Evaluation of texture analysis for the differential diagnosis of focal nodular hyperplasia from hepatocellular adenoma on contrast-enhanced CT images

Radiomics in CT and MR imaging of the liver and pancreas: tools with potential for clinical application

Radiomics allows the extraction of quantitative imaging features from clinical magnetic resonance imaging (MRI) and computerized tomography (CT) studies. The advantages of radiomics have primarily been exploited in oncological applications, including better characterization and staging of oncological lesions and prediction of patient outcomes and treatment response. The potential introduction of radiomics in the clinical setting requires the establishment of a standardized radiomics pipeline and a quality assurance program. Radiomics and texture analysis of the liver have improved the differentiation of hypervascular lesions such as adenomas, focal nodular hyperplasia, and hepatocellular carcinoma (HCC) during the arterial phase, and in the pretreatment determination of HCC prognostic factors (e.g., tumor grade, microvascular invasion, Ki-67 proliferation index). Radiomics of pancreatic CT and MR images has enhanced pancreatic ductal adenocarcinoma detection and its differentiation from pancreatic neuroendocrine tumors, mass-forming chronic pancreatitis, or autoimmune pancreatitis. Radiomics can further help to better characterize incidental pancreatic cystic lesions, accurately discriminating benign from malignant intrapancreatic mucinous neoplasms. Nonetheless, despite their encouraging results and exciting potential, these tools have yet to be implemented in the clinical setting. This non-systematic review will describe the essential steps in the implementation of the radiomics and feature extraction workflow from liver and pancreas CT and MRI studies for their potential clinical application. A succinct overview of reported radiomics applications in the liver and pancreas and the challenges and limitations of their implementation in the clinical setting is also discussed, concluding with a brief exploration of the future perspectives of radiomics in the gastroenterology field.

Focal Lesions of the Liver and Radiomics: What Do We Know?

Despite differences in pathological analysis, focal liver lesions are not always distinguishable in contrast-enhanced magnetic resonance imaging (MRI), contrast-enhanced computed tomography (CT), and positron emission tomography (PET). This issue can cause problems of differential diagnosis, treatment, and follow-up, especially in patients affected by HBV/HCV chronic liver disease or fatty liver disease. Radiomics is an innovative imaging approach that extracts and analyzes non-visible quantitative imaging features, supporting the radiologist in the most challenging differential diagnosis when the best-known methods are not conclusive. The purpose of this review is to evaluate the most significant CT and MRI texture features, which can discriminate between the main benign and malignant focal liver lesions and can be helpful to predict the response to pharmacological or surgical therapy and the patient's prognosis.

Impact of deep learning image reconstruction algorithms on CT radiomic features in patients with liver tumors

Objective

To evaluate the impact of deep learning image reconstruction (DLIR) and adaptive statistical iterative reconstruction-Veo (ASIR-V) on abdominal CT radiomic features acquired in portal venous phase in liver tumor patients.

Methods

Sixty patients with liver tumors who underwent contrast-enhanced abdominal CT were retrospectively enrolled. Six groups including filtered back projection (FBP), ASIR-V (30%, 70%) and DLIR at low (DLIR-L), medium (DLIR-M and high (DLIR-H), were reconstructed using portal venous phase data. CT-based radiomic features (first-order, texture and wavelet features) were extracted from 2D and 3D liver tumors, peritumor and liver parenchyma. All features were analyzed for comparison. P < 0.05 indicated statistically different. The consistency of 3D lesion feature extraction was assessed by calculating intraclass correlation coefficient (ICC).

Results

Different reconstruction algorithms influenced most radiomic features. The percentages of first-order, texture and wavelet features without statistical difference among 2D and 3D lesions, peritumor and liver parenchyma for all six groups were 27.78% (5/18), 5.33% (4/75) and 5.56% (1/18), respectively (all p > 0.05), and they decreased while the level of reconstruction strengthened for both ASIR-V and DLIR. Compared with FBP, the features of ASIR-V30% and 70% without statistical difference decreased from 71.31% to 23.95%, and DLIR-L, DLIR-M, and DLIR-H decreased from 31.65% to 27.11% and 23.73%. Among texture features, unaffected features of peritumor were larger than those of lesions and liver parenchyma, and unaffected 3D lesions features were larger than those of 2D lesions. The consistency of 3D lesion first-order features was excellent, with intra- and inter-observer ICCs ranging from 0.891 to 0.999 and 0.880 to 0.998.

