Optimized Feature Extraction for Radiomics Analysis of 18F-FDG PET Imaging

Enhancing Interoperability and Harmonisation of Nuclear Medicine Image Data and Associated Clinical Data

Nuclear imaging techniques such as positron emission tomography (PET) and single photon emission computed tomography (SPECT) in combination with computed tomography (CT) are established imaging modalities in clinical practice, particularly for oncological problems. Due to a multitude of manufacturers, different measurement protocols, local demographic or clinical workflow variations as well as various available reconstruction and analysis software, very heterogeneous datasets are generated. This review article examines the current state of interoperability and harmonisation of image data and related clinical data in the field of nuclear medicine. Various approaches and standards to improve data compatibility and integration are discussed. These include, for example, structured clinical history, standardisation of image acquisition and reconstruction as well as standardised preparation of image data for evaluation. Approaches to improve data acquisition, storage and analysis will be presented. Furthermore, approaches are presented to prepare the datasets in such a way that they become usable for projects applying artificial intelligence (AI) (machine learning, deep learning, etc.). This review article concludes with an outlook on future developments and trends related to AI in nuclear medicine, including a brief research of commercial solutions.

Impact of Tracer Dose Reduction in [18 F]-Labelled Fluorodeoxyglucose-Positron Emission Tomography ([18 F]-FDG)-PET) on Texture Features and Histogram Indices: A Study in Homogeneous Tissues of Phantom and Patient

BACKGROUND

Histogram indices (HIs) and texture features (TFs) are considered to play an important role in future oncologic PET-imaging and it is unknown how these indices are affected by changes of tracer doses. A randomized undersampling of PET list mode data enables a simulation of tracer dose reduction. We performed a phantom study to compare HIs/TFs of simulated and measured tracer dose reductions and evaluated changes of HIs/TFs in the liver of patients with PETs from simulated reduced tracer doses. Overall, 42 HIs/TFs were evaluated in a NEMA phantom at measured and simulated doses (stepwise reduction of [18 F] from 100% to 25% of the measured dose). [18 F]-FDG-PET datasets of 15 patients were simulated from 3.0 down to 0.5 MBq/kgBW in intervals of 0.25 MBq/kgBW. HIs/TFs were calculated from two VOIs placed in physiological tissue of the right and left liver lobe and linear correlations and coefficients of variation analysis were performed.

RESULTS

All 42 TFs did not differ significantly in measured and simulated doses (p > 0.05). Also, 40 TFs showed the same behaviour over dose reduction regarding differences in the same group (measured or simulated), and for 26 TFs a linear behaviour over dose reduction for measured and simulated doses could be validated. Out of these, 13 TFs could be identified, which showed a linear change in TF value in both the NEMA phantom and patient data and therefore should maintain the same informative value when transferred in a dose reduction setting. Out of this Homogeneity 2, Entropy and Zone size non-uniformity are of special interest because they have been described as preferentially considerable for tumour heterogeneity characterization.

CONCLUSIONS

We could show that there was no significant difference of measured and simulated HIs/TFs in the phantom study and most TFs reveal a linear behaviour over dose reduction, when tested in homogeneous tissue. This indicates that texture analysis in PET might be robust to dose modulations.

Development of a nomogram based on radiomics and semantic features for predicting chromosome 7 gain/chromosome 10 loss in IDH wild-type histologically low-grade gliomas

Purpose

To predict chromosome 7 gain and chromosome 10 loss (+7/-10) in IDH wild-type (IDH-wt) histologically low-grade gliomas (LGG) by machine learning models based on MRI radiomics and semantic features.

Methods

A total of 122 patients diagnosed as IDH-wt histologically LGG were retrospectively included in this study. The patients were randomly divided into a training group and a test group in a ratio of 7:3. The radiomics features were extracted from axial T1WI, T2WI, FLAIR and CET1 sequences, respectively. The distance correlation (DC) and least absolute shrinkage and selection operator (LASSO) were used to select the radiomics signatures. Three machine learning algorithms including neural network (NN), support vector machine (SVM), and linear discriminant analysis (LDA) were used to construct radiomics models. In addition, a nomogram was developed by combining the optimal radiomics signature with clinical risk factors, and the potential clinical utility of the nomogram was evaluated using decision curve analysis.

Results

The LDA+DC model was identified as the optimal classifier among the six radiomics models. Necrosis was determined as a risk factor for +7/-10 in IDH-wt histologically LGG. The nomogram achieved the best performance, with an AUC of 0.854 and an accuracy of 0.778 in the independent test group. The decision curve of the nomogram confirmed its clinical usefulness in a wide range of thresholds.

Conclusion

The nomogram combining radiomics and semantic features can predict the +7/-10 status effectively, which may contribute to the risk stratification and individualized treatment planning of patients with IDH-wt histologically LGG.

Feasibility of radiomic feature harmonization for pooling of [18F]FET or [18F]GE-180 PET images of gliomas

INTRODUCTION

Large datasets are required to ensure reliable non-invasive glioma assessment with radiomics-based machine learning methods. This can often only be achieved by pooling images from different centers. Moreover, trained models should perform with high accuracy when applied to data from different centers. In this study, the impact of reconstruction settings and segmentation methods on radiomic features derived from amino acid and TSPO PET images of glioma patients was examined. Additionally, the ability to model and thus reduce feature differences was investigated.

METHODS

[¹⁸F]FET and [¹⁸F]GE-180 PET data were acquired from 19 glioma patients. For each acquisition, 10 reconstruction settings and 9 segmentation methods were included to emulate multicentric data. Statistical robustness measures were calculated before and after ComBat harmonization. Differences between features due to setting variations were assessed using Friedman test, coefficient of variation (CV) and inter-rater reliability measures, including intraclass and Spearman's rank correlation coefficients and Fleiss' Kappa.

RESULTS

According to Friedman analyses, most features (>60%) showed significant differences. Yet, CV and inter-rater reliability measures indicated higher robustness. ComBat resulted in almost complete harmonization (>87%) according to Friedman test and little to no improvement according to CV and inter-rater reliability measures. [¹⁸F]GE-180 features were more sensitive to reconstruction settings than [¹⁸F]FET features.

CONCLUSIONS

According to Friedman test, feature distributions could be successfully aligned using ComBat. However, depending on settings, changes in patient ranks were observed for some features and could not be eliminated by harmonization. Thus, for clinical utilization it is recommended to exclude affected features.

Impact of image quality on radiomics applications

Radiomics features extracted from medical images have been widely reported to be useful in the patient specific outcome modeling for variety of assessment and prediction purposes. Successful application of radiomics features as imaging biomarkers, however, is dependent on the robustness of the approach to the variation in each step of the modeling workflow. Variation in the input image quality is one of the main sources that impacts the reproducibility of radiomics analysis when a model is applied to broader range of medical imaging data. The quality of medical image is generally affected by both the scanner related factors such as image acquisition/reconstruction settings and the patient related factors such as patient motion. This article aimed to review the published literatures in this field that reported the impact of various imaging factors on the radiomics features through the change in image quality. The literatures were categorized by different imaging modalities and also tabulated based on the imaging parameters and the class of radiomics features included in the study. Strategies for image quality standardization were discussed based on the relevant literatures and recommendations for reducing the impact of image quality variation on the radiomics in multi-institutional clinical trial were summarized at the end of this article.

