Detection of pancreatic neuroendocrine tumors (PNET) using semi-quantitative [68Ga]DOTATOC PET in combination with multiphase contrast-enhanced CT.

AIM: Pancreatic neuroendocrine tumors (PNETs) pose a diagnostic challenge with respect to the physiologic somatostatin receptor expression in the uncinate process representing a potential pitfall for receptor imaging with PET/CT. METHODS: We identified 49 PNETs from a total of 316 consecutive [68Ga]DOTATOC PET/CT examinations for whom the detections rates of PET and multiphase contrast enhanced (CE-) CT could be retrospectively compared and 38 PNETs for which SUVmax and SUVmax target-to-liver ratios could be calculated for the tumors and the uncinate process. RESULTS: The detection rate of PET (83.7%) was higher than of the different CT phases (arterial: 59.2%, P=0.017; portal-venous: 38.8%, P<0.001; venous: 46.9%, P=0.001; multiphase: 71.4%, P=0.286). Compared to the other method PET revealed 28.6% additional lesions and multiphase CE-CT 16.3%. The portal-venous phase revealed only lesions that were also detected in the arterial or venous phase. The detection rate for PNETs in the uncinate process (N.=9) was 66.7% for PET versus 55.6% for multiphase CE-CT. SUVmax and SUVmax target-to-liver ratios differed significantly (P<0.001) for PNETs (mean, range: SUVmax, 14.6, 1.4-69.3; SUVmax target-to-liver ratio, 3.2, 0.69-23.1) and uncinate process (4.32, 0.8-13.5; 0.94, 0.51-1.56), however with a wide overlap. CONCLUSION: Patients with PNETs should undergo [68Ga]DOTATOC PET/CT with at least an arterial and venous phase CT scan. SUVmax and SUVmax target-to-liver ratios provide additional information but do no reliably separate PNETs from normal tracer uptake in the uncinate process.

Detection of pancreatic neuroendocrine tumors (PNET) using semi-quantitative [68Ga]DOTATOC PET in combination with multiphase contrast-enhanced CT.

Aim: Pancreatic neuroendocrine tumors (PNETs) pose a diagnostic challenge with respect to the physiologic somatostatin receptor expression in the uncinate process representing a potential pitfall for receptor imaging with PET/CT. Methods: We identified 49 PNETs from a total of 316 consecutive [68Ga]DOTATOC PET/CT examinations for whom the detections rates of PET and multiphase contrast enhanced (CE-) CT could be retrospectively compared and 38 PNETs for which SUV max and SUV max target-to-liver ratios could be calculated for the tumors and the uncinate process. Results: The detection rate of PET (83.7%) was higher than of the different CT phases (arterial: 59.2%, P=0.017; portal-venous: 38.8%, P<0.001; venous: 46.9%, P=0.001; multiphase: 71.4%, P=0.286). Compared to the other method PET revealed 28.6% additional lesions and multiphase CE-CT 16.3%. The portal-venous phase revealed only lesions that were also detected in the arterial or venous phase. The detection rate for PNETs in the uncinate process (N.=9) was 66.7% for PET versus 55.6% for multiphase CE-CT. SUV max and SUV max target-to-liver ratios differed significantly (P<0.001) for PNETs (mean, range: SUV max,14.6, 1.4-69.3; SUV max target-to-liver ratio, 3.2, 0.69-23.1) and uncinate process (4.32, 0.8-13.5; 0.94, 0.51-1.56), however with a wide overlap. Conclusion: Patients with PNETs should undergo [68Ga]DOTATOC PET/CT with at least an arterial and venous phase CT scan. SUV max and SUV max target-to-liver ratios provide additional information but do no reliably separate PNETs from normal tracer uptake in the uncinate process.

MRI-guided and CT-guided cervical nerve root infiltration therapy: a cost comparison.

