Purpose/Objective(s)
Traditional methods of evaluating cardiotoxicity focus solely on radiation doses to the heart and do not incorporate functional imaging information. Functional imaging has great potential to improve the ability to provide early prediction for cardiotoxicity for lung cancer patients undergoing radiotherapy. FDG-based PET/CT imaging is routinely obtained as part of standard staging work up for lung cancer patients. Although FDG PET/CT scans are typically used to evaluate the tumor, imaging guidelines note that FDG PET/CT scans are an FDA-approved method to image for cardiac inflammation, and studies have noted that the PET cardiac signal can be predictive of clinical outcomes. The purpose of this work was to develop a radiomics model to predict clinical cardiac assessment of standard of care FDG PET/CT scans.
Materials/Methods
The study included 100 consecutive lung cancer patients treated with radiotherapy who underwent standard pre-treatment FDG-PET/CT staging scans. A clinician reviewed the PET/CT scans per clinical cardiac assessment guidelines and classified the cardiac uptake as: 0 = uniform diffuse, 1 = absent, 2 = heterogeneous, with event rates of 20%, 44%, and 35%, respectively. The heart was delineated and 200 novel functional radiomics features were selected to classify cardiac FDG uptake patterns. We divided the data into an 80% training set and a 20% test set to train and evaluate the classification models. Feature reduction was carried out using the Wilcoxon test (with Bonferroni adjusted p<0.05), hierarchical clustering, and Recursive Feature Elimination. Two automatic machine learning (AutoML) frameworks were used to determine classification models: a Random Forest Classifier (Tree-based Pipeline Optimization Tool, TPOT) and Linear Discriminant Analysis (AutoSklearn). 10-fold cross validation was carried out for training and the accuracy of the ability of the models to predict for clinical cardiac assessment is reported.
Results
Fifty-one independent radiomics features were reduced to 3 clinically pertinent features (PET 2D Skewness, PET Grey Level Co-occurrence Matrix Correlation, and PET Median) using feature reduction techniques. The model selected by TPOT showed 89.8% predictive accuracy in the cross validation of the training set and 85% predictive accuracy on the test set. The model selected by AutoSklearn showed 89.7% predictive accuracy in the cross validation of the training set and 80% predictive accuracy on the test set.
Conclusion
The novelty of this work is that it is the first study to develop and evaluate functional cardiac radiomic features from standard of care FDG PET/CT scans with the data showing good predictive accuracy with clinical imaging evaluation. If validated, the current work provides automated methods to provide functional cardiac information using standard of care imaging that can be used as an imaging biomarker for early clinical toxicity prediction for lung cancer patients.
Call Girl Hyderabad Madhuri 9907093804 Independent Escort Service Hyderabad
Novel Functional Radiomics for Prediction of Cardiac PET Avidity in Lung Cancer Radiotherapy
1. Novel Functional Radiomics for
Prediction of Cardiac PET Avidity
in Lung Cancer Radiotherapy
W. Choi1, Y. Jia1, J. Kwak2, A. P. Dicker1, N. L. Simone1, E. Storozynsky3, V. Jain1, and Y. Vinogradskiy
1Department of Radiation Oncology, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia,
PA,
2University of Colorado School of Medicine, Aurora, CO,
3Department of Cardiology, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, PA
2. • Thomas Jefferson University
• A research grant from ViewRay, Inc.
• A research grant from the PROPEL Center
• Research SW license for INT Contour from Carina Medical
2
Disclosure
3. • Traditional methods of evaluating cardiotoxicity:
Focus solely on radiation doses to the heart.
Do not incorporate functional imaging information.
• Functional imaging has potential to improve cardiotoxicity prediction for
lung cancer patients undergoing radiotherapy:
FDG PET/CT imaging is routinely obtained for lung cancer staging.
PET scans can be used to evaluate cardiac inflammation.
Studies indicate that PET cardiac signal can predict clinical outcomes.
• The purpose of this work:
Develop a radiomics model to predict clinical cardiac assessment using FDG PET/CT
scans.
Introduction
3
4. • The study included 209 pre-treatment 18F-FDG PET-CT images of lung cancer patients from 3
different study populations (TJU: 100, ACRIN: 70, CU: 39).
Low quality images were excluded 7 (TJU: 1, ACRIN: 4, CU:2).
Dataset
CLASS
LV MYOCARDIAL
UPTAKE PATTERN
TJU ACRIN CU TOTAL
0 Uniform 19 (19.2%) 21 (31.8%) 11 (29.8%) 51 (25.2%)
1 Absent 44 (44.4%) 23 (34.8%) 13 (35.1%) 80 (39.6%)
2 Heterogeneous 36 (36.4%) 22 (33.3%) 13 (35.1%) 71 (35.1%)
ALL 99 (100.0%) 66 (100.0%) 37 (100.0%) 202 (100.0%)
Table 1. Statistics of LV FDG Uptake Pattern Classification
Figure 2. Cardiac uptake classifications, Uniform, Absent, and Heterogeneous (Non-uniform and Focal )
4
5. 5
Method - Study Flow Diagram
Figure 1. Study and cardiac model diagram, training dataset: TJU (N=99), and external validation data sets: ACRIN (N=66) and CU (N=37).
