High quality machine-robust image features: identification in nonsmall cell lung cancer computed tomography images

Research on multi-model imaging machine learning for distinguishing early hepatocellular carcinoma

OBJECTIVE

To investigate the value of differential diagnosis of hepatocellular carcinoma (HCC) and non-hepatocellular carcinoma (non-HCC) based on CT and MR multiphase radiomics combined with different machine learning models and compare the diagnostic efficacy between different radiomics models.

BACKGROUND

Primary liver cancer is one of the most common clinical malignancies, hepatocellular carcinoma (HCC) is the most common subtype of primary liver cancer, accounting for approximately 90% of cases. A clear diagnosis of HCC is important for the individualized treatment of patients with HCC. However, more sophisticated diagnostic modalities need to be explored.

METHODS

This retrospective study included 211 patients with liver lesions: 97 HCC and 124 non-hepatocellular carcinoma (non-HCC) who underwent CT and MRI. Imaging data were used to obtain imaging features of lesions and radiomics regions of interest (ROI). The extracted imaging features were combined to construct different radiomics models. The clinical data and imaging features were then combined with radiomics features to construct the combined models. Support Vector Machine (SVM), K-nearest Neighbor (KNN), RandomForest (RF), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), Multilayer Perceptron (MLP) six machine learning models were used for training. Five-fold cross-validation was used to train the models, and ROC curves were used to analyze the diagnostic efficacy of each model and calculate the accuracy rate. Model training and efficacy test were performed as before.

RESULTS

Statistical analysis showed that some clinical data (gender and concomitant cirrhosis) and imaging features (presence of envelope, marked enhancement in the arterial phase, rapid contouring in the portal phase, uniform density/signal and concomitant steatosis) were statistical differences (P < 0.001). The results of machine learning models showed that KNN had the best diagnostic efficacy. The results of the combined model showed that SVM had the best diagnostic efficacy, indicating that the combined model (accuracy 0.824) had better diagnostic efficacy than the radiomics-only model.

CONCLUSIONS

Our results demonstrate that the radiomic features of CT and MRI combined with machine learning models enable differential diagnosis of HCC and non-HCC (malignant, benign). The diagnostic model with dual radiomic had better diagnostic efficacy. The combined model was superior to the radiomic model alone.

Radiomics signature based on robust features derived from diffusion data for differentiation between benign and malignant solitary pulmonary lesions

BACKGROUND

Classifying and characterizing pulmonary lesions are critical for clinical decision-making process to identify optimal therapeutic strategies. The purpose of this study was to develop and validate a radiomics nomogram for distinguishing between benign and malignant pulmonary lesions based on robust features derived from diffusion images.

MATERIAL AND METHODS

The study was conducted in two phases. In the first phase, we prospectively collected 30 patients with pulmonary nodule/mass who underwent twice EPI-DWI scans. The robustness of features between the two scans was evaluated using the concordance correlation coefficient (CCC) and dynamic range (DR). In the second phase, 139 patients who underwent pulmonary DWI were randomly divided into training and test sets in a 7:3 ratio. Maximum relevance minimum redundancy, least absolute shrinkage and selection operator, and logistic regression were used for feature selection and construction of radiomics signatures. Nomograms were established incorporating clinical features, radiomics signatures, and ADC(0, 800). The diagnostic efficiency of different models was evaluated using the area under the curve (AUC) and decision curve analysis.

RESULTS

Among the features extracted from DWI and ADC images, 42.7% and 37.4% were stable (both CCC and DR ≥ 0.85). The AUCs for distinguishing pulmonary lesions in the test set for clinical model, ADC, ADC radiomics signatures, and DWI radiomics signatures were 0.694, 0.802, 0.885, and 0.767, respectively. The nomogram exhibited the best differentiation performance (AUC = 0.923). The decision curve showed that the nomogram consistently outperformed ADC value and clinical model in lesion differentiation.

CONCLUSION

Our study demonstrates the robustness of radiomics features derived from lung DWI. The ADC radiomics nomogram shows superior clinical net benefits compared to conventional clinical models or ADC values alone in distinguishing solitary pulmonary lesions, offering a promising tool for noninvasive, precision diagnosis in lung cancer.

DiffusionCT: Latent Diffusion Model for CT Image Standardization

Computed tomography (CT) is one of the modalities for effective lung cancer screening, diagnosis, treatment, and prognosis. The features extracted from CT images are now used to quantify spatial and temporal variations in tumors. However, CT images obtained from various scanners with customized acquisition protocols may introduce considerable variations in texture features, even for the same patient. This presents a fundamental challenge to downstream studies that require consistent and reliable feature analysis. Existing CT image harmonization models rely on GAN-based supervised or semi-supervised learning, with limited performance. This work addresses the issue of CT image harmonization using a new diffusion-based model, named DiffusionCT, to standardize CT images acquired from different vendors and protocols. DiffusionCT operates in the latent space by mapping a latent non-standard distribution into a standard one. DiffusionCT incorporates a U-Net-based encoder-decoder, augmented by a diffusion model integrated into the bottleneck part. The model is designed in two training phases. The encoder-decoder is first trained, without embedding the diffusion model, to learn the latent representation of the input data. The latent diffusion model is then trained in the next training phase while fixing the encoder-decoder. Finally, the decoder synthesizes a standardized image with the transformed latent representation. The experimental results demonstrate a significant improvement in the performance of the standardization task using DiffusionCT.

A study on the differential of solid lung adenocarcinoma and tuberculous granuloma nodules in CT images by Radiomics machine learning

To study the classification efficiency of using texture feature machine learning method in distinguishing solid lung adenocarcinoma (SADC) and tuberculous granulomatous nodules (TGN) that appear as solid nodules (SN) in non-enhanced CT images. 200 patients with SADC and TGN who underwent thoracic non-enhanced CT examination from January 2012 to October 2019 were included in the study, 490 texture eigenvalues of 6 categories were extracted from the lesions in the non-enhanced CT images of these patients for machine learning, the classification prediction model is established by using relatively the best classifier selected according to the fitting degree of learning curve in the process of machine learning, and the effectiveness of the model was tested and verified. The logistic regression model of clinical data (including demographic data and CT parameters and CT signs of solitary nodules) was used for comparison. The prediction model of clinical data was established by logistic regression, and the classifier was established by machine learning of radiologic texture features. The area under the curve was 0.82 and 0.65 for the prediction model based on clinical CT and only CT parameters and CT signs, and 0.870 based on Radiomics characteristics. The machine learning prediction model developed by us can improve the differentiation efficiency of SADC and TGN with SN, and provide appropriate support for treatment decisions.

A Comprehensive Survey on the Progress, Process, and Challenges of Lung Cancer Detection and Classification

Lung cancer is the primary reason of cancer deaths worldwide, and the percentage of death rate is increasing step by step. There are chances of recovering from lung cancer by detecting it early. In any case, because the number of radiologists is limited and they have been working overtime, the increase in image data makes it hard for them to evaluate the images accurately. As a result, many researchers have come up with automated ways to predict the growth of cancer cells using medical imaging methods in a quick and accurate way. Previously, a lot of work was done on computer-aided detection (CADe) and computer-aided diagnosis (CADx) in computed tomography (CT) scan, magnetic resonance imaging (MRI), and X-ray with the goal of effective detection and segmentation of pulmonary nodule, as well as classifying nodules as malignant or benign. But still, no complete comprehensive review that includes all aspects of lung cancer has been done. In this paper, every aspect of lung cancer is discussed in detail, including datasets, image preprocessing, segmentation methods, optimal feature extraction and selection methods, evaluation measurement matrices, and classifiers. Finally, the study looks into several lung cancer-related issues with possible solutions.

Reproducibility of radiomics features from ultrasound images: influence of image acquisition and processing

OBJECTIVES

To systematically assess the reproducibility of radiomics features from ultrasound (US) images during image acquisition and processing.

MATERIALS AND METHODS

A standardized phantom was scanned to obtain US images. Reproducibility of radiomics features from US images, also known as ultrasomics features, was explored via (a) intra-US machine: changing the US acquisition parameters including gain, focus, and frequency; (b) inter-US machine: comparing three different scanners; (c) changing segmentation locations; and (d) inter-platform: comparing features extracted by the Ultrasomics and PyRadiomics algorithm platforms. Reproducible ultrasomics features were selected based on coefficients of variation.

RESULTS

A total of 108 US images from three scanners were obtained; 5253 ultrasomics features including seven categories of features were extracted and evaluated for each US image. From intra-US machine analysis, 37.0-38.8% of features showed good reproducibility. From inter-US machine analysis, 42.8% (2248/5253) of features exhibited good reproducibility. From segmentation location analysis, 55.7-57.6% of features showed good reproducibility. No significant difference in the normalized feature ranges was found between the 100 features extracted by the Ultrasomics and PyRadiomics platforms with the same algorithm (p = 0.563). A total of 1452 (27.6%) ultrasomics features were reproducible whenever intra-/inter-US machine or segmentation location were changed, most of which were wavelet and shearlet features.

