Radiomics analysis enables fatal outcome prediction for hospitalized patients with coronavirus disease 2019 (COVID-19).

BACKGROUND: In December 2019, a rare respiratory disease named coronavirus disease 2019 (COVID-19) broke out, leading to great concern around the world. PURPOSE: To develop and validate a radiomics nomogram for predicting the fatal outcome of COVID-19 pneumonia. MATERIAL AND METHODS: The present study consisted of a training dataset (n = 66) and a validation dataset (n = 30) with COVID-19 from January 2020 to March 2020. A radiomics signature was generated using the least absolute shrinkage and selection operator (LASSO) Cox regression model. A radiomics score (Rad-score) was developed from the training cohort. The radiomics model, clinical model, and integrated model were built to assess the association between radiomics signature/clinical characteristics and the mortality of COVID-19 cases. The radiomics signature combined with the Rad-score and the independent clinical factors and radiomics nomogram were constructed. RESULTS: Seven stable radiomics features associated with the mortality of COVID-19 were finally selected. A radiomics nomogram was based on a combined model consisting of the radiomics signature and the clinical risk factors indicating optimal predictive performance for the fatal outcome of patients with COVID-19 with a C-index of 0.912 (95% confidence interval [CI] 0.867-0.957) in the training dataset and 0.907 (95% CI 0.849-0.966) in the validation dataset. The calibration curves indicated optimal consistency between the prediction and the observation in both training and validation cohorts. CONCLUSION: The CT-based radiomics nomogram indicated favorable predictive efficacy for the overall survival risk of patients with COVID-19, which could help clinicians intensively follow up high-risk patients and make timely diagnoses.

Development and multicenter validation of a CT-based radiomics signature for predicting severe COVID-19 pneumonia.

OBJECTIVES: To develop and validate a radiomics nomogram for timely predicting severe COVID-19 pneumonia. MATERIALS AND METHODS: Three hundred and sixteen COVID-19 patients (246 non-severe and 70 severe) were retrospectively collected from two institutions and allocated to training, validation, and testing cohorts. Radiomics features were extracted from chest CT images. Radiomics signature was constructed based on reproducible features using the least absolute shrinkage and selection operator (LASSO) logistic regression algorithm with 5-fold cross-validation. Logistic regression modeling was employed to build different models based on quantitative CT features, radiomics signature, clinical factors, and/or the former combined features. Nomogram performance for severe COVID-19 prediction was assessed with respect to calibration, discrimination, and clinical usefulness. RESULTS: Sixteen selected features were used to build the radiomics signature. The CT-based radiomics model showed good calibration and discrimination in the training cohort (AUC, 0.9; 95% CI, 0.843-0.942), the validation cohort (AUC, 0.878; 95% CI, 0.796-0.958), and the testing cohort (AUC, 0.842; 95% CI, 0.761-0.922). The CT-based radiomics model showed better discrimination capability (all p < 0.05) compared with the clinical factors joint quantitative CT model (AUC, 0.781; 95% CI, 0.708-0.843) in the training cohort, the validation cohort (AUC, 0.814; 95% CI, 0.703-0.897), and the testing cohort (AUC, 0.696; 95% CI, 0.581-0.796). Decision curve analysis demonstrated that in terms of clinical usefulness, the radiomics model outperformed the clinical factors model and quantitative CT model alone. CONCLUSIONS: The CT-based radiomics signature shows favorable predictive efficacy for severe COVID-19, which might assist clinicians in tailoring precise therapy. KEY POINTS: • Radiomics can be applied in CT images of COVID-19 and radiomics signature was an independent predictor of severe COVID-19. • CT-based radiomics model can predict severe COVID-19 with satisfactory accuracy compared with subjective CT findings and clinical factors. • Radiomics nomogram integrated with the radiomics signature, subjective CT findings, and clinical factors can achieve better severity prediction with improved diagnostic performance.

A strategy for quality control of ginkgo biloba preparations based on UPLC fingerprint analysis and multi-component separation combined with quantitative analysis.