A multi-modality physical phantom for mimicking tumour heterogeneity patterns in PET/CT and PET/MRI

Background

Hybrid imaging (e.g., positron emission tomography [PET]/computed tomography [CT], PET/magnetic resonance imaging [MRI]) helps one to visualize and quantify morphological and physiological tumor characteristics in a single study. The noninvasive characterization of tumor heterogeneity is essential for grading, treatment planning, and following‐up oncological patients. However, conventional (CONV) image‐based parameters, such as tumor diameter, tumor volume, and radiotracer activity uptake, are insufficient to describe tumor heterogeneities. Here, radiomics shows promise for a better characterization of tumors. Nevertheless, the validation of such methods demands imaging objects capable of reflecting heterogeneities in multi‐modality imaging. We propose a phantom to simulate tumor heterogeneity repeatably in PET, CT, and MRI.

Methods

The phantom consists of three 50‐ml plastic tubes filled partially with acrylic spheres of S1: 1.6 mm, S2: 50%(1.6 mm)/50%(6.3 mm), or S3: 6.3‐mm diameter. The spheres were fixed to the bottom of each tube by a plastic grid, yielding one sphere free homogeneous region and one heterogeneous (S1, S2, or S3) region per tube. A 3‐tube phantom and its replica were filled with a fluorodeoxyglucose (18F) solution for test–retest measurements in a PET/CT Siemens TPTV and a PET/MR Siemens Biograph mMR system. A number of 42 radiomic features (10 first order and 32 texture features) were calculated for each phantom region and imaging modality. Radiomic features stability was evaluated through coefficients of variation (COV) across phantoms and scans for PET, CT, and MRI. Further, the Wilcoxon test was used to assess the capability of stable features to discriminate the simulated phantom regions.

Results

The different patterns (S1–S3) did present visible heterogeneity in all imaging modalities. However, only for CT and MRI, a clear visual difference was present between the different patterns. Across all phantom regions in PET, CT, and MR images, 10, 16, and 21 features out of 42 evaluated features in total had a COV of 10% or less. In particular, CONV, histogram, and gray‐level run length matrix features showed high repeatability for all the phantom regions and imaging modalities. Several of repeatable texture features allowed the image‐based discrimination of the different phantom regions (p < 0.05). However, depending on the feature, different pattern discrimination capabilities were found for the different imaging modalities.

Conclusion

The proposed phantom appears suitable for simulating heterogeneities in PET, CT, and MRI. We demonstrate that it is possible to select radiomic features for the readout of the phantom. Most of these features had been shown to be relevant in previous clinical studies.

Improved Prognosis of Treatment Failure in Cervical Cancer with Nontumor PET/CT Radiomics

Radiomics has been applied to predict recurrence in several disease sites, but current approaches are typically restricted to analyzing tumor features, neglecting nontumor information in the rest of the body. The purpose of this work was to develop and validate a model incorporating nontumor radiomics, including whole-body features, to predict treatment outcomes in patients with previously untreated locoregionally advanced cervical cancer. Methods: We analyzed 127 cervical cancer patients treated definitively with chemoradiotherapy and intracavitary brachytherapy. All patients underwent pretreatment whole-body ¹⁸F-FDG PET/CT. To quantify effects due to the tumor itself, the gross tumor volume (GTV) was directly contoured on the PET/CT image. Meanwhile, to quantify effects arising from the rest of the body, the planning target volume (PTV) was deformably registered from each planning CT to the PET/CT scan, and a semiautomated approach combining seed-growing and manual contour review generated whole-body muscle, bone, and fat segmentations on each PET/CT image. A total of 965 radiomic features were extracted for GTV, PTV, muscle, bone, and fat. Ninety-five patients were used to train a Cox model of disease recurrence including both radiomic and clinical features (age, stage, tumor grade, histology, and baseline complete blood cell counts), using bagging and split-sample-validation for feature reduction and model selection. To further avoid overfitting, the resulting models were tested for generalization on the remaining 32 patients, by calculating a risk score based on Cox regression and evaluating the c-index (c-index > 0.5 indicates predictive power). Results: Optimal performance was seen in a Cox model including 1 clinical biomarker (whether or not a tumor was stage III-IVA), 2 GTV radiomic biomarkers (PET gray-level size-zone matrix small area low gray level emphasis and zone entropy), 1 PTV radiomic biomarker (major axis length), and 1 whole-body radiomic biomarker (CT bone root mean square). In particular, stratification into high- and low-risk groups, based on the linear risk score from this Cox model, resulted in a hazard ratio of 0.019 (95% CI, 0.004, 0.082), an improvement over stratification based on clinical stage alone, which had a hazard ratio of 0.36 (95% CI, 0.16, 0.83). Conclusion: Incorporating nontumor radiomic biomarkers can improve the performance of prognostic models compared with using only clinical and tumor radiomic biomarkers. Future work should look to further test these models in larger, multiinstitutional cohorts.

Analysis of Cross-Combinations of Feature Selection and Machine-Learning Classification Methods Based on [18F]F-FDG PET/CT Radiomic Features for Metabolic Response Prediction of Metastatic Breast Cancer Lesions

Simple Summary

Breast cancer is a leading cause of morbidity and mortality worldwide. The metastatic disease is largely responsible for cancer patient deaths, and its treatment implies usually different therapies. In this context, the prediction of response to treatment or depiction of treatment-resistant phenotypes is essential in clinical practice, especially in the new era of precision medicine. In this study, we used several combinations of feature selection methods and machine-learning classifiers to construct predictive models of the metabolic response to the treatment of metastatic breast cancer lesions. These models were based on clinical variables and radiomic features extracted from 2-deoxy-2-[¹⁸F]fluoro-D-glucose positron emission tomography/computed tomography ([¹⁸F]F-FDG PET/CT) images, obtained prior to the treatment. Our main goal was to know if this prediction was feasible and to identify those combinations with better predictive performance. We found that several combinations were successful to predict the metabolic response to treatment, of which the least absolute shrinkage and selection operator (Lasso) + support vector machines (SVM) had the best mean performance in terms of area under the curve, in both training and validation cohorts. Model performances depended largely on the selected combinations.