CONCLUSIONS

Different acquisition parameters, US scanners, segmentation locations, and feature extraction platforms affected the reproducibility of ultrasomics features. Wavelet and shearlet features showed the best reproducibility across all procedures.

KEY POINTS

• Different acquisition parameters, US scanners, segmentation locations, and feature extraction platforms affected the reproducibility of ultrasomics features. • A total of 1452 (27.6%) ultrasomics features were reproducible whenever intra-/inter-US machine or segmentation location were changed. • Wavelet and shearlet features showed the best reproducibility across all procedures.

Cross-Vendor CT Image Data Harmonization Using CVH-CT

While remarkable advances have been made in Computed Tomography (CT), most of the existing efforts focus on imaging enhancement while reducing radiation dose. How to harmonize CT image data captured using different scanners is vital in cross-center large-scale radiomics studies but remains the boundary to explore. Furthermore, the lack of paired training image problem makes it computationally challenging to adopt existing deep learning models. We propose a novel deep learning approach called CVH-CT for harmonizing CT images captured using scanners from different vendors. The generator of CVH-CT uses a self-attention mechanism to learn the scanner-related information. We also propose a VGG feature based domain loss to effectively extract texture properties from unpaired image data to learn the scanner based texture distributions. The experimental results show that CVH-CT is clearly better than the baselines because of the use of the proposed domain loss, and CVH-CT can effectively reduce the scanner-related variability in terms of radiomic features.

Non-small cell lung carcinoma histopathological subtype phenotyping using high-dimensional multinomial multiclass CT radiomics signature

OBJECTIVE

The aim of this study was to identify the most important features and assess their discriminative power in the classification of the subtypes of NSCLC.

METHODS

This study involved 354 pathologically proven NSCLC patients including 134 squamous cell carcinoma (SCC), 110 large cell carcinoma (LCC), 62 not other specified (NOS), and 48 adenocarcinoma (ADC). In total, 1433 radiomics features were extracted from 3D volumes of interest drawn on the malignant lesion identified on CT images. Wrapper algorithm and multivariate adaptive regression splines were implemented to identify the most relevant/discriminative features. A multivariable multinomial logistic regression was employed with 1000 bootstrapping samples based on the selected features to classify four main subtypes of NSCLC.

RESULTS

The results revealed that the texture features, specifically gray level size zone matrix features (GLSZM), were the significant indicators of NSCLC subtypes. The optimized classifier achieved an average precision, recall, F1-score, and accuracy of 0.710, 0.703, 0.706, and 0.865, respectively, based on the selected features by the wrapper algorithm.

CONCLUSIONS

Our CT radiomics approach demonstrated impressive potential for the classification of the four main histological subtypes of NSCLC, It is anticipated that CT radiomics could be useful in treatment planning and precision medicine.

Deep Feature Stability Analysis Using CT Images of a Physical Phantom Across Scanner Manufacturers, Cartridges, Pixel Sizes, and Slice Thickness

Image acquisition parameters for computed tomography scans such as slice thickness and field of view may vary depending on tumor size and site. Recent studies have shown that some radiomics features were dependent on voxel size (= pixel size × slice thickness), and with proper normalization, this voxel size dependency could be reduced. Deep features from a convolutional neural network (CNN) have shown great promise in characterizing cancers. However, how do these deep features vary with changes in imaging acquisition parameters? To analyze the variability of deep features, a physical radiomics phantom with 10 different material cartridges was scanned on 8 different scanners. We assessed scans from 3 different cartridges (rubber, dense cork, and normal cork). Deep features from the penultimate layer of the CNN before (pre-rectified linear unit) and after (post-rectified linear unit) applying the rectified linear unit activation function were extracted from a pre-trained CNN using transfer learning. We studied both the interscanner and intrascanner dependency of deep features and also the deep features' dependency over the 3 cartridges. We found some deep features were dependent on pixel size and that, with appropriate normalization, this dependency could be reduced. False discovery rate was applied for multiple comparisons, to mitigate potentially optimistic results. We also used stable deep features for prognostic analysis on 1 non-small cell lung cancer data set.

Reproducibility of lung nodule radiomic features: Multivariable and univariable investigations that account for interactions between CT acquisition and reconstruction parameters

PURPOSE

Recent studies have demonstrated a lack of reproducibility of radiomic features in response to variations in CT parameters. In addition, reproducibility of radiomic features has not been well established in clinical datasets. We aimed to investigate the effects of a wide range of CT acquisition and reconstruction parameters on radiomic features in a realistic setting using clinical low dose lung cancer screening cases. We performed univariable and multivariable explorations to consider the effects of individual parameters and the simultaneous interactions between three different acquisition/reconstruction parameters of radiation dose level, reconstructed slice thickness, and kernel.

METHOD

A cohort of 89 lung cancer screening patients were collected that each had a solid lung nodule >4mm diameter. A computational pipeline was used to perform a simulation of dose reduction of the raw projection data, collected from patient scans. This was followed by reconstruction of raw data with weighted filter back projection (wFBP) algorithm and automatic lung nodule detection and segmentation using a computer-aided detection tool. For each patient, 36 different image datasets were created corresponding to dose levels of 100%, 50%, 25%, and 10% of the original dose level, three slice thicknesses of 0.6 mm, 1 mm, and 2 mm, as well as three reconstruction kernels of smooth, medium, and sharp. For each nodule, 226 well-known radiomic features were calculated at each image condition. The reproducibility of radiomic features was first evaluated by measuring the intercondition agreement of the feature values among the 36 image conditions. Then in a series of univariable analyses, the impact of individual CT parameters was assessed by selecting subsets of conditions with one varying and two constant CT parameters. In each subset, intraparameter agreements were assessed. Overall concordance correlation coefficient (OCCC) served as the measure of agreement. An OCCC ≥ 0.9 implied strong agreement and reproducibility of radiomic features in intercondition or intraparameter comparisons. Furthermore, the interaction of CT parameters in impacting radiomic feature values was investigated via ANOVA.

RESULTS

All included radiomic features lacked intercondition reproducibility (OCCC < 0.9) among all the 36 conditions. Out of 226 radiomic features analyzed, only 17 and 18 features were considered reproducible (OCCC ≥ 0.9) to dose and kernel variation, respectively, within the corresponding condition subsets. Slice thickness demonstrated the largest impact on radiomic feature values where only one to five features were reproducible at a few condition subsets. ANOVA revealed significant interactions (P < 0.05) between CT parameters affecting the variability of >50% of radiomic features.

CONCLUSION

We systematically explored the multidimensional space of CT parameters in affecting lung nodule radiomic features. Univariable and multivariable analyses of this study not only showed the lack of reproducibility of the majority of radiomic features but also revealed existing interactions among CT parameters, meaning that the effect of individual CT parameters on radiomic features can be conditional upon other CT acquisition and reconstruction parameters. Our findings advise on careful radiomic feature selection and attention to the inclusion criteria for CT image acquisition protocols within the datasets of radiomic studies.

Reproducibility of radiomics features derived from intravoxel incoherent motion diffusion-weighted MRI of cervical cancer

BACKGROUND

The reproducibility of intravoxel incoherent motion (IVIM)-based radiomics studies in humans has not been reported.

PURPOSE

To determine the inter- and intra-observer variability on the reproducibility of IVIM-based radiomics features in cervical cancer (CC).

MATERIAL AND METHODS

The IVIM images of 25 patients with CC were retrospectively collected. Based on the high-resolution T2-weighted images, the regions of interest (ROIs) were independently delineated twice in diffusion-weighted images at a b value of 1000 s/mm² (interval time was one month) by two radiologists. This was done at the largest transversal cross-sections of the tumors. The ROI was subsequently used in apparent diffusion coefficient (ADC), true diffusion coefficient (D), pseudo-diffusion coefficient (D*), and perfusion fraction (f) maps derived from IVIM images. In total, 105 radiomics features were then finally extracted from the IVIM-derived maps. The inter- and intra-observer reproducibility of IVIM-derived features was then evaluated using the intraclass correlation coefficient.

RESULTS

Inter- and intra-observer variability affected the reproducibility of radiomics features. D* map had 100% and 95% reproducible features, ADC map had 89% and 93%, D map had 97% and 86%, while f map had 54% and 62% reproducible features with good to excellent reliability in the intra-observer analysis. Similarly, D* map had 90% and 94%, ADC map had 85% and 70%, D map had 81% and 78%, while f map had 41% and 93% reproducible features with good to excellent reliability in the inter-observer analysis.

CONCLUSION

Inter- and intra-observer variability can affect radiomics analysis. Cognizant to this, multicenter studies should pay more attention to intra- and inter-observer variability.

