Embed Size (px)
Citation preview
How To Beat Cancer
Jay M. Tenenbaum
HIAI 2013
Summer, 2011
Median Survival
Metastatic Melanoma
Time’s Up
Year 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0
Drug Discovery
FDA Approval
Trials Phase 1
Trials Phase 2
Trials Phase 3
Melanoma Targeted Therapies
Ipilimumab (approved March 2011)
Vemurafenib (approved August 2011)
Cancer in the Genomics Age
• 1000’s of rare diseases
• Treated with 100’s of targeted therapies
• Drug cocktails
• Randomized trials are slow, expensive, & fail
30,000 Oncologists, 1000s of “N of 1” Trials
Vision
A rapid learning community for cancer
• Obtain the best possible outcomes
• Learn as much as possible
• Disseminate rapidly
10
mTOR
AKT
PI3K
PTEN
NRAS
BRAF
MEK
ERK
Bcl-2, Bcl-xL, Mcl-1
BAK
BAX
NOXA, PUMA BIM, BID, BAD
p53
MDM2
p14ARF
CDK4/6 p16
Cyclin D
MITF
MAPK 1
2 NRAS
3 MITF
4 PI3K
5 CDK
6 c-KIT
7 Bcl-2
8 8 MAPK/
PI3K
9 9 MAPK/
CDK
10 10 10
NRAS/
MAPK/
PI3K
Melanoma Reference Model Subtypes Cell Signaling Pathways
Rapid Learning Platform
Users
Doctor Patient Researcher
Partners
Services
Pharma Hospital Insurance Labs
Cancer Commons
Biopsy
Sequence
Compare
Target Test
Treat
Monitor
Precision Oncology 2013
In silico In vivo In vitro
Normal skin cell
Sequencing Machines
Chromosomes
Normal cell
Cancer cell
Treated cell
Scans
Patient
Biomarkers
Original cancer cell
Adapted from NY Times
Biopsy
Sequence
Compare
Target Test
Treat
Monitor
Precision Oncology 2013
In silico In vivo In vitro
Treated cell
Original cancer cell
Sequencing Machines
Chromosomes
Normal cell
Cancer cell
Treated cell
Scans
Patient
Biomarkers
Adapted from NY Times
Beyond Standard of Care
Case Studies in
Precision Oncology
Lukas Wartman, Washington U.
• Diagnosis: ALL (Acute
Lymphoblastic Leukemia)
• Driver: over-expressed
FLT3
• Drug: Sutent (kidney
cancer)
Patient 102 - Seattle Oncologist
• Diagnosis: pancreatic cancer
• Driver: PIK3CA amplification
• Test: in-vitro HTS, mouse avatar
• Drug: PF-05212384 (PIK3/mTOR
inhibitor for solid tumors)
Beating Cancer with Knowledge
Imatinib approvals and trials
• CML, 2001 (Target: BCR-ABL)
• GIST 2003 (Target C-Kit)
• Melanoma 2009 Phase I/II trial
for tumors with C-Kit mutations
• 6 soft tissue sarcomas (C-Kit
Basket trial)
Beating Cancer with Learning
Vemurafenib approvals and trials
• Melanoma, 2011 (Target: v600e
BRAF mutation)
• Colorectal cancer, 2012 (failed
phase 1/2 trial)
Sequence
Compare
Target Test
Treat
Monitor
Precision Oncology 2020
In silico In vivo In vitro
Normal skin cell
Sequencing Machines
Chromosomes
Normal cell
Cancer cell
Treated cell
Scans
Patient
Biomarkers
Adapted from NY Times
Original cancer cell
Biopsies Biopsy
Panomics
Networks
Targets
Network Analysis
Combo Therapies
Serum Markers
Summer, 2011
• Organize the world’s knowledge of cancer biology and therapeutics
• Adaptively plan thousands of ethical treatment “experiments”
• Integrate the resulting data to infer the true causal mechanisms of tumors and drugs
• Generalize the resulting knowledge so that it can be applied to new cases.
