Search

Purpose: To determine the clinical presentation in cases of adult intussusception demonstrated at computed tomography (CT) or magnetic resonance (MR) imaging and to correlate the imaging appearance with clinical diagnosis., Materials and Methods: Retrospective review of CT and MR images and clinical records of all patients with an intussusception demonstrated on CT or MR images from January 1, 1991, through April 30, 1998., Results: Thirty-three patients had one or more intussusceptions demonstrated on CT (n = 30) or MR (n = 3) images. Twenty-nine patients had enteroenteric intussusceptions, and four had intussusceptions involving the colon. Ten patients (30%) had a neoplastic lead point, including all four of the intussusceptions involving the colon (benign mass, n = 3; malignant mass, n = 7). In 23 cases (70%), no neoplastic lead point was identified. A variety of causes were implicated in these cases, with 16 cases (48%) classified as idiopathic. Enteric intussusceptions in the nonneoplastic group were shorter in length (median, 4 vs 10.8 cm; P = .002), smaller in diameter (median, 3 vs 4 cm; P = .002), and less likely to be associated with obstruction (4.3% vs 50%; P = .02)., Conclusion: Less than one-third of adult intussusceptions demonstrated at CT or MR imaging were caused by a neoplastic lead point. Almost half of adult cases in this series were idiopathic.

Purpose: To determine the effect of glucagon and orally administered contrast material on hepatic enhancement and metastatic lesion conspicuity on computed tomographic (CT) scans., Materials and Methods: Nine patients with a history of hepatic metastasis underwent two CT examinations with intravenously administered contrast material. Each patient was given orally administered CT contrast material for the first examination. No oral contrast material was given in the second examination. Five patients underwent the initial CT with intravenous administration of 1 mg of glucagon; the second examination was performed without glucagon. Four patients were administered glucagon before the second examination, but no glucagon was administered before the first. Attenuation in the liver, portal vein, and aorta was measured by observers blinded to whether the patient had been given glucagon. Lesion conspicuity was rated on a continuous scale., Results: Greater mean hepatic enhancement was noted on scans of patients in whom oral contrast material was administered (mean, 52 HU) versus those in whom no oral contrast material was administered (mean, 47 HU; P = .019). Glucagon was not associated with greater hepatic enhancement. Neither oral contrast material nor glucagon had a significant effect on lesion conspicuity., Conclusion: Oral CT contrast material is associated with a small increase in hepatic enhancement that does not appear to be clinically important. Glucagon does not appear to affect hepatic enhancement or lesion conspicuity in humans.

Purpose: To investigate the effect of glucagon on the timing and degree of hepatic enhancement at computed tomography., Materials and Methods: Each of 11 dogs underwent injection of contrast material at two uniphasic rates (0.5 mL/sec and 1.5 mL/sec) with without previous intravenous administration of 0.5 mg of glucagon. Scans were obtained at a single intrahepatic level every 3.5 seconds for 120-160 seconds. Attenuation values in the liver, aorta, portal vein, and inferior vena cava (IVC) were measured., Results: Glucagon administration was associated with greater hepatic enhancement throughout the study: Peak hepatic enhancement was greater at contrast injection rates of 0.5 mL/sec (P = .021) and 1.5 mL/sec (P = .0001). Peak hepatic enhancement also occurred earlier during the glucagon runs. Portal vein enhancement was greater during the glucagon runs at an injection rate of 1.5 mL/sec (P = .032). IVC enhancement was greater during the nonglucagon runs at 0.5 mL/sec (P = .013) and at 1.5 mL/sec (P = .005)., Conclusion: Intravenous administration of glucagon before contrast material injection produces greater and more rapid hepatic enhancement in a canine model.

Purpose: To evaluate the magnetic resonance (MR) imaging findings of os acromiale in rotator cuff disease., Materials and Methods: Retrospective analysis was performed of 10 shoulder MR studies of 10 patients older than 25 years with os acromiale. The authors subdivided these patients into three types on the basis of seven possible configurations., Results: Eight of the 10 patients with os acromiale showed failure of fusion between the mesoacromion and meta-acromion (type A), one between the preacromion and mesoacromion (type B), and one between the meta-acromion and basiacromion (type C). Osteophytic lipping was seen at the margins of the acromial gap in nine cases. The os acromiale could be detected on the oblique sagittal image, oblique coronal image, or both, and in each case, there was a lesion in the supraspinatus tendon: tendinitis in four patients and tendon tear in six., Conclusion: Os acromiale can be an important cause of rotator cuff impingement and can be detected with routine MR imaging of the shoulder.

