Search

Purpose: This study investigated the clinical application of a magnetic-resonance (MR)-guided breast biopsy (MRBB) system consisting of a nonmagnetic coaxial needle and a ferromagnetic core biopsy needle., Materials and Methods: MRBB was performed on 70 breast lesions. The biopsy device consisted of a nonmagnetic 14- to 16-gauge coaxial needle and a ferromagnetic 16- to 18-gauge biopsy needle., Results: Of the 70 lesions, 29 were malignant and 41 nonmalignant. All 29 malignant lesions underwent surgery and were confirmed as malignant at final histology. Of the 41 nonmalignant lesions, 35 underwent follow-up breast MR imaging (mean, 26 ± 19 months), which demonstrated no lesions changes; six lesions underwent surgery because of poor radiological-pathological correlation; of these 6 lesions, 3 were nonmalignant, one was borderline (lobular carcinoma in situ) and two were malignant (well-differentiated tubular carcinoma and infiltrating ductal carcinoma). Sensitivity, specificity, positive and negative predictive values and diagnostic accuracy were, respectively, 93.5%, 100%, 100%, 95.1% and 97.1% if the lobular carcinoma in situ was considered a nonmalignant histological result, and 90.6%, 100%, 100%, 92.7% and 95.7% if the lobular carcinoma in situ was considered malignant., Conclusions: MRBB with a ferromagnetic-nonmagnetic coaxial system represented an easy way to perform a biopsy procedure and was easily applicable in the routine clinical setting.

Purpose: The aim of our study was to review CT findings concerning secondary aorto-enteric fistulae (AEF). In particular, we aimed to evaluate signs of contrast medium extravasation from the aortic graft into the bowel (active bleeding), in correlation with clinical and surgical reports., Materials and Methods: Clinical and CT findings were retrospectively evaluated in 13 surgically proven cases of AEF. All patients underwent spiral CT examination with biphasic contrast technique, before and 30 and 80 seconds after intravenous injection of 120-150 ml of contrast medium (Ultravist 370; flow rate 2.5-3 ml/sec). Late scans (240 seconds) were also acquired on surgical anastomoses. We used 3 mm (arterial phase) and 5 mm (venous phase) slices., Results: At surgery, all 13 patients presented a communication between the bowel and the aortic graft. At CT examination, all 13 patients presented one or more CT signs indicating AEF (perigraft soft tissue, perigraft fluid, ectopic air or bowel wall thickening). Moreover, in 6 out of 13 patients, contrast medium extravasation from the aortic graft into the small bowel (active bleeding) was detected with CT. Detection of active bleeding was possible because CT examinations were performed without oral administration of contrast medium., Conclusions: CT is easily and readily available and provides an accurate evaluation of the aorta and surrounding retroperitoneal tissues. In our study we were able to identify the CT sign of active bleeding in more than a half of the patients with acute gastrointestinal bleeding (66%). Hence, we suggest that patients undergo CT examination without oral contrast medium administration in order to better appreciate the presence of active bleeding.

Purpose: To assess the accuracy of magnetic resonance (MR) with Mangafodipir (Mn-DPDP) in the identification of focal solid pancreatic lesions. The possibility of lesion characterisation based on quantification of Mn uptake was also investigated., Materials and Methods: Thirty-four patients (11 females, 23 males, aged 21-75, mean age 57) selected at sonography (US) and spiral CT (SCT) for clinically suspected focal solid pancreatic lesion, were studied between June 2000 and July 2001. Patients eligible for surgery underwent MR imaging (1.5 T) before and after infusion of 5 micromol/kg of Mn-DPDP (0.5 ml/kg Teslascan, Nycomed Amersham Health, Oslo, Norway). The baseline examination included FSE T2-weighted sequences and fat-saturated and non fat-saturated breath-hold gradient-echo T1-weighted sequences (FMP SPGR). After Teslascan infusion, fat-saturated and non fat-saturated FMP SPGR sequences were repeated. Thirty of the 34 patients had a definitive diagnosis provided by reference standards such as post-operative histology (22 cases), cytology (FNAB) and/or a follow-up period of at least 6 months (8 patients). As regards lesion characterisation, the signal-to-noise ratio (S/N ratio) and contrast-to-noise ratio (C/N ratio) of lesions compared to the pancreatic parenchyma were calculated using ROIs before and after contrast infusion., Results: The definitive diagnosis was pancreatic malignancy in 18 patients, focal pancreatitis in 5 and neuroendocrine tumours in 3. Four patients with suspected lesions at US and/or SCT were free of focal pancreatic disease. Mn-DPDP MR identified 17/18 malignancies, 2/3 endocrine neoplasms 5/5 focal pancreatitis; the 4 patients with no pancreatic lesions were correctly identified. The Mn-DPDP MR accuracy in detecting focal pancreatic solid lesions was 93%. MR missed 1 small adenocarcinoma (the only pT1 in our group) and 1 insulinoma (with Mn uptake similar to the surrounding parenchyma). Due to the small number of inflammatory lesions included in the study, no significant differences were found in signal intensity and Mn-DPDP uptake between focal pancreatitis and neoplasms., Conclusions: MR with Mn-DPDP is very accurate in the identification of focal solid pancreatic lesions. Mangafodi-pir is also very useful for excluding the presence of pancreatic lesions suspected at US or CT. The characterisation of lesions, in particular of inflammatory versus neoplastic lesions, remains problematic and requires further investigation.

