Artificial intelligence in a prediction model for postendoscopic retrograde cholangiopancreatography pancreatitis.

OBJECTIVES: In this study we aimed to develop an artificial intelligence-based model for predicting postendoscopic retrograde cholangiopancreatography (ERCP) pancreatitis (PEP). METHODS: We retrospectively reviewed ERCP patients at Nagoya University Hospital (NUH) and Toyota Memorial Hospital (TMH). We constructed two prediction models, a random forest (RF), one of the machine-learning algorithms, and a logistic regression (LR) model. First, we selected features of each model from 40 possible features. Then the models were trained and validated using three fold cross-validation in the NUH cohort and tested in the TMH cohort. The area under the receiver operating characteristic curve (AUROC) was used to assess model performance. Finally, using the output parameters of the RF model, we classified the patients into low-, medium-, and high-risk groups. RESULTS: A total of 615 patients at NUH and 544 patients at TMH were enrolled. Ten features were selected for the RF model, including albumin, creatinine, biliary tract cancer, pancreatic cancer, bile duct stone, total procedure time, pancreatic duct injection, pancreatic guidewire-assisted technique without a pancreatic stent, intraductal ultrasonography, and bile duct biopsy. In the three fold cross-validation, the RF model showed better predictive ability than the LR model (AUROC 0.821 vs. 0.660). In the test, the RF model also showed better performance (AUROC 0.770 vs. 0.663, P = 0.002). Based on the RF model, we classified the patients according to the incidence of PEP (2.9%, 10.0%, and 23.9%). CONCLUSION: We developed an RF model. Machine-learning algorithms could be powerful tools to develop accurate prediction models.

In-Hospital Cancer Mortality Prediction by Multimodal Learning of Non-English Clinical Texts.

Predicting important outcomes in patients with complex medical conditions using multimodal electronic medical records remains challenge. We trained a machine learning model to predict the inpatient prognosis of cancer patients using EMR data with Japanese clinical text records, which has been considered difficult due to its high context. We confirmed high accuracy of the mortality prediction model using clinical text in addition to other clinical data, suggesting applicability of this method to cancer.

Prognostic Factors for Covid-19 on Admission Profile and Air Pollutants.

It has been reported that the severity and lethality of Covid-19 are associated with coexisting underlying diseases (hypertension, diabetes, etc.) and cardiovascular diseases (coronary artery disease, atrial fibrillation, heart failure, etc.) that increase with age, but environmental exposure such as air pollutants may also be a risk factor for mortality. In this study, we investigated patient characteristics at admission and prognostic factors of air pollutants in Covid-19 patients using a machine learning (random forest) prediction model. Age, Photochemical oxidant concentration one month prior to admission, and level of care required were shown to be highly important for the characteristics, while the cumulative concentrations of air pollutants SPM, NO2, and PM2.5 one year prior to admission were the most important characteristics for patients aged 65 years and older, suggesting the influence of long-term exposure.

Predicting oxygen requirements in patients with coronavirus disease 2019 using an artificial intelligence-clinician model based on local non-image data.

Background: When facing unprecedented emergencies such as the coronavirus disease 2019 (COVID-19) pandemic, a predictive artificial intelligence (AI) model with real-time customized designs can be helpful for clinical decision-making support in constantly changing environments. We created models and compared the performance of AI in collaboration with a clinician and that of AI alone to predict the need for supplemental oxygen based on local, non-image data of patients with COVID-19. Materials and methods: We enrolled 30 patients with COVID-19 who were aged >60 years on admission and not treated with oxygen therapy between December 1, 2020 and January 4, 2021 in this 50-bed, single-center retrospective cohort study. The outcome was requirement for oxygen after admission. Results: The model performance to predict the need for oxygen by AI in collaboration with a clinician was better than that by AI alone. Sodium chloride difference >33.5 emerged as a novel indicator to predict the need for oxygen in patients with COVID-19. To prevent severe COVID-19 in older patients, dehydration compensation may be considered in pre-hospitalization care. Conclusion: In clinical practice, our approach enables the building of a better predictive model with prompt clinician feedback even in new scenarios. These can be applied not only to current and future pandemic situations but also to other diseases within the healthcare system.

