Machine learning‐based classification of diffuse large B‐cell lymphoma patients by eight gene expression profiles

Abstract

Gene expression profiling (GEP) had divided the diffuse large B‐cell lymphoma (DLBCL) into molecular subgroups: germinal center B‐cell like (GCB), activated B‐cell like (ABC), and unclassified (UC) subtype. However, this classification with prognostic significance was not applied into clinical practice since there were more than 1000 genes to detect and interpreting was difficult. To classify cancer samples validly, eight significant genes (MYBL1, LMO2, BCL6, MME, IRF4, NFKBIZ, PDE4B, and SLA) were selected in 414 patients treated with CHOP/R‐CHOP chemotherapy from Gene Expression Omnibus (GEO) data sets. Cutoffs for each gene were obtained using receiver–operating characteristic curves (ROC) new model based on the support vector machine (SVM) estimated the probability of membership into one of two subgroups: GCB and Non‐GCB (ABC and UC). Furtherly, multivariate analysis validated the model in another two cohorts including 855 cases in all. As a result, patients in the training and validated cohorts were stratified into two subgroups with 94.0%, 91.0%, and 94.4% concordance with GEP, respectively. Patients with Non‐GCB subtype had significantly poorer outcomes than that with GCB subtype, which agreed with the prognostic power of GEP classification. Moreover, the similar prognosis received in the low (0–2) and high (3–5) IPI scores group demonstrated that the new model was independent of IPI as well as GEP method. In conclusion, our new model could stratify DLBCL patients with CHOP/R‐CHOP regimen matching GEP subtypes effectively.

Cell‐of‐origin determined by both gene expression profiling and immunohistochemistry is the strongest predictor of survival in patients with diffuse large B‐cell lymphoma

“…The UC cases in our study had an outcome comparable to patients classified as GCB. This is in contrast to other studies showing that the UC cases are more likely to be grouped together with the ABC subtype with an inferior patient outcome, although others have shown results similar to ours . The UC patients in our cohort did not differ significantly from GCB patients regarding clinical characteristics.…”

Phosphodiesterase 4 inhibitors have wide-ranging activity in B-cell malignancies

“…20,21 The data discussed in the previous paragraphs showed that this second messenger is also an important negative regulator of the BCR, a physiologic safeguard that is lost in B-cell malignancies that display elevated PDE4 expression and activity. 5,6,13,14,16 PDE4 and B-cell lymphoma angiogenesis…”

“…5 The association between high PDE4B expression and poor DLBCL outcome was subsequently confirmed in larger independent series. 6,13,14 To advance these early observations, preclinical models were used to show that genetic or pharmacological inhibition of PDE4 results in growth suppression and apoptosis in DLBCL. 6 Mechanistically, PDE4 inhibition resulted in elevation of intracellular cyclic-AMP levels and suppression of PI3K and AKT activity.…”

Safety and Pharmacodynamics of the PDE4 Inhibitor Roflumilast in Advanced B-cell Malignancies

“…This observation, which has been confirmed in independent tumor series(16-18), is biologically relevant because PDE4 functions by hydrolyzing and terminating the activity of the second messenger cAMP, which has long been known to deliver potent growth inhibitory signals towards immune cells(1). Thus, malignant lymphocytes with elevated expression/activity of PDE4B lose the physiologic growth inhibitory effects of cAMP, which we propose are restored by PDE4 inhibitors.…”

“…Indeed, we and others showed that PDE4B was expressed at significantly higher levels in biopsies from fatal diffuse large B cell lymphoma (DLBCL) than in tumors from patients that survived their disease(14-18). Further, we used in vitro and in vivo preclinical models to meticulously charter how cAMP may suppress the growth of DLBCL cells, and how PDE4 expression/activity would abrogate these anti-lymphoma effects(17, 19, 20).…”