SS 1001 - New techniques in abdominal imaging
Purpose: To prospectively compare two different approaches for calculating the lean body weight (LBW) and the amount of intravenous contrast media (CM) for MDCT of the abdomen in obese patients.
Methods and Materials: Nineteen patients (9 men, 10 women) with a BMI greater than 35 kg/m2 were included in this prospective study and underwent MDCT of the abdomen. The amount of CM injected was computed according to the patient’s LBW which was estimated using either the Boer formula (Group A) or the James formula (Group B). The following variables were compared and analysed between the two groups: patient’s characteristics, CM volume and iodine dose. CNR of liver, kidney, portal vein, aorta and pancreas was also calculated.
Results: Patients in Group A has a superior BMI than Group B (38.56±6.31 vs. 37.23±5.54 kg/m2) while an inferior LBW was observed for Group A vs Group B (59.15±12.23 vs 59.79±12.68 kg). Group A provided greater CNR (liver 3.64±1.12; kidney 13.53±7.2; portal vein 10.22±6.34; aorta 11.93±3.45; pancreas 2.51±0.98) during the portal venous phase compared to CNR of Group B (liver 3.38±2.11; kidney 9.87±6.55; portal vein 7.70±3.43; aorta 8.07±2.69; pancreas 2.35±1.12). A significant difference was observed for CNR of kidney, portal vein and aorta (p<0.05) although a lower amount of CM was administered in Group A compared to Group B (115.45±13.3 ml vs 116.64±12.56, p>0.05).
Conclusion: The optimisation of contrast media volume using Boer formula significantly improves parenchymal enhancement in obese patients although a lower amount of contrast media was administered.
Purpose: To assess the robustness of accelerated and optimised diffusion-weighted imaging (DWI) in routine abdominal MRI using the simultaneous-multislice (SMS) technique for scan time reduction and a 3D Diagonal diffusion mode to improve image quality.
Methods and Materials: A total of 152 patients (88 male, 64 female; mean age 62 years) were included in this study. All received the institutional standard abdominal MRI protocol including an optimized diffusion-weighted sequence (DWIOPT: TR/TE 3100/56ms; b-values 50, 400 and 800s/mm²; diffusion mode 3D Diagonal; SMS factor 2; scan time 1:44min). A subgroup of 41 patients (mean age 62 years) additionally received standard DWI (DWISTD: TR/TE 5100/60ms; b-values 50, 400 and 800s/mm²; diffusion mode 4-scan-Trace; scan time 2:35min) as reference. Image quality criteria, lesion conspicuity (rated on a 5-point Likert-scale; 5=excellent) and ADC-values (measured in a region-of-interest-analysis) in DWISTD and DWIOPT were compared interindividually using dedicated statistics. P-values <.05 were considered significant.
Results: Interindividually DWIOPT proved superior to DWISTD in comparison of overall image quality (DWIOPT: 4.6 ±0.7; DWISTD: 4.2 ±0.8; p=.025); lesion conspicuity was comparable (p=.461); ADC-values showed no statistically significant difference (right hepatic lobe p=.107; kidney p=.098). The patients receiving DWIOPT only showed very high overall image quality and lesion conspicuity (4.7±0.6 and 4.7±0.6, respectively).
Conclusion: Integrating the SMS technique- and a 3D-Diagonal diffusion mode abdominal DWI can be accelerated and optimised with superior image quality and similar ADC-values compared to standard DWI. This approach proved stable in a larger patient cohort, so that we can recommend it for clinical routine application.
Diagnostic value and radiation dose reduction of model-based iterative reconstruction compared with hybrid iterative reconstruction in routinely upper abdominal CT study
Purpose: To evaluate dose reduction and image quality of upper abdominal CT images reconstructed with model-based iterative reconstruction (IMR) compared with hybrid iterative reconstruction (iDose4).
Methods and Materials: We prospectively enrolled 72 patients who underwent to upper abdominal CT scan;36 patients were examined using a low-kV setting (100kV) while the other 36 patients were investigated using standard kV setting (120kV), on the same 256-row CT. In the first group (Study Group) images were reconstructed using IMR algorithm while in the second group (Control Group) we used iDose4 technique. CTDIvol, DLP and Estimate Doses (ED) were evaluated. Region of interests were drawn in three different points in the liver and spleen; signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated. A four-point scale was used to subjectively evaluate the imageS in both groups.
