Neighborhood Gray-tone Difference Matrix Research Articles

CT radiomics-based biomarkers can predict response to immunotherapy in hepatocellular carcinoma.

Hepatocellular Carcinoma (HCC) remains a leading cause of cancer deaths. Despite the rise of immunotherapies, many HCC patients don't benefit. There's a clear need for biomarkers to guide treatment decisions. This research aims to identify such biomarkers by combining radiological data and machine learning. We analyzed clinical and CT imaging data of 54 HCC patients undergoing immunotherapy. Radiologic features were examined to develop a model predicting short-term immunotherapy effects. We utilized 9 machine learning and 2 ensemble learning techniques using RapidMiner for model construction. We conducted the validation of the above feature combination using 29 HCC patients who received immunotherapy from another hospital, and tested and validated it using XGBoost combined with a genetic algorithm. In 54 HCC patients, radiomics features varied significantly between those with partial response (PR) and stable disease (SD). Key features in Gray Level Run Length Matrix (GLRLM) and in adjacent tissues' Intensity Direct, Neighborhood Gray Tone Difference Matrix (NGTDM), and Shape correlated with short-term immunotherapy efficacy. Selected feature combinations of 15, 19, and 8/15 features yielded better outcomes. Deep learning, random forest, and naive bayes outperformed other methods, with Bagging being more reliable than Adaboost. In the validation set of 29 HCC patients, the mentioned feature combination also demonstrated favorable performance. Furthermore, we achieved improved results when employing XGBoost in conjunction with a genetic algorithm for testing and validation. The machine learning model built with CT image features derived from GLCM, GLRLM, IntensityDirect, NGTDM, and Shape can accurately forecast the short-term efficacy of immunotherapy for HCC.

Radiomic survival prediction with liver metastases in colorectal cancer.

e15584 Background: The survival outcome of colorectal cancer depends upon staging and the presence of metastases. Nonetheless, there isn’t a reliable prediction model for overall survival (OS) specifically tailored to patients with colorectal cancer who have liver metastases. Consequently, this study aims to find radiomic features that can be utilized to predict OS of colorectal cancer with liver metastases. Methods: We utilized the Colorectal Liver Metastases data set in The Cancer Imaging Archive (TCIA). Overall, 197 patients with colorectal cancer metastatic to the liver were identified. Semi-automatic segmentation of liver metastases was used to extract radiomic features including Gray-Level Co-occurrence (glcm), Gray-Level Run Length Matrix (glrlm), Gray-Level Size Zone Matrix (glszm), Gray-Level Dependence Matrix (gldm), and Neighborhood Gray-Tone Difference Matrix (ngtdm). 3D Slicer was used to extract radiomic features. We tested a Cox proportional hazards model in each radiomics feature to predict OS. Results: Total 107 radiomic features were extracted from pre-segmented liver metastases on patients’ Computed Tomography images. Among them, original shape sphericity, original glcm correlation, original gldm DependenceEntropy, and original glrlm RunEntropy were statistically significant for the prediction of OS, with a hazard ratio of -2.16 (CI 0.01-0.92; p value 0.04), 1.54 (CI 1.5-14.0; p value 0.006), 0.51 (CI 1.0-2.5; p value 0.02), and 0.52 (CI 1.0-2.6; p value 0.03), respectively. Conclusions: Our data analysis indicates crucial radiomic features that hold promise for predicting the clinical outcome of patients with colorectal cancer who have liver metastases. Future research with a larger sample size is warranted for the external validation of our results.

Robustness of radiomic features in healthy abdominal parenchyma of patients with repeated examinations on dual-layer dual-energy CT

ObjectivesRobustness of radiomic features in physiological tissue is an important prerequisite for quantitative analysis of tumor biology and response assessment. In contrast to previous studies which focused on different tumors with mostly short scan-re-scan intervals, this study aimed to evaluate the robustness of radiomic features in cancer-free patients and over a clinically encountered inter-scan interval. Materials and methodsPatients without visible tumor burden who underwent at least two portal-venous phase dual energy CT examinations of the abdomen between May 2016 and January 2020 were included, while macroscopic tumor burden was excluded based upon follow-up imaging for all patients (≥3 months). Further, patients were excluded if no follow-up imaging was available, or if the CT protocol showed deviations between repeated examinations. Circular regions of interest were placed and proofread by two board-certified radiologists (4 years and 5 years experience) within the liver (segments 3 and 6), the psoas muscle (left and right), the pancreatic head, and the spleen to obtain radiomic features from normal-appearing organ parenchyma using PyRadiomics. Radiomic feature robustness was tested using the concordance correlation coefficient with a threshold of 0.75 considered indicative for deeming a feature robust. ResultsIn total, 160 patients with 480 repeated abdominal CT examinations (range: 2–4 per patient) were retrospectively included in this single-center, IRB-approved study. Considering all organs and feature categories, only 4.58 % (25/546) of all features were robust with the highest rate being found in the first order feature category (20.37 %, 22/108). Other feature categories (grey level co-occurrence matrix, grey level dependence matrix, grey level run length matrix, grey level size zone matrix, and neighborhood gray-tone difference matrix) yielded an overall low percentage of robust features (range: 0.00 %-1.19 %). A subgroup analysis revealed the reconstructed field of view and the X-ray tube current as determinants of feature robustness (significant differences in subgroups for all organs, p < 0.001) as well as the size of the region of interest (no significant difference for the pancreatic head with p = 0.135, significant difference with p < 0.001 for all other organs). ConclusionRadiomic feature robustness obtained from cancer-free subjects with repeated examinations using a consistent protocol and CT scanner was limited, with first order features yielding the highest proportion of robust features.

