PET-adapted Strategies in Advanced Stage Hodgkin Lymphoma: Big to Small, or Small to Big, or Does It Even Matter?

The preferred treatment for advanced-stage Hodgkin lymphoma (HL) has historically differed with geography, with oncologists in the United States favoring chemotherapy with ABVD (doxorubicin, bleomycin, vinblastine, dacarbazine), while Europeans have favored escalated-BEACOPP (escBEACOPP; bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, prednisone). ABVD is associated with an inferior progression-free survival (PFS), but its proponents argue that 1) this decrement in PFS is less important than the increase in toxicity, including serious infection, infertility, and secondary malignancies, seen with escBEACOPP and 2) there is no overall survival (OS) benefit attached to the more toxic regimen. In an attempt to avoid excessive toxicity in patients expected to do well with ABVD, the Risk-Adapted Treatment of Advanced Hodgkin Lymphoma (RATHL) and Southwest Oncology Group (SWOG) S0816 trials used a PET-adapted treatment escalation design in which patients with positive positron emission tomography (PET) scans following two cycles (PET2) of ABVD chemotherapy had their therapy intensified to BEACOPP or escBEACOPP.^1,2  Patients with a negative PET2 either continued ABVD for six cycles (SWOG S0816) or were randomized to complete six cycles of ABVD or to have bleomycin removed from subsequent cycles (RATHL). With this strategy, two- and three-year PFS rates for patients on the SWOG S0816 and RATHL trials overall were 79 percent and 67.5 percent, respectively. Dr. René-Olivier Casasnovas and colleagues reported the opposite strategy in Lancet Oncology this year; they discussed treatment de-escalation based on a reassuring PET response after two cycles of escBEACOPP in an attempt to limit toxicity in patients for whom it is not needed.

In the AHL2011 study, reviewed here in the context of the existing literature, all patients with advanced-stage HL were started on escBEACOPP. Patients were randomized to standard treatment (n=413) consisting of six cycles of escBEACOPP, or to PET-driven treatment (n=410) where the results of PET2 were used to determine subsequent therapy. PET2-positive patients (Deauville 4 or 5) received additional escBEACOPP and PET2-negative patients (Deauville 1-3) deintensified therapy to ABVD. PET2 was negative in 88 percent and 87 percent of patients in the standard treatment and PET-driven treatment groups, respectively. The primary endpoint of the study was five-year PFS, which was similar in both groups (86.2% vs. 85.7%, respectively) and compares favorably with the PFS results of studies of ABVD alone (2- to 5-year PFS 77-79%), strategies of PET-driven treatment intensification such as RATHL (3-year PFS 67.5%), and a new standard for advanced-stage HL, AVD+brentuximab, from the ECHELON-1 study (2-year PFS 82.1%).^1-4  When considering PET2-positive patients in each group, five-year PFS was 73.2 percent in the standard treatment arm and 68.2 percent in the PET-driven treatment arm. Sixty-two patients with a positive PET2 converted to a negative PET scan after four cycles of escBEACOPP (PET4); these patients had an intermediate five-year PFS compared to PET2/PET4-negative patients and PET2/PET4-positive patients (75.4% vs. 90.9% vs. 46.5%, respectively). There were 13 patients with a negative PET2 who relapsed with a positive PET4 scan; these were equally distributed between the two treatment arms. In contrast, on the RATHL study, 84 percent of patients had a negative PET2 after two cycles of ABVD, demonstrating that starting with ABVD yields a similarly favorable early metabolic response. In the PET2-negative group, the three-year PFS was 85.7 percent, and was 84.4 percent in the groups that continued ABVD and AVD, respectively, which was comparable to AHL2011. Among the patients with a positive PET2, 74.4 percent converted to a negative PET scan subsequently. The three-year PFS for the PET2-positive group was 67.5 percent, again comparable to the 73.2 percent five-year PFS in the PET2-positive group on AHL2011.

OS on the RATHL trial was quite good at 87.8 percent at three years. Five-year OS was similarly good at greater than 90 percent for all patients on AHL2011, regardless of PET2 and PET4 results. This becomes important when one considers the differential safety and toxicity of these therapies. Historically, there are more high-grade cytopenias with escBEACOPP compared with ABVD chemotherapy, and this is associated with increased incidence of infection. Furthermore, concerns about the differential impact on fertility and risk of secondary malignancies with escBEACOPP and ABVD are important considerations in this disease that largely affects adolescents and young adults. In this study, the incidence of infection was 22 percent in the standard treatment group compared with 11 percent in the PET-driven treatment group, demonstrating a positive impact of limiting the exposure to escBEACOPP on the risk of infection. There were eight deaths on this study, six in the standard treatment group due to septic shock (n=2), pneumonitis (n=2), heart failure (n=1), and acute leukemia (n=1), compared with two in the PET-driven treatment group due to septic shock after the first cycle of escBEACOPP (n=1) and acute leukemia (n=1). This is significant when, regardless of response to therapy, OS is estimated to be greater than 90 percent for these patients; therefore, consideration and avoidance of treatment-related deaths must be prioritized when comparing treatment regimens. The authors do report on rates of secondary malignancies (n=15, 2%); however, follow up for this endpoint is relatively short. They also report on the rate of post-treatment pregnancies; however, the denominator reflecting the number of patients who tried to become pregnant but were unable is unknown.