Use of a picture archiving and communication system (PACS) for clinical review requires rapid access to patient images. The authors have developed a second-generation archive system for a departmental PACS that optimizes access time for both current and historical images. The system is based on a composite staging mechanism that uses multiple storage media: magnetic disks, erasable magneto-optical disks, and write-once-read-many (WORM) disks. A two-tiered communication network composed of a standard Ethernet network and a gigabit bandwidth fiberoptic UltraNet network is used for distributing images to remote display stations. The system currently archives 1.5 Gbytes of radiologic images daily from three magnetic resonance imagers, two computed tomographic scanners, one computed radiographic system, and one laser film digitizer. The system features (a) multiple optical drives and robotic arms, which support concurrent archival and retrieval operations; (b) a stacking mechanism, which provides fast retrieval of current images; (c) a pre-fetch mechanism, which minimizes on-line image retrieval, hence relieving peak-hour workload of the optical disk library and the networks; (d) studies grouping and platter management, which optimize retrieval of a patient's images from multiple examinations; and (e) a job-prioritizing control, which minimizes waiting time for radiologists and referring physicians to review images at the display stations. This system is suitable for a large-scale PACS to serve the clinical and research staff.

Purpose: To compare the appearance of hypervascular liver lesions on gadolinium-enhanced fast low-angle shot (FLASH) imaging with T2-weighted fat-suppressed spin-echo imaging, dynamic nonequilibrium-phase FLASH imaging, and dynamic nonequilibrium-phase iodine-enhanced computed tomography (CT) and to characterize the appearance of lesions on serial postgadolinium FLASH images., Materials and Methods: Twenty-nine patients with hypervascular malignant liver lesions were examined with dynamic contrast-enhanced CT and magnetic resonance (MR) imaging within a 1-month interval. MR sequences included T2-weighted fat-suppression, precontrast FLASH, and postgadolinium FLASH at 1 second (sinusoid phase), 45 seconds (nonequilibrium phase), and 10 minutes., Results: More than five lesions were detected in 12 patients with CT, 15 patients with T2-weighted fat-suppression imaging, 16 with sinusoid-phase FLASH imaging, and 11 with nonequilibrium-phase FLASH imaging. In six patients, a statistically significant (P = .03) increase in the number of lesions detected, by category, was observed on sinusoid-phase FLASH images compared with CT images., Conclusion: Sinusoid-phase FLASH imaging is superior to nonequilibrium-phase imaging with MR or CT for the demonstration of hypervascular malignant lesions.

Fifty-eight patients suspected of having focal hepatic disease were studied prior to and following the intravenous administration of manganese (II) N,N'-dipyridoxylethylenediamine-N,N'-diacetate 5,5'-bis(phosphate) (DPDP), a hepatobiliary magnetic resonance (MR) contrast agent. Four doses (3, 5, 8, or 10 mumol/kg) of Mn-DPDP were used to test the hypothesis that Mn-DPDP-enhanced MR imaging would display enhancement in tumors of hepatocellular origin. A total of 203 lesions were evaluated. Histologic proof was available in 32 cases, and in 26 cases lesions were evaluated on the basis of characteristic imaging findings. Statistical calculations for distinction of tumors of hepatocellular origin yielded a sensitivity of 100%, a specificity of 92.0%, an accuracy of 93.6%, a positive predictive value of 75.9%, and a negative predictive value of 100%. The authors conclude that the presence and patterns of enhancement at Mn-DPDP-enhanced MR imaging permit reliable distinction between hepatocellular and nonhepatocellular tumors.

Ninety-six patients with known or suspected focal hepatic disease were evaluated in a multiinstitutional study of manganese (II) N,N'-dipyridoxylethylenediamine-N,N'-diacetate 5,5'bis(phosphate) (DPDP) as a hepatic-specific contrast agent for magnetic resonance (MR) imaging. The patients were divided into four dose groups, receiving 3, 5, 8, or 10 mol/kg of Mn-DPDP. Half of the patients in each dose group received Mn-DPDP as an intravenous bolus (0.25 mL/sec) and the other half as an infusion (1 mL/min). Patients were evaluated with T1-weighted imaging parameters. No serious side effects were noted. In 76 patients, both Mn-DPDP-enhanced and nonenhanced T1-weighted images depicted the same number of lesions, but one additional lesion was depicted with enhanced imaging in 12 patients, two additional lesions in three patients, and three additional lesions in three patients. Enhanced, T1-weighted images depicted no more lesions than nonenhanced, T2-weighted images in 77 patients, but one more lesion was depicted in nine patients, two more lesions in two patients, three more lesions in one patient, and four more lesions in one patient.

The utility of perfluoroctylbromide (PFOB) as a gastrointestinal contrast agent for magnetic resonance (MR) imaging was evaluated with MR examinations performed in 30 subjects (16 healthy volunteers and 14 patients). Transaxial T1-, proton density-, and T2-weighted MR images were acquired in each subject before and after the administration of PFOB. The healthy volunteers each underwent two sets of post-PFOB MR examinations, one before and one after glucagon administration. The degree of bowel marking, clarity of bowel-wall visualization, ability to distinguish bowel from adjacent parenchymal organs, and severity of phase-encoding artifacts were independently analyzed by two reviewers. Oral administration of PFOB significantly (P less than .001) increased the percentage of bowel loops with low signal intensity. Subcutaneous administration of glucagon significantly (P less than .001) increased the clarity of bowel-wall visualization on post-PFOB MR studies. The severity of phase-encoding artifacts did not change substantially after administration of PFOB or glucagon.

