Implementing Precision

Medicine in Oncology:

The IMPACT Clinical

Trials at MD Anderson

Cancer Center

Apostolia M. Tsimberidou, MD,

PhD

Professor

Department of Investigational

Cancer Therapeutics

Disclosures

Research Funding (my institution): Foundation

Medicine, Immatics, Merck/EMD Serono, Boston

Biomedical, Onyx, Bayer, OBI Pharmaceuticals,

Karus, Tvardi, Parker Institute

Advisory Board: Roche, Europe

Hypothesis, 2007

▪Selection of therapy based on patients’ tumor molecular analysis

will improve clinical outcomes compared to the standard approach

Methods

▪Patients who exhausted standard treatment options or had

incurable rare cancers were referred to our Phase I program for

treatment.

▪CLIA-certified tumor molecular testing in consecutive patients

referred for treatment.

▪Genes analyzed: 1-50, depending on time of testing

▪Trials available against various targets

▪Treatment: matched targeted therapy, if available; if unavailable,

non-matched.

▪Retrospective analysis, exploratory.

www.clinicaltrials.gov NCT00851032

Initiative for Molecular Profiling in Advanced Cancer Therapy (IMPACT)

▪Molecular testing: N = 3,743 (2007-2013)

▪ 1,307 (34.9%): ≥1 targetable molecular alteration

▪ 711 (54.4%): matched targeted therapy; 596 (45.6%) non-matched therapy.

▪Median age: 57 yrs (range, 16-86); 39%, men.

▪Median no. of prior therapies, 4 (range, 0-16); previously untreated = 2.8%

▪Cancers: gastrointestinal, 24.2%; gynecological, 19.4%; breast, 13.5%;

melanoma, 11.9%; lung, 8.7%.

Response, evaluable Matched,

N = 697

Non-matched,

N= 571P

Objective response, % 16.2 5.4

Stable disease ≥ 6 months, % 18.7 14.7

Total, % 34.9 20.1 <.001

IMPACT: Results

ASCO 2018, Press Briefing Presentation AM Tsimberidou, MD, PhD

20

40

60

80

100

12 24 36 48 60 72 84 96 108 120 132

Months

No: 596 511 2.8 2.4-3.0

Yes: 711 597 4.0 3.7-4.4

Matched: Total Event Median 95%CI

P < .001HR = .67

Pro

gre

ssio

n-F

ree

Su

rviv

al, %

Progression-Free Survival by Type of Therapy

ASCO 2018, Press Briefing Presentation AM Tsimberidou, MD, PhD

20

40

60

80

100

12 24 36 48 60 72 84 96 108 120 132

Months

No: 596 559 7.3 6.5-8.0

Yes: 711 629 9.3 8.4-10.5Matched: Total Died Median 95% CI

P < .001HR = .72

Overa

ll S

urv

ival, %

Overall Survival by Type of Therapy

ASCO 2018, Press Briefing Presentation AM Tsimberidou, MD, PhD

Risk Factor (vs. other) HR 95% CI P

PAM pathway alterations 1.22 1.08-1.38 .002

Liver metastases 1.46 1.29-1.65 <.001

LDH > ULN 1.66 1.47-1.88 <.001

PS >1 2.15 1.72-2.68 <.001

Albumin < ULN 1.44 1.18-1.76 <.001

PLT > ULN 1.53 1.15-2.04 .003

Age ≥ 60 yrs 1.15 1.03-1.29 .02

CI, confidence interval; HR, hazard ratio; LDH, lactate dehydrogenase; PAM,

PI3K/Akt/mTOR, PLT, platelet count; PS, performance status; ULN, upper limit of normal

Multivariate Analysis, Overall Survival (N = 1,307)