Aim: To assess the feasibility and accuracy of cytologic and microhistologic breast biopsy using a MR imaging-guided stereotactic system with MR-compatible non-magnetic needles., Materials and Methods: Between December 2001 and September 2002, cytologic and microhistologic sampling of 14 lesions (12 patients) was performed in our radiology department using a commercially available MR-guided stereotactic device. MR-compatible non-magnetic needles or mixed kit (non-magnetic coaxial needle + conventional ferromagnetic needle) were used. Of the 12 patients examined, 2 were undergoing MR examination because of genetic/familial risk, 4 during post-operative follow-up and the remaining 6 for contradictory mammographic and sonographic findings. Ten of the 14 lesions were visible on MR alone. The mean lesion diameter was 12 mm (range 5-30 mm)., Results: The procedure was relatively simple and feasible. The procedure took 45 minutes on average. The cytologic samples were adequate in 4/14 cases (29%). The diagnosis was absence of malignant cells in one case, ductal carcinoma in one case and suspected carcinoma in two cases. The cytologic diagnosis was confirmed by core biopsy, and by post-operative histology in the malignant lesions. The remaining 10/14 (71%) cytologic samples were inadequate for diagnosis. All the microhistologic samples (100%) were considered sufficient for diagnosis and yielded diagnosis of benignity in 9/14 (60%) lesions and malignancy in 5/14 (40%). Two patients with benign diagnosis underwent surgery, which confirmed the diagnosis. The remaining seven patients were evaluated by follow-up MRI; the first follow-up at 3 months showed no significant changes. Post-operative histology of the 5 lesions with malignant microhistologic diagnosis confirmed the nature of the lesions., Conclusions: MR-guided stereotactic biopsy is a simple, fast and safe procedure comparable to the interventional breast procedures performed under mammography and ultrasound guidance. The stereotactic device used ensures correct positioning of the needle into the lesion. The new dedicated non-magnetic needles and the mixed kit provide quantitatively and qualitatively adequate tissue for the histologic analysis. On the basis of our initial experience, we conclude that the procedure is feasible and accurate and can therefore be recommended for routine clinical use.

The aim of this work is to describe the characteristics of Magnetic Resonance contrast media available or experimental in evaluation of focal liver diseases: non specific extracellular contrast media, reticoloendothelial and hepatobiliary-specific contrast media. A rational use of these different categories is suggested by analysis of their molecular and pharmacokinetic features.