RFE: Recursive Feature Elimination, SFS: Sequential Feature Selection
1. Pre-processing
LV FDG uptake patter
review
By Radiation Oncologist
Manual
Heart Contouring
By Medical Physicist
Automatic
Heart Contouring
By 1Commercial Software
Training
Data set
(N=99)
2. Feature Analysis
Feature Extraction
- PyRadiomics
- In-house SW
Feature Reduction
- Wilcoxon test
- Clustering
Feature Selection
- RFE
- SFS
3. Model Building
Model Exploration
- TPOT
Optimization
- Hyper-parameters
- Feature Selection
Evaluation
- Validation
- External validation
Predicted
LV Uptake
Pattern
External
Validation
Data sets
(N=103)
1INTContour, Carina Medical https://www.carinaai.com/intcontour.html
6. • Training and Validation Data:
Training dataset: TJU (N=100), 80% training set (TPOT optimization), 20% validation set.
External validation datasets: ACRIN (N=70), CU (N=39)
• Radiomics Feature Extraction
Heart delineation: Manual and Auto
Selected 200 novel functional radiomics features – In-house: 103, PyRadiomics: 97
Feature categories include Shape (2D and 3D), First-order (2D and 3D), GLCM, GLSZM, GLRLM, NGTDM,
GLDM.
• Feature Reduction
Feature robustness test using ICC between features from Auto and Manual delineations
Wilcoxon test (Bonferroni adjusted p<0.05)
Hierarchical clustering
Method – Radiomics Feature Extraction and Reduction
6
7. • TPOT (Tree-Based Pipeline Optimization Tool)
100 populations, 10 generations
Evaluated 10,000 pipeline configurations to identify the best model.
• Pipeline Construction
Three component template: Feature Selector—Feature Transformer—Classifier
o Feature Selector: P-value (ANOVA F-statistic), Family Wise Error Rate, Percentile, Low-Variance Feature Removal, RFE
with Extra Trees Classifier
o Feature Transformer: Binarization, ICA, Feature Agglomeration (Various Linkage & Affinity Methods), Scaling (Max
Absolute Value, Min-Max Values), Normalization (l1, l2, max norm) ,Kernel Approximation (RBF, Cosine, χ2, etc.), PCA,
Polynomial Features, RBF Sampler, Robust Scaler, Standard Scaler, Zero Count, One Hot Encoder
o Classifier: Naïve Bayesian (Bernoulli, Multinomial), Decision Tree Classifier, Extra Trees Classifier, Random Forest
Classifier, Gradient Boosting Classifier, XGBoost, KNN Classifier, Linear SVM Classifier, Logistic Regression, SGD, MLP
• Hyper-parameter Optimization: optimize parameters of the discovered pipeline
• Feature Optimization: Utilized RFE for feature ranking followed by SFS with TPOT optimization.
Method - Machine Learning Model Pipeline Selection and Optimization
7
8. 8
Results: Pipeline Optimization using TPOT
Feature Selector
Variance Threshold
Threshold: 0.0001
56 features
Threshold: 0.005
52 features
RFE ranking and
SFS with TPOT optimization
Top 9 features
Feature Transformer
One Hot Encoder
Minimum_fraction: 0.15,
Sparse: False,
Threshold: 10
Minimum_fraction: 0.25
Minimum_fraction: 0.15
Classifier
Extra Trees Classifier
Bootstrap: False,
criterion: gini,
max_features: 0.95,
min_samples_leaf: 3,
min_samples_split: 20
Criterion: entropy,
max_features: 0.75,
min_samples_leaf: 2,
min_samples_split: 17
Bootstrap: True,
criterion: gini,
max_features: 0.8,
min_samples_split: 10
Model
Discovery
Hyper-
parameter
Optimization
Feature
Optimization
Validation Accuracy (%)
TJU ACRIN CU
90.0 78.8 90.2
90.0 78.8 91.9
95.0 80.3 91.9
9. Results – Feature reduction and Prediction results
Model iteration and explanation
Number of
features
10-fold cross
validation
accuracy (%)
Training
accuracy (%)
Validation
accuracy (%)
External
Validation
accuracy -
ACRIN (%)
External
Validation
accuracy – CU
(%)
Iteration 1:
Model discovery 56 88.9 94.9 90.0 78.8 89.2
Iteration 2:
Hyperparameter optimization
52 89.9 96.2 90.0 78.8 91.9
Iteration 3:
Feature optimization 9 89.8 92.4 95.0 80.3 91.9
9 Selected Features: 2D Maximum, 2D Skewness, Mean, GLCM Cluster Prominence, GLCM Sum
Average, GLCM Cluster Shade, GLDM Dependence Variance, Maximum, 2D Kurtosis
Feature reduction The number of remaining Features
ICC 153
Wilcoxon Test 141
Hierarchical Clustering 62
Table 3. Model pipeline optimization results and validation results.
Table 2. Feature reduction results
9
10. • Utilized dataset of 209 patients for functional cardiac radiomics model
development.
• Achieved high accuracy:
9 optimal features were selected
Training set validation: TJU: 95%
External validation : ACRIN: 80.3%, CU: 91.1%
• Clinical Application:
Automated prediction of existing cardiac conditions.
Early biomarker for post-radiotherapy cardiac complications.
Conclusion
10