Abstract

Background: This study aimed to identify optimal combinations between feature selection methods and machine-learning classifiers for predicting the metabolic response of individual metastatic breast cancer lesions, based on clinical variables and radiomic features extracted from pretreatment [¹⁸F]F-FDG PET/CT images. Methods: A total of 48 patients with confirmed metastatic breast cancer, who received different treatments, were included. All patients had an [¹⁸F]F-FDG PET/CT scan before and after the treatment. From 228 metastatic lesions identified, 127 were categorized as responders (complete or partial metabolic response) and 101 as non-responders (stable or progressive metabolic response), by using the percentage changes in SULpeak (peak standardized uptake values normalized for body lean body mass). The lesion pool was divided into training (n = 182) and testing cohorts (n = 46); for each lesion, 101 image features from both PET and CT were extracted (202 features per lesion). These features, along with clinical and pathological information, allowed the prediction model’s construction by using seven popular feature selection methods in cross-combination with another seven machine-learning (ML) classifiers. The performance of the different models was investigated with the receiver-operating characteristic curve (ROC) analysis, using the area under the curve (AUC) and accuracy (ACC) metrics. Results: The combinations, least absolute shrinkage and selection operator (Lasso) + support vector machines (SVM), or random forest (RF) had the highest AUC in the cross-validation, with 0.93 ± 0.06 and 0.92 ± 0.03, respectively, whereas Lasso + neural network (NN) or SVM, and mutual information (MI) + RF, had the higher AUC and ACC in the validation cohort, with 0.90/0.72, 0.86/0.76, and 87/85, respectively. On average, the models with Lasso and models with SVM had the best mean performance for both AUC and ACC in both training and validation cohorts. Conclusions: Image features obtained from a pretreatment [¹⁸F]F-FDG PET/CT along with clinical vaiables could predict the metabolic response of metastatic breast cancer lesions, by their incorporation into predictive models, whose performance depends on the selected combination between feature selection and ML classifier methods.

Development and validation of a radiomic model for the diagnosis of dopaminergic denervation on [18F]FDOPA PET/CT

Purpose

FDOPA PET shows good performance for the diagnosis of striatal dopaminergic denervation, making it a valuable tool for the differential diagnosis of Parkinsonism. Textural features are image biomarkers that could potentially improve the early diagnosis and monitoring of neurodegenerative parkinsonian syndromes. We explored the performances of textural features for binary classification of FDOPA scans.

Methods

We used two FDOPA PET datasets: 443 scans for feature selection, and 100 scans from a different PET/CT system for model testing. Scans were labelled according to expert interpretation (dopaminergic denervation versus no dopaminergic denervation). We built LASSO logistic regression models using 43 biomarkers including 32 textural features. Clinical data were also collected using a shortened UPDRS scale.

Results

The model built from the clinical data alone had a mean area under the receiver operating characteristics (AUROC) of 63.91. Conventional imaging features reached a maximum score of 93.47 but the addition of textural features significantly improved the AUROC to 95.73 (p < 0.001), and 96.10 (p < 0.001) when limiting the model to the top three features: GLCM_Correlation, Skewness and Compacity. Testing the model on the external dataset yielded an AUROC of 96.00, with 95% sensitivity and 97% specificity. GLCM_Correlation was one of the most independent features on correlation analysis, and systematically had the heaviest weight in the classification model.

Conclusion

A simple model with three radiomic features can identify pathologic FDOPA PET scans with excellent sensitivity and specificity. Textural features show promise for the diagnosis of parkinsonian syndromes.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00259-022-05816-7.

Noise-Based Image Harmonization Significantly Increases Repeatability and Reproducibility of Radiomics Features in PET Images: A Phantom Study

For multicenter clinical studies, characterizing the robustness of image-derived radiomics features is essential. Features calculated on PET images have been shown to be very sensitive to image noise. The purpose of this work was to investigate the efficacy of a relatively simple harmonization strategy on feature robustness and agreement. A purpose-built texture pattern phantom was scanned on 10 different PET scanners in 7 institutions with various different image acquisition and reconstruction protocols. An image harmonization technique based on equalizing a contrast-to-noise ratio was employed to generate a "harmonized" alongside a "standard" dataset for a reproducibility study. In addition, a repeatability study was performed with images from a single PET scanner of variable image noise, varying the binning time of the reconstruction. Feature agreement was measured using the intraclass correlation coefficient (ICC). In the repeatability study, 81/93 features had a lower ICC on the images with the highest image noise as compared to the images with the lowest image noise. Using the harmonized dataset significantly improved the feature agreement for five of the six investigated feature classes over the standard dataset. For three feature classes, high feature agreement corresponded with higher sensitivity to the different patterns, suggesting a way to select suitable features for predictive models.

Repeatability and Reproducibility of Computed Tomography Radiomics for Pulmonary Nodules: A Multicenter Phantom Study

BACKGROUND

Radiomics can yield minable information from medical images, which can facilitate computer-aided diagnosis. However, the lack of repeatability and reproducibility of radiomic features (RFs) may hinder their generalizability in clinical applications.

OBJECTIVES

The aims of this study were to explore 3 main sources of variability in RFs, investigate their influencing magnitudes and patterns, and identify a subset of robust RFs for further studies.

MATERIALS AND METHODS

A chest phantom with nodules was scanned with different computed tomography (CT) scanners repeatedly with varying acquisition and reconstruction parameters (April-May 2019) to evaluate 3 sources of variability: test-retest, inter-CT, and intra-CT protocol variability. The robustness of the RFs was measured using the concordance correlation coefficient, dynamic range, and intraclass correlation coefficient (ICC). The influencing magnitudes and patterns were analyzed using the Friedman test and Spearman rank correlation coefficient. Stable and informative RFs were selected, and their redundancy was eliminated using hierarchical clustering. Clinical validation was also performed to verify the clinical effectiveness and potential enhancement of the generalizability of radiomics research.

RESULTS

A total of 1295 RFs that showed all 3 sources of variability were included. The reconstruction kernel and the iteration level showed the greatest (ICC, 0.35 ± 0.31) and the least (ICC, 0.63 ± 0.27) influence on magnitudes. The different sources of variability showed relatively consistent patterns of influence (false discovery rate <0.001). Finally, we obtained a subset of 19 stable, informative, and nonredundant RFs under all 3 sources of variability. These RFs exhibited clinical effectiveness and showed better prediction performance than unstable RFs in the validation dataset (P = 0.017, Delong test).

CONCLUSIONS

The stability of RFs was affected to different degrees by test-retest and differences in CT manufacturers and models and CT acquisition and reconstruction parameters, but the influences of these factors showed relatively consistent patterns. We also obtained a subset of 19 stable, informative, and nonredundant RFs that should be preferably used to enhance the generalizability of further radiomics research.

AI in spotting high-risk characteristics of medical imaging and molecular pathology

Medical imaging provides a comprehensive perspective and rich information for disease diagnosis. Combined with artificial intelligence technology, medical imaging can be further mined for detailed pathological information. Many studies have shown that the macroscopic imaging characteristics of tumors are closely related to microscopic gene, protein and molecular changes. In order to explore the function of artificial intelligence algorithms in in-depth analysis of medical imaging information, this paper reviews the articles published in recent years from three perspectives: medical imaging analysis method, clinical applications and the development of medical imaging in the direction of pathological molecular prediction. We believe that AI-aided medical imaging analysis will be extensively contributing to precise and efficient clinical decision.

Advances in PET-CT technology: An update

This article reviews the current evolution and future directions in PET-CT technology focusing on three areas: time of flight, image reconstruction, and data-driven gating. Image reconstruction is considered with advances in point spread function modelling, Bayesian penalised likelihood reconstruction, and artificial intelligence approaches. Data-driven gating is examined with reference to respiratory motion, cardiac motion, and head motion. For each of these technological advancements, theory will be briefly discussed, benefits of their use in routine practice will be detailed and potential future developments will be discussed. Representative clinical cases will be presented, demonstrating the huge opportunities given to the PET community by hardware and software advances in PET technology when it comes to lesion detection, disease characterization, accurate quantitation and quicker scans. Through this review, hospitals are encouraged to embrace, evaluate and appropriately implement the wide range of new PET technologies that are available now or in the near future, for the improvement of patient care.