PURPOSE: To evaluate and compare the costs of MRI-guided and CT-guided cervical nerve root infiltration for the minimally invasive treatment of radicular neck pain. MATERIALS AND METHODS: Between September 2009 and April 2012, 22 patients (9 men, 13 women; mean age: 48.2 years) underwent MRI-guided (1.0 Tesla, Panorama HFO, Philips) single-site periradicular cervical nerve root infiltration with 40 mg triamcinolone acetonide. A further 64 patients (34 men, 30 women; mean age: 50.3 years) were treated under CT fluoroscopic guidance (Somatom Definition 64, Siemens). The mean overall costs were calculated as the sum of the prorated costs of equipment use (purchase, depreciation, maintenance, and energy costs), personnel costs and expenditure for disposables that were identified for MRI- and CT-guided procedures. Additionally, the cost of ultrasound guidance was calculated. RESULTS: The mean intervention time was 24.9 min. (range: 12 - 36 min.) for MRI-guided infiltration and 19.7 min. (range: 5 - 54 min.) for CT-guided infiltration. The average total costs per patient were EUR 240 for MRI-guided interventions and EUR 124 for CT-guided interventions. These were (MRI/CT guidance) EUR 150/60 for equipment use, EUR 46/40 for personnel, and EUR 44/25 for disposables. The mean overall cost of ultrasound guidance was EUR 76. CONCLUSION: Cervical nerve root infiltration using MRI guidance is still about twice as expensive as infiltration using CT guidance. However, since it does not involve radiation exposure for patients and personnel, MRI-guided nerve root infiltration may become a promising alternative to the CT-guided procedure, especially since a further price decrease is expected for MRI devices and MR-compatible disposables. In contrast, ultrasound remains the less expensive method for nerve root infiltration guidance.

[Costs and role of ultrasound follow-up of polytrauma patients after initial computed tomography]. / Kosten und Stellenwert von Ultraschallverlaufskontrollen bei polytraumatisierten Patienten nach initialer Computertomografie.

PURPOSE: To assess the costs and diagnostic gain of abdominal ultrasound follow-up of polytrauma patients initially examined by whole-body computed tomography (CT). MATERIALS AND METHODS: A total of 176 patients with suspected multiple trauma (126 men, 50 women; age 43.5 ± 17.4 years) were retrospectively analyzed with regard to supplementary and new findings obtained by ultrasound follow-up compared with the results of exploratory FAST (focused assessment with sonography for trauma) at admission and the findings of whole-body CT. A process model was used to document the staff, materials, and total costs of the ultrasound follow-up examinations. RESULTS: FAST yielded 26 abdominal findings (organ injury and/or free intra-abdominal fluid) in 19 patients, while the abdominal scan of whole-body CT revealed 32 findings in 25 patients. FAST had 81 % sensitivity and 100 % specificity. Follow-up ultrasound examinations revealed new findings in 2 of the 25 patients with abdominal injuries detected with initial CT. In the 151 patients without abdominal injuries in the initial CT scan, ultrasound follow-up did not yield any supplementary or new findings. The total costs of an ultrasound follow-up examination were EUR 28.93. The total costs of all follow-up ultrasound examinations performed in the study population were EUR 5658.23. CONCLUSION: Follow-up abdominal ultrasound yields only a low overall diagnostic gain in polytrauma patients in whom initial CT fails to detect any abdominal injuries but incurs high personnel expenses for radiological departments.

New insights into peripartum cardiomyopathy using cardiac magnetic resonance imaging.

PURPOSE: The aim of this study was to evaluate a comprehensive cardiac magnetic resonance (MR) imaging approach in patients with peripartum cardiomyopathy (PPCM). The focus was on inflammatory myocardial changes. MATERIALS AND METHODS: Retrospective analysis of 12 cardiac MR examinations was performed in 6 patients with PPCM. The protocol comprised cine sequences for the determination of chamber sizes and function. T 2-weighted sequences for determination of edema (T 2 ratio), T 1-weighted images for measurement of early gadolinium enhancement ratio (EGER), and late gadolinium enhancement (LGE) sequences were used for tissue characterization. 5 examinations were performed during the acute stage, and 7 examinations were performed during the course of the disease. RESULTS: Initially, 3 of 5 patients presented with an elevated left ventricular end-diastolic volume (LVEDV); in one patient, the LVEDV was in the upper range. In 4 of 5 subjects, the left ventricular ejection fraction (LVEF) was decreased. The T 2 ratio and EGER values were initially elevated in all women. No LGE was detected in initial scans. In follow-up examinations, the LVEDV decreased and the LVEF increased in all patients. Tissue-characterizing parameters decreased to normal in all but 1 patient. 2 patients showing LGE did not present a favorable clinical course. CONCLUSION: Myocardial inflammation was detected in the acute stage of PPCM, which was mostly transient. In our small group, patients showing LGE had a non-favorable clinical course. Future studies should include tissue-characterizing parameters, such as T 2 ratio and EGER. Thus, further insights into pathophysiology can be gained and therapeutic effects can be measured in a more extensive manner.

