outcomes in patients with PTCL was conducted. Methods: Consecutive adult patients with pathologically confirmed PTCL diagnosed between January 2004 and December 2022 across eight institutions in the Gifu Hematology Study Group (N = 185) were retrospectively reviewed. After excluding eight patients due to insufficient data for comorbidity scoring, 177 patients were included in the final analysis. The CCI was calculated at diagnosis and analyzed as both a categorical (0, 1, 2, or ≥3) and a dichotomous (0 vs ≥1) variable. The primary endpoint was overall survival (OS). Survival analyses were performed using Kaplan-Meier methods with log-rank tests. Multivariable Cox proportional hazards models were constructed, adjusting for sex, age ≥61 years, ALK-positive anaplastic large-cell lymphoma (ALCL), ECOG performance status (PS) 2-4, B symptoms, LDH elevation, Ann Arbor stage III-IV, extranodal sites ≥2, bone marrow involvement, and receipt of best supportive care (BSC). Results: The cohort included PTCL not otherwise specified (n = 112, 63.3%), angioimmunoblastic T-cell lymphoma (n = 33, 18.6%), ALK-negative ALCL (n = 17, 9.6%), and other subtypes (n =15, 8.5%). The median follow-up for survivors was 7.0 (interquartile range 2.7–12.2) years. The CCI distribution was as follows: CCI 0 (n = 111, 62.7%), CCI 1 (n = 40, 22.6%), CCI 2 (n = 21, 11.9%), and CCI ≥3 (n = 5, 2.8%). The median age increased across the CCI groups (CCI 0: 67 years; CCI 1: 72 years; CCI 2: 78 years; CCI ≥3: 75 years). The overall median OS was 2.1 (95% confidence interval [CI] 1.6-4.7) years, with a 5-year OS rate of 41.8%. Of various thresholds, dichotomization at CCI = 0 versus CCI ≥1 demonstrated the clearest survival separation (median OS 5.3 vs 1.4 years; 5-year OS 50.6% vs 26.7%; log-rank p = 0.007). A significant trend was also observed across increasing CCI levels (p for trend = 0.001), with a 1-year OS of 40% and median OS of 140 days in patients with CCI ≥3. On multivariable analysis, CCI ≥1 remained independently associated with inferior OS (hazard ratio [HR] 1.54, 95% CI 1.02-2.33, p = 0.039), along with ECOG PS 2-4 (HR 2.17, 95% CI 1.38-3.40, p < 0.001) and BSC (HR 3.18, 95% CI 1.56-6.48, p = 0.002). Compared with patients with CCI = 0, those with CCI ≥1 were more likely to receive BSC as initial therapy (15.2% vs. 3.6%, p = 0.009) and also at relapse (32.4% vs. 15.4%, p = 0.078), and they were less likely to undergo upfront autologous stem-cell transplantation (3.0% vs. 10.8%, p = 0.084). Although not statistically significant, the complete response rate tended to be higher in CCI 0 than in CCI ≥1 (53.3% vs 46.4%, p = 0.42). On exploratory receiver-operating characteristic curve analysis, the combination of PS and CCI yielded a significantly higher area under the curve (AUC) (0.725, 95% CI 0.651–0.799) than PS alone (0.683, 95% CI 0.610–0.757; p = 0.048). Although the difference was not statistically significant, the PS + CCI model showed a higher AUC than the IPI (0.700, 95% CI 0.625-0.775) and the PIT (0.681, 95% CI 0.606-0.757). Conclusions: This multicenter, real-world study demonstrated that the CCI is an independent and readily available predictor of OS in PTCL. The comorbidity burden significantly affects both survival and treatment selection, particularly in elderly or frail patients. In addition, the combination of PS and the CCI, both non-tumor-related factors, provided prognostic stratification comparable to established scores such as the IPI and the PIT. Given the limited availability of effective treatments specifically for PTCL, incorporating comorbidity assessment into routine clinical evaluation, alongside PS, may enhance predictive ability and guide more individualized treatment strategies in PTCL.