Conclusions

Both ASIR-V and DLIR algorithms with different strengths influenced the radiomic features of abdominal CT images in portal venous phase, and the influences aggravated as reconstruction strength increased.

Advances in liver US, CT, and MRI: moving toward the future

Over the past two decades, the epidemiology of chronic liver disease has changed with an increase in the prevalence of nonalcoholic fatty liver disease in parallel to the advent of curative treatments for hepatitis C. Recent developments provided new tools for diagnosis and monitoring of liver diseases based on ultrasound (US), computed tomography (CT), and magnetic resonance imaging (MRI), as applied for assessing steatosis, fibrosis, and focal lesions. This narrative review aims to discuss the emerging approaches for qualitative and quantitative liver imaging, focusing on those expected to become adopted in clinical practice in the next 5 to 10 years. While radiomics is an emerging tool for many of these applications, dedicated techniques have been investigated for US (controlled attenuation parameter, backscatter coefficient, elastography methods such as point shear wave elastography [pSWE] and transient elastography [TE], novel Doppler techniques, and three-dimensional contrast-enhanced ultrasound [3D-CEUS]), CT (dual-energy, spectral photon counting, extracellular volume fraction, perfusion, and surface nodularity), and MRI (proton density fat fraction [PDFF], elastography [MRE], contrast enhancement index, relative enhancement, T1 mapping on the hepatobiliary phase, perfusion). Concurrently, the advent of abbreviated MRI protocols will help fulfill an increasing number of examination requests in an era of healthcare resource constraints.

Differentiating infected focal liver lesions from malignant mimickers: value of ultrasound-based radiomics

AIM

To establish an ultrasound-based radiomics model through machine learning methods and then to assess the ability of the model to differentiate infected focal liver lesions from malignant mimickers.

MATERIALS AND METHODS

A total of 104 patients with infected focal liver lesions and 485 patients with malignant hepatic tumours were included, consisting of hepatocellular carcinoma (HCC), cholangiocarcinoma (CC), combined hepatocellular-cholangiocarcinoma (cHCC-CC), and liver metastasis. Radiomics features were extracted from grey-scale ultrasound images. Feature selection and predictive modelling were carried out by dimensionality reduction methods and classifiers. The diagnostic effect of the prediction mode was assessed by receiver operating characteristic (ROC) curve analysis.

RESULTS

In total, 5,234 radiomics features were extracted from grey-scale ultrasound image of every focal liver lesion. The ultrasound-based radiomics model had a favourable predictive value for differentiating infected focal liver lesions from malignant hepatic tumours, with an area under the curve (AUC) of 0.887 and 0.836 (HCC group), 0.896 and 0.766 (CC group), 0.944 and 0.754 (cHCC-CC group), 0.918 and 0.808 (liver metastasis group), and 0.949 and 0.745 (malignant hepatic tumour group) for the training set and validation set, respectively.

CONCLUSIONS

Ultrasound-based radiomics is helpful in differentiating infected focal liver lesions from malignant mimickers and has the potential for use as a supplement to conventional grey-scale ultrasound and contrast-enhanced ultrasound (CEUS).

Computed Tomography Image Texture under Feature Extraction Algorithm in the Diagnosis of Effect of Specific Nursing Intervention on Mycoplasma Pneumonia in Children

To evaluate the effect of specific nursing intervention in children with mycoplasma pneumonia (MP), a feature extraction algorithm based on gray level co-occurrence matrix (GLCM) was proposed and combined with computed tomography (CT) image texture features. Then, 98 children with MP were rolled into the observation group with 49 cases (specific nursing) and the control group with 49 cases (routine nursing). CT images based on feature extraction algorithm of optimized GLCM were used to examine the children before and after nursing intervention, and the recovery of the two groups of children was discussed. The results showed that the proportion of lung texture increase, rope shadow, ground glass shadow, atelectasis, and pleural effusion in the observation group (24.11%, 3.86%, 8.53%, 15.03%, and 3.74%) was significantly lower than that in the control group (28.53%, 10.23%, 13.34%, 21.15%, and 8.13%) after nursing (P < 0.05). There were no significant differences in the proportion of small patchy shadows, large patchy consolidation shadows, and bronchiectasis between the observation group and the control group (P > 0.05). In the course of nursing intervention, in the observation group, the disappearance time of cough, normal temperature, disappearance time of lung rales, and absorption time of lung shadow (2.15 ± 0.86 days, 4.81 ± 1.14 days, 3.64 ± 0.55 days, and 5.96 ± 0.62 days) were significantly shorter than those in the control group (2.87 ± 0.95 days, 3.95 ± 1.06 days, 4.51 ± 1.02 days, and 8.14 ± 1.35 days) (P < 0.05). After nursing intervention, the proportion of satisfaction and total satisfaction in the experimental group (67.08% and 28.66%) was significantly higher than that in the control group (40.21% and 47.39%), while the proportion of dissatisfaction (4.26%) was significantly lower than that in the control group (12.4%) (P < 0.05). To sum up, specific nursing intervention was more beneficial to improve the progress of characterization recovery and the overall recovery effect of children with MP relative to conventional nursing. CT image based on feature extraction algorithm of optimized GLCM was of good adoption value in the diagnosis and treatment of MP in children.

