PF308 PROGNOSTIC FACTORS AND IMMUNOHISTOCHEMICAL PROFILE IN PRIMARY REFRACTORY AND RECURRENT DIFFUSE LARGE B‐CELL LYMPHOMAS

• Published in: HemaSphere Vol. 3; no. S1; pp. 106 - 107
• Main Authors: Lampikoski, H., Drakos, E., Nummisalo, E., Tzoras, E., Kokaraki, G., Harrysson, S., Ekström Smedby, K., Rassidakis, G.
• Format: Journal Article
• Language: English
• Published: 01.06.2019
• ISSN: 2572-9241; 2572-9241
• DOI: 10.1097/01.HS9.0000559444.84225.ec

Background: One‐third of diffuse large B‐cell lymphomas (DLBCL) are either refractory to first‐line treatment or relapse after complete remission. Although various clinical and biological prognostic factors have been identified in DLBCL cohorts of general patient populations, their significance has not been fully elucidated specifically for primary refractory and recurrent DLBCL. Moreover, little is known about possible alterations of the immunohistochemical (IHC) profile of these difficult‐to‐treat cases at first relapse or at the first post‐treatment biopsy in primary refractory DLBCLs. Aims: This study aimed to retrospectively determine 1) whether a selection of known prognostic factors in DLBCL (age, International Prognostic Index [IPI], cell of origin phenotype, BCL2, BCL6, CD5, CD30, and p53) were significant in a selected cohort of refractory and recurrent cases, 2) whether the expression patterns of IHC markers differed from those reported in unselected de novo DLBCL cohorts, and 3) whether expression of IHC markers changed between the biopsy at diagnosis and at first relapse in relapsing DLBCL or first post‐treatment biopsy in primary refractory DLBCL. Methods: 122 patients with primary refractory or recurrent de novo DLBCL, not otherwise specified (NOS) were identified from the Swedish National Quality Register for Lymphomas. All patients were diagnosed with their primary tumor between 2007 and 2014 and treated at Stockholm hospitals. All patients were treated with a curative intention, most commonly with R‐CHOP immunochemotherapy. The data were collected from patient records and tissue microarrays constructed from archival tissue using duplicate tumor cores. Survival, from primary diagnosis for refractory cases and the first relapse for recurrent ones, was assessed using the Kaplan‐Meier method with log‐rank tests in univariate analysis and Cox proportional hazard models in multivariate analysis. Results: Age above 60 years, a high IPI score (R‐IPI>2, NCCN‐IPI>3) and high expression of MYC (40% and 70% cutoffs) conferred an adverse prognosis (p < 0.001 for age and IPI, 0.040 and 0.005 for MYC). When adjusted for age, R‐IPI, NCCN‐IPI and MYC (both cutoffs), significant or borderline p‐values were obtained for BCL6, suggesting a favorable prognostic significance (p = 0.054, 0.043, 0.089, 0.050, and 0.060, respectively). Cell of origin, BCL2, CD5, CD30 and p53 had no significant impact. Patients with recurrent disease showed a statistical trend toward worse outcome than those who were treatment‐refractory (p = 0.089). CD5 was expressed by the lymphoma cells in 25% (17/69) of cases with available data at primary presentation, more frequently than in three large unselected Western cohorts (7%, N = 2618: International DLBCL Rituximab‐CHOP Consortium Program, Lunenburg Trial, and RICOVER‐60). Between primary presentation and confirmed refractoriness or relapse, 42% of the cases with available data (n = 19) lost their expression of BCL6 while 16% gained BCL6‐positivity. The expression of other markers rarely changed. Summary/Conclusion: Our results suggest that age, IPI and MYC are adverse prognostic factors in primary refractory and recurrent DLBCL. Moreover, we showed for the first time that BCL6 was a favorable prognostic factor in multivariate analysis and that expression of BCL6 was frequently lost at first relapse or first post‐treatment tissue biopsy in refractory cases. The other IHC markers studied seem to lack prognostic significance in refractory and recurrent DLBCL. Aberrant expression of CD5 was more frequent compared to unselected DLBCL cohorts and may be of potential predictive value.