Personalized intrahepatic cholangiocarcinoma prognosis prediction using radiomics: Application and development trend.

Radiomics was proposed by Lambin et al. in 2012 and since then there has been an explosion of related research. There has been significant interest in developing high-throughput methods that can automatically extract a large number of quantitative image features from medical images for better diagnostic or predictive performance. There have also been numerous radiomics investigations on intrahepatic cholangiocarcinoma in recent years, but no pertinent review materials are readily available. This work discusses the modeling analysis of radiomics for the prediction of lymph node metastasis, microvascular invasion, and early recurrence of intrahepatic cholangiocarcinoma, as well as the use of deep learning. This paper briefly reviews the current status of radiomics research to provide a reference for future studies.

Preoperative prediction for pathological grade of hepatocellular carcinoma via machine learning-based radiomics.

OBJECTIVE: To investigate the efficacy of contrast-enhanced computed tomography (CECT)-based radiomics signatures for preoperative prediction of pathological grades of hepatocellular carcinoma (HCC) via machine learning. METHODS: In this single-center retrospective study, data collected from 297 consecutive subjects with HCC were allocated to training dataset (n = 237) and test dataset (n = 60). Manual segmentation of lesion sites was performed with ITK-SNAP, the radiomics features were extracted by the Pyradiomics, and radiomics signatures were synthesized using recursive feature elimination (RFE) method. The prediction models for pathological grading of HCC were established by using eXtreme Gradient Boosting (XGBoost). The performance of the models was evaluated using the AUC along with 95% confidence intervals (CIs) and standard deviation, sensitivity, specificity, and accuracy. RESULTS: The radiomics signatures were found highly efficient for machine learning to differentiate high-grade HCC from low-grade HCC. For the clinical factors, when they were merely applied to train a machine learning model, the model achieved an AUC of 0.6698, along with 95% CI and standard deviation of 0.5307-0.8089 and 0.0710, respectively (sensitivity, 0.6522; specificity, 0.4595; accuracy, 0.5333). Meanwhile, when the radiomics signatures were applied in association with clinical factors to train a machine learning model, the performance of the model remarkably increased with AUC of 0.8014, along with 95% CI and standard deviation of 0.6899-0.9129 and 0.0569, respectively (sensitivity, 0.6522; specificity, 0.7297; accuracy, 0.7000). CONCLUSIONS: The radiomics signatures could non-invasively explore the underlying association between CECT images and pathological grades of HCC. KEY POINTS: • The radiomics signatures may non-invasively explore the underlying association between CECT images and pathological grades of HCC via machine learning. • The radiomics signatures of CECT images may enhance the prediction performance of pathological grading of HCC, and further validation is required. • The features extracted from arterial phase CECT images may be more reliable than venous phase CECT images for predicting pathological grades of HCC.

Application of CT radiomics in prediction of early recurrence in hepatocellular carcinoma.

PURPOSE: To appraise the ability of the computed tomography (CT) radiomics signature for prediction of early recurrence (ER) in patients with hepatocellular carcinoma (HCC). METHODS: A set of 325 HCC patients were enrolled in this retrospective study and the whole dataset was divided into 2 cohorts, including "training set" (225 patients) and "test set" (100 patients). All patients who underwent partial hepatectomy were followed up at least within 1 year. 656 Radiomics features were extracted from arterial-phase and portal venous-phase CT images. Lasso regression model was used for data dimension reduction, feature selection, and radiomics signature building. Univariate analysis was used to identify clinical and radiomics significant features. Models (radiomics signature, clinical model, and combined model) were evaluated by area under the curve (AUC) of receiver operating characteristic curve. The models' performances for prediction of ER were assessed. RESULTS: The radiomics signature was built by 14 selected radiomics features and was significantly associated with ER (P < 0.001); the AUCs of the "train set" and the "test set" were 0.818 (95% CI 0.760-0.865) and 0.719 (95% CI 0.621-0.805), respectively. The tumor size, tumor capsule, and γ-glutamyl transferase (GGT) were significantly associated with ER in the clinical model (P < 0.05). The combined model showed incremental prognostic value, with the AUCs of "training dataset" and "test dataset" were 0.846 (95% CI 0.792-0.890) and 0.737 (95% CI 0.640-0.820), respectively. The radiomics signature, tumor size, and the level of GGT were independent predictors of ER (P < 0.05). CONCLUSIONS: The CT radiomics signature can be conveniently used to predict the ER in patient with HCC. The combined model performed better for prediction of ER than radiomics signature or clinical model.

