Introduction: Decision-making for the treatment of ES HL patients continues to involve consideration of the tradeoff between shorter-term progression-free survival (PFS) and the impact of late effects (LE) on overall survival (OS). To evaluate the complex decision process arising from these factors, there is a desire to obtain integrated data that assimilates short-term outcomes, LE, and quality adjusted life expectancy (QALE) to help aid treatment choices.
Methods: We analyzed PFS and OS from 4 large, modern randomized trials comparing chemotherapy alone with combined modality therapy (CMT) for ES HL (Meyer NEJM 2012; Wolden JCO 2012; Raemaekers JCO 2014; Radford NEJM 2015). We constructed a detailed Markov model comparing QALE after initial therapy for ES HL using chemotherapy alone vs CMT. The model incorporated: 3-year PFS from PET- patients; quality of life associated with relapse and cure; cure rates after relapse; 15-year latency period before onset of increased mortality from LE; and “discounting” of future life years. We also performed sensitivity analyses to determine the amount that CMT (with contemporary radiotherapy) must increase LE-associated mortality for that impact to outweigh its early PFS advantage.
Results: All 4 trials demonstrated that short-term PFS was superior with CMT, including PET response-adapted paradigms, however OS was at least equivalent in each study. Collectively, the analyses showed that if CMT-associated LE increased deaths >15-20% (relative) compared with chemotherapy alone, then treatment with the latter was associated with overall improved QALE. This finding was consistent across a range of patient ages and among varied estimates for SMR after cure for those treated with chemotherapy alone (see Figure). Furthermore, sensitivity analyses demonstrated that these findings were most sensitive to assumptions regarding the impact of CMT on mortality after cure, and insensitive to moderate changes in the assumed short-term impact of CMT on PFS.
Conclusions: We identified the “break-even point” whereby chemotherapy alone was associated with superior QALE vs CMT (ie, 15-20% relative increased death from LE). This model provides the base analytic framework to aid the complex decision of treating ES HL. Future model iterations may incorporate patient & tumor characteristics, patient preferences, and cost-effectiveness of varied therapeutics in order to help guide individualized decisions for providers and patients with HL.