Radiomics in PET Imaging:: A Practical Guide for Newcomers

Radiomics has undergone considerable development in recent years. In PET imaging, very promising results concerning the ability of handcrafted features to predict the biological characteristics of lesions and to assess patient prognosis or response to treatment have been reported in the literature. This article presents a checklist for designing a reliable radiomic study, gives an overview of the steps of the pipeline, and outlines approaches for data harmonization. Tips are provided for critical reading of the content of articles. The advantages and limitations of handcrafted radiomics compared with deep-learning approaches for the characterization of PET images are also discussed.

Morpho-Molecular Metabolic Analysis and Classification of Human Pituitary Gland and Adenoma Biopsies Based on Multimodal Optical Imaging

Pituitary adenomas count among the most common intracranial tumors. During pituitary oncogenesis structural, textural, metabolic and molecular changes occur which can be revealed with our integrated ultrahigh-resolution multimodal imaging approach including optical coherence tomography (OCT), multiphoton microscopy (MPM) and line scan Raman microspectroscopy (LSRM) on an unprecedented cellular level in a label-free manner. We investigated 5 pituitary gland and 25 adenoma biopsies, including lactotroph, null cell, gonadotroph, somatotroph and mammosomatotroph as well as corticotroph. First-level binary classification for discrimination of pituitary gland and adenomas was performed by feature extraction via radiomic analysis on OCT and MPM images and achieved an accuracy of 88%. Second-level multi-class classification was performed based on molecular analysis of the specimen via LSRM to discriminate pituitary adenomas subtypes with accuracies of up to 99%. Chemical compounds such as lipids, proteins, collagen, DNA and carotenoids and their relation could be identified as relevant biomarkers, and their spatial distribution visualized to provide deeper insight into the chemical properties of pituitary adenomas. Thereby, the aim of the current work was to assess a unique label-free and non-invasive multimodal optical imaging platform for pituitary tissue imaging and to perform a multiparametric morpho-molecular metabolic analysis and classification.

Evaluation and optimization of radiomics features stability to respiratory motion in 18 F-FDG 3D PET imaging

PURPOSE

To evaluate the impact of respiratory motion on radiomics features in ¹⁸ F-fluoro-2-deoxy-D-glucose three dimensional positron emission tomography (¹⁸ F-FDG 3D PET) imaging and optimize feature stability by combining preprocessing configurations and aggregation strategies.

METHODS

An in-house developed respiratory motion phantom was imaged in 3D PET scanner under nine respiratory patterns including one reference pattern. In total, 487 radiomics features were extracted for each respiratory pattern. Feature stability to respiratory motion was first evaluated by metrics of coefficient of variation (COV) and relative difference (RD) in a fixed preprocessing configuration. Further, one-way ANOVA and trend analysis were performed to evaluate the impact of preprocessing configuration (voxel size, discretization scheme) and aggregation strategy on feature stability. Finally, an optimization framework was proposed by selected feature-specific configurations with minimum COV value, and the diagnostic performance was validated in stable versus unstable features and fixed versus optimal features by a set of 46 patients with lung disease.

RESULTS

PET radiomics features were sensitive to respiratory motion, only 79/487 (16%) features were identified to be very stable in the fixed configuration. Preprocessing configuration and aggregation strategy had an impact on feature stability. For different voxel size, bin number, bin size and aggregation strategy, 188/487 (39%), 70/487 (15%), 148/487 (30%), and 38/95 (29%) features appeared significant changes in feature stability. The optimized configuration had the potential to improve feature stability compared to fixed configuration, with the COV decreased from 22% ±24% to 16% ±20%. Regarding the diagnostic performance, the stable and optimal configuration features outperformed the unstable and fixed configuration features, respectively (AUC 0.88, 0.87 vs. 0.83, 0.85).

CONCLUSIONS

Respiratory motion shows considerable impact on feature stability in 3D PET imaging, while optimizing preprocessing configuration may improve feature stability and diagnostic performance.

Supervised machine learning enables non-invasive lesion characterization in primary prostate cancer with [68Ga]Ga-PSMA-11 PET/MRI

PURPOSE

Risk classification of primary prostate cancer in clinical routine is mainly based on prostate-specific antigen (PSA) levels, Gleason scores from biopsy samples, and tumor-nodes-metastasis (TNM) staging. This study aimed to investigate the diagnostic performance of positron emission tomography/magnetic resonance imaging (PET/MRI) in vivo models for predicting low-vs-high lesion risk (LH) as well as biochemical recurrence (BCR) and overall patient risk (OPR) with machine learning.

METHODS

Fifty-two patients who underwent multi-parametric dual-tracer [18F]FMC and [68Ga]Ga-PSMA-11 PET/MRI as well as radical prostatectomy between 2014 and 2015 were included as part of a single-center pilot to a randomized prospective trial (NCT02659527). Radiomics in combination with ensemble machine learning was applied including the [68Ga]Ga-PSMA-11 PET, the apparent diffusion coefficient, and the transverse relaxation time-weighted MRI scans of each patient to establish a low-vs-high risk lesion prediction model (MLH). Furthermore, MBCR and MOPR predictive model schemes were built by combining MLH, PSA, and clinical stage values of patients. Performance evaluation of the established models was performed with 1000-fold Monte Carlo (MC) cross-validation. Results were additionally compared to conventional [68Ga]Ga-PSMA-11 standardized uptake value (SUV) analyses.

RESULTS

The area under the receiver operator characteristic curve (AUC) of the MLH model (0.86) was higher than the AUC of the [68Ga]Ga-PSMA-11 SUVmax analysis (0.80). MC cross-validation revealed 89% and 91% accuracies with 0.90 and 0.94 AUCs for the MBCR and MOPR models respectively, while standard routine analysis based on PSA, biopsy Gleason score, and TNM staging resulted in 69% and 70% accuracies to predict BCR and OPR respectively.

CONCLUSION

Our results demonstrate the potential to enhance risk classification in primary prostate cancer patients built on PET/MRI radiomics and machine learning without biopsy sampling.