The objective of this study was to quantitatively and qualitatively determine contrast enhancement patterns of normal abdominal organs with dynamic gadolinium-enhanced magnetic resonance (MR) imaging. Dynamic gadolinium-enhanced, T1-weighted, spin-echo imaging was performed during a 23-second breath hold in 38 patients, with images acquired before, during, and at 1,2, and 5 minutes after bolus injection of gadopentetate dimeglumine. Enhancement patterns of normal liver, spleen, pancreas, adrenal gland, kidney, aorta, inferior vena cava, and fat were determined by visual evaluation and by performance of signal intensity measurements with an electronic cursor. Time-intensity curves demonstrated peak enhancement of all abdominal organs during or immediately after bolus injection of gadopentetate dimeglumine. MR enhancement patterns included visualization of renal cortical nephrogram and heterogeneous enhancement of the spleen during the bolus phase of contrast material administration. Peak enhancement of normal liver was 72%; spleen, 172%; pancreas, 82%; adrenal gland, 85%; and kidney, 291%. This study established reference data regarding abdominal organ enhancement that will be useful as dynamic gadolinium-enhanced MR imaging becomes clinically implemented.

Magnetic resonance (MR) examinations of 43 patients with 95 focal hepatic lesions (diameter, greater than 1 cm) were analyzed for lesion shape, homogeneity, and relative signal intensity compared with normal liver parenchyma, spleen, and skeletal muscle. On T1-weighted, balanced, and T2-weighted images, most metastases (74%), cavernous hemangiomas (76%), and cysts (82%) were smooth and round or oval, while the hepatocellular carcinomas all had irregular borders (40%) or were lobulated (60%). All lesions with irregular borders were malignant. Seventy percent of metastatic lesions, 85% of cavernous hemangiomas, and 100% of simple hepatic cysts were of homogeneous signal intensity, while 60% of hepatocellular carcinomas were inhomogeneous. Logistic regression analysis of multiple lesion characteristics showed that inhomogeneous lesions had a high likelihood of malignancy, while markedly hyperintense lesions had a very low probability of being malignant, regardless of other traits. Homogeneous lesions that were isointense or hyperintense compared with spleen on balanced images but were not markedly hyperintense on T2-weighted images also had a high likelihood of malignancy.

The authors studied the ability to improve detection of splenic lesions during suspended respiration with dynamic gadolinium-enhanced T1-weighted spin-echo magnetic resonance (MR) imaging. In the first phase of the study, normal splenic contrast material enhancement patterns were assessed in 10 control patients without splenic lesions. A heterogeneous signal intensity pattern was observed in 11 patients with splenic lesions during bolus injection of gadopentetate dimeglumine, with conversion to homogeneous enhancement 1 minute later. Mean splenic enhancement was 321% during bolus injection, with a rapid return toward baseline signal intensity thereafter. In the second phase, evaluation of 18 splenic lesions detected with contrast-enhanced computed tomography in 11 patients revealed that dynamic gadolinium-enhanced MR pulse sequences significantly improved lesion conspicuity and detectability compared with conventional T1-and T2-weighted pulse sequences. Contrast-to-artifact ratio measurements were 0.5, 3.7, and 9.3 for conventional T1-weighted, T2-weighted, and dynamic gadolinium-enhanced MR images, respectively.

Rapid acquisition spin-echo (RASE) magnetic resonance (MR) imaging allows for coverage of the entire liver with highly T1-weighted SE images during a single 23-second breath-holding period. The RASE sequence was implemented in conjunction with rapid intravenous injection of gadopentetate dimeglumine to enable performance of dynamic contrast material-enhanced MR imaging of the liver. Prospective evaluation of 24 patients with 62 liver lesions 1 cm or greater in diameter was performed. Images obtained with RASE were devoid of respiratory-related ghost artifacts or edge blurring. The dynamic contrast-enhanced RASE technique resulted in contrast-to-noise and contrast-to-artifact values and time efficiency measures significantly greater (P less than .05) than those obtained with use of conventional T1- and T2-weighted pulse sequences, indicating a higher likelihood for lesion detectability. Lesion conspicuity was maximal during or immediately following bolus administration of gadopentetate dimeglumine, with lesions often becoming obscured at delayed postcontrast imaging.

Magnetic resonance (MR) imaging was used in 40 renal transplant recipients to determine whether this modality can enable distinction of acute tubular necrosis (ATN) and acute rejection by means of corticomedullary differentiation (CMD). Each patient underwent initial MR imaging after allograft renal transplantation. Twenty-nine of these 40 patients (72%) also underwent subsequent follow-up MR imaging. Seventeen studies were obtained during episodes of ATN; 12 of these studies (71%) showed poor CMD. Eleven studies were obtained during episodes of acute rejection; eight of these studies (73%) showed poor CMD. In addition, six of seven studies (86%) showing various combinations of renal disease (ATN, acute rejection, chronic rejection, and cyclosporine toxicity) also showed poor CMD. Loss of CMD is reversible after improvement of ATN and acute rejection. Because loss of CMD is a nonspecific though sensitive sign reflecting renal transplant dysfunction, MR imaging is of limited value in the differentiation of ATN from acute rejection.