A comprehensible machine learning tool to differentially diagnose idiopathic pulmonary fibrosis from other chronic interstitial lung diseases.

BACKGROUND AND OBJECTIVE: Idiopathic pulmonary fibrosis (IPF) has poor prognosis, and the multidisciplinary diagnostic agreement is low. Moreover, surgical lung biopsies pose comorbidity risks. Therefore, using data from non-invasive tests usually employed to assess interstitial lung diseases (ILDs), we aimed to develop an automated algorithm combining deep learning and machine learning that would be capable of detecting and differentiating IPF from other ILDs. METHODS: We retrospectively analysed consecutive patients presenting with ILD between April 2007 and July 2017. Deep learning was used for semantic image segmentation of HRCT based on the corresponding labelled images. A diagnostic algorithm was then trained using the semantic results and non-invasive findings. Diagnostic accuracy was assessed using five-fold cross-validation. RESULTS: In total, 646,800 HRCT images and the corresponding labelled images were acquired from 1068 patients with ILD, of whom 42.7% had IPF. The average segmentation accuracy was 96.1%. The machine learning algorithm had an average diagnostic accuracy of 83.6%, with high sensitivity, specificity and kappa coefficient values (80.7%, 85.8% and 0.665, respectively). Using Cox hazard analysis, IPF diagnosed using this algorithm was a significant prognostic factor (hazard ratio, 2.593; 95% CI, 2.069-3.250; p < 0.001). Diagnostic accuracy was good even in patients with usual interstitial pneumonia patterns on HRCT and those with surgical lung biopsies. CONCLUSION: Using data from non-invasive examinations, the combined deep learning and machine learning algorithm accurately, easily and quickly diagnosed IPF in a population with various ILDs.

Smart hospital infrastructure: geomagnetic in-hospital medical worker tracking.

PURPOSE: Location visualization is essential for locating people/objects, improving efficiency, and preventing accidents. In hospitals, Wi-Fi, Bluetooth low energy (BLE) Beacon, indoor messaging system, and similar methods have generally been used for tracking, with Wi-Fi and BLE being the most common. Recently, nurses are increasingly using mobile devices, such as smartphones and tablets, while shifting. The accuracy when using Wi-Fi or BLE may be affected by interference or multipath propagation. In this research, we evaluated the positioning accuracy of geomagnetic indoor positioning in hospitals. MATERIALS AND METHODS: We compared the position measurement accuracy of a geomagnetic method alone, Wi-Fi alone, BLE beacons alone, geomagnetic plus Wi-Fi, and geomagnetic plus BLE in a general inpatient ward, using a geomagnetic positioning algorithm by GiPStech. The existing Wi-Fi infrastructure was used, and 20 additional BLE beacons were installed. Our first experiment compared these methods' accuracy for 8 test routes, while the second experiment verified a combined geomagnetic/BLE beacon method using 3 routes based on actual daily activities. RESULTS: The experimental results demonstrated that the most accurate method was geomagnetic/BLE, followed by geomagnetic/Wi-Fi, and then geomagnetic alone. DISCUSSION: The geomagnetic method's positioning accuracy varied widely, but combining it with BLE beacons reduced the average position error to approximately 1.2 m, and the positioning accuracy could be improved further. We believe this could effectively target humans (patients) where errors of up to 3 m can generally be tolerated. CONCLUSION: In conjunction with BLE beacons, geomagnetic positioning could be sufficiently effective for many in-hospital localization tasks.

Use of in-Hospital Geomagnetic Fingerprinting Localization.

Recently, visualizing location of people and things in a hospital has become an issue particularly for improving work efficiency and incident prevention. Although radio frequency waves such as Wi-Fi and Bluetooth are commonly used in indoor positioning, they have several limitations owing to their physical characteristics. We proposed in-hospital hybrid positioning technique, involving a combination of radio waves and geomagnetic fingerprinting techniques. We compared accuracy of proposed technique with that of Wi-Fi- and BLE-based techniques.