BACKGROUND: Many studies have assessed the fingerprint and quantitative analysis of Ginkgo biloba preparations, but the fingerprint mainly focuses on flavonoid glycosides. However, according to our previous study, the differences among diverse manufacturers mainly involve organic acids. METHODS: A novel reverse-phase liquid chromatography assay using diode array detection was developed for evaluating Ginkgo biloba preparations for quality based on a chromatographic fingerprint allowing the simultaneous assessment of eleven compounds, including four organic acids, six flavonol glycosides and one flavonoid aglycone. And the method was applied to 51 batches of Ginkgo biloba preparations from manufacturers in China. Chemometric approaches were performed for evaluating 51 batches of Ginkgo biloba preparations from various manufacturers. RESULTS: The similarity values among the chromatograms of 51 samples ranged from 0.45 to 1.00, showing that the quality of Ginkgo biloba preparations produced by different manufacturers varied greatly. Data analysis of the 51 batches of GBP samples suggested significant variations of the total contents of all 11 targets, also demonstrating the quality difference of GBP samples. There were significant differences in organic acids in particular. CONCLUSION: Combining the chemical fingerprint and quantitative assessment revealed significant variations in the examined commercial products with regard to organic acids. Thus, this study provided a more comprehensive tool for monitoring the quality consistency of Ginkgo biloba preparations.

Validation of SE-EPI-based T2 mapping for characterization of prostate cancer: a new method compared with the traditional CPMG method.

PURPOSE: We aim to compare the results of spin echo-echo planar imaging (SE-EPI)-based T2 mapping with those of the conventional Carr-Purcell-Meiboom-Gill (CPMG) method and to investigate the potential validity of SE-EPI-T2 mapping for the characterization of prostate cancer (PCa). METHODS: Our retrospective study included 42 PCa patients and 42 noncancer patients who underwent 3.0T MRI with b values ranging from 0 to 2000 s/mm2 and echo times (TEs) ranging from 32 to 100 ms before biopsies. Bland-Altman analysis was used to compare the agreement between the two methods. The correlations between CPMG-T2 values and SE-EPI-T2 values at different b values were determined by Spearman's rho analysis or Pearson analysis. The Mann-Whitney U test and two-sample t tests were used to analyze the differences between the cancerous and noncancerous groups. RESULTS: Substantial agreement regarding the measurements was observed between the two methods. The average correlation between the CPMG-T2 values and SE-EPI-T2 values was moderate and positive, and the best correlations were found at b = 200 s/mm2 in the noncancer group (r = 0.557, P = 0.000) and at b = 100 s/mm2 in the cancer group (r = 0.537, P = 0.000). In addition, statistically significant differences were found between the noncancer and cancer groups in T2 values and ADC values (diff TEs) (P = 0.000). CONCLUSIONS: Substantial agreement in the measurements was found between the SE-EPI method and CPMG method. SE-EPI-based T2 mapping has potential clinical value for the prostate and can be considered an alternative to the traditional CPMG-T2 mapping method.

Multi-parametric MRI-based radiomics signature for discriminating between clinically significant and insignificant prostate cancer: Cross-validation of a machine learning method.

PURPOSE: To evaluate the performance of a multi-parametric MRI (mp-MRI)-based radiomics signature for discriminating between clinically significant prostate cancer (csPCa) and insignificant PCa (ciPCa). MATERIALS AND METHODS: Two hundred and eighty patients with pathology-proven PCa were enrolled and were randomly divided into training and test cohorts. Eight hundred and nineteen radiomics features were extracted from mp-MRI for each patient. The minority group in the training cohort was balanced via the synthetic minority over-sampling technique (SMOTE) method. We used minimum-redundancy maximum-relevance (mRMR) selection and the LASSO algorithm for feature selection and radiomics signature building. The classification performance of the radiomics signature for csPCa and ciPCa was evaluated by receiver operating characteristic curve analysis in the training and test cohorts. RESULTS: Nine features were selected for the radiomics signature building. Significant differences in the radiomics signature existed between the csPCa and ciPCa groups in both the training and test cohorts (p < 0.01 for both). The AUC, sensitivity and specificity of the radiomics signature were 0.872 (95% CI: 0.823-0.921), 0.883, and 0.753, respectively, in the training cohort, and 0.823 (95% CI: 0.669-0.976), 0.841, and 0.727, respectively, in the test cohort. CONCLUSION: Mp-MRI-based radiomics signature have the potential to noninvasively discriminate between csPCa and ciPCa.