[Diagnostic value of multislice perfusion CT in dementia patients]. / Bedeutung der zerebralen Perfusionsmessung am Mehrzeilenspiral-CT in der Demenzdiagnostik.

BACKGROUND: The diagnostic value of perfusion CT in dementia patients was retrospectively evaluated in comparison to the Mini-Mental State Examination (MMSE). MATERIAL AND METHODS: The perfusion CT database of 55 subjects was investigated. All patients underwent an unenhanced, contrast-enhanced, and perfusion CT of the head. The clinical evaluation of the degree of dementia was performed with the MMSE. In the perfusion CT data, 24 ROI were marked. Then blood volume (BV), blood flow (BF), and mean transit time (MTT) were calculated and compared with the ipsilateral and contralateral regions and with the degree of dementia. RESULTS: With increasing degree of dementia, a significant decrease of the occipital and temporal BV was found. A significant decrease of the BF in the frontal lobe, basal ganglia, and occipital region was found. Concurrently, the MTT increased significantly in the basal ganglia region. The group with Alzheimer's disease showed significant regional hypoperfusion compared with the group of cognitively normal subjects in the frontal, basal ganglia, occipital, and temporal region. CONCLUSIONS: The cerebral perfusion decreased with increasing degree of dementia. The inexpensive and widely available perfusion CT reveals information about regional differences of cerebral perfusion, which may be useful in differentiating severity and types of dementia.

Magnetic resonance imaging in the assessment of cystic pancreatic lesions: differentiation of benign and malignant lesion status.

PURPOSE: To characterize cystic pancreatic lesions and tumors with magnetic resonance imaging (MRI), and to assess the value of morphological criteria in differentiating pseudocysts versus cystic tumors and benign versus malignant cystic tumors. MATERIAL AND METHODS: Twenty-three patients with cystic pancreatic tumors or lesions underwent plain and contrast-material-enhanced MRI, including magnetic resonance cholangiopancreatography (MRCP). The MR findings were characterized and analyzed by two readers, and the role of various imaging criteria and combinations thereof for final lesion assignment were assessed. Final diagnoses were obtained from the results of open surgery (n = 19) and/or biopsy (n = 4). RESULTS: Final diagnoses included cystic tumors (n = 11) and pseudocysts (n = 12). The lesions were located in the head (56%) and body or tail (44%). Lesion diameters ranged from 7 to 50 mm. Various lesion contrast enhancement patterns were observed for both benign and malignant lesions. Serous cystadenomas were located in the head, they were lobulated, and had wall diameters < or = 2 mm; with the combination of these characteristics all patients with serous cystadenoma could be identified, whereas in no other patient was this constellation observed. CONCLUSION: MRI facilitated the diagnosis of serous cystadenomas, although no definite morphologic criterion for the differentiation between pseudocysts and mucinous cystadenomas was identified. In consideration of the substantial therapeutic consequences, either diagnostics in unclear cystic pancreatic lesions should comprise cyst fluid analysis if necessary, or eligible patients should be referred for surgical resection.

Magnetic resonance imaging including magnetic resonance cholangiopancreatography for tumor localization and therapy planning in malignant hilar obstructions.

