ObjectivesTo compare the efficacy of two quantitative methods for discrimination between benign and malignant focal liver lesions (FLLs): apparent diffusion coefficient (ADC) values and T2 relaxation times.MethodsSeventy-three patients with 215 confirmed FLLs (115 benign, 100 malignant) underwent 1.5-T MRI with respiratory-triggered single-shot SE DWI (b = 50, 400, 800) and dual-echo T2TSE (TR = 3,000 ms; TE1 = 84 ms; TE2 = 228 ms). ADC values and T2 relaxation times of FLLs were calculated. Sensitivity, specificity and accuracy of both techniques in diagnosing malignancy were assessed.ResultsThe mean ADC value of malignant tumours (1.07 × 10−3 mm2/s) was significantly lower (P < 0.05) than that of benign lesions (1.86 × 10−3 mm2/s ); however, with the use of the optimal cut-off value of 1.25 × 10−3 mm2/s, 20 false positive (FP) and 20 false negative (FN) diagnoses of malignancy were noted, generating 79 % sensitivity, 82.6 % specificity and 80.9 % accuracy. The mean T2 relaxation time of malignant tumours (64.4 ms) was significantly lower (P < 0.05) than that of benign lesions (476.1 ms). At the threshold of 107 ms 22 FP and 1 FN diagnoses were noted; the sensitivity was 99 %, specificity 80.9 % and accuracy 89.3 %.ConclusionsQuantitative analysis of T2 relaxation times yielded significantly higher sensitivity and accuracy in diagnosing malignant liver tumour than ADC values.Key Points• Diffusion-weighted magnetic resonance imaging is increasingly used for liver lesions.• But ADC values demonstrated only moderate accuracy for differentiation of liver lesions.• T2 relaxation times yielded higher accuracy in diagnosing malignant liver tumours.• Both ADC and T2 values overlapped between focal nodular hyperplasia and malignant lesions.• Nevertheless T2 liver mapping could be valuable for evaluating focal liver lesions.

SummaryBackgroundDiaphragmatic injuries occur in 0.8–8% of patients with blunt trauma. The clinical diagnosis of diaphragmatic rupture is difficult and may be overshadowed by associated injuries.Diaphragmatic rupture does not resolve spontaneously and may cause life-threatening complications.The aim of this study was to present radiological findings in patients with diaphragmatic injury.Material/MethodsThe analysis of computed tomography examinations performed between 2007 and 2012 revealed 200 patients after blunt thoraco-abdominal trauma. Diaphragmatic rupture was diagnosed in 13 patients. Twelve of these patients had suffered traumatic injuries and underwent a surgical procedure that confirmed the rupture of the diaphragm. Most of diaphragmatic ruptures were left-sided (10) while only 2 of them were right-sided. In addition to those 12 patients there, another patient was admitted to the emergency department with left-sided abdominal and chest pain. That patient had undergone a blunt thoracoabdominal trauma 5 years earlier and complained of recurring pain. During surgery there was only partial relaxation of the diaphragm, without rupture.The most important signs of the diaphragmatic rupture in computed tomography include: segmental discontinuity of the diaphragm with herniation through the rupture, dependent viscera sign, collar sign and other signs (sinus cut-off sign, hump sign, band sign).ResultsIn our study blunt diaphragmatic rupture occurred in 6% of cases as confirmed intraoperatively.In all patients, coronal and sagittal reformatted images showed herniation through the diaphragmatic rupture. In left-sided ruptures, herniation was accompanied by segmental discontinuity of the diaphragm and collar sign. In right-sided ruptures, predominance of hump sign and band sign was observed.Other signs were less common.ConclusionsThe knowledge of the CT findings suggesting diaphragmatic rupture improves the detection of injuries in thoraco-abdominal trauma patients.

PurposeTo optimise the intravoxel incoherent motion (IVIM) imaging of the liver on a 3.0T scanner by assessing parameter reproducibility on free-breathing (FB) and respiratory-triggered (RT) sequences acquired with different numbers of signal averages (NSA).Material and methodsIn this prospective study 20 subjects (M/F: 10/10; age: 25-62 years, mean: 39 years) underwent IVIM magnetic resonance imaging (MRI) on a 3.0T scanner using an 18-channel phase-arrayed coil and four different echo-planar sequences, each with 10 b values: 0, 10, 30, 50, 75, 100, 150, 200, 500, and 900 s/mm2. Images were acquired with FB and RT with NSA = 1-4 (FBNSA1-4, RTNSA1-4) and with NSA = 3-6 (FBNSA3-6, RTNSA3-6). Subsequently, for the assessment of reproducibility of IVIM-derived parameters (f, D, D*), each subject was scanned again with an identical protocol during the same session. IVIM parameters were calculated. The distribution of IVIM-parameters for each DWI sequence were given as the median value with first and third quartile. Inter-scan reproducibility for each IVIM parameter was evaluated using coefficient of variance and Bland-Altman difference. Differences between FB sequence and RT sequence were tested using non-parametric Wilcoxon signed-rank test.ResultsMean coefficient of variance (%) for f, D, and D* ranged from 60 to 64, from 58 to 84, and from 82 to 99 for FBNSA1-4 sequence; from 50 to 69, from 41 to 97, and from 80 to 82 for RTNSA1-4 sequence; from 22 to 27, 15, and from 70 to 80 for FBNSA3-6 sequence; and from 21 to 32, from 12 to, and from 50 to 80 for RTNSA3-6 sequence, respectively.ConclusionsIncreasing the number of signal averages for IVIM acquisitions allows us to improve the reproducibility of IVIM-derived parameters. The sequence acquired during free-breathing with NSA = 3-6 was optimal in terms of reproducibility and acquisition time.

No abstract

No abstract