Los días 15 y 16 de octubre de 2014, la Fundación Ramón Areces y la Real Academia Nacional de Farmacia, en colaboración con la Fundación de la Innovación Bankinter, reunieron en Madrid a algunos de los mayores expertos mundiales en nuevas terapias contra el cáncer. El Simposio Internacional, coordinado por la profesora y académica María José Alonso, analizó el momento actual de la lucha contra esta enfermedad. También fue un punto de encuentro para científicos de los más innovadores institutos de investigación en oncología, quienes debatieron sobre tres grandes temas: la Medicina Personalizada contra el cáncer, los nanomedicamentos en la terapia del cáncer y las terapias basadas en la inmunomodulación.

1. Personalize Medicine in PDAC
Manuel Hidalgo, M.D., Ph.D.

2. Biomarkers in PDAC
• CA199 is the only used biomarker:
– Stage of disease.
• High levels heralds advanced disease
• Inclusion criterium in adjuvant and neoadjuvant studies (100 U/ml)
– Poor prognosis.
– Decrement associated with better survival.
Reni et al. Cancer 2009;115:2630–9

3. Multifaceted Biology of PDA
M Hidalgo & D Von Hoff. CCR 2012;18:4249-56

4. hENT1 Expression in PDAC
Poplin et al. JCO 2013;31:4453-61

5. LEAP Trial in PDAC
hENT1 low All patients
Poplin et al. JCO 2013;31:4453-61

6. hENT1 Expression in ESPAC-3
Greenhalf et al, JNCI 2013;106:djt347

7. Overall Survival
Von Hoff et al. NEJM 2013; 369:1691-703

8. Nab paclitaxel Transport
1. Albumin is a natural carrier of hydrophobic
molecules (like paclitaxel)
2. There is active transport of albumin-carrying the
drug across the vascular endothelium into tumors
(microenvironment) via a receptor mediated pathway
involving the following proteins
• Gp60 (albumin receptor)
• Caveolin 1 (protein forms invagination
and carries albumin through endothelial
cells –transcytosis)
• SPARC – binds albumin to create sink in
stroma and on tumor cells (both high in
SPARC)

9. Paclitaxel albumin + gemcitabine in patients
with metastatic pancreatic cancer: SPARC
• SPARC status was evaluated in
36 patients
• A significantly longer OS was
reported in the high SPARC vs
low SPARC group
– Median OS: 17.8 vs 8.1
mo, p=0.0431
• SPARC level remained a
significant predictor for OS
after adjusting for clinical
covariates (eg age, sex, race,
baseline CA 19-9) (p=0.041)
• Stromal SPARC correlated with
OS (p=0.013) but SPARC in
tumour cells did not (p=0.15)
0 3 6 9 12 15 18 21 24 27 30
Von Hoff et all. JCO 2011 Oct 3. Epub ahead of print.
100
75
50
25
0
Time (months)
Probability of survival (%)
Average z score ≥0,
high SPARC (n=19)
Average z score <0,
low SPARC (n=17)
p=0.0431
17.8 months
8.1 months

10. SPARC Expression in Tumor Stroma is
Associated with Worse Outcome
Overall: log-rank p<0.01
Intratumour SPARC only: log-rank p=0.13
Stromal SPARC only: log-rank p<0.001
Tumour-/-Stroma
Tumour+/-Stroma
Tumour-/+Stroma
Tumour+/+Stroma
0 12 24 36 48
Time following surgery (months)
Infante et al. J Clin Oncol 2007;25:319-25.
1.00
0.75
0.50
0.25
0
Proportion surviving
No. at risk
Tumour -/- Stroma 49 37 21 10 7
Tumour -/+ Stroma 50 35 15 9 7
Tumour +/- Stroma156 74 17 7 3
Tumour +/+ Stroma44 24 9 5 4
Infante et al, JCO 2007;25:319-325