PURPOSE: To assess image quality and overall accuracy of magnetic resonance imaging (MRI), including two magnetic cholangiopancreatography (MRCP) techniques, for the diagnostics and preoperative work-up of malignant hilar obstructions. MATERIAL AND METHODS: Thirty-one patients with malignant hilar obstructions (hilar cholangiocarcinoma, n=30; hepatocellular carcinoma, n=1) received MRCP by two techniques (single-shot thick-slab and multisection thin-slice MRCP) and unenhanced and contrast material-enhanced MRI. MR assessment included the evaluation of image quality and visualization of bile ducts (5-point scale), and the classification of tumor status. MR results were subsequently correlated with the results from surgery and pathology. RESULTS: The maximum intensity projections of multisection thin-slice MRCP had significantly more artifacts compared to MRCP in the single-shot thick-slab technique, and overall image quality of single-shot thick-slab MRCP was rated significantly superior compared to multisection thin-slice MRCP (4.4 +/- 0.7 and 4.1 +/- 0.9, respectively). Moreover, ductal visualization of different parts of the biliary system was rated superior with single-shot thick-slab MRCP. In contrast, the original data from multisection thin slice MRCP facilitated visualization of periductal lesions and adjacent structures. Overall MR accuracy for the assessment of tumor status, periductal infiltration, and lymph node metastases was 90%, 87%, and 66%, respectively. CONCLUSION: For evaluation of malignant hilar obstructions, MRCP by the single-shot thick-slab technique had superior image quality and fewer artifacts; in contrast, besides sole biliary visualization, multisection MRCP depicted complementary adjacent parenchymal and periductal structures. We therefore recommend MRI, with a combination of both MRCP techniques, for the diagnostic work-up and therapy planning of malignant hilar obstructions.

[Combination of dynamic MR enteroclysis (Sellink) and MR colonography to diagnose Crohn's disease]. / Kombination von dynamischem MR-Sellink und MR-Kolonographie zur Diagnostik des Morbus Crohn.

PURPOSE: To investigate whether a combined examination with MRI enteroclysis and MRI colonography is practicable and would provide additional findings in the regions seen with ileo-colonoscopy in the work-up of patients with Crohn's disease. MATERIAL AND METHODS: Twenty-five consecutive patients with Crohn's disease (CD) (age range 19 to 42 years, mean age 29.2 years, gender ratio male:female 8:17) were retrospectively studied. All patients underwent conventional ileo-colonoscopy as a standard of reference followed by a combined examination of MRI enteroclysis and MRI colonography at 1.5T within 15 days. Two blinded radiologists evaluated the MRI examinations and compared them with the colonoscopic results. RESULTS: The MRI examination detected 31 inflamed bowel segments in 25 patients. In comparison to colonoscopy, 5 additional, endoscopically inaccessible lesions were found by MRI and only 2 lesions were missed by MRI. Compared to colonoscopy, MRI found 7 of 10 fistulas detected by colonoscopy, and 3 otherwise indiscernible abscesses. The detection of inflamed bowel segments by means of MRI (endoscopy) revealed a sensitivity of 88.8 % (100 %), specificity of 80 % (100 %) and an overall accuracy of 96 % (100 %). CONCLUSION: This study provides strong evidence that the combination of MRI enteroclysis and MRI colonography is practicable and supplies additional results regarding the regions which are not seen with ileo-colonoscopy in the work-up of patients with Crohn's disease.

Colon dissection: a new three-dimensional reconstruction tool for computed tomography colonography.

PURPOSE: To improve the sensitivity of computed tomography (CT) colonography in the detection of polyps by comparing the 3D reconstruction tool "colon dissection" and endoluminal view (virtual colonoscopy) with axial 2D reconstructions. MATERIAL AND METHODS: Forty-eight patients (22 M, 26 F, mean age 57 +/- 21) were studied after intra-anal air insufflation in the supine and prone positions using a 16-slice helical CT (16 x 0.625 mm, pitch 1.7; detector rotation time 0.5 s; 160 mAs und 120 kV) and conventional colonoscopy. Two radiologists blinded to the results of the conventional colonoscopy analyzed the 3D reconstruction in virtual-endoscopy mode, in colon-dissection mode, and axial 2D slices. RESULTS: Conventional colonoscopy revealed a total of 35 polyps in 15 patients; 33 polyps were disclosed by CT methods. Sensitivity and specificity for detecting colon polyps were 94% and 94%, respectively, when using the "colon dissection", 89% and 94% when using "virtual endoscopy", and 62% and 100% when using axial 2D reconstruction. Sensitivity in relation to the diameter of colon polyps with "colon dissection", "virtual colonoscopy", and axial 2D-slices was: polyps with a diameter >5.0 mm, 100%, 100%, and 71%, respectively; polyps with a diameter of between 3 and 4.9 mm, 92%, 85%, and 46%; and polyps with a diameter < 3 mm, 89%, 78%, and 56%. The difference between "virtual endoscopy" and "colon dissection" in diagnosing polyps up to 4.9 mm in diameter was statistically significant. CONCLUSION: 3D reconstruction software "colon dissection" improves sensitivity of CT colonography compared with the endoluminal view.

