Three-dimensional MR imaging in the assessment of physeal growth arrest

The purpose of this study is to describe an imaging method for identifying and characterising physeal growth arrest following physeal plate aggression. The authors describe the use of three-dimensional MRI performed with fat-suppressed three-dimensional spoiled gradient-recalled echo sequences followed by manual image reconstruction to create a 3D model of the physeal plate. This retrospective series reports the analysis of 33 bony physeal bridges in 28 children (mean age 10.5 years) with the use of fat-suppressed three-dimensional spoiled gradient-recalled echo imaging and 3D reconstructions from the source images. 3D reconstructions were obtained after the outlining was done manually on each source image. Files of all patients were reviewed for clinical data at the time of MRI, type of injury, age at MRI and bone bridge characteristics on reconstructions. Twenty-one (63%) of the 33 bridges were post-traumatic and were mostly situated in the lower extremities (19/21). The distal tibia was involved in 66% (14/21) of the cases. Bridges due to causes other than trauma were located in the lower extremities in 10/12 cases, and the distal femur represented 60% of these cases. Of the 28 patients, five presented with two bridges involving two different growth plates making a total of 33 physeal bone bars. The location and shape of each bridge was accurately identified in each patient, and in post-traumatic cases, 89% of bone bars were of Ogden type III (central) or I (peripheral). Reconstructions were obtained in 15 min and are easy to interpret. Volumes of the physeal bone bridge(s) and of the remaining normal physis were calculated. The bone bridging represented less than 1% to 47% of the total physeal plate volume. The precise shape and location of the bridge can be visualised on the 3D reconstructions. This information is useful in the surgical management of these deformities; as for the eight patients who underwent bone bar resection, an excellent correspondence was found by the treating surgeon between the MRI 3D model and the per-operative findings. Accurate 3D mapping obtained after manual reconstruction can also visualise very small physeal plates and bridges such as in cases of finger physeal disorders. MR imaging with fat-suppressed three-dimensional spoiled gradient-recalled echo sequences can be used to identify patterns of physeal growth arrest. 3D reconstructions can be obtained from the manual outlining of source images to provide an accurate representation of the bony bridge that can be a guide during surgical management.     

Internal derangement of the knee after ipsilateral femoral shaft fracture: MR imaging findings

Objective. This study uses magnetic resonance (MR) imaging to delineate the types and frequencies of injuries seen in the knee after ipsilateral femoral shaft fracture. We also compare the results of the orthopedic knee examination with the MR findings. Design and patients. MR imaging of the ipsilateral knee was performed on 34 patients with closed femoral shaft fractures. Indications for knee MR imaging included knee pain at the time of fracture, soft tissue swelling or an effusion of the knee, or a positive knee examination under anesthesia. The patients had a mean age of 27 years and all were stabilized with intramedullary nails. Imaging was performed a mean time of 2.5 days after surgery. All patients had knee examinations done under anesthesia, and the MR results were compiled and compared with the clinical examinations. Results. Ninety-seven percent of patients demonstrated knee effusions. Twenty-seven percent of patients demonstrated meniscal tears, with the posterior horn of the medial meniscus most frequently torn. The medial collateral ligament was the most frequent site of ligamentous injury (38%) followed by the posterior cruciate ligament (21%). Fifty percent of patients had injuries of the extensor mechanism. Bone bruises were noted in 32% of patients. Articular cartilage injuries were confined to the patella in four cases. One occult tibial plateau fracture and one meniscocapsular separation were seen. Conclusions. There is a common incidence of both ligamentous and meniscal injury to the knee after ipsilateral femoral shaft fracture. MR imaging can be useful in assessing the extent of injury, and may reveal findings unsuspected after clinical examination of the knee.     