11. 30% of MPACT Patients Were Evaluable for the
Stromal SPARC Biomarker Analysis
861
All patients (ITT population) 100%
376
256
131
nab-P +
Gem
125
Gem
12.8% pancreas
50.4% liver mets
20.0% other mets
16.8% NOS
Stromal SPARC evaluable
44%
30%
9.9% pancreas
55.0% liver mets
17.6% other mets
Sample tissue of origin
17.6% not otherwise specified (NOS)
Hidalgo M, et al. Oral presentation at: World GI 2014 [abstract O-0004]

12. Stromal SPARC Not Prognostic of Overall Survival in
MPACT with Clinical Trial Assay
Stromal SPARC IHC SCORE
SPARC Level HIGH LOW
n (%) 71 (28%) 185 (72%)
HR high vs low (95% CI) 1.019 (0.750, 1.386)
P-value (log-rank) 0.9026
Proportion of Survival
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0 6 12 18 24 30 36
Time, months
Hidalgo M, et al. Oral presentation at: World GI 2014 [abstract O-0004]

13. Stromal SPARC - Multivariate Analysis in SPARC
Evaluable Population
Covariatea HR 95% CI P Value
Treatment Group (nab-P + Gem vs Gem) 0.65 0.47 - 0.89 0.007
KPS (70-80 vs 90-100) 1.50 1.09 - 2.06 0.012
Presence of Liver Metastases (Yes vs No) 2.12 1.31 - 3.41 0.002
• In a multivariate analysis, stromal SPARC was not a
significant, independent predictor of overall survival1
• Treatment, Karnofsky performance status, and presence of
liver metastases were significant predictors of OS,
consistent with the ITT analysis2
A stepwise procedure was carried out using the following factors: treatment group, age, sex, KPS, Geographic region,
pancreatic cancer primary location, presence of biliary stent, previous whipple procedure, presence of liver metastasis,
presence of pulmonary metastasis, peritoneal carcinomatosis, stage at diagnosis, number of metastatic sites, baseline
level of CA 19-9, and stromal SPARC.
1. Hidalgo M, et al. Oral presentation at: World GI 2014 [abstract O-0004];
2. Von Hoff DD, et al. N Engl J Med. 2013;369:1691-1703.

14. Cytokeratin 19
Sparc (+/+)
Sparc (-/-)
100
80
60
40
20
0
Positive pixels (%)
Collagen I
Sparc (+/+)
Sparc (-/-)
100
80
60
40
20
0
Positive pixels (%)
Masson
Sparc (+/+)
Sparc (-/-)
50
40
30
20
10
0
Positive pixels (%)
a-SMA
Sparc (+/+)
Sparc (-/-)
10
8
6
4
2
0
Positive pixels (%)
Vimentin
Sparc (+/+)
Sparc (-/-)
100
80
60
40
20
0
Positive pixels (%)
Sparc (+/+) Sparc (-/-)
Sparc
Ck-19
Collagen I
Masson
a-Sma
Vimentin
No difference in IHC markers
between SPARC null and
SPARC wt K-Ras_G12V p53
KO driven tumors

15. No difference in Overall Survival (OS) between SPARC
null and SPARC wt K-Ras_G12V p53 KO mice
Unpaired t-Test; two-tailed:
Sparc+/+ vs. Sparc-/+ p=0.5960
Sparc+/+ vs. Sparc-/-p=0.6407
Group 1: Control (vehicle)
Group 2: Abi as single agent (50mg PTX/kg, once a week
x 4, i.v.)
Group 3: Gem (100mg/kg, twice a week x 4, i.p.)
Group 4: Abi + Gem at the above mentioned doses

16. Genomic Diversity of Pancreas Cancer
Jones et al, Science 2008;321:1801

17. EGFR Expression is Needed for
KRAS Driven PDA Tumorigenesis
Navas et al. Can Cell 2012;22:318–330

18. Loss of EGFR Delays but does not
Prevent PDA Tumorigenesis in p53 -/-
Navas et al. Can Cell 2012;22:318–330

19. Striking Pan-HER Response in Pancreatic PDX
Models
 Patient-derived xenograft models with mutated KRAS and known
resistance to HER family targeted therapeutics selected
Pan-HER dosed at 50 mg/kg, 3xqw, 10 doses in total
 Target expression and/or modulation not indicative of responsiveness

