HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Masi, J. N. Articles by Link, T. M. Search for Related Content PubMed PubMed Citation Articles by Masi, J. N. Articles by Link, T. M.

Cartilage MR Imaging at 3.0 versus That at 1.5 T: Preliminary Results in a Porcine Model¹

¹ From the Department of Radiology, University of California, San Francisco, 400 Parnassus Ave, A 367, Box 0628, San Francisco, CA 94143-0628 (J.N.M., C.A.S., C.P., D.N., L.S., S.M., T.M.L.); and GE Medical Systems, Milwaukee, Wis (E.H.). Received April 24, 2004; revision requested July 12; revision received August 5; accepted September 2. Address correspondence to T.M.L. (e-mail: tmlink{at}radiology.ucsf.edu).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
PURPOSE: To compare 1.5- and 3.0-T magnetic resonance (MR) images of porcine knee specimens containing artificial cartilage lesions in terms of accuracy of lesion depiction, image quality, and signal-to-noise ratio (SNR).

MATERIALS AND METHODS: This Health Insurance Portability and Accountability Act–compliant study had institutional review board approval, and informed consent was obtained from the human volunteers. Two fat-saturated cartilage MR imaging sequences (an intermediate-weighted fast spin-echo [SE] sequence and a spoiled gradient-echo [GRE] sequence) were optimized for imaging at 3.0 T in two human volunteers and then used to image 10 porcine knees in which 29 artificial cartilage lesions had been created. Corresponding sequences were used at 1.5 T for all specimens. Images were assessed by two radiologists in consensus, and diagnostic performance in lesion depiction was determined by using macroscopic findings in specimen slices as a reference standard. SNRs were also calculated. For statistical analysis, the McNemar test of discordant pairs was used with a level of significance of P < .05.

RESULTS: The best diagnostic performance for both the intermediate-weighted fast SE and the spoiled GRE sequences was achieved at 3.0 T. With use of corresponding fat-saturated intermediate-weighted fast SE sequences with an identical acquisition time (9 minutes 44 seconds), 26 (90%) of 29 lesions were detected at 3.0 T, while 18 (62%) were detected at 1.5 T. With use of fat-saturated spoiled GRE sequences, 24 (83%) of 29 lesions were detected at 3.0 T (acquisition time, 8 minutes 48 seconds), and 23 (79%) lesions were detected at 1.5 T (acquisition time, 11 minutes 14 seconds). The rate of correct lesion grade assessment was 65% (17 of 26 lesions) at 3.0 T and 61% (11 of 18 lesions) at 1.5 T with the intermediate-weighted fast SE sequences and 83% (20 of 24 lesions) at 3.0 T and 70% (16 of 23 lesions) at 1.5 T with the spoiled GRE sequences. Both subjective evaluation of image quality and SNR values were significantly higher at 3.0 T (P < .05).

CONCLUSION: In this animal model, MR imaging at 3.0 T increased the accuracy of cartilage lesion assessment when compared with imaging at 1.5 T. Image quality and SNR were highest at 3.0 T.

© RSNA, 2005

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
Therapy of focal cartilage lesions has changed dramatically in recent years, with new articular cartilage repair options becoming increasingly used. Autologous chondrocyte implantation is currently used for articular cartilage repair (1–5), and osteochondral autografting (so-called mosaicplasty) has shown positive results in restoring articular surfaces and preventing the onset of early arthritis in young patients (1–8). Microfracture of subchondral bone, cartilage débridement, and drilling also represent viable treatment possibilities for patients with cartilage defects (2,9). In addition, investigational procedures such as growth factor– and gene transfection–based techniques have shown encouraging treatment potential (10).

If available therapies are to be optimally used and the management of cartilage defects is to be improved, accurate diagnostic procedures are required. Arthroscopy, while considered the reference standard for this application, is invasive, time consuming, and expensive. Magnetic resonance (MR) imaging, on the other hand, is noninvasive and has been shown to be a sensitive and accurate tool for depicting cartilage morphologic features (11–18).

Unfortunately, MR imaging at 1.5 T has limitations in the detection of subtle morphologic changes and lesions. The high spatial resolution and image contrast required for the depiction of subtle structural changes and early lesions often exceed those of currently used MR imaging units (12,19–23). Although full- and partial-thickness defects are sufficiently well diagnosed at 1.5 T, the depiction of surface fibrillation and small abnormalities at 1.5 T remains limited (21,24,25). Even current MR imaging techniques that are specialized for cartilage do not reliably depict early cartilage damage (26). In addition, the clinical usefulness of MR imaging is currently restricted by the long time it takes to obtain the high-spatial-resolution images necessary for adequate cartilage depiction (27). Therefore, to optimize the diagnosis of subtle cartilage changes, further improvements in MR imaging techniques are needed.

Enhancements in spatial resolution, signal-to-noise ratio (SNR), and image contrast, along with the use of optimized high-spatial-resolution sequences are likely to markedly enhance the diagnostic capabilities of MR imaging for cartilage lesions. SNR-efficient sequences, in particular, yield markedly improved contrast between cartilage and synovial fluid (20). The use of higher magnetic field strengths such as 3.0 T may also improve imaging of cartilage because use of such field strengths increases SNR and spatial resolution. At present, to our knowledge, results of studies in which the potential of 3.0-T MR imaging for depicting cartilage lesions was compared with the potential of 1.5-T MR imaging have not been published. In addition, MR imaging parameters for articular cartilage evaluation at 1.5 T have been extensively studied, but a paucity of data exists regarding optimal parameters for imaging cartilage at 3.0 T. The purpose of this study was to compare 1.5- and 3.0-T MR images of porcine knee specimens containing artificial cartilage lesions in terms of accuracy of lesion depiction, image quality, and SNR.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
Optimization Studies
Sequences for imaging cartilage were performed at 3.0 T in two human volunteers by using a Signa MR imaging unit (GE Medical Systems, Milwaukee, Wis) and a standard quadrature knee coil (GE Medical Systems). The 3.0-T system was equipped with 4 x 10^–4 T/cm gradients. The knees of the volunteers were placed in a supine, head-first orientation within the centers of the coils, which were lined up with the inferior margin of the patella during imaging. These examinations were performed after institutional review board approval was granted and informed consent was obtained from both volunteers. This investigation was compliant with the Health Insurance Portability and Accountability Act. Both volunteers were men (25 and 40 years of age) and had no subjective symptoms or clinical signs that would suggest acute injury or degenerative joint disease. The sequences to be used in the subsequent experimental study were chosen by two radiologists (T.M.L., who had 12 years of experience in musculoskeletal MR imaging, and C.P., who had 4 years of experience), who based their assessments on (a) image quality and (b) acquisition time considerations.