Current Status of Radiomics and Deep Learning in Liver Imaging

ABSTRACT

Artificial intelligence is poised to revolutionize medical image. It takes advantage of the high-dimensional quantitative features present in medical images that may not be fully appreciated by humans. Artificial intelligence has the potential to facilitate automatic organ segmentation, disease detection and characterization, and prediction of disease recurrence. This article reviews the current status of artificial intelligence in liver imaging and reviews the opportunities and challenges in clinical implementation.

CT-Based Radiomics Nomogram: A Potential Tool for Differentiating Hepatocellular Adenoma From Hepatocellular Carcinoma in the Noncirrhotic Liver

RATIONALE AND OBJECTIVES

To evaluate the value of a radiomics nomogram for preoperative differentiating hepatocellular adenoma (HCA) from hepatocellular carcinoma (HCC) in the noncirrhotic liver.

MATERIALS AND METHODS

One hundred and thirty-one patients with HCA (n = 46) and HCC (n = 85) were divided into a training set (n = 93) and a test set (n = 38). Clinical data and CT findings were analyzed. Radiomics features were extracted from the triphasic contrast CT images. A radiomics signature was constructed with the least absolute shrinkage and selection operator algorithm and a radiomics score was calculated. Combined with the radiomics score and independent clinical factors, a radiomics nomogram was developed by multivariate logistic regression analysis. The performance of the radiomics nomogram was assessed by calibration, discrimination and clinical usefulness.

RESULTS

Gender, age, and enhancement pattern were the independent clinical factors. Three thousand seven hundred and sixty-eight features were extracted and reduced to 7 features as the optimal discriminators to build the radiomics signature. The radiomics nomogram (area under the curve [AUC], 0.96; 95% confidence interval [CI], 0.93-0.99) and the clinical factors model (AUC, 0.93; 95%CI, 0.88-0.99) showed better discrimination capability (p = 0.001 and 0.047) than the radiomics signature (AUC, 0.83; 95%CI, 0.74-0.92) in the training set. In the test set, the radiomics nomogram (AUC, 0.94; 95%CI, 0.87-1.00) performed better (p = 0.013) than the radiomics signature (AUC, 0.75; 95%CI, 0.59-0.91). Decision curve analysis showed the radiomics nomogram outperformed the clinical factors model and the radiomics signature in terms of clinical usefulness.

CONCLUSION

The CT-based radiomics nomogram has the potential to accurately differentiate HCA from HCC in the noncirrhotic liver.

CT Texture Analysis for Differentiating Bronchiolar Adenoma, Adenocarcinoma In Situ, and Minimally Invasive Adenocarcinoma of the Lung

Purpose: This study aimed to investigate the potential of computed tomography (CT) imaging features and texture analysis to distinguish bronchiolar adenoma (BA) from adenocarcinoma in situ (AIS)/minimally invasive adenocarcinoma (MIA).

Materials and Methods: Fifteen patients with BA, 38 patients with AIS, and 36 patients with MIA were included in this study. Clinical data and CT imaging features of the three lesions were evaluated. Texture features were extracted from the thin-section unenhanced CT images using Artificial Intelligence Kit software. Then, multivariate logistic regression analysis based on selected texture features was employed to distinguish BA from AIS/MIA. Receiver operating characteristics curves were performed to determine the diagnostic performance of the features.

Results: By comparison with AIS/MIA, significantly different CT imaging features of BA included nodule type, tumor size, and pseudo-cavitation sign. Among them, pseudo-cavitation sign had a moderate diagnostic value for distinguishing BA and AIS/MIA (AUC: 0.741 and 0.708, respectively). Further, a total of 396 quantitative texture features were extracted. After comparation, the top six texture features showing the most significant difference between BA and AIS or MIA were chosen. The ROC results showed that these key texture features had a high diagnostic value for differentiating BA from AIS or MIA, among which the value of a comprehensive model with six selected texture features was the highest (AUC: 0.977 or 0.976, respectively) for BA and AIS or MIA. These results indicated that texture analyses can effectively improve the efficacy of thin-section unenhanced CT for discriminating BA from AIS/MIA.

