Clinical Imaging Hypoxia

Imaging of hypoxia

Philippe Lambin

Octobre 12th 2012

This course is funded with the support of the METOXIA project under the FP7 Programme.

Learning objectives

1. Know the advantages and disadvantages of hypoxia imaging

2. Be able to name the various hypoxia imaging modalities with some basic features

Advantage of imaging: 3D & non invasive

Advantageous compared to IHC, gene-miRNA signatures because tumours are heterogeneous in space

Courtesy of B. van der Kogel

4

Disadvantage: The oxygen distribution inside a voxel

Measured voxel pO2 is average pO2 of the cells inside it.

Each voxel contains a mixture of cells at different oxygen concentrations.

Ignoring this will affect the correct choice of dose.

Why should we do it?

For prognosis (Give prognosis of disease outcome regardless of therapy)

For prediction (Predict outcome with a specific therapy & allow to adapt treatment (drugs, boost BTV…)

As endpoint (secondary or primary: e.g. reduction of hypoxia with nitroglycerine patch, pattern of relapse…)

For research purposes (no immediate clinical benefit)

How to measure tumor oxygenation ?

Non Imaging methods

• Polarographic measurements• Eppendorf®

• Optical measurements• Phosphorescence• Fluorescence (OxyLite ®)

• Hypoxia markers• Nitroimidazoles

• Immunohistochemistry

• Gene signatures

• Blood biomarkers

Imaging methods

• Hypoxia markers• Nitroimidazoles PET/SPECT• CA9 ligands PET/SPECT

• MR• 19F relaxation• BOLD• DCE-MRI• 1H relaxation

• EPR-related methods

B. Gallez, NMR Biomed. 2004, 17, 240

How to measure tumor oxygenation ?

Real oxygenmeasurements

• Polarographic measurements• Eppendorf®

• Optical measurements• Fluorescence (OxyLite ®)

• EPR-related methods

• MRI• 19F relaxation

Oxygen-sensitivemeasurements

• NMR• DCE-MRI• BOLD• 1H relaxation

• Hypoxia markers• Nitroimidazoles

• Immunohistochemistry• PET/SPECT

• CA9 ligands PET/SPECT

• Gene signatures

• Blood biomarkers

How to measure tumor oxygenation ?

InvasiveOr not immediately

clinically applicable

• Polarographic measurements• Eppendorf®

• Optical measurements• Fluorescence (OxyLite ®)

• EPR-related methods

• MRI• 19F relaxation

ImagingClinically applicable

• NMR• DCE-MRI• BOLD• 1H relaxation

• Hypoxia markers• Nitroimidazoles PET/SPECT• CA9 ligands PET/SPECT

Used to qualify

the clinically applicable

oxygen-sensitive

techniques

Real oxygen

measurements

Ideal clinical oxygen imaging modality

• Able to distinguish normoxia/hypoxia/anoxia/necrosis

• Able to distinguish between perfusion-related and diffusion-related hypoxia (sensitive to changes of hypoxia)

• Able to reflect cellular oxygenation in preference to vascular oxygenation

• Be applicable to any tumor site

• Simple to perform, non toxic and allowing repeated measurements

• Sensitive at pO2 relevant to tumor therapies

• Results independant of the timing of imagingA. Padhani, Eur. Radiol. 2007, 17, 861.

EPR (Electron Paramagnetic Resonance) Oximetry

O2 dependent broadening of the EPR

linewidth (LW) of a paramagnetic O2

sensor implanted in the tumor

A particular material can be calibrated in terms of the effect of oxygen on the LW

When introduced in vivo, the measurement of LW can be interpreted in terms of oxygenation in the vicinity of the probe

3168 3318 3468

Magnetic Field (G)

0

10

20

30

40

0 7 14 21

% O2

LW

(G

)

air

nitrogen

B. Gallez, NMR Biomed. 2004,17, 240

Non-invasive imaging of chronic and cycling tumour hypoxia in xenografts with

EPR

Yasui et al., 2010, Can Res, 70:6427-6436.