Machine learning approach using 18F-FDG-PET-radiomic features and the visibility of right ventricle 18F-FDG uptake for predicting clinical events in patients with cardiac sarcoidosis

ObjectivesTo investigate the usefulness of machine learning (ML) models using pretreatment 18F-FDG-PET-based radiomic features for predicting adverse clinical events (ACEs) in patients with cardiac sarcoidosis (CS).Materials and methodsThis retrospective study included 47 patients with CS who underwent 18F-FDG-PET/CT scan before treatment. The lesions were assigned to the training (n = 38) and testing (n = 9) cohorts. In total, 49 18F-FDG-PET-based radiomic features and the visibility of right ventricle 18F-FDG uptake were used to predict ACEs using seven different ML algorithms (namely, decision tree, random forest [RF], neural network, k-nearest neighbors, Naïve Bayes, logistic regression, and support vector machine [SVM]) with tenfold cross-validation and the synthetic minority over-sampling technique. The ML models were constructed using the top four features ranked by the decrease in Gini impurity. The AUCs and accuracies were used to compare predictive performances.ResultsPatients who developed ACEs presented with a significantly higher surface area and gray level run length matrix run length non-uniformity (GLRLM_RLNU), and lower neighborhood gray-tone difference matrix_coarseness and sphericity than those without ACEs (each, p < 0.05). In the training cohort, all seven ML algorithms had a good classification performance with AUC values of > 0.80 (range: 0.841–0.944). In the testing cohort, the RF algorithm had the highest AUC and accuracy (88.9% [8/9]) with a similar classification performance between training and testing cohorts (AUC: 0.945 vs 0.889). GLRLM_RLNU was the most important feature of the modeling process of this RF algorithm.ConclusionML analyses using 18F-FDG-PET-based radiomic features may be useful for predicting ACEs in patients with CS.

Synthesis of virtual monoenergetic images from kilovoltage peak images using wavelet loss enhanced CycleGAN for improving radiomics features reproducibility.

Dual-energy computed tomography (CT) can provide a range of image information beyond conventional CT through virtual monoenergetic images (VMIs). The purpose of this study was to investigate the impact of material decomposition in detector-based spectral CT on radiomics features and effectiveness of using deep learning-based image synthesis to improve the reproducibility of radiomics features. In this paper, spectral CT image data from 45 esophageal cancer patients were collected for investigation retrospectively. First, we computed the correlation coefficient of radiomics features between conventional kilovoltage peak (kVp) CT images and VMI. Then, a wavelet loss-enhanced CycleGAN (WLL-CycleGAN) with paired loss terms was developed to synthesize virtual monoenergetic CT images from the corresponding conventional single-energy CT (SECT) images for improving radiomics reproducibility. Finally, the radiomic features in 6 different categories, including gray-level co-occurrence matrix (GLCM), gray-level difference matrix (GLDM), gray-level run-length matrix (GLRLM), gray-level size-zone matrix (GLSZM), neighborhood gray-tone difference matrix (NGTDM), and wavelet, were extracted from the gross tumor volumes from conventional single energy CT, synthetic virtual monoenergetic CT images, and virtual monoenergetic CT images. Comparison between errors in the VMI and synthetic VMI (sVMI) suggested that the performance of our proposed deep learning method improved the radiomic feature accuracy. Material decomposition of dual-layer dual-energy CT (DECT) can substantially influence the reproducibility of the radiomic features, and the degree of impact is feature dependent. The average reduction of radiomics errors for 15 patients in testing sets was 96.9% for first-order, 12.1% for GLCM, 12.9% for GLDM, 15.7% for GLRLM, 50.3% for GLSZM, 53.4% for NGTDM, and 6% for wavelet features. The work revealed that material decomposition has a significant effect on the radiomic feature values. The deep learning-based method reduced the influence of material decomposition in VMIs and might improve the robustness and reproducibility of radiomic features in esophageal cancer. Quantitative results demonstrated that our proposed wavelet loss-enhanced paired CycleGAN outperforms the original CycleGAN.

Technical and clinical considerations of a physical liver phantom for CT radiomics analysis.

