This document discusses the potential of personalized medicine using genomic and other patient data. It notes that current medicine treats all patients as averages, but personalized approaches could use large databases to predict individual risk. New technologies like genome sequencing, biosensors, and artificial intelligence could enable superconvergence of data to precisely tailor treatment. This would increase efficiency and allow treatment of interconnected biological systems rather than reductionist views. The document advocates for integrated electronic medical records and biobanks to enable these personalized approaches.

1. Personalised medicine
Medicine for the Quantum age….
Dr. Patrick Gladding, MBChB, PhD
Cardiologist, WDHB

2. Problems with contemporary medicine
• Clinical decisions are based on historical clinical
trial data
• Clinical trials have traditionally taken
heterogeneous populations and demonstrated
results only for the “average patient”
• Personalized risk predictions could be made
using databases containing heterogeneous data,
e.g. genomic data, imaging data (EMR) and
cardiac models

3. Current medicine is reductionistic
“The whole is more than the sum of its
parts.”
– Aristotle, Metaphysica
Interconnected and Interdependent Systems
3

4. 4

5. 5

6. Personalised Medicine
• Its about:
– populations as
much as
individuals
– Increasing
efficiency and
resource
utilisation
– In a world of
limited resources
only option is to
personalise

7. Genomic and Personalised Medicine
• Genomics,
proteomics,
metabolomics
• Information
technology
• Artificial intelligence
• Supercomputing
• Molecular imaging
• Nanomedicine
• Biosensors

8. • Superconvergence
– “Ubiquitous” Cloud
supercomputing
– Low-cost genome sequencing
– Pervasive connectivity
– Digital medicine
– mHealth

9. tgaagccctctttttctctccttctatttctctctagagcactcaagactttactgacgaaaa
ctcaggaaatcctctatcacaaagaggtttggcaactaaactaagacattaaaagga
aaataccacatgccactctgcaggttgcaataactactacttactggatacattcaaa
ccctccagaatcaacagttatcaggtaaccaacaagaaatgcaagccgtcgacaa
cctcacctctgcgcctggtaacaccagtctgtgcaccagagactacaaaatcaccca
ggtcctcttcccactgctctacactgtcctgttttttgttggacttatcacaaatggcctggc
gatgaggattttctttcaaatccggagtaaatcaaactttattatttttcttaagaacacag
tcatttctgatcttctcatgattctgacttttccattcaaaattcttagtgatgccaaactggg
aacaggaccactgagaacttttgtgtgtcaagttacctccgtcatattttatttcacaatgt
atatcagtatttcattcctgggactgataactatcgatcgctaccagaagaccaccagg
ccatttaaaacatccaaccccaaaaatctcttgggggctaagattctctctgttgtcatct
gggcattcatgttcttactctctttgcctaacatgattctgaccaacaggcagccgagag
acaagaatgtgaagaaatgctctttccttaaatcagagttcggtctagtctggcatgaa
atagtaaattacatctgtcaagtcattttctggattaatttcttaattgttattgtatgttataca
ctcattacaaaagaactgtaccggtcatacgtaagaacgaggggtgtaggtaaagtc
cccaggaaaaaggtgaacgtcaaagttttcattatcattgctgtattctttatttgttttgttc
ctttccattttgcccgaattccttacaccctgagccaaacccgggatgtctttgactgcac
tgctgaaaatactctgttctatgtgaaagagagcactctgtggttaacttccttaaatgca
tgcctggatccgttcatctattttttcctttgcaagtccttcagaaattccttgataagtatgc
tgaagtgccccaattctgcaacatctctgtcccaggacaataggaaaatagaacag
gatggtggtgacccaaatgaagagactccaatgtaaacaaattaactaaggaaat