Repeatability of 18F-FDG PET Radiomic Features in Cervical Cancer

Knowledge of the intrinsic variability of radiomic features is essential to the proper interpretation of changes in these features over time. The primary aim of this study was to assess the test-retest repeatability of radiomic features extracted from ¹⁸F-FDG PET images of cervical tumors. The impact of different image preprocessing methods was also explored. Methods: Patients with cervical cancer underwent baseline and repeat ¹⁸F-FDG PET/CT imaging within 7 d. PET images were reconstructed using 2 methods: ordered-subset expectation maximization (PET_OSEM) or ordered-subset expectation maximization with point-spread function (PET_PSF). Tumors were segmented to produce whole-tumor volumes of interest (VOI_WT) and 40% isocontours (VOI₄₀). Voxels were either left at the default size or resampled to 3-mm isotropic voxels. SUV was discretized to a fixed number of bins (32, 64, or 128). Radiomic features were extracted from both VOIs, and repeatability was then assessed using the Lin concordance correlation coefficient (CCC). Results: Eleven patients were enrolled and completed the test-retest PET/CT imaging protocol. Shape, neighborhood gray-level difference matrix, and gray-level cooccurrence matrix features were repeatable, with a mean CCC value of 0.81. Radiomic features extracted from PET_OSEM images showed significantly better repeatability than features extracted from PET_PSF images (P < 0.001). Radiomic features extracted from VOI₄₀ were more repeatable than features extracted from VOI_WT (P < 0.001). For most features (78.4%), a change in bin number or voxel size resulted in less than a 10% change in feature value. All gray-level emphasis and gray-level run emphasis features showed poor repeatability (CCC values < 0.52) when extracted from VOI_WT but were highly repeatable (mean CCC values > 0.96) when extracted from VOI₄₀ Conclusion: Shape, gray-level cooccurrence matrix, and neighborhood gray-level difference matrix radiomic features were consistently repeatable, whereas gray-level run length matrix and gray-level zone length matrix features were highly variable. Radiomic features extracted from VOI₄₀ were more repeatable than features extracted from VOI_WT Changes in voxel size or SUV discretization parameters typically resulted in relatively small differences in feature value, though several features were highly sensitive to these changes.

Imager-4D: New Software for Viewing Dynamic PET Scans and Extracting Radiomic Parameters from PET Data

Extensive research is currently being conducted into dynamic positron emission tomography (PET) acquisitions (including dynamic whole-body imaging) as well as extraction of radiomic features from imaging modalities. We describe a new PET viewing software known as Imager-4D that provides a facile means of viewing and analyzing dynamic PET data and obtaining associated quantitative metrics including radiomic parameters. The Imager-4D was programmed in the Java language utilizing the FX extensions. It is executable on any system for which a Java w/FX compliant virtual machine is available. The software incorporates the ability to view and analyze dynamic data acquired with different types of dynamic protocols. For image display, the program maintains a built-in library of 62 different lookup tables with monochromatic and full-color distributions. The Imager-4D provides multiple display layouts and can display fused images. Multiple methods of volume-of-interest (VOI) selection are available. Dynamic analysis features, such as image summation and full Patlak analysis, are also available. The user interface includes window width and level, blending, and zoom functionality. VOI sizes are adjustable and data from VOIs can either be displayed numerically or graphically within the software or exported. An example case of a 50-year-old woman with metastatic colorectal cancer and thyroiditis is included and demonstrates the steps for a user to obtain standard PET parameters, dynamic data, and radiomic features using selected VOIs. The Imager-4D represents a novel PET viewer that allows the user to view dynamic PET data, to derive dynamic and radiomic parameters from that data, and to combine dynamic data with radiomics ("dynomics"). The Imager-4D is available as a free download. This software has the potential to speed the adoption of advanced analysis of dynamic PET data into routine clinical use.

TSPO PET imaging of natalizumab-associated progressive multifocal leukoencephalopathy

Progressive multifocal leukoencephalopathy (PML) is a severe infection of the central nervous system caused by the polyomavirus JC (JCV) that can occur in multiple sclerosis (MS) patients treated with natalizumab. Clinical management of patients with natalizumab-associated PML is challenging not the least because current imaging tools for the early detection, longitudinal monitoring and differential diagnosis of PML lesions are limited. Here we evaluate whether TSPO positron emission tomography (PET) imaging can be applied to monitor the inflammatory activity of PML lesions over time and differentiate them from MS lesions. For this monocenter pilot study we followed 8 patients with natalizumab-associated PML with PET imaging using the TSPO radioligand [18F]GE-180 combined with frequent 3 T MRI imaging. In addition we compared TSPO PET signals in PML lesions with the signal pattern of MS lesions from 17 independent MS patients. We evaluated the standardized uptake value ratio (SUVR) as well as the morphometry of the TSPO uptake for putative PML and MS lesions areas compared to a radiologically unaffected pseudo-reference region in the cerebrum. Furthermore TSPO expression in situ was immunohistochemically verified by determining the density and cellular identity of TSPO-expressing cells in brain sections from four patients with early natalizumab-associated PML as well as five patients with other forms of PML and six patients with inflammatory demyelinating CNS lesions (clinically isolated syndrome/MS). Histological analysis revealed a reticular accumulation of TSPO expressing phagocytes in PML lesions, while such phagocytes showed a more homogenous distribution in putative MS lesions. TSPO PET imaging showed an enhanced tracer uptake in natalizumab-associated PML lesions that was present from the early to the chronic stages (up to 52 months after PML diagnosis). While gadolinium enhancement on MRI rapidly declined to baseline levels, TSPO tracer uptake followed a slow one phase decay curve. A TSPO-based 3-dimensional diagnostic matrix taking into account the uptake levels as well as the shape and texture of the TSPO signal differentiated more than 96% of PML and MS lesions. Indeed, treatment with rituximab after natalizumab-associated PML in three patients did not affect tracer uptake in the assigned PML lesions but reverted tracer uptake to baseline in the assigned active MS lesions. Taken together our study suggests that TSPO PET imaging can reveal CNS inflammation in natalizumab-associated PML. TSPO PET may facilitate longitudinal monitoring of disease activity and help to distinguish recurrent MS activity from PML progression.

Breast Tumor Characterization Using [18F]FDG-PET/CT Imaging Combined with Data Preprocessing and Radiomics

Simple Summary

Breast cancer is the second most common diagnosed malignancy in women worldwide. In this study, we examine the feasibility of breast tumor characterization based on [¹⁸F]FDG-PET/CT images using machine learning (ML) approaches in combination with data-preprocessing techniques. ML prediction models for breast cancer detection and the identification of breast cancer receptor status, proliferation rate, and molecular subtypes were established and evaluated. Furthermore, the importance of most repeatable features was investigated. Results displayed high performance of malignant/benign tumor differentiation and triple negative tumor subtype ML models. We observed high repeatability of radiomic features for both high performing predictive models.

Abstract

Background: This study investigated the performance of ensemble learning holomic models for the detection of breast cancer, receptor status, proliferation rate, and molecular subtypes from [¹⁸F]FDG-PET/CT images with and without incorporating data pre-processing algorithms. Additionally, machine learning (ML) models were compared with conventional data analysis using standard uptake value lesion classification. Methods: A cohort of 170 patients with 173 breast cancer tumors (132 malignant, 38 benign) was examined with [¹⁸F]FDG-PET/CT. Breast tumors were segmented and radiomic features were extracted following the imaging biomarker standardization initiative (IBSI) guidelines combined with optimized feature extraction. Ensemble learning including five supervised ML algorithms was utilized in a 100-fold Monte Carlo (MC) cross-validation scheme. Data pre-processing methods were incorporated prior to machine learning, including outlier and borderline noisy sample detection, feature selection, and class imbalance correction. Feature importance in each model was assessed by calculating feature occurrence by the R-squared method across MC folds. Results: Cross validation demonstrated high performance of the cancer detection model (80% sensitivity, 78% specificity, 80% accuracy, 0.81 area under the curve (AUC)), and of the triple negative tumor identification model (85% sensitivity, 78% specificity, 82% accuracy, 0.82 AUC). The individual receptor status and luminal A/B subtype models yielded low performance (0.46–0.68 AUC). SUV_max model yielded 0.76 AUC in cancer detection and 0.70 AUC in predicting triple negative subtype. Conclusions: Predictive models based on [¹⁸F]FDG-PET/CT images in combination with advanced data pre-processing steps aid in breast cancer diagnosis and in ML-based prediction of the aggressive triple negative breast cancer subtype.