[Quantitative analysis of hepatocellular carcinomas pathological grading in non-contrast magnetic resonance images].

In order to solve the pathological grading of hepatocellular carcinomas (HCC) which depends on biopsy or surgical pathology invasively, a quantitative analysis method based on radiomics signature was proposed for pathological grading of HCC in non-contrast magnetic resonance imaging (MRI) images. The MRI images were integrated to predict clinical outcomes using 328 radiomics features, quantifying tumour image intensity, shape and text, which are extracted from lesion by manual segmentation. Least absolute shrinkage and selection operator (LASSO) were used to select the most-predictive radiomics features for the pathological grading. A radiomics signature, a clinical model, and a combined model were built. The association between the radiomics signature and HCC grading was explored. This quantitative analysis method was validated in 170 consecutive patients (training dataset: n = 125; validation dataset, n = 45), and cross-validation with receiver operating characteristic (ROC) analysis was performed and the area under the ROC curve (AUC) was employed as the prediction metric. Through the proposed method, AUC was 0.909 in training dataset and 0.800 in validation dataset, respectively. Overall, the prediction performances by radiomics features showed statistically significant correlations with pathological grading. The results showed that radiomics signature was developed to be a significant predictor for HCC pathological grading, which may serve as a noninvasive complementary tool for clinical doctors in determining the prognosis and therapeutic strategy for HCC.

Predicting the grade of hepatocellular carcinoma based on non-contrast-enhanced MRI radiomics signature.

PURPOSE: This study was conducted in order to investigate the value of magnetic resonance imaging (MRI)-based radiomics signatures for the preoperative prediction of hepatocellular carcinoma (HCC) grade. METHODS: Data from 170 patients confirmed to have HCC by surgical pathology were divided into a training group (n = 125) and a test group (n = 45). The radiomics features of tumours based on both T1-weighted imaging (WI) and T2WI were extracted by using Matrix Laboratory (MATLAB), and radiomics signatures were generated using the least absolute shrinkage and selection operator (LASSO) logistic regression model. The predicted values of pathological HCC grades using radiomics signatures, clinical factors (including age, sex, tumour size, alpha fetoprotein (AFP) level, history of hepatitis B, hepatocirrhosis, portal vein tumour thrombosis, portal hypertension and pseudocapsule) and the combined models were assessed. RESULTS: Radiomics signatures could successfully categorise high-grade and low-grade HCC cases (p < 0.05) in both the training and test datasets. Regarding the performances of clinical factors, radiomics signatures and the combined clinical and radiomics signature (from the combined T1WI and T2WI images) models for HCC grading prediction, the areas under the curve (AUCs) were 0.600, 0.742 and 0.800 in the test datasets, respectively. Both the AFP level and radiomics signature were independent predictors of HCC grade (p < 0.05). CONCLUSIONS: Radiomics signatures may be important for discriminating high-grade and low-grade HCC cases. The combination of the radiomics signatures with clinical factors may be helpful for the preoperative prediction of HCC grade. KEY POINTS: • The radiomics signature based on non-contrast-enhanced MR images was significantly associated with the pathological grade of HCC. • The radiomics signatures based on T1WI or T2WI images performed similarly at predicting the pathological grade of HCC. • Combining the radiomics signature and clinical factors (including age, sex, tumour size, AFP level, history of hepatitis B, hepatocirrhosis, portal vein tumour thrombosis, portal hypertension and pseudocapsule) may be helpful for the preoperative prediction of HCC grade.

The impact of computed high b-value images on the diagnostic accuracy of DWI for prostate cancer: A receiver operating characteristics analysis.