ASCO 2018, Press Briefing Presentation AM Tsimberidou, MD, PhD

Apostolia-Maria Tsimberidou, MD, PhD

20

40

60

80

100

12 24 36 48 60 72 84 96 108 120 132

Months

Score Total Died Median 95% CI

0 204 171 13.7 10.9-17.6

1 420 363 11.0 9.8-12.6

2 403 382 7.4 6.2-8.4

3-4 262 255 4.6 4.2-5.7

5-7 18 18 1.9 1.1-3.4

P < .001

Prognostic Score. Overall Survival (N = 1,307)O

vera

ll S

urv

ival, %

ASCO 2018, Press Briefing Presentation AM Tsimberidou, MD, PhD

Risk Factor (vs. other) HR 95% CI P

Non-Matched therapy 1.30 1.16-1.46 <.001

PAM alterations 1.25 1.10-1.42 <.001

Liver metastases 1.45 1.28-1.64 <.001

LDH > ULN 1.61 1.42-1.83 <.001

PS >1 2.12 1.71-2.64 <.001

Albumin < ULN 1.41 1.15-1.72 .001

PLT > ULN 1.54 1.16-2.04 .003

Age ≥ 60 yrs 1.14 1.02-1.28 .02CI, confidence interval; HR, hazard ratio; LDH, lactate dehydrogenase; PAM, PI3K/Akt/mTOR;

PLT, platelet count; PS, performance status; ULN, upper limit of normal

Multivariate Analysis, Overall Survival Therapy Added (N = 1,307)

ASCO 2018, Press Briefing Presentation AM Tsimberidou, MD, PhD

10

▪ Long-term overall survival was noted in the matched therapy group. The 3-yr

overall survival rate was 15% in the matched group compared to 7% in the

non-matched group. The 10-yr overall survival rate was 6% vs. 1%,

respectively.

▪ Matched therapy was an independent factor predicting longer survival in

multivariate analysis.

▪ PI3K/Akt/mTOR pathway abnormalities were associated with inferior

outcomes compared to other alterations.

▪ We developed a prognostic score for overall survival including molecular

pathway abnormalities.

IMPACT 1: Conclusions

ASCO 2018, Press Briefing Presentation AM Tsimberidou, MD, PhD

▪ Precision Medicine uses targeted therapy, immunotherapy, and other

strategies to target specific biological abnormalities causing

carcinogenesis in individual patients.

▪ Precision Medicine in cancer requires:

1. Complete understanding of tumor biology, including immune

features, that drives carcinogenesis

2. Use of effective drugs and therapeutic strategies that inhibit

carcinogenesis (rigorous definition)

3. Access to testing and effective drugs for all patients starting at

diagnosis and during the course of their disease

Implementation of Precision Medicine

ASCO 2018, Press Briefing Presentation AM Tsimberidou, MD, PhD

Precision Medicine in a Patient with Salivary

Cancer (BRAF V600E Mutation, Vemurafenib)

•The MD Anderson program pooled 1,144 patients in a phase I study after profiling their tumors for mutations that might be targets of the tested drugs.

•Apostolia Tsimberidou, the researcher who led the study, reported that 40% had mutations in 10 molecular pathways that were targeted by the experimental compounds.

•Tumors in 27% of those given agents that targeted their mutations responded to treatment compared to 5% for those with unmatched therapies.

THE WALL STREET JOURNAL June 6, 2011

Major Shift in War on Cancer

THE ECONOMIST June 9, 2011

Taking aim sooner

If personalized medicine is to achieve its full

potential, it should be used earlier on in clinical trials

Many scientists … believe that matching volunteers'

genetic profiles to the drugs being tested will not only

be better for the volunteers, but may also speed up the

trials, and save millions of dollars in the process.

One such is Apostolia-Maria Tsimberidou of the

University of Texas's MD Anderson Cancer Center, in

Houston. And her preliminary results, presented at a

meeting of the American Society of Clinical Oncology in Chicago, suggest she is right.