Gadoxetate disodium-enhanced MR imaging: differentiation between early-enhancing non-tumorous lesions and hypervascular hepatocellular carcinomas.

PURPOSE: To retrospectively assess imaging features that help differentiate early-enhancing non-tumorous (EN) hepatic lesions from hepatocellular carcinomas (HCCs) on gadoxetate disodium-enhanced MR imaging. MATERIALS AND METHODS: Our institutional review board approved this retrospective study. We reviewed the studies of 158 patients (92 men and 65 women; age range: 29-91; mean age: 65.6 years) with chronic liver damage, who underwent gadoxetate disodium-enhanced MR imaging at 3T MR scanner. Hypervascular lesions identified during the hepatic artery phase were selected for a study cohort. The location, shape, size (maximum diameter and maximum area), and contrast enhancement signal intensity characteristics of the lesions were evaluated, then compared between the EN and HCC lesions. RESULTS: A total of 65 EN lesions (range: 3-60mm, mean: 13.6 ± 10.6 mm) from 35 patients and 33 HCCs (range: 9-61 mm, mean: 19.3 ± 12.6 mm) from 20 patients were identified. Lesions were more frequently round or oval in shape for HCCs (n=29; 88%) than ENs (n=26; 40%) (P<0.01). Unexpectedly, some ENs (n=12; 18%) showed hypointensity on hepatocyte-phase, and 6 (50%) of them were T2 hyperintense. For lesions smaller than 2 cm (9 ENs and 21 HCCs) on hepatic arterial-phase images, the mean area of hypointensity in hepatocyte-phase (54.2 ± 33.1 mm(2)) was significantly smaller than those of the corresponding hyperintensity in hepatic arterial-phase (97.1 ± 42.0 mm(2)) for EN lesions (P=0.019), whereas no significant difference in area was found for HCCs. CONCLUSION: EN lesions may occasionally present with hypointensity during the hepatocyte-phase; presenting a diagnostic dilemma. In this situation, EN lesions may be differentiated from HCCs when a hypointense area in hepatocyte-phase is smaller than the corresponding hypervascular area in hepatic-arterial phase.

CT of the pancreas: comparison of anatomic structure depiction, image quality, and radiation exposure between 320-detector volumetric images and 64-detector helical images.

PURPOSE: To prospectively compare 320-detector volumetric and 64-detector helical computed tomographic (CT) images of the pancreas for depiction of anatomic structures, image noise, and radiation exposure. MATERIALS AND METHODS: This study was approved by the institutional review board, and written informed consent was obtained. A total of 154 patients (85 men, 69 women; age range, 26-85 years; mean age, 67 years) who underwent biphasic (arterial and pancreatic phase) contrast material-enhanced CT performed with a 320-detector scanner were randomized into two groups: the 320-detector group and the 64-detector group. Biphasic transaxial multiplanar reformatted images and volume-rendered CT angiograms were obtained. CT numbers in the abdominal aorta, pancreas, and abdominal wall fat tissue; signal-to-noise ratio (SNR); and dose-length product (DLP) were compared. In addition, image quality and focal lesion depiction (n = 35) were qualitatively determined in the two groups. Unpaired t and Mann-Whitney tests were used for quantitative and qualitative assessment, respectively. RESULTS: No significant difference in CT numbers of the abdominal aorta and pancreas was noted between the two groups. Mean DLP was 43% lower in the 320-detector group (675.4 mGy·cm) than in the 64-detector group (1187.8 mGy·cm) (P < .001). SNR of the abdominal aorta, pancreas, and abdominal wall fat on biphasic images was significantly lower in the 320-detector group than in the 64-detector group (P < .001). Image quality was acceptable in both groups and was slightly better in the 64-detector group for pancreatic phase axial images (P = .02) and arterial phase multiplanar reformatted images (P < .01). No significant difference was found in the depiction of pancreatic parenchyma, main pancreatic duct, focal pancreatic lesions, splanchnic arteries, or most of the small splanchnic arterial branches. CONCLUSION: A 320-detector CT scan facilitates fast volumetric contrast-enhanced CT of the entire pancreas with acceptable image quality, even though SNR was significantly lower at 320-detector volumetric scanning.