Introduction: Cardiovascular disease remains the leading cause of death in the world. Invasive coronary angiography is currently the only clinical method available to visualize the coronary arteries, with up to 20% of the procedures demonstrating no evidence of severe coronary artery stenoses. We investigated the role of two-dimensional (2D) coronary magnetic resonance angiography (MRA) in patients with suspected coronary arteries disease and to check the placement and the patency of previously placed coronary artery stents., Material and Methods: Eleven patients with suspected coronary artery disease who underwent elective cardiac catheterization with coronary angiography were examined with 2D coronary MRA to detect coronary artery stenoses. Other 11 patients with 13 stented coronary arteries (6 RCA, 5 LAD, 2 Lcx) were prospectively examined with MRA one day to 8 months after stent placement. Eighteen amagnetic stents were imaged. Imaging was performed with a 1.5 T MR unit (GE Signa Horizon Echo Speed) with a phased array multicoil. Segmented k-space fast GE sequences were acquired with and without fat suppression at several cardiac cycle phases within a single breath-hold. Correlation with coronary angiography was performed in all patients., Results: Thirteen significant stenoses were found at coronary angiography in 11 coronary arteries. The sensitivity and specificity of MR coronary angiography, as compared with conventional angiography, in correctly identifying the single vessels with > 50% angiographic stenoses were 73% and 94%, respectively. The corresponding positive and negative predictive values and accuracy were 88%; 84% and 86%, respectively. As far as the study of coronary artery stents is concerned, no MR-related adverse events were observed. The stents were visualized as signal loss areas. The length of the signal loss corresponded to the length of the stents in all 18 cases (r = 97). The patent blood flow distal to the stents appeared as a high-signal band distal to the signal void, corresponding to stent patency at coronary angiography., Conclusions: Although in an early stage of technical development, 2D coronary MRA can depict 73% of hemodynamically severe coronary artery stenoses. Moreover breath-hold coronary cine MRA is a safe technique to visualize coronary artery stents. Stent site and patency can be noninvasively studied with this technique.

Purpose: The methods to measure intraabdominal fat amount and to distinguish visceral from subcutaneous fat are useful and needed because visceral obese people are at risk of developing cardiovascular disorders. We investigated US capabilities in measuring intraabdominal fat thickness and distribution distinguishing visceral from subcutaneous fat. The results were compared with those obtained with CT, the gold standard, and with the waist-hip ratio (W/H)., Material and Methods: Thirty obese women admitted to the Internal Medicine I Department, Ospedale S. Raffaele (Milan, Italy) were examined. The patients, aged 18-60 years and with BMI ranging 29.0-47.3, were submitted to consecutive double blind measurements with US and CT. The following anthropometric values were compared for every patient: W/H, US visceral/subcutaneous thickness, CT visceral/subcutaneous thickness, CT visceral area, CT subcutaneous adipose area and CT visceral/subcutaneous adipose area., Results: The classification of visceral obesity by W/H (> .85) was confirmed by CT visceral/subcutaneous adipose area (> .491). The W/H correlated significantly with CT visceral/subcutaneous adipose thickness and CT visceral/subcutaneous adipose area (r = .52, p < .004; r = .51, p < .004), but not with US visceral/subcutaneous adipose thickness (r = .42, p < .06). Significant correlations were found between Ct visceral/subcutaneous adipose area and with both US and CT visceral/subcutaneous adipose thickness (r = .59, p < .006; = .71, p < .0001). A high correlation was found between US visceral/subcutaneous adipose thickness and CT visceral/subcutaneous adipose thickness (r = .96, p < .0001)., Conclusion: Analyzing the results of the different methods, we conclude that US can always be used to study abdominal fat amount and distribution in obese women because this method exhibits significant correlations with CT, the gold standard. The W/H is not sufficient to distinguish visceral from subcutaneous intraabdominal fat.

Introduction: We report our experience with the superselective embolization of iatrogenic vascular injuries. We used new coaxial catheters and new microcoils to assess their technical features and efficacy., Materials and Methods: Diagnostic angiography by catheterization of the main renal artery was performed in 7 patients with suspected renal vascular injury to detect injury presence and site. After super-selective catheterization of the feeding arteries with a coaxial catheter (Tracker 18), we performed embolization with platinum microcoils inserted using a coil pusher. The procedure efficacy was assessed at angiography and by comparing serum hemoglobin and hematocrit levels before and after embolization; posttreatment renal function was also studied., Results: Selective embolization permitted complete occlusion of the arteries feeding the vascular injury in all patients, preserving as much parenchyma as possible. Hemoglobin and hematocrit levels were improved or unchanged after the maneuver (average hemoglobin level before embolization: 9.01 mg/dl versus 9.95 mg/dl after embolization; average hematocrit level before embolization: 27.3% versus 30.06% after embolization). We had no difficulty during microcoil insertion and the coaxial catheter was always stable., Conclusion: Superselective embolization of iatrogenic vascular injuries is an effective and definitive treatment. We recommend it in the patients with renal bleeding due to iatrogenic vascular injury after percutaneous diagnostic or therapeutic procedures.