Results: Mean CTDI Volume was 9,2 ± 1,2 mGy and 17,7 ± 2,5 mGy for low-kV IMR CT scans and standard dose iDose4 CT scans, respectively. Compared to iDose4-CT, low kV IMR CT showed significant lower DLP (570,5 ± 8,3 mGy x cm) and ED (6,3 ± 1,8 mSv).Low kV IMR images yielded higher attenuation values (HU) and higher SNR and CNR in the liver and spleen, compared with standard dose iDose4 images.The subjective image quality of both groups was comparable.
Conclusion: Low-kV MBIR technique allows significant reduction of radiation doses exposure maintaining high image quality (i.e. attenuation values, SNR and CNR in the liver and in the spleen) at routine abdominal CT, compared with standard-dose HIR technique.
Contrast dose variability depending on morphometric values: a retrospective analysis on patients undergoing multi-phase abdominal CT
Purpose: The dose of contrast material for multi-phase abdominal CT is usually based on patient body weight (BW). Inter-patient morphometric variables, such as muscular and adipose tissues or blood pool extent, are implicated in variations of parenchymal contrast enhancement (CE). We assessed their impact on liver CE.
Methods and Materials: We retrospectively evaluated 200 abdominal CT exams in 106 males and 94 females, intravenously injected with Iopamidol (370mg/ml). Body mass index (BMI), lean body weight (LBW) and body surface area (BSA) were estimated using validated formulas. Liver CE, defined as the absolute difference in HU between portal and basal phases, was measured. Data were reported as mean±standard deviation; bivariate correlation and multivariate regression analyses were performed.
Results: Patient age was 66±13 years, with a BW of 72±15kg and a LBW of 53±11kg. The contrast dose was 1.2±0.2 ml/kg of BW or 1.7±0.2 ml/kg of LBW; the liver CE resulted 43±9 HU. A negative correlation was found between BW and dose (r=-0.686, p<0.001). Liver CE weighted on iodine grams positively correlated with the BW (r=0.704, p<0.001), LBW (r=0.665, p<0.001) and BSA (r=0.712, p<0.001). At multivariate regression analysis, the contrast dose per kg of BW was the only independent predictor of contrast enhancement variability, with a standardised correlation coefficient of 0.268 (p=0.002).
Conclusion: We found a high variability of administered dose due to radiologist’s adjustment based on patient morphometry and comorbidities. To reduce liver CE variability, further studies are needed to find out a more patient-centered way to calculate contrast dose.
Purpose: To investigate the feasibility of material suppressed iodine (MSI) images derived from spectral CT to replace true noncontrast-enhanced (TNC) images and assess the effect of contrast agent dose and body mass index (BMI) on MSI images.
Methods and Materials: Sixty patients who underwent spectral CT were divided into two scan protocols based on BMI (n = 30 each; protocol A, 80 kVp/300 mgI/kg, 18-23.9kg/m2; protocol B, 100 kVp/400 mgI/kg, 24-29kg/m2). TNC images (A1 and B1) were obtained while MSI images with adaptive statistical iterative reconstruction (ASIR) were reconstructed during the arterial phase (A2 and B2) and venous phase (A3 and B3). Image noise, CT value and contrast-to-noise-ratio (CNR) of liver, aorta and portal vein were compared between MSI images and TNC images and between protocols A and B.
Results: Intra-group: image noise in A3> A2>A1 and B3>B2, in B1 was similar to those of B2 and B3. The CT values of liver in A1 were similar to A3 but higher than A2, in B1>B3>B2. The CT values of vessels in MSI images were similar to TNC images except B2> B1 in aorta and A3>A1 in portal vein.The CNR values of liver and vessels were similar except A1>A2 and A1>A3 in aorta and portal vein. Inter-group: compared with protocol B, protocol A showed higher image noise but lower CT values of vessels.
Conclusion: The image quality of MSI images might be able to replace TNC images; the increase of contrast material dose and BMI might improve the image quality of MSI images.
Feasibility of material suppressed iodine using dual-energy spectral CT with optimised adaptive statistical iterative reconstruction (ASIR) in abdominal CT
Purpose: To assess the image quality of material suppressed iodine (MSI) images derived from spectral CT imaging with different percentages of ASIR.
Methods and Materials: Forty-six patients underwent conventional non-contrast-enhanced (TNC) and arterial-phase (AP) and venous phase (VP) with CT spectral imaging mode. MSI images generated from spectral CT were reconstructed with different percentages of ASIR (0%, 30%,50%, 70%) during the arterial phase (B1, C1, D1, E1) and venous phase (B2, C2, D2, E2). TNC images (A) were reconstructed with 0% ASIR and compared with MSI images in image noise, CT value and contrast-to-noise ratio (CNR) of liver, aorta and portal vein. Volume CT dose index (CTDIvol) was recorded.