3D 1 HMRSI对外周带前列腺癌鉴别诊断价值的初步研究

目的:探讨3D1HMRSI(three dimensional proton MR spectroscopic imaging)研究外周带前列腺癌(PCa)的代谢特点,及对外周带PCa的鉴别诊断价值。方法:分析18例良性前列腺增生(BPH)和21例PCa的3D MRSI表现。测量外周带PCa体素及BPH外周带体素的(Cho Cre)/Cit及Cho/Cre,并比较它们之间的差异。结果:BPH的外周带波谱主要表现为Cit高耸,Cho次之,Cre最低;PCa波谱主要表现为Cit峰明显下降甚至消失,Cho显著升高,两者波峰可呈现倒置,(Cho Cre)/Cit>0.86。PCa体素的(Cho Cre)/Cit显著高于BPH体素,且两者仅存在小部分重叠,PCa体素的Cho/Cre也显著高于BPH体素,但两者存在较大部分重叠;经ROC曲线分析得出(Cho Cre)/Cit对外周带PCa的鉴别诊断效能高于Cho/Cre,(Cho Cre)/Cit对外周带PCa的最佳诊断阈值为1.16。结论:外周带PCa的3D MRSI表现具有特征性,(Cho Cre)/Cit为诊断外周带PCa的有效指标,以(Cho Cre)/Cit>1.16作为诊断标准,可获得较高的敏感度、特异度和准确度(分别为99.3%、99.2%、99.3%)。     

Clinical comparison of three-dimensional MP-RAGE and FLASH techniques for MR imaging of the head.

Three-dimensional (3D) MP-RAGE (magnetization-prepared rapid gradient-echo) imaging was evaluated as a high-resolution 3D T1-weighted brain imaging technique for patients with suspected neurologic disease. Fourteen patients were studied. In five, 3D MP-RAGE images were compared with 3D FLASH (fast low-angle shot) images. Signal difference--to-noise ratios and T1 contrast were not statistically different for 3D MP-RAGE images as opposed to 3D FLASH images. Advantages intrinsic to the application of 3D MP-RAGE sequences include decreased imaging time and decreased motion artifact. With this technique, it is possible to perform a relatively motion-insensitive, T1-weighted screening brain study with voxel resolution of 1.0 x 1.4 x 2.0 mm or smaller, in an imaging time of 5.9 minutes or less--permitting offline (poststudy) reconstruction of high-resolution images in any desired plane.     

Detection of hepatocellular carcinoma: comparison of ferumoxides-enhanced and gadolinium-enhanced dynamic three-dimensional volume interpolated breath-hold MR imaging

The purpose was to compare the diagnostic accuracy of ferumoxides-enhanced MR imaging and gadolinium-enhanced dynamic MR imaging using three-dimensional (3D) volume interpolated breath-hold examination (VIBE) for the detection of hepatocellular carcinoma (HCC). Forty-nine patients with 61 HCCs, who underwent ferumoxides-enhanced and gadolinium-enhanced dynamic MR imaging, were included prospectively in this study. Ferumoxides-enhanced MR imaging was performed 24 h after completion of the dynamic study using 3D-VIBE. Three radiologists independently interpreted the images. The diagnostic accuracy was evaluated using the receiver-operating characteristic method, and the sensitivity of each imaging technique was compared using McNemars test. The mean diagnostic accuracy of dynamic MR imaging (Az=0.95) was higher than that of ferumoxides-enhanced MR imaging (Az=0.90), but failed to reach a statistical significance (P=0.057). The mean sensitivity of dynamic MR imaging (90.7%) was significantly superior to that of ferumoxides-enhanced MR imaging (80.9%, P=0.03). Furthermore, for lesions smaller than 15 mm, the mean sensitivity of dynamic MR imaging was significantly higher than that of ferumoxides-enhanced MR imaging (85.2% vs. 69.2%, P<0.05). Dynamic MR imaging showed a trend toward better diagnostic accuracy for than ferumoxides-enhanced MR imaging for the detection of HCCs.     

3D dynamic pituitary MR imaging with CAIPIRINHA: Initial experience and comparison with 2D dynamic MR imaging

Objectives

To evaluate the validity of 3D dynamic pituitary MR imaging with controlled aliasing in parallel imaging results in higher acceleration (CAIPIRINHA), with special emphasis on demarcation of pituitary posterior lobe and stalk.

Methods

Participants comprised 32 patients who underwent dynamic pituitary MR imaging due to pituitary or parasellar lesions. 3D dynamic MR with CAIPIRINHA was performed at 3 T with 20-s-interval, precontrast, 1st to 5th dynamic images. Normalized values and enhanced ratios (dynamic postcontrast image values divided by precontrast ones) were compared between 3D and 2D dynamic MR imaging for patients with visual identification of posterior lobe and stalk.