To evaluate the performance of computed high b value diffusion-weighted images (DWI) in prostate cancer detection. 97 consecutive patients who had undergone multiparametric MRI of the prostate followed by biopsy were reviewed. Five radiologists independently scored 138 lesions on native high b-value images (b = 1200 s/mm2), apparent diffusion coefficient (ADC) maps, and computed high b-value images (contrast equivalent to b = 2000 s/mm2) to compare their diagnostic accuracy. Receiver operating characteristic (ROC) analysis and McNemar's test were performed to assess the relative performance of computed high b value DWI, native high b-value DWI and ADC maps. No significant difference existed in the area under the curve (AUC) for ROCs comparing B1200 (b = 1200 s/mm2) to computed B2000 (c-B2000) in 5 readers. In 4 of 5 readers c-B2000 had significantly increased sensitivity and/or decreased specificity compared to B1200 (McNemar's p < 0.05), at selected thresholds of interpretation. ADC maps were less accurate than B1200 or c-B2000 for 2 of 5 readers (P < 0.05). This study detected no consistent improvement in overall diagnostic accuracy using c-B2000, compared with B1200 images. Readers detected more cancer with c-B2000 images (increased sensitivity) but also more false positive findings (decreased specificity).

Esophageal carcinoma: Ex vivo evaluation by high-spatial-resolution T2 -mapping MRI compared with histopathological findings at 3.0T.

PURPOSE: To prospectively determine the feasibility of T2 -mapping magnetic resonance imaging (MRI) to quantitatively describe the signal characteristics of the normal esophageal wall and assess the depth of esophageal wall invasion by carcinoma at 3.0T. MATERIALS AND METHODS: Thirty-two patient specimens, each having foci of carcinoma, were studied using 3.0T MR. Freehand regions of interest were placed to measure the T2 value of the normal esophageal layers and were compared with the regions of carcinoma. Three independent readers reviewed the MR images to evaluate the depth of carcinoma invasion; when the three radiologists could not fully agree with each other, the final stage was determined by consensus. The Games-Howell test was used to compare the difference between the normal esophageal layers and carcinoma. Spearman correlation coefficient analysis was used to compare the stage at MRI with that at histopathological analysis. The interobserver agreement was compared with Cohen's kappa. The sensitivity, specificity, and accuracy for detecting carcinoma invasion were calculated. RESULTS: The T2 values between the carcinoma and normal esophageal layers were different (all P < 0.01), except for the inner circular muscle (P = 0.511). The T2 value of each layer of the normal esophageal wall was also different from that of the adjacent layer (all P < 0.01). In 29 of 32 lesions, the depth of the esophageal wall invasion determined by MR was consistent with the histopathological stage (r = 0.969, P < 0.001). The sensitivity, specificity, and accuracy were 80%, 96.3%, and 93.8%, respectively, for invasion into the mucosa; 77.8%, 95.7%, and 90.6%, respectively, for invasion into submucosa; 100%, 95.8%, and 96.9%, respectively, for invasion into muscularis propria; and 100%, 100%, and 100%, respectively, for invasion into the adventitia. CONCLUSION: T2 -mapping MR images obtained using a 3.0T MR scanner can be used to depict the precise histopathological layers of the esophageal wall clearly and provide excellent diagnostic accuracy for assessing esophageal carcinoma invasion. LEVEL OF EVIDENCE: 1 Technical Efficacy: Stage 2 J. MAGN. RESON. IMAGING 2017;45:1609-1616.

Oesophageal carcinoma: comparison of ex vivo high-resolution 3.0 T MR imaging with histopathological findings.

High-resolution magnetic resonance (MR) images clearly depict the normal oesophageal wall as consisting of eight layers, which correlates well with histopathological findings. In 56 (91.8%) of 61 lesions, the depth of oesophageal wall invasion determined through MR imaging was consistent with histopathological staging (r = 0.975, P < 0.001). The sensitivity, specificity and accuracy for the mucosa were 71.4%, 98.1%, and 95.1%, respectively, and the corresponding values for the submucosa were 82.4%, 95.5%, and 91.8%; for the muscularis propria, the sensitivity, specificity and accuracy were 100%, 95.7%, and 96.7%, respectively, and for the adventitia, these values were 100%, 100%, and 100%. The Cohen k values for interobserver agreement were excellent: K = 0.839, P < 0.001 (observer 1 vs. observer 2); K = 0.908, P < 0.001 (observer 1 vs. observer 3); and K = 0.885, P < 0.01 (observer 2 vs. observer 3). High-resolution ex vivo MR images obtained with a 3.0 T scanner can be used to precisely evaluate oesophageal carcinoma invasion and provide good diagnostic sensitivity, specificity and accuracy.