Computed-Tomography-Based Radiomics Model for Predicting the Malignant Potential of Gastrointestinal Stromal Tumors Preoperatively: A Multi-Classifier and Multicenter Study

Background: Our goal was to establish and verify a radiomics risk grading model for gastrointestinal stromal tumors (GISTs) and to identify the optimal algorithm for risk stratification. Methods: We conducted a retrospective analysis of 324 patients with GISTs, the presence of which was confirmed by surgical pathology. Patients were treated at three different hospitals. A training cohort of 180 patients was collected from the largest center, while an external validation cohort of 144 patients was collected from the other two centers. To extract radiomics features, regions of interest (ROIs) were outlined layer by layer along the edge of the tumor contour on CT images of the arterial and portal venous phases. The dimensionality of radiomic features was reduced, and the top 10 features with importance value above 5 were selected before modeling. The training cohort used three classifiers [logistic regression, support vector machine (SVM), and random forest] to establish three GIST risk stratification prediction models. The receiver operating characteristic curve (ROC) was used to compare model performance, which was validated by external data. Results: In the training cohort, the average area under the curve (AUC) was 0.84 ± 0.07 of the logistic regression, 0.88 ± 0.06 of the random forest, and 0.81 ± 0.08 of the SVM. In the external validation cohort, the AUC was 0.85 of the logistic regression, 0.90 of the random forest, and 0.80 of the SVM. The random forest model performed the best in both the training and the external validation cohorts and could be generalized. Conclusion: Based on CT radiomics, there are multiple machine-learning models that can predict the risk of GISTs. Among them, the random forest algorithm had the highest prediction efficiency and could be readily generalizable. Through external validation data, we assume that the random forest model may be used as an effective tool to guide preoperative clinical decision-making.

The impact of the variation of imaging parameters on the robustness of Computed Tomography radiomic features: A review

The field of radiomics is at the forefront of personalized medicine. However, there is concern that high variation in imaging parameters will impact robustness of radiomic features and subsequently the performance of the predictive models built upon them. Therefore, our review aims to evaluate the impact of imaging parameters on the robustness of radiomic features. We also provide insights into the validity and discrepancy of different methodologies applied to investigate the robustness of radiomic features. We selected 47 papers based on our predefined inclusion criteria and grouped these papers by the imaging parameter under investigation: (i) scanner parameters, (ii) acquisition parameters and (iii) reconstruction parameters. Our review highlighted that most of the imaging parameters are disruptive parameters, and shape along with First order statistics were reported as the most robust radiomic features against variation in imaging parameters. This review identified inconsistencies related to the methodology of the reviewed studies such as the metrics used for robustness, the feature extraction techniques, the reporting style, and their outcome inclusion. We hope this review will aid the scientific community in conducting research in a way that is more reproducible and avoids the pitfalls of previous analyses.

Understanding Sources of Variation to Improve the Reproducibility of Radiomics

Radiomics is the method of choice for investigating the association between cancer imaging phenotype, cancer genotype and clinical outcome prediction in the era of precision medicine. The fast dispersal of this new methodology has benefited from the existing advances of the core technologies involved in radiomics workflow: image acquisition, tumor segmentation, feature extraction and machine learning. However, despite the rapidly increasing body of publications, there is no real clinical use of a developed radiomics signature so far. Reasons are multifaceted. One of the major challenges is the lack of reproducibility and generalizability of the reported radiomics signatures (features and models). Sources of variation exist in each step of the workflow; some are controllable or can be controlled to certain degrees, while others are uncontrollable or even unknown. Insufficient transparency in reporting radiomics studies further prevents translation of the developed radiomics signatures from the bench to the bedside. This review article first addresses sources of variation, which is illustrated using demonstrative examples. Then, it reviews a number of published studies and progresses made to date in the investigation and improvement of feature reproducibility and model performance. Lastly, it discusses potential strategies and practical considerations to reduce feature variability and improve the quality of radiomics study. This review focuses on CT image acquisition, tumor segmentation, quantitative feature extraction, and the disease of lung cancer.

Ultrasensitive 3D Aerosol-Jet-Printed Perovskite X-ray Photodetector

X-ray photon detection is important for a wide range of applications. The highest demand, however, comes from medical imaging, which requires cost-effective, high-resolution detectors operating at low-photon flux, therefore stimulating the search for novel materials and new approaches. Recently, hybrid halide perovskite CH₃NH₃PbI₃ (MAPbI₃) has attracted considerable attention due to its advantageous optoelectronic properties and low fabrication costs. The presence of heavy atoms, providing a high scattering cross-section for photons, makes this material a perfect candidate for X-ray detection. Despite the already-successful demonstrations of efficiency in detection, its integration into standard microelectronics fabrication processes is still pending. Here, we demonstrate a promising method for building X-ray detector units by 3D aerosol jet printing with a record sensitivity of 2.2 × 10⁸ μC Gy_air^-1 cm^-2 when detecting 8 keV photons at dose rates below 1 μGy/s (detection limit 0.12 μGy/s), a 4-fold improvement on the best-in-class devices. An introduction of MAPbI₃-based detection into medical imaging would significantly reduce health hazards related to the strongly ionizing X-rays' photons.

Distinguishing granulomas from adenocarcinomas by integrating stable and discriminating radiomic features on non-contrast computed tomography scans

OBJECTIVE

To identify stable and discriminating radiomic features on non-contrast CT scans to develop more generalisable radiomic classifiers for distinguishing granulomas from adenocarcinomas.

METHODS

In total, 412 patients with adenocarcinomas and granulomas from three institutions were retrospectively included. Segmentations of the lung nodules were performed manually by an expert radiologist in a 2D axial view. Radiomic features were extracted from intra- and perinodular regions. A total of 145 patients were used as part of the training set (S_tr), whereas 205 patients were used as part of test set I (S_te¹) and 62 patients were used as part of independent test set II (S_te²). To mitigate the variation of CT acquisition parameters, we defined 'stable' radiomic features as those for which the feature expression remains relatively unchanged between different sites, as assessed using a Wilcoxon rank-sum test. These stable features were used to develop more generalisable radiomic classifiers that were more resilient to variations in lung CT scans. Features were ranked based on two criteria, firstly based on discriminability (i.e. maximising AUC) alone and subsequently based on maximising both feature stability and discriminability. Different machine-learning classifiers (Linear discriminant analysis, Quadratic discriminant analysis, Support vector machines and random forest) were trained with features selected using the two different criteria and then compared on the two independent test sets for distinguishing granulomas from adenocarcinomas, in terms of area under the receiver operating characteristic curve.

RESULTS

In the test sets, classifiers constructed using the criteria involving maximising feature stability and discriminability simultaneously achieved higher AUC compared with the discriminating alone criteria (S_te¹ [n = 205]: maximum AUCs of 0.85versus . 0.80; p-value = 0.047 and S_te² [n = 62]: maximum AUCs of 0.87 versus. 0.79; p-value = 0.021). These differences held for features extracted from scans with <3 mm slice thickness (AUC = 0.88 versus. 0.80; p-value = 0.039, n = 100) and for the ≥3 mm cases (AUC = 0.81 versus. 0.76; p-value = 0.034, n = 105). In both experiments, shape and peritumoural texture features had a higher stability compared with intratumoural texture features.

CONCLUSIONS

Our study suggests that explicitly accounting for both stability and discriminability results in more generalisable radiomic classifiers to distinguish adenocarcinomas from granulomas on non-contrast CT scans. Our results also showed that peritumoural texture and shape features were less affected by the scanner parameters compared with intratumoural texture features; however, they were also less discriminating compared with intratumoural features.

Robustness of magnetic resonance radiomic features to pixel size resampling and interpolation in patients with cervical cancer

Background

Radiomics is a promising field in oncology imaging. However, the implementation of radiomics clinically has been limited because its robustness remains unclear. Previous CT and PET studies suggested that radiomic features were sensitive to variations in pixel size and slice thickness of the images. The purpose of this study was to assess robustness of magnetic resonance (MR) radiomic features to pixel size resampling and interpolation in patients with cervical cancer.

Methods

This retrospective study included 254 patients with a pathological diagnosis of cervical cancer stages IB to IVA who received definitive chemoradiation at our institution between January 2006 and June 2020. Pretreatment MR scans were analyzed. Each region of cervical cancer was segmented on the axial gadolinium-enhanced T1- and T2-weighted images; 107 radiomic features were extracted. MR scans were interpolated and resampled using various slice thicknesses and pixel spaces. Intraclass correlation coefficients (ICCs) were calculated between the original images and images that underwent pixel size resampling (OP), interpolation (OI), or pixel size resampling and interpolation (OP+I) as well as among processed image sets with various pixel spaces (P), various slice thicknesses (I), and both (P + I).

Results

After feature standardization, ≥86.0% of features showed good robustness when compared between the original and processed images (OP, OI, and OP+I) and ≥ 88.8% of features showed good robustness when processed images were compared (P, I, and P + I). Although most first-order, shape, and texture features showed good robustness, GLSZM small-area emphasis-related features and NGTDM strength were sensitive to variations in pixel size and slice thickness.