Potential Value of Radiomics in the Identification of Stage T3 and T4a Esophagogastric Junction Adenocarcinoma Based on Contrast-Enhanced CT Images

Purpose

This study was designed to evaluate the predictive performance of contrast-enhanced CT-based radiomic features for the personalized, differential diagnosis of esophagogastric junction (EGJ) adenocarcinoma at stages T3 and T4a.

Methods

Two hundred patients with T3 (n = 44) and T4a (n = 156) EGJ adenocarcinoma lesions were enrolled in this study. Traditional computed tomography (CT) features were obtained from contrast-enhanced CT images, and the traditional model was constructed using a multivariate logistic regression analysis. A radiomic model was established based on radiomic features from venous CT images, and the radiomic score (Radscore) of each patient was calculated. A combined nomogram diagnostic model was constructed based on Radscores and traditional features. The diagnostic performances of these three models (traditional model, radiomic model, and nomogram) were assessed with receiver operating characteristics curves. Sensitivity, specificity, accuracy, positive predictive value, negative predictive value, and areas under the curve (AUC) of models were calculated, and the performances of the models were evaluated and compared. Finally, the clinical effectiveness of the three models was evaluated by conducting a decision curve analysis (DCA).

Results

An eleven-feature combined radiomic signature and two traditional CT features were constructed as the radiomic and traditional feature models, respectively. The Radscore was significantly different between patients with stage T3 and T4a EGJ adenocarcinoma. The combined nomogram performed the best and has potential clinical usefulness.

Conclusions

The developed combined nomogram might be useful in differentiating T3 and T4a stages of EGJ adenocarcinoma and may facilitate the decision-making process for the treatment of T3 and T4a EGJ adenocarcinoma.

Standardization of imaging methods for machine learning in neuro-oncology

Radiomics is a novel technique in which quantitative phenotypes or features are extracted from medical images. Machine learning enables analysis of large quantities of medical imaging data generated by radiomic feature extraction. A growing number of studies based on these methods have developed tools for neuro-oncology applications. Despite the initial promises, many of these imaging tools remain far from clinical implementation. One major limitation hindering the use of these models is their lack of reproducibility when applied across different institutions and clinical settings. In this article, we discuss the importance of standardization of methodology and reporting in our effort to improve reproducibility. Ongoing efforts of standardization for neuro-oncological imaging are reviewed. Challenges related to standardization and potential disadvantages in over-standardization are also described. Ultimately, greater multi-institutional collaborative effort is needed to provide and implement standards for data acquisition and analysis methods to facilitate research results to be interoperable and reliable for integration into different practice environments.

Radiomics Features of 18F-Fluorodeoxyglucose Positron-Emission Tomography as a Novel Prognostic Signature in Colorectal Cancer

The aim of this study was to investigate the prognostic value of radiomics signatures derived from ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron-emission tomography (PET) in patients with colorectal cancer (CRC). From April 2008 to Jan 2014, we identified CRC patients who underwent ¹⁸F-FDG-PET before starting any neoadjuvant treatments and surgery. Radiomics features were extracted from the primary lesions identified on ¹⁸F-FDG-PET. Patients were divided into a training and validation set by random sampling. A least absolute shrinkage and selection operator Cox regression model was applied for prognostic signature building with progression-free survival (PFS) using the training set. Using the calculated radiomics score, a nomogram was developed, and its clinical utility was assessed in the validation set. A total of 381 patients with surgically resected CRC patients (training set: 228 vs. validation set: 153) were included. In the training set, a radiomics signature labeled as a rad_score was generated using two PET-derived features, such as gray-level run length matrix long-run emphasis (GLRLM_LRE) and gray-level zone length matrix short-zone low-gray-level emphasis (GLZLM_SZLGE). Patients with a high rad_score in the training and validation set had a shorter PFS. Multivariable analysis revealed that the rad_score was an independent prognostic factor in both training and validation sets. A radiomics nomogram, developed using rad_score, nodal stage, and lymphovascular invasion, showed good performance in the calibration curve and comparable predictive power with the staging system in the validation set. Textural features derived from ¹⁸F-FDG-PET images may enable detailed stratification of prognosis in patients with CRC.

Preoperative prediction for lauren type of gastric cancer: A radiomics nomogram analysis based on CT images and clinical features

PURPOSE

To investigate feasibility of predicting Lauren type of gastric cancer based on CT radiomics nomogram before operation.

MATERIALS AND METHODS

The clinical data and pre-treatment CT images of 300 gastric cancer patients with Lauren intestinal or diffuse type confirmed by postoperative pathology were retrospectively analyzed, who were randomly divided into training set and testing set with a ratio of 2:1. Clinical features were compared between the two Lauren types in the training set and testing set, respectively. Gastric tumors on CT images were manually segmented using ITK-SNAP software, and radiomic features of the segmented tumors were extracted, filtered and minimized using the least absolute shrinkage and selection operator (LASSO) regression to select optimal features and develop radiomics signature. A nomogram was constructed with radiomic features and clinical characteristics to predict Lauren type of gastric cancer. Clinical model, radiomics signature model, and the nomogram model were compared using the receiver operating characteristic (ROC) curve analysis with area under the curve (AUC). The calibration curve was used to test the agreement between prediction probability and actual clinical findings, and the decision curve was performed to assess the clinical usage of the nomogram model.

RESULTS

In clinical features, Lauren type of gastric cancer relate to age and CT-N stage of patients (all p < 0.05). Radiomics signature was developed with the retained 10 radiomic features. The nomogram was constructed with the 2 clinical features and radiomics signature. Among 3 prediction models, performance of the nomogram was the best in predicting Lauren type of gastric cancer, with the respective AUC, accuracy, sensitivity and specificity of 0.864, 78.0%, 90.0%, 70.0%in the testing set. In addition, the calibration curve showed a good agreement between prediction probability and actual clinical findings (p > 0.05).

CONCLUSION

The nomogram combining radiomics signature and clinical features is a useful tool with the increased value to predict Lauren type of gastric cancer.

Application of radiomics and machine learning in head and neck cancers

With the continuous development of medical image informatics technology, more and more high-throughput quantitative data could be extracted from digital medical images, which has resulted in a new kind of omics-Radiomics. In recent years, in addition to genomics, proteomics and metabolomics, radiomic has attracted the interest of more and more researchers. Compared to other omics, radiomics can be perfectly integrated with clinical data, even with the pathology and molecular biomarker, so that the study can be closer to the clinical reality and more revealing of the tumor development. Mass data will also be generated in this process. Machine learning, due to its own characteristics, has a unique advantage in processing massive radiomic data. By analyzing mass amounts of data with strong clinical relevance, people can construct models that more accurately reflect tumor development and progression, thereby providing the possibility of personalized and sequential treatment of patients. As one of the cancer types whose treatment and diagnosis rely on imaging examination, radiomics has a very broad application prospect in head and neck cancers (HNC). Until now, there have been some notable results in HNC. In this review, we will introduce the concepts and workflow of radiomics and machine learning and their current applications in head and neck cancers, as well as the directions and applications of artificial intelligence in the treatment and diagnosis of HNC.