Image quality was assessed by using the following subjective criteria: (a) delineation of cartilage from joint fluid and subchondral bone, (b) visualization of bones and growth plates, (c) amount of noise, and (d) artifacts. The following two sequences were used: a fat-saturated intermediate-weighted fast spin-echo (SE) sequence and a fat-saturated spoiled gradient-echo (GRE) sequence. Image quality for both sequences was graded by both radiologists in consensus according to a four-level scale in which a score of 4 indicated excellent image quality; a score of 3, good image quality; a score of 2, satisfactory image quality; and a score of 1, poor image quality.

A standard fat-saturated intermediate-weighted fast SE sequence that was optimized for imaging at 1.5 T (repetition time msec/echo time msec, 2000/35; echo train length, four; bandwidth, 15.63 kHz; field of view, 10 cm; section thickness, 3 mm; number of signals acquired, two; matrix, 320 x 224; acquisition time, 3 minutes 48 seconds) was chosen initially. Parameters were modified to meet the following goals: (a) accurate depiction of cartilage for the purpose of assessing focal lesions, (b) clinically usable acquisition times, and (c) comparable sequences between 1.5 and 3.0 T with respect to imaging parameters and acquisition time. Repetition time was varied from 2000 to 5000 msec, followed by echo time, which was varied from 20 to 60 msec. Echo train length was varied from four to eight, section thickness was varied from 1.5 to 3.0 mm, and the number of signals acquired was varied from one to four. The field of view was kept constant at 14 cm.

Images acquired with the various sequences were analyzed with an emphasis on the contrast between cartilage and joint fluid, the clarity of cartilage layers and internal structure, and the edge sharpness of the articular cartilage surface. In addition to subjective image assessment and time considerations, the final selection of repetition time and bandwidth was also influenced by previous recommendations from Gold et al (28). A sequence with the following parameters was finally selected for imaging the pig knees at 1.5 and 3.0 T: 4000/35; echo train length, eight; bandwidth, 31.25 kHz; field of view, 10 cm; section thickness, 2 mm; number of signals acquired, three; matrix, 512 x 384; and acquisition time, 9 minutes 44 seconds.

The optimal fat-saturated spoiled GRE sequence was determined in a similar fashion. For imaging at 1.5 T, a standard spoiled GRE sequence was used with the following parameters: 32.9/16; flip angle, 30°; bandwidth, 15.63 kHz; field of view, 14 cm; section thickness, 1.5 mm; number of signals acquired, two; matrix, 512 x 256; and acquisition time, 11 minutes 14 seconds. Because spoiled GRE sequences with high spatial resolution usually involve long acquisition times, the aim was to find a 3.0-T sequence that had a fairly short imaging time relative to that for the 1.5-T sequence. So that we could create a sequence that had a shorter acquisition time, we varied repetition time from 32.9 to 15.0 msec and flip angle from 15.0° to 31.0°. On the basis of analysis of images acquired by using these parameters, a 3.0-T sequence with an acquisition time that was approximately 25% shorter than the acquisition time of the corresponding 1.5-T sequence was chosen for imaging the pig knees. The final parameters of this 3.0-T sequence were as follows: 22/10.6; flip angle, 30°; bandwidth, 15.63 kHz; field of view, 10 cm; section thickness, 1.5 mm; number of signals acquired, two; matrix, 512 x 256; and acquisition time, 8 minutes 48 seconds.

Animal Studies and Reference Standard
Ten fresh porcine knees were obtained from a local slaughterhouse. Specimens were frozen for storage and thawed to room temperature before the articular cartilage was exposed with a lateral peripatellar approach. The meniscus was incised to improve condylar access but preserved so as not to allow joint space reduction. Care was also taken to preserve the anterior cruciate ligaments, the posterior cruciate ligaments, and the patellar tendon in all procedures.

Cartilage lesions were then created by using a ceramic scalpel (Fine Science Tools, San Francisco, Calif) so that metal artifacts could be avoided. All experimental procedures were performed by two authors (J.N.M. and C.A.S.). Figure 1 depicts a specimen with two representative lesions in the patella and the femoral lateral condyle. The focal defects (n = 27) and fissures (n = 2) that were created simulated the morphologic features of in vivo cartilage defects depicted on arthroscopic images (29) and were graded according to a modified Noyes classification (30). Each knee was treated as a particular defect configuration, and creation of the 29 defects at the 50 joint surfaces was random. As is the case in vivo, more lesions were located at the femoral condyles (the larger joint surfaces) than at the patella and the tibia. Eight of the lesions were located at the patella; 15, at the femoral condyles; and six, at the tibia. We randomly created 50% full-thickness (extending to the bone surface) and 50% partial-thickness lesions.

View larger version (132K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Photograph of porcine knee specimen shows patellar (left) and femoral (right) cartilage surfaces. A patellar cartilage lesion (dashed arrow) and an anterior femoral condylar lesion (solid arrow) were created. Both lesions are grade 2 lesions in that they have depths that are greater than 50% of the cartilage thickness but do not expose the subchondral bone. With close inspection, the irregular borders of these lesions can be appreciated.

After lesion creation and measurement, the joints were filled with a mixture of two parts ultrasonographic gel (Parker Laboratories, Fairfield, NJ) and one part distilled water while we carefully attempted to remove remaining air. The signal intensity of this mixture was compared with that of synovial fluid on the knee MR images that were used to determine sequence parameters, and the two substances were found to have similar signal intensities. The knee was reassembled, with special attention paid to restoring the proper physiologic alignment of the articular surfaces and menisci. A transparent latex bag filled with the gel-and-water mixture was pulled over the knee from the tibial side up to and over the femoral head. The knee was flexed and extended as air was manipulated out of the mixture. Once no air was apparent within the bag, it was sealed at the femoral shaft by using micropore tape. Some fluid was allowed to escape from the bag during air removal and during the sealing process. The knees were wrapped in parafilm (SPI Supplies, West Chester, Pa) from the tibial shaft all the way up to the femoral shaft. Care was taken to manually keep the patella in its normal alignment during this step because the tissue that normally performs this function had been partially severed at surgery. Each knee was placed in a sealed and labeled plastic bag.

MR Imaging
In the human volunteers, imaging was performed as previously described in the section on optimization studies. Sequence parameters that yielded high image quality with a reasonable imaging time at 3.0 T were identified (Table 1). Increasing bandwidth and repetition time mildly improved the depiction of cartilage with the fat-saturated intermediate-weighted fast SE sequence. For this sequence, we therefore decided to use the same imaging parameters at 1.5 T and at 3.0 T to achieve most-comparable sequences. For the fat-saturated spoiled GRE sequence, repetition time was shortened at 3.0 T to reduce acquisition time; no substantial differences in image quality between a repetition time of 22 and one of 32.9 msec were observed.