This study identifies key characteristics to help build a physical liver computed tomography (CT) phantom for radiomics harmonization; particularly, the higher-order texture metrics. CT scans of a radiomics phantom comprising of 18 novel 3D printed inserts with varying size, shape, and material combinations were acquired on a 64-slice CT scanner (Brilliance 64, Philips Healthcare). The images were acquired at 120kV, 250mAs, CTDIvol of 16.36mGy, 2mm slice thickness, and iterative noise-reduction reconstruction (iDose, Philips Healthcare, Andover, MA). Radiomics analysis was performed using the Cancer Imaging Phenomics Toolkit (CaPTk), following automated segmentation of 3D regions of interest (ROI) of the 18 inserts. The findings were compared to three additional ROI obtained of an anthropomorphic liver phantom, a patient liver CT scan, and a water phantom, at comparable imaging settings. Percentage difference in radiomic metrics values between phantom and tissue was used to assess the biological equivalency and <10% was used to claim equivalent. The HU for all 18 ROI from the phantom ranged from -30 to 120 which is within clinically observed HU range of the liver, showing that our phantom material (T3-6B) is representative of biological CT tissue densities (liver) with >50% radiomic features having <10% difference from liver tissue. Based on the assessment of the Neighborhood Gray Tone Difference Matrix (NGTDM) metrics it is evident that the water phantom ROI show extreme values compared to the ROIs from the phantom. This result may further reinforce the difference between a structureless quantity such as water HU values and tissue HU values found in liver. The 3-D printed patterns of the constructed radiomics phantom cover a wide span of liver tissue textures seen in CT images. Using our results, texture metrics can be selectively harmonized to establish clinically relevant and reliable radiomics panels.

A comprehensive study on automatic non‐informative frame detection in colonoscopy videos

AbstractDespite today's developing healthcare technology, conventional colonoscopy is still a gold‐standard method to detect colon abnormalities. Due to the folded structure of the intestine and visual disturbances caused by artifacts, it can be hard for specialists to detect abnormalities during the procedure. Frames that include artifacts such as specular reflection, improper contrast levels from insufficient or excessive illumination gastric juice, bubbles, or residuals should be detected to increase an accurate diagnosis rate. In this work, both conventional machine learning and transfer learning methods have been used to detect non‐informative frames in colonoscopy videos. The conventional machine learning part consists of 5 different types of texture features, which are gray level co‐occurrence matrix (GLCM), gray level run length matrix (GLRLM), neighborhood gray‐tone difference matrix (NGTDM), focus measure operators (FMOs), and first‐order statistics. In addition to these methods, we utilized 8 different transfer learning models: AlexNet, SqueezeNet, GoogleNet, ShuffleNet, ResNet50, ResNet18, NasNetMobile, and MobileNet. The results showed that FMOs and decision tree combination gave the best accuracy and f‐measure values with almost 89% and 0.79%, respectively, for the conventional machine learning part. When the transfer learning part is taken into account, AlexNet (99.85%) and SqueezeNet (98.80%) have the highest performance metric results. This study shows the potential of both transfer learning and conventional machine learning algorithms to provide fast and accurate non‐informative frame detection to be used during a colonoscopy, which may be considered the initial step in identifying and classifying colon‐related diseases automatically to help guide physicians.

The Role of Radiomics Model to Predict Early Xerostomia in Head and Neck Cancer after Radiotherapy

Based on the patients' characteristics, the most treated patients suffering from early xerostomia during radiation therapy treatment and in the early months after treatment in head and neck cancer. The aim of this study is to provide an image-based model to predict early xerostomia in these patients before radiotherapy. The dosimetric and clinical factors (age, sex, history of disease) and computed tomography (CT) features are also combined with the image data in resolving xerostomia detection. Clinical factors and CT features were considered for 62 patients with head and neck cancer, and the complication of parotid and salivary glands was evaluated. For feature extraction from CT images, 3D Slicer software was used to extract 107 features. The features type included shape, gray level dependence matrix, gray level co-occurrence matrix, first order gray level run length matrix, gray level size zone matrix, and neighborhood gray-tone difference matrix. Four different datasets; demographic + sum dose volume histogram (DVH), demographic + ipsilateral and contralateral DVH, demographic + CT features + sum DVH, and demographic + CT features + ipsilateral and contralateral DVH; were used as model inputs. Four algorithms: Random Tree (RT), Linear Support Vector Machine (LSVM), Neural Network (NN), and Bayesian Network (BN), were applied to analyze the datasets to predict xerostomia. The metrics of the models used sensitivity, specificity, accuracy, and area-under-the-curve (AUC) to assess the performance. The AUC has its highest value for the RT model among the aforementioned models and its value for the first and second data sets were 0.84 and 0.64, respectively. In general, model performance was improved when the sum of the parotids DVH data was used. Furthermore, the RT model showed the best performance when using radiomics features in the third and fourth data sets with the AUC of 0.86 and 0.89, respectively. Among the RT, LSVM, NN, and BN models, RT has the highest performance in predicting radiation-induced early xerostomia in patients who underwent head and neck radiotherapy. Model performance was improved when the sum of the parotids DVH data was used in addition to the dose data and CT features. This study showed that combining the dosimetric data with radiomics features may improve the model performance for predicting early xerostomia.