Artificial intelligence applications for oncological positron emission tomography imaging

Positron emission tomography (PET), a functional and dynamic molecular imaging technique, is generally used to reveal tumors' biological behavior. Radiomics allows a high-throughput extraction of multiple features from images with artificial intelligence (AI) approaches and develops rapidly worldwide. Quantitative and objective features of medical images have been explored to recognize reliable biomarkers, with the development of PET radiomics. This paper will review the current clinical exploration of PET-based classical machine learning and deep learning methods, including disease diagnosis, the prediction of histological subtype, gene mutation status, tumor metastasis, tumor relapse, therapeutic side effects, therapeutic intervention and evaluation of prognosis. The applications of AI in oncology will be mainly discussed. The image-guided biopsy or surgery assisted by PET-based AI will be introduced as well. This paper aims to present the applications and methods of AI for PET imaging, which may offer important details for further clinical studies. Relevant precautions are put forward and future research directions are suggested.

Performance of 18F-FDG PET/CT Radiomics for Predicting EGFR Mutation Status in Patients With Non-Small Cell Lung Cancer

Objective

To assess the performance of pretreatment ¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) radiomics features for predicting EGFR mutation status in patients with non-small cell lung cancer (NSCLC).

Patients and Methods

We enrolled total 173 patients with histologically proven NSCLC who underwent preoperative ¹⁸F-FDG PET/CT. Tumor tissues of all patients were tested for EGFR mutation status. A PET/CT radiomics prediction model was established through multi-step feature selection. The predictive performances of radiomics model, clinical features and conventional PET-derived semi-quantitative parameters were compared using receiver operating curves (ROCs) analysis.

Results

Four CT and two PET radiomics features were finally selected to build the PET/CT radiomics model. Compared with area under the ROC curve (AUC) equal to 0.664, 0.683 and 0.662 for clinical features, maximum standardized uptake values (SUV_max) and total lesion glycolysis (TLG), the PET/CT radiomics model showed better performance to discriminate between EGFR positive and negative mutations with the AUC of 0.769 and the accuracy of 67.06% after 10-fold cross-validation. The combined model, based on the PET/CT radiomics and clinical feature (gender) further improved the AUC to 0.827 and the accuracy to 75.29%. Only one PET radiomics feature demonstrated significant but low predictive ability (AUC = 0.661) for differentiating 19 Del from 21 L858R mutation subtypes.

Conclusions

EGFR mutations status in patients with NSCLC could be well predicted by the combined model based on ¹⁸F-FDG PET/CT radiomics and clinical feature, providing an alternative useful method for the selection of targeted therapy.

Reinventing radiation therapy with machine learning and imaging bio-markers (radiomics): State-of-the-art, challenges and perspectives

Radiation therapy is a pivotal cancer treatment that has significantly progressed over the last decade due to numerous technological breakthroughs. Imaging is now playing a critical role on deployment of the clinical workflow, both for treatment planning and treatment delivery. Machine-learning analysis of predefined features extracted from medical images, i.e. radiomics, has emerged as a promising clinical tool for a wide range of clinical problems addressing drug development, clinical diagnosis, treatment selection and implementation as well as prognosis. Radiomics denotes a paradigm shift redefining medical images as a quantitative asset for data-driven precision medicine. The adoption of machine-learning in a clinical setting and in particular of radiomics features requires the selection of robust, representative and clinically interpretable biomarkers that are properly evaluated on a representative clinical data set. To be clinically relevant, radiomics must not only improve patients' management with great accuracy but also be reproducible and generalizable. Hence, this review explores the existing literature and exposes its potential technical caveats, such as the lack of quality control, standardization, sufficient sample size, type of data collection, and external validation. Based upon the analysis of 165 original research studies based on PET, CT-scan, and MRI, this review provides an overview of new concepts, and hypotheses generating findings that should be validated. In particular, it describes evolving research trends to enhance several clinical tasks such as prognostication, treatment planning, response assessment, prediction of recurrence/relapse, and prediction of toxicity. Perspectives regarding the implementation of an AI-based radiotherapy workflow are presented.

FDG PET versus CT radiomics to predict outcome in malignant pleural mesothelioma patients

BACKGROUND

Careful selection of malignant pleural mesothelioma (MPM) patients for curative treatment is of highest importance, as the multimodal treatment regimen is challenging for patients and harbors a high risk of substantial toxicity. Radiomics-a quantitative method for image analysis-has shown its prognostic ability in different tumor entities and could therefore play an important role in optimizing patient selection for radical cancer treatment. So far, radiomics as a prognostic tool in MPM was not investigated.

MATERIALS AND METHODS

This study is based on 72 MPM patients treated with surgery in a curative intent at our institution between 2009 and 2017. Pre-treatment Fluorine-18 fluorodeoxyglucose (FDG) PET and CT scans were used for radiomics outcome modeling. After extraction of 1404 CT and 1410 FDG PET features from each image, a preselection by principal component analysis was performed to include only robust, non-redundant features for the cox regression to predict the progression-free survival (PFS) and the overall survival (OS). Results were validated on a separate cohort. Additionally, SUVmax and SUVmean, and volume were tested for their prognostic ability for PFS and OS.

RESULTS

For the PFS a concordance index (c-index) of 0.67 (95% CI 0.52-0.82) and 0.66 (95% CI 0.57-0.78) for the training cohort (n = 36) and internal validation cohort (n = 36), respectively, were obtained for the PET radiomics model. The PFS advantage of the low-risk group translated also into an OS advantage. On CT images, no radiomics model could be trained. SUV max and SUV mean were also not prognostic in terms of PFS and OS.

CONCLUSION

We were able to build a successful FDG PET radiomics model for the prediction of PFS in MPM. Radiomics could serve as a tool to aid clinical decision support systems for treatment of MPM in future.

Functional imaging using radiomic features in assessment of lymphoma

Lymphomas are typically large, well-defined, and relatively homogeneous tumors, and therefore represent ideal targets for the use of radiomics. Of the available functional imaging tests, [18F]FDG-PET for body lymphoma and diffusion-weighted MRI (DWI) for central nervous system (CNS) lymphoma are of particular interest. The current literature suggests that two main applications for radiomics in lymphoma show promise: differentiation of lymphomas from other tumors, and lymphoma treatment response and outcome prognostication. In particular, encouraging results reported in the limited number of presently available studies that utilize functional imaging suggest that (1) MRI-based radiomics enables differentiation of CNS lymphoma from glioblastoma, and (2) baseline [18F]FDG-PET radiomics could be useful for survival prognostication, adding to or even replacing commonly used metrics such as standardized uptake values and metabolic tumor volume. However, due to differences in biological and clinical characteristics of different lymphoma subtypes and an increasing number of treatment options, more data are required to support these findings. Furthermore, a consensus on several critical steps in the radiomics workflow -most importantly, image reconstruction and post processing, lesion segmentation, and choice of classification algorithm- is desirable to ensure comparability of results between research institutions.