Image Analysis
All four sequences were used to obtain images of the 10 porcine knee specimens, each of which contained one to four lesions. The images were reviewed in a random order at a picture archiving and communication system workstation (Agfa, Ridgefield Park, NJ) by two radiologists (C.P. and T.M.L., working in consensus) over the course of four reading sessions that occurred on separate days to prevent a learning bias. Neither radiologist had been involved in the experimental procedures, and both were blinded to the location of the lesions. All information pertaining to the specimen number, sequence protocol, and field strength was masked in each image. Images with excessive air artifacts owing to preparation difficulties were excluded.

The radiologists were asked to search for lesions in an approximated clinical scenario, with each radiologist having control over section selection and window leveling parameters. When a lesion was detected, each radiologist then assigned a grade indicating his or her level of confidence in lesion presence. A grade of 1 indicated that the radiologist was uncertain whether a lesion was present; a grade of 2, that a lesion was probably present; and a grade of 3, that a lesion was definitely present. Location had to be stated for each lesion, and in cases of discordance between the lesion location as determined by both radiologists and the true lesion location, radiologist determinations were rated as false-positive.

The radiologists then measured the anterior-to-posterior length of the lesions and scored lesion depth by using the previously described modified Noyes classification. In cases in which lesions were irregular and displayed different depths, the readers were asked to indicate the largest depth. Measurement accuracy was determined by using results of gross pathologic examination of the specimens as a standard of reference.

Additionally, image quality was scored by each radiologist according to a four-level scale in which a grade of 4 indicated excellent image quality and a grade of 1 indicated poor image quality. Images were given a score of 4 when cartilage was sharply delineated from joint fluid and subchondral bone, when bones and growth plates were clearly visualized, and when noise and artifacts were minimal. Images were given a score of 3 when one or two of the above-mentioned criteria were less than optimal (eg, mild artifacts were present); however, where cartilage lesions were concerned, the diagnostic quality of the images was not limited. A score of 2 corresponded to more substantial limitations that mildly affected the diagnostic quality of the images. Substantial limitations in diagnostic quality of the images, extensive artifacts, and noise resulted in a score of 1.

Finally, SNRs were calculated (by J.N.M.) for all images with the following equation: SNR = SI_cart/SD_bn, where SI_cart is the signal intensity of the cartilage and SD_bn is the standard deviation of the background noise. The corresponding regions of interest are shown in Figure 2. In an attempt to avoid partial volume effects, the region placed in the cartilage was 0.25–0.50 mm². Three central images per sequence were used for the measurement of signal intensity for each specimen. For each image, three regions of interest were placed within the articular cartilage: one on the patella, one on the anterior femoral condyle, and one on the posterior femoral condyle. One background region of interest was placed for each image at a point inferior to the patella, between the anterior femoral condyle and the edge of the image.

View larger version (133K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Sagittal fat-saturated spoiled GRE MR image (22/10.6; flip angle, 30°) depicts the regions of interest used for the calculation of the SNR measurements in the cartilage (small black circles with arrows) and background (large white circle).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
MR Imaging at 3.0 T versus That at 1.5 T
So that we could evaluate the performance between the individual sequences, we first compared their false-positive rates and found that they were low and comparable: For the fat-saturated intermediate-weighted fast SE sequence at each field strength, one probable lesion that was actually not created was detected (a false-positive lesion); for the fat-saturated spoiled GRE sequence at 1.5 T, one false-positive lesion was found, and for the spoiled GRE sequence at 3.0 T, two false-positive lesions were found. Therefore, our statistical analyses focused on the comparison of sensitivities.

An improvement in sensitivity of 28% (26 of 29 lesions detected at 3.0 T vs 18 of 29 lesions detected at 1.5 T) was observed for 3.0 T relative to 1.5 T with fat-saturated intermediate-weighted fast SE sequences that had identical acquisition times (Table 2). According to results of the McNemar test of discordant pairs, this difference was statistically significant (P < .008). Representative images are shown in Figures 3–5. Most of the lesions were clearly depicted at 3.0 T, but a number of them were not sufficiently visualized at 1.5 T (Figs 4, 5). The fat-saturated spoiled GRE sequence at 3.0 T was 25% shorter than that at 1.5 T but still yielded a 4% higher sensitivity (24 of 29 lesions detected at 3.0 T vs 23 of 29 lesions detected at 1.5 T); this difference in sensitivity, was, however, not significant (P > .05) (Figs 3, 4).

View larger version (151K): [in this window] [in a new window] [Download PPT slide]   Figure 3a. Representative sagittal MR images of grade 3 (full-thickness) lesion. The lesion is depicted on corresponding sections obtained at (a, c) 1.5 T and (b, d) 3.0 T with (a, b) fat-saturated intermediate-weighted fast SE (4000/35) and (c, d) fat-saturated spoiled GRE (32.9/16 with a flip angle of 30° [at 1.5 T] or 22/10.6 with a flip angle of 30° [at 3.0 T]) sequences. Note the differences in image quality between sequences at 1.5 and 3.0 T, with substantially lower noise at 3.0 T. Note also the apparent depth difference between c and d. In c, the lesion appears to be of less than full thickness, while in d, the disruption of cartilage all the way to the subchondral bone is evident, indicating a full-thickness (grade 3) lesion.

View larger version (126K): [in this window] [in a new window] [Download PPT slide]   Figure 3b. Representative sagittal MR images of grade 3 (full-thickness) lesion. The lesion is depicted on corresponding sections obtained at (a, c) 1.5 T and (b, d) 3.0 T with (a, b) fat-saturated intermediate-weighted fast SE (4000/35) and (c, d) fat-saturated spoiled GRE (32.9/16 with a flip angle of 30° [at 1.5 T] or 22/10.6 with a flip angle of 30° [at 3.0 T]) sequences. Note the differences in image quality between sequences at 1.5 and 3.0 T, with substantially lower noise at 3.0 T. Note also the apparent depth difference between c and d. In c, the lesion appears to be of less than full thickness, while in d, the disruption of cartilage all the way to the subchondral bone is evident, indicating a full-thickness (grade 3) lesion.