Challenges: Current Barriers

Actual Goal

Biopsy for molecular profile Not standard Standard of care

Tumor biology, markers Limited Complete

Bioinformatics Limited Optimized

Tumor heterogeneity Tumor/blood Validated cell-free DNA analysis

Drug discovery Limited More and effective drugs

Time to analysis 10-60 days 1-3 days

Clinical trial, drug available 5-30% of patients All patients

Timing (course of disease) Advanced, metastatic Starting at diagnosis

“Targeted drug” definition Imprecise Precise

Selection of optimal therapy Subjective Evidence-based, tumor board,

artificial intelligence

Adaptive learning, “N of 1” <10% 100%

Primary Objective

To determine whether patients treated with a targeted

therapy selected on the basis of mutational analysis of the

tumor have longer progression-free survival from the time of

randomization than those whose treatment is not selected

based on alteration analysis

PI: Tsimberidou, AM

www.clinicaltrials.gov NCT02152254

Supported in part by a research grant from Foundation Medicine

Randomized Study Evaluating Molecular Profiling and

Targeted Agents in Metastatic Cancer (IMPACT 2)

IMPACT 2. Study Design (I)

Metastatic disease

Tumor biopsy for molecular profiling, 100%

Targetable molecular aberrations (≥1 aberration)

Yes, 50% No, 50%

FDA-approved drugs within labeled indication

Yes, 30% No, 70%

Excluded; patient

followed for

progression but

not randomized

Is there a clinical trial

or commercially

available targeted

therapy?

Yes, 70%

Randomize

IMPACT 2. Study Design (II)

Targeted

therapy

Treatment

not selected

based on

molecular

analysis

1

1

CrossoverIf:

•Progressive disease

•Toxicity

0

50

100

150

200

250

300

350

400

450

PATIENTS ENROLLED IN IMPACT2

Cumulative plot of patients enrolled in IMPACT2

0

10

20

30

40

50

60

70

PATIENTS RANDOMIZED IN IMPACT2

Cumulative plot of patients randomized in IMPACT2

Genomic Alterations

FGF19 amplification

FGF4 amplification

FGF23 amplification

FGF3 amplification

FGF6 amplification

CCND1 amplification

CCND2 amplification

CDKN2A/B loss

CHD2 D213N

CREBBP R1392*

EMSY amplification

KDM5A amplification

KRAS amplification

MYC duplication exons 2-3

TP53 E204*

Head and Neck Squamous Cell Carcinoma with FGF Amplifications: CR to FGFR Inhibitor

Dumbrava I, … Tsimberidou, AM, JCO Precision Oncology – In Press

A C

DB

Head and Neck Squamous Cell Carcinoma with FGFAmplifications: CR to FGFR Inhibitor

Dumbrava I, … Tsimberidou, AM, JCO Precision Oncology – In Press

Mutation in Potentially

Actionable Gene

Underwent Genomic Testing

N = 2000

Genotype-matched trial after

genomic testing?

No (1211)

Genotype-Selected

Trial N = 54

Genotype-Relevant

Trial N = 29

Yes (789)

No (706)Yes (83)

Enrollment on Genotype-Matched Trials

11% of pts with mutations

in actionable genes went

on genotype-matched trials

Meric-Bernstam et al, JCO, 2015

Workflow: Development of Personalized Cancer Therapy Gene Data

Kurnit et al. Cancer Res 2017

1. Molecular profile is ordered as:

▪ Standard of care or

▪ For clinical trials: i.e. IMPACT2, NCI-MATCH, MPACT

▪ Interpretation of molecular profile:

▪ Precision Oncology Decision Support team

▪ Expert oncologists in precision medicine

▪ Selection and treatment on clinical trials is based on:

▪ Recommendation of tumor molecular board

▪ Clinical trial availability

▪ Patient preference and eligibility

▪ Study sponsor and insurance approval

2. Stringent regulatory CRC/IRB/DSMB review and trial

prioritization

Clinical Practice: Phase I Program, MD Anderson

ASCO’s Targeted Agent and Profiling Utilization Registry (TAPUR)

➢ To describe the anti-tumor activity and toxicity of commercially available, targeted therapy of patients with advanced cancer whose tumor harbors a genomic variant known to be a drug target.

