Managing multicentre clinical trials with open source.

BACKGROUND: Multicentre clinical trials are challenged by high administrative burden, data management pitfalls and costs. This leads to a reduced enthusiasm and commitment of the physicians involved and thus to a reluctance in conducting multicentre clinical trials. OBJECTIVE: The purpose of this study was to develop a web-based open source platform to support a multi-centre clinical trial. METHODS: We developed on Drupal, an open source software distributed under the terms of the General Public License, a web-based, multi-centre clinical trial management system with the design science research approach. RESULTS: This system was evaluated by user-testing and well supported several completed and on-going clinical trials and is available for free download. CONCLUSION: Open source clinical trial management systems are capable in supporting multi-centre clinical trials by enhancing efficiency, quality of data management and collaboration.

Liver: segment-specific analysis of B1 field homogeneity at 3.0-T MR imaging with single-source versus dual-source parallel radiofrequency excitation.

PURPOSE: To measure B1 field distribution in different liver segments with and without dual transmission and to quantify the contrast-to-noise ratio (CNR) between normal liver tissue and segmental venous vessels on standard clinical 3.0-T liver magnetic resonance (MR) images. MATERIALS AND METHODS: This prospective study was approved by the local ethics committee. All subjects gave written informed consent. Six patients with liver lesions and nine healthy volunteers were included. Average hepatic B1 field values in all Couinaud liver segments were assessed by using actual flip-angle imaging (first and second repetition times msec/echo time msec: 72, 192/2.2; transmission angle: 60°) for both single and dual transmission in a 3.0-T MR imaging unit that allowed both transmission modes. Additionally, two-dimensional T1-weighted gradient-echo (repetition time msec/echo time msec, 180/2.3; transmission angle, 55°) and T2-weighted single-shot fast spin-echo images (1501/80) were acquired. Average CNR between liver parenchyma and segmental veins were measured in each segment. Two-sided paired Student t tests were used for statistical evaluation. Two blinded radiologists independently identified lesions in images from acquisitions in both transmission modes. RESULTS: Mean flip angles achieved with conventional single transmission were 44%-53% of the nominal value in segments II-IV and 67% and 63% of the nominal value in segments VI and VII, respectively, and were less than 77% in all segments. Mean actual flip angles measured for dual transmission were between 82% and 100% of the nominal value in all segments. T1-weighted single-transmission images exhibited areas of low B1 field strength with reduced image contrast. T2-weighted single-transmission images displayed significantly reduced signal intensity but nearly unchanged contrast weighting in these areas. On T1-weighted dual-transmission images, the two readers detected 22 and 14 additional lesions that they did not identify on the single-transmission images. On the dual-transmission T2-weighted images, they detected 11 and five additional lesions, respectively. CONCLUSION: Dual transmission can generate a B1 field with significantly improved homogeneity over all liver segments at a field strength of 3.0 T.

Flat-panel CT arthrography: feasibility study and comparison to multidetector CT arthrography.

OBJECTIVES: To show the feasibility of flat-panel computed tomography (FPCT) arthrography and quantitatively and qualitatively compare different FPCT protocols with standard multidetector computed tomography (MDCT). MATERIALS AND METHODS: First, a phantom simulating joint space with increasing iodine concentrations was scanned using a standard MDCT and 3 different FPCT protocols. Quantitative analyses were performed by measuring CT numbers of iodine dilutions, radiation dose, and image noise as well as signal-to-noise ratio and contrast-to-noise ratio. Second, FPCT arthrographies of 4 animal joint specimens were performed and analyzed qualitatively by 2 independent readers who evaluated image artifacts, image noise, overall image quality and anatomic depiction of bone, cartilage, and soft tissue. Kappa values were calculated for inter-reader agreement. Pearson's correlation coefficient (r) and Wilcoxon signed-ranks test with Bonferroni corrections for multiple comparisons were used to compare MDCT and FPCT. RESULTS: In phantoms, all CT scans showed a linear correlation between increasing iodine concentrations and mean HU values of contrast media and radiation dose, respectively (r = 0.98-0.99, P < 0.01). Dose-length product remained constant for MDCT scans. Signal-to-noise ratio for phantom water linearly decreased in all FPCT scans with increasing iodine concentrations. Contrast-to-noise ratio curves showed reduced slope at iodine concentrations higher than 75 mg/mL. FPCT arthrography after intra-articular administration of 5 to 6 mL of a 25% dilution of iopromide (Ultravist 300 mg/mL, Bayer HealthCare, Berlin, Germany) was successfully performed in all 4 animal joint specimens. Kappa values for inter-reader agreement of qualitative image analyses were 0.62 to 0.91. Image and depiction quality of 20-s FPCT scans were similar or superior compared with standard MDCT (P < 0.005). CONCLUSION: FPCT arthrography is feasible and may allow similar image quality compared with standard MDCT arthrography.