Prognostic Value of Weekly Delta-Radiomics during MR-Linac Radiotherapy of Glioblastoma

MRI after chemoradiotherapy (chemoRT) shows areas of presumed tumor growth in ≤ 50% of glioblastoma (GBM) patients, which can be true progression (TP) - tumor growth with poor treatment response, or pseudoprogression (PP) - edema and tumor necrosis with favorable treatment response. Patients with TP have median overall survival (OS) of only 7 months, while patients with PP have median OS of 36 months. However, on imaging, TP and PP are usually not discernible during treatment, making it difficult to adapt radiation for poor responders. The purpose of this study was to investigate the prognostic value of delta radiomic features from MR-Linac for GBM. Using an IRB-approved prospective cohort of GBM patients undergoing 30 fractions of chemoRT to 60 Gy on a 0.35T MR-Linac, 2 regions of interest (ROI) were contoured on daily T2-weighted treatment set-up scans: 1) tumor/edema (lesion) and 2) post-surgical resection cavity (RC). The lesion ROI were used to calculate texture features: second order radiomics features based on the gray-level co-occurrence matrix (GLCM), gray-level size zone matrix (GLSZM), gray-level run length matrix (GLRLM), and neighborhood gray-tone difference matrix (NGTDM). Each of these describe the probability of spatial relationships of gray levels occurring within the ROI. Features from fraction 1 (pre-radiation) were subtracted from fractions 5, 10, 15, 25, and 30 to create delta features at 5 timepoints (D5-D30). Patient response was retrospectively defined as no progression (NP), TP, or PP. Supervised machine learning was utilized using a 500-tree random forest (RF) classification model with TP or PP as the outcome. Variable importance analysis was conducted by calculating the out-of-bag errors with multiple bootstrapped data sets. The most prognostic features were selected using the RF importance scores. Thirty-six patients were screened for inclusion: 9 were excluded due to no T2 lesion (RC ROI only). Of the remaining 27 patients: 10 had NP, 11 had TP, and 6 had PP. Thirty-nine texture features, plus lesion volume and mean lesion intensity (for a total of 41 variables per time point) were calculated and included in the model. Of the 10 most prognostic features, 6 were from D10, suggesting that prognostic changes in the underlying lesion microenvironment are occurring within the first 10 fractions of treatment. The model selected GLSZM high gray-level zone emphasis (HGZE) D10, IBSI code 5GN9, as the most prognostic feature. The receiver operator characteristic (ROC) area under the curve (AUC) for GLSZM HGZE D10 was 0.94 (95% CI = 0.81-1.00). Delta radiomic features extracted from MR-Linac imaging may predict between PP and TP in GBM patients during treatment, which is earlier than current methods. This could allow physicians to adapt/intensify treatment in real time for poorly responding patients. Future directions include analysis with a larger patient cohort and with additional MRI contrasts (MR-Linac multiparametric MRI).

Prognostic Value of Metabolic Parameters and Textural Features in Pretreatment 18F-FDG PET/CT of Primary Lesions for Pediatric Patients with Neuroblastoma

Our study evaluated the prognostic value of the metabolic parameters and textural features in pretreatment 18F-Fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) of primary lesions for pediatric patients with neuroblastoma. In total, 107 pediatric patients with neuroblastoma who underwent pretreatment 18F-FDG PET/CT were retrospectively included and analyzed. All patients were diagnosed by pathology, and baseline characteristics and clinical data were collected. The four metabolic parameters and 43 textural features of 18F-FDG PET/CT of the primary lesions were measured. The prognostic significance of metabolic parameters and other clinical variables was assessed using Cox proportional hazards regression models. Differences in progression-free survival (PFS) and overall survival (OS) in relation to parameters were examined using the Kaplan-Meier method. During a median follow-up period of 34.3 months, 45 patients (42.1%) experienced tumor recurrence or progression, and 21 patients (19.6%) died of cancer. In univariate Cox regression analysis, age, location of disease, International Neuroblastoma Risk Group Staging System (INRGSS) stage M, neuron-specific enolase (NSE), lactate dehydrogenase (LDH), four positron emission tomography (PET) metabolic parameters, and 33 textural features were significant predictors of PFS. In multivariate analysis, INRGSS stage M (hazard ratio [HR]=19.940, 95% confidence interval [CI]=2.733-145.491, P=0.003), skewness (＞0.173; PET first-order features; HR=2.938, 95% CI=1.389-6.215, P=0.005), coarseness (＞0.003; neighborhood gray-tone difference matrix; HR=0.253, 95% CI=0.132-0.484, P＜0.001), and variance (＞103.837; CT first-order gray histogram parameters; HR=2.810, 95% CI=1.160-6.807, P=0.022) were independent predictors of PFS. In univariate Cox regression analysis, gender, INRGSS stage M, MYCN amplification, NSE, LDH, two PET metabolic parameters, and five textural features were significant predictors of OS. In multivariate analysis, INRGSS stage M (HR=7.704, 95% CI=1.031-57.576, P=0.047), MYCN amplification (HR=3.011, 95% CI=1.164-7.786, P=0.023), and metabolic tumor volume (＞138.788; HR=3.930, 95% CI=1.317-11.727, P=0.014) were independent predictors of OS. The metabolic parameters and textural features in pretreatment 18F-FDG PET/CT of primary lesions are predictive of survival in pediatric patients with neuroblastoma.

Prediction of MYCN Gene Amplification in Pediatric Neuroblastomas: Development of a Deep Learning-Based Tool for Automatic Tumor Segmentation and Comparative Analysis of Computed Tomography-Based Radiomics Features Harmonization.

