Incorporating pharmacodynamic, response and patient selection biomarkers. Paul Elvin PhD Chief Translational Science Officer Aptus Clinical

Transcription

1 Incorporating pharmacodynamic, response and patient selection biomarkers Paul Elvin PhD Chief Translational Science Officer Aptus Clinical 22

2 Oncology drug development Biomarkers key for: Strong hypothesis Dose/schedule selection Target engagement/ phenotypic change (eg Ki67) Patient selection Resistance pathways Right target Right dose/schedule Right patient Target selection/hypothesis (unmet need) Cell lines Xenograft tumour models Patient derived xenograft tumours 23

3 Biomarkers and target selection/validation Mutation Genomics Expression Amplification Deletion Disease linkage Analysis of clinical samples corresponding to target population Right Target Evidence from competitor compounds Biomarkers to measure target engagement/inhibition in preclinical models; ideally compatible with use in clinical studies 24

4 Pharmacodynamic biomarkers confirm osimertinib activity against T790M megfr Osimertinib inhibited EGFR phosphorylation in cell lines with EGFR sensitising mutations - PC-9 (exon 19del); H3255 (L858R); H1975 (L858R/T790M); PC9VanR (exon19del/t790m) and poor activty vs wtegfr Pharmacodynamic confirmation of target inhibition in vivo Pathway biomarkers inhibited in vivo - 6hr post dose in lung tumours in L858R/T790M transgenic model Cross et al, Cancer Discov; 4(9);

5 Phase 1 studies underestimate toxicity of recommended Phase II dose 45% patients treated with small molecule targeted agents required dose modifications in Phase III trials Use preclinical PK-PD, dose and schedule modelling in conjunction with biomarkers to support dose selection Roda et al, Clin Cancer Res; 22:

6 H score Target engagement proof of mechanism demonstrated in AZD5363 Phase 1 Preclinical PK-PD at tolerated dose 100mg per kg - biomarker knockdown indicated twice daily dosing was required Phase 1 - Biomarker knockdown in paired tumour biopsy samples IHC analysis ppras40 pgsk3β * 320 mg cont 360 mg cont 480 mg 4/7 800 mg 2/7 Twice daily dosing Davies et al,mol Cancer Ther : 873. Elvin et al, ASCO Meeting abstracts May 2014:

7 Biomarkers in clinical development What sample Invasive vs non-invasive - blood/urine vs tumour biopsy Biomarker stability - sample logistics What technology Quantifiable - assay validation and qualification Samples are precious, often limited multiplex where possible IHC linking target to pathology/histology Target population data Variability of biomarker in disease prevalence Level of target inhibition setting expectations 28

8 ctdna avoids potential biopsy sample error Mutational heterogeneity in primary breast cancer Clinical response to AZD9291 according to baseline T790M status concordance between tissue and ctdna Jensen et al, Clin Cancer Res, 2010; 17(4); Thress et al, Lung Cancer 90 (2015)

9 ctdna serial monitoring of treatment response Olsson et al, EMBO Mol Med (2015) 7:

10 CTCs predict survival benefit in prostate cancer Multiple technology platforms CellSearch FDA approved EpCAM+ve, CD45-ve cells EPIC Sciences CK/DAPI+ve, CD45-ve all cells captured. IF analysis incorporates target specific markers Genomic analysis from CTCs provides additional avenue to address tumour heterogeneity de Bono et al, Clin Cancer Res 2008;14(19)

