Morphological and biochemical T2 evaluation of cartilage repair tissue based on a hybrid double echo at steady state (DESS-T2d) approach




To use a new approach which provides, based on the widely used three-dimensional double-echo steady-state (DESS) sequence, in addition to the morphological information, the generation of biochemical T2 maps in one hybrid sequence.

Materials and Methods:

In 50 consecutive MRIs at 3.0 Tesla (T) after matrix-associated autologous chondrocyte transplantation (MACT) of the knee, by the use this new DESS-T2d approach, the morphological Magnetic resonance Observation of CArtilage Repair Tissue (MOCART) score, as well as biochemical T2d values were assessed. Furthermore, these results were correlated to standard morphological sequences as well as to standard multi-echo spin-echo T2 mapping.


The MOCART score correlated (Pearson:0.945; P < 0.001) significantly as assessed with standard morphological sequences (68.8 ± 13.2) and the morphological images of the DESS T2d sequence (68.7 ± 12.6). T2 and T2d relaxation times (ms) were comparable in between the control cartilage (T2: 52.5 ± 11.4; T2d: 46.6 ± 10.3) and the repair tissue (T2: 54.4 ± 11.4; T2d: 47.5 ± 13.0) (T2: P = 0.157; T2d: P = 0.589). As expected, T2d values were lower than the standard-T2 values, however, both functional relaxation times correlated significantly (Pearson:0.429; P < 0.001).


The presented hybrid approach provides the possibility to combine morphological and biochemical MRI in one fast 3D sequence, and thus, may attract for the clinical use of biochemical MRI. J. Magn. Reson. Imaging 2011;. © 2011 Wiley-Liss, Inc.

ARTICULAR CARTILAGE LESIONS are a common pathology of the knee joint and many patients could benefit from cartilage repair (1, 2). Such surgical treatment options may offer the possibility for young active patients with cartilage defects to avoid the development of osteoarthritis or delay its progression. Newly developed cartilage repair techniques, including arthroscopic or open surgical approaches, as well as marrow-stimulation techniques, osteochondral grafting, and chondrocyte implantation/transplantation, require noninvasive and high quality follow-up studies to monitor the efficacy of the therapy.

Morphological and biochemical MRI is capable of visualizing cartilage repair tissue and the adjacent native cartilage, in vivo, at high-resolution, and in clinically applicable scan times (3–6). Standard morphological approaches can demonstrate the constitution of cartilage repair tissue, particularly the filling of the defect and the integration with adjacent native cartilage and bone. Newer three-dimensional (3D) sequences can be used for improved morphological cartilage repair imaging in three dimensions, and also for the diagnosis of surrounding pathologies in the knee joint (5, 7–10). An accepted morphological grading system to depict cartilage repair tissue and the adjacent structures on MRI is the Magnetic resonance Observation of CArtilage Repair Tissue (MOCART) score (11, 12). In addition to morphological MRI, biochemical MR approaches, such as delayed gadolinium-enhanced MRI of cartilage (dGEMRIC), or T2 relaxation time mapping, are able to provide a specific measure of the structure of cartilage (13, 14). Cartilage physiology and composition can be determined in vivo and cartilage repair tissue can thus be evaluated, different repair tissues can be differentiated, and the maturation after transplantation techniques can be assessed (15–20). In cartilage defects, and following nonsurgical and surgical cartilage repair, morphological MRI provides the basis for diagnosis and follow-up evaluation, whereas biochemical MRI provides deeper insight into the composition of cartilage and cartilage repair tissue. A combination of both, together with clinical evaluation, may, in the future, represent a desirable multimodal approach to diagnosis, as well as for routine clinical follow-up after cartilage repair procedures.

Recently, the 3D double-echo steady state (DESS) sequence, enabling an accurate and precise analysis of cartilage morphology (21, 22), was reported to permit the generation of biochemical T2 maps (23). Hence, morphological and biochemical information about the articular cartilage and cartilage repair tissue may be provided by only one hybrid sequence.

The aim of this study was to use this new DESS-T2d approach in an initial study to assess the morphological MOCART score, as well as biochemical T2 values in patients after matrix-associated autologous chondrocyte transplantation (MACT) of the knee with only one sequence. We also wanted to compare the performance of this new hybrid approach with standard morphological sequences, as well as standard multi-slice multi-echo spin-echo (MSME) T2 mapping. Additionally the maturation of the cartilage repair tissue, as assessed by morphological and biochemical MRI, should be evaluated.



The study protocol was approved by the by the medical university ethics commission, and written, informed consent was obtained from all patients before enrollment in the study.

Fifty consecutive MR scans were prospectively included in this study. MRI was performed during clinical routine standard follow-up postoperatively after MACT of the knee joint, including intervals of 1, 3, 6, 12, 24, 36, 60, and 96 months. The 50 MR scans were performed in 43 patients, with a mean age of 36.1 ± 9.3 years (ranging from 18 years to 53 years), at the time point of the MRI exam. There were 10 female and 33 male patients included, 18 left knees and 25 right knees. There were 36 patients who underwent the MR measurement once, and seven patients who underwent the MR measurements twice.