10. tgaagccctctttttctctccttctatttctctctagagcactcaagactttactgacgaaaa
ctcaggaaatcctctatcacaaagaggtttggcaactaaactaagacattaaaagga
aaataccacatgccactctgcaggttgcaataactactacttactggatacattcaaa
ccctccagaatcaacagttatcaggtaaccaacaagaaatgcaagccgtcgacaa
cctcacctctgcgcctggtaacaccagtctgtgcaccagagactacaaaatcaccca
ggtcctcttcccactgctctacactgtcctgttttttgttggacttatcacaaatggcctggc
gatgaggattttctttcaaatccggagtaaatcaaactttattatttttcttaagaacacag
tcatttctgatcttctcatgattctgacttttccattcaaaattcttagtgatgccaaactggg
aacaggaccactgagaacttttgtgtgtcaagttacctccgtcatattttatttcacaatgt
Single Nucleotide Polymorphisms (SNPs)
atatcagtatttcattcctgggactgataactatcgatcgctaccagaagaccaccagg
ccatttaaaacatccaaccccaaaaatctcttgggggctaagattctctctgttgtcatct
~1 letter every 1,200
gggcattcatgttcttactctctttgcctaacatgattctgaccaacaggcagccgagag
acaagaatgtgaagaaatgctctttccttaaatcagagttcggtctagtctggcatgaa
atagtaaattacatctgtcaagtcattttctggattaatttcttaattgttattgtatgttataca
ctcattacaaaagaactgtaccggtcatacgtaagaacgaggggtgtaggtaaagtc
cccaggaaaaaggtgaacgtcaaagttttcattatcattgctgtattctttatttgttttgttc
ctttccattttgcccgaattccttacaccctgagccaaacccgggatgtctttgactgcac
tgctgaaaatactctgttctatgtgaaagagagcactctgtggttaacttccttaaatgca
tgcctggatccgttcatctattttttcctttgcaagtccttcagaaattccttgataagtatgc
tgaagtgccccaattctgcaacatctctgtcccaggacaataggaaaatagaacag
gatggtggtgacccaaatgaagagactccaatgtaaacaaattaactaaggaaat

11. 10101011010010101010001010010101001010010000100011
00101001010101001010010000010111110010101010011010
01001010001010100010100101001001001010111111111111
11111111111111000000101001000100100000100010011100
01010010101001010010100101010011001010000100111111
11111111111111100000100101101111110011001010101011
01011110101010010101010010101101001010010010101001
01111111100101010101010101010100101010010100101001
01111111111111111110010010101010100101010100101010
Single Nucleotide Polymorphisms (SNPs)
01111111111111111111010101001010010101001010100100
10100101010010100100101001010101001010010101001001
~1 letter every 1,200
00101001001010010010101001010100101011111111111111
11111111111100101010101001010101001010010010101010
00001010111111111111111111010100101001000000000000
00000001010100101001010100101010100101010101100100
00101010100101010010101000101010100010101001010010
00101010111101000100010001111100000100111110001010
10001010111100100101010010101001010011111001010101
01001010010010001000111111111110001010101000001011
11111000101110101010001010101000101111100100101001
11110010100101110100101111001

12. Genome sequencing > Exponential reduction in cost
NEJM 2012

13. Handheld minisequencing
• $1 per test
• 20mins for result
• Not yet available

14. Feedback loops – giving patients their
information
Action Personalised data
Relevance
Choices
A Learning system 14
TED talks: Thomas Goetz

16. • An integrated EMR &
biobank is essential
• Semantic
interoperability
• Machine processable
databases

17. Gladding et al. Personalized Medicine (2010) 7(4)

18. Deep computing – data mining

19. Network
Medicine
• Nodes, hubs
• Topology
• Scale free
• Self-organising
• Small world
• Holographic

20. Social networks and molecular medicine

21. Reclassifying disease – “diseasome”
• Linking
diseases
through
networks,
closest
neighbour
based on
genomics
http://diseasome.eu/

22. Drug repositioning

23. Non-profit
Socially responsible
Virtual company
Jobs and NZ knowledge economy

24. Disclosures - IP
CardioMPO
Australasia
Clopidogrel
Pharmacogenomics
PCT phase US, China, Japan,
Europe, Australia, New Zealand
• Founder of non-profit Theranostics Laboratory
• USPTO issued patent; Licensing fees and
royalties