Conclusion

Most MR radiomic features in patients with cervical cancer were robust after pixel size resampling and interpolation following the feature standardization process. The understanding regarding the robustness of individual features after pixel size resampling and interpolation could help future radiomics research.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40644-021-00388-5.

Comparison of Radiomic Models Based on Low-Dose and Standard-Dose CT for Prediction of Adenocarcinomas and Benign Lesions in Solid Pulmonary Nodules

Objectives

This study aimed to develop radiomic models based on low-dose CT (LDCT) and standard-dose CT to distinguish adenocarcinomas from benign lesions in patients with solid solitary pulmonary nodules and compare the performance among these radiomic models and Lung CT Screening Reporting and Data System (Lung-RADS). The reproducibility of radiomic features between LDCT and standard-dose CT were also evaluated.

Methods

A total of 141 consecutive pathologically confirmed solid solitary pulmonary nodules were enrolled including 50 adenocarcinomas and 48 benign nodules in primary cohort and 22 adenocarcinomas and 21 benign nodules in validation cohort. LDCT and standard-dose CT scans were conducted using same acquisition parameters and reconstruction method except for radiation dose. All nodules were automatically segmented and 104 original radiomic features were extracted. The concordance correlation coefficient was used to quantify reproducibility of radiomic features between LDCT and standard-dose CT. Radiomic features were selected to build radiomic signature, and clinical characteristics and radiomic signature were combined to develop radiomic nomogram for LDCT and standard-dose CT, respectively. The performance of radiomic models and Lung-RADS was assessed by area under curve (AUC) of receiver operating characteristic curve, sensitivity, and specificity.

Results

Shape and first order features, and neighboring gray tone difference matrix features were highly reproducible between LDCT and standard-dose CT. No significant differences of AUCs were found among radiomic signature and nomogram of LDCT and standard-dose CT in both primary and validation cohort (0.915 vs. 0.919 vs. 0.898 vs. 0.909 and 0.976 vs. 0.976 vs. 0.985 vs. 0.987, respectively). These radiomic models had higher specificity than Lung-RADS (all correct P < 0.05), while there were no significant differences of sensitivity between Lung-RADS and radiomic models.

Conclusions

The diagnostic performance of LDCT-based radiomic models to differentiate adenocarcinomas from benign lesions in solid pulmonary nodules were equivalent to that of standard-dose CT. The LDCT-based radiomic model with higher specificity and lower false-positive rate than Lung-RADS might help reduce overdiagnosis and overtreatment of solid pulmonary nodules in lung cancer screening.

STAN-CT: Standardizing CT Image using Generative Adversarial Networks

Computed Tomography (CT) plays an important role in lung malignancy diagnostics, therapy assessment, and facilitating precision medicine delivery. However, the use of personalized imaging protocols poses a challenge in large-scale cross-center CT image radiomic studies. We present an end-to-end solution called STAN-CT for CT image standardization and normalization, which effectively reduces discrepancies in image features caused by using different imaging protocols or using different CT scanners with the same imaging protocol. STAN-CT consists oftwo components: 1)a Generative Adversarial Networks (GAN) model where a latent-feature-based loss function is adopted to learn the data distribution of standard images within a few rounds of generator training, and 2) an automatic DICOM reconstruction pipeline with systematic image quality control that ensures the generation ofhigh-quality standard DICOM images. Experimental results indicate that the training efficiency and model performance of STAN-CT have been significantly improved compared to the state-of-the-art CT image standardization and normalization algorithms.

Enhanced computed tomography radiomics-based machine-learning methods for predicting the Fuhrman grades of renal clear cell carcinoma

OBJECTIVE

To develop and test an optimal machine learning model based on the enhanced computed tomography (CT) to preoperatively predict pathological grade of clear cell renal cell carcinoma (ccRCC).

METHODS

A retrospective analysis of 53 pathologically confirmed cases of ccRCC was performed and 25 consecutive ccRCC cases were selected as a prospective testing set. All patients underwent routine preoperative abdominal CT plain and enhanced scans. Renal tumor lesions were segmented on arterial phase images and 396 radiomics features were extracted. In the training set, seven discrimination classifiers for high- and low-grade ccRCCs were constructed based on seven different machine learning models, respectively, and their performance and stability for predicting ccRCC grades were evaluated through receiver operating characteristic (ROC) analysis and cross-validation. Prediction accuracy and area under ROC curve were used as evaluation indices. Finally, the diagnostic efficacy of the optimal model was verified in the testing set.

RESULTS

The accuracies and AUC values achieved by support vector machine with radial basis function kernel (svmRadial), random forest and naïve Bayesian models were 0.860±0.158 and 0.919±0.118, 0.840±0.160 and 0.915±0.138, 0.839±0.147 and 0.921±0.133, respectively, which showed high predictive performance, whereas K-nearest neighborhood model yielded lower accuracy of 0.720±0.188 and lower AUC value of 0.810±0.150. Additionally, svmRadial had smallest relative standard deviation (RSD, 0.13 for AUC, 0.17 for accuracy), which indicates higher stability.

CONCLUSION

svmRadial performs best in predicting pathological grades of ccRCC using radiomics features computed from the preoperative CT images, and thus may have high clinical potential in guiding preoperative decision.

Survival Prediction in Gallbladder Cancer Using CT Based Machine Learning

Objective

To establish a classifier for accurately predicting the overall survival of gallbladder cancer (GBC) patients by analyzing pre-treatment CT images using machine learning technology.

Methods

This retrospective study included 141 patients with pathologically confirmed GBC. After obtaining the pre-treatment CT images, manual segmentation of the tumor lesion was performed and LIFEx package was used to extract the tumor signature. Next, LASSO and Random Forest methods were used to optimize and model. Finally, the clinical information was combined to accurately predict the survival outcomes of GBC patients.

Results

Fifteen CT features were selected through LASSO and random forest. On the basis of relative importance GLZLM-HGZE, GLCM-homogeneity and NGLDM-coarseness were included in the final model. The hazard ratio of the CT-based model was 1.462(95% CI: 1.014–2.107). According to the median of risk score, all patients were divided into high and low risk groups, and survival analysis showed that high-risk groups had a poor survival outcome (P = 0.012). After inclusion of clinical factors, we used multivariate COX to classify patients with GBC. The AUC values in the test set and validation set for 3 years reached 0.79 and 0.73, respectively.

Conclusion

GBC survival outcomes could be predicted by radiomics based on LASSO and Random Forest.

Are quantitative features of lung nodules reproducible at different CT acquisition and reconstruction parameters?

Consistency and duplicability in Computed Tomography (CT) output is essential to quantitative imaging for lung cancer detection and monitoring. This study of CT-detected lung nodules investigated the reproducibility of volume-, density-, and texture-based features (outcome variables) over routine ranges of radiation dose, reconstruction kernel, and slice thickness. CT raw data of 23 nodules were reconstructed using 320 acquisition/reconstruction conditions (combinations of 4 doses, 10 kernels, and 8 thicknesses). Scans at 12.5%, 25%, and 50% of protocol dose were simulated; reduced-dose and full-dose data were reconstructed using conventional filtered back-projection and iterative-reconstruction kernels at a range of thicknesses (0.6-5.0 mm). Full-dose/B50f kernel reconstructions underwent expert segmentation for reference Region-Of-Interest (ROI) and nodule volume per thickness; each ROI was applied to 40 corresponding images (combinations of 4 doses and 10 kernels). Typical texture analysis metrics (including 5 histogram features, 13 Gray Level Co-occurrence Matrix, 5 Run Length Matrix, 2 Neighboring Gray-Level Dependence Matrix, and 3 Neighborhood Gray-Tone Difference Matrix) were computed per ROI. Reconstruction conditions resulting in no significant change in volume, density, or texture metrics were identified as "compatible pairs" for a given outcome variable. Our results indicate that as thickness increases, volumetric reproducibility decreases, while reproducibility of histogram- and texture-based features across different acquisition and reconstruction parameters improves. To achieve concomitant reproducibility of volumetric and radiomic results across studies, balanced standardization of the imaging acquisition parameters is required.

Radiomics in medical imaging-"how-to" guide and critical reflection

Radiomics is a quantitative approach to medical imaging, which aims at enhancing the existing data available to clinicians by means of advanced mathematical analysis. Through mathematical extraction of the spatial distribution of signal intensities and pixel interrelationships, radiomics quantifies textural information by using analysis methods from the field of artificial intelligence. Various studies from different fields in imaging have been published so far, highlighting the potential of radiomics to enhance clinical decision-making. However, the field faces several important challenges, which are mainly caused by the various technical factors influencing the extracted radiomic features.The aim of the present review is twofold: first, we present the typical workflow of a radiomics analysis and deliver a practical "how-to" guide for a typical radiomics analysis. Second, we discuss the current limitations of radiomics, suggest potential improvements, and summarize relevant literature on the subject.