Results

In 3D, stalk was identified in 29 patients and unidentified in 3, and posterior lobe was identified in 28 and unidentified in 4. In 2D, stalk was identified in 26 patients and unidentified in 6 patients, and posterior lobe was identified in 15 and unidentified in 17. Normalized values of pituitary posterior lobe and stalk were higher in 3D than 2D (P < 0.001). No significant difference in enhancement ratio was seen between 3D and 2D.

Conclusions

3D dynamic pituitary MR provided better identification and higher normalized values of pituitary posterior lobe and stalk than 2D.     

MR imaging of knee hyaline cartilage: Evaluation of two- and three-dimensional sequences

Magnetic resonance (MR) imaging of cadaveric knees was performed to determine optimal sequences for visualization of hyaline cartilage. Full-thickness cartilage lesions ranging in diameter from 1 to 5 mm and a partial-thickness cartilage lesion 15 mm in diameter were created in the femoral articular surfaces of three cadaveric knees. The knees were then imaged with a 1.5-T imager with various two-dimensional and high-resolution three-dimensional (3D) techniques. After imaging, the knee specimens were sectioned for evaluation. Measurements of cartilage thickness in fast spin-echo images correlated best with those in the gross specimen. Diffuse areas of cartilage thinning were also most accurately identified with fast spinecho images. Small, focal cartilage defects were best delineated in 3D SPGR (spoiled GRASS [gradient-recalled acquisition in the steady state]) images.     

Experimental hyaline cartilage lesions: two-dimensional spin-echo versus three-dimensional gradient-echo MR imaging.

The value of magnetic resonance (MR) imaging, with two-dimensional (2D) spin-echo and FISP (fast imaging with steady-state precession) and FLASH (fast low-angle shot) three-dimensional (3D) gradient-echo sequences, for the detection of hyaline cartilage defects of the femoral condyle and the tibial plateau, was investigated in an animal model. In eight dogs, the anterior cruciate ligament was transected in one knee joint, resulting in rapid development of osteoarthritis with degeneration of the hyaline cartilage. At autopsy, 24 cartilage lesions were found, which were classified into four grades. The overall detection of cartilage lesions with MR imaging was poor. Only five of the 24 lesions were visible on 2D spin-echo images, while 11 of 24 were visible on 3D FISP images and 15 of 24 were seen on 3D FLASH images. The best results were obtained in advanced stages of cartilage degeneration, involving ulceration and complete abrasion of the cartilage layer. Signal loss or signal intensity increase in the cartilage layer was seen inconsistently in grades 3 and 4 degeneration. In this animal model, 2D spin-echo imaging was inadequate for the diagnosis of hyaline cartilage lesions, while 3D gradient-echo imaging permitted satisfactory diagnosis in only grade 4 cartilage disease.     

Pancreatic adenocarcinoma: combination of MR imaging, MR angiography and MR cholangiopancreatography for the diagnosis and assessment of resectability

The purpose of this study was to determine the possibility of integrating MR cholangiopancreatography (MRCP) and MR angiography (MRA) to conventional MR images in the diagnosis and assessment of resectability of pancreatic adenocarcinoma. Twenty-three patients with pancreatic adenocarcinoma were prospectively examined with MR. Conventional MR images were acquired in all patients. Three-dimensional MRCP and MRA images were acquired in all patients with suspected biliary and vascular involvement. Acquisition time was less than 45 min in all cases. Images were independently evaluated by two radiologists, with final reading decided by consensus among readers. Diagnosis was confirmed with surgery in 16 patients and with percutaneous biopsy in 7. Concordance among readers was high with a kappa value of 0.83. Pancreatic adenocarcinoma was observed in all patients. Correct assessment of unresectability due to vascular involvement was found in 22 of 23 patients. Biliary obstruction was evident in 13 patients, involving the biliary and pancreatic ducts in 9 and the biliary ducts only in 4. Technical advances permit extensive use of MRI in the evaluation of abdominal pathologies. The combination of MR imaging, MRCP, and MRA can provide sufficient information for the diagnosis and assessment of resectability of pancreatic adenocarcinoma, which otherwise would require three different exams. Received 22 August 1996; Revision received 3 June 1997; Accepted 19 August 1997     