20. PALB2 Mutations
Tischkowitz and Xia, CR 2010; 70:7353-7359

21. Targeting PALB2 Mutations
Villarroel et al, MCT 2011;10:3-8

22. TSC2 Mutation
Houghton, CCR 2010

23. Ranking Rx Candidates
1400
1200
1000
800
600
400
200
0
JH033
1 4 8 11 15 18 22 25 28
Day
Tumor volume (mm3)
Control
Rapamycin
MMC
Rapamycin + MMC

24. Salient Features of Avatar Models

25. A Molecularly Annotated Platform of
Patient-Derived Xenografts
(“Xenopatients”) Identifies HER2 as an
Effective Therapeutic Target in Cetuximab-
Resistant Colorectal Cancer, 2011
Tumor grafts derived from women with
breast cancer authentically reflect tumor
pathology, growth, metastasis and disease
outcomes, 2011
A Validated Tumorgraft Model Reveals
Activity of Dovitinib Against Renal Cell
Carcinoma, 2012

26. Application of Avatar Models to
Personalize Medicine

27. JJ00550077 DDeessiiggnn
Surgery
Xenograft
PD
Rx Effective in
Avatar Mode
Garrido-Laguna, CCR 2011

28. J0507 Flow Chart
Garrido-Laguna, CCR 2011

29. DPC4 Status Affects Engrafment
Garrido-Laguna, CCR 2011

30. Engraftment is Associated with
Poor Prognosis
Garrido-Laguna, CCR 2011

31. Xenograft Response to Gemcitabine
Garrido-Laguna, CCR 2011

32. Prospective Cases Series
• To guide patient treatment by selecting
the most effective regimen from a panel
of agents tested against the patient own
tumorgraft.
• To study mechanisms underlying drug
sensitivity in these tumors.

33. Overall Results
Tumor Type Number of
Drugs
Tested
Predicted Clinical
Sensitivity
(Yes/No)
Predicted
Clinical
Resistance
(Yes/No)
Treatment Course &
Duration of Response
PDA 17 Yes, three times Yes, once 2nd line, CR, 46 + mo
PDA 5 No active agent Yes, once 1st line, PD
PDA 4 Yes, once Not tested 1st line, SD, 6 mo
PDA 1 Yes, once Not tested 1st line, SD, 6 mo
LMS 28 Yes, twice Yes, twice 4th line, PR, 9 mo
MCS 21 Yes, once Yes, once 3rd line, PR, 9 mo
NSCLC 25 Yes, once Yes, twice 3rd line, PR, 9 mo
Esophageal 24 Yes, three times Not tested 3rd, 4th, & 5th, PR 48 + mo
Myoepithelioma 13 No active agent Yes, twice 1st and 2nd line, PD
CRC 16 Yes, once Not tested 3rd line, PR, 12 + mo
Breast Cancer 12 Yes, once Not tested Not treated
11 166 14/14 9/9 N/A
Hidalgo et al, MCT 2011

34. Case Studies
Paz et al, CBI, Unpublished

35. Overall Results
Paz et al, CBI, Unpublished

36. Break Down Results
Paz et al, CBI, Unpublished

37. Metrics
PPV % NPV %
Overall 89 96
First Line 95 100
Recurrent 87 93
Paz et al, CBI, Unpublished

38. Exome
sequencing
Normal DNA
(Blood)
Tumor
Aim 1
Genomic profiling
Detecting Biomarkers
Transcriptional
Profiling
Copy number and
Sequencing (somatic
Mutations)
SENSITIVE RESISTANT
Potential drugs
based on Genomic
profiling
Drug Response in
Avatar mouse
Aim 2
Identify drug sensitivity based
on tumor genomics alterations
Patient 1
Personalized drug therapy
Tumor
Aim 3
Genomic profiling
Predictive &Prognostic
methodology
Personalize
d drug
therapy
Custom
Database with
Response
Biomarkers
Avatar