Even though digital mammography might potentially yield major advantages in management, biology and diagnosis, state-of-the-art digitalization is still in a clinical experimental phase. To assess the diagnostic accuracy of a digital mammography system with storage phosphors, we analyzed 320 digital and 320 conventional mammograms acquired in lateral-oblique projection in the same patients. Digital mammography capabilities in identifying and characterizing breast lesions were compared with those of conventional mammography; the presence/absence of lesions and their benign/malignant nature were investigated. Complete conventional mammography was our gold standard. The digital system did not miss any malignant lesion but it did miss some benign lesions (focal masses) which had been depicted with conventional mammography, especially small low-contrast opacities. Microcalcifications were better depicted on digital images which showed, at the same time, tissues of different density, thanks to their wider dynamic range. Digital mammography yielded a false positive result on a cluster of microcalcifications, because its spatial resolution is lower. The diagnostic yield of digital mammography was poorer only in the detection of small low-contrast lesions and in the characterization of microcalcifications. To conclude, in our experience, the storage phosphor system seems to be suitable for clinical mammography, but only with careful monitoring; in contrast, we think it is not yet suitable for screening purposes.

We studied the main ocular and retrobulbar vessels with color-Doppler US and report on examination technique, detectability of the vessels and their flow characteristics to define the normal ranges of Doppler spectra for each artery and vein. We comparatively examined both eyes of 20 healthy subjects. Of each eye we studied the ophthalmic artery, the central retinal artery, the ciliary artery, the central retinal vein and the superior ophthalmic vein. The following flow parameters were considered: peak systolic velocity, end-diastolic velocity and resistive index for arteries; maximum and minimum velocity for veins. The examination lasted about 10 minutes per eye--20 minutes for each subject. In all subjects the five investigated vessels were identified in both eyes, and adequate Doppler spectra were obtained. The average peak systolic velocities of ophthalmic, central retinal and ciliary arteries were respectively about 35, 12 and 10 cm/s, with low resistance patterns (resistive index: 0.75, 0.72 and 0.68, respectively). The venous flow, which is usually continuous, may be sometimes influenced by cardiac and respiratory activities: the maximum velocities of superior ophthalmic and central retinal veins were about 6 and 5 cm/s, respectively. Color-Doppler US noninvasively visualizes both ocular and retrobulbar vessels, providing major hemodynamic information from different flow parameters; the knowledge of these parameters in normal conditions can be the basis of hemodynamic studies in many abnormal orbital conditions.

The recent developments in thoracic surgery have broadened the application spectrum of thoracoscopy. Video-assisted thoracoscopic lung resection is now available both to diagnose and to treat peripheral pulmonary nodules. A major factor limiting the success of thoracoscopic lung resection is the difficult localization of the pulmonary nodule. When lesions are very small and deep inside the lung, their preoperative detection may be difficult. This study was aimed at investigating the effectiveness of hookwires for CT-guided localization of peripheral pulmonary lesions requiring thoracoscopic resection. In 12 patients 12 nodules, previously detected with chest radiography, were localized with hookwires under CT guidance. Ten of 12 patients underwent thoracoscopic resection for diagnostic purposes, while the extant two patients underwent therapeutic wedge resection. In all patients prethoracoscopic hookwire positioning was successful. In 11 of 12 patients the lesion was accurately localized, while in the extant patient the wire tip was inserted deeper than the nodule. In 3 patients the wire was dislodged before surgery: in all of them a small hemorrhage developed on the visceral pleura and wedge resection was likewise possible. Four patients had a small pneumothorax which was not treated before surgery. In conclusion, percutaneous hookwire localization of peripheral lung lesions is an easy and efficacious technique and promotes thoracoscopy as a less invasive alternative to thoracotomy.

March, 1991, to June, 1992, five lung transplantations for end-stage lung disease were successfully performed at the Ospedale Maggiore Policlinico in Milan. All patients underwent high-resolution CT (HRCT) of the lung in a complex follow-up program to identify specific abnormalities of acute and chronic rejection (bronchiolitis obliterans) and to monitor the resolution of the bronchial anastomosis. Twenty-two HRCT exams were performed. In patients with acute rejection HRCT failed to identify specific abnormalities of lung parenchyma. In contrast, in one patient with pathological evidence of early bronchiolitis obliterans HRCT showed decreased peripheral vascularization. In the study of the bronchial anastomosis, HRCT showed optimal anastomosis resolution in 4 patients, whereas in one patient with a granuloma demonstrated by fibrobronchoscopy it confirmed the lesion showing also a small pneumomediastinum. Even though the HRCT finding of decreased peripheral vascularization does not appear to be specific for bronchiolitis obliterans, it may be of value in suggesting the diagnosis of early bronchiolitis obliterans in lung transplant. HRCT should be used in all patients with bronchoscopic diagnosis of bronchial complication to study the lesion and its mediastinal spread.