Coarctation of the aorta: comparison of aortic dimensions between conventional MR imaging, 3D MR angiography,and conventional angiography

Magnetic resonance angiography is increasingly used as a non-invasive method in the evaluation of coarctation of the aorta. The aim of this study was to compare aortic dimensions calculated by MR angiography and those obtained by more conventional MR sequences and conventional angiography. Twenty-six consecutive patients with coarctation underwent three-dimensional MR angiography. Two independent observers retrospectively evaluated three aortic segments, site of coarctation, presence of aneurysm and existence of collateral circulation. Three aortic segments were also compared with those obtained on classical MR sequences and conventional angiography. The MR angiography was successfully performed in all showing 1 aneurysm and collateral circulation in 8 patients. Almost perfect intraobserver ( r(2)>0.91) and excellent interobserver ( r(2)>0.80) reliabilities were obtained for each aortic segment no matter which MR sequence was employed. Similarly, mainly excellent ( r(2)>0.80) concordance analysis was observed between MR angiography measurements and those calculated by either spin-echo/gradient-echo sequences or conventional angiography. This study demonstrates that MR angiography is a fast, accurate and reproducible method in the evaluation of coarctation of the aorta. It provides excellent anatomic information and reliably detects collateral vessels. Magnetic resonance angiography could probably replace the conventional angiography and will provide an additional diagnostic value in combination with turbo spin-echo sequence.     

The role of MR imaging in staging femoral head osteonecrosis

Reliability, accuracy and prognostic value of any classification system are important in evaluation and treatment of femoral head osteonecrosis. The purpose of the present study was to correlate the plain radiographs with MRI in femoral head osteonecrosis. Between 2000 and 2005, 115 hips (72 patients) were evaluated and classified according to the ARCO classification criteria with the use of plain radiographs and additional application of MRI. Classification was performed by consensus between a musculoskeletal radiologist and an orthopaedic surgeon. Sensitivity (SEN), specificity (SP), positive (PPV) and negative (NPV) predictive value of X-rays were estimated. According to MRI, 17 hips were classified as stage I, 25 as stage II, 48 as stage III and 25 as stage IV. The SEN, SP, PPV and NPV of plain radiographs were for stage II 88%, 90.5%, 78.6% and 95%; for stage III 79.2% 82%, 80.8% and 87.2%; for stage IV 76%, 100%, 100% and 90.9%, respectively. The agreement between plain radiographs and MRI was 80.6% for staging the disease, 71.2% for recording the location of the osteonecrotic lesion, 67.1% for evaluating the size of the lesion, 79.2% for the presence of collapse of the articular surface and 56.3% for the degree of collapse. In conclusion, the ARCO classification could miss important information in stages II and III, where treatment aims at preservation of the hip joint integrity. The results of the present study suggest that MRI should be incorporated in the classification of osteonecrosis (stages II and III), to add accuracy and prognostic value.     

Unbiased and efficient estimation of bladder volume with MR imaging

Almost all methods of measuring residual volume of urine in the bladder of patients undergoing urologic assessment are biased to an unknown extent. The authors describe the application of an unbiased stereologic technique for estimating the volume of bladder urine known as the Cavalieri method. The method requires imaging of a series of systematic (ie, equally spaced) parallel sections through the bladder. Such data can be conveniently obtained with magnetic resonance (MR) imaging. If sampling begins at a position randomly chosen within the distance corresponding to the section interval. bladder volume is estimated without bias as the sum of the areas of the bladder sections on the images multiplied by the section interval. Computer-aided point-counting techniques represent an efficient means of obtaining the required section area estimates. Optimum sectioning and point counting densities for estimating bladder volume were established by analyzing detailed data sets obtained in five volunteers. It was shown that if an average of only 20 points were counted in each of only five systematic sections through the bladder, the volume of bladder urine was estimated with a coefficient of error of about 5%. By studying these five volunteers and an additional 13 with MR imaging and the Cavalieri method, the authors showed that the difference between the volume of urine in the bladder before and after micturition is unbiased (ie, shows no systematic difference) with respect to the volume of urine voided by the subjects.     