MYCN oncogene amplification is closely linked to high-grade neuroblastoma with poor prognosis. Accurate quantification is essential for risk assessment, which guides clinical decision making and disease management. This study proposes an end-to-end deep-learning framework for automatic tumor segmentation of pediatric neuroblastomas and radiomics features-based classification of MYCN gene amplification. Data from pretreatment contrast-enhanced computed tomography scans and MYCN status from 47 cases of pediatric neuroblastomas treated at a tertiary children's hospital from 2009 to 2020 were reviewed. Automated tumor segmentation and grading pipeline includes (1) a modified U-Net for tumor segmentation; (2) extraction of radiomic textural features; (3) feature-based ComBat harmonization for removal of variabilities across scanners; (4) feature selection using 2 approaches, namely, (a) an ensemble approach and (b) stepwise forward-and-backward selection method using logistic regression classifier; and (5) radiomics features-based classification of MYCN gene amplification using machine learning classifiers. Median train/test Dice score for modified U-Net was 0.728/0.680. The top 3 features from the ensemble approach were neighborhood gray-tone difference matrix (NGTDM) busyness, NGTDM strength, and gray-level run-length matrix (GLRLM) low gray-level run emphasis, whereas those from the stepwise approach were GLRLM low gray-level run emphasis, GLRLM high gray-level run emphasis, and NGTDM coarseness. The top-performing tumor classification algorithm achieved a weighted F1 score of 97%, an area under the receiver operating characteristic curve of 96.9%, an accuracy of 96.97%, and a negative predictive value of 100%. Harmonization-based tumor classification improved the accuracy by 2% to 3% for all classifiers. The proposed end-to-end framework achieved high accuracy for MYCN gene amplification status classification.

Predicting Survival for Hepatic Arterial Infusion Chemotherapy of Unresectable Colorectal Liver Metastases: Radiomics Analysis of Pretreatment Computed Tomography.

ObjectiveHepatic arterial infusion chemotherapy (HAIC) is an effective treatment for advanced unresectable colorectal cancer liver metastases (CRLM). This study was conducted to predict the efficacy of HAIC in patients with unresectable CRLM by radiomics methods based on pretreatment computed tomography (CT) examinations and clinical data.Materials and MethodsA total of 63 patients were included in this study (41 in the training group and 22 in the validation group). All these patients underwent CT examination before HAIC. During the follow-up period, CT scans and laboratory examinations were performed regularly. Eighty-five radiological features were extracted from the regions of interest (ROIs) of CT images using the PyRadiomics program. The t-test and correlation were applied to select features. These features were analyzed using LASSO-Cox regression, and a linear model was developed to predict overall survival (OS).ResultsAfter reducing features by t-test and correlation test, seven features remained. After LASSO-Cox cross-validation, four features remained at λ = 0.232. They were gray level co-occurrence matrix (GLCM), gray level run length matrix (GLRLM), neighborhood gray tone difference matrix (NGTDM), and the location of the primary tumor. The C-index was 0.758 in the training group and 0.743 in the test group. Nomograms predicting 1-, 2-, and 3-year survival were established.ConclusionOur study demonstrates that a radiomics approach based on pretreatment CT texture analysis has the ability to predict early the outcome of HAIC in patients with advanced unresectable colorectal cancer with a high degree of accuracy and feasibility.

Impact of ComBat Harmonization on PET Radiomics-Based Tissue Classification: A Dual-Center PET/MRI and PET/CT Study.

Our purpose was to determine whether ComBat harmonization improves 18F-FDG PET radiomics-based tissue classification in pooled PET/MRI and PET/CT datasets. Methods: Two hundred patients who had undergone 18F-FDG PET/MRI (2 scanners and vendors; 50 patients each) or PET/CT (2 scanners and vendors; 50 patients each) were retrospectively included. Gray-level histogram, gray-level cooccurrence matrix, gray-level run-length matrix, gray-level size-zone matrix, and neighborhood gray-tone difference matrix radiomic features were calculated for volumes of interest in the disease-free liver, spleen, and bone marrow. For individual feature classes and a multiclass radiomic signature, tissue was classified on ComBat-harmonized and unharmonized pooled data, using a multilayer perceptron neural network. Results: Median accuracies in training and validation datasets were 69.5% and 68.3% (harmonized), respectively, versus 59.5% and 58.9% (unharmonized), respectively, for gray-level histogram; 92.1% and 86.1% (harmonized), respectively, versus 53.6% and 50.0% (unharmonized), respectively, for gray-level cooccurrence matrix; 84.8% and 82.8% (harmonized), respectively, versus 62.4% and 58.3% (unharmonized), respectively, for gray-level run-length matrix; 87.6% and 85.6% (harmonized), respectively, versus 56.2% and 52.8% (unharmonized), respectively, for gray-level size-zone matrix; 79.5% and 77.2% (harmonized), respectively, versus 54.8% and 53.9% (unharmonized), respectively, for neighborhood gray-tone difference matrix; and 86.9% and 84.4% (harmonized), respectively, versus 62.9% and 58.3% (unharmonized), respectively, for radiomic signature. Conclusion: ComBat harmonization may be useful for multicenter 18F-FDG PET radiomics studies using pooled PET/MRI and PET/CT data.

