Clinical trial design in the current age of immunotherapy and targeted therapy Martijn Lolkema, MD/PhD Medical Oncologist Erasmus MC Cancer Institute
Content Clinical trial design: the old paradigm The challenges posed by immunotherapy The challenges posed by targeted therapy How to deal with the exploding trial portfolio
CLINICAL TRIAL DESIGN: THE OLD PARADIGM
Drug development in oncology Oncology Drug Development Preclinical hypothesis/ compound generation Phase I Phase II Phase III/IV
Phase I studies Dose selection Toxicity determination PK analysis PD analysis
Phase I: dose selection
PK = Pharmacokinetics What does the body do to the drug?
PD = Pharmacodynamics What does drug to its target?
Phase II study
Phase 3 studies Standard of care R (Standard of care +) novel treatment
Conclusies
HOW TO TACKLE THE MOST IMPORTANT PROBLEMS IN ONCOLOGY DRUG DEVELOPMENT
There are >900 drugs in development Only 5% of novel drugs in cancer treatment become FDA/EMA approved Success rate 100% x x x = ~30% ~40% ~70% ~5% ~60% 90 80 70 60 50 40 30 20 10 0 Phase I Phase II Phase III Registration Overall Phases in the drug development process SOURCE: I. Kola, J. Landis “Can the pharmaceutical industry reduce attrition rates?” Nature Reviews drug discovery, ‘04
Where should we look for improvements?
Success in phase I essential, phase II less important 100 75 %Change From Baseline 50 (Sum of Lesion Size) 25 0 -25 -50 -75 Onderzoeker bepaald -100 Inclusief bevestigde & onbevestigde tumorrespons
How to improve translation Efficacy/ Efficacy/ Efficacy Adjust the clinical trial paradigm
Drug development in oncology Efficacy Safety Preclinical hypothesis/ compound generation Phase I Phase II Phase III/IV
Fase II onderzoek Traditional Modern Phase III Phase II Phase I
Focussed phase I trials Proof of concept! Kwak EL, et al. N Engl J Med. 2010;363:1693-703 . Anaplastic lymphoma kinase inhibition in non-small-cell lung cancer. After screening tumor samples from approximately 1500 patients with non– small cell lung cancer for the presence of ALK rearrangements, we identified 82 patients with advanced ALK -positive disease The overall response rate was 57%; 27 patients (33%) had stable disease. A total of 63 of 82 patients (77%) were continuing to receive crizotinib at the time of data cutoff, and the estimated probability of 6-month progression-free survival was 72%, with no median for the study reached.
Challenges in performing these studies All cancer patients without standard treatment options Disease and All-comer Disease Mutation Mutation specific specific specific
Patient Centered studies Patient population: Verkoop In study In study A TEST Patient In study B etc In study X
Basket and Umbrella studies
SOME THOUGHTS ON BIOMARKERS AND PHARMACODYNAMICS
Patient population RCT Chemotherapy RCT Targeted therapy placebo Active agent placebo Active agent 5% response cannot be 50% response can be detected easily! detected, but: 50% response in the blue dots could!
Biomarkers: I-SPY trial an example of how to perform biomarker studies Rugo HS et al. N Engl J Med 2016;375:23-34.
Results from I-SPY study: Her2 negative patients were selected for randomization Rugo HS et al. N Engl J Med 2016;375:23-34.
Outcome: based on pathologic complete response Rugo HS et al. N Engl J Med 2016;375:23-34.
The strength of Bayesian statistics
Conclusion on I-SPY trial Well-designed studies that include biomarker based stratification and randomization are essential in determining predictive biomarkers Early and well-defined outcome variable are essential in producing results that can be interpreted BUT: we need confirmation in a phase III trial…….
MSI: the first approved context independent, predictive biomarker
Anti-PD1 in MSI high patients
Why is MSI the first? Strong biological rationale (Lynch syndrome/ high immune infiltrate/ high mutational load) Can be detected easily (routine diagnostics exist/ IHC and PCR based) Correlation with response robust among different tumor types Biology over ontology!
So do we incorporate PD/ biomarker research in all early phase trials? Sweiss et al JCO 2016
Impact of biopsies A statistically significant biomarker result was reported in 17% of studies (n = 12) Sweiss et al JCO 2016
Biomarkers in clinical practice If you add invasive procedures: make them obligatory so you get the numbers Think about ways to obviate the need for invasive procedures The best place to do this would be at the end of phase I studies. Once you have a biomarkers make sure you get the patient numbers
Overall conclusions Old fashioned oncology drug development paradigm is linear in its thinking Biggest problem we face in oncology drug development is the lack of early signs of efficacy The trial design has developed into a more parallel thinking model: early studies gain in impact Selection is the main item: we need to use better algorithms to select our patients for treatment
Thank you for your attention
Recommend
More recommend