EPR (Electron Paramagnetic Resonance) Oximetry

• Absolute measurement of pO2

• High sensitivity at low pO2: variations

lower than 1 mm Hg can be measured

• Minimally invasive: few microparticles should be introduced in the tissue (invasive only the first time)

• Measurements can be repeated from the same site over long periods of time (hours, days, months)

• 1GHz: penetration depth of 1 cm

• Not applicable immediately into the clinic: pionneer clinical studies ongoing in Dartmouth Medical School on superficial tumors

N. Khan, Antiox. Redox. Signal. 2007, 9, 1169

19F-relaxometry

Mason RP et al, IJROBP 1998, 42, 747; and following works of Mason’s group

• Intratumoral injection of PFC (i.e. hexafluorobenzene, single 19F line)• Measurement of the R1 relaxation of 19F relaxation that is strongly dependent on pO2

• Maps of pO2

19F-relaxometry

Rapid estimation of R1 using SNAP-IR sequence

Simultaneous monitoring using 19F-MRI and fiber optic probes

B. JordanMRM 2009, 61, 634-638

19F-relaxometry

Problem to inject the PFC in the

entire tumor

The technique is likely to slightly overestimate pO2

Not applicable for chronic purposes (only acute studies)

Toxicity concerns with some PFCs

Lack of translation into the clinic (invasiveness, 19F coils)

Rather sensitive method

Quantitative method: pO2

values

No 19F NMR background signal in tissues

Good temporal & spatial resolutions for O2 mapping

DCE-MRI

Attempt to correlate DCE-MRI parameters with tumor hypoxia

(Gribbestad., Dynamic Contrast-Enhanced Magnetic Resonance Imaging in Oncology, 2006)

Ktrans significantly higher for radiation sensitive tumors

without hypoxia than for radiation resistant

tumors with hypoxic regions

K Gullisksrud, Radiother. Oncol. 2011, 98, 360

Dynamic MRI of cervix carcinoma

C DModified from:J. Barentsz

A. T2-weighted spin echo imageB. [Gd] max imageC. [Gd] time-to-peak image

3

2

a

4

1A B

C Dt

[Gd] 123

4

a

Tumor Oxygenation

Perfusion Oxygen consumption

High O2 consumption rate by tumor cells

Cancer cells consume oxygen at high rate, even if they generally present a highly glycolytic metabolism

This high consumption rate significantly contributes to the tumor hypoxia resulting from the imbalance between oxygen delivery and oxygen consumption

Strategies to decrease the oxygen consumption by tumor cells and potentiate radiotherapy

• Meta-iodobenzylguanidine JE Biaglow, IJROBP 1998, 42, 871

• Nitric oxide donors B. Jordan, Int. J. Cancer 2004, 109, 768

• Insulin B. Jordan, Cancer Res. 2002, 62, 3555

• NSAIDs N. Crokart, Cancer Res. 2005, 65, 7911

• Glucocorticoids N. Crokart, Clin. Cancer Res. 2007, 13, 630

• SU5416, ZD6474 R. Ansiaux, Cancer Res. 2006, 66, 9698; Radiat. Res. 2009, 172, 584

• Propylthiouracil B. Jordan, Radiat. Res. 2007, 168, 428

• Arsenic trioxide C. Diepart, submitted

Theoretical simulations:

To alleviate tumor hypoxia, decreasing the O2 consumption rate of tumor is more effective than increasing

oxygen deliverySecomb, Acta Oncol. 1995 34, 313

Pre-clinical studies

DCE-MRI

Trend of perfusion

parameters to differentiate between hypoxic vs normoxic tumors

Clinically usable

No absolute values of pO2

It is unlikely that a treshold value of a DCE-MRI parameter will predict the radiosensitivity

Tumor perfusion is not the main or sole factor that is responsible for the tumor oxygenation (oxygen consumption)

BOLD-MRI – R2*

Oxy-Hemoglobin: diamagneticDeoxy-Hemoglobin: Paramagnetic

S. Ogawa et al, PNAS 1990, 87, 9868

Carbogenbreathing

From fMRI …… to tumor oxygenation

GS Karczmar, NMR Biomed 1994, 12, 881SP Robinson, IJROBP 1995, 33, 855F Howe, MRI 1999, 17, 1307