11 approaches can be explored early avoiding the toxicity of ineffective standar Methods Methods for CTC for Detection; CTC Detection; henotypic, Genomic Characterization Methods for CTC Detec typic, CTC biomarkers Genomic - EPIC Characterization Sciences 1) Epic CTC Detection Platform Phenotypic, Charac Relocation 3) and Relocation Capture of and Single Capture Cells of for Single Genotyping Cells for and Genotyping Genomic Instability and Genomic Instabi Identifies all CTC populations 3) with opportunity to quantify additional target related biomarkers 3) Relocation and Capture of S 10ml blood sample 2) 4) Algorithm 4) Development: Algorithm Development: Single Cell 2) Features Single Cell Features Multivariate classifier Multivariate to predict classifier Genomic to predict Instability Genomic from Insta Phenotype Phenotype Protein Biomarker Features Protein Biomarker Features CK cratio (protein expression) CK cratio (protein expression) AR cratio (protein expression) AR cratio (protein expression) Digital Pathology Features Digital Pathology Features Nuclear Area (um 2 ) Nuclear Area (um 2 ) Cytoplasmic Area(um Cytoplasmic 2 ) Area(um 2 ) Nuclear Convex Area (um 2 ) Nuclear Cytoplasmic Convex Area (um 2 Convex Area (um 2 ) ) Cytoplasmic Convex Area (um 2 ) Nuclear Major Axis (um) Nuclear Major Axis (um) Cytoplasmic Major Axis (um) Cytoplasmic Major Axis (um) Nuclear Minor Axis (um) Nuclear Minor Axis (um) Cytoplasmic Minor Axis (um) Nuclear Circularity Cytoplasmic Minor Axis (um) Cytoplasmic Circularity Nuclear Circularity Nuclear Solidity Cytoplasmic Circularity Cytoplasmic Solidity Nuclear Solidity Nuclear Entropy Cytoplasmic Solidity Nuclear to Cytoplasmic Convex Nuclear Entropy Area RatioNuclear to Cytoplasmic Convex Nucleoli Area Ratio CK Speckles Nucleoli 32 Nuclear Speckles CK Speckles Phenotypes Phenotypes 2) orm meration, CTC Enumeration, r orkflow: Analyses Workflow: o le slides placed and onto stored slides in a and -80 stored C in a -80 C tin (CK), CD45, DAPI, AR N-term and with cytokeratin (CK), CD45, DAPI, AR N-term and lti-parametric digital Schematic pathology of Epic CTC Platform CTC Enumeration, ected by a multi-parametric digital pathology tion. ader confirmation. Morphology, and Biomarker Analyses Workflow: and AR channels and ~20 nuclear and the DAPI, CK, 1) and Nucleated AR channels cells and from ~20 blood nuclear sample and sified. placed onto slides and stored in a -80 C Prediction d, cted shotgun and classified. Source: EPIC libraries Sciences biorepository. constructed, Slides and are stained with cytokeratin (CK), CD45, DAPI, AR N-term and ome V categorizes amplified, 1Mb scanned. shotgun segments libraries CTC for candidates constructed, are detected and by a multi-parametric digital pathology l le breakpoints genome CNV for algorithm categorizes regions >10Mb followed 1Mb are segments by human for reader confirmation. chromosomal breakpoints for regions >10Mb are 2) CTCs are then segmented within the DAPI, CK, and AR channels and ~20 nuclear and Genomic Instability Genomic Ins Single Cell Features Protein Biomarker Features CK cratio (protein expression) AR cratio (protein expression) Digital Pathology Features Nuclear Area (um 2 ) Cytoplasmic Area(um 2 ) Nuclear Convex Area (um 2 ) Cytoplasmic Convex Area (um 2 ) Nuclear Major Axis (um) Cytoplasmic Major Axis (um) Nuclear Minor Axis (um) Cytoplasmic Minor Axis (um) Nuclear Circularity Cytoplasmic Circularity Nuclear Solidity Single CTC genome analysis Prediction

12 Right patient disease fragmentation and targeted therapy Colorectal Breast Lung - adenocarcinoma Zhang et al, Nature 2014;513: ; Pereira et al, Nature Comm 7:

13 Vandetanib targeted for RET-fusion positive NSCLC Vandetanib - an inhibitor of VEGFR2, EGFR and RET Approved for medullary thyroid cancer - RET, VEGFR and EGFR contributing to disease progression Activity in Phase II in differentiated thyroid cancer - RET rearrangements contribute to disease progression RET-fusions identified in NSCLC reported in Feb Prevalence in vandetanib Phase III NSCLC tumour samples ~0.7% Preclinical activity in LC-2 NSCLC cells Endogenous RET fusion FDG-PET/CT responses in patient with poorly differentiated lung adenocarcinoma harbouring a RET-KIF5B rearrangement Baseline Day 8 Day 28 RET-KIF5B fusion by FISH Matsubara et al, J Thoracic Oncol 7: Gautschi et al, J Thoracic Oncol 8: e

14 Biomarker strategy key points Biomarker selection to address key questions for pharmacodynamics and patient selection Assay validation/qualification and compatibility with clinical deployment Plan for success ensure protocol anticipates potential sample use Sample tracking and recovery importance of sample labelling and documentation Patient selection/diagnostic development needs to be in place for Phase 1 avoid unnecessary delays 35