Multi-model Analysis of Diffusion-weighted Imaging of Normal Testes at 3.0 T: Preliminary Findings.

RATIONALE AND OBJECTIVES: This study aimed to establish diffusion quantitative parameters (apparent diffusion coefficient [ADC], DDC, α, Dapp, and Kapp) in normal testes at 3.0 T. MATERIALS AND METHODS: Sixty-four healthy volunteers in two age groups (A: 10-39 years; B: ≥ 40 years) underwent diffusion-weighted imaging scanning at 3.0 T. ADC1000, ADC2000, ADC3000, DDC, α, Dapp, and Kapp were calculated using the mono-exponential, stretched-exponential, and kurtosis models. The correlations between parameters and the age were analyzed. The parameters were compared between the age groups and between the right and the left testes. RESULTS: The average ADC1000, ADC2000, ADC3000, DDC, α, Dapp, and Kapp values did not significantly differ between the right and the left testes (P > .05 for all). The following significant correlations were found: positive correlations between age and testicular ADC1000, ADC2000, ADC3000, DDC, and Dapp (r = 0.516, 0.518, 0.518, 0.521, and 0.516, respectively; P < .01 for all) and negative correlations between age and testicular α and Kapp (r = -0.363, -0.427, respectively; P < .01 for both). Compared to group B, in group A, ADC1000, ADC2000, ADC3000, DDC, and Dapp were significantly lower (P < .05 for all), but α and Kapp were significantly higher (P < .05 for both). CONCLUSIONS: Our study demonstrated the applicability of the testicular mono-exponential, stretched-exponential, and kurtosis models. Our results can help establish a baseline for the normal testicular parameters in these diffusion models. The contralateral normal testis can serve as a suitable reference for evaluating the abnormalities of the other side. The effect of age on these parameters requires further attention.

Effects of Echo Time on IVIM Quantification of the Normal Prostate.

The two-compartment intravoxel incoherent motion (IVIM) theory assumes that the transverse relaxation time is the same in both compartments. However, blood and tissue have different T2 values, and echo time (TE) may thus have an effect on the quantitative parameters of IVIM. The purpose of this study was to investigate the effects of TE on IVIM-DWI-derived parameters of the prostate. In total, 17 healthy volunteers underwent two repeat examinations. IVIM-DWI data were scanned 6 times with variable TE values of 60, 70, 80, 90, 100, and 120 ms. The ADC of a mono-exponential model and the D, D*, and f parameters of the IVIM model were calculated separately for each TE. Repeat measures were assessed by calculating the coefficient of variation and Bland-Altman limits of agreement for each parameter. Spearman's rho test was used to analyse relationships between IVIM indices and TE. Our results showed that TE had an effect on IVIM quantification, which should be kept constant in the examination protocol at each individual institution. Alternatively, an extended IVIM could be used to eliminate the effect of the TE value on the quantitative parameters of IVIM. This may be helpful for guiding clinical research, especially for longitudinal studies.

Characterization of testicular germ cell tumors: Whole-lesion histogram analysis of the apparent diffusion coefficient at 3T.