View larger version (156K): [in this window] [in a new window] [Download PPT slide]   Figure 3c. Representative sagittal MR images of grade 3 (full-thickness) lesion. The lesion is depicted on corresponding sections obtained at (a, c) 1.5 T and (b, d) 3.0 T with (a, b) fat-saturated intermediate-weighted fast SE (4000/35) and (c, d) fat-saturated spoiled GRE (32.9/16 with a flip angle of 30° [at 1.5 T] or 22/10.6 with a flip angle of 30° [at 3.0 T]) sequences. Note the differences in image quality between sequences at 1.5 and 3.0 T, with substantially lower noise at 3.0 T. Note also the apparent depth difference between c and d. In c, the lesion appears to be of less than full thickness, while in d, the disruption of cartilage all the way to the subchondral bone is evident, indicating a full-thickness (grade 3) lesion.

View larger version (154K): [in this window] [in a new window] [Download PPT slide]   Figure 3d. Representative sagittal MR images of grade 3 (full-thickness) lesion. The lesion is depicted on corresponding sections obtained at (a, c) 1.5 T and (b, d) 3.0 T with (a, b) fat-saturated intermediate-weighted fast SE (4000/35) and (c, d) fat-saturated spoiled GRE (32.9/16 with a flip angle of 30° [at 1.5 T] or 22/10.6 with a flip angle of 30° [at 3.0 T]) sequences. Note the differences in image quality between sequences at 1.5 and 3.0 T, with substantially lower noise at 3.0 T. Note also the apparent depth difference between c and d. In c, the lesion appears to be of less than full thickness, while in d, the disruption of cartilage all the way to the subchondral bone is evident, indicating a full-thickness (grade 3) lesion.

View larger version (167K): [in this window] [in a new window] [Download PPT slide]   Figure 4a. Sagittal MR images of a grade 2 lesion (arrow) in the tibia. The lesion is depicted on corresponding sections obtained at (a, c) 1.5 T and (b, d) 3.0 T with (a, b) fat-saturated intermediate-weighted fast SE (4000/35) and (c, d) fat-saturated spoiled GRE (32.9/16 with a flip angle of 30° [at 1.5 T] or 22/10.6 with a flip angle of 30° [at 3.0 T]) sequences. No lesion was diagnosed at the tibia with the intermediate-weighted fast SE sequence at 1.5 T, while a definite lesion was diagnosed at 3.0 T. Also note that although the lesion is visualized at both field strengths with the spoiled GRE sequences, the borders of the lesion are substantially better aligned at 3.0 T (d).

View larger version (194K): [in this window] [in a new window] [Download PPT slide]   Figure 4b. Sagittal MR images of a grade 2 lesion (arrow) in the tibia. The lesion is depicted on corresponding sections obtained at (a, c) 1.5 T and (b, d) 3.0 T with (a, b) fat-saturated intermediate-weighted fast SE (4000/35) and (c, d) fat-saturated spoiled GRE (32.9/16 with a flip angle of 30° [at 1.5 T] or 22/10.6 with a flip angle of 30° [at 3.0 T]) sequences. No lesion was diagnosed at the tibia with the intermediate-weighted fast SE sequence at 1.5 T, while a definite lesion was diagnosed at 3.0 T. Also note that although the lesion is visualized at both field strengths with the spoiled GRE sequences, the borders of the lesion are substantially better aligned at 3.0 T (d).

View larger version (179K): [in this window] [in a new window] [Download PPT slide]   Figure 4c. Sagittal MR images of a grade 2 lesion (arrow) in the tibia. The lesion is depicted on corresponding sections obtained at (a, c) 1.5 T and (b, d) 3.0 T with (a, b) fat-saturated intermediate-weighted fast SE (4000/35) and (c, d) fat-saturated spoiled GRE (32.9/16 with a flip angle of 30° [at 1.5 T] or 22/10.6 with a flip angle of 30° [at 3.0 T]) sequences. No lesion was diagnosed at the tibia with the intermediate-weighted fast SE sequence at 1.5 T, while a definite lesion was diagnosed at 3.0 T. Also note that although the lesion is visualized at both field strengths with the spoiled GRE sequences, the borders of the lesion are substantially better aligned at 3.0 T (d).

View larger version (190K): [in this window] [in a new window] [Download PPT slide]   Figure 4d. Sagittal MR images of a grade 2 lesion (arrow) in the tibia. The lesion is depicted on corresponding sections obtained at (a, c) 1.5 T and (b, d) 3.0 T with (a, b) fat-saturated intermediate-weighted fast SE (4000/35) and (c, d) fat-saturated spoiled GRE (32.9/16 with a flip angle of 30° [at 1.5 T] or 22/10.6 with a flip angle of 30° [at 3.0 T]) sequences. No lesion was diagnosed at the tibia with the intermediate-weighted fast SE sequence at 1.5 T, while a definite lesion was diagnosed at 3.0 T. Also note that although the lesion is visualized at both field strengths with the spoiled GRE sequences, the borders of the lesion are substantially better aligned at 3.0 T (d).

View larger version (160K): [in this window] [in a new window] [Download PPT slide]   Figure 5a. Sagittal fat-saturated intermediate-weighted fast SE MR images (4000/35) of a grade 3 lesion at (a) 1.5 T and (b) 3.0 T. The artificial cartilage lesion (arrow) created in the patella is not sufficiently visualized at 1.5 T (and was classified as uncertain), while it is well depicted at 3.0 T. Also note the higher SNR and better depiction of anatomic structures such as the growth plate (*) at 3.0 T.

View larger version (181K): [in this window] [in a new window] [Download PPT slide]   Figure 5b. Sagittal fat-saturated intermediate-weighted fast SE MR images (4000/35) of a grade 3 lesion at (a) 1.5 T and (b) 3.0 T. The artificial cartilage lesion (arrow) created in the patella is not sufficiently visualized at 1.5 T (and was classified as uncertain), while it is well depicted at 3.0 T. Also note the higher SNR and better depiction of anatomic structures such as the growth plate (*) at 3.0 T.

