FDG-PET SUV Correlates with Expression of Genes Reflecting Proliferation, Metabolism, and Oncogene Activity in Mantle Cell Lymphoma (MCL),

Mantle cell lymphoma (MCL) is a biologically heterogeneous disease with marked variation in clinical behavior. Tumor proliferation on gene expression profiling (GEP) is prognostic in MCL but is not available to the clinician. The % of Ki-67 expressing cells correlates with proliferation, but the test suffers inter-observer variability. We previously reported that SUV on FDG PET correlates to the tumor proliferation score (PS), the % Ki67 positive cells, and serum LDH levels in untreated MCL (Roschewski ASH 2010). Here, we expand this investigation to a larger cohort and present results of gene expression analysis of MCL lymph node biopsies in relation to FDG uptake.

Untreated MCL patients underwent baseline PET/CT followed by a nodal biopsy. Analysis of the PET/CT scans was done by two experts in nuclear medicine who were blinded to the outcome of the patients and pathology findings. SUVnode and SUVmax normalized to body weight were recorded. For MCL patients, the expression of 18 proliferation signature genes were averaged to generate a PS as previously described (Rosenwald Cancer Cell 2003). The %Ki67 staining was calculated using computer-based image analysis software and recorded as the number of positive cells/total number of cells. Serum LDH on the day of treatment was recorded. Gene expression between tumor with high (> 5 SUVnode) and low (< 5 SUVnode) FDG uptake was compared. Gene expression and SUV as a continuous variable were also analyzed by Spearman correlation.

41 MCL patients were evaluated between September 2005 and March 2011. Median age was 58 years (41–73) and 78% of patients were male. All patients were stage IV and MIPI risk groups included 21 low, 15 intermediate and 5 high. PET/CT scans were available in 39, biopsies in 28, and adequate RNA for GEP in 27. The median SUVnode was 4.94 (1.81–12.89) and the median SUVmax 6.40 (2.22–17.5). SUVnode and SUVmax had virtually identical correlations to PS: r=0.45 (p < 0.016) and r=0.46 (p < 0.015). % Ki67 in MCL lymph node biopsies correlated with both SUVnode, r=0.58 (p < 0.002) and PS, r=0.59 (p < 0.001). LDH correlated with PS, r=0.55 (p < 0.002) but less so with SUVmax, r=0.35 (p < 0.03). We sought to determine which cellular processes correlate best with FDG avidity. We first separated MCL biopsies into two groups using the median SUV node. 88 probesets representing 79 genes were differentially expressed with a fold change ≥ 2. Of 32 upregulated genes, 14 were associated with cell proliferation and in Ingenuity Pathway analysis (IPA) the top network was cellular growth and proliferation. Next, we used Spearman correlation to better represent the full dynamic range of the FDG-PET measurement. 69 probesets genes were identified (FDR < 0.2) representing 61 genes, 47 of which were positively correlated with SUV PET. A key network function of these genes by IPA was carbohydrate and lipid metabolism. Among the genes with the highest correlation with SUV PET (rho > 0.72, p < 0.0001) were two members of the RAS Oncogene family: KRAS and RAB11A. There was also a remarkable representation of genes involved in protein homeostasis including PSMB5 that encodes the 26S proteasome subunit carrying the chymotrypsin like activity, Hsp40-family chaperones, and an ubiquitin conjugating enzyme (UBE2S). Finally, PRMT5 a protein arginine methyltransferase that has been shown to cooperate with cyclin D1 in promoting tumor growth also showed a high correlation with FDG-PET (rho = 0.65, p = 0.002).

High FDG PET SUVs are typically seen in the quartile of patients with the highest tumor proliferation. Across all patients increasing SUV values correlate well with increased metabolism that intriguingly also includes the ubiquitin-proteasome system and PSMB5, the target of bortezomib. Thus, FDG PET might have some correlation with response to bortezomib treatment and should be further studied in this context. FDG PET integrates measures of proliferation and cellular metabolism and reflects disease specific activity of oncogenic pathways.

No relevant conflicts of interest to declare.