Prenatal MR imaging of Dandy-Walker complex: midline sagittal area analysis

Objective

To measure the mid-sagittal areas of vermis (VA) and of posterior fossa (PFA) and determine their differences among fetuses with various Dandy-Walker (DW) entities and control subjects.

Methods

We reviewed data in 25 fetal patients with a MR diagnosis of DW complex including hypoplastic vermis (HV), HV with rotation (HVR), and mega cistern magna (MCM), and in 85 fetal controls with normal CNS. PFA and VA of each subject were manually traced on mid-sagittal MR images. Regarding each of VA and PFA, after age correction, we determined statistically significant differences among HVR, HV, MCM, and control groups.

Results

The mean VA residue of MCM was greater than that of the control, which was in turn greater than those of HVR and HV. The mean PF residue of the control was smaller than all other groups.

Conclusion

Fetuses with HVR or HV had smaller VA than fetuses with MCM or control subjects. Fetuses with MCM, HVR, or HV had larger PFA than control subjects. These results may be an early step leading to better understanding of the confusion about the PF anomalies in future.     

正常乳猪骨骺及干骺端骨髓的增强MR成像研究

目的 :探讨正常乳猪骨骺及干骺端骨髓的增强MR表现特征及不同年龄阶段的骨髓转化特征。方法 :8头乳猪 ,分别在出生 2、4、6和 8周的时间对双侧股骨近段及远段进行增强前后T1WIMR扫描 ,分析不同时间骨骺二次骨化中心骨髓和干骺端骨髓增强MR表现特征及增强率的变化 ,并与相应组织学发现进行对照研究。结果 :干骺端骨髓和骨骺二次骨化中心周边骨髓为含有丰富血管的红骨髓 ,增强后呈显著强化 ,且干骺端增强率较骨骺骨髓更高 ,但随年龄增长 ,两者的增强率均逐渐减低。骨干中央和二次骨化中心中央为黄骨髓 ,强化不明显。结论 :Gd DTPA增强在干骺端红骨髓与二次骨化中心周边红骨髓比在骨骺二次骨化中心中央区域黄骨髓强化更明显 ,其强化程度随年龄的增长逐渐减弱。     

MR imaging of tibial collateral ligament injury: comparison with clinical examination

We sought to clarify the capability of routine magnetic resonance (MR) imaging of the knee to detect and grade tibial collateral ligament (TCL) injury. We also wanted to define the exact MR findings that were the most important for diagnosis. Sixty-three patients were clinically evaluated for TCL injury prior to MR imaging of the knee on a 1.5-tesla system. MR studies were scored with respect to multiple direct and indirect findings. By discriminant analysis, the overall accuracy of classification (into clinical grades 1, 2, 3 and normal) on the basis of MR was 65%. Accuracy for detection of abnormal TCLs was 87%. The most useful MR findings for detection and grading of TCL injury were direct signs: subcutaneous edema over the TCL, focal disruption of the TCL, increased signal in the TCL on T2-weighted images, and longitudinal striations within the TCL. Indirect signs such as trabecular trauma were less useful for TCL evaluation. The presence of a tear of the anterior cruciate ligament, however, correlated with TCL injury of a higher grade. T2-weighted coronal images, by better depicting intraligamentous edema and fiber disruption, improved the MR assessment of the TCL.     

Gender differences in knee joint cartilage thickness, volume and articular surface areas: assessment with quantitative three-dimensional MR imaging