View larger version (16K): [in this window] [in a new window] [Download PPT slide]   Figure 6a. (a) Bar graph shows percentages of detected lesions at 1.5 and 3.0 T for each imaging sequence (Scan Type). FSE = fast SE, SPGR = spoiled GRE. Of the 29 lesions created, 18 were detected at 1.5 T and 26 were detected at 3.0 T with the fat-saturated intermediate-weighted fast SE sequence, while 23 were detected at 1.5 T and 24 were detected at 3.0 T with the fat-saturated spoiled GRE sequence. (b) Bar graph shows percentages of detected lesions at 1.5 T and 3.0 T for the fat-saturated intermediate-weighted fast SE sequence according to the grade of the cartilage lesion (grade 1 = less than 50% cartilage defect [n = 7], grade 2 = more than 50% cartilage defect [n = 7], grade 3 = full-thickness lesion [n = 15]). The grade 1, 2, and 3 lesions created were detected with frequencies of 29% (two of seven lesions), 86% (six of seven lesions), and 67% (10 of 15 lesions), respectively, while at 3.0 T detection sensitivity was higher at all grades, with frequencies of 71% (five of seven lesions), 100% (seven of seven lesions), and 93% (14 of 15 lesions), respectively. (c) Bar graph shows percentages of detected lesions at 1.5 T and 3.0 T for the fat-saturated spoiled GRE sequence according to the grade of the cartilage lesion. A sensitivity difference in lesion detection with this sequence only became apparent when grade 1 lesions were assessed, with three (43%) of seven grade 1 lesions detected at 1.5 T and four (57%) detected at 3.0 T. At both 1.5 and 3.0 T, six (86%) of seven grade 2 lesions and 14 (93%) of 15 grade 3 lesions were detected.

View larger version (17K): [in this window] [in a new window] [Download PPT slide]   Figure 6b. (a) Bar graph shows percentages of detected lesions at 1.5 and 3.0 T for each imaging sequence (Scan Type). FSE = fast SE, SPGR = spoiled GRE. Of the 29 lesions created, 18 were detected at 1.5 T and 26 were detected at 3.0 T with the fat-saturated intermediate-weighted fast SE sequence, while 23 were detected at 1.5 T and 24 were detected at 3.0 T with the fat-saturated spoiled GRE sequence. (b) Bar graph shows percentages of detected lesions at 1.5 T and 3.0 T for the fat-saturated intermediate-weighted fast SE sequence according to the grade of the cartilage lesion (grade 1 = less than 50% cartilage defect [n = 7], grade 2 = more than 50% cartilage defect [n = 7], grade 3 = full-thickness lesion [n = 15]). The grade 1, 2, and 3 lesions created were detected with frequencies of 29% (two of seven lesions), 86% (six of seven lesions), and 67% (10 of 15 lesions), respectively, while at 3.0 T detection sensitivity was higher at all grades, with frequencies of 71% (five of seven lesions), 100% (seven of seven lesions), and 93% (14 of 15 lesions), respectively. (c) Bar graph shows percentages of detected lesions at 1.5 T and 3.0 T for the fat-saturated spoiled GRE sequence according to the grade of the cartilage lesion. A sensitivity difference in lesion detection with this sequence only became apparent when grade 1 lesions were assessed, with three (43%) of seven grade 1 lesions detected at 1.5 T and four (57%) detected at 3.0 T. At both 1.5 and 3.0 T, six (86%) of seven grade 2 lesions and 14 (93%) of 15 grade 3 lesions were detected.

View larger version (18K): [in this window] [in a new window] [Download PPT slide]   Figure 6c. (a) Bar graph shows percentages of detected lesions at 1.5 and 3.0 T for each imaging sequence (Scan Type). FSE = fast SE, SPGR = spoiled GRE. Of the 29 lesions created, 18 were detected at 1.5 T and 26 were detected at 3.0 T with the fat-saturated intermediate-weighted fast SE sequence, while 23 were detected at 1.5 T and 24 were detected at 3.0 T with the fat-saturated spoiled GRE sequence. (b) Bar graph shows percentages of detected lesions at 1.5 T and 3.0 T for the fat-saturated intermediate-weighted fast SE sequence according to the grade of the cartilage lesion (grade 1 = less than 50% cartilage defect [n = 7], grade 2 = more than 50% cartilage defect [n = 7], grade 3 = full-thickness lesion [n = 15]). The grade 1, 2, and 3 lesions created were detected with frequencies of 29% (two of seven lesions), 86% (six of seven lesions), and 67% (10 of 15 lesions), respectively, while at 3.0 T detection sensitivity was higher at all grades, with frequencies of 71% (five of seven lesions), 100% (seven of seven lesions), and 93% (14 of 15 lesions), respectively. (c) Bar graph shows percentages of detected lesions at 1.5 T and 3.0 T for the fat-saturated spoiled GRE sequence according to the grade of the cartilage lesion. A sensitivity difference in lesion detection with this sequence only became apparent when grade 1 lesions were assessed, with three (43%) of seven grade 1 lesions detected at 1.5 T and four (57%) detected at 3.0 T. At both 1.5 and 3.0 T, six (86%) of seven grade 2 lesions and 14 (93%) of 15 grade 3 lesions were detected.

Confidence Scores
A review of the radiologists' confidence scores in diagnosing the artificial cartilage lesions revealed that lesions were classified more frequently as definite (ie, detected with high certainty) at 3.0 T than at 1.5 T (Table 2, Fig 7). With fat-saturated intermediate-weighted fast SE sequences, 18 (69%) of 26 lesions observed at 3.0 T were detected with a high degree of certainty, whereas only nine (50%) of 18 lesions observed at 1.5 T were detected with a high degree of certainty. In spite of the fact that improvement in lesion detection was only slightly higher at 3.0 T than at 1.5 T with the fat-saturated spoiled GRE sequences, detection of lesions with a high degree of certainty with this sequence was significantly improved at the higher field strength, as is indicated by 15 of 24 lesions being detected with high certainty at 3.0 T and only 10 of 23 lesions being detected with high certainty at 1.5 T. Figure 7 further clarifies this improvement, showing a comparable increase at 3.0 T in lesion detection confidence between the spoiled GRE and intermediate-weighted fast SE sequences (a score improvement of 0.28 from 1.5 to 3.0 T with spoiled GRE sequences vs a score improvement of 0.25 with intermediate-weighted fast SE sequences). Table 3 summarizes the relationship between lesion detection and confidence scores by grade and shows higher confidence scores for grade 2 and grade 3 lesions than for grade 1 lesions.

View larger version (20K): [in this window] [in a new window] [Download PPT slide]   Figure 7. Bar graph shows mean confidence scores and standard deviations for the lesions according to field strength and imaging sequence (Scan type). FSE = fast SE, SPGR = spoiled GRE. These scores were determined by dividing the sum of the confidence scores for each lesion by the number of lesions detected for each sequence. Standard deviations were 0.77 at 1.5 T and 0.70 at 3.0 T for the fat-saturated intermediate-weighted fast SE sequence and 0.74 at 1.5 T and 0.66 at 3.0 T for the fat-saturated spoiled GRE sequence.