Conclusion: CT texture analysis can effectively improve the efficacy of thin-section unenhanced CT for discriminating BA from AIS/MIA, which has a potential clinical value and helps pathologist and clinicians to make diagnostic and therapeutic strategies.

Robotic resection of liver focal nodal hyperplasia guided by indocyanine green fluorescence imaging: A preliminary analysis of 23 cases

BACKGROUND

Focal nodal hyperplasia (FNH) is a common benign tumor of the liver. It occurs mostly in people aged 40-50 years and 90% of the patients are female. FNH can be cured by local resection. How to locate and judge the tumor boundary in real time is often a challenge for surgeons.

AIM

To summarize the technique and feasibility of robotic resection of FNH guided by indocyanine green (ICG) fluorescence imaging.

METHODS

The demographics and perioperative outcomes of a consecutive series of patients who underwent robotic resection of liver FNH guided by ICG fluorescence imaging between May 1, 2018 and September 30, 2019 were retrospectively analyzed. ICG was injected through the median elbow vein in all the patients at a dose of 0.25 mg/kg 48 h before the operation. During the operation, the position of FNH in the liver was located in the fluorescence mode of the Da Vinci Si robot operating system and the tumor boundary was determined during the resection.

RESULTS

Among the 23 patients, there were 11 males and 12 females, with a mean age of 30.5 ± 9.3 years. Twenty-two cases completed robotic resection, while one (4.3%) case converted to open surgery. In the robotic surgery group, the operation time was 35-340 min with a median of 120 min, the intraoperative bleeding was 10-800 mL with a median of 50 mL, and the postoperative hospital stay was 1-7 d with a median of 4 d. Biliary fistula occurred in two (8.7%) patients after robotic operation and they both recovered after conservative treatment. One (4.3%) patient received blood transfusion and there was no death in this study. The postoperative hospital stay in the small tumor group was significantly shorter than that in the large tumor group (P < 0.05).

CONCLUSION

ICG fluorescence imaging can guide the surgeon to perform robotic resection of liver FNH by locating the tumor and displaying the tumor boundary in real time. It is a safe and feasible method to ensure the complete resection of the tumor.

Texture analysis on preoperative contrast-enhanced magnetic resonance imaging identifies microvascular invasion in hepatocellular carcinoma

BACKGROUND

Radiomic texture analysis quantifies tumor heterogeneity. The aim of this study is to determine if radiomics can predict biologic aggressiveness in HCC and identify tumors with MVI.

METHODS

Single-center, retrospective review of HCC patients undergoing resection/ablation with curative intent from 2009 to 2017. DICOM images from preoperative MRIs were analyzed with texture analysis software. Texture analysis parameters extracted on T1, T2, hepatic arterial phase (HAP) and portal venous phase (PVP) images. Multivariate logistic regression analysis evaluated factors associated with MVI.

RESULTS

MVI was present in 52.2% (n = 133) of HCCs. On multivariate analysis only T1 mean (OR = 0.97, 95%CI 0.95-0.99, p = 0.043) and PVP entropy (OR = 4.7, 95%CI 1.37-16.3, p = 0.014) were associated with tumor MVI. Area under ROC curve was 0.83 for this final model. Empirical optimal cutpoint for PVP tumor entropy and T1 tumor mean were 5.73 and 23.41, respectively. At these cutpoint values, sensitivity was 0.68 and 0.5, respectively and specificity was 0.64 and 0.86. When both criteria were met, the probability of MVI in the tumor was 87%.

CONCLUSION

Tumor entropy and mean are both associated with MVI. Texture analysis on preoperative imaging correlates with microscopic features of HCC and can be used to predict patients with high-risk tumors.

Contrast-Enhanced CT Texture Analysis for Distinguishing Fat-Poor Renal Angiomyolipoma From Chromophobe Renal Cell Carcinoma

Objective:

To evaluate the value of 2-dimensional (2D) and 3-dimensional (3D) computed tomography texture analysis (CTTA) models in distinguishing fat-poor angiomyolipoma (fpAML) from chromophobe renal cell carcinoma (chRCC).