The bursa of the tibial collateral ligament (TCL) may be visualized at magnetic resonance (MR) imaging when it becomes distended with fluid. In the authors' experience, this finding signifies a pathologic condition either in the medial capsuloligamentous complex or in the bursa itself. Such a finding may indicate TCL bursitis. TCL bursitis can be suspected in the setting of isolated pain in the medial joint line in the absence of mechanical symptoms. Prolonged relief of symptoms after injection of steroid into the bursa is supportive of the diagnosis. Seven cases are presented in which a fluid-filled TCL bursa was identified at MR imaging. In five cases, TCL bursitis was suspected. The differential diagnosis for the MR findings is discussed.

Results of 53 thoracic magnetic resonance (MR) imaging examinations were reviewed to determine the prevalence and severity of artifacts and pitfalls that may occur in the evaluation of acute aortic dissection. Grade 1 artifacts and pitfalls were mimics of aortic dissection on individual images but could be demonstrated not to represent a dissection when other images from the same sequence were evaluated. Grade 2 artifacts and pitfalls required the use of images from other planes or sequences to distinguish them from a dissection. Grade 3 artifacts and pitfalls could not be distinguished from a dissection without the use of other imaging modalities. Of the 53 cases examined, 34 (64%) had artifacts or pitfalls of grade 1 or higher, 10 (19%) had artifacts or pitfalls of grade 2 or higher, and one case (2%) had grade 3 artifacts or pitfalls. Sixteen cases had more than one artifact or pitfall. Pitfalls and artifacts that mimic aortic dissection occur in a significant percentage of thoracic MR imaging examinations. An awareness of their existence, knowledge of normal anatomy, the use of axial images in all cases with the addition of images in other planes as needed, rotation of phase and frequency gradients as needed, and clinical correlation may avert misinterpretation in nearly all cases.

Dynamic contrast material-enhanced gradient-echo magnetic resonance (MR) imaging was performed on 15 patients with 18 renal masses (seven simple renal cysts, nine renal cell carcinomas, one angiomyolipoma, and one oncocytoma). Fifteen sequential images were obtained while the patients held their breath during a 2.5-3.5-minute interval during and immediately after the intravenous administration of gadolinium diethylenetriaminepentaacetic acid (DTPA); delayed images were also obtained for 15 minutes. Time-intensity curves showed that renal cortical enhancement reached maximal intensity 80 seconds after the injection of Gd-DTPA. Medullary enhancement reached maximal intensity at 120 seconds. None of the simple renal cysts showed enhancement; each cyst displayed a signal intensity less than that of the renal cortex on precontrast images. All renal cell carcinomas were isointense with the renal cortex and demonstrated variable enhancement. Three patterns of enhancement were observed: predominantly peripheral, heterogeneous, and homogeneous. Both the angiomyolipoma and the oncocytoma showed brisk, homogeneous enhancement. This MR imaging technique appears to be useful in the detection and characterization of simple renal cysts and solid neoplasms.

The diagnosis of valvular regurgitation and stenosis with cardiac cine magnetic resonance (MR) imaging is based on observation of localized regions of signal void. However, areas of signal void frequently are present in patients without cardiac abnormalities and may closely resemble signal-void areas that occur secondary to valvular pathologic conditions. Twenty healthy volunteers were prospectively evaluated with gradient-echo cine MR imaging to determine the frequency, distribution, and timing of appearance of regions of signal void in the healthy population. The most frequent sites included the regions adjacent to the atrioventricular valves, the left ventricular outflow tract, atrial appendages, and areas of venous inflow, including the coronary sinus, superior vena cava, and pulmonary veins. Awareness of normal signal-void areas and their characteristics is critical in the assessment of cardiac cine MR examinations.

The rapid acquisition spin-echo (RASE) technique combines a short repetition time, a short echo time, and a single excitation pulse sequence with half-Fourier data sampling. This allows for acquisition of 11 strongly T1-weighted sections during a single 23-second breath-holding period. Measurements obtained from volunteers and with phantoms reveal that RASE images have a lower signal-to-noise ratio and contrast-to-noise ratio than do conventional multiacquisition spin-echo (SE) images due to reduced data acquisition. However, liver-spleen contrast and spatial resolution are not affected. Moreover, contrast-to-artifact (C/A) measurements are 77% greater with RASE. When normalized for imaging time, all parameters are significantly higher with RASE, with a C/A per unit time that was 338% higher. Randomized, blinded review of RASE and SE sequences from 20 patients was conducted to evaluate qualitative performance. Excellent to good performances for phase-encoding artifact reduction, edge sharpness, and overall image quality were recorded for 89%, 88%, and 86% of RASE examinations, respectively, versus 41%, 59%, and 47% of conventional SE examinations, respectively. All results were statistically significant with P less than .001. RASE is an easily implemented imaging technique that utilizes widely available existing technology. Its major benefits relate to significant reduction in imaging time, elimination of respiratory artifacts, and the potential for performing dynamic contrast material-enhanced screening examinations.