Image Quality
Subjective evaluation of image quality was higher at 3.0 T with both the intermediate-weighted fast SE and the spoiled GRE sequences (P < .05) (Table 5). Overall, subjective evaluation of cartilage was higher at 3.0 than at 1.5 T for both sequence types, and the femoral cartilage was consistently given better scores than the tibial cartilage. Edge sharpness scores were also significantly higher at 3.0 than at 1.5 T for both sequence types, although the disparity between femoral and tibial cartilage was minimal.

View larger version (148K): [in this window] [in a new window] [Download PPT slide]   Figure 8. Sagittal MR image obtained with a fat-saturated spoiled GRE sequence (4000/35) shows a markedly high cartilage signal intensity band (arrows) on the anterior aspect of the femoral condyle between the lower border of the patella and the anterior border of the tibial plateau; this is an artifact.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
Studies have consistently revealed that the potential of MR imaging to accurately depict early chondral degeneration, reveal the location of cartilage lesions, and help determine lesion volume and cartilage thickness makes it a viable and potentially superior alternative to the invasive procedures currently used for assessing cartilage (12–18). However, limitations of imaging at 1.5 T and lower field strengths also have repeatedly been noted (12,19–25,27). In this preliminary study, we compared imaging of artificial cartilage lesions at 3.0 T to that at 1.5 T with fat-saturated intermediate-weighted fast SE and spoiled GRE sequences and found that imaging at 3.0 T yielded a higher diagnostic performance, and, in many cases, allowed lesions to be detected and characterized with greater confidence. This improved performance at 3.0 T likely is due to the substantially increased SNR yielded by using the higher field strength magnet.

Although MR imaging of cartilage has been performed by using many sequence types, fat-saturated intermediate-weighted fast SE and spoiled GRE sequences have been cited as being among the most appropriate for this application (14). Sonin et al (31) reported articular cartilage grading sensitivities of 59%–73.5% and specificities of 86.7%–90.5% with intermediate-weighted fast SE sequences that were similar to those found in our study. Fat-saturated spoiled GRE sequences, like fat-saturated intermediate-weighted fast SE sequences, have also been found to have a high sensitivity in the detection of cartilage lesions (32,33). Considering the results of these previous studies at 1.5 T, fat-saturated intermediate-weighted fast SE and fat-saturated spoiled GRE sequences were chosen for this study because, to our knowledge, results of studies analyzing imaging sequences for cartilage defects at 3.0 T have not yet been published.

Previous work has shown that increases in spatial resolution enhance cartilage depiction and lesion detection (19). Even though higher spatial resolution is possible at 3.0 T with optimized coils, we chose to keep spatial resolution identical at both 3.0 and 1.5 T, thereby isolating potential advantages in soft-tissue contrast and SNR at higher field strengths. Future studies are required to push the frontiers of spatial resolution at 3.0 T. Currently, coils for these applications are a limiting factor.

Because we wanted to evaluate potential advantages of 3.0 T imaging in the context of patient imaging, time considerations were emphasized. For the fat-saturated spoiled GRE sequences, we wished to validate the proposed time-saving advantage of using the 3.0 T imaging unit while also making comparisons of diagnostic capabilities. Therefore, the rationale for the fat-saturated spoiled GRE sequence selection process was to choose a sequence with a relatively short acquisition time and compare it with a standardized fat-saturated spoiled GRE sequence with a longer acquisition time at 1.5 T. Even though the acquisition time was 25% shorter (8 minutes 48 seconds at 3.0 T vs 11 minutes 14 seconds at 1.5 T), a mildly higher diagnostic performance was observed by using the fat-saturated spoiled GRE sequence at 3.0 T. For the detection of suspected subtle lesions (ie, grade 1 lesions in our study), maximizing the diagnostic potential of 3.0 T images by using imaging times comparable to those typical at 1.5 T may be more beneficial than reducing the acquisition time, given the relative difficulty in detecting these defects. For suspected larger lesions (ie, grade 2 or 3 lesions in our study) or follow-up of previously viewed lesions, sequences with shorter acquisition times than those typical at 1.5 T are still likely to result in adequate lesion depiction.

Like several previous investigators, we performed our experiments by using a porcine animal model (34,35). Given the control we had over lesion creation in our cadaveric specimens, it was possible to introduce subtle lesions in a wide variety of anatomic regions at sizes that are encountered in clinical studies. These lesions were challenging for the individual radiologist to assess with the different field strengths and imaging sequences.

A focus of this study was the assessment of lesions similar to those that occur naturally in vivo, with an emphasis on clinical utility. Therefore, the created defects had irregular borders, nonuniform depths, and a variety of shapes. This inherent irregularity added a dimension of uncertainty to dimension analysis. With smaller lesions and greater section thicknesses, this uncertainty tends to increase. Additionally, as lesion shapes become more irregular, the possibility that images do not depict the actual size of the lesion may be more frequent. Please note that the fat-saturated spoiled GRE sequences tended to be better suited for the assessment of lesion depth, perhaps in part because they had a thinner section thickness of 1.5 mm (versus 2 mm for the fat-saturated intermediate-weighted fast SE sequences) and because of the relatively high signal intensity of cartilage in these sequences. A measurement accuracy with fat-saturated spoiled GRE sequences that was similar to that seen in this study was reported for another study of cartilage lesion assessment (31).

A potential limitation of this study was that we were not able to exactly simulate in vivo conditions: Although motion artifacts are common among humans, they were obviously absent in our specimens. Because even slight motion has the potential to compromise the detection of small lesions, the use of optimal bracing techniques would be recommended in future studies involving humans. Additionally, due to the nature of specimen preparation, air artifacts that may have mildly limited diagnostic performance and had more of an effect on the spoiled GRE sequence than on the intermediate-weighted fast SE sequence were introduced. We therefore highlight the importance in future studies of investing time in developing a technique capable of minimizing air artifacts, as was done in this study.

Also of consideration is the difference in cartilage thickness between pig and human knees, which may have made conspicuity of the lesions in this model more difficult. This limitation presented a special challenge in the creation and assessment of lesions in the tibial plateau, which is very thin in porcine knees. The fact that we used identical imaging parameters at 1.5 and 3.0 T for the fat-saturated intermediate-weighted fast SE sequence may have been another limitation. However, when we initially compared standard 1.5-T fat-saturated intermediate-weighted fast SE images with the ones obtained at 1.5 T by using the parameters optimized for 3.0 T, image quality was fairly similar. A marked effect on detection rate was therefore not expected. In this preliminary study, consensus readings were performed, so future studies with larger numbers of radiologists are required to assess interobserver variation.