39. N# TUMOR TYPE N# MUTS RELEVANT SOMATIC MUTS N# CNV RELEVANT CNV PUTATIVE
TARGETS.
1 Neuroendocrine
tumor 5 CREB3L3, ITPR2, MYO5B 0 0 CREB3L3
2 Glioblastoma 63 EPHA3, NF1, PTPN11 ,FAS , CDKN2A 0 0 NF1
3
High grade
pancreatic
neuroendocrine
62 ARID1A, ARID1B, JAKMIP2, JARID2
PIK3C2A, PIK3CA, SSTR2, DDR2, TP53 6 GNG11 PI3KCA, DDR2
4 PDAC 38
KRAS, UBA1, FAM83H,
SMAD4, SLC15A2, PIWIL3, SLC3A2, SLC22A17,
TP53.
10 0 -
5 Melanoma Uveal 5 GNA11, TAOK3 0 0 GNA11
6 Colon cancer 71 APC, DICER 1, TP53, CHEK1, SOS1 63 0 CHEK1
7 PDAC 952 BRCA1, EZH2, FGFR2, FN1, IGF1R, KDR, KRAS,
MET, MPL, PRKCB, PIK3C2G, PTK2B. 0 0
FGFR2, IGF1R,
PIK3C2G, MET,
BRCA1
8 Melanoma 29 BAI3, DNAH5, MDN1, NRAS 2 SKT19 NRAS
9 PDAC 18 SMAD4, KRAS, ERBB2IP 0 0 ERBB2IP
10 PDAC 21 KRAS, XPC, P53 0 0 XPC
11 PDAC 29 KRAS. P53,SMAD4 3 0 -
12 Renal Carcinoma 25 BAP1 965
ZAP70, FGFR3, NOTCH1,
TERT, STK11, GNA11,
ZNF668, SOCS1, IRS2
BAP1, FGFR3,
NOTCH1, STK11,
GNA11.
13 Glioblastoma 64 MLLT10, PBRM1 27 EGFR, CDKN2A, ERRFI1 EGFR
14 Cervix Cancer 73 ALK, NOTCH, FANCB, MAOA 186 ERBB2, JAK3, AKT ALK, AKT, ERRB2
15 PDAC 573 BRCA2,GAK 0 0 BRCA2
16 PDAC 48 KRAS, NOTCH1, NF2, MMP21 35 0 NF2, NOTCH1
Garralda et al, CCR 2014

40. Primary tumor Gene/Pathway
Targeted Matched treatment
Best
Response
(RECIST)
Time on
treatment
(months)
Present
status
Neuroendocrine
tumor
CREB3L3
mutation Sandostatin + Metformine EE.
CR by PET 18 On
treatment
Glioblastoma NF1 mutation Everolimus + Erlotinib +
Bevacizumab PD 3 Exitus
High grade pancreatic
neuroendocrine
tumor
PI3KCA, DDR2
mutations Dasatinib PD 3 Exitus
Uveal melanoma GNA11 mutation
1st: Protein Kinase C
inhibitor.
2nd: Carboplatin +
Paclitaxel + Pi3k inhibitor
EE 4 Exitus
PDAC XPC mutation Mytomicin C EE (clinical
benefit) 3 On
treatment
Renal BAP1 mutation Mytomicin C + Irinotecan EE 4 On
treatment
Glioblastoma EGFR
amplification Erlotinib EE 3 On
treatment
PDAC BRCA2 Mytomicin C EE 5 On
treatment

41. Panc 031 Accionable Targets
Saal L H et al. Cancer Res 2005;65:2554-2559
Hammerman et al. Cancer Discovery 2011;1:78-89
Point mutation in PIK3CA
gene: 909F>C.
Point mutation in DDR2
protein(discoidin domain
receptor tyrosine kinase 2).
Amplification GNG11
Garralda et al, CCR 2014