Magnetic resonance cholangiopancreatography: image quality, ductal morphology, and value of additional T2- and T1-weighted sequences for the assessment of suspected pancreatic cancer.

PURPOSE: To assess image quality and duct morphology on magnetic resonance cholangiopancreatography (MRCP) and also the value of additional T2- and T1-weighted sequences for differentiation of benignity and malignancy in patients with suspected pancreatic tumors. MATERIAL AND METHODS: One-hundred-and-fourteen patients received MRCP and unenhanced and contrast material-enhanced MR imaging. MR results were analyzed independently by two blinded readers, and subsequently correlated with the results from surgery, biopsy, and follow-up. Assessment included the evaluation of image quality, duct visualization and morphology, and the differentiation of pancreatic lesion status (benign versus malignant). RESULTS: Overall, 49 patients had benign final diagnoses, while 65 had a malignant diagnosis. Image quality of single-shot thick-slab MRCP was rated significantly better than the MIP images of multisection MRCP. With MRCP alone, the two readers' accuracy in the assessment of pancreatic lesion status was 72% (95% CI, 64% to 83%) and 69% (95% CI, 56% to 77%), respectively; with the addition of T2- and T1-weighted images the accuracy significantly improved to 89% (95% CI, 82% to 95%) and 84% (95% CI, 77% to 92%) for readers 1 and 2, respectively. CONCLUSION: Single-shot thick-slab MRCP and multisection MRCP provide complementary results; however, single-shot MRCP had superior image quality. Moreover, assessment of ductal morphology with MRCP alone facilitated the diagnosis of different pathologic conditions of the pancreatobiliary system in the majority of patients. However, with the addition of T2- and T1-weighted sequences the overall diagnostic accuracy was significantly improved and thus we consider that a comprehensive MR approach should comprise both MRCP techniques and parenchymal sequences.

[Computed tomographic criteria as expected effect to inhaled nitric oxide in patients with severe acute respiratory distress syndrome]. / Computertomographische Kriterien für den zu erwartenden Effekt von inhaliertem Stickstoffmonoxid bei Patienten mit schwerem akutem Lungenversagen.

PURPOSE: Inhaled nitric oxide (iNO) is an effective therapy for severe hypoxemia in most patients with acute respiratory distress syndrome (ARDS). For unknown reason, a subset of ARDS patients does not respond favorably to iNO therapy. We hypothesized that radiological manifestation of lung injury may be related to iNO response. MATERIALS AND METHODS: We retrospectively analyzed data from n = 25 ARDS patients who received iNO, and underwent chest CT within 72 h prior to inhaled treatment. The morphology of coherently pathologic lung tissue was characterized by the length of the borderline between consolidated, infiltrated and atelectatic lung tissue and radiologically normal lung tissue. This quantity was expressed as relative fraction of the visceral pleural circumference and averaged over all CT slices. Furthermore we semiquantitatively determined the total volume of consolidated lung tissue as part of the whole lung. RESULTS: In n = 6 non-responders to iNO (DeltaPaO2 < 10 %), we determined a short relative borderline between normal and consolidated lung tissue due to the presence of large and coherently consolidated lung regions. In n = 19 iNO responders (DeltaPaO2 > 10 %), we found significantly less coherently consolidated lung tissue evidenced by an increased relative borderline when compared to iNO non-responders (0.09 +/- 0.02 vs. 0.1 +/- 0.01; P < 0.05). Moreover, there was a moderate and significant correlation between DeltaPaO2 induced by iNO and the relative borderline in all patients studied (R = 0.59; P < 0.05). Total fraction of consolidated lung tissue volume was not different between iNO non-responders and responders (60 +/- 3 % vs. 54 +/- 2 % n. s.). CONCLUSION: Our data demonstrate that the gross morphological distribution of pathological lung tissue influences iNO response in ARDS. Inhaled NO was most beneficial in injured lungs characterized by many small consolidated areas surrounded by normal lung tissue. The increased borderline between pathologic and normal lung tissue offers additional possibility for iNO to divert blood flow from shunt areas to ventilated lung regions, which consequently improves arterial oxygenation.