Practical application: The results of our study show that MR imaging at 3.0 T may play an important role in clinical musculoskeletal imaging, particularly in the assessment of cartilage. To fully take advantage of the higher field strength, however, further optimization of sequences for specific applications may be necessary. With development of optimized coils, maximized use of parallel imaging methods, spatial resolution enhancement, and improvement of pulse sequences, higher diagnostic utility and further acquisition time reduction at 3.0 T can be expected.

In conclusion, the results of this study show that the diagnostic performance of MR imaging in depicting hyaline articular cartilage lesions may be substantially enhanced by imaging at 3.0 T rather than at 1.5 T. In this experimental study, more lesions were detected and graded correctly at 3.0 T than at 1.5 T. Image quality and SNR measurements were also significantly higher at 3.0 T than at 1.5 T. Also in this study, fat-saturated intermediate-weighted fast SE sequences were more sensitive for detecting cartilage lesions than were fat-saturated spoiled GRE sequences. Fat-saturated spoiled GRE sequences, however, tended to be more accurate for lesion grade assessment.

	ACKNOWLEDGMENTS

 
We express our gratitude to Niles Bruce, RT, for assisting with the imaging procedures and to Ying Lu, PhD, for his help with the statistical analysis of the data.

	FOOTNOTES

 

Abbreviations: GRE = gradient echo • SE = spin echo • SNR = signal-to-noise ratio

Author contributions: Guarantors of integrity of entire study, T.M.L., S.M., L.S.; study concepts, T.M.L., J.N.M., S.M.; study design, T.M.L., L.S., S.M., C.P.; literature research, J.N.M., C.A.S., T.M.L.; experimental studies, J.N.M., C.A.S., T.M.L., E.H., D.N., C.P.; data acquisition, C.A.S., J.N.M., D.N., E.H., C.P.; data analysis/interpretation, J.N.M., C.A.S., T.M.L., S.M., D.N., C.P.; statistical analysis, C.A.S., J.N.M., T.M.L.; manuscript preparation, J.N.M., T.M.L., C.P.; manuscript definition of intellectual content, J.N.M., T.M.L., S.M., L.S.; manuscript editing, T.M.L., S.M., L.S., E.H.; manuscript revision/review and final version approval, T.M.L., S.M., L.S

	References

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 

1. Roberts S, McCall IW, Darby AJ, et al. Autologous chondrocyte implantation for cartilage repair: monitoring its success by magnetic resonance imaging and histology. Arthritis Res Ther 2003; 5:R60–R73.[CrossRef][Medline]
2. Andres BM, Mears SC, Wenz JF. Surgical treatment options for cartilage defects within the knee. Orthop Nurs 2001; 20:27–31.
3. Bentley G, Biant LC, Carrington RW, et al. A prospective, randomised comparison of autologous chondrocyte implantation versus mosaicplasty for osteochondral defects in the knee. J Bone Joint Surg Br 2003; 85:223–230.
4. Breinan HA, Minas T, Hsu HP, Nehrer S, Shortkroff S, Spector M. Autologous chondrocyte implantation in a canine model: change in composition of reparative tissue with time. J Orthop Res 2001; 19:482–492.[CrossRef][Medline]
5. Minas T, Nehrer S. Current concepts in the treatment of articular cartilage defects. Orthopedics 1997; 20:525–538.[Medline]
6. Koulalis D, Schultz W, Heyden M, Konig F. Autologous osteochondral grafts in the treatment of cartilage defects of the knee joint. Knee Surg Sports Traumatol Arthrosc 2004; 12:329–334.[Medline]
7. Hangody L, Fules P. Autologous osteochondral mosaicplasty for the treatment of full-thickness defects of weight-bearing joints: ten years of experimental and clinical experience. J Bone Joint Surg Am 2003; 85-A(suppl 2):25–32.
8. Hangody L, Rathonyi GK, Duska Z, Vasarhelyi G, Fules P, Modis L. Autologous osteochondral mosaicplasty: surgical technique. J Bone Joint Surg Am 2004; 86-A(suppl 1):65–72.
9. Gross AE. Cartilage resurfacing: filling defects. J Arthroplasty 2003; 18:14–17.[CrossRef][Medline]
10. Hunziker EB. Articular cartilage repair: basic science and clinical progress—a review of the current status and prospects. Osteoarthritis Cartilage 2002; 10:432–463.[CrossRef][Medline]
11. Macarini L, Murrone M, Marini S, Moretti B, Patella V. Aspects of magnetic resonance in the surgical treatment of osteochondral lesions of the knee. Radiol Med (Torino) 2003; 106:74–86.
12. Schafer FK, Muhle C, Heller M, Brossmann J. Gelenkknorpel in der MR-Tomographie. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 2001; 173:279–288.[Medline]
13. McCauley T, Disler D. Magnetic resonance imaging of articular cartilage of the knee. J Am Acad Orthop Surg 2001; 9:2–8.[Abstract/Free Full Text]
14. Trattnig S, Mlynarik V, Huber M, Ba-Ssalamah A, Puig S, Imhof H. Magnetic resonance imaging of articular cartilage and evaluation of cartilage disease. Invest Radiol 2000; 35:595–601.[CrossRef][Medline]
15. Recht MP, Resnick D. Magnetic resonance imaging of articular cartilage: an overview. Top Magn Reson Imaging 1998; 9:328–336.[Medline]
16. Bredella MA, Tirman PF, Peterfy CG, et al. Accuracy of T2-weighted fast spin-echo MR imaging with fat saturation in detecting cartilage defects in the knee: comparison with arthroscopy in 130 patients. AJR Am J Roentgenol 1999; 172:1073–1080.[Abstract/Free Full Text]
17. Peterfy CG, Genant HK. Emerging applications of magnetic resonance imaging in the evaluation of articular cartilage. Radiol Clin North Am 1996; 34:195–213.[Medline]
18. Vallotton JA, Meuli RA, Leyvraz PF, Landry M. Comparison between magnetic resonance imaging and arthroscopy in the diagnosis of patellar cartilage lesions: a prospective study. Knee Surg Sports Traumatol Arthrosc 1995; 3:157–162.[CrossRef][Medline]
19. Link TM, Majumdar S, Peterfy C, et al. High resolution MRI of small joints: impact of spatial resolution on diagnostic performance and SNR. Magn Reson Imaging 1998; 16:147–155.[CrossRef][Medline]
20. Hargreaves BA, Gold GE, Beaulieu CF, Vasanawala SS, Nishimura DG, Pauly JM. Comparison of new sequences for high-resolution cartilage imaging. Magn Reson Med 2003; 49:700–709.[CrossRef][Medline]
21. Hodler J, Resnick D. Current status of imaging of articular cartilage. Skeletal Radiol 1996; 25:703–709.[CrossRef][Medline]
22. McNicholas MJ, Brooksbank AJ, Walker CM. Observer agreement analysis of MRI grading of knee osteoarthritis. J R Coll Surg Edinb 1999; 44:31–33.
23. Matsui N, Kobayashi M. Application of MR imaging for internal derangement of the knee (orthopedic surgeon's view). Semin Musculoskelet Radiol 2001; 5:139–141.[CrossRef][Medline]
24. Rubenstein JD, Li JG, Majumdar S, Henkelman RM. Image resolution and signal-to-noise ratio requirements for MR imaging of degenerative cartilage. AJR Am J Roentgenol 1997; 169:1089–1096.[Abstract/Free Full Text]
25. Macarini L, Murrone M, Marini S, Mariano M, Zaccheo N, Moretti B. MR in the study of knee cartilage pathologies: influence of location and grade on the effectiveness of the method. Radiol Med (Torino) 2003; 105:296–307.
26. Uetani M. MR imaging of cartilage lesions of the knee: what is the clinical indication (radiologist's view)? Semin Musculoskelet Radiol 2001; 5:147–149.[CrossRef][Medline]
27. Hargreaves B, Gold G, Lang P, et al. MR imaging of articular cartilage using driven equilibrium. Magn Reson Med 1999; 42:695–703.[CrossRef][Medline]
28. Gold G, Han E, Stainsby J, Wright G, Brittain J, Beaulieu C. Relaxation times and contrast in musculoskeletal MR imaging at 3.0 Tesla (abstr). In: Radiological Society of North America scientific assembly and annual meeting program. Oak Brook, Ill: Radiological Society of North America, 2003; 655–656.
29. McGinty JB. Operative arthroscopy. New York, NY: Raven, 1996.
30. Noyes FR, Stabler CL. A system for grading articular cartilage lesions at arthroscopy. Am J Sports Med 1989; 17:505–513.[Abstract/Free Full Text]
31. Sonin AH, Pensy RA, Mulligan ME, Hatem S. Grading articular cartilage of the knee using fast spin-echo proton density-weighted MR imaging without fat suppression. AJR Am J Roentgenol 2002; 179:1159–1166.[Abstract/Free Full Text]
32. Tan TC, Wilcox DM, Frank L, et al. MR imaging of articular cartilage in the ankle: comparison of available imaging sequences and methods of measurement in cadavers. Skeletal Radiol 1996; 25:749–755.[CrossRef][Medline]
33. Disler DG, McCauley TR, Wirth CR, Fuchs MD. Detection of knee hyaline cartilage defects using fat-suppressed three-dimensional spoiled gradient-echo MR imaging: comparison with standard MR imaging and correlation with arthroscopy. AJR Am J Roentgenol 1995; 165:377–382.[Abstract/Free Full Text]
34. Woertler K, Strothmann M, Tombach B, Reimer P. Detection of articular cartilage lesions: experimental evaluation of low- and high-field-strength MR imaging at 0.18 and 1.0 T. J Magn Reson Imaging 2000; 11:678–685.[CrossRef][Medline]
35. van der Linden E, Kroon HM, Doornbos J, Hermans J, Bloem JL. MR imaging of hyaline cartilage at 0.5 T: a quantitative and qualitative in vitro evaluation of three types of sequences. Skeletal Radiol 1998; 27:297–305. [CrossRef][Medline]