Robustness Assessment of Images From a 0.35T Scanner of an Integrated MRI-Linac: Characterization of Radiomics Features in Phantom and Patient Data.

Purpose: Radiomics entails the extraction of quantitative imaging biomarkers (or radiomics features) hypothesized to provide additional pathophysiological and/or clinical information compared to qualitative visual observation and interpretation. This retrospective study explores the variability of radiomics features extracted from images acquired with the 0.35 T scanner of an integrated MRI-Linac. We hypothesized we would be able to identify features with high repeatability and reproducibility over various imaging conditions using phantom and patient imaging studies. We also compared findings from the literature relevant to our results. Methods: Eleven scans of a Magphan® RT phantom over 13 months and 11 scans of a ViewRay Daily QA phantom over 11 days constituted the phantom data. Patient datasets included 50 images from ten anonymized stereotactic body radiation therapy (SBRT) pancreatic cancer patients (50 Gy in 5 fractions). A True Fast Imaging with Steady-State Free Precession (TRUFI) pulse sequence was selected, using a voxel resolution of 1.5 mm × 1.5 mm × 1.5 mm and 1.5 mm × 1.5 mm × 3.0 mm for phantom and patient data, respectively. A total of 1087 shape-based, first, second, and higher order features were extracted followed by robustness analysis. Robustness was assessed with the Coefficient of Variation (CoV < 5%). Results: We identified 130 robust features across the datasets. Robust features were found within each category, except for 2 second-order sub-groups, namely, Gray Level Size Zone Matrix (GLSZM) and Neighborhood Gray Tone Difference Matrix (NGTDM). Additionally, several robust features agreed with findings from other stability assessments or predictive performance studies in the literature. Conclusion: We verified the stability of the 0.35 T scanner of an integrated MRI-Linac for longitudinal radiomics phantom studies and identified robust features over various imaging conditions. We conclude that phantom measurements can be used to identify robust radiomics features. More stability assessment research is warranted.

NIMG-56. USING RADIOMIC FEATURES FROM DAILY MAGNETIC RESONANCE IMAGING TO PREDICT RESPONSE TO RADIATION THERAPY IN GLIOBLASTOMA PATIENTS: A PILOT STUDY

Abstract A dilemma in treating glioblastoma is that MRI after chemotherapy and radiation therapy (chemoRT) shows areas of presumed tumor growth in up to 50% of patients. These areas can represent true progression (TP), tumor growth with tumors non-responsive to treatment, or pseudoprogression (PP), edema and tumor necrosis with favorable treatment response. On imaging, TP and PP are usually not discernable. Patients in this study undergo six weeks of chemoRT on a combination MRI/RT device, receiving daily MRIs. The goal of this study is to explore the correlation of radiomics features with progression. The tumor lesion and surrounding areas of growth/edema were manually outlined as regions of interest (ROIs) for each daily T2-weighted MRI scan. The ROIs were used to calculate texture features: statistical features based on the gray-level co-occurrence matrix (GLCM), the gray-level zone size matrix (GLZSM), the gray-level run length matrix (GLRLM), and the neighborhood gray-tone difference matrix (NGTDM). Each of these matrix classes describe the probability of spatial relationships of gray levels occurring within the ROI. Daily texture features were averaged per week of treatment for each patient. Patient response was retrospectively defined as no progression (NP), TP, or PP. A Kruskal-Wallis test was performed to identify texture features that correlated most strongly with patient response. Forty texture features were calculated for 12 patients (19 treated, 7 excluded due to no T2 lesion or progression status unknown, 6 NP, 3 TP, 3 PP). There was a trend of more texture features correlating significantly with response in weeks 4-6 of treatment, compared to weeks 1-3. A particular texture feature, GLSZM Small Zone Low Gray-Level Emphasis, showed increasing difference between PP and TP over time, with significant difference during week 6 of treatment (p=0.0495). Future directions include correlating early outcomes with greater numbers of patients and daily multiparametric MRI.

MRI-Based Radiomics Signature for the Prediction of Response of Lung Cancer Brain Metastases After Whole-Brain Radio-Therapy