This study was aimed at assessing the Magnetic Resonance (MR) features of persistent hip pain in children. Twenty-six patients aged 1.6 to 15.2 years (mean: 6.5 years) were clinically selected for the study; at clinics, all patients had persistent hip pain after 10 days' therapy. All patients were examined with radiography, US and MRI at 0.5 T. SE T1-weighted sequences, with and without fat suppression (FS), SE T2-weighted and gradient echo (GE) T1-weighted-like (T1*) sequences were acquired on the coronal plane. Slices were 5 and 3 mm thick on SE and GE T1* sequences, respectively. Morphology and signal intensity of epiphysis, growth plate and metaphysis were prospectively studied with MRI. Clinical and/or imaging follow-up (3 months) was the reference standard in our study. Final diagnoses were: no evidence of alteration (n = 3), transient synovitis (n = 6), rheumatic fever (n = 3), Perthes' disease (n = 7), Meyer's dysplasia (dysplasia epiphysealis capitis femoris, DECF) (n = 2), early slipped capital femoral epiphysis (n = 2), incomplete fracture (n = 1), extraarticular cause of pain (muscular abscess, osteomyelitis) (n = 2). In 23 of 26 patients MRI confirmed clinical, radiographic and US findings. MRI was particularly helpful in making an unquestionable diagnosis in the other 3 cases; in a patient with suspected slipped capital femoral epiphysis MRI revealed an incomplete fracture, in a patient with suspected Meyer's dysplasia MRI revealed early Perthes' disease and finally in a patient with suspected transient synovitis MRI revealed Perthes' disease. To conclude, MRI allows the condition causing persistent hip pain to be assessed and accurately depicted, integrating clinical, radiographic and US findings and in some cases also changing diagnosis and therapy.

The purpose of this study was to assess Magnetic Resonance Imaging (MRI) patterns of hepatocellular carcinoma (HCC) treated with percutaneous ethanol injection (PEI) or Transarterial Chemoembolization (TACE) and, consequently, the potential role of MR Imaging in the follow-up of these lesions. HCC treated with PEI. Thirty-one patients with a single small HCC lesion underwent MR Imaging at 0.5 T before and after PEI. In all cases before and after treatment contrast enhanced Computed Tomography (CT) and US-guided fine-needle biopsy were performed. Twenty-seven of 31 HCC lesions in which complete tumor necrosis was obtained with PEI showed homogeneous hypointensity on SE T2-weighted MR images. This feature corresponded to an unenhanced and low-attenuation area on follow-up contrast-enhanced CT scans. All these lesions were negative for malignant cells at fine-needle biopsy follow-up. In four HCCs, high-signal areas on SE T2-weighted images and high-attenuation areas on contrast-enhanced CT scans were observed, suggesting the presence of residual tumor tissue; these lesions were positive for malignant cells at 6-month fine-needle biopsy. In each case, incomplete tumor necrosis was confirmed at pathologic examination of the surgical specimen. HCC treated with TACE. Twenty-one patients with a total of 36 HCC lesions underwent plain and Gadolinium-enhanced MR Imaging before and after TACE. 10 HCC lesions were later surgically resected; 26/36 lesions underwent MR, CT and angiographic follow-up. At short-term follow-up exams (15-30 days), hypointensity was present on enhanced SE T1 weighted sequences in those lesions (5/10) in which complete tumor necrosis was histologically confirmed. In the remaining 5/10 HCC lesions, persistent viable tumor portions were found at pathology. These areas corresponded to areas on hyperintensity of Gadolinium-enhanced SE T1-weighted images. Hypointensity on both SE T2-weighted and enhanced SE T1-weighted images was a characteristic pattern on long-term follow-up MR images in 21/26 unresected lesions; this finding was correlated with devascularization at angiography; the presence of hyperintense areas on SE T2 weighted and enhanced SE T1-weighted images corresponded to the persistence of hypervascular (viable) areas at angiography.