Results: Image noise in E1<D1<C1<B1<A, C2<D2<B2<E2<A (all P values <0.05). The CT values of the liver in B1<C1<D1<E1<A (all P values <0.05), C2<B2<D2<E2<A (all P values> 0.05). The CT values of the aorta and portal vein of MSI images were similar to those of TNC images, and the CT value of MSI with 70% ASIR was close to the TNC images. There were significant differences in CNR of liver: E1>B1>C1> D1>A, C2>E2>D2>B2>A. There was no significant difference between MSI images and TNC images in CNR of aorta and portal vein (all P values >0.05).The CTDIvol could be reduced by 9.92mGy (34%) without TNE scan.
Conclusion: MSI images with 70% ASIR might replace TNC images and could reduce 34% radiation dose.
Fast abdominal imaging with high parallel-imaging factors: comparative study of a 60-channel receiver coil with the standard coil setup
Purpose: To assess a novel 60-channel receiver coil setup regarding signal-to-noise ratio (SNR) and image quality in fast abdominal imaging as compared to a standard 30-channel coil setup.
Methods and Materials: All imaging data were acquired on a 3T MR scanner (Magnetom Skyra, Siemens Healthineers). 3D-T1-GRE imaging with different PAT factors (CAIPIRINHA: none,2,3,2x2,3x2,2x3,3x3) was performed in a water-phantom and in 5 volunteers using a novel 60-channel (30-channel anterior+posterior) receiver coil setup and standard 30-channel (18-channel-body and 12-channel-spine coil)-setup. SNR was measured on phantom images. Image quality assessed by two radiologists (5-point Likert scale; 5=excellent quality). In a further step, standard contrast-enhanced abdominal 3D-T1-GRE imaging with PAT 2 and accelerated imaging with PAT 2x2 and 3x2 was performed in 17 patients who were assigned into two groups (60-channel, n=8; 30-channel, n=9). Image quality was compared between the two patient groups.
Results: In comparison to the 30-channel coil, SNR gain for the 60-channel setup was observed for all acquisitions except for PAT=none. Relative SNR gain for the 60-channel coil ranged between 2.21% for PAT=2 and 22.69% for PAT=3x2. Comparison of image quality in healthy participants revealed significantly higher ratings of the 60-channel setup for all acquisitions (p≤.046), except for PAT=none (p≥.180). In patients, similar results were observed with comparable image quality at lower PAT factors (PAT=2; p≥.069) and significantly superior image quality of the 60-channel setup for higher PAT factors (2x2 and 3x2; p≤.036).
Conclusion: The 60-channel coil setup improves image quality and, therefore, enables fast abdominal imaging at high PAT factors with diagnostic image quality.
Assessment of bolus tracking for the acquisition of a single dual-energy iodine map as a quantitative imaging biomarker replacing abdominal CT perfusion
Purpose: To quantitatively evaluate dual-energy (DE) computed tomography (CT) iodine concentration maps acquired using bolus tracking with different trigger delays as a dose-reduced replacement of CT-perfusion maps for diagnosis and treatment response assessment of abdominal tumours, such as pancreatic carcinoma.
Methods and Materials: For 22 patients with pancreatic carcinoma, DECT perfusion sequences were dynamically acquired using dual-source DECT at tube potentials of 80kVp and 140kVp with tin filtration. After deformable motion-correction, perfusion maps were calculated from 80kVp image series and DECT iodine maps were calculated for each of the 34 DECT acquisitions per patient. Measurements of perfusion and iodine concentration were performed in regions of interest in healthy pancreatic tissue and carcinoma. For each patient, 34 iodine maps at different acquisition delays were evaluated with regard to correlation to perfusion and inter-group differences between healthy tissue and carcinoma, simulating acquisitions triggered by bolus tracking by using the timing of bolus tracking of prior three-phase contrast-enhanced CT as a reference (threshold 120HU in abdominal aorta).
Results: Average perfusion measured in healthy pancreatic tissue and carcinoma was 87.6±28.4ml/100ml/min and 38.6±22.2ml/100ml/min, respectively. Correlation between iodine concentrations and perfusion was statistically significant for bolus tracking with trigger delay greater than 0s (rmax=0.89). Differences in iodine concentrations between healthy pancreatic tissue and carcinoma were statistically significant for DECT acquisitions corresponding to trigger delay of 15-21s.