[The effect of the field strength on standardized MRI of the brain to demonstrate cranial nerves and vessels: a comparison of 1.5 and 3.0 Tesla]. / Auswirkung der Feldstärke bei der standardisierten Kernspintomographie des Gehirns am Beispiel der Darstellung von Hirnnerven und Gefässen in den basalen Zisternen: Vergleich zwischen 1.5 und 3.0 Tesla.

PURPOSE: Comparison of MR images acquired as routine examinations at a field strength of 3.0 T and 1.5 T to determine whether and to which degree the image quality improves at the higher field strength of 3.0 T. MATERIALS AND METHODS: Routine MR images of 200 patients were examined retrospectively, with 100 images obtained at 1.5 T and 100 obtained at 3.0 T. The examinations were performed with a quadrature head coil and focused on the basal cisterns because of the abundance of small distinct structures in this region. We selected the T2-weighted 2D-FSE sequence in transverse direction for comparison. At both field strengths, the same section thickness of 5 mm and a matrix of 512 x 388 (FOV: 220 mm) were used. The quality of the images was evaluated with regard to depicting the cranial nerves N. III, V - X, the AICA and PICA. For comparison, image quality was rated with a score from 1 (well defined) to 5 (not depicted). RESULTS: A score of 1 was obtained in 46 % of the anatomic structures examined at 3.0 T and in only 9.2 % at 1.5 T. A score of 2 was given in 27.6 % of the anatomic structures at 3.0 T vs. 23.5 % at 1.5 T, a score of 3 in 17.2 % vs. 28.1 %, a score of 4 in 8.6 % vs. 28.7 %, and a score of 5 in 0.4 % vs. 10.3 %, respectively. The Mann-Whitney U test showed significance at p < 0.001 for the comparison of images at 1.5 and 3.0 Tesla. CONCLUSION: Routine magnetic resonance imaging using the same quadrature coil technique and similar acquisition times at 3.0 T and 1.5 T shows an improvement for T2-weighted images at the higher field strength.

[Detection of airways stenoses: comparison of virtual and flexible bronchoscopy]. / Detektion von Atemwegsstenosen: Vergleich der virtuellen mit der flexiblen Bronchoskopie.

PURPOSE: To compare virtual with flexible bronchoscopy for the detection of bronchial stenoses. MATERIALS AND METHODS: In a retrospective study, we compared the results of 26 patients, who had clinical suspected pathologies of the tracheobronchial airways and underwent both flexible bronchoscopy and multislice CT with 3D surface rendering of the airways. Flexible bronchoscopy and virtual bronchoscopy were compared as to the rate of detecting bronchial stenoses. For statistical analysis, we divided the tracheobronchial tree in the following sections: trachea, 2 main bronchi, 6 lobar bronchi, 18 segmental bronchi and 36 subsegmental bronchi, corresponding to 63 bronchial sections for each patient (on average) and a total of 1638 bronchial sections for all 26 patients. We graded the bronchial stenosis as less than 50 %, as 50 to 95 % and as complete obstruction. RESULTS: Virtual bronchoscopy detected 25 bronchial stenoses, while flexible bronchoscopy only revealed 17 stenoses. Stenoses with a diameter less than 50 % were found with virtual bronchoscopy 14 times and with flexible bronchoscopy 10 times. Stenoses with a diameter between 50 and 95 % were detected 7 and 4 times, respectively, and complete obstructions 4 and 3 times, respectively. Tracheobronchial stenoses were well recognized with virtual bronchoscopy. Moreover, the virtual method enabled the visualization of high-grade stenoses and post-stenotic areas that could not be passed by the fiberoptic bronchoscope. Virtual bronchoscopy detected stenoses at a higher rate but the difference was not statistically significant (stenoses < 50 %: p = 0.352, 50 - 95 %: p = 0.339, complete obstruction: p = 0.696). CONCLUSION: Virtual bronchoscopy is a useful non-invasive method for the diagnostic evaluation of the tracheobronchial tree. In comparison with flexible bronchoscopy, virtual bronchoscopy is superior in revealing high-grade stenoses and visualizing post-stenotic areas.