Methods:

We retrospectively enrolled 32 fpAMLs and 24 chRCCs. Texture features were extracted from 2D and 3D regions of interest in triphasic CT images. The 2D and 3D CTTA models were constructed with the least absolute shrinkage and selection operator algorithm and texture scores were calculated. The diagnostic performance of the 2D and 3D CTTA models was evaluated with respect to calibration, discrimination, and clinical usefulness.

Results:

Of the 177 and 183 texture features extracted from 2D and 3D regions of interest, respectively, 5 2D features and 8 3D features were selected to build 2D and 3D CTTA models. The 2D CTTA model (area under the curve [AUC], 0.811; 95% confidence interval [CI], 0.695-0.927) and the 3D CTTA model (AUC, 0.915; 95% CI, 0.838-0.993) showed good discrimination and calibration (P > .05). There was no significant difference in AUC between the 2 models (P = .093). Decision curve analysis showed the 3D model outperformed the 2D model in terms of clinical usefulness.

Conclusions:

The CTTA models based on contrast-enhanced CT images had a high value in differentiating fpAML from chRCC.

Diagnostic accuracy of unenhanced CT texture analysis to differentiate mass-forming pancreatitis from pancreatic ductal adenocarcinoma

PURPOSE

To investigate the value of texture analysis on unenhanced computed tomography (CT) to potentially differentiate mass-forming pancreatitis (MFP) from pancreatic ductal adenocarcinoma (PDAC).

METHODS

A retrospective study consisting of 109 patients (30 MFP patients vs 79 PDAC patients) who underwent preoperative unenhanced CT between January 2012 and December 2017 was performed. Synthetic minority oversampling technique (SMOTE) algorithm was adopted to reconstruct and balance MFP and PDAC samples. A total of 396 radiomic features were extracted from unenhanced CT images. Mann-Whitney U test and minimum redundancy maximum relevance (MRMR) methods were used for the purpose of dimension reduction. Predictive models were constructed using random forest (RF) method, and were validated using leave group out cross-validation (LGOCV) method. Diagnostic performance of the predictive model, including sensitivity, specificity, accuracy, positive predicting value (PPV), and negative predicting value (NPV), was recorded.

RESULTS

We applied 200% of SMOTE to MFP and PDAC patients, resulting in 90 MFP patients compared with 120 PDAC patients. Dimension reduction steps yielded 30 radiomic features using Mann-Whitney U test and MRMR methods. Ten radiomic features were retained using RF method. Four most predictive parameters, including GreyLevelNonuniformity_angle90_offset1, VoxelValueSum, HaraVariance, and ClusterProminence_AllDirection_offset1_SD, were used to generate the predictive model with preferable 92.2% sensitivity, 94.2% specificity, 93.3% accuracy, 92.2% PPV, and 94.2% NPV. Finally, in LGOCV analysis, a high pooled mean sensitivity, specificity, and accuracy (82.6%, 80.8%, and 82.1%, respectively) indicate a relatively reliable and stable predictive model.

CONCLUSIONS

Unenhanced CT texture analysis can be a promising noninvasive method in discriminating MFP from PDAC.

A CT-based radiomics nomogram for differentiation of focal nodular hyperplasia from hepatocellular carcinoma in the non-cirrhotic liver

Background

The purpose of this study was to develop and validate a radiomics nomogram for preoperative differentiating focal nodular hyperplasia (FNH) from hepatocellular carcinoma (HCC) in the non-cirrhotic liver.

Methods

A total of 156 patients with FNH (n = 55) and HCC (n = 101) were divided into a training set (n = 119) and a validation set (n = 37). Radiomics features were extracted from triphasic contrast CT images. A radiomics signature was constructed with the least absolute shrinkage and selection operator algorithm, and a radiomics score (Rad-score) was calculated. Clinical data and CT findings were assessed to build a clinical factors model. Combined with the Rad-score and independent clinical factors, a radiomics nomogram was constructed by multivariate logistic regression analysis. Nomogram performance was assessed with respect to discrimination and clinical usefulness.

Results

Four thousand two hundred twenty-seven features were extracted and reduced to 10 features as the most important discriminators to build the radiomics signature. The radiomics signature showed good discrimination in the training set (AUC [area under the curve], 0.964; 95% confidence interval [CI], 0.934–0.995) and the validation set (AUC, 0.865; 95% CI, 0.725–1.000). Age, Hepatitis B virus infection, and enhancement pattern were the independent clinical factors. The radiomics nomogram, which incorporated the Rad-score and clinical factors, showed good discrimination in the training set (AUC, 0.979; 95% CI, 0.959–0.998) and the validation set (AUC, 0.917; 95% CI, 0.800–1.000), and showed better discrimination capability (P < 0.001) compared with the clinical factors model (AUC, 0.799; 95% CI, 0.719–0.879) in the training set. Decision curve analysis showed the nomogram outperformed the clinical factors model in terms of clinical usefulness.