Hemodynamic safety and efficacy of ferumoxytol as an intravenous contrast agents in pediatric patients and young adults.

PURPOSE: To evaluate the safety and feasibility of off-label use of ferumoxytol as an intravenous MRI contrast agents in pediatric patients and young adults. MATERIALS AND METHODS: With HIPAA compliance and IRB approval, 86 consecutive patients who had undergone 3 T or 1.5 T MRI with ferumoxytol were retrospectively identified. The blood pressure and heart rate of patients before and after ferumoxytol injection were compared. The overall image quality was evaluated independently by two radiologists with a four-point scale. Interobserver agreement was calculated using weighted kappa statistics. RESULTS: The mean±standard deviation (SD) pre and post-contrast systolic blood pressures (SBP) were 101±18 and 95±20, respectively. There was a statistically significant difference between pre-SBP and post-SBP (P=0.003). The pre-contrast diastolic blood pressure (DBP) and the post-contrast diastolic blood pressure (DBP) were 60±14 and 51±17, respectively. There was a statistically significant difference between pre-DBP and post-DBP (P<0.001). The number of patients with SBP and DBP increase, SBP increase and DBP decrease, SBP decrease and DBP increase, SBP and DBP decrease, SBP increase and DBP unchanged were 14 (16%), 9 (10%), 6 (7%), 56 (65%), 1 (1%), respectively. There was moderate agreement on all individual assessments of image quality (kappa=0.45). Eighty-two of 86 (95.4%) studies were scored 3 or above (at least diagnostic quality) by both readers, with 90% confidence interval of 92-99%. CONCLUSION: Ferumoxytol is effective as an MR contrast agent. In our sample, there was on average a small but clinically insignificant drop in SBP and DBP post-contrast injection. Large, randomized, controlled trials are needed to establish optimal dosing, imaging procedures, and safety monitoring.

A semiflexible 64-channel receive-only phased array for pediatric body MRI at 3T.

PURPOSE: To design, construct, and validate a semiflexible 64-channel receive-only phased array for pediatric body MRI at 3T. METHODS: A 64-channel receive-only phased array was developed and constructed. The designed flexible coil can easily conform to different patient sizes with nonoverlapping coil elements in the transverse plane. It can cover a field of view of up to 44 × 28 cm(2) and removes the need for coil repositioning for body MRI patients with multiple clinical concerns. The 64-channel coil was compared with a 32-channel standard coil for signal-to-noise ratio and parallel imaging performances on different phantoms. With IRB approval and informed consent/assent, the designed coil was validated on 21 consecutive pediatric patients. RESULTS: The pediatric coil provided higher signal-to-noise ratio than the standard coil on different phantoms, with the averaged signal-to-noise ratio gain at least 23% over a depth of 7 cm along the cross-section of phantoms. It also achieved better parallel imaging performance under moderate acceleration factors. Good image quality (average score 4.6 out of 5) was achieved using the developed pediatric coil in the clinical studies. CONCLUSION: A 64-channel semiflexible receive-only phased array has been developed and validated to facilitate high quality pediatric body MRI at 3T. Magn Reson Med 76:1015-1021, 2016. © 2015 Wiley Periodicals, Inc.

Association of SPP1 promoter variants with hip osteoarthritis susceptibility in Chinese population.

