TUMOR GENE EXPRESSION PROFILES ASSOCIATED WITH SINGLE-AGENT COPANLISIB ACTIVITY IN HEAVILY PRETREATED PATIENTS WITH INDOLENT AND AGGRESSIVE NON-HODGKIN LYMPHOMA (NHL)
(Abstract release date: 05/19/16)
EHA Library. Liu L. 06/09/16; 132929; E1380

Dr. Li Liu
Contributions
Contributions
Abstract
Abstract: E1380
Type: Eposter Presentation
Background
Copanlisib, a novel pan-class I PI3K inhibitor with predominant activity against α and δ isoforms, has shown promising single agent anti-tumor activity in a phase 2 study (NCT01660451, part A) in heavily pretreated patients with indolent and aggressive lymphoma.
Aims
Tumor gene expression profiling was used as a discovery tool to identify genes or pathways associated with copanlisib treatment outcomes and provide potential predictive markers and rationale for novel combinations.
Methods
Single gene multivariate approach and gene set enrichment analysis (GSEA) using RNA expression array data from archival tumors from patients enrolled in the phase 2 trial were performed to identify genes and pathways associated with copanlisib efficacy. A weighted gene expression score (WGS) reflecting overall expression level for each pathway was generated from logistic regression and Cox proportional hazards models to assess association with response status and progression-free survival (PFS), respectively.
Results
A total of 24 patients including both indolent and aggressive lymphoma were available for analysis. Patients had either complete response (n=3), partial response (n=5), stable disease (n=11), or progressive disease (n=5) as best clinical response; follicular lymphoma (FL, n=10), marginal zone lymphoma (n=2), mantle cell lymphoma (n=2), diffuse large B-cell lymphoma (DLBCL, n=5), transformed indolent lymphoma (n=2) or chronic lymphocytic leukemia (n=3). The GSEA gene sets ranked highest for correlation with response were those reflecting upregulated B-cell receptor (BCR) signaling and a PI3K signature (including both α and δ isoforms), with normalized enrichment scores of 1.92 and 1.62 and false discovery rates of 0.014 and 0.087, respectively. Response rates were increased in patients with high BCR and PI3K WGS compared to low (nominal p=0.06 and 0.07, AUC=0.81 and 0.75, respectively). PFS was longer in copanlisib-treated patients with high (above the median) BCR WGS (377 vs 62 days, HR=0.035, nominal p<0.0001) and in patients with high PI3K WGS (288 vs 104 days, HR=0.24, nominal p=0.02) compared to those with low. In contrast, upregulation of NFҡB, IL6/JAK/STAT3, stromal and inflammatory process gene sets and the individual genes NOP10, MT2A, and CSTB were potentially associated with lower likelihood of response to copanlisib.
Conclusion
Durable response to single-agent copanlisib is associated with tumors with PI3K pathway activation and/or strong BCR signaling. A strong NFҡB expression profile is associated with lower likelihood of response to single-agent copanlisib. The higher response rate in FL as compared to DLBCL is consistent with a predominant BCR/PI3K expression profile in FL, with low NFҡB activation. These results suggest that FL even in advanced stage and prior extensive treatment can respond to single-agent copanlisib. Efficacy in DLBCL may require combination therapy, adding agents that target NFҡB pathway to copanlisib.
Session topic: E-poster
Keyword(s): B cell lymphoma, Gene expression, NHL, PI3K
Type: Eposter Presentation
Background
Copanlisib, a novel pan-class I PI3K inhibitor with predominant activity against α and δ isoforms, has shown promising single agent anti-tumor activity in a phase 2 study (NCT01660451, part A) in heavily pretreated patients with indolent and aggressive lymphoma.
Aims
Tumor gene expression profiling was used as a discovery tool to identify genes or pathways associated with copanlisib treatment outcomes and provide potential predictive markers and rationale for novel combinations.
Methods
Single gene multivariate approach and gene set enrichment analysis (GSEA) using RNA expression array data from archival tumors from patients enrolled in the phase 2 trial were performed to identify genes and pathways associated with copanlisib efficacy. A weighted gene expression score (WGS) reflecting overall expression level for each pathway was generated from logistic regression and Cox proportional hazards models to assess association with response status and progression-free survival (PFS), respectively.
Results
A total of 24 patients including both indolent and aggressive lymphoma were available for analysis. Patients had either complete response (n=3), partial response (n=5), stable disease (n=11), or progressive disease (n=5) as best clinical response; follicular lymphoma (FL, n=10), marginal zone lymphoma (n=2), mantle cell lymphoma (n=2), diffuse large B-cell lymphoma (DLBCL, n=5), transformed indolent lymphoma (n=2) or chronic lymphocytic leukemia (n=3). The GSEA gene sets ranked highest for correlation with response were those reflecting upregulated B-cell receptor (BCR) signaling and a PI3K signature (including both α and δ isoforms), with normalized enrichment scores of 1.92 and 1.62 and false discovery rates of 0.014 and 0.087, respectively. Response rates were increased in patients with high BCR and PI3K WGS compared to low (nominal p=0.06 and 0.07, AUC=0.81 and 0.75, respectively). PFS was longer in copanlisib-treated patients with high (above the median) BCR WGS (377 vs 62 days, HR=0.035, nominal p<0.0001) and in patients with high PI3K WGS (288 vs 104 days, HR=0.24, nominal p=0.02) compared to those with low. In contrast, upregulation of NFҡB, IL6/JAK/STAT3, stromal and inflammatory process gene sets and the individual genes NOP10, MT2A, and CSTB were potentially associated with lower likelihood of response to copanlisib.
Conclusion
Durable response to single-agent copanlisib is associated with tumors with PI3K pathway activation and/or strong BCR signaling. A strong NFҡB expression profile is associated with lower likelihood of response to single-agent copanlisib. The higher response rate in FL as compared to DLBCL is consistent with a predominant BCR/PI3K expression profile in FL, with low NFҡB activation. These results suggest that FL even in advanced stage and prior extensive treatment can respond to single-agent copanlisib. Efficacy in DLBCL may require combination therapy, adding agents that target NFҡB pathway to copanlisib.
Session topic: E-poster
Keyword(s): B cell lymphoma, Gene expression, NHL, PI3K
Abstract: E1380
Type: Eposter Presentation
Background
Copanlisib, a novel pan-class I PI3K inhibitor with predominant activity against α and δ isoforms, has shown promising single agent anti-tumor activity in a phase 2 study (NCT01660451, part A) in heavily pretreated patients with indolent and aggressive lymphoma.
Aims
Tumor gene expression profiling was used as a discovery tool to identify genes or pathways associated with copanlisib treatment outcomes and provide potential predictive markers and rationale for novel combinations.
Methods
Single gene multivariate approach and gene set enrichment analysis (GSEA) using RNA expression array data from archival tumors from patients enrolled in the phase 2 trial were performed to identify genes and pathways associated with copanlisib efficacy. A weighted gene expression score (WGS) reflecting overall expression level for each pathway was generated from logistic regression and Cox proportional hazards models to assess association with response status and progression-free survival (PFS), respectively.
Results
A total of 24 patients including both indolent and aggressive lymphoma were available for analysis. Patients had either complete response (n=3), partial response (n=5), stable disease (n=11), or progressive disease (n=5) as best clinical response; follicular lymphoma (FL, n=10), marginal zone lymphoma (n=2), mantle cell lymphoma (n=2), diffuse large B-cell lymphoma (DLBCL, n=5), transformed indolent lymphoma (n=2) or chronic lymphocytic leukemia (n=3). The GSEA gene sets ranked highest for correlation with response were those reflecting upregulated B-cell receptor (BCR) signaling and a PI3K signature (including both α and δ isoforms), with normalized enrichment scores of 1.92 and 1.62 and false discovery rates of 0.014 and 0.087, respectively. Response rates were increased in patients with high BCR and PI3K WGS compared to low (nominal p=0.06 and 0.07, AUC=0.81 and 0.75, respectively). PFS was longer in copanlisib-treated patients with high (above the median) BCR WGS (377 vs 62 days, HR=0.035, nominal p<0.0001) and in patients with high PI3K WGS (288 vs 104 days, HR=0.24, nominal p=0.02) compared to those with low. In contrast, upregulation of NFҡB, IL6/JAK/STAT3, stromal and inflammatory process gene sets and the individual genes NOP10, MT2A, and CSTB were potentially associated with lower likelihood of response to copanlisib.
Conclusion
Durable response to single-agent copanlisib is associated with tumors with PI3K pathway activation and/or strong BCR signaling. A strong NFҡB expression profile is associated with lower likelihood of response to single-agent copanlisib. The higher response rate in FL as compared to DLBCL is consistent with a predominant BCR/PI3K expression profile in FL, with low NFҡB activation. These results suggest that FL even in advanced stage and prior extensive treatment can respond to single-agent copanlisib. Efficacy in DLBCL may require combination therapy, adding agents that target NFҡB pathway to copanlisib.
Session topic: E-poster
Keyword(s): B cell lymphoma, Gene expression, NHL, PI3K
Type: Eposter Presentation
Background
Copanlisib, a novel pan-class I PI3K inhibitor with predominant activity against α and δ isoforms, has shown promising single agent anti-tumor activity in a phase 2 study (NCT01660451, part A) in heavily pretreated patients with indolent and aggressive lymphoma.
Aims
Tumor gene expression profiling was used as a discovery tool to identify genes or pathways associated with copanlisib treatment outcomes and provide potential predictive markers and rationale for novel combinations.
Methods
Single gene multivariate approach and gene set enrichment analysis (GSEA) using RNA expression array data from archival tumors from patients enrolled in the phase 2 trial were performed to identify genes and pathways associated with copanlisib efficacy. A weighted gene expression score (WGS) reflecting overall expression level for each pathway was generated from logistic regression and Cox proportional hazards models to assess association with response status and progression-free survival (PFS), respectively.
Results
A total of 24 patients including both indolent and aggressive lymphoma were available for analysis. Patients had either complete response (n=3), partial response (n=5), stable disease (n=11), or progressive disease (n=5) as best clinical response; follicular lymphoma (FL, n=10), marginal zone lymphoma (n=2), mantle cell lymphoma (n=2), diffuse large B-cell lymphoma (DLBCL, n=5), transformed indolent lymphoma (n=2) or chronic lymphocytic leukemia (n=3). The GSEA gene sets ranked highest for correlation with response were those reflecting upregulated B-cell receptor (BCR) signaling and a PI3K signature (including both α and δ isoforms), with normalized enrichment scores of 1.92 and 1.62 and false discovery rates of 0.014 and 0.087, respectively. Response rates were increased in patients with high BCR and PI3K WGS compared to low (nominal p=0.06 and 0.07, AUC=0.81 and 0.75, respectively). PFS was longer in copanlisib-treated patients with high (above the median) BCR WGS (377 vs 62 days, HR=0.035, nominal p<0.0001) and in patients with high PI3K WGS (288 vs 104 days, HR=0.24, nominal p=0.02) compared to those with low. In contrast, upregulation of NFҡB, IL6/JAK/STAT3, stromal and inflammatory process gene sets and the individual genes NOP10, MT2A, and CSTB were potentially associated with lower likelihood of response to copanlisib.
Conclusion
Durable response to single-agent copanlisib is associated with tumors with PI3K pathway activation and/or strong BCR signaling. A strong NFҡB expression profile is associated with lower likelihood of response to single-agent copanlisib. The higher response rate in FL as compared to DLBCL is consistent with a predominant BCR/PI3K expression profile in FL, with low NFҡB activation. These results suggest that FL even in advanced stage and prior extensive treatment can respond to single-agent copanlisib. Efficacy in DLBCL may require combination therapy, adding agents that target NFҡB pathway to copanlisib.
Session topic: E-poster
Keyword(s): B cell lymphoma, Gene expression, NHL, PI3K
{{ help_message }}
{{filter}}