Exploring technical issues in personalized medicine: NSCLC survival prediction by quantitative image analysis-usefulness of density correction of volumetric CT data

The aim of this study was to apply density correction method to the quantitative image analysis of non-small cell lung cancer (NSCLC) computed tomography (CT) images, determining its influence on overall survival (OS) prediction of surgically treated patients. Clinicopathological (CP) data and preoperative CT scans, pre- and post-contrast medium (CM) administration, of 57 surgically treated NSCLC patients, were retrospectively collected. After CT volumetric density measurement of primary gross tumour volume (GTV), aorta and tracheal air, density correction was conducted on GTV (reference values: aortic blood and tracheal air). For each resulting data set (combining CM administration and normalization), first-order statistical and textural features were extracted. CP and imaging data were correlated with patients 1-, 3- and 5-year OS, alone and combined (uni-/multivariate logistic regression and Akaike information criterion). Predictive performance was evaluated using the ROC curves and AUC values and compared among non-normalized/normalized data sets (DeLong test). The best predictive values were obtained when combining CP and imaging parameters (AUC values: 1 year 0.72; 3 years 0.82; 5 years 0.78). After normalization resulted an improvement in predicting 1-year OS for some of the grey level size zonebased features (large zone low grey level emphasis) and for the combined CP-imaging model, a worse performance for grey level co-occurrence matrix (cluster prominence and shade) and first-order statistical (range) parameters for 1- and 5-year OS, respectively. The negative performance of cluster prominence in predicting 1-year OS was the only statistically significant result (p value 0.05). Density corrections of volumetric CT data showed an opposite influence on the performance of imaging quantitative features in predicting OS of surgically treated NSCLC patients, even if no statistically significant for almost all predictors.

Radiomics: from qualitative to quantitative imaging

Historically, medical imaging has been a qualitative or semi-quantitative modality. It is difficult to quantify what can be seen in an image, and to turn it into valuable predictive outcomes. As a result of advances in both computational hardware and machine learning algorithms, computers are making great strides in obtaining quantitative information from imaging and correlating it with outcomes. Radiomics, in its two forms "handcrafted and deep," is an emerging field that translates medical images into quantitative data to yield biological information and enable radiologic phenotypic profiling for diagnosis, theragnosis, decision support, and monitoring. Handcrafted radiomics is a multistage process in which features based on shape, pixel intensities, and texture are extracted from radiographs. Within this review, we describe the steps: starting with quantitative imaging data, how it can be extracted, how to correlate it with clinical and biological outcomes, resulting in models that can be used to make predictions, such as survival, or for detection and classification used in diagnostics. The application of deep learning, the second arm of radiomics, and its place in the radiomics workflow is discussed, along with its advantages and disadvantages. To better illustrate the technologies being used, we provide real-world clinical applications of radiomics in oncology, showcasing research on the applications of radiomics, as well as covering its limitations and its future direction.

Assessment of Stability and Discrimination Capacity of Radiomic Features on Apparent Diffusion Coefficient Images

The objectives of the study are to develop a new way to assess stability and discrimination capacity of radiomic features without the need of test-retest or multiple delineations and to use information obtained to perform a preliminary feature selection. Apparent diffusion coefficient (ADC) maps were computed from diffusion-weighted magnetic resonance images (DW-MRI) of two groups of patients: 18 with soft tissue sarcomas (STS) and 18 with oropharyngeal cancers (OPC). Sixty-nine radiomic features were computed, using three different histogram discretizations (16, 32, and 64 bins). Geometrical transformations (translations) of increasing entity were applied to the regions of interest (ROIs), and the intra-class correlation coefficient (ICC) was used to compare the features computed on the original and modified ROIs. The distribution of ICC values for minimal and maximal entity translations (ICC₁₀ and ICC₁₀₀, respectively) was used to adjust thresholds of ICC (ICC_min and ICC_max) used to discriminate between good, unstable (ICC₁₀ < ICC_min), and non-discriminative features (ICC₁₀₀ > ICC_max). Fifty-four and 59 radiomic features passed the stability-based selection for all the three histogram discretizations for the OPC and STS datasets, respectively. The excluded features were similar across the different histogram discretizations (Jaccard’s index 0.77 ± 0.13 and 0.9 ± 0.1 for OPC and STS, respectively) but different between datasets (Jaccard’s index 0.19 ± 0.02). The results suggest that the observed radiomic features are mainly stable and discriminative, but the stability depends on the region of the body under observation. The method provides a way to assess stability without the need of test-retest or multiple delineations.

Application of Radiomics in Predicting the Malignancy of Pulmonary Nodules in Different Sizes

OBJECTIVE. The purpose of this study was to investigate the utility of radiomics for predicting the malignancy of pulmonary nodules (PNs) of different sizes using unenhanced, thin-section CT images. MATERIALS AND METHODS. Patients with a single PN (n = 373) who underwent a preoperative chest CT were recruited retrospectively at Beijing Friendship Hospital from March 2015 to March 2018. Of the 373 PNs studied, 192 were benign and 181 were malignant. The lesions were classified into three groups (T1a, T1b, or T1c according to the 8th edition of the TNM staging system for lung cancer) on the basis of lesion diameters: T1a (diameter, 0-1 cm), T1b (1 cm < diameter ≤ 2 cm) and T1c (2 cm < diameter ≤ 3 cm). A total of 1160 radiomic features were extracted from PN segmentation on unenhanced CT images. We developed three radiomic models to predict PN malignancy in each group on the basis of the extracted radiomic features. Fivefold cross-validation was used to estimate AUC, accuracy, sensitivity, and specificity for indicating the performance of prediction models. RESULTS. The AUC, accuracy, sensitivity, and specificity for predicting PN malignancy in each group were 0.84, 0.77, 0.89, and 0.74 with the T1a model; 0.78, 0.73, 0.74, and 0.71 with the T1b model, and 0.79, 0.76, 0.77, and 0.73 with the T1c model, respectively. The most contributive radiomic features for predicting PN malignancy for groups T1a, T1b, and T1c were LoG_X_Uniformity, Intensity_Minimum, and Shape_SI9, respectively. CONCLUSION. Radiomic features based on unenhanced CT images can be used to predict the malignancy of pulmonary nodules. The radiomic T1a model showed superior prediction performance to the T1b and T1c models, and the best performance in terms of AUC and sensitivity was found for predicting the malignancy of T1a PN.

Radiomics and Texture Analysis in Laryngeal Cancer. Looking for New Frontiers in Precision Medicine through Imaging Analysis

Radiomics and texture analysis represent a new option in our biomarkers arsenal. These techniques extract a large number of quantitative features, analyzing their properties to incorporate them in clinical decision-making. Laryngeal cancer represents one of the most frequent cancers in the head and neck area. We hypothesized that radiomics features can be included as a laryngeal cancer precision medicine tool, as it is able to non-invasively characterize the overall tumor accounting for heterogeneity, being a prognostic and/or predictive biomarker derived from routine, standard of care, imaging data, and providing support during the follow up of the patient, in some cases avoiding the need for biopsies. The larynx represents a unique diagnostic and therapeutic challenge for clinicians due to its complex tridimensional anatomical structure. Its complex regional and functional anatomy makes it necessary to enhance our diagnostic tools in order to improve decision-making protocols, aimed at better survival and functional results. For this reason, this technique can be an option for monitoring the evolution of the disease, especially in surgical and non-surgical organ preservation treatments. This concise review article will explain basic concepts about radiomics and discuss recent progress and results related to laryngeal cancer.

Systematic analysis of bias and variability of texture measurements in computed tomography

Texture is a key radiomics measurement for quantification of disease and disease progression. The sensitivity of the measurements to image acquisition, however, is uncertain. We assessed bias and variability of computed tomography (CT) texture feature measurements across many clinical image acquisition settings and reconstruction algorithms. Diverse, anatomically informed textures (texture A, B, and C) were simulated across 1188 clinically relevant CT imaging conditions representing four in-plane pixel sizes (0.4, 0.5, 0.7, and 0.9 mm), three slice thicknesses (0.625, 1.25, and 2.5 mm), three dose levels ( CTDI vol 1.90, 3.75, and 7.50 mGy), and 33 reconstruction kernels. Imaging conditions corresponded to noise and resolution properties representative of five commercial scanners (GE LightSpeed VCT, GE Discovery 750 HD, GE Revolution, Siemens Definition Flash, and Siemens Force) in filtered backprojection and iterative reconstruction. About 21 texture features were calculated and compared between the ground-truth phantom (i.e., preimaging) and its corresponding images. Each feature was measured with four unique volumes of interest (VOIs) sizes (244, 579, 1000, and 1953    mm 3 . To characterize the bias, the percentage relative difference [PRD(%)] in each feature was calculated between the imaged scenario and the ground truth for all VOI sizes. Feature variability was assessed in terms of (1)  σ PRD ( % ) indicating the variability between the ground truth and simulated image scenario based on the PRD(%), (2)  COV f indicating the simulation-based variability, and (3)  COV T indicating the natural variability present in the ground-truth phantom. The PRD ranged widely from - 97 % to 1220%, with an underlying variability ( σ ) of up to 241%. Features such as gray-level nonuniformity, texture entropy, sum average, and homogeneity exhibited low susceptibility to reconstruction kernel effects ( PRD < 3 % ) with relatively small σ PRD ( % ) ( ≤ 5 % ) across imaging conditions. The dynamic range of results indicates that image acquisition and reconstruction conditions of in-plane pixel sizes, slice thicknesses, dose levels, and reconstruction kernels can lead to significant bias and variability in feature measurements.