Three genetic variants in the promoter of SPP1 (secreted phosphoprotein 1) gene have been reported to affect transcriptional activity of SPP1, thus conferring an increased risk for some diseases. To testify if these variants are associated with risk of hip osteoarthritis (OA) as well, we performed a case-control study including 389 hip OA patients and 315 healthy controls. Genotypes of SPP1 were determined by DNA sequencing, and differential expressions of SPP1 in relation with genotypes were evaluated by RT-PCR and ELISA. The results showed that rs17524488 (delG>insG) increased the risk of hip OA, with the adjusted OR 1.48 (95% CI 1.18-1.85, P<0.01) for risk allele insG, 1.90 (95% CI 1.35-2.66, P<0.01) for delG/insG and 2.04 (95% CI 1.20-3.49, P<0.01) for insG/insG respectively. However, as for rs11730582 (T>C), the adjusted ORs were 1.18 (95% CI 0.94-1.49, P=0.148) for allele C, 1.26 (95% CI 0.90-1.75, P=0.158) for TC, and 1.31 (95% CI 0.77-2.24, P=0.293) for CC, indicating no association of rs11730582 with hip OA risk. The variant rs28357094 was not observed in the tested subjects. Furthermore, the delG/insG and insG/insG genotypes of rs17524488 both correlated with higher levels of SPP1 expression in articular cartilage (P<0.01 for all comparisons) as well as in in synovial fluid (P<0.01 for all comparisons) compared with delG/delG, while rs11730582 had no effect on the SPP1 expression (P>0.05 for all comparisons). These results collectively indicate that the genetic variant rs17524488 in SPP1 promoter confers high risk for hip OA in a Chinese population, possibly through enhancing SPP1 expression.

X-ray dose reduction in abdominal computed tomography using advanced iterative reconstruction algorithms.

OBJECTIVE: This work aims to explore the effects of adaptive statistical iterative reconstruction (ASiR) and model-based iterative reconstruction (MBIR) algorithms in reducing computed tomography (CT) radiation dosages in abdominal imaging. METHODS: CT scans on a standard male phantom were performed at different tube currents. Images at the different tube currents were reconstructed with the filtered back-projection (FBP), 50% ASiR and MBIR algorithms and compared. The CT value, image noise and contrast-to-noise ratios (CNRs) of the reconstructed abdominal images were measured. Volumetric CT dose indexes (CTDIvol) were recorded. RESULTS: At different tube currents, 50% ASiR and MBIR significantly reduced image noise and increased the CNR when compared with FBP. The minimal tube current values required by FBP, 50% ASiR, and MBIR to achieve acceptable image quality using this phantom were 200, 140, and 80 mA, respectively. At the identical image quality, 50% ASiR and MBIR reduced the radiation dose by 35.9% and 59.9% respectively when compared with FBP. CONCLUSIONS: Advanced iterative reconstruction techniques are able to reduce image noise and increase image CNRs. Compared with FBP, 50% ASiR and MBIR reduced radiation doses by 35.9% and 59.9%, respectively.

Analysis on the imaging features of AIDS with pulmonary fungal infection.

BACKGROUND: Pulmonary fungal infection is one type of the common opportunistic infections in AIDS patients. The disease is hard to diagnose because of its complicated imaging features. The objective of this study was to investigate the imaging performance characteristics of pulmonary fungal infection in AIDS patients. METHODS: Fifty-one patients with AIDS complicated with pulmonary fungal infection and 56 patients of non-AIDS with pulmonary fungal infection were examined by CT scans and high-resolution CT scans. The contrast enhanced scans were performed in patients with the mass or suspected enlarged mediastinal lymph nodes. Results were compared between the two groups. RESULTS: The most common fungal infection in the two groups of patients was Candida albicans. The infection rates were 54.8% (28 cases) in the group (AIDS patients with pulmonary fungal infection) and 58.3% (32 cases) in another group (non-AIDS patients with pulmonary fungal infection). In the two groups, the difference in diffuse distribution and the difference in incidence of affected upper and lower lobes in the bilateral lung fields were statistically significant. The differences in patchy or large consolidation shadow, cavitas, enlarged lymph nodes in mediastinum and pleural effusion were also significant when comparing the two groups. CONCLUSIONS: The lesion in most of AIDS patients with pulmonary fungal infection tends to exhibit diffuse distribution, patchy or large consolidation shadow covering a more extensive region. The differences between AIDS with pulmonary fungal infection and non-AIDS with pulmonary fungal infection are statistically significant in lesion location and complicated imaging features. The most common fungal infection in AIDS patients is Candida albicans.