Dimitris Visvikis Director of Research National Institute of Health and Medical Research INSERM LaTIM UMR 1101 Brest France Cancer Oncology Gold standard for diagnosis ID: 540620
Download Presentation The PPT/PDF document "Multi-parametric multi-modality imaging ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Multi-parametric multi-modality imaging for prognostic and predictive modeling in oncology
Dimitris Visvikis
Director
of
Research
National Institute of
Health
and
Medical
Research
(INSERM),
LaTIM
, UMR 1101
Brest, FranceSlide2
Cancer
Oncology
Gold standard for diagnosis
Other applications of interest:
Radiotherapy planning
Prognosis, therapy assessment
PET/CT multimodality imaging
Quantification
uptake measurement
functional tumor volume
uptake distribution analysis
IntroductionSlide3
Activity concentration indices
Semi-quantitative index of concentration of activity in a given ROI:
ROI choise
Robustness
Reproducibility
Repeatibility
No ROI
Robustness
Reproducibility
Repeatibility
ROI fixed
Robustness
Reproducibility
RepeatibilitySlide4
Dynamic acquisitions:
Analysis of dynamic sequence of images
Arterial sampling
Mathematical modelling
Kinetic behaviour of FDG in a particular ROI
Input function
Parameters adjustment
Acquisition protocols Slide5
Other popular PET image derived
indices
Tumour / background Metabolically active tumor volume (MATV, cm3)Total lesion glycolysis (TLG = MATV ×
SUVmean)
SUV
max
SUV
peak
MATV
SUV
mean
TLGSlide6
Image noise
(differences in acquisition protocols)
Partial Volume effects
(spatial resolution)
Voxel size
(4-5 mm)
Tumour heterogeneity
Functional
volume segmentationSlide7
Nestle U et al., J Nucl Med, 2005
T
hresholding
approachesSlide8
Probability of observation
P(Y|X)
Spatial correlation probability
P(X)
1
2
3
4
5
6
7
8
9
[0 , 1]
(
μ
,
σ
)
tumor
(
μ
,
σ
)
Physiological background
(
μ
,
σ
)
Fuzzy transitions
Final map
Hatt M et al. IEEE TMI 2009; Int J Rad Onc Biol Phys 2010
FLAB:
Fuzzy
Localy
Adaptive
BayesianSlide9
Predictive
and prognostic value of functional volume for different cancer models Initial FDG PET scan
Better patient stratification and management (
survival, therapy response) Example: 45 esophageal cancer patient
Therapy response
Survival
Hatt et al EJNM 2010
Accurate
MATV:
prediction
,
prognosisSlide10
Lungs
Heart
Spinal Cord
PTV
FLAB
PTV
THRES
Accurate
MATV:
radiotherapy
Le Maitre A et al., Phys Med Biol, 2012Slide11
Tumor characterisation: geometric forms
Hypotheses: associated with tumor aggresivity, metastasis potential…
morphological, functional and/or morpho-functional:Form descriptors(a)sphericity, solidity, convexity, rectangularity, excentricity…Intra-tumor PET activity distribution form : independent prognostic value demonstrated onH&N cancer1Lung cancer2Sarcoma3
CT
PET - FDG
Fusion PET/CT
Apostolova I,
et al
.
BMC Cancer
. 2014
Hofheinz F,
et al
.
Eur J Nucl Med Mol Imaging
. 2015
Eary J,
et al.
J Nucl Med 2008Slide12
18
FLT PET during chemo-radiotherapy
1
Hoeben
BA,
et al
. J Nucl Med. 2013Majdoub M, et al. (submitted) 2016
High sphericity
and low SUV
Low sphericity
et high SUV
Time (months)
12
24
36
4860
72Survival probability (%)
Predictive value of sphericity
2
Hazard ratio = 6,7 (p<0,0001)
(with SUV only HR = 4,1, p=0,02)
(with sphericity only HR = 4,2, p=0,01)
Tumor
characterisation
: geometric formsSlide13
Activity conc, volume, … Global mesures
Intra-tumor activity distribution heterogeneity characterisation
Tumor
heterogeneity
characterisationSlide14
Textural features can quantify different types of voxels intensity variability in the tumor volume, at different scales
M.
Tixier F et al, J Nucl Med, 2011
Tumor
heterogeneity
characterisationSlide15
Jensen. Introductory Image Processing 3rd ed. Upper Saddle River, NJ: Prentice-Hall 2005
Histogram analysis
No spatial information
Co-occurrence
Matrices
Local spatial information
Regional spatial information
Complexity
Versatility and potential
Tumor
heterogeneity
characterisationSlide16
PET/CT image derived features
Parameters for tumor heterogeneity characterisation
Textural features can quantify different types of voxels intensity variability in the tumor volume, at different scales.
Chicklore, et al.
Quantifying tumour heterogeneity in 18F-FDG PET/CT imaging by texture analysis. Eur J Nucl Med Mol Imaging 2013
First-order parameters:
a = b = c = d
Second-order features:
a
# (
b
=
c
= d).Third-order features:a # b #
c #
dSlide17
Heterogeneity
quantification
1. Tumor delineation [1-3]
[1] Hatt, et al. IEEE Trans Med Imaging. 2009;28:881-893
[2] Hatt, et al.