Acid sphingomyelinase regulates glucose and lipid metabolism in hepatocytes through AKT activation and AMP-activated protein kinase suppression.

Acid sphingomyelinase (ASM) regulates the homeostasis of sphingolipids, including ceramides and sphingosine-1-phosphate (S1P). Because sphingolipids regulate AKT activation, we investigated the role of ASM in hepatic glucose and lipid metabolism. Initially, we overexpressed ASM in the livers of wild-type and diabetic db/db mice by adenovirus vector (Ad5ASM). In these mice, glucose tolerance was improved, and glycogen and lipid accumulation in the liver were increased. Using primary cultured hepatocytes, we confirmed that ASM increased glucose uptake, glycogen deposition, and lipid accumulation through activation of AKT and glycogen synthase kinase-3ß. In addition, ASM induced up-regulation of glucose transporter 2 accompanied by suppression of AMP-activated protein kinase (AMPK) phosphorylation. Loss of sphingosine kinase-1 (SphK1) diminished ASM-mediated AKT phosphorylation, but exogenous S1P induced AKT activation in hepatocytes. In contrast, SphK1 deficiency did not affect AMPK activation. These results suggest that the SphK/S1P pathway is required for ASM-mediated AKT activation but not for AMPK inactivation. Finally, we found that treatment with high-dose glucose increased glycogen deposition and lipid accumulation in wild-type hepatocytes but not in ASM(-/-) cells. This result is consistent with glucose intolerance in ASM(-/-) mice. In conclusion, ASM modulates AKT activation and AMPK inactivation, thus regulating glucose and lipid metabolism in the liver.

Hepatic hemangioma and metastasis: differentiation with gadoxetate disodium-enhanced 3-T MRI.

OBJECTIVE: The purpose of this study was to evaluate the gadoxetate disodium-enhanced MRI findings of hepatic hemangioma and to investigate the diagnostic performance in differentiating hepatic hemangioma and metastasis. MATERIALS AND METHODS: Images of 32 hepatic hemangiomas in 25 patients and of 29 hepatic metastatic lesions in 20 patients were retrospectively reviewed. Two independent readers interpreted hepatobiliary phase images alone, dynamic extracellular phase images alone, and combined hepatobiliary and dynamic extracellular phase images. MRI findings and performance with respect to the differential diagnosis of hemangioma and metastasis were assessed. RESULTS: During the hepatic arterial phase, 11 of the 32 hemangiomas (34%) exhibited early total enhancement, and nine (28%) exhibited peripheral nodular enhancement. A bright dot sign or minimal peripheral enhancement during the late dynamic phase was observed for a small number of lesions (6% and 28%, respectively). Twenty-three of the 29 metastatic lesions (79%) exhibited ring enhancement during the hepatic arterial phase. Twenty-nine hemangiomas (91%) and all of the metastatic lesions exhibited homogeneous or heterogeneous hypointensity during the hepatobiliary phase. The sensitivity, specificity, and area under the receiver operating characteristic curve for the detection of hemangioma were 76%, 81%, and 0.87 for the hepatobiliary phase alone; 97%, 88%, and 0.97 for the dynamic extracellular phase alone; and 97%, 88%, and 0.98 for the combination. Five nodules smaller than 1 cm (four hemangiomas, one metastatic lesion) that exhibited no enhancement during the arterial phase and minimal enhancement during the late dynamic phase were not differentiated. CONCLUSION: Gadoxetate disodium-enhanced MRI was found useful for differentiating hepatic hemangiomas and metastatic lesions, especially during the dynamic extracellular phase. Only a limited number of lesions smaller than 1 cm in diameter, which exhibited minimal enhancement on late dynamic phase images, were difficult to diagnose.

Body size indexes for optimizing iodine dose for aortic and hepatic enhancement at multidetector CT: comparison of total body weight, lean body weight, and blood volume.