25. Pharmacogenomics
Tailoring drug treatment to genotype

26. 0.75c/d $4/d
Prasugrel
$550k $2.9 million
Ticagrelor
$6/d
$4.4 million
26
Pharmgkb

27. Ethnic rates of nonresponders
*2/wt = higher dose
*2/*2 = alternative drug
27

28. Rapid low-cost genotyping - NSH
1
• MALDI-TOF MS
• Cheap <$30
• Nanosphere
2 o Rapid 2hrs
o Blood->result
o Clopidogrel/warfarin
PGx
o Norovirus
o Hypervirulent C. diff

29. Mass Spectrometry MALDI-TOF

30. Mass Spectrometry MALDI-TOF
2
• Pattern recognition of spectral output
• ‘Machine learning’ 3
• Dendrogram:
assignment of
species pattern
1
• Culture specimen
• Ionise
• Analyse 4
• Species identification based on probabilistics

31. Cardiac mHealth projects
Community focused
31

32. Cardiac ultrasound project
• Modeling astronaut’s hearts on ISS
• NZ Employment

33. ICMA v 1.0
User ID
Password

34. Jagir Hussan ICMA v 1.0
Auckland 03 Sept 12 10:31 NZST

35. Jagir Hussan ICMA v 1.0
Auckland 03 Sept 12 10:47 NZST
MR. John E. Dtotu
Referring Physician: Dr James Thomas
5543AT75
1071, Kna street, Male 163 cm
04-971-3343 90 kgs
Kapiti Coast 10-12-1975
Scans Models
4Views Stress14 Pretrial P4LVX
Created on: 31 Sept 2012 Created by: Dr. Patrick Gladding Status: Validated (Dr James Thomas)
Analysis Workshee
t
APLAX Normal Fiber
4CH
2CH
SAX
Surface
Lines
Epi Endo
CLEAR ALL
CLEAR ALL SAVE
SAVE

36. Advanced ECG for general practice
Digital ECG e.g. XML
Internet
• Spectralised, digital ECG
• Artificial intelligence
• Metric for health

37. Advanced ECG
• Standard 12L snap-shot resting ECG from
Mortara, Philips, Cardiax machine
• 12L ECG is ‘spectralised’ into multiple
parameters
• Pattern recognition, artificial intelligence applied
to resulting parameters
• Higher diagnostic yield than standard 12L ECG
for CAD, HCM, NICM, ICM

38. Enhancement of exercise treadmill testing
• Problem: Limited access to exercise
treadmill testing for patients with chest pain
• ETT Sensitivity 67%, specificity 78%
• n=58 patients (10 normal, 48 severe disease
on coronary angio)
• All with positive treadmill “coronary disease”
• One excluded due to noise and six due to
other cardiac abnormalities on echo
38

39. A-ECG results
• Lime - Healthy
• Red – CAD
• Blue - HCM
• Aqua - LVH
• Purple - NICM
• Orange - ICM
39

40. Enhancement of exercise treadmill testing -Results
• A-ECG sensitivity of 97.2% and specificity
of 53% for the angiography results
• 8 healthy patients with abnormal ETT
would have been identified correctly with
A-ECG
• Theoretically reducing unnecessary
invasive coronary angiograms by 53%
• Cost saving $528/A-ECG
40

41. Screening programs Population Genomics
Targeted resource allocation Modeling
Community oriented Disease Simulation
Patient-centric Cost-effectiveness Analysis
Diagnostics Lab
Wireless Biosensor Genebank
High risk individual
Digital Avatar
Electronic Healthcare Database
Population 4.2 million

42. Conclusion
• Personalised medicine is emerging as the
fastest evolving aspect of medicine
• Ushering in a new era in preventative medicine
with low cost molecular diagnostics, advanced
informatics
• Will be applicable to every specialty within
medicine
• Need for industry partnerships (e.g. IBM, Orion)
& HINZ members
• Needs your help!
– Acknowledgements: Innovation Hub, North Shore
hospital laboratory staff, CEO