Background: Inferior outcomes have been reported for adolescents and young adults (AYAs) across several cancer types, as compared to younger pediatric or older patients (pts). While lack of access to clinical trials has been implicated, we described inferior outcomes for AYAs treated on a large US Hodgkin Lymphoma (HL) adult study (Henderson, 2018). Earlier this year, in an analysis of three US phase 3 pediatric HL trials, worse outcomes were reported for pts >15 years (yrs) compared with younger pts (Kahn, 2022). Little is known about the impact of age on outcomes from recent adult clinical trials.
Methods: Individual patient data from eight advanced stage clinical trials (ECOG2496, SWOG0816, HD2000, HD9601, HD0607, HD0801, UK Stanford V, and RATHL), conducted from 1996-2012, were harmonized as part of the HoLISTIC Consortium. Pts with classic HL, stage IIb, III or IV disease, ages 14-65 yrs, and treated at adult centers were included. Outcomes were 5-yr progression-free survival (PFS) and 5-yr overall survival (OS), which were estimated in univariable and multivariable models. Age at diagnosis and 10 other clinical factors, were examined with multivariable adjusted plots and piecewise linear splines to identify functional forms of the relationship with PFS or OS. Multiple imputation was used for missing data.
Results: Data on 3893 HL patients were included. Across all studies, 5-year PFS was 76.5% (95% CI=75.1%, 77.9%); 5-year OS was 91.6% (95% CI=90.7%, 92.6%). Median age was 32 yrs and 41% of patients were <30 yrs at diagnosis. Associations between age (analyzed as a continuous variable) and PFS displayed a distinct piecewise linear relationship with an inflection point at age 30 (Figure A). In multivariable analyses, PFS improved from ages 14 to 30, and then declined after age 30 (Figure B). The association between age and OS was not significant <30 yrs, advancing age >30 yrs was associated with worse OS. These patterns were seen broadly across studies and not dominated by one trial.
Conclusion: The association between age and HL survival in the modern era appears to be more nuanced than the dichotomous variable of age at 45 yrs, as used in previous models. While patients under 30 yrs have worse short-term disease outcomes than age 30 yrs, 5-yr OS results suggest that younger patients may be amenable to successful salvage. Further research is needed to understand these differences to optimize outcomes.