PURPOSE: To assess the values of parameters derived from whole-lesion histograms of the apparent diffusion coefficient (ADC) at 3T for the characterization of testicular germ cell tumors (TGCTs). MATERIALS AND METHODS: A total of 24 men with TGCTs underwent 3T diffusion-weighted imaging. Fourteen tumors were pathologically confirmed as seminomas, and ten tumors were pathologically confirmed as nonseminomas. Whole-lesion histogram analysis of the ADC values was performed. A Mann-Whitney U test was employed to compare the differences in ADC histogram parameters between seminomas and nonseminomas. Receiver operating characteristic analysis was used to identify the cutoff values for each parameter for differentiating seminomas from nonseminomas; furthermore, the area under the curve (AUC) was calculated to evaluate the diagnostic accuracy. RESULTS: The median of 10th, 25th, 50th, 75th, and 90th percentiles and mean, minimum and maximum ADC values were all significantly reduced for seminomas compared with nonseminomas (p<0.05 for all). In contrast, the median of kurtosis and skewness of ADC values of seminomas were both significantly increased compared with those of nonseminomas (p=0.003 and 0.001, respectively). For differentiating nonseminomas from seminomas, the 10th percentile ADC yielded the highest AUC with a sensitivity and specificity of 100% and 92.86%, respectively. CONCLUSION: Whole-lesion histogram analysis of ADCs might be used for preoperative characterization of TGCTs.

Diagnostic Performance and Interobserver Consistency of the Prostate Imaging Reporting and Data System Version 2: A Study on Six Prostate Radiologists with Different Experiences from Half a Year to 17 Years.

BACKGROUND: One of the main aims of the updated Prostate Imaging Reporting and Data System Version 2 (PI-RADS v2) is to diminish variation in the interpretation and reporting of prostate imaging, especially among readers with varied experience levels. This study aimed to retrospectively analyze diagnostic consistency and accuracy for prostate disease among six radiologists with different experience levels from a single center and to evaluate the diagnostic performance of PI-RADS v2 scores in the detection of clinically significant prostate cancer (PCa). METHODS: From December 2014 to March 2016, 84 PCa patients and 99 benign prostatic shyperplasia patients who underwent 3.0T multiparametric magnetic resonance imaging before biopsy were included in our study. All patients received evaluation according to the PI-RADS v2 scale (1-5 scores) from six blinded readers (with 6 months and 2, 3, 4, 5, or 17 years of experience, respectively, the last reader was a reviewer/contributor for the PI-RADS v2). The correlation among the readers' scores and the Gleason score (GS) was determined with the Kendall test. Intra-/inter-observer agreement was evaluated using κ statistics, while receiver operating characteristic curve and area under the curve analyses were performed to evaluate the diagnostic performance of the scores. RESULTS: Based on the PI-RADS v2, the median κ score and standard error among all possible pairs of readers were 0.506 and 0.043, respectively; the average correlation between the six readers' scores and the GS was positive, exhibiting weak-to-moderate strength (r = 0.391, P = 0.006). The AUC values of the six radiologists were 0.883, 0.924, 0.927, 0.932, 0.929, and 0.947, respectively. CONCLUSION: The inter-reader agreement for the PI-RADS v2 among the six readers with different experience is weak to moderate. Different experience levels affect the interpretation of MRI images.

Using the prostate imaging reporting and data system version 2 (PI-RIDS v2) to detect prostate cancer can prevent unnecessary biopsies and invasive treatment.