Examinations of the liver using magnetic resonance (MR) and computed tomography (CT) were performed on 50 patients with hepatic metastases. MR and CT were comparable in their ability to detect metastases, which generally appeared hypointense compared with normal liver parenchyma on T1-weighted MR images and hyperintense on T2-weighted images. The MR imaging techniques that were most reliable in detecting metastases were inversion recovery and a relatively T2-weighted, spin-echo technique (TR = 1,500 msec, TE = 60 msec). We conclude that CT, because of its shorter imaging time, greater spatial resolution, and lower cost, should remain the preferred screening test for hepatic metastases. MR imaging should be reserved for patients with equivocal CT findings and for patients in whom there is persistent clinical suspicion of hepatic metastases despite a negative CT examination.

We describe a simple method that uses patient positioning to obtain oblique views of the human heart by magnetic resonance (MR) imaging. These views provide MR images that directly correlate with conventional anterior or posterior oblique radiography.

The sensitivities of contrast medium-enhanced computed tomography (CT), delayed CT (DCT), CT during arterial portography (CTAP), and magnetic resonance (MR) imaging for detecting focal liver lesions were prospectively evaluated in eight patients who subsequently underwent hepatic lobectomy or transplantation. Pathologic evaluation of the resected liver specimens demonstrated 37 lesions. The sensitivities were 81% (30 of 37 lesions) for CTAP, 57% (21 of 37 lesions) for MR imaging, 52% (12 of 23 lesions) for DCT, and 38% (14 of 37 lesions) for contrast-enhanced CT. The difference between the sensitivity of CTAP and the sensitivities of the other imaging tests was statistically significant (P less than .004). Of the lesions smaller than 1 cm in diameter, CTAP depicted 61% (11 of 18 lesions), MR imaging 17% (three of 18 lesions), CT 0% (zero of 18 lesions), and DCT 0% (zero of nine lesions). It is concluded that for preoperative detection of focal hepatic masses, CTAP is the most accurate technique available to most radiologists. Patients with primary or secondary hepatic neoplasms who are being considered for hepatic resection should undergo CTAP as part of their preoperative examination.

Magnetic resonance (MR) imaging was performed in 23 patients (25 tumors) with proved bladder neoplasms. MR studies were retrospectively evaluated and compared with computed tomographic (CT) and pathologic findings. Bladder neoplasms, having a signal intensity intermediate between those of urine and perivesical fat, were best seen on T1-weighted and proton-density images. MR imaging was as accurate as technically well-performed CT studies in detecting extravesical tumor extension. MR could additionally be used to assess the integrity of the bladder wall. On T2-weighted images the normal bladder wall appeared as a thin, linear, low-intensity structure. The disruption of this low-intensity line was indicative of deep muscle invasion, whereas preservation of this low intensity line implied a more localized lesion. Although chemical shift artifacts might cause apparent disruption of the bladder wall, knowledge of this artifact coupled with additional imaging along different planes helps avoid misinterpretation of this artifact as deep muscle invasion.

Preoperative staging of transitional-cell carcinoma of the upper urinary tract is important for identification of those tumors amenable to limited resection. Twenty-two patients were examined using computed tomography (CT), and three patterns were noted: (a) a focal intraluminal mass, (b) ureteral wall thickening with luminal narrowing, and (c) an infiltrating mass. In most cases, attenuation was similar to that of soft tissue; one tumor was calcified. Tumors of the renal pelvis may exhibit contrast enhancement. In 11 cases, excretory urography was inadequate or not attempted. CT demonstrated the site and cause of obstruction in all cases and proved to be a useful noninvasive staging procedure for suspected or proved transitional-cell carcinoma of the upper urinary tract.

Computed tomographic scans in 23 patients who had undergone total laryngectomy were analyzed retrospectively to determine normal postoperative appearance and to evaluate the role of CT in assessing recurrent neoplasm. Nine patients without clinical evidence of recurrence illustrated the normal postoperative changes: a round or ovoid neopharynx connecting the base of the tongue with the cervical esophagus and intact fat planes surrounding the neopharynx, neurovascular bundles, and sternocleidomastoid muscles. In the 12 patients with recurrent neoplasm, the CT manifestations included masses involving the internal jugular lymph node chain (adjacent to the neopharynx, neurovascular bundles, or sternocleidomastoid muscles), tracheostomy site, or paratracheal region. Recurrence was mimicked on CT in two patients, one with an abscess and one with metastases from an adenocarcinoma of unknown primary site. In eight patients, a distended neopharyngeal lumen correlated with benign or malignant stricture. CT supplemented physical examination and indirect mirror examination, providing data regarding presence and extent of recurrent tumor and aiding in planning the mode and scope of therapy.