MRI: the new reference standard in quantifying hepatic steatosis?

OBJECTIVE: The purpose of this study was to assess non-invasive imaging modalities including MRI and CT and compare the quantitative amount of fat with data provided by the pathologist and a chemical lipid assay in leptin-deficient mouse livers. METHODS: A liver/fat phantom was first used to assess the accuracy of small-animal MRI and human MRI and CT, followed by correlation analysis with ob/ob mouse liver fat quantified by an accurate chemical lipid assay. Similarly, the authors compared the pathologist's quantification and the automated software quantification of fat with the lipid assay. The authors then investigated whether hepatic steatosis assessed by MRI correlates with the degree of liver injury in a model of ischaemia/reperfusion in leptin-deficient mice as well as with serious postoperative complications in patients undergoing major liver resection (NCT01234714). RESULTS: The authors designed lipid/liver mixtures at various ratios to mimic a wide range of fat liver contents. Small-animal and human MRI detected this fat with a high correlation to the actual fat contents. Mouse livers assessed by human MRI correlated best with total intrahepatic fat by chemical lipid analysis (r=0.975). Human CT, the pathologist's assessment and the automated software were less reliable (r=-0.873, 0.512 and 0.873, respectively). There was a significant correlation of the MRI fat quantification with several parameters of liver injury, and MRI data could predict mouse survival after ischaemia/reperfusion injury. In patients undergoing major liver resection, higher liver fat content was associated with more serious postoperative complications, such as liver or multiorgan failure and sepsis, necessitating admission to the intensive care unit. CONCLUSIONS: With the use of a well-defined set of biological standards, MRI can predict intrahepatic fat with high accuracy. In contrast to biopsies, this method is non-invasive, giving a representative assessment of the whole liver.

High-pitch dual-source CT coronary angiography: systolic data acquisition at high heart rates.

OBJECTIVE: To assess the effect of systolic data acquisition for electrocardiography (ECG)-triggered high-pitch computed tomography (CT) on motion artefacts of coronary arteries in patients with high heart rates (HRs). METHODS: Eighty consecutive patients (15 women, age 67 ± 14 years) with HR ≥ 70 bpm underwent CT angiography of the thoracic aorta (CTA) on 128-slice dual-source CT in ECG-triggered high-pitch acquisition mode (pitch = 3.2) set at 60% (group A, n = 40) or 30% (group B, n = 40) of the RR interval. Two blinded readers graded coronary artery image quality on a three-point scale. Radiation doses were calculated. RESULTS: Inter-observer agreement in grading image quality of the 1,154 coronary segments was good (κ = 0.62). HRs were similar in groups A and B (85 ± 13 bpm vs 85 ± 14 bpm, p not significant). Significantly fewer coronary segments with non-diagnostic image quality occurred (i.e. score 3) in group B than in group A [2.8% (16/579) vs 8.3% (48/575), p < 0.001]. Seventeen patients (42.5%) of group A and 12 patients (30.0%) of group B had at least one non-diagnostic segment. Effective radiation doses were 2.3 ± 0.3 mSv for chest CTA. CONCLUSION: A systolic acquisition window for high-pitch dual-source CTA in patients with high HRs (≥ 70 bpm) significantly improves coronary artery image quality at a low radiation dose.