[The diagnostic value of radial multiplanar reformatting (MPR) in the CT-diagnosis of pulmonary embolism]. / Stellenwert radiärer Reformatierungen in der CT-Diagnostik der Lungenarterienembolie.

PURPOSE: To evaluate the diagnostic value of radial reformatting of axial multislice CT (MS-CT) data sets in the presumptive diagnosis of pulmonary embolism. MATERIALS AND METHODS: In 126 cases with the presumptive diagnosis of an acute pulmonary embolism, 4- and 8-slice axial MS-CT data sets were radially reformatted. The pulmonary vessels were evaluated by five experienced radiologists who determined the number of thrombi at the level of the segmental and subsegmental arteries. A pulmonary artery was considered as thrombosed if it showed at least one unambiguous filling defect on two consecutive sections. It was determined whether the artery was uniformly opacified without filling defects, contained thrombotic material or was completely occluded by a blood clot. Pulmonary vessels that could not be evaluated because of moving artifacts or inadequate opacification were classified as negative. Finally, all experts gave a subjective estimate concerning the simplicity of the diagnosis for both reconstructions. RESULTS: The diagnosis of thromboembolism at a subsegmental level was significantly different for axial sections and radial reconstruction: at 4-slice CT, 77 subsegmental thromboembolisms were found in axial sections vs. 98.6 in radial reconstructions; at 8-slice CT, 23.6 in axial sections vs. 31.2 in radial reconstructions; and using the combined evaluation, 103.2 in axial sections vs. 130.6 in radial reconstructions, with a significance of p = 0.043. The five judges did not find any significant differences in the number of the blood clots in each pulmonary segment. Concerning the simplicity of the diagnosis, the subjective judgment classified the radial reconstructions as better in 49.7 % and the axial sections as better in 22 %. The radial reconstructions of the 4-slice CT were classified as better in 45.7 % vs. 25 % and the 8-slice CT as better in 66 % vs. 16 %. CONCLUSION: Using radial multiplanar reformatting of the MS-CT data sets improves significantly the recognition of subsegmental pulmonary embolism.

Computed tomography-guided pulmonary nodule localization before thoracoscopic resection.

PURPOSE: To assess the success rate and complication rate of a CT-guided pulmonary nodule-marker system before thoracoscopic resection. MATERIAL AND METHODS: In 24 patients (15 M, 9 F; age range, 18-71 years) a total of 25 pulmonary nodules (in 1 patient 2 lesions simultaneously) were marked with a special wire under CT-guidance and then thoracoscopically resected. We evaluated lesion size, lesion distance to the pleura, the time of intervention, complications, and thoracoscopic success rate. RESULTS: Mean lesion size was 7 mm (range 4-15 mm) and mean lesional distance to the pleura was 13 mm (range 2-31 mm). The pulmonary nodule-marker system was positioned successfully in all 25 pulmonary nodules within 5-11 min (mean 7.5 min). Minimal pneumothoraces were observed in five patients with no requirements of chest drains. In addition, no bleeding complications or hematothorax were observed. All 25 pulmonary nodules could be resected thoracoscopically. However, in one patient (4%), the guide-wire dislocated during thoracoscopy, but the lesion could be successfully resected during thoracoscopy. CONCLUSION: The CT-guided placement of the pulmonary nodule-marker system used here offers a safe and accurate guide for the localization of small pulmonary nodules during thoracoscopic resection.