Conclusion: DECT iodine maps acquired using bolus tracking with appropriate trigger delay could serve as a quantitative imaging biomarker in the abdomen, providing a dose-efficient replacement of CT-perfusion.
Image quality comparison of low-dose contrast-enhanced abdominal CT with different protocols: reduced tube voltage or reduced tube current?
Purpose: To determine a better optimised scan protocol between reduced tube voltage and reduced tube current for low-dose-enhanced abdominal CT.
Methods and Materials: 41 patients who were randomly assigned into 2 groups underwent enhanced abdominal CT. The scan protocols were 100kV/150mAs with ATCM for group 1, and 120kV/100mAs with ATCM for group 2. Images were reconstructed with filtered back projection (FBP), hybrid iterative reconstruction (HIR) and iterative model reconstruction (IMR) algorithms, respectively. Images acquired in portal vein phase were used for analysis. CT attenuation, image noise and CNR of liver parenchyma were measured. Subjective image quality assessments were performed according to the features of structure delineation, noise, artefacts, image distortion, and diagnostic confidence.
Results: No difference was found in mean CTDIvol between two groups. Image noise was higher in group 1 than group 2 in FBP and HIR images; however, no difference of image noise was observed in IMR images. No difference was found in CNR between groups. FBP images in both groups failed to reach acceptable subjective image quality, although images in group 2 exhibited better scores than group 1. HIR images in both groups reached acceptable image quality with no difference between each other. IMR images in both groups reached excellent image quality with better score in group 1 than group 2.
Conclusion: Both reduced tube voltage and tube current protocols can be used in low-dose abdominal CT with HIR and IMR. Reduced tube voltage protocol tends to exhibit better image quality when using IMR.
Purpose: To correlate spleen stiffness (SS) measured by ARFI (acoustic radiation force impulse) with upper GI endoscopy for the presence of esophageal varices (EV) in patients of cirrhosis.
Methods and Materials: Our study included 78 patients of cirrhosis (66 males, 12 females). We performed 5 measurements in each and a median value was obtained. Patients were examined by endoscopy and grade of EV was noted.
Results: A cut-off value of 3.04 m/s for spleen stiffness by ARFI showed sensitivity of 78.2%, specificity of 78.3%, PPV of 89.5% and NPV of 60% for differentiating those with and without EV. SS in patients with grade 3 EV was significantly higher than those with grade 1 EV. Spleen size in patients with EV was significantly higher than those without EV. Within the group of patients with EV, those with higher grades of EV had higher values of spleen size.
Conclusion: 1) SS is significantly higher in patients of cirrhosis than in normal individuals. 2) In patients of liver cirrhosis, SS by ARFI can be used to predict the presence of EV; and endoscopy being an invasive procedure can be avoided in rest of the patients who do not have EV. 3) SS in patients with grade 3 EV was significantly higher than those with grade 1 EV. Thus it could predict the severity of EV.
Purpose: To assess the important CT findings and clinical features for differentiating malignant from benign focal splenic lesions.
Methods and Materials: From 2003 to 2015, among 673 patients with splenectomy in two institutes, we included 114 patients with focal splenic lesion who underwent preoperative CT scans (malignant=66, benign=48). Two radiologists retrospectively accessed CT findings including size of spleen, multiplicity, nature (mainly cyst vs mainly solid), margin (well- vs ill-defined), presence of wall, calcification, and enhancement pattern. They also graded the possibility of malignancy. Clinical features including presence of underlying malignancy, fever, and leukocytosis were assessed. Multivariate logistic regression analysis was performed to identify significant predictors for malignant lesion. We used receiver operating curve analysis for diagnostic performance.
Results: Statistically common features for malignant lesions included enhanced, mainly solid nature, ill-defined margin, absence of splenomegaly, absence of wall and calcification, and presence of underlying malignancy (p<0.05). In multivariate analysis, mainly cystic nature (OR, 7.428, P=0.005) and presence of underlying malignancy (OR, 15.017, P=0.001) were significant predictors for malignant lesion. Diagnostic performance for differentiating malignant lesions were 0.856 and 0.893 for each reviewers with substantial interobserver agreement (κ=0.6336). Malignant lesion (3.9+3.2cm) was smaller than benign lesion (5.9+3.3cm). Using the 4.25cm as a cut-off value, sensitivity and specificity were 66.7% and 72.7%.
Conclusion: CT and clinical features were useful for prediction of malignant splenic lesion. Mainly solid nature and presence of underlying malignancy can be helpful when differentiating malignant splenic lesion.