PURPOSE: To evaluate and compare total body weight (TBW), lean body weight (LBW), and estimated blood volume (BV) for the adjustment of the iodine dose required for contrast material-enhanced multidetector computed tomography (CT) of the aorta and liver. MATERIALS AND METHODS: Institutional review committee approval and written informed consent were obtained. One hundred twenty patients (54 men, 66 women; mean age, 64.1 years; range, 19-88 years) who underwent multidetector CT of the upper abdomen were randomized into three groups of 40 patients each: (a) TBW group (0.6 g of iodine per kilogram of TBW), (b) LBW group (0.821 g of iodine per kilogram of LBW), and (c) BV group (men, 8.6 g of iodine per liter of BV; women, 9.9 g of iodine per liter of BV). Change in CT number between unenhanced and contrast-enhanced images per gram of iodine and maximum hepatic enhancement (MHE) adjusted for iodine dose were examined for correlation with TBW, LBW, and BV by using linear regression analysis. RESULTS: In the portal venous phase, correlation coefficients for the correlation of change in CT number per gram of iodine with TBW for the aorta and liver were -0.71 and -0.79, respectively, in the TBW group; -0.80 and -0.86, respectively, in the LBW group; and -0.68 and -0.66, respectively, in the BV group. In the liver, they were marginally higher in the LBW group than in the BV group (P = .03). Adjusted MHE remained constant at 77.9 HU +/- 10.2 (standard deviation) in the LBW group with respect to TBW, but it increased in the TBW (r = 0.80, P < .001) and BV (r = 0.70, P < .001) groups as TBW increased. CONCLUSION: When LBW, rather than TBW or BV, is used, the iodine dose required to achieve consistent hepatic enhancement may be estimated more precisely and with reduced patient-to-patient variability.

Detection and grading for esophageal varices in patients with chronic liver damage: comparison of gadolinium-enhanced and unenhanced steady-state coherent MR images.

The purpose of this study was to compare observer interpreted steady-state coherent coronal images and gadolinium-enhanced axial images in terms of the detection and grading of esophageal varices. Magnetic resonance imaging (MRI) and gastrointestinal endoscopy were performed within 2 weeks in 90 patients with chronic liver damage, including 55 with untreated esophageal varices, for periodic screening purposes. Two blinded readers retrospectively reviewed T1- and T2-weighted images with gadolinium-enhanced (gadolinium image set) and steady-state coherent (coherent image set) images. Sensitivity for the detection of esophageal varices was higher (P<.001) in the gadolinium image set (76%) than in the coherent image set (35%); on the other hand, specificity was higher (P<.001) in the coherent image set (91%) than in the gadolinium image set (66%). Furthermore, area under the ROC curve was higher for the gadolinium image set (Az=0.823) than the coherent image set (Az=0.761) (P=.48). Moderate and weak positive correlations with endoscopic grades were found for the gadolinium image (r=0.48, P<.01) and coherent image sets (r=0.34, P=.018). The addition of steady-state coherent imaging to the current routine liver imaging protocol did not improve the detection or grading of esophageal varices, whereas gadolinium-enhanced imaging was found to be potentially valuable. Nevertheless, endoscopy was confirmed to be mandatory in patients with esophageal varices suspected by MRI of the liver.

Optimal acquisition delay for dynamic contrast-enhanced MRI of hypervascular hepatocellular carcinoma.

