Ratios of T-cell immune effectors and checkpoint molecules as prognostic biomarkers in diffuse large B-cell lymphoma: a population-based study

Abstract

Background: Risk-stratification of diffuse large B-cell lymphoma (DLBCL) requires identification of patients with disease that is not cured, despite initial treatment with R-CHOP. The prognostic importance of the revised International Prognostic Index (R-IPI) and cell of origin of the malignant B cell are established in DLBCL. We aimed to develop a novel, easily applicable, tissue-based prognostic biomarker based on quantification of the tumour microenvironment that is independent of and additive to the R-IPI and cell of origin.

Methods: We performed digital hybridisation on the NanoString platform to assess the relation between immune effector and inhibitory (checkpoint) genes in 252 formalin-fixed, paraffin-embedded DLBCL tissue specimens obtained from patients treated with R-CHOP. We used a tree-based survival model to quantify net antitumoral immunity (using ratios of immune effector to checkpoint genes) and to generate a cutoff as an outcome predictor in 158 of the 252 patients. We validated this model in tissue (n=233) and blood (n=140) samples from two independent cohorts treated with R-CHOP.

Findings: T-cell and NK-cell immune effector molecule expression correlated with tumour-associated macrophage and PD-1/PD-L1 axis markers, consistent with malignant B cells triggering a dynamic checkpoint response to adapt to and evade immune surveillance. The ratio of CD4*CD8 to (CD163:CD68[M2])*PD-L1 was better able to stratify overall survival than was any one immune marker or combination, distinguishing groups with disparate 4-year overall survival. 94 (59%) of 158 patients had a score above the cutoff and 4-year overall survival of 92·1% (95% CI 82·9-96·7), and the remaining 64 (41%) patients had a score below the cutoff and 4-year overall survival of 47·0% (32·8-60·5; hazard ratio [HR] 8·3, 95% CI 4·3-17·3; p<0·0001). The CD4*CD8:M2*PD-L1 immune ratio was independent of and added to the R-IPI and cell of origin. Tissue findings in the independent tissue cohort accorded with those in our initial tissue cohort. 139 (60%) of 233 patients had a score above the cutoff and 4-year overall survival of 75·6% (95% CI 64·6-83·6), with the remaining 94 (40%) patients having a score below the cutoff (63·5% [52·5-72·7]; HR 1·9, 95% CI 1·1-3·3; p=0·0067).

Interpretation: Ratios of immune effectors to checkpoints augment the cell of origin and R-IPI in DLBCL and are applicable to paraffin-embedded biopsy specimens. These findings might have potential implications for selection of patients for checkpoint blockade within clinical trials.

Funding: Leukaemia Foundation of Queensland, Kasey-Anne Oklobdzijato Memorial Fund, the Australasian Leukaemia and Lymphoma Group (Malcolm Broomhead Bequest), the Australian Cancer Research Foundation, and the Cancer Council of Queensland.