Conclusions

The CT-based radiomics nomogram, a noninvasive preoperative prediction tool that incorporates the Rad-score and clinical factors, shows favorable predictive efficacy for differentiating FNH from HCC in the non-cirrhotic liver, which might facilitate clinical decision-making process.

Evaluation of Texture Analysis for the Differential Diagnosis of Mass-Forming Pancreatitis From Pancreatic Ductal Adenocarcinoma on Contrast-Enhanced CT Images

Purpose: To investigate the potential of computed tomography (CT) imaging features and texture analysis to differentiate between mass-forming pancreatitis (MFP) and pancreatic ductal adenocarcinoma (PDAC).

Materials and Methods: Thirty patients with pathologically proved MFP and 79 patients with PDAC were included in this study. Clinical data and CT imaging features of the two lesions were evaluated. Texture features were extracted from arterial and portal phase CT images using commercially available software (AnalysisKit). Multivariate logistic regression analyses were used to identify relevant CT imaging and texture parameters to discriminate MFP from PDAC. Receiver operating characteristic curves were performed to determine the diagnostic performance of predictions.

Results: MFP showed a larger size compared to PDAC (p = 0.009). Cystic degeneration, pancreatic ductal dilatation, vascular invasion, and pancreatic sinistral portal hypertension were more frequent and duct penetrating sign was less frequent in PDAC compared to MFP. Arterial CT attenuation, arterial, and portal enhancement ratios of MFP were higher than PDAC (p < 0.05). In multivariate analysis, arterial CT attenuation and pancreatic duct penetrating sign were independent predictors. Texture features in arterial phase including SurfaceArea, Percentile40, InverseDifferenceMoment_angle90_offset4, LongRunEmphasis_angle45_offset4, and uniformity were independent predictors. Texture features in portal phase including LongRunEmphasis_angle135_offset7, VoxelValueSum, LongRunEmphasis_angle135_offset4, and GLCMEntropy_angle45_offset1 were independent predictors. Areas under the curve of imaging feature-based, texture feature-based in arterial and portal phases, and the combined models were 0.84, 0.96, 0.93, and 0.98, respectively.

Conclusions: CT texture analysis demonstrates great potential to differentiate MFP from PDAC.

Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection

PURPOSE

To evaluate the feasibility of using computed tomography texture analysis (CTTA) parameters for predicting malignant risk grade and mitosis index of gastrointestinal stromal tumors (GISTs), compared with visual inspection.

METHOD AND MATERIALS

CTTA was performed on portal phase CT images of 145 surgically confirmed GISTs (mean size: 42.9 ± 37.5 mm), using TexRAD software. Mean, standard deviation, entropy, mean of positive pixels (MPP), skewness, and kurtosis of CTTA parameters, on spatial scaling factor (SSF), 2-6 were compared by risk grade, mitosis rate, and the presence or absence of necrosis on visual inspection. CTTA parameters were correlated with risk grade. Diagnostic performance was evaluated with receiver operating characteristic curve analysis. Enhancement pattern, necrosis, heterogeneity, calcification, growth pattern, and mucosal ulceration were subjectively evaluated by two observers.

RESULTS

Three to four parameters at different scales were significantly different according to the risk grade, mitosis rate, and the presence or absence of necrosis (p < 0.041). MPP at fine or medium scale (r = - 0.547 to - 393) and kurtosis at coarse scale (r = 0.424-0.454) correlated significantly with risk grade (p < 0.001). HG-GIST was best differentiated from LG-GIST by MPP at SSF 2 (AUC, 0.782), and kurtosis at SSF 4 (AUC, 0.779) (all p < 0.001). CT features predictive of HG-GIST were density lower than or equal to that of the erector spinae muscles on enhanced images (OR 2.1; p = 0.037; AUC, 0.59), necrosis (OR, 6.1; p < 0.001; AUC, 0.70), heterogeneity (OR, 4.3; p < 0.001; AUC, 0.67), and mucosal ulceration (OR, 3.3; p = 0.002; AUC, 0.62).

CONCLUSION

Using TexRAD, MPP and kurtosis are feasible in predicting risk grade and mitosis index of GISTs. CTTA demonstrated meaningful accuracy in preoperative risk stratification of GISTs.