ALBERT

Description: Background: Mantle cell lymphoma (MCL) is an aggressive NHL subtype that remains incurable in most cases with conventional chemotherapy. However, individual patient survival can vary considerably, and studies have identified potential markers of prognostic subgroups, including Ki-67 and the cyclin-dependent kinase (CDK) inhibitor p27. Genomic studies further highlight tumor proliferation status as the major driver of differential prognosis in MCL (Rosenwald et al, Cancer Cell 2003). Such studies were conducted prior to the introduction of bortezomib, a first-in-class proteasome inhibitor that was recently approved for the treatment of relapsed MCL. Plausible hypotheses for the activity of this novel agent in MCL include inhibition of NF-κB activity and disruption of the cell cycle via stabilization of CDK inhibitors such as p27. In order to test these hypotheses directly and also examine how this new agent impacts prognostic subgroups of MCL, the PINNACLE trial, the largest phase 2 study of bortezomib in MCL, specified the collection of archived tumor specimens for immunohistochemical (IHC) analyses. In this trial 155 relapsed/refractory MCL patients received single-agent bortezomib at a dose of 1.3 mg/m2 on days 1, 4, 8, and 11 of 21-day cycles. A 33% overall response rate was observed (8% CR/CRu), with a median time to progression (TTP) of 6.2 months and a responder TTP of 10.6 months (Fisher et al, J Clin Oncol 2006). Methods: A series of unstained tumor slides from the diagnostic tumor specimens were collected from a subset of 70 evaluable patients. This subset formed the basis of the current IHC analysis and the clinical parameters of this subset appeared similar to those of the study as a whole. Several proteins were assessed via IHC; these included p27, SKP2, cyclin D1, and Ki-67, as well as candidate markers correlated with bortezomib efficacy in a genomic study of multiple myeloma, including NF-κB and proteasome subunits (Mulligan et al, Blood 2007). Association with response was evaluated using Fisher’s Exact Test. A Cox proportional hazard linear model adjusted for clinical covariates was used to determine the effect of these parameters on TTP. Results: In this subset analysis no protein markers were significantly associated with response to therapy. However, patients with 〉50% Ki-67-positive tumor cells were more likely to progress than patients with