OBJECTIVE: The purpose of this study was to prospectively determine the optimal acquisition delay for imaging of hypervascular hepatocellular carcinoma with multiphasic dynamic contrast-enhanced MRI. SUBJECTS AND METHODS: One hundred twenty patients with chronic hepatic disease underwent three-phase dynamic contrast-enhanced MRI of the liver, which revealed 49 hypervascular hepatocellular carcinomas. Abdominal aortic contrast arrival time was determined with test bolus imaging. Patients were assigned to one of the following four groups according to acquisition delay determined from abdominal aortic contrast arrival time to the middle of the k-space for the early, late hepatic arterial, and portal venous phases: 0, 12, and 49 seconds (group 1); 3, 15, and 52 seconds (group 2); 6, 18, and 55 seconds (group 3); and 9, 21, and 58 seconds (group 4). Each phase of imaging took 12 seconds. Contrast enhancement in the abdominal aorta, portal vein, hepatic parenchyma, and hepatocellular carcinoma was evaluated. Peritumoral sinusoidal enhancement (i.e., coronal enhancement) also was assessed. RESULTS: Intense enhancement of hepatocellular carcinoma with little background parenchymal enhancement occurred at 9-12 seconds (p < 0.05) after arrival of contrast material in the abdominal aorta. Hepatocellular carcinoma-to-liver contrast began to decline at 15 seconds and decreased to less than zero at 49 seconds. The conspicuity of coronal enhancement was greater 21 seconds after contrast administration than in earlier phases. CONCLUSION: With the injection protocol used in this study, optimal acquisition delay-determined from abdominal aortic contrast arrival time to the middle of the k-space acquisition-for imaging of hypervascular hepatocellular carcinoma was 9-12, 21 or more, and 49 seconds for the early, late hepatic arterial, and portal venous phases.

Hepatic hemangioma: correlation of enhancement types with diffusion-weighted MR findings and apparent diffusion coefficients.

PURPOSE: To correlate hepatic hemangioma enhancement types in gadolinium-enhanced magnetic resonance (MR) images with diffusion-weighted MR findings and apparent diffusion coefficients (ADCs). MATERIALS AND METHODS: Respiratory-triggered diffusion-weighted MR images (TR/TE, 2422/46 ms; parallel imaging factor, 2; b factor, 500 s/mm(2); number of averaging, 6) obtained in 35 patients with 44 hepatic hemangiomas diagnosed by gadolinium-enhanced MR and by follow-up imaging were retrospectively evaluated. Hemangiomas were classified into three enhancement types based on gadolinium-enhanced MR imaging findings: type I, early-enhancement type; type II, peripheral nodular enhancement type; type III, delayed enhancement type. Two blinded readers qualitatively assessed lesion sizes and signal intensities on T2-weighted turbo spin-echo and diffusion-weighted images. The ADCs of hemangiomas were also measured. RESULTS: No significant difference was observed between the three enhancement types in terms of signal intensities on T2-weighted images. Signal intensities on diffusion-weighted images were lower in the order type I to III (P<.01), and mean ADCs were 2.18 x 10(-3), 1.86 x 10(-3), and 1.71 x 10(-3) mm(2)/s for types I, II, and III, respectively (P<.01). No correlation was found between lesion sizes and ADCs. CONCLUSION: Hepatic hemangiomas were found to have enhancement type dependent signal intensities and ADCs on diffusion-weighted MR images. Further studies will have to substantiate that these diffusion patterns might reflect intratumoral blood flow or perfusion.

Abdominal multidetector CT in patients with varying body fat percentages: estimation of optimal contrast material dose.

PURPOSE: To determine if contrast material dose for abdominal multidetector computed tomography (CT), as determined by using body weight (BW), overestimates the amount of contrast material required in heavier patients. MATERIALS AND METHODS: Institutional review committee approval and patients' written informed consent were obtained. CT images of the abdomen were obtained by using 2 mL per kilogram of BW of intravenous contrast material (300 mg/mL iodine) injected at 4 mL/sec in 161 consecutive patients (age range, 28-90 years; mean age, 63 years; 95 men, 66 women). CT scans were initiated 45 and 150 seconds after aortic enhancement increased by 50 HU. The patients were divided into low (37-54 kg) and high (55-75 kg) BW groups. The DeltaHU/I, where DeltaHU is change in CT number and I is iodine dose in grams, and adjusted maximum hepatic enhancement (DeltaHU/[I/kg]) were assessed for correlation with BW, body mass index (BMI), and body fat percentage (BFP) by using linear regression. RESULTS: DeltaHU/I correlated (P < .001) inversely with BW in the aorta (r = -0.78) and liver (r = -0.80) and with BMI in the aorta (r = -0.59) and liver (r = -0.61) on portal venous phase images. Regression formula for the low BW group was DeltaHU/I = 4.1 - .044 x BW (P < .001) and for the high BW group was DeltaHU/I = 2.7 - .017 x BW (P < .001), suggesting that the amount of contrast material required with increased BW is less in the high BW group. Adjusted maximum hepatic enhancement directly correlated with BFP (r = 0.25, P < .01). CONCLUSION: Excessive contrast material may inadvertently be given in heavier patients when the dose is determined by patient BW. Contrast material dose may need to be tailored in individual patients by using BFP.