Prostate cancer (PCa) is one of the most common cancers among men globally. The authors aimed to evaluate the ability of the Prostate Imaging Reporting and Data System version 2 (PI-RADS v2) to classify men with PCa, clinically significant PCa (CSPCa), or no PCa, especially among those with serum total prostate-specific antigen (tPSA) levels in the "gray zone" (4-10 ng ml-1). A total of 308 patients (355 lesions) were enrolled in this study. Diagnostic efficiency was determined. Univariate and multivariate analyses, receiver operating characteristic curve analysis, and decision curve analysis were performed to determine and compare the predictors of PCa and CSPCa. The results suggested that PI-RADS v2, tPSA, and prostate-specific antigen density (PSAD) were independent predictors of PCa and CSPCa. A PI-RADS v2 score ≥4 provided high negative predictive values (91.39% for PCa and 95.69% for CSPCa). A model of PI-RADS combined with PSA and PSAD helped to define a high-risk group (PI-RADS score = 5 and PSAD ≥0.15 ng ml-1 cm-3, with tPSA in the gray zone, or PI-RADS score ≥4 with high tPSA level) with a detection rate of 96.1% for PCa and 93.0% for CSPCa while a low-risk group with a detection rate of 6.1% for PCa and 2.2% for CSPCa. It was concluded that the PI-RADS v2 could be used as a reliable and independent predictor of PCa and CSPCa. The combination of PI-RADS v2 score with PSA and PSAD could be helpful in the prediction and diagnosis of PCa and CSPCa and, thus, may help in preventing unnecessary invasive procedures.

Evaluation of different mathematical models and different b-value ranges of diffusion-weighted imaging in peripheral zone prostate cancer detection using b-value up to 4500 s/mm2.

OBJECTIVES: To evaluate the diagnostic performance of different mathematical models and different b-value ranges of diffusion-weighted imaging (DWI) in peripheral zone prostate cancer (PZ PCa) detection. METHODS: Fifty-six patients with histologically proven PZ PCa who underwent DWI-magnetic resonance imaging (MRI) using 21 b-values (0-4500 s/mm2) were included. The mean signal intensities of the regions of interest (ROIs) placed in benign PZs and cancerous tissues on DWI images were fitted using mono-exponential, bi-exponential, stretched-exponential, and kurtosis models. The b-values were divided into four ranges: 0-1000, 0-2000, 0-3200, and 0-4500 s/mm2, grouped as A, B, C, and D, respectively. ADC, , D*, f, DDC, α, Dapp, and Kapp were estimated for each group. The adjusted coefficient of determination (R2) was calculated to measure goodness-of-fit. Receiver operating characteristic curve analysis was performed to evaluate the diagnostic performance of the parameters. RESULTS: All parameters except D* showed significant differences between cancerous tissues and benign PZs in each group. The area under the curve values (AUCs) of ADC were comparable in groups C and D (p = 0.980) and were significantly higher than those in groups A and B (p< 0.05 for all). The AUCs of ADC and Kapp in groups B and C were similar (p = 0.07 and p = 0.954), and were significantly higher than the other parameters (p< 0.001 for all). The AUCs of ADC in group D was slightly higher than Kapp (p = 0.002), and both were significantly higher than the other parameters (p< 0.001 for all). CONCLUSIONS: ADC derived from conventional mono-exponential high b-value (3200 s/mm2) models is an optimal parameter for PZ PCa detection.

Searching for prostate cancer by fully automated magnetic resonance imaging classification: deep learning versus non-deep learning.

Prostate cancer (PCa) is a major cause of death since ancient time documented in Egyptian Ptolemaic mummy imaging. PCa detection is critical to personalized medicine and varies considerably under an MRI scan. 172 patients with 2,602 morphologic images (axial 2D T2-weighted imaging) of the prostate were obtained. A deep learning with deep convolutional neural network (DCNN) and a non-deep learning with SIFT image feature and bag-of-word (BoW), a representative method for image recognition and analysis, were used to distinguish pathologically confirmed PCa patients from prostate benign conditions (BCs) patients with prostatitis or prostate benign hyperplasia (BPH). In fully automated detection of PCa patients, deep learning had a statistically higher area under the receiver operating characteristics curve (AUC) than non-deep learning (P = 0.0007 < 0.001). The AUCs were 0.84 (95% CI 0.78-0.89) for deep learning method and 0.70 (95% CI 0.63-0.77) for non-deep learning method, respectively. Our results suggest that deep learning with DCNN is superior to non-deep learning with SIFT image feature and BoW model for fully automated PCa patients differentiation from prostate BCs patients. Our deep learning method is extensible to image modalities such as MR imaging, CT and PET of other organs.