Int J Radiat Oncol Biol Phys. 2010;77(1):301-8[3] Hatt, et al. Eur J Nucl Mol Imaging. 2011;38:663-672
Co-occurrence matrix
Size-zone matrix
2. Heterogeneity quantification using textural features [4-6]
[4] Tixier, et al. JNM. 2011;52:369-378
[5] Tixier, et al. JNM. 2012;53:693-700
[6] Hatt, et al. EJNM. 2013;40:1662-1671
2. Quantization Slide18
M.
Tixier F et al, J Nucl Med, 2011
Tumor
heterogeneity
characterisationSlide19
Example
: NSCLC
NSCLC without metastasis (n=100)Stade I (n=18), II (n=29), III n=(53)Males (n=78), Females (n=22)Age 64±9Treatment :Surgery (n=47)Chemotherapy (n=82)Radiotherapy (n=51)
18
19
9
12
41
1
0
Surgery
Chemo
Radio
Tumor
heterogeneity
characterisationSlide20
12
24
36
48
60
3 year survival ~ 30%
Médian 18 months
Global survival
Time (months)
Survival probabiity (%)
Tumor
heterogeneity
characterisationSlide21
12
24
36
48
60
Stade I : médiane -
Stade III : médian 18 months
p = 0.008
Stade II : médian 21 months
Time (months)
Survival probability (%)
Tumor
heterogeneity
characterisationSlide22
Multimodality tumor characterisation
Complimentarity
: functional volume – PET heterogeneity
M. Hatt,
et al. J Nuc Med 2015
N=34, 17 deaths (50%)
Median 45 months
N=38, 25 deaths (66%)
Median 21 months, HR=2.3
N=29, 27 deaths (93%)
Median 9 mo, HR=3.8
Probabilité de survie (%)
Time (months)
Volume PET (cm
3
) logarithmic scale
Entropy
1
2
3
4Slide23
PET/CT image derived features
Heterogeneity characterization
What about CT or other modalities (MRI)?Several studies have investigated the use of textural features or other metrics to characterize CT or MRI tumor volumes1
Colon cancer
Contrast CT1. Davnall, et al. Assessment of tumor heterogeneity: an emerging imaging tool for clinical practice? Insights Imaging 2013Slide24
PET/CT image derived features
Heterogeneity characterization
What about CT or other modalities (MRI)?
Rectal cancer
T2-weighted MRI1. Davnall, et al. Assessment of tumor heterogeneity: an emerging imaging tool for clinical practice? Insights Imaging 2013Slide25
PET/CT image derived features
Heterogeneity characterization
What about CT or other modalities (MRI)?
Pre-treatment (baseline)
Post-treatment (neoadjuvant chemo)
1. Davnall, et al.
Assessment of tumor heterogeneity: an emerging imaging tool for clinical practice? Insights Imaging 2013
Esophageal cancer - CTSlide26
Multi-parametric & Multi-modality
Prognostic
model PET/CT (116 NSCLC patients)Stage, PET volume and heterogeneity, CT heterogeneity
Desseroit
et al Eur J Nucl
Med 2016Slide27
Perspectives
Future disease characterization with imaging
Multi tracers
MRI T2, diffusion, elastography…
Metabolism, hypoxia, proliferation
Multi modalities
PET / CT
PET / MRI
Temporal data
Therapy follow-up, adaptive
Tracer kinetics
Full tumor « signature » in the multi modal spectrumSlide28
Perspectives
Actual
ParadigmProposed Paradigm
Ex.
Astronomy
images
(3 observations)
FusionSlide29
Perspectives
Future disease characterization
Combination and correlation of multimodal tumor signatures with other informationInformation fusion → Building predictive modelsSlide30
Conclusions
PET can provide
useful information for early response assessmentextract more information from PET requiresstandardization, image enhancement accurate, robust, reproducible automated delineationnew features such as texture or shapeIn the future:Multi-center trials and large patient cohortscombination of omics data with multimodal anatomo-functional characterization (tumor “signature”)→ efficient predictive models for improved patient managementSlide31
Thank you for your attentionSlide32Slide33
PET tumor heterogeneity: reproducibilitySlide34
34
M. Hatt,
et al
.
J Nuc Med 2015
Corrélation = 0.98
Corrélation = 0.93
Corrélation = 0,56
Local entropy (co-occurrence matrix)
Volume (log, cm
3
)
Quantization
= 128
13 matrices
+ mean
Quantization
= 64
13 matrices
+ mean
Quantization
= 64
1 matrice
PET tumor heterogeneity: minimum volumeSlide35
PET tumor heterogeneity
Method
Textural features analysis Validation Reproducibility, robustness vs. reconstruction, partial volume effects, tumor delineation [6-8] ApplicationPrediction of therapy response/patient’s outcome from baseline scan[1-5] What does PET heterogeneity represent?Hypothesis: associated with tumor physiology (glucose metabolism, hypoxia, angiogenesis, …)
[1]
Tixier F et al. J Nucl Med. 2011; 52: 369-378[2] George J et al. ISBI 2012[3] Cook GJR et al. J Nucl Med. 2013; 54: 1-8[4] Willaime JMY et al. Phys. Med. Biol. 2013; 58: 187-203[5] Huang B et al. AJR Am J Roentgenol. 2012; 199: 169-174[6] Galavis PE et al. Acta Oncologica. 2010; 49: 1012-1016
[7]
Tixier F et al. J Nucl Med. 2012; 53: 693-700[8] Hatt M et al. Eur J Nucl Med and Mol Imaging. 2013 ahead of print