Level of blood oxygenation DeoxyHb /OxyHb blood content SI

BOLD-MRI

T2* is sensitive to the relative Hb/HbO2 ratio in vessels:

Blood oxygen saturationHematocrit

Blood volume

Basal R2* and tumor oxygenation

Correlation between pimonidazole uptakeand high R2* in prostate cancer

PJ Hoskin, IJROBP 2007, 68, 1065

Inverse correlation between pimoninazoleuptake and R2* values in mammary tumors

McPhail, Radiology 2010, 254, 110

R2*: phenotype specific ?Requires simultaneous measurements of vasculature function ?

AR Padhani, Radiology 2010, 254, 1

Change in R2* and change in tumor oxygenation

T2*w SI T2*

Local

Whole tumor

pO2

BOLD signal response correctly reflected the evolution of tumor oxygenation in carbogen challenge

BOLD signal

C. Baudelet and B. GallezMagn.Reson. Med.

2002;48:980.

Comparison with OxyLite: simultaneous measurement of R2* and pO2

Tumor Oxygenation

Perfusion Oxygen consumption

Change in R2* and change in tumor oxygenation

Increase in pO2 through changes in O2 consumption

Insulin infusion Control NS-398

Changes in BOLD signal and R2* in tumors do not depend uniquely on changes in oxygenation status

B. Jordan, MRM 2006, 56, 637

BOLD-MRI and R2*

Variations in tumor oxygenation

can be qualitatively measured during carbogen breathing

Rapid dynamic measurement

Clinically usable

No absolute values of pO2

The value of basal R2* is

debatable

Variation in R2* cannot predict

changes in oxygenation induced by treatments modulating the oxygen consumption

T1-based measurements

Carbogenbreathing

O2 O2

O2

O2 O2

O2

H20H20

H20

H20H20

O2 O2

H20 H20

H20H20

H20

Dissolved oxygen acts as a T1-shortening paramagnetic contrast agent

Oxygen produces changes in relaxation rate R1 of water

H20

H20

KI Matsumoto, MRM 2006, 56, 240

air

carb

ogen-5

0

5

10

15

20

R1 (H2O)

R1 (Lipids @ ~ 3.5 ppm)

rela

tive

ch

ang

e (%

)

0 50 100

0

10

20

30

40

50

carbogen

postmortem

air

time (min)

pO

2(m

mH

g)

Pooled resultsN=5

T1-based measurementsMOBILE

Mapping of Oxygen By Imaging Lipids Relaxation Enhancement

air carbogen postmortem

-0.4

-0.2

0.0

0.2 R1 (H2O)

R1 (Lipids @ ~ 3.5 ppm)

rela

tive

ch

ang

e (%

)

Nitro-imidazoles mechanism of uptake

• Initial distribution is flow dependent• Local oxygen tension = main determinant for long-term retention

• NO2 reduction in radical anion;

• if O2 is present, back to the original structure

• in absence of O2, second reduction with product binding to macromolecules

• depends on the nitroreductase activity• generally assumed to have retention under 10 mm Hg

RNO2 RNO2

necrosisRNO2

-e- reduction

normoxia

RNH2

e- reduction

hypoxia

retentionmetabolites

vascular space cellular compartment

Nitroimidazole-based PET: How does it work?

Rat rhabdomyosarcoma: optimizing imaging conditions with HX4

0 1 2 3 4 5 6

0

2

4

6

8

Time (hours) after injection

max

Tiss

ue to

Blo

od ra

tio

*

**

*

NS NS

4h p.i.

Dubois et al. PNAS 2011

Is there a significant correlation between pimonidazole staining (“the gold standard”) and HX4 uptake?