➢ To learn from the real world practice of prescribing targeted therapies and to educate oncologists about implementation of precision medicine in clinical practice

Richard Schilsky, CMO and VP ASCO

Intra-tumor Heterogeneity: How to Resolve it?

Circulating free tumor DNA

Targeted Therapy and Immunotherapy

• Unresectable/metastatic, MSI-H or mismatch repair deficient

(dMMR) solid tumors that have progressed following prior

treatment or with MSI-H or dMMR colorectal cancer that has

progressed following treatment with a fluoropyrimidine,

oxaliplatin, and irinotecan.

• PD-L1–positive recurrent or advanced gastric or

gastroesophageal junction (GEJ) adenocarcinoma with ≥ 2

lines of chemotherapy, including fluoropyrimidine- and

platinum-containing chemotherapy, and, if appropriate,

HER2/neu-targeted therapy: KEYNOTE-059 study: 143 of

259 patients had PD–L1-positive tumors (combined positive

score ≥1), non-MSI high; ORR: 13.3% (CR 1.4%; PR 11.9%);

Duration of response: 2.8+ to 19.4+ months

Pembrolizumab: FDA Approval Based on Tumor Markers

Depth of response

Nivolumab in Mismatch-Repair Deficient Non-Colorectal Cancers: NCI-MATCH Trial:N= 34: PR, 24%; SD ≥ 2 months, 32%

Mutational Burden and ORR to PD1/PDL1

Inhibitors in Selected Tumor Types

Slide 4

Study of Immuno-Markers That Predict Response to Immunotherapy

Presented By Lisa Butterfield at 2016 ASCO Annual Meeting

Presented By Mary Disis at 2018 ASCO-SITC Clinical Immuno-Oncology Symposium

Hot and Cold Tumors

Screening Production phase Treatment/Observation Follow-up

NCT02876510, Immatics PI, AM Tsimberidou; Co-PI, Borje Andersson

ACTolog: Endogenous CD8+ T cells in Advanced Cancer

HLA phenotype HLA-A*02:01

Precision Medicine 2018

▪ Multiple alterations, complex molecular networks,

immune mechanisms, transcriptomic, proteomic and

epigenetic changes can be identified in individual

patients

▪ These markers should be integrated into clinical practice

to select optimal therapy

▪ Complexity of biomarkers is increasing

▪ Development of infrastructure is needed to use artificial

intelligence to integrate all available patient data to

perform algorithm analysis in decision making for

optimal drug selection, for More Effective Drugs, For

More Patients, Faster.

Acknowledgements

MD Anderson or WIN Leadership

• Dr. John Mendelsohn

• Dr. Razelle Kurzrock

• Dr. Richard Schilsky

• Dr. Patrick Hwu

Investigational Cancer Therapeutics

• Dr. Funda Meric, Chair

• Dr. David Hong

• Dr. Filip Janku

• Dr. Aung Naing

• Dr. Siqing Fu

• Dr. Sarina Piha-Paul

• Dr. Vivek Subbiah

• Dr. Jordi Rodon

• Dr. Timothy Yap

• Dr. Jennifer Wheler (former faculty)

• Dr. Gerald Falchook (former faculty)

Pathology

•Dr. Stanley Hamilton

•Dr. Russell Broaddus

Biostatistics

•Dr. Donald Berry (IMPACT 2)

•Dr. Jack Lee (IMPACT 1)

•Graciela Nogueras (IMPACT 1)

Institute for Personalized Cancer Therapy

Funding

•Donors: Alberto Barretto Alberto

Barretto, Jamie Hope, and Mr. and Mrs.

Zane W. Arrott (IMPACT 1)

•Foundation Medicine (IMPACT 2)

•Multiple pharmaceutical Companies

(individual clinical trials)

Patients and Families

Department of Investigational Cancer Therapeutics

Thank you!

atsimber@mdanderson.org