A total of 103 consecutive patients with suspected biliary obstruction were studied using both computed tomography (CT) and ultrasound (US) to evaluate the relative accuracy of the methods. In 47 patients with confirmed obstruction, CT and US were comparable accurate in differentiating obstruction from nonobstruction. The precise level of obstruction was identified by CT in 88% and by US in 60%; the cause of obstruction was accurately predicted by CT in 70% and by US in 38%. Both methods detected useful additional information, such as cholelithiasis or retroperitoneal adenopathy. The authors use US as a screening examination; if there is doubt about the level and cause of sonographically demonstrated obstruction, CT has proved to be an accurate means of further evaluation.

Two hundred consecutive patients were evaluated for renal obstruction using both static B-scanning and high resolution real-time scanning. The overall diagnostic accuracy was 93% for the static B-scanner and 94% for the real-time unit, with a false positive diagnostic rate of 8% for each. Because of its greater flexibility and shorter scanning time, real-time can replace static B-scanning in the diagnosis of renal obstruction. It has also become the preferred method of guidance for interventional uroradiological procedures. The pitfalls of real-time scanning in evaluating renal obstruction are also discussed.

The authors investigated the utility of magnetic resonance (MR) imaging in identifying the normal adrenal gland in 100 patients as well as in distinguishing adrenal adenomas (n = 12) from malignant neoplasms (n = 14). The left adrenal gland was seen in 99 of 100 cases and the right in 91 of 100 cases. The adrenals were most easily seen with T1-weighted spin-echo pulse sequences. The ratio of the intensity of the adrenal mass to that of fat at 2,100/90 (repetition time msec/echo time msec) was most helpful in distinguishing adrenal adenomas from malignant neoplasms. In contrast to other studies, the adrenal mass/liver intensity ratios were not helpful. All ten lesions with adrenal mass/fat ratios at 2,100/90 of 0.8 or greater were malignant, whereas all eight adrenal masses with a ratio less than 0.6 were adrenal adenomas. However, eight (31%) of the masses (four adenomas and four malignant neoplasms) had ratios between 0.6 and 0.8. Although MR imaging has considerable potential in characterizing adrenal masses, larger studies are needed to determine its true sensitivity and specificity.

Thirty-three patients with a variety of disorders of the thoracic aorta (aneurysm, dissection, Marfan syndrome, coarctation/pseudocoarctation, L-transposition, and Takayasu disease) were evaluated with magnetic resonance (MR) imaging. MR imaging delineated the presence and extent of thoracic aortic aneurysms and showed the relationship of the aneurysm to arch vessels; it also demonstrated intimal flaps and individual lumina in types A and B aortic dissection. Dilation of the ascending aorta in Marfan syndrome and focal narrowing of the aorta in coarctation were well visualized. The anteroposterior and side-to-side relationships of the aorta and pulmonary artery in L-transposition were demonstrated, as were aortic wall thickening and branch vessel narrowing in Takayasu arteritis. Initial experience suggests that MR imaging may provide a noninvasive method for evaluating thoracic aortic disease. Limitations include inferior spatial resolution, occasional difficulty in imaging the entire region of interest in one section, lack of signal from calcifications, and inability to monitor critically ill patients.

Thoracic computed tomographic (CT) scans of 250 patients with newly diagnosed or recurrent lymphoma revealed thoracic wall involvement in 24 patients (11 with Hodgkin disease, 13 with non-Hodgkin lymphoma). Thoracic wall involvement occurred without contiguous mediastinal or parenchymal involvement in 17 patients. Of these, 13 patients had masses beneath the pectoralis muscles or within the breast, and four had masses arising from the ribs. Five additional patients had mediastinal masses with thymic involvement and parasternal extension through the thoracic wall. Pulmonary parenchymal lymphoma with thoracic wall invasion was noted in the remaining two patients. In five of nine patients receiving radiation therapy, treatment plans were modified by CT demonstration of thoracic wall lymphoma.

The authors studied the computed tomographic (CT) images obtained in 56 patients with pathologically proved biliary cancer and 75 patients with no evidence of biliary disease, attempting to define the normal anatomy of the lymphatic system draining the bile ducts and the prevalence of extrahepatic spread of primary biliary cancer into these retroperitoneal planes. Of 20 patients with gallbladder cancer, 14 (70%) had proved adenopathy and nine (45%) had peritoneal spread at presentation, and another three later developed carcinomatosis. Of 22 patients with proximal cholangiocarcinoma, 16 (73%) had nodal involvement at presentation, four later developed adenopathy, and five had peritoneal dissemination. Distal or diffuse cholangiocarcinomas were less associated with metastatic nodes or peritoneal spread. For all biliary cancers, the nodes most commonly involved were the node of the foramen of Winslow, the superior pancreatoduodenal node, and the posterior pancreatoduodenal chain. Extrahepatic tumor spread produced proximal intestinal obstruction in 13 patients (23%). CT reliably demonstrates lymphatic or other extrahepatic spread of biliary cancers, which may have an important bearing on management decisions.