Validation of 18F-HX4 uptake using IHC (setup)

region selection based on CT %HX4 per region

DORSAL

HEAD

TAIL

VENTRAL

a b c d

ab

cd

Dubois et al. PNAS 2011

Validation with pimonidazole IHC: first results

Courtesy of B. van der KogelDubois et al. PNAS 2011

Validation of 18F-HX4 uptake using IHC (results)

n = 76

0.07330.6627III-3

0.00020.8202II-3

< 0.00010.9046I-2

0.00550.6588III-2

0.00280.7334III-1

< 0.00010.7222regions

0.28480.6000total

p-valueSpearman R

%PIMO vs %HX4 summary

Dubois et al. PNAS 2011

Is there a causal relationship between hypoxia and HX4 uptake?

[18F]HX4 accumulation is oxygen dependent

Basal scan Carbogen/Nicotinamide

7% oxygen breathingBasal scanDubois et al. PNAS 2011

Which hypoxic biomarker is the best?

Hypoxia PET tracers

18F-FMISO 18F-FAZA 18F-HX4

Nitro-imidazoles:

• Clearance:

• Hydrophilicity:

Liver –intestine – kidney

Kidney – intestine – liver

Kidney – bladder

Positron-emitting

radionuclide

Hypoxia sensitive

partRest group

Tumor to blood

S. Peeters et al. In preparation

Imaging HypoxiaPrognostic value of outcome

F-MISO PET in 73 patientswith H&N cancer

FMISO Tumor/Bloodis a prognostic measure

of the outcome

JG Rajendran, Clin. Cancer Res. 2006, 12, 5435

Correlation also found in:

FMISO PET in 40 patients with H&N cancerSM Eschmann, J. Nucl. Med. 2005, 46, 253

No correlation found in:

FMISO PET in 20 patients with H&N cancerNL Lee, IJROBP 2009, 75, 201

Copyright © American Society of Clinical Oncology

Rischin, D. et al. J Clin Oncol; 24:2098-2104 2006

(A) Baseline [18F]-fluorodeoxyglucose (FDG) positron emission tomography (PET) of patient with T2N2b squamous cell carcinoma of the pyriform fossa with left nodal mass

(A) Baseline [18F]-fluorodeoxyglucose (FDG) positron emission tomography (PET) of patient with T2N2b squamous cell carcinoma of the pyriform fossa with left nodal mass.

(B) (B) [18F]-fluoromisonidazole (FMISO) -PET at baseline, nonhypoxic primary tumor, and hypoxic node.

(C) C) FDG-PET 12 weeks after chemoboost, complete response in nonhypoxic primary tumor, and poor response in hypoxic node. Residual tumor in nodal mass was confirmed pathologically after neck

dissection.

Phase 1 HX4 imaging: clinically feasible?Van Loon et al, EJNM 2010

CT delineated tumor [18F]HX4 accumulation

0

0,4

0,8

1,2

1,6

30 60 120

Time (min) after injection

T/M

Van Loon J. et al. EJNM 2010

NSCLC Stage IIIB

18HX4-PETCT

Van Loon et al. Eur J Nucl Med Mol Imaging. 2010

Copyright © American Society of Clinical Oncology

Rischin, D. et al. J Clin Oncol; 24:2098-2104 2006

Time to local failure (Kaplan-Meier method) by treatment arm and hypoxia in the primary tumor (censored times are indicated as tick marks on the curves)

Zips et al. Radiother Oncol 2012

Zips et al. Radiother Oncol 2012

Zips et al[18F]Misnidazole

0 Gy 10 Gy 20 Gy 40 Gy

[18F]Misonidazol

[18F]FDG

Zips , Kotzerke, Baumann et al.

Biomarker: Hypoxia (F-MISO PET)

Department of Radiation Oncology M. Baumann |Regaud Lecture 2012

Radiolabelled nitroimidazoles

Hypoxia-sensitive method

Relevant to estimate radioresistant relevant hypoxia (< 10 mm Hg)

Clinically usable

Prognostic value (especially if repeated)

No absolute values of pO2

Poor correlation with pO2 values

for some tracers

Accumulation dependent on the level of nitroreductases

Imaging of tumor acute hypoxiaSpontaneous fluctuations in tumor oxygenation/perfusion

Technique Spatial resolution

Temporal resolution

Characteristics Reference

NITRO-PET 4.2 mm Day 1,2,5 Hypoxia (More or less

than 10 mm Hg)

Wang,Med. Phys.