Diffusion-weighted imaging of the liver: optimizing b value for the detection and characterization of benign and malignant hepatic lesions.

PURPOSE: To determine the optimal b values required for diffusion-weighted (DW) imaging of the liver in the detection and characterization of benign and malignant hepatic lesions. MATERIALS AND METHODS: MR images obtained in 76 patients including 28 malignant hepatic lesions (21 hepatocellular carcinomas and 7 metastases) and 27 benign lesions (12 hemangiomas and 15 cysts) were reviewed. DW-echo planner images (EPIs; b values with 100, 200, 400, and 800 s/mm2) were reviewed solely first, and then with T2-weighted EPIs (b=0 s/mm2). RESULTS: Sensitivity for malignant lesions (74%) was highest on DW-EPIs with b value of 100 s/mm2 and T2-weighted EPIs combined (P<0.05), and sensitivity for benign lesions (87%) was highest on DW-EPIs with b value of 800 s/mm2 and T2-weighted EPIs (P<0.05). Specificities were comparably high for all sequences. The Az values for malignant lesions were 0.94, 0.90, 0.87, and 0.89, and those for benign lesions were 0.91, 0.89, 0.87, and 0.94 on DW-EPIs with b values of 100, 200, 400, and 800 and T2-weighted EPIs combined, respectively. Hepatic cysts were clearly distinguished with the cutoff ADC value of 2.5x10(-3) mm2/s using a b value of 400 s/mm2 or greater. CONCLUSION: DW-EPIs with middle b values were not required in the detection and characterization of benign and malignant hepatic lesions.

Abdomen: angiography with 16-detector CT--comparison of image quality and radiation dose between studies with 0.625-mm and those with 1.25-mm collimation.

PURPOSE: To prospectively compare image quality and volume computed tomographic (CT) dose index (CTDI(vol)) of 16-detector CT angiograms of the abdomen acquired with 0.625-mm collimation with those of images acquired with 1.25-mm collimation. MATERIALS AND METHODS: This study had institutional review board approval, and informed consent was obtained from all patients. Dual-phase contrast material-enhanced CT was performed in 78 patients (48 men and 30 women; age, 34-91 years; mean age, 64.8 years) by using a 16-detector CT scanner. Patients were prospectively randomized into two equal-sized groups: those who underwent CT with 0.625-mm collimation and nonoverlapped reconstruction and those who underwent CT with 1.25-mm collimation and 50% overlapped reconstruction. Scan acquisition time was 7.5 seconds in both groups. CTDI(vol) was recorded. Arterial phase volume-rendered, arterial phase multiplanar reformatted, and portal venous phase multiplanar reformatted CT angiograms were generated. Qualitative assessment was performed for image quality and for depiction of splanchnic, intercostal, and lumbar arteries and veins. The unpaired t test was used for statistical comparison. RESULTS: On the arterial phase CT angiograms, there was no difference between the two collimation groups for the depiction of proximal splanchnic arteries, while the dorsal pancreatic, intercostal, and lumbar arteries and some peripheral splanchnic arterial branches were better delineated on CT scans obtained with 0.625-mm collimation than on scans obtained with 1.25-mm collimation (P < .05). Regarding the portal venous phase CT angiograms, no difference between the two groups was found in most veins, except the right adrenal vein (P = .003). Image quality was superior for 1.25-mm collimation (P < .001). CTDI(vol) values were positively correlated with patient body weight (r = 0.34, P < .001) but had no correlation with collimation size (P = .24). CONCLUSION: Scanning with 1.25-mm collimation seems adequate for a routine CT angiography examination of most arteries and veins at 16-detector CT, while scanning with 0.625-mm collimation facilitates improved delineation of fine vessels. CTDI(vol) values correlate positively with body weight but have no correlation with collimation size.