Recognition of variations in size, shape, and density of the normal thymus on computed tomographic (CT) scans is of paramount importance, lest it be misinterpreted as an abnormal mediastinal mass. Studying patients subsequently proved free of active chest disease, we analyzed 154 CT scans of the mediastinum, performed on a fourth-generation scanner, to determine the incidence of visualization and appearance of the normal thymus. The thymus was seen in 100% of patients under age 30, 73% of patients between 30 and 49 years, and in 17% of patients over 49 years of age. The thickness of the thymus showed a definite decrease in size with increasing age; although the width showed a similar general tendency, a wide variation was noted within each age group. In younger patients, the density of the thymus was similar to that of muscle; the attenuation values progressively decreased in older patients, finally approaching that of fat.

Two sports-related knee injuries resulted in small fractures at the posteromedial corner of the tibial plateau. Magnetic resonance imaging demonstrated findings consistent with an avulsion injury at the semimembranosus insertion, as well as a similar pattern of internal derangement in both cases.

A comparison was made between post-lymphangiogram abdominal radiographs and computed tomographic (CT) scans in 18 lymphoma patients. When radiographs were obtained more than one year after the original lymphangiogram, only a small percentage of the patients had adequate contrast medium remaining in the lymph nodes for a definitive diagnosis. In three cases with apparently adequate residual contrast material, the radiograph did not provide accurate information on the presence or extent of recurrent lymphoma. The authors conclude that CT is the preferred method for the initial staging of patients with lymphoma, as well as for follow-up studies after one year.

Magnetic resonance (MR) images of 12 fresh uterine specimens obtained with a high-resolution surface coil were correlated with findings on gross and microscopic pathologic examination. The uterine wall can be differentiated into three distinct zones: a central high-intensity zone, a junctional low-intensity band, and a peripheral medium-intensity area. Whereas the endometrium (basale and functionale) corresponded to the high-intensity zone, the myometrium correlated best with the combined width of the areas of low and medium signal intensity. The signal intensity of a leiomyoma varied, depending on its cellularity. Primary endometrial and cervical carcinomas were clearly shown on MR images, as were other abnormalities such as adenomyosis and Nabothian cysts. The potential usefulness of MR imaging in detecting and staging uterine neoplasms is discussed.

A simple method of improving visualization of the posterior horn of the medial meniscus during fluoroscopic spot-filming in arthrography of the knee is described. Since it results in increased diagnostic accuracy at the most common site of occult pathology, the authors recommend that it be part of the routine examination.

Magnetic resonance (MR) imaging studies were performed in 29 patients with clinical findings of stage B prostatic carcinoma with use of both T1 and T2-weighted spin echo sequences. MR imaging findings were correlated with surgical/pathologic results in 18 patients who underwent an operation. Periprostatic fat, periprostatic venous plexus, seminal vesicles, and lymph nodes were assessed from MR imaging studies. When each was assessed independently, the periprostatic fat signal had a sensitivity of 29%, specificity of 100%, and accuracy of 85% in detecting extracapsular tumor extension, whereas the periprostatic venous plexus had a sensitivity of 57%, specificity of 86%, and accuracy of 80%. MR imaging had a sensitivity of 50%, specificity of 97%, and accuracy of 89% for detecting seminal vesicular involvement. When all four anatomic features were taken into consideration, MR imaging had an accuracy of 89% (16 of 18 patients) in differentiating stage B from stage C or D disease (sensitivity 87%, specificity 90%). Our data indicate that MR imaging is a promising method for staging prostatic carcinoma.

In a review of 200 consecutive CT scans of the upper abdomen, the structures within the gastrohepatic ligament (GHL) were well seen in 182 (91%). In 85% of these 182 patients, the largest structure visible within the GHL was 6 mm or smaller. A total of 27 patients had a structure larger than 6 mm within the GHL; this finding could be explained in 13 by the presence of a normal anatomic variant. Of the 14 others, 12 had known tumor arising in or known to have spread to the upper abdomen. Two patients had no obvious explanation. Fourteen patients with cancers of the stomach (9 patients), pancreas (3 patients), and esophagus (2 patients) had 57 intact nodes that were evaluated pathologically. Of these 40/40 benign nodes and 10/17 malignant nodes were less than or equal to 8 mm in size. When anatomic variants are excluded, the finding of rounded structures greater than 8 mm in the GHL is a reliable indicator of left gastric node involvement by carcinoma or lymphoma or of coronary venous dilatation.

Compensatory hypertrophy of bone developed in 5 patients following surgery on the foot. The typical finding of diffuse cortical thickening of the entire ray is thought to represent a static compensatory response which occurs when the usual dynamic response of altered weight-bearing is inadequate or overridden.

Sonographic and computed-tomographic (CT) findings in two patients with mesenteric desmoid tumors are reviewed. On ultrasound scans, mesenteric desmoids appear as well-circumscribed masses with scattered dense high-level echoes which correspond to mesenteric fat and vessels on the CT scan. Although these tumors appear well marginated on both ultrasound and CT scans, the pathologic examination shows microscopic invasion at the tumor margin in most cases. Differentiation from other mesenteric masses is discussed.