This work was aimed at assessing Magnetic Resonance (MR) accuracy in the preoperative investigation of abdominal aortic aneurysms, with a view to suggesting MRI as the method of choice for surgical planning, replacing other relatively invasive investigation techniques like angiography and enhanced CT. In the last 3 years 80 patients with abdominal aortic aneurysms identified with US or CT were examined with 0.5-T MRI and underwent surgical repair within 15 days. Spin-echo (SE) T1-weighted axial, sagittal and coronal sequences were always performed. In 18 patients gradient-echo (GE) flow sequences were also acquired; SE T2-weighted sequences were used to study parietal thrombi in 10 patients and finally Gd-DTPA T1-weighted sequences were obtained in inflammatory aneurysms (3 patients). MR images were blindly evaluated by 2 radiologists. The following variables which are useful for surgical planning were considered for each patient: aneurysm extent, characteristics of parietal thrombi, dissections, fixurations, inflammation signs, involvement of renal arteries, vena cava, ureters, duodenum, the presence of retroaortic renal vein or of other anomalies or associated conditions. MR images were always compared with intraoperative findings, since surgery was considered as the gold standard. MR findings were in agreement with surgical findings in the evaluation of cranio-caudal aneurysm extent (78/80), parietal thrombus features (80/80), dissection (1/1) and fixuration signs (8/9), origin of renal arteries (80/80), inferior vena cava involvement (3/3), other anatomical anomalies or conditions (6/6). Inflammation signs were overestimated (14/8) as well as adhesion between aneurysm and duodenum (17/13). The presence of distal renal polar arteries was underestimated (2/4). In conclusion, MRI proves to be a highly sensitive method to demonstrate abdominal aortic aneurysms and is therefore suggested as the examination of choice before surgery.

To assess the efficacy of a new US contrast agent [SH U 508 A (Levovist), Schering] in evaluating hepatocellular carcinoma (HCC) vascularization, 38 patients with 43 lesions were submitted to color Doppler US before and after i.v. contrast medium administration. Four patients were studied after arterial chemoembolization. The patients had been selected on the basis of suboptimal color Doppler signals on baseline images. Each patient received two to four injections of Levovist in standard doses. Tumor vascularization was qualitatively graded on a 0-3 scale. Twelve tumors (27.9%) appeared avascular at baseline examinations, while 31 (72.1%) exhibited low to moderate flow signals. After contrast agent administration, color Doppler signals were markedly enhanced in 35/43 lesions (81.4%), lasting 40 to 240 seconds. The lack of enhancement was related to tumor hypovascularity (necrosis at CT), portal vein thrombosis, deep location and successful chemoembolization. The detection of flow signals in chemoembolized tumors was explained by the persistence of viable tumor tissue. After Levovist administration, flow signals were detectable in 97.6% of the HCCs. Therefore, Levovist proved to be an effective tool for color Doppler evaluation of HCC vascularization.

The correct staging of renal neoplasms is critical for both prognosis and subsequent treatment planning. This study was aimed at evaluating the accuracy of contrast-enhanced CT, of plain MR and of contrast-enhanced MR in the staging of renal cell carcinomas (RCC). Forty-four consecutive patients with renal masses were evaluated. All of them underwent plain and enhanced CT and 0.5 T plain and Gadolinium-enhanced MR. T1, T2 and Gd-DTPA T1-weighted MR sequences were acquired (500/20; 2000/100). The diagnostic studies were prospectively and separately interpreted by three radiologists. Imaging findings were then correlated with pathologic findings in all cases. The Robson Staging System was used for pathologic classification. The anatomical staging was correctly performed by CT and MR in 36/44 patients (82%) and by Gd-MR in 39/44 (89%). Seven stage-I and 1 stage-IIIC patients were overstaged by enhanced CT and by plain MR (3 patients to stage II and 5 to stage IV A); 5 stage-I patients were over-staged by Gd-MR (4 patients to stage II and 1 to stage IV A). In four cases, in which CT and MR misdiagnosed stage IV A, Gd-MR correctly identified the lack of adjacent organ infiltration. Perinephric spread, venous invasion, metastatic lymphadenopathy and distant metastases were correctly evaluated by the three imaging modalities. Our results suggest that Gd-DTPA could be useful in identifying direct tumor spread to adjacent organs, because of the different enhancement between the cancer and the other tissues.