Evaluating local hepatocellular carcinoma recurrence post-transcatheter arterial chemoembolization: is diffusion-weighted MRI reliable as an indicator?

PURPOSE: To evaluate the detectability of local hepatocellular carcinoma (HCC) recurrence after transcatheter arterial chemoembolization (TACE) by diffusion-weighted MR imaging in correlation with those of gadolinium-enhanced MR imaging. MATERIALS AND METHODS: Respiratory-triggered diffusion-weighted MR images (b factor, 500 s/mm(2); number of averaging, six were obtained in 25 patients with 39 HCCs. Two independent radiologists evaluated diffusion-weighted MR images, gadolinium-enhanced MR images after TACE, and assigned confidence levels for postoperative HCC recurrence. Apparent diffusion coefficients (ADCs) in HCCs were also measured. Sensitivities and specificities were compared using an extension of the McNemar test. Observer performance was also determined by ROC curve analysis. RESULTS: Local recurrences in 14 HCCs and complete tumor necrosis in 25 HCCs after TACE were determined. Sensitivity for the detection of local HCC recurrence was higher on gadolinium-enhanced MR imaging (82%) than on diffusion-weighted MR imaging (60.7%) for the two readers in combination and separately (P < 0.05). Specificities were comparably high for both sequences. Az values were higher for gadolinium-enhanced MR images (0.92) than for diffusion-weighted MR images (0.74) for readers in combination and separately (P < 0.05). Mean ADC values showed an increase after TACE (P < 0.001). CONCLUSION: Diffusion-weighted MR imaging was not found to be a reliable predictor of local HCC recurrence after TACE as compared with gadolinium-enhanced MR imaging.

Multi-detector row CT of the kidney: optimizing scan delays for bolus tracking techniques of arterial, corticomedullary, and nephrographic phases.

PURPOSE: To determine optimal scan delays for renal arterial-, corticomedullary-, and nephrographic-phase imaging with multi-detector row computed tomography (MDCT) of the kidney using a bolus-tracking technique. METHODS AND MATERIALS: One hundred and twenty-eight patients underwent three-phase CT scan of the kidney with eight-row MDCT after receiving 2 mL/kg of 300 mgI/mL contrast medium at 4 mL/s. Patients were prospectively randomized into three groups with different scan delays for the three scan phases (arterial, corticomedullary, and nephrographic) after bolus-tracking triggered at 50 HU of aortic contrast enhancement: group 1 (5, 20, 45 s); group 2 (10, 25, 50s); and group 3 (15, 30, 55 s). Mean CT values (HU) of the abdominal aorta, renal artery, renal vein, renal cortex, and renal medulla were measured; increases in CT values pre- to post-contrast were assessed as contrast enhancement. Renal artery-to-vein and renal cortex-to-medulla contrast differences were also assessed. Qualitative analysis was also performed. RESULTS: Mean renal artery enhancement was 240-288 HU at 5-15s after the trigger and peaked at 10s (P<.001). Mean renal cortical enhancement was 195-217 HU at 10-30s and peaked at 25s (P<.01). Contrast enhancement in the renal medulla increased gradually and reached mean 145 HU at 55 s. Cortex-to-medulla contrast difference was high (110-140 HU) at 5-30s and decreased below 30 HU at 45 s after the trigger. Renal artery-to-vein contrast difference was high (121-125 HU) at 5-10s. Qualitative results correlated well with quantitative results. CONCLUSION: For the injection protocol used in this study, optimal scan delays after the bolus-tracking trigger were 5-10 s for renal arterial, 15-25 s for corticomedullary, and 50-55 s for nephrographic phases.