Local response prediction for brain metastases (BMs) from lung cancer after Whole-Brain Radiotherapy (WBRT) is challenging, as existing criteria are based solely on unidimensional measurements. This study sought to determine whether radiomic features of lung cancer BMs derived from pre-treatment magnetic resonance imaging (MRI) could be used to predict local response following WBRT.A total of 88 Lung Cancer patients with BMs treated with WBRT were analyzed. After volumes of interest were drawn, 944 radiomic features including first-order, shape, Gray Level Co-occurrence Matrix (GLCM), Gray Level Dependence Matrix (GLDM), Gray Level Run Length Matrix (GLRLM), Gray Level Size Zone Matrix (GLSZM), Neighborhood Gray Tone Difference Matrix (NGTDM), and Laplacian of Gaussian (LoG) features were extracted, using the baseline pre-treatment post-contrast T1 (T1c) and T2 fluid-attenuated inversion recovery (FLAIR) MRI sequences, respectively. Local response status was determined by contrasting the baseline and follow-up MRI according to the RANO-BM criteria. The independent samples t test or Mann-Whitney U test, and then least absolute shrinkage and selection operator (LASSO) were used for dimensionality reduction and feature selection. An adaboost classifier was trained using the selected radiomic features and evaluated using the area under the receiver operating characteristic curve (AUC) in both the training and testing sets. Other discrimination metrics, including classification accuracy, positive predictive value (PPV), negative predictive value (NPV), sensitivity, and specificity, were also calculated.The optimal radiomics signature was developed based on a multivariable logistic regression with 4, 5, 6 radiomic features on T1c, T2 FLAIR and T1c+T2 FLAIR, respectively. The radiomics model based on T1c features presented the AUC of (0.920 vs. 0.805, respectively) for both the training and testing sets, followed by T2 FLAIR features (0.893 vs. 0.701, respectively), and T1c+T2 FLAIR features (0.971 vs. 0.857, respectively). The classification accuracy of the radiomics model also well predicted the local response of BMs for both the models), and training and testing sets (T1c: 82.9% vs. 77.8%, T2 FLAIR: 82.9% vs. 77.8%, T1c+T2 FLAIR: 90.0% vs. 77.8%, respectively).Radiomics holds promise for predicting local tumor response following WBRT in patients with lung cancer and brain metastases. A predictive model built on radiomic features from an institutional cohort performed well on cross-validation testing. These results warrant further validation in independent datasets. Such work could prove invaluable for guiding management of individual patients and assessing outcomes of novel interventions.

Primary clinical study of radiomics for diagnosing simple bone cyst of the jaw.

To screen the radiomic features of simple bone cysts of the jaws and explore the potential application of radiomics in pre-operative diagnosis of jaw simple bone cysts. The investigators designed and implemented a case-control study. 19 patients with simple bone cysts who were admitted to the Department of Maxillofacial Surgery, Sun Yat-sen University Affiliated Stomatology Hospital from 2013 to 2019 were included in this study. Their clinical data and cone-beam computed tomography (CBCT) images were examined. The control group consisted of patients with odontogenic keratocyst. CBCT imaging features were analyzed and compared between the patient and control groups. Overall, 10,323 image features were extracted through feature analysis. A subset of 25 radiomic features obtained after feature selection were analyzed further. These 25 features were significantly different between the 2 groups (p < 0.05). The absolute value of correlation coefficient was 0.487-0.775. Gray-level co-occurrence matrix (GLCM) contrast, neighborhood gray tone difference matrix (NGTDM) contrast, and GLCM variance were the features with the highest correlation coefficients. Pre-operative radiomics analysis showed the differences between simple bone cysts and odontogenic keratocysts, can help to diagnose simple bone cysts. Three specific texture features-GLCM contrast, NGTDM contrast, and GLCM variance-may be the characteristic imaging features of simple bone cysts of the jaw.

Statistical harmonization can improve the development of a multicenter CT-based radiomic model predictive of nonresponse to induction chemotherapy in laryngeal cancers.

To develop a radiomic model predicting nonresponse to induction chemotherapy in laryngeal cancers, from multicenter pretherapeutic contrast-enhanced computed tomography (CE-CT) and evaluate the benefit of feature harmonization in such a context. Patients (n = 104) eligible for laryngeal preservation chemotherapy were included in five centers. Primary tumor was manually delineated on the CE-CT images. The following radiomic features were extracted with an in-house software (MIRAS v1.1, LaTIM UMR 1101): intensity, shape, and textural features derived from Gray-Level Co-occurrence Matrix (GLCM), Neighborhood Gray Tone Difference Matrix (NGTDM), Gray-Level Run Length Matrix (GLRLM), and Gray-Level Size Zone Matrix (GLSZM). Harmonization was performed using ComBat after unsupervised hierarchical clustering, used to determine labels automatically, given the high heterogeneity of imaging characteristics across and within centers. Patients with similar feature distributions were grouped with unsupervised clustering into an optimal number of clusters (2) determined with "silhouette scoring." Statistical harmonization was then carried out with ComBat on these 2 identified clusters. The cohort was split into training/validation (n = 66) and testing (n = 32) sets. Area under the receiver operating characteristics curves (AUC) were used to evaluate the ability of radiomic features (before and after harmonization) to predict nonresponse to chemotherapy, and specificity (Sp) and sensitivity (Se) were used to quantify their performance in the testing set. Without harmonization, none of the features identified as predictive in the training set remained significant in the testing set. After ComBat, one textural feature identified in the training set keeps a predictive trend in the testing set-Zone Percentage, derived from the GLSZM, was predictive of nonresponse in the training set (AUC = 0.62, Se = 70%, Sp = 64%, P = 0.04) and obtained a satisfactory performance in the testing set (Se = 80%, Sp = 67%, P = 0.03), although significance was limited by the size of the testing set. These results are consistent with previously published findings in head and neck cancers. Radiomic features from CE-CT could help in the selection of patients for induction chemotherapy in laryngeal cancers, with relatively good sensitivity and specificity in predicting lack of response. Statistical harmonization with ComBat and unsupervised clustering seems to improve the predictive value of features extracted in such a heterogeneous multicenter setting.