42. In Vivo Testing
Garralda et al, CCR 2014

43. In Vivo Testing
Garralda et al, CCR 2014

44. The Avatar Clinical Trial

45. Trial Progress
Patient Code Arm Sex Age Xenograft
Code
Avatar-01 Conventional M 76
Avatar-02 Experimental M 76 Panc-079
Avatar-03 Experimental F 43 Panc-078
Avatar-04 Conventional F 44
Avatar-05 Conventional F 71
Avatar-06 Experimental M 49 Panc-080
Avatar-07 Experimental F 49 Panc-081
Avatar-08 Experimental M 55 Panc-082
Avatar-09 Experimental F 63 Panc-083
Avatar-10 Experimental F 59 Panc-084

46. Sequencing
Ion Torrent Proton
Ion AmpliSeq™ Comprehensive Cancer Panel - 409 Genes
Features:
• Broad survey of 409 key genes in a simple PCR reaction, no
additional capital equipment required
• Unmatched plexy of 16,000 primer pairs in only four pools with
Ion AmpliSeq™ technology
• Low DNA input of only 40 ng DNA and short amplicons enable
FFPE samples and needle biopsies
Comprehensive Cancer Panel was designed to target all
exons of key tumor suppressor genes and oncogenes most
frequently cited and most frequently mutated.
Strategically designed to interrogate CDS and splice
variants across multiple gene families simultaneously,
pathway-based gene selection profiles mutational
spectrum in cancer driver genes and drug targets, along
with signaling cascades, apoptosis, DNA repair,
transcription regulators, inflammatory response, and
growth factor genes in a single assay.

47. Bioinformatics Analysis
Changes Changes
409 genes Targeted
sequencing of exons
Remove non-coding mutations
ccooddiningg S SNNVVss/i/ninddivivididuuaal l
Remove synonymous variants
n noonn-s-syynnoonnyymmoouuss c cooddiningg S SNNVVss
Reference
Human DB
nnoovveel ln noonn-s-syynnoonnyymmoouuss c cooddiningg S SNNVVss
22/1/100 c clinlinicicaal lt atargrgeet tg geenneess
Reference
Human DB
Patient’s
Normal DNA
TTuummoor r
FFinindd t utummoor-rs-sppeeccifiifcic ( s(soommaatitcic) )m muutatatitoionnss
Goals
1.- To identify relevant somatic
mutations (SNVs and CNVs)
2.- To select actionable
mutations
3.- To propose approved and/or
clinical trials drugs using our
Pan-Drugs DB
Remove known SNPs
Select deleterious non-synonymous / nonsense /
frameshifts mutations

48. Example Patient 02
Distribution of drug categories for the
selected relevant mutations
Unique drugs
Approved
Clinical
Experimental
FDA-cancer
Relevant missense mutations
predicted as deleterious
SNVs
KRAS p.Gly12Asp
TP53 p.Phe109Val
IRS2 p.Gly1143Ala
CDK12 p.Leu584Trp
CCDN1 p.Glu252Gln
Copy Number
CDKN2A del
PALB2 del

49. Conclusions
• So far, no biomarker, other than CA19.9, has
gained clinical acceptance.
• Discordant results in IHC-based biomarkers
• Genomic analysis may point out biomarkers with
predictive role.
• Global genomic analysis should be tested in
efficacy oriented clinical studies,

50. GI Group at CNIO
D. Spas N. Baños
P. Lopez-Casas M. Muñoz
L. Fernandez E. Garralda
V. Moreno L. Moreno
R. Martinez C. Menendez
SU2C Team at JHU
C. Iacobuzio-Donahue
R. Hruban
A. Maitra
R. Kumar
E. Oliveira
V. Velculescu
C. Dang
PCRT
A.Stoll
C. Moriarty
PCRT Investigators
TGen
D. Von Hoff
R. Ramanathan
GI Group at CIOCC
E. Vicente
Y. Quijano
A. Cubillo
J. Rodriguez
R. Alvarez
O. Hernando
E. Vega
Pancreas Team at CNIO
M. Barbacid
C Guerra
P. Real
C. Heeschen
N. Malats
LDT Lab at CIOCC
F. Lopez-Rios
Imaging Unit at CNIO
P. Mulero
CBI
D. Sidransky