Grand Challenge: Beat Cancer
• Organize the world’s knowledge of cancer biology and therapeutics
• Adaptively plan thousands of ethical treatment “experiments”
• Integrate the resulting data to infer the true causal mechanisms of tumors and drugs
• Generalize the resulting knowledge so that it can be applied to new cases.
Grand Challenge: Beat Cancer
Rapid Learning Platform
Users
Doctor Patient Researcher
Partners
Services
Pharma Hospital Insurance Labs
Cancer Commons
Consensus Modeling Process
Evidence Review
Rapid Learning Process
Publishing Knowledge Engineering Rapid Learning Platform
V1 V2
V3 V4
V5
Literature Cases News Conferences
Journal of Precision Oncology
48 hour publication of short form articles
PubMed indexed
Semi-structured for computation and search
Discussion forums
Versioned revisions and updates
All contributions linked to Evidence Review
Rapid Learning Platform
Data and Knowledge
Services
Pharma Hospital Insurance Labs
Users
Doctor Patient Researcher
Partners
Apps
Donate Your Data
• Organize the world’s knowledge of cancer biology and therapeutics
• Adaptively plan thousands of ethical treatment “experiments”
• Integrate the resulting data to infer the true causal mechanisms of tumors and drugs
• Generalize the resulting knowledge so that it can be applied to new cases.
Grand Challenge: Beat Cancer
Planning Challenge
Adaptively plan individual treatment
protocols to achieve optimal outcomes
while
Maximizing the learnings for other
patients and cancer research
Individual Treatment Planning
Efficacy
Safety
Expert
knowledge
Treatment data
Melanoma Pathways (mmmp.org)
DrugRank: Rational Therapy Selection
Drugs
Targets
Pathways
Proteins
Genes
Global Adaptive Trial Presentation
-omics
Standard
treatment
available
?
Treat Response
? Monitor
Compute choices
Best choice
Or patient
preference?
Choose best choice for
experimental purposes
yes yes
yes
no
no no
No options
("dead end") “N of 1”
discovery
mode
recurrence
Global Adaptive Trial Presentation
-omics
Standard
treatment
available
?
Treat Response
? Monitor
Compute choices
Best choice
Or patient
preference?
yes yes
yes
no no
No options
("dead end")
recurrence
Global Adaptive Trial Presentation
-omics
Standard
treatment
available
?
Treat Response
? Monitor
Compute choices
Best choice
Or patient
preference?
Choose best choice for
experimental purposes
yes yes
yes
no
no no
No options
("dead end")
recurrence
Global Adaptive Trial Presentation
-omics
Standard
treatment
available
?
Treat Response
? Monitor
Compute choices
Best choice
Or patient
preference?
Choose best choice for
experimental purposes
yes yes
yes
no
no no
No options
("dead end") “N of 1”
discovery
mode
recurrence
• Organize the world’s knowledge of cancer biology and therapeutics
• Adaptively plan thousands of ethical treatment “experiments”
• Integrate the resulting data to infer the true causal mechanisms of tumors and drugs
• Generalize the resulting knowledge so that it can be applied to new cases.
Grand Challenge: Beat Cancer
Big Data in Oncology
Learning from Cases
Activitome
(adapted from T. Goldstein)
Learning from an ‘N of 1’
Drugs
Targets
Pathways
Proteins
Genes
Multiple
specimens
Panomics &
hypoth. valid.
Public
models
Sequential
biopsies
Response to
therapy
• Organize the world’s knowledge of cancer biology and therapeutics
• Adaptively plan thousands of ethical treatment “experiments”
• Integrate the resulting data to infer the true causal mechanisms of tumors and drugs
• Generalize the resulting knowledge so that it can be applied to new cases.
Grand Challenge: Beat Cancer