This article has been cited by other articles:

				J. Zhao, R. Krug, D. Xu, Y. Lu, and T. M. Link MRI of the Spine: Image Quality and Normal-Neoplastic Bone Marrow Contrast at 3 T Versus 1.5 T Am. J. Roentgenol., April 1, 2009; 192(4): 873 - 880. [Abstract] [Full Text] [PDF]

				R. Kijowski, D. G. Blankenbaker, K. W. Davis, K. Shinki, L. D. Kaplan, and A. A. De Smet Comparison of 1.5- and 3.0-T MR Imaging for Evaluating the Articular Cartilage of the Knee Joint Radiology, March 1, 2009; 250(3): 839 - 848. [Abstract] [Full Text] [PDF]

				J. Y. Jung, Y. C. Yoon, S.-H. Choi, J. W. Kwon, J. Yoo, and B.-K. Choe Three-dimensional Isotropic Shoulder MR Arthrography: Comparison with Two-dimensional MR Arthrography for the Diagnosis of Labral Lesions at 3.0 T Radiology, February 1, 2009; 250(2): 498 - 505. [Abstract] [Full Text] [PDF]

				C. K. Kuhl, F. Traber, J. Gieseke, W. Drahanowsky, N. Morakkabati-Spitz, W. Willinek, M. von Falkenhausen, C. Manka, and H. H. Schild Whole-Body High-Field-Strength (3.0-T) MR Imaging in Clinical Practice * Part II. Technical Considerations and Clinical Applications Radiology, April 1, 2008; 247(1): 16 - 35. [Abstract] [Full Text] [PDF]

				J. S. Bauer, S. Banerjee, T. D. Henning, R. Krug, S. Majumdar, and T. M. Link Fast High-Spatial-Resolution MRI of the Ankle with Parallel Imaging Using GRAPPA at 3 T Am. J. Roentgenol., July 1, 2007; 189(1): 240 - 245. [Abstract] [Full Text] [PDF]

				N. Saupe, C. W. A. Pfirrmann, M. R. Schmid, T. Schertler, M. Manestar, and D. Weishaupt MR Imaging of Cartilage in Cadaveric Wrists: Comparison between Imaging at 1.5 and 3.0 T and Gross Pathologic Inspection Radiology, April 1, 2007; 243(1): 180 - 187. [Abstract] [Full Text] [PDF]

				J. S. Bauer, S. J. Krause, C. J. Ross, R. Krug, J. Carballido-Gamio, E. Ozhinsky, S. Majumdar, and T. M. Link Volumetric Cartilage Measurements of Porcine Knee at 1.5-T and 3.0-T MR Imaging: Evaluation of Precision and Accuracy. Radiology, November 1, 2006; 241(2): 399 - 406. [Abstract] [Full Text] [PDF]

				C. M. Phan, M. Matsuura, J. S. Bauer, T. C. Dunn, D. Newitt, E. M. Lochmueller, F. Eckstein, S. Majumdar, and T. M. Link Trabecular Bone Structure of the Calcaneus: Comparison of MR Imaging at 3.0 and 1.5 T with Micro-CT as the Standard of Reference Radiology, May 1, 2006; 239(2): 488 - 496. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Masi, J. N. Articles by Link, T. M. Search for Related Content PubMed PubMed Citation Articles by Masi, J. N. Articles by Link, T. M.