Reproducibility and Generalizability in Radiomics Modeling: Possible Strategies in Radiologic and Statistical Perspectives

Radiomics, which involves the use of high-dimensional quantitative imaging features for predictive purposes, is a powerful tool for developing and testing medical hypotheses. Radiologic and statistical challenges in radiomics include those related to the reproducibility of imaging data, control of overfitting due to high dimensionality, and the generalizability of modeling. The aims of this review article are to clarify the distinctions between radiomics features and other omics and imaging data, to describe the challenges and potential strategies in reproducibility and feature selection, and to reveal the epidemiological background of modeling, thereby facilitating and promoting more reproducible and generalizable radiomics research.

Repeatability of Multiparametric Prostate MRI Radiomics Features

In this study we assessed the repeatability of radiomics features on small prostate tumors using test-retest Multiparametric Magnetic Resonance Imaging (mpMRI). The premise of radiomics is that quantitative image-based features can serve as biomarkers for detecting and characterizing disease. For such biomarkers to be useful, repeatability is a basic requirement, meaning its value must remain stable between two scans, if the conditions remain stable. We investigated repeatability of radiomics features under various preprocessing and extraction configurations including various image normalization schemes, different image pre-filtering, and different bin widths for image discretization. Although we found many radiomics features and preprocessing combinations with high repeatability (Intraclass Correlation Coefficient > 0.85), our results indicate that overall the repeatability is highly sensitive to the processing parameters. Neither image normalization, using a variety of approaches, nor the use of pre-filtering options resulted in consistent improvements in repeatability. We urge caution when interpreting radiomics features and advise paying close attention to the processing configuration details of reported results. Furthermore, we advocate reporting all processing details in radiomics studies and strongly recommend the use of open source implementations.

Radiomics robustness assessment and classification evaluation: A two-stage method demonstrated on multivendor FFDM

PURPOSE

Radiomic texture analysis is typically performed on images acquired under specific, homogeneous imaging conditions. These controlled conditions may not be representative of the range of imaging conditions implemented clinically. We aim to develop a two-stage method of radiomic texture analysis that incorporates the reproducibility of individual texture features across imaging conditions to guide the development of texture signatures which are robust across mammography unit vendors.

METHODS

Full-field digital mammograms were retrospectively collected for women who underwent screening mammography on both a Hologic Lorad Selenia and GE Senographe 2000D system. Radiomic features were calculated on manually placed regions of interest in each image. In stage one (robustness assessment), we identified a set of nonredundant features that were reproducible across the two different vendors. This was achieved through hierarchical clustering and application of robustness metrics. In stage two (classification evaluation), we performed stepwise feature selection and leave-one-out quadratic discriminant analysis (QDA) to construct radiomic signatures. We refer to this two-state method as robustness assessment, classification evaluation (RACE). These radiomic signatures were used to classify the risk of breast cancer through receiver operator characteristic (ROC) analysis, using the area under the ROC curve as a figure of merit in the task of distinguishing between women with and without high-risk factors present. Generalizability was investigated by comparing the classification performance of a feature set on the images from which they were selected (intravendor) to the classification performance on images from the vendor on which it was not selected (intervendor). Intervendor and intravendor performances were also compared to the performance obtained by implementing ComBat, a feature-level harmonization method and to the performance by implementing ComBat followed by RACE.

RESULTS

Generalizability, defined as the difference between intervendor and intravendor classification performance, was shown to monotonically decrease as the number of clusters used in stage one increased (Mann-Kendall P < 0.001). Intravendor performance was not shown to be statistically different from ComBat harmonization while intervendor performance was significantly higher than ComBat. No significant difference was observed between either of the single methods and the use of ComBat followed by RACE.

CONCLUSIONS

A two-stage method for robust radiomic signature construction is proposed and demonstrated in the task of breast cancer risk assessment. The proposed method was used to assess generalizability of radiomic texture signatures at varying levels of feature robustness criteria. The results suggest that generalizability of feature sets monotonically decreases as reproducibility of features decreases. This trend suggests that considerations of feature robustness in feature selection methodology could improve classifier generalizability in multifarious full-field digital mammography datasets collected on various vendor units. Additionally, harmonization methods such as ComBat may hold utility in classification schemes and should continue to be investigated.

Reproducibility and non-redundancy of radiomic features extracted from arterial phase CT scans in hepatocellular carcinoma patients: impact of tumor segmentation variability

Background

The reproducibility and non-redundancy of radiomic features are challenges in accelerating the clinical translation of radiomics. In this study, we focused on the robustness and non-redundancy of radiomic features extracted from computed tomography (CT) scans in hepatocellular carcinoma (HCC) patients with respect to different tumor segmentation methods.

Methods

Arterial enhanced CT images were retrospectively randomly obtained from 106 patients. As a training data set, 26 HCC patients were used to calculate the features' reproducibility and redundancy. Another data set (55 HCC patients and 25 healthy volunteers) was used for classification. The GrowCut and GraphCut semiautomatic segmentation methods were implemented in 3D Slicer software by two independent observers, and manual delineation was performed by five abdominal radiation oncologists to acquire the gross tumor volume (GTV). Seventy-one radiomic features were extracted from GTVs using Imaging Biomarker Explorer (IBEX) software, including 17 tumor intensity statistical features, 16 shape features and 38 textural features. For each radiomic feature, intraclass correlation coefficient (ICC) and hierarchical clustering were used to quantify its reproducibility and redundancy. Features with ICC values greater than 0.75 were considered reproducible. To generate the number of non-redundancy feature subgroups, the R² statistic method was used. Then, a classification model was built using a support vector machine (SVM) algorithm with 10-fold cross validation, and area under ROC curve (AUC) was used to evaluate the utility of non-redundant feature extraction by hierarchical clustering.

Results

The percentages of excellent reproducible features in the manual delineation group, GraphCut and GrowCut segmentation group were 69% [49], 73% [52] and 79% [56], respectively. Sixty-five percent [46] of the features showed strong robustness for all segmentation methods. The optimal number of cluster subgroup were 9, 13 and 11 for manual delineation, GraphCut and GrowCut segmentation, respectively. The optimal cluster subgroup number was 6 for all groups when the collectively high reproducibility features were selected for clustering. The receiver operating characteristic (ROC) analysis of radiomics classification model with and without feature reduction for healthy liver and HCC had an AUC value of 0.857 and 0.721 respectively.

Conclusions

Our study demonstrates that variations exist in the reproducibility of quantitative imaging features extracted from tumor regions segmented using different methods. The reproducibility and non-redundancy of the radiomic features rely greatly on the tumor segmentation in HCC CT images. We recommend that the most reliable and uniform radiomic features should be selected in the clinical use of radiomics. Classification experiments with feature reduction showed that radiomic features were effective in identifying healthy liver and HCC.

The Feasibility Study of Megavoltage Computed Tomographic (MVCT) Image for Texture Feature Analysis

Purpose: To determine whether radiomics texture features can be reproducibly obtained from megavoltage computed tomographic (MVCT) images acquired by Helical TomoTherapy (HT) with different imaging conditions.

Methods: For each of the 195 textures enrolled, the mean intrapatient difference, which is considered to be the benchmark for reproducibility, was calculated from the MVCT images of 22 patients with early-stage non-small-cell lung cancer. Test–retest MVCT images of an in-house designed phantom were acquired to determine the concordance correlation coefficient (CCC) for these 195 texture features. Features with high reproducibility (CCC > 0.9) in the phantom test–retest set were investigated for sensitivities to different imaging protocols, scatter levels, and motion frequencies using a wood phantom and in-vitro animal tissues.

Results: Of the 195 features, 165 (85%) features had CCC > 0.9. For the wood phantom, 124 features were reproducible in two kinds of scatter materials, and further investigations were performed on these features. For animal tissues, 108 features passed the criteria for reproducibility when one layer of scatter was covered, while 106 and 108 features of in-vitro liver and bone passed with two layers of scatter, respectively. Considering the effect of differing acquisition pitch (AcP), 97 features extracted from wood passed, while 103 and 59 features extracted from in-vitro liver and bone passed, respectively. Different reconstruction intervals (RI) had a small effect on the stability of the feature value. When AcP and RI were held consistent without motion, all 124 features calculated from wood passed, and a majority (122 of 124) of the features passed when imaging with a “fine” AcP with different RIs. However, only 55 and 40 features passed with motion frequencies of 20 and 25 beats per minute, respectively.