MRI-based radiomics signature for the prediction of response of lung cancer brain metastases after whole-brain radiotherapy.

e21096 MRI-based Radiomics signature for the Prediction of Response of Lung Cancer Brain Metastases After Whole-Brain Radiotherapy Background: Local response prediction for brain metastases (BMs) from lung cancer after Whole-Brain Radiotherapy (WBRT) is challenging, as existing criteria are based solely on unidimensional measurements. This study sought to determine whether radiomic features of lung cancer BMs derived from pre-treatment magnetic resonance imaging (MRI) could be used to predict local response following WBRT. Methods: A total of 88 Lung Cancer patients with BMs treated with WBRT were analyzed. After volumes of interest were drawn, 944 radiomic features including first-order, shape, Gray Level Co-occurrence Matrix (GLCM), Gray Level Dependence Matrix (GLDM), Gray Level Run Length Matrix (GLRLM), Gray Level Size Zone Matrix (GLSZM), Neighborhood Gray Tone Difference Matrix (NGTDM), and Laplacian of Gaussian (LoG) features were extracted, using the baseline pre-treatment post-contrast T1 (T1c) and T2 fluid-attenuated inversion recovery (FLAIR) MRI sequences, respectively. Local response status was determined by contrasting the baseline and follow-up MRI according to the RANO-BM criteria. The independent samples t test or Mann-Whitney U test, and then least absolute shrinkage and selection operator (LASSO) were used for dimensionality reduction and feature selection. An adaboost classifier was trained using the selected radiomic features and evaluated using the area under the receiver operating characteristic curve (AUC) in both the training and testing sets. Other discrimination metrics, including classification accuracy, positive predictive value (PPV), negative predictive value (NPV), sensitivity, and specificity, were also calculated. Results: The optimal radiomics signature was developed based on a multivariable logistic regression with 4, 5, 6 radiomic features on T1c, T2 FLAIR and T1c+T2 FLAIR, respectively. The radiomics model based on T1c features presented the AUC of (0.920 vs. 0.805, respectively) for both the training and testing sets, followed by T2 FLAIR features (0.893 vs. 0.701, respectively), and T1c+T2 FLAIR features (0.971 vs. 0.857, respectively). The classification accuracy of the radiomics model also well predicted the local response of BMs for both the the training and testing sets (T1c: 82.9% vs. 77.8%, T2 FLAIR: 82.9% vs. 77.8%, T1c+T2 FLAIR: 90.0% vs. 77.8%, respectively). Conclusions: Radiomics holds promise for predicting local tumor response following WBRT in patients with lung cancer and brain metastases. A predictive model built on radiomic features from an institutional cohort performed well on cross-validation testing. These results warrant further validation in independent datasets. Such work could prove invaluable for guiding management of individual patients and assessing outcomes of novel interventions.

Are computed-tomography-based hematoma radiomics features reproducible and predictive of intracerebral hemorrhage expansion? an in vitro experiment and clinical study.

To identify reproducible hematoma radiomics features (RFs) for use in predicting hematoma expansion (HE) in patients with acute intracerebral hemorrhage (ICH). For test-retest analysis, three syringes with different volumes of blood collected at the same time (to mimic homogeneous hematoma) and a phantom (FT/HK 2000; Huake, Szechwan, China) containing three cylindrical inserts were scanned seven times within 6 h on the same CT scanner. Three additional syringes with mixed blood collected at different time points (to mimic heterogeneous hematoma) were tied together with the first three syringes as well as the phantom were scanned using modified CT acquisition parameters for intra CT analysis. A coefficient of variation below 10% served as the cutoff value for reproducibility. Finally, reproducible and potentially useful RFs were used to predict HE in 144 acute ICH patients, with the area under the receiver operating characteristic curves (AUC) used to evaluate their diagnostic performance. A total of 630 RFs including 18 first-order, 24 gray-level co-occurrence matrix (GLCM), 16 gray-level run length matrix (GLRLM), five neighborhood gray-tone difference matrix (NGTDM), 63 Laplacian of Gaussian (LoG), and 504 Wavelet features were evaluated. In the test-retest analysis, the percentages of reproducible RFs ranged from 42.54% (268/630) to 45.4% (286/630) for the three homogeneous hematoma samples and 79.05% (498/630) to 81.43% (513/630) for the phantom. In the intra-CT analysis, the percentages varied from 31.43% (198/630) to 42.38% (267/630) for the six hematoma samples and 48.89% (308/630) to 53.97% (340/630) for the phantom. In the in vitro experiment, 148 RFs were reproducible for all hematoma samples in both the test-retest and intra-CT analyses; however, only 80 were statistically different between homogeneous and heterogeneous hematoma samples. Finally, HE occurred in 25% (growth >6 ml, 36/144) to 31.94% (growth >3 ml or 33%, 46/144) of the patients. The AUCs in predicting HE ranged from 0.625 to 0.703. Only a few CT-based RFs from the in vitro hematoma were reproducible and can distinguish between homogeneous and heterogeneous hematomas. The use of RFs alone to predict HE in acute ICH showed only a moderate performance. Using an in vitro experiment and clinical validation, this study demonstrated for the first time that CT-based hematoma RFs can be used to predict HE in acute ICH; nonetheless, only a few RFs are reproducible and can be used for prediction.