Page 16 - Haematologica May 2022

Editorials
Stem cell transplant for lymphoma - never too late?
Amanda F. Cashen and Nancy L. Bartlett
Washington University School of Medicine and Siteman Cancer Center, St. Louis, MO, USA E-mail: AMANDA CASHEN - acashen@wustl.edu
doi:10.3324/haematol.2021.280519
In this issue, using the Center for International Blood and Marrow Transplant Research (CIBMTR) registry,
1
Mei et al. provide a retrospective outcomes analysis of
285 patients with relapsed or refractory (R/R) diffuse large B-cell lymphoma (DLBCL) who underwent autolo- gous hematopoietic cell transplantation (autoHCT) after achieving complete (CR) or partial response (PR) to their third or higher line of chemotherapy. Sixty three percent of patients had primary refractory disease or relapsed within 1 year of initial diagnosis. For perspective, during the same interval 577 patients with R/R DLBCL in the reg- istry underwent autoHCT after second line therapy. Patients in the registry with detailed data including prior lines of therapy comprise approximately 8-10% of all patients in the CIBMTR registry, which in other analyses have been representative of the entire dataset (personal communication with corresponding author). Mei et al. report 5-year overall survival (OS) and progression-free survival (PFS) for the study population of 51% and 38%. As expected, patients in CR at the time of transplant had better OS and PFS, but a substantial minority of patients with PR also benefited from autoHCT, with a 5-year PFS of 34%. Relapses in patients with a PR occurred early, with a near plateau in the relapse/progression curve after 1 year. In contrast, patients in CR had a significantly lower 1-year risk of relapse but a continuous risk of relapse for at least 4 years after autoHCT, resulting in a non-significant difference in the 5-year risk of relapse between the two groups (CR 45% vs. PR 54%, P=0.14).
Importantly, while nearly half of all patients were with- out relapse at 5 years, the 5-year PFS of 38% reflects a 5- year non-relapse mortality of 12%. After lymphoma, the second most common cause of death was second malig- nancy, in 10% of the study cohort during the entire fol- low-up period.
How does this retrospective registry analysis inform our decision whether to pursue autoHCT for a given patient with R/R DLBCL? Primarily, it confirms that patients with chemosensitive disease can achieve long- term remissions after autoHCT, including those who have early failure of frontline immunochemotherapy, those who achieve less than a CR to salvage chemother- apy, and those who have been treated with more than one line of salvage chemotherapy. Consistent with prior reports, Mei et al. found that, while patients in CR have a better PFS than those in PR, approximately a third of PR patients remain alive and progression-free after autoHCT. Their results mirror another CIBMTR analysis2 that reported 5-year PFS of 41% for patients with R/R DLBCL who had PET-positive PR prior to autoHCT. Even patients with primary refractory disease or early relapse have a 40-50% chance of durable PFS after autoHCT, if they respond to salvage therapy.3,4,5
Unfortunately, many patients with R/R DLBCL do not achieve a CR or PR to one or more lines of salvage
chemotherapy. In the prospective, randomized CORAL study,4 63% of all patients, 51% of patients previously treated with rituximab, and 46% of patients with primary refractory disease or early relapse achieved CR or PR to R- ICE or R-DHAP. Of the patients who did not respond to the study-assigned salvage regimen, only 32% responded to subsequent therapy and underwent a stem cell trans- plant.6 In two randomized studies of anti-CD19 chimeric antigen receptor T-cell therapy (CAR-T) versus standard of care in patients with early failure of frontline immunochemotherapy, the response rates to standard second line therapies were 50%7 and 43%.8 Clearly, inad- equate response to second- or third-line therapy is a sig- nificant barrier to use of autoHCT in many patients with R/R DLBCL. New agents with encouraging activity in R/R DLBCL, including bispecific T-cell engagers9 and anti- body-drug conjugates,10 may improve the number of patients who achieve PR or CR to second- or third-line therapy making autoHCT an option for more patients. However, whether response to novel therapies with unique mechanisms of action will predict response to autoHCT is unknown and will be extremely challenging to study.
While CAR-T is the obvious choice for patients who do not respond to standard second- or third-line regimens, the choice between autoHCT and CAR-T can be difficult for patients with chemosensitive R/R DLBCL. The poten- tial approval of CAR-T in the second line will increase the challenge of weighing the relative benefits of autoHCT versus CAR-T for a given patient. Now more than 4 years after approval of the first CAR-T for R/R DLBCL, signifi- cant hurdles remain, including complicated and delayed insurance approvals, limitations on manufacturing slots, and the potential for progressive disease and worsening performance status during the “waiting period”. Rigorous cost-benefit analysis of CAR-T versus autoHCT for patients with chemosensitive disease in second and third line would be informative. Additionally, late effects of CAR-T and autoHCT need to be compared and consid- ered in the treatment decision. Post-CAR-T complica- tions, especially prolonged cytopenias and infections, can be more challenging than post-autoHCT in a significant minority of patients, and, as Mei’s study demonstrates, second malignancies are a concerning cause of non- relapse mortality post-autoHCT.
Although we have ample data on the tolerability and efficacy of CAR-T following relapse after autoHCT, there is no data on autoHCT following CAR-T cell failure. In the current use of CAR-T for chemorefractory disease or relapse after autoHCT, patients would not be candidates for autoHCT after CAR-T. However, if approvals are forthcoming for CAR-T as second-line therapy, under- standing the feasibility and efficacy of autoHCT after CAR-T would help inform the decision of sequencing these therapies in patients responding to second- or third-
1024
haematologica | 2022; 107(5)