Contributions
Abstract: PB1447
Type: Publication Only
Session title: Aggressive Non-Hodgkin lymphoma - Clinical
Background
Diffuse large B-cell lymphoma (DLBCL) is characterized by molecular biological heterogeneity and variability of the clinical course, which casts doubt on the use of a single standard of patient care and promotes the search for risk-adapted approaches in the first line of therapy for patients with DLBCL. The search for new predictors of effectiveness of induction therapy and survival are of significant scientific and practical importance
Aims
The research is aimed at the assessment of prognostic value of PD-L1 protein expression in a combined model for predicting the course of DLBCL in patients receiving induction therapy according to the RCHOP therapy
Methods
A retrospective analysis of the data of 85 DLBCL patients who were treated at the Kirov Scientific Re-search Institute of Hematology and Blood Transfusion was carried out. The median age was 59 (18-80) years. All the patients received no less than 4-6 courses of RСНОР chemotherapy. The median follow-up was 17 (1-168) months. The optimal cut-off threshold for assessing the proportion of cells expressing the PD-L1 protein was determined using the CART (Classification and Regression Tree) machine learning method
Results
The patients were divided into three groups according to the immunohistochemical subtype and IPI risk. In group 1 with GCB type tumors and any risk except high 21 (84%) and 4 (16%) patients, respectively, showed low or higher PD-L1 expression. Low level of PDL1 provided two-year PFS in 76% of all the cases, while higher level of PDL1 – in 100% (р=0,628). Two-year OS was 100% in both groups.
In Group 2 with nonGCB subtype and any risk except high 27 (67,5%) and 13 (32,5%) patients had low and high PD-L1 expression, respectively. Two-year PFS didn’t differ much: 46% - in high PDL1 expression level and 49% - in low PD-L1 expression level. At the same time the patients with high PD-L1 expression had shorter OS than those with low expression: 52% and 87%, respectively (р=0,049).
In Group 3 with high IPI risk 9 (45%) and 11 (55%) patients had high and low PD-L1 expression level, respectively. In cases of high PD-L1 two-year OS was 0%, with 46% in cases of low PD-L1 (р=0,002). Two-year OS was 66% in the patients with low PD-L1 expression level (p=0,008).
Conclusion
High PD-L1 expression level in combination with high IPI and non-GCB tumor subtype is associated with low OS and PFS in DLBCL patients. This is probably due to the low efficacy of induction therapy according to the RCHOP regimen in patients with high IPI risk. This allows us to consider the assessment of the proportion of PD-L1 expressing cells as an additional criterion for stratification of patients into risk groups for predicting the response to treatment and a differentiated approach to the choice of therapeutic tactics at the onset of the pathological process
Keyword(s): Diffuse large B cell lymphoma, Survival
Abstract: PB1447
Type: Publication Only
Session title: Aggressive Non-Hodgkin lymphoma - Clinical
Background
Diffuse large B-cell lymphoma (DLBCL) is characterized by molecular biological heterogeneity and variability of the clinical course, which casts doubt on the use of a single standard of patient care and promotes the search for risk-adapted approaches in the first line of therapy for patients with DLBCL. The search for new predictors of effectiveness of induction therapy and survival are of significant scientific and practical importance
Aims
The research is aimed at the assessment of prognostic value of PD-L1 protein expression in a combined model for predicting the course of DLBCL in patients receiving induction therapy according to the RCHOP therapy
Methods
A retrospective analysis of the data of 85 DLBCL patients who were treated at the Kirov Scientific Re-search Institute of Hematology and Blood Transfusion was carried out. The median age was 59 (18-80) years. All the patients received no less than 4-6 courses of RСНОР chemotherapy. The median follow-up was 17 (1-168) months. The optimal cut-off threshold for assessing the proportion of cells expressing the PD-L1 protein was determined using the CART (Classification and Regression Tree) machine learning method
Results
The patients were divided into three groups according to the immunohistochemical subtype and IPI risk. In group 1 with GCB type tumors and any risk except high 21 (84%) and 4 (16%) patients, respectively, showed low or higher PD-L1 expression. Low level of PDL1 provided two-year PFS in 76% of all the cases, while higher level of PDL1 – in 100% (р=0,628). Two-year OS was 100% in both groups.
In Group 2 with nonGCB subtype and any risk except high 27 (67,5%) and 13 (32,5%) patients had low and high PD-L1 expression, respectively. Two-year PFS didn’t differ much: 46% - in high PDL1 expression level and 49% - in low PD-L1 expression level. At the same time the patients with high PD-L1 expression had shorter OS than those with low expression: 52% and 87%, respectively (р=0,049).
In Group 3 with high IPI risk 9 (45%) and 11 (55%) patients had high and low PD-L1 expression level, respectively. In cases of high PD-L1 two-year OS was 0%, with 46% in cases of low PD-L1 (р=0,002). Two-year OS was 66% in the patients with low PD-L1 expression level (p=0,008).
Conclusion
High PD-L1 expression level in combination with high IPI and non-GCB tumor subtype is associated with low OS and PFS in DLBCL patients. This is probably due to the low efficacy of induction therapy according to the RCHOP regimen in patients with high IPI risk. This allows us to consider the assessment of the proportion of PD-L1 expressing cells as an additional criterion for stratification of patients into risk groups for predicting the response to treatment and a differentiated approach to the choice of therapeutic tactics at the onset of the pathological process
Keyword(s): Diffuse large B cell lymphoma, Survival