Inclusion criteria was a MACT procedure due to a full-thickness cartilage defect of the femoral condyle of the knee joint; the lateral femoral condyle was treated in 12 patients and the medial femoral condyle was treated in 31 patients. MACT was performed due to cartilage injury or osteochondritis dissecans. The singular nature of the cartilage defect was the precondition for the surgical therapy and was documented during the first arthroscopic surgical step, where chondrocytes were harvested. During a second surgical step, a mini-arthrotomy, after debridement, the biomatrix with seeded cells was trimmed to exactly match the defect size and was implanted with no periosteal cover and fixed only by fibrin glue (24). The defect size was 4.8 cm2 (range: 1.5–11.2 cm2).

Inclusion criteria also dictated that subjects had no severe osteoarthritis, no meniscal injury, instability of the knee joint, or deformity. These criteria were preoperatively proven by clinical examination, conventional radiographs and MRI, and documented during the initial surgery.

For further subdivision of the patients according to the postoperative interval (time between MACT surgery and follow-up MRI), the patients were divided into four groups, with a follow-up period of 1–6 months (group I; n = 12; age = 37.1 ± 11.4years), 12 months (group II; n = 10; age = 38.1 ± 11.2 years), 24 months (group III; n = 16; 33.3 ± 8.2 years), and 36–96 months (group IV; n = 12; age = 37.6 ± 7.5 years). The follow-up intervals were chosen to gain information on the maturation of cartilage repair tissue and to assess if the standard as well as the hybrid DESS-T2d approach provide comparable information. No significant difference in age was observed when comparing the different follow-up groups (P ≥ 0.05).

Image Acquisition

MR imaging was performed on a 3 Tesla (T) MR scanner (Magnetom Trio, Siemens Medical Solutions, Erlangen, Germany) with a gradient strength of 40 mT/m, using a dedicated eight-channel knee coil (In vivo, Gainesville, FL). All patients were positioned consistently with the joint space in the middle of the coil and the knee extended in the coil. Patients were scanned after at least half an hour of rest to avoid changes in T2 or T2d relaxation because of different loading before MR measurement (25).

The MR protocol was identical for all included MR measurements. For morphological MOCART scoring, the protocol consisted of a high-resolution proton-density turbo spin-echo (PD-TSE) sequence, a T2-weighted dual fast spin-echo (dual-FSE) sequence, and a T1-weighted turbo inversion recovery magnitude (TIRM) sequence, as recommended for the evaluation of the MOCART score. In addition, the new DESS T2d sequence (23) was prepared and the morphological DESS images were used for the evaluation of the MOCART score. For the evaluation of biochemical T2 relaxation times, a standard multi-slice multi-echo spin-echo (MSME) sequence was performed. The new DESS T2d sequence (23) was also used to assess quantitative T2 values.

The sagittal PD-TSE sequence was obtained with a repetition time (TR) of 2400 ms, an echo time (TE) of 38 ms, and a flip angle of 160°. Field of view (FOV) was 120 × 120 mm, pixel matrix was 512 × 512, and slice thickness was 2 mm, which resulted in a voxel size of 0.2 × 0.2 × 2 mm. No fat suppression was used, the parallel acquisition technique (PAT) was turned off, and bandwidth was 244 Hz/Px. Total scan time for 32 slides was 6:11 min. The sagittal dual-FSE sequence was performed with a TR of 5120, a first TE of 9.5 ms, and a second TE of 67 ms; the flip angle was 140°. FOV was 180 × 180 mm, pixel matrix was 448 × 448, and slice thickness was 3.0 mm (0.4 × 0.4 × 3 mm). No fat suppression was applied, PAT was off, bandwidth was 203 Hz/Px, and 30 slices were achieved in 6:46 min. The coronal TIRM sequence was performed with a TR of 7690, a TE of 41 ms, an inversion time (TI) of 220 ms, and a flip angle of 150°. FOV was 150 × 150 mm, pixel matrix was 256 × 256, and slice thickness was 3.0 mm (0.6 × 0.6 × 3 mm). Bandwidth was 250 Hz/Px and, with PAT turned on, with an acceleration factor of two using a generalized auto calibrating partially parallel acquisition (GRAPPA) technique, 36 slices were acquired in 2:35 min. The sagittal MSME sequence for quantitative T2 mapping was performed with a TR of 1200 ms, TEs of 13.8 ms, 27.6 ms, 41.4 ms, 55.2 ms, 69 ms, and 82.8 ms, and a flip-angle of 180°. The FOV was 160 × 160mm, the pixel matrix was 384 × 384, and the slice thickness was 3.0 mm, resulting in a voxel size of 0.4 × 0.4 × 3.0 mm. The bandwidth was 228 Hz/Px, and the data acquisition time for this sequence was 4:09 min for 12 slides. The new hybrid DESS T2d sequence, for morphological and biochemical imaging, was performed sagittally, with a TR of 19.9 ms. The TEs were 4.2 ms for the S+ echo and 35.6 ms for the S− echo, with the flip-angle at 33° for the optimal combination of morphological and biochemical information, as described in the study by Welsch et al (23). To ensure an optimal comparability for the quantitative T2 measurements, the FOV with 160 × 160 mm, the pixel matrix with 384 × 384 and the slice thickness with 3.0 mm (resulting in a voxel size of 0.4 × 0.4 × 3