2009, 36, 4400

DCE-MRI 0.5x0.2 mm3 15 min Flow variation Brurberg,MRM

2007, 58, 473

T2*w-GE MRI 470x470 µm2 12.8s Oxygen/flow variation

Baudelet, Phys. Med. Biol. 2004, 49, 3389

19F-MRI 1.88 mm 1.5 min Quantitative pO2

Magat,Med. Phys.

2010, 37, 5434

EPRI 1.8 mm 3 min Quantitative pO2

Yasui,Cancer Res.

2010, 70, 6427

*P < 0.05; **P < 0.01

In vivo CAIX specific sulfonamide accumulation is reversible upon reoxygenation

Dubois et al, Radiother & Oncol 2009

Dubois et al, Radiother & Oncol 2009

PET images of mice with tumor located subcutaneously on right hind leg at 4 (A) and 24 h (B) after injection. p.i. = after

injection; SUV = standardized uptake value.This course is funded with the support of

the METOXIA project under the FP7 Programme.

MAb-N-succinyldesferal-89Zr (MAb-N-sucDf-89Zr) (*)

Imaging of CAIX with antibody

Hoeben, Kaanders et al. 2010 Jul;51(7):1076-83.

How to include hypoxia imaging in the clinic?

Exploiting intra patient heterogeneity for dose painting of radiation

Intensity modulated radiation therapy (IMRT)

Possibility to sculpt the doses as a function of possible needs

Galvin, J. Clin. Oncol. 2007, 25, 924

Hypoxia Guided IMRT: dose escalation in hypoxic areas

Dose Painting by Contours

Dose Painting by Numbers

0.4

0.6

1.1

1.9

2.7

2.9

Proof-of-concept using Cu-ASTM

KSC Chao, IJROBP 2001, 49, 1171

Theoretical feasibilityusing 18F-FMISO

Dose prescription based on tumor hypoxia

D. Thorwarth, IJROBP 2007, 68, 291Z. Lin, IJROBP 2008, 70, 1219NY Lee, IJROBP 2008, 70, 2

I. Toma-Dasu, Acta Oncol. 2009, 48, 1181W. Choi, Radiother. Oncol. 2010, 97, 176

Hypoxia Guided IMRT: dose escalation in hypoxic areas

Dose Painting by Contours

Dose Painting by Numbers

0.4

0.6

1.1

1.9

2.7

2.9

Proof-of-concept using Cu-ASTM

KSC Chao, IJROBP 2001, 49, 1171

Theoretical feasibilityusing 18F-FMISO

Dose prescription based on tumor hypoxia

D. Thorwarth, IJROBP 2007, 68, 291Z. Lin, IJROBP 2008, 70, 1219NY Lee, IJROBP 2008, 70, 2

I. Toma-Dasu, Acta Oncol. 2009, 48, 1181W. Choi, Radiother. Oncol. 2010, 97, 176

Conclusions

• To bridge the gap between hypoxia-induced radioresistance and optimized radiotherapeutic treatment with drugs

– Oxygenation imaging is mandatory

– Qualification of oxygenation biomarkers is still mandatory at the pre-clinical and the clinical level

– There is a crucial need to validate the value of hypoxia-gimaging inprospective trials with interventions

ISMRM 2011: Clinical Needs and Research Promises

AcknowledgementsUniversity of Nijmegen (The Netherlands)

– Albert van der Kogel– Jan Bussink– Hans Kaanders

University of Amsterdam (VUmc)– Guus van Dongen– Bert Windhorts– Jonas Eriksson

University of Florence (Italy)– Andrea Scozzafava– Claudiu Supuran

University of Brussels (UCL)– Bernard Gallez*– Vincent Grégoire

Our patients (No immediate benefit for them)

Euroxy-Metoxia 6th & 7th Framework

NIH (USA)

Siemens MI

University of Maastricht– Ludwig Dubois *– Judith van Loon – Sarah Peeters – Karen Zeghers