Conclusion: Motion frequency has a significant impact on MVCT texture features, and features from MVCT were more reproducibility in different scatter conditions than those from CBCT. Considering the effects of AcP and RI, the scanning protocols should be kept consistent when MVCT images are used for feature analysis. Some radiomics features from HT MVCT images are reproducible and could be used for creating clinical prediction models in the future.

Variation in algorithm implementation across radiomics software

Given the increased need for consistent quantitative image analysis, variations in radiomics feature calculations due to differences in radiomics software were investigated. Two in-house radiomics packages and two freely available radiomics packages, MaZda and IBEX, were utilized. Forty 256 × 256 - pixel regions of interest (ROIs) from 40 digital mammograms were studied along with 39 manually delineated ROIs from the head and neck (HN) computed tomography (CT) scans of 39 patients. Each package was used to calculate first-order histogram and second-order gray-level co-occurrence matrix (GLCM) features. Friedman tests determined differences in feature values across packages, whereas intraclass-correlation coefficients (ICC) quantified agreement. All first-order features computed from both mammography and HN cases (except skewness in mammography) showed significant differences across all packages due to systematic biases introduced by each package; however, based on ICC values, all but one first-order feature calculated on mammography ROIs and all but two first-order features calculated on HN CT ROIs showed excellent agreement, indicating the observed differences were small relative to the feature values but the bias was systematic. All second-order features computed from the two databases both differed significantly and showed poor agreement among packages, due largely to discrepancies in package-specific default GLCM parameters. Additional differences in radiomics features were traced to variations in image preprocessing, algorithm implementation, and naming conventions. Large variations in features among software packages indicate that increased efforts to standardize radiomics processes must be conducted.

Short-term reproducibility of radiomic features in liver parenchyma and liver malignancies on contrast-enhanced CT imaging

PURPOSE

To evaluate the short-term reproducibility of radiomic features in liver parenchyma and liver cancers in patients who underwent consecutive contrast-enhanced CT (CECT) with intravenous iodinated contrast within 2 weeks by chance.

METHODS

The Institutional Review Board approved this HIPAA-compliant retrospective study and waived the requirement for patients' informed consent. Patients were included if they had a liver malignancy (liver metastasis, n = 22, intrahepatic cholangiocarcinoma, n = 10, and hepatocellular carcinoma, n = 6), had two consecutive CECT within 14 days, and had no prior or intervening therapy. Liver tumors and liver parenchyma were segmented and radiomic features (n = 254) were extracted. The number of reproducible features (with concordance correlation coefficients > 0.9) was calculated for patient subgroups with different variations in contrast injection rate and pixel resolution.

RESULTS

The number of reproducible radiomic features decreased with increasing variations in contrast injection rate and pixel resolution. When including all CECTs with injection rates differences of less than 15% vs. up to 50%, 63/254 vs. 0/254 features were reproducible for liver parenchyma and 68/254 vs. 50/254 features were reproducible for malignancies. When including all CT with pixel resolution differences of 0-5% or 0-15%, 20/254 vs. 0/254 features were reproducible for liver parenchyma; 34/254 liver malignancy features were reproducible with pixel differences up to 15%.

CONCLUSION

A greater number of liver malignancy radiomic features were reproducible compared to liver parenchyma features, but the proportion of reproducible features decreased with increasing variations in contrast injection rates and pixel resolution.

Lung tumor segmentation methods: Impact on the uncertainty of radiomics features for non-small cell lung cancer

Purpose

To evaluate the uncertainty of radiomics features from contrast-enhanced breath-hold helical CT scans of non-small cell lung cancer for both manual and semi-automatic segmentation due to intra-observer, inter-observer, and inter-software reliability.

Methods

Three radiation oncologists manually delineated lung tumors twice from 10 CT scans using two software tools (3D-Slicer and MIM Maestro). Additionally, three observers without formal clinical training were instructed to use two semi-automatic segmentation tools, Lesion Sizing Toolkit (LSTK) and GrowCut, to delineate the same tumor volumes. The accuracy of the semi-automatic contours was assessed by comparison with physician manual contours using Dice similarity coefficients and Hausdorff distances. Eighty-three radiomics features were calculated for each delineated tumor contour. Informative features were identified based on their dynamic range and correlation to other features. Feature reliability was then evaluated using intra-class correlation coefficients (ICC). Feature range was used to evaluate the uncertainty of the segmentation methods.

Results

From the initial set of 83 features, 40 radiomics features were found to be informative, and these 40 features were used in the subsequent analyses. For both intra-observer and inter-observer reliability, LSTK had higher reliability than GrowCut and the two manual segmentation tools. All observers achieved consistently high ICC values when using LSTK, but the ICC value varied greatly for each observer when using GrowCut and the manual segmentation tools. For inter-software reliability, features were not reproducible across the software tools for either manual or semi-automatic segmentation methods. Additionally, no feature category was found to be more reproducible than another feature category. Feature ranges of LSTK contours were smaller than those of manual contours for all features.

Conclusion

Radiomics features extracted from LSTK contours were highly reliable across and among observers. With semi-automatic segmentation tools, observers without formal clinical training were comparable to physicians in evaluating tumor segmentation.

The utility of quantitative CT radiomics features for improved prediction of radiation pneumonitis

PURPOSE

The purpose of this study was to explore gains in predictive model performance for radiation pneumonitis (RP) using pretreatment CT radiomics features extracted from the normal lung volume.

METHODS

A total of 192 patients treated for nonsmall cell lung cancer with definitive radiotherapy were considered in the current study. In addition to clinical and dosimetric data, CT radiomics features were extracted from the total lung volume defined using the treatment planning scan. A total of 6851 features (15 clinical, 298 total lung and heart dosimetric, and 6538 image features) were gathered and considered candidate predictors for modeling of RP grade ≥3. Models were built with the least absolute shrinkage and selection operator (LASSO) logistic regression and applied to the set of candidate predictors with 50 iterations of tenfold nested cross-validation.

RESULTS

In the current cohort, 30 of 192 patients (15.6%) presented with RP grade ≥3. Average cross-validated AUC (CV-AUC) using only the clinical and dosimetric parameters was 0.51. CV-AUC was 0.68 when total lung CT radiomics features were added. Analysis with the entire set of available predictors revealed seven different image features selected in at least 40% of the model fits.

CONCLUSIONS

We have successfully incorporated CT radiomics features into a framework for building predictive RP models via LASSO logistic regression. Addition of normal lung image features produced superior model performance relative to traditional dosimetric and clinical predictors of RP, suggesting that pretreatment CT radiomics features should be considered in the context of RP prediction.

Technical Note: Identification of CT Texture Features Robust to Tumor Size Variations for Normal Lung Texture Analysis

Normal lung CT texture features have been used for the prediction of radiation-induced lung disease (RILD). For these features to be clinically useful, they should be robust to tumor size variations and not correlated with the normal lung volume of interest, i.e., the volume of the peri-tumoral region (PTR). CT images of 14 lung cancer patients were studied. Different sizes of gross tumor volumes (GTVs) were simulated and placed in the lung contralateral to the tumor. 27 texture features [nine from intensity histogram, eight from the gray-level co-occurrence matrix (GLCM) and ten from the gray-level run-length matrix (GLRM)] were extracted from the PTR. The Bland-Altman analysis was applied to measure the normalized range of agreement (nRoA) for each feature when GTV size varied. A feature was considered as robust when its nRoA was less than the threshold (100%). Sixteen texture features were identified as robust. None of the robust features was correlated with the volume of the PTR. No feature showed statistically significant differences (P<0.05) on GTV locations. We identified 16 robust normal lung CT texture features that can be further examined for the prediction of RILD.

Stability assessment of first order statistics features computed on ADC maps in soft-tissue sarcoma

Radiomics extracts a large number of features from medical images to perform a quantitative characterization. Aim of this study was to assess radiomic features stability and relevance. Apparent diffusion coefficient (ADC) maps were computed from diffusion-weighted magnetic resonance images (DW-MRI) of 18 patients diagnosed with soft-tissue sarcomas (STSs). Thirty-seven intensity-based features were computed on the regions of interest (ROIs). First, ROIs of the images were subjected to translations and rotations in specific ranges. The 37 features computed on the original and transformed ROIs were compared in terms of percentage of variations. The intra-class correlation coefficient (ICC) was computed. To be accepted, a feature should satisfy the following conditions: the ICC after a minimum entity transformation is > 0.6 and the ICC after a maximum entity translation is <; 0.4. In total, 31 features out of 37 were accepted by the algorithm. This stability analysis can be used as a first step in the features selection process.