Seventy-nine magnetic resonance (MR) studies of the knee were reviewed in an evaluation of the ability of MR imaging to demonstrate arthroscopically proved anterior cruciate ligament (ACL) tears. MR findings were also compared with the findings of two commonly applied clinical tests of ACL instability: the Lachman test and the anterior drawer test. The sensitivity of MR imaging was 94% (17 of 18), compared with 78% (14 of 18) for the anterior drawer test and 89% (16 of 18) for the Lachman test. The specificity was 100% for all three. Three MR criteria were applied: irregularity or a wavy contour of the anterior margin of the ligament, high-signal-intensity change within the substance of the ligament on T2-weighted images, and discontinuity of that substance. The sagittal T2-weighted image was especially helpful, producing an "arthrographic" effect, in which the anterior margin of the ACL is outlined by high-signal-intensity joint fluid. By demonstrating ACL and other extrameniscal lesions, MR imaging may help clarify the mechanisms of knee injury.

The authors report their experience with magnetic resonance imaging (MR) in 20 patients with CT-proved lymphadenopathy of the abdomen and pelvis. Signal intensity from abnormal nodes was intermediate between fat and muscle (T1 longer than fat and muscle; T2 equal to or slightly shorter than fat but longer than muscle) and was shown best in the spin-echo mode with a TR of 900 msec. and a TE of 30 msec. Abnormal nodes were identified by their increased size and other morphological changes rather than by tissue characteristics established by signal intensity information. Lymphadenopathy due to lymphoma could not be differentiated from that due to metastases. MR detected all lymphadenopathy originally documented by CT and in addition easily distinguished blood vessels from lymph nodes; on the other hand, CT offered better spatial resolution and could be completed in less time.

Magnetic resonance (MR) imaging of the scrotum with a high-resolution surface coil was performed in ten healthy volunteers and 20 patients with scrotal abnormalities demonstrated by high-resolution real-time ultrasound (US). Four patients had an abnormal testis (two tumors, one cyst, one testicular atrophy), and 16 patients had extratesticular abnormalities (four hydroceles, five epididymal cysts, one hernia, and six cases of epididymitis). The normal structures of the scrotum were depicted clearly on MR images. In all cases, the tunica albuginea was easily differentiated from the testis and epididymis. MR imaging enabled one to distinguish intratesticular from extratesticular lesions and to determine whether a lesion was solid or cystic. Complicated and simple fluid collections could also be differentiated. In general, MR imaging and US scanning provided similar information. A potential advantage of MR imaging is in the evaluation of patients with painful scrotal lesions that may limit US evaluation.

With the advent of computed tomography, lymphomatous involvement of sites other than lymph nodes is being seen with increasing frequency. Review of computed tomographic scans in 400 patients with newly diagnosed or recurrent non-Hodgkin lymphoma revealed 37 patients to have involvement of 56 unusual sites below the diaphragm: psoas/iliacus muscle (16 patients), kidney (13 patients), pancreas (5 patients), adrenal (4 patients), skin/subcutaneous tissue (4 patients), abdominal wall musculature (4 patients), peritoneum (4 patients), omentum (3 patients), and female reproductive tract (3 patients). These were mostly seen in patients with lymphomas of diffuse architecture, especially diffuse histiocytic lymphoma. Concomitant retroperitoneal and/or mesenteric adenopathy was very common; extranodal involvement was rarely the only site of initial or recurrent lymphoma.

Magnetic resonance (MR) studies were performed on 20 healthy volunteers and 41 patients with proved cervical and uterine neoplasms. MR imaging demonstrated normal uterine landmarks in all patients. On T2-weighted images, the normal uterine wall could be differentiated into three distinct layers: a central high-intensity zone, a junctional low-intensity band, and a peripheral medium-intensity area. While most of the normal cervices had only two distinct zones (central high-intensity zone and peripheral low-intensity zone), a small percentage had three layers of signal intensity, similar to the uterine body. Primary cervical and uterine neoplasms could be identified on MR images. In 18 of 22 patients with proved carcinoma, a mass with a signal intensity higher than that of normal cervical lips was seen on T2-weighted images. Endometrial carcinoma was most often identified as expansion of the central high-intensity area; discrete tumor nodules were visible in nine of 15 patients. Mixed müllerian sarcoma appeared as a large pelvic mass with complete obliteration of normal uterine landmarks. MR imaging delineates primary cervical and endometrial carcinoma better than computed tomography does.

The mediastinal vessels in 27 patients who had previously undergone vascular surgery were evaluated with use of magnetic resonance (MR) imaging for patency and the presence of stenosis. Thirty-six postoperative vessels or shunts were studied, including 11 aortas after coarctation or stenosis repair, 18 systemic-pulmonary shunts, four pulmonary artery or vein repairs, and three spiral vein grafts. Angiographic (n = 21), direct surgical (n = 1), contrast material-enhanced computed tomography (n = 1), or clinical (n = 13) findings were correlated with the MR findings. Overall, the accuracy of MR imaging for determination of patency was 100%. In patients with angiographic correlation and technically adequate MR examinations, MR imaging results were used to identify or exclude stenoses correctly in 66.7% of patent vessels. The four missed or underestimated stenoses occurred in vessels 10 mm or less in diameter.