Objective: To compare the cartilage thickness, volume, and articular surface areas of the knee joint between young healthy, non-athletic female and male individuals. Subjects and design. MR imaging was performed in 18 healthy subjects without local or systemic joint disease (9 female, age 22.3±2.4 years, and 9 male, age 22.2±1.9 years.), using a fat-suppressed FLASH 3D pulse sequence (TR=41 ms, TE=11 ms, FA=30°) with sagittal orien- tation and a spatial resolution of 2×0.31×0.31 mm³. After three- dimensional reconstruction and triangulation of the knee joint cartilage plates, the cartilage thickness (mean and maximal), volume, and size of the articular surface area were quantified, independent of the original section orientation. Results and conclusions: Women displayed smaller cartilage volumes than men, the percentage difference ranging from 19.9% in the patella, to 46.6% in the medial tibia. The gender differences of the cartilage thickness were smaller, ranging from 2.0% in the femoral trochlea to 13.3% in the medial tibia for the mean thickness, and from 4.3% in the medial femoral condyle to 18.3% in the medial tibia for the maximal cartilage thickness. The differences between the cartilage surface areas were similar to those of the volumes, with values ranging from 21.0% in the femur to 33.4% in the lateral tibia. Gender differences could be reduced for cartilage volume and surface area when normalized to body weight and body weight×body height. The study demonstrates significant gender differences in cartilage volume and surface area of men and women, which need to be taken into account when retrospectively estimating articular cartilage loss in patients with symptoms of degenerative joint disease. Differences in cartilage volume are primarily due to differences in joint surface areas (epiphyseal bone size), not to differences in cartilage thickness. Received: 19 June 2000 Revision requested: 4 August 2000 Revision received: 30 November 2000 Accepted: 6 December 2000     

van der Knaap syndrome: MR imaging findings including FLAIR, diffusion imaging, and proton MR spectroscopy

A patient is reported with diffuse leukoencephalopathy associated with cystic degeneration of the white matter of the brain (van der Knaap syndrome). The changes were studied by fluid attenuated inversion recovery (FLAIR), and diffusion-weighted MR imaging. The FLAIR sequence revealed suppressed signal of the cysts, and widespread high-signal white matter changes associated with thinned cortices. On diffusion-weighted MR imaging, apparent diffusion coefficient (ADC) values ranged from 3.0 × 10^–3 to 2.7 × 10^–3 mm²/s in the temporal cysts, similar to that of CSF. The ADC values within the parenchyma ranged between 2 × 10^–3 and 2.1 × 10^–3 mm²/s, a value falling between normal parenchyma and cerebrospinal fluid, compared with a control group of three healthy subjects. The changes were also evaluated by proton MR spectroscopy, and were compared with a control group of 12 cases. Magnetic resonance spectroscopy revealed apparently increased NAA/Cr ratios in most parts of the brain. The NAA/Cho ratios were either high or low, and the Cho/Cr ratios were increased or normal in different regions. Received: 27 October 1999; Revised: 9 December 1999; Accepted: 20 December 1999     

T2-weighted MR imaging of the liver: Qualitative and quantitative comparison of SPACE MR imaging with turbo spin-echo MR imaging

Objective

To qualitatively and quantitatively compare T2-weighted MR imaging of the liver using volumetric spin-echo with sampling perfection with application-optimized contrast using different flip angle evolutions (SPACE) with conventional turbo spin-echo (TSE) sequence for fat-suppressed T2-weighted MR imaging of the liver.

Materials and methods

Thirty-three patients with suspected focal liver lesions had SPACE MR imaging and conventional fat-suppressed TSE MR imaging. Images were analyzed quantitatively by measuring the lesion-to-liver contrast-to-noise ratio (CNR), and the signal-to-noise ratio (SNR) of main focal hepatic lesions, hepatic and splenic parenchyma and qualitatively by evaluating the presence of vascular, respiratory motion and cardiac artifacts. Wilcoxon signed rank test was used to search for differences between the two sequences.

Results

SPACE MR imaging showed significantly greater CNR for focal liver lesions (median = 22.82) than TSE MR imaging (median = 14.15) (P < .001). No differences were found for SNR of hepatic parenchyma (P = .097), main focal hepatic lesions (P = .35), and splenic parenchyma (P = .25). SPACE sequence showed less artifacts than TSE sequence (vascular, P < .001; respiratory motion, P < .001; cardiac, P < .001) but needed a longer acquisition time (228.4 vs. 162.1 s; P < .001).

Conclusion

SPACE MR imaging provides a significantly increased CNR for focal liver lesions and less artifacts by comparison with the conventional TSE sequence. These results should stimulate further clinical studies with a surgical standard of reference to compare the two techniques in terms of sensitivity for malignant lesions.