Radiomics of CT Features May Be Nonreproducible and Redundant: Influence of CT Acquisition Parameters

Purpose To identify the reproducible and nonredundant radiomics features (RFs) for computed tomography (CT). Materials and Methods Two phantoms were used to test RF reproducibility by using test-retest analysis, by changing the CT acquisition parameters (hereafter, intra-CT analysis), and by comparing five different scanners with the same CT parameters (hereafter, inter-CT analysis). Reproducible RFs were selected by using the concordance correlation coefficient (as a measure of the agreement between variables) and the coefficient of variation (defined as the ratio of the standard deviation to the mean). Redundant features were grouped by using hierarchical cluster analysis. Results A total of 177 RFs including intensity, shape, and texture features were evaluated. The test-retest analysis showed that 91% (161 of 177) of the RFs were reproducible according to concordance correlation coefficient. Reproducibility of intra-CT RFs, based on coefficient of variation, ranged from 89.3% (151 of 177) to 43.1% (76 of 177) where the pitch factor and the reconstruction kernel were modified, respectively. Reproducibility of inter-CT RFs, based on coefficient of variation, also showed large material differences, from 85.3% (151 of 177; wood) to only 15.8% (28 of 177; polyurethane). Ten clusters were identified after the hierarchical cluster analysis and one RF per cluster was chosen as representative. Conclusion Many RFs were redundant and nonreproducible. If all the CT parameters are fixed except field of view, tube voltage, and milliamperage, then the information provided by the analyzed RFs can be summarized in only 10 RFs (each representing a cluster) because of redundancy.

Can we trust the calculation of texture indices of CT images? A phantom study

PURPOSE

Texture analysis is an emerging tool in the field of medical imaging analysis. However, many issues have been raised in terms of its use in assessing patient images and it is crucial to harmonize and standardize this new imaging measurement tool. This study was designed to evaluate the reliability of texture indices of CT images on a phantom including a reproducibility study, to assess the discriminatory capacity of indices potentially relevant in CT medical images and to determine their redundancy.

METHODS

For the reproducibility and discriminatory analysis, eight identical CT acquisitions were performed on a phantom including one homogeneous insert and two close heterogeneous inserts. Texture indices were selected for their high reproducibility and capability of discriminating different textures. For the redundancy analysis, 39 acquisitions of the same phantom were performed using varying acquisition parameters and a correlation matrix was used to explore the 2 × 2 relationships. LIFEx software was used to explore 34 different parameters including first order and texture indices.

RESULTS

Only eight indices of 34 exhibited high reproducibility and discriminated textures from each other. Skewness and kurtosis from histogram were independent from the six other indices but were intercorrelated, the other six indices correlated in diverse degrees (entropy, dissimilarity, and contrast of the co-occurrence matrix, contrast of the Neighborhood Gray Level difference matrix, SZE, ZLNU of the Gray-Level Size Zone Matrix).

CONCLUSIONS

Care should be taken when using texture analysis as a tool to characterize CT images because changes in quantitation may be primarily due to internal variability rather than from real physio-pathological effects. Some textural indices appear to be sufficiently reliable and capable to discriminate close textures on CT images.

Influence of CT acquisition and reconstruction parameters on radiomic feature reproducibility

High-dimensional imaging features extracted from diagnostic imaging, called radiomics, are increasingly reported for diagnosis, prognosis, and response to therapy. Establishing the sensitivity of radiomic features to variation in scan protocols is necessary because acquisition and reconstruction parameters can vary widely across and within institutions. Our objective was to assess the reproducibility of radiomic features derived from computed tomography (CT) images by varying tube current (mA), noise index, and reconstruction [adaptive statistical iterative reconstruction (ASiR)], parameters increasingly varied by institutions seeking to reduce radiation dose in their patients. We extracted radiomic features from CT images of a uniform water phantom, anthropomorphic phantom, and a human scan. Scans were acquired from the phantoms with six tube currents (50, 100, 200, 300, 400, and 500 mA) and five noise index levels (12, 14, 16, 18, and 20), respectively. Scans of the phantoms and patient were reconstructed from 0% ASiR (i.e., filtered back projection) to 100% ASiR in increments of 10%. Two hundred and forty-eight well-known radiomic features were extracted from all scans. The concordance correlation coefficient was used to assess agreement of features. Our analysis suggests that image acquisition parameters (tube current, noise index) as well as the reconstruction technique strongly influence radiomic feature reproducibility and demonstrate a subset of reproducible features potentially usable in clinical practice.

Reproducibility with repeat CT in radiomics study for rectal cancer

Purpose

To evaluate the reproducibility of radiomics features by repeating computed tomographic (CT) scans in rectal cancer. To choose stable radiomics features for rectal cancer.

Results

Volume normalized features are much more reproducible than unnormalized features. The average value of all slices is the most reproducible feature type in rectal cancer. Different filters have little effect for the reproducibility of radiomics features. For the average type features, 496 out of 775 features showed high reproducibility (ICC ≥ 0.8), 225 out of 775 features showed medium reproducibility (0.8 > ICC ≥ 0.5) and 54 out of 775 features showed low reproducibility (ICC < 0.5).

Methods

40 rectal cancer patients with stage II were enrolled in this study, each of whom underwent two CT scans within average 8.7 days. 775 radiomics features were defined in this study. For each features, five different values (value from the largest slice, maximum value, minimum value, average value of all slices and value from superposed intermediate matrix) were extracted. Meanwhile a LOG filter with different parameters was applied to these images to find stable filter value. Concordance correlation coefficients (CCC) and inter-class correlation coefficients (ICC) of two CT scans were calculated to assess the reproducibility, based on original features and volume normalized features.

Conclusions

Features are recommended to be normalized to volume in radiomics analysis. The average type radiomics features are the most stable features in rectal cancer. Further analysis of these features of rectal cancer can be warranted for treatment monitoring and prognosis prediction.

Guidelines and Experience Using Imaging Biomarker Explorer (IBEX) for Radiomics

Imaging Biomarker Explorer (IBEX) is an open-source tool for medical imaging radiomics work. The purpose of this paper is to describe how to use IBEX's graphical user interface (GUI) and to demonstrate how IBEX calculated features have been used in clinical studies. IBEX allows for the import of DICOM images with DICOM radiation therapy structure files or Pinnacle files. Once the images are imported, IBEX has tools within the Data Selection GUI to manipulate the viewing of the images, measure voxel values and distances, and create and edit contours. IBEX comes with 27 preprocessing and 132 feature choices to design feature sets. Each preprocessing and feature category has parameters that can be altered. The output from IBEX is a spreadsheet that contains: 1) each feature from the feature set calculated for each contour in a data set, 2) image information about each contour in a data set, and 3) a summary of the preprocessing and features used with their selected parameters. Features calculated from IBEX have been used in studies to test the variability of features under different imaging conditions and in survival models to improve current clinical models.

Towards precision medicine: from quantitative imaging to radiomics

Radiology (imaging) and imaging-guided interventions, which provide multi-parametric morphologic and functional information, are playing an increasingly significant role in precision medicine. Radiologists are trained to understand the imaging phenotypes, transcribe those observations (phenotypes) to correlate with underlying diseases and to characterize the images. However, in order to understand and characterize the molecular phenotype (to obtain genomic information) of solid heterogeneous tumours, the advanced sequencing of those tissues using biopsy is required. Thus, radiologists image the tissues from various views and angles in order to have the complete image phenotypes, thereby acquiring a huge amount of data. Deriving meaningful details from all these radiological data becomes challenging and raises the big data issues. Therefore, interest in the application of radiomics has been growing in recent years as it has the potential to provide significant interpretive and predictive information for decision support. Radiomics is a combination of conventional computer-aided diagnosis, deep learning methods, and human skills, and thus can be used for quantitative characterization of tumour phenotypes. This paper discusses the overview of radiomics workflow, the results of various radiomics-based studies conducted using various radiological images such as computed tomography (CT), magnetic resonance imaging (MRI), and positron-emission tomography (PET), the challenges we are facing, and the potential contribution of radiomics towards precision medicine.

Investigating the Robustness Neighborhood Gray Tone Difference Matrix and Gray Level Co-occurrence Matrix Radiomic Features on Clinical Computed Tomography Systems Using Anthropomorphic Phantoms: Evidence From a Multivendor Study

OBJECTIVE

The aim of this study was to determine if optimized imaging protocols across multiple computed tomography (CT) vendors could result in reproducible radiomic features calculated from an anthropomorphic phantom.

METHODS

Materials with varying degrees of heterogeneity were placed throughout the lungs of the phantom. Twenty scans of the phantom were acquired on 3 CT manufacturers with chest CT protocols that had optimized protocol parameters. Scans were reconstructed using vendor-specific standards and lung kernels. The concordance correlation coefficient (CCC) was used to calculate reproducibility between features. For features with high CCC values, Bland-Altman analysis was also used to quantify agreement.

RESULTS

The mean Hounsfield unit (HU) was 32.93 HU (141.7 to -26.5 HU) for the rubber insert and 347.2 HU (-320.9 to -347.7 HU) for the wood insert. Low CCC values of less than 0.9 were calculated for all features across all scans.

CONCLUSIONS

Radiomic features that are derived from the spatial distribution of voxel intensities should be particularly scrutinized for reproducibility in a multivendor environment.