Introduction to Radiomics

Radiomics is a rapidly evolving field of research concerned with the extraction of quantitative metrics-the so-called radiomic features-within medical images. Radiomic features capture tissue and lesion characteristics such as heterogeneity and shape and may, alone or in combination with demographic, histologic, genomic, or proteomic data, be used for clinical problem solving. The goal of this continuing education article is to provide an introduction to the field, covering the basic radiomics workflow: feature calculation and selection, dimensionality reduction, and data processing. Potential clinical applications in nuclear medicine that include PET radiomics-based prediction of treatment response and survival will be discussed. Current limitations of radiomics, such as sensitivity to acquisition parameter variations, and common pitfalls will also be covered.

Robustness of radiomic features in magnetic resonance imaging: review and a phantom study

Radiomic analysis has exponentially increased the amount of quantitative data that can be extracted from a single image. These imaging biomarkers can aid in the generation of prediction models aimed to further personalized medicine. However, the generalizability of the model is dependent on the robustness of these features. The purpose of this study is to review the current literature regarding robustness of radiomic features on magnetic resonance imaging. Additionally, a phantom study is performed to systematically evaluate the behavior of radiomic features under various conditions (signal to noise ratio, region of interest delineation, voxel size change and normalization methods) using intraclass correlation coefficients. The features extracted in this phantom study include first order, shape, gray level cooccurrence matrix and gray level run length matrix. Many features are found to be non-robust to changing parameters. Feature robustness assessment prior to feature selection, especially in the case of combining multi-institutional data, may be warranted. Further investigation is needed in this area of research.

11C-methionine-PET for differentiating recurrent brain tumor from radiation necrosis: radiomics approach with random forest classifier

Differentiating recurrent brain tumor from radiation necrosis is often difficult. This study aims to investigate the efficacy of 11C-methionine (MET)-PET radiomics for distinguishing recurrent brain tumor from radiation necrosis, as compared with conventional tumor-to-normal cortex (T/N) ratio evaluation. We enrolled 41 patients with metastatic brain tumor or glioma treated using radiation therapy who underwent MET-PET. The area with a standardized uptake value > 1.3 times that of the normal brain cortex was contoured. Forty-two PET features were extracted and used in a random forest classifier and the diagnostic performance was evaluated using a 10-fold cross-validation scheme. Gini index was measured to identify relevant PET parameters for classification. The reference standard was surgical histopathological analysis or more than 6 months of follow-up with MRI. Forty-four lesions were used for the analysis. Thirty-three and 11 lesions were confirmed as recurrent brain tumor and radiation necrosis, respectively. Radiomics and T/N ratio evaluation showed sensitivities of 90.1% and 60.6%, and specificities of 93.9% and 72.7% with areas under the curve of 0.98 and 0.73, respectively. Gray level co-occurrence matrix dissimilarity was the most pertinent feature for diagnosis. MET-PET radiomics yielded excellent outcome for differentiating recurrent brain tumor from radiation necrosis, which outperformed T/N ratio evaluation.

Experimental Multicenter and Multivendor Evaluation of the Performance of PET Radiomic Features Using 3-Dimensionally Printed Phantom Inserts

The sensitivity of radiomic features to several confounding factors, such as reconstruction settings, makes clinical use challenging. To investigate the impact of harmonized image reconstructions on feature consistency, a multicenter phantom study was performed using 3-dimensionally printed phantom inserts reflecting realistic tumor shapes and heterogeneity uptakes. Methods: Tumors extracted from real PET/CT scans of patients with non-small cell lung cancer served as model for three 3-dimensionally printed inserts. Different heterogeneity pattern were realized by printing separate compartments that could be filled with different activity solutions. The inserts were placed in the National Electrical Manufacturers Association image-quality phantom and scanned various times. First, a list-mode scan was acquired and 5 statistically equal replicates were reconstructed. Second, the phantom was scanned 4 times on the same scanner. Third, the phantom was scanned on 6 PET/CT systems. All images were reconstructed using EANM Research Ltd. (EARL)-compliant and locally clinically preferred reconstructions. EARL-compliant reconstructions were performed without (EARL1) or with (EARL2) point-spread function. Images were analyzed with and without resampling to 2-mm cubic voxels. Images were discretized with a fixed bin width (FBW) of 0.25 and a fixed bin number (FBN) of 64. The intraclass correlation coefficient (ICC) of each scan setup was calculated and compared across reconstruction settings. An ICC above 0.75 was regarded as high. Results: The percentage of features yielding a high ICC was largest for the statistically equal replicates (70%-91% for FBN; 90%-96% for FBW discretization). For scans acquired on the same system, the percentage decreased, but most features still resulted in a high ICC (FBN, 52%-63%; FBW, 75%-85%). The percentage of features yielding a high ICC decreased more in the multicenter setting. In this case, the percentage of features yielding a high ICC was larger for images reconstructed with EARL-compliant reconstructions: for example, 40% for EARL1 and 60% for EARL2 versus 21% for the clinically preferred setting for FBW discretization. When discretized with FBW and resampled to isotropic voxels, this benefit was more pronounced. Conclusion: EARL-compliant reconstructions harmonize a wide range of radiomic features. FBW discretization and a sampling to isotropic voxels enhances the benefits of EARL-compliant reconstructions.

Radiomics study for differentiating gastric cancer from gastric stromal tumor based on contrast-enhanced CT images

PURPOSE

To test the feasibility of differentiate gastric cancer from gastric stromal tumor using a radiomics study based on contrast-enhanced CT images.

MATERIALS AND METHODS

The contrast-enhanced CT image data of 60 patients with gastric cancer and 40 patients with gastric stromal tumor confirmed by postoperative pathology were retrospectively analyzed. First, CT images were read by two senior radiologists to acquire subjective CT signs model, including perigastric fatty infiltration, perigastric enlarged lymph nodes, the enhancement and growth modes of gastric tumors. Second, the manual segmentation of gastric tumors from the CT images was performed by the two radiologists to extract radiomics features via ITK-SNAP software, and to construct radiomics signature model. Finally, a diagnostic model integrated with subjective CT signs and radiomics signatures was constructed. The diagnostic efficacy of three models in differentiating gastric cancer from gastric stromal tumor was compared by using receiver operating characteristic curves (ROC).

RESULTS

There are statistically significant differences between the gastric cancer and gastric stromal tumor in the perigastric enlarged lymph nodes, growth mode and radiomics signature (p < 0.05). The area under ROC curve (AUC), sensitivity and accuracy of subjective CT signs model were the lowest among the three models. While the combined model yields the highest AUC value (0.903), specificity (93.33%) and accuracy (86.00%) among the three models (p = 0.03).

CONCLUSION

The diagnostic model integrating subjective CT signs and radiomics signature can improve the diagnostic accuracy of gastric tumors.