Virtual Library

Start Your Search

Aaron Lisberg



Author of

  • +

    P1.04 - Immunooncology (Not CME Accredited Session) (ID 936)

    • Event: WCLC 2018
    • Type: Poster Viewing in the Exhibit Hall
    • Track:
    • Presentations: 1
    • Moderators:
    • Coordinates: 9/24/2018, 16:45 - 18:00, Exhibit Hall
    • +

      P1.04-14 - HLA B44 Supertype Associated with Less Favorable Neoantigen Binding in Non-Small Cell Lung Cancer Treated with Immunotherapy (ID 12285)

      16:45 - 18:00  |  Author(s): Aaron Lisberg

      • Abstract
      • Slides

      Background

      Human leukocyte antigen (HLA) supertypes may influence immunotherapy efficacy, particularly HLA class 1 B44 supertype (B44), which is found in 35-55% of the population irrespective of race (Sidney, BMC Immunol). In melanoma patients treated with immune checkpoint inhibitors, presence of B44 correlated with improved survival (Chowell, Science), but in a cohort of 58 non-small cell lung cancer (NSCLC) patients treated with single-agent pembrolizumab, B44 was associated with poorer outcomes (Lu, ASCO 2018). Given B44’s small electropositive binding pocket, it was hypothesized that the transversions that predominate in NSCLC result in more positive tumor variant amino acids (vAAs) and that these neoantigens would have decreased binding affinity and/or HLA B44:peptide stability.

      a9ded1e5ce5d75814730bb4caaf49419 Method

      58 advanced NSCLC patients treated with pembrolizumab had germline and tumor multiplexed paired-end Illumina WES performed. HLA typing used BWA-ALN and Athlates software; supertype was determined by 2008 criteria (Sidney, BMC Immunol). Six subjects with at least one strong HLA B44 supertype allele had nonsynonymous coding mutations identified with Genome Toolkit Analysis (GATK) best practices utilizing the Hg38 genome reference. PvacSeq software used NetMHC algorithms to identify tumor neoantigens 9 base pairs in length matched to their corresponding HLA B44 allele (B44-specific neoantigens, BSNs). Missense BSNs were classified by transition (Ti) transversion (Tv) ratios and vAA charge change. TiTv was compared with matched pairs analysis. Predicted NetMHC IC50 binding affinities were compared with student’s t-tests. All statistical analyses were performed with SAS JMP, Version 13.0 (Cary, NC).

      4c3880bb027f159e801041b1021e88e8 Result

      Among the 6 subjects, 3073 BSNs were identified; 1090 were unique. There were no common BSNs among subjects. Subject tumor TiTv median was 1.58 (95% CI 1.09-1.94), mean difference compared to germline was -0.62 (95% CI -1.11 to -0.12, p=0.02). BSNs with vAAs that changed charge represented 13.5% of all BSNs. Positive vAA charge changes were as expected based on theoretical distribution (12.5% Tv vs 6.3% Ti, p=0.02). In aggregate, there were 205 BSNs with negative charge change (-BSN) and 204 with positive charge change (+BSN). The anticipated HLA B44 binding affinity was lower for +BSN, with median NetMHC IC50 binding 176.2 (95% CI 178.0-217.9) vs 258.6 (95% CI 216.9-257.7) for -BSN, p<0.01.

      8eea62084ca7e541d918e823422bd82e Conclusion

      A potential mechanism for decreased survival in B44 NSCLC subjects treated with immunotherapy is unfavorable neoantigen binding related to increased transversions leading to tumor vAAs with positive charge changes and poorer HLA B44 binding. All subjects from this cohort will be evaluated for BSNs 8-12 base-pairs long to confirm these findings.

      6f8b794f3246b0c1e1780bb4d4d5dc53

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.

  • +

    P2.09 - Pathology (Not CME Accredited Session) (ID 958)

    • Event: WCLC 2018
    • Type: Poster Viewing in the Exhibit Hall
    • Track:
    • Presentations: 1
    • Moderators:
    • Coordinates: 9/25/2018, 16:45 - 18:00, Exhibit Hall
    • +

      P2.09-05 - Evaluation of PD-L1-Stained Tumor Cells via the 22C3 and SP-142 Antibodies in Cohort of Patients Treated on KEYNOTE-001 (ID 12976)

      16:45 - 18:00  |  Presenting Author(s): Aaron Lisberg

      • Abstract

      Background

      Four PD-L1 antibodies have been utilized in NSCLC clinical trials, with an analytical comparison demonstrating a high level of concordance between the percentage of PD-L1–stained tumor cells with three of these antibodies (22C3, 22C3, 28-8, and SP263), but not a fourth, SP-142 (Hirsch et al, JTO 2017). This finding led us to evaluate the relationship between the percentage of PD-L1–stained tumor cells with 22C3 and SP-142, as well as the association between the PD-L1 levels identified by each antibody and clinical outcomes in 28 NSCLC patients treated on KEYNOTE-001.

      a9ded1e5ce5d75814730bb4caaf49419 Method

      We performed a retrospective analysis of 28 NSCLC patients treated with pembrolizumab on the KEYNOTE-001 trial at UCLA (23pts) or MSKCC (5pts) with data cut-off 12/2017. Patients had PD-L1–stained tumor cell levels assessed by both the 22C3 antibody (Dako), via central evaluation as previously described (Garon et al, NEJM 2015) and the SP-142 antibody (Spring Bioscience) at UCLA in accordance with established methods (Zaretsky et al, NEJM 2016). Survival curves for PFS and OS were estimated using the Kaplan-Meier method and formally compared between groups using the log-rank test. The association between PD-L1–stained tumor cell levels identified by the two antibodies was assessed using the Pearson correlation coefficient.

      4c3880bb027f159e801041b1021e88e8 Result

      In 61% (17/28) of patients, PD-L1 levels were grouped similarly (either <1%, 1-49%, or >50%) by both antibodies. Specifically, compared to 22C3 staining, SP-142 led to the same grouping for 63% (5/8) pts with >50% staining, 85% (11/13) pts with 1-49% staining, and 14% (1/7) pts w <1% staining. Evaluating the relationship between PD-L1 grouping and clinical outcomes via the SP-142 antibody revealed improved PFS and OS in pts with higher PD-L1 expression levels, while the 22C3 antibody predicted for improved PFS in these patients, but not improved OS [SP142 (PFS,OS): (p=0.0039, p=0.0425)][22C3 (PFS,OS): p=0.0121, p=0.1222). The PD-L1 results from the SP-142 and 22C3 antibodies were strongly associated (r =0.58, p=0.001).

      8eea62084ca7e541d918e823422bd82e Conclusion

      The PD-L1–stained tumor cell levels in the majority of patients evaluated were similarly grouped into one of three categories (<1%, 1-49%, or >50%) by both 22C3 and SP142. This analysis is limited by small patient number, but suggests that the number of PD-L1–stained tumor cells identified by each antibody is similar and a higher PD-L1 level identified by either antibody predicts for improved clinical outcomes with pembrolizumab.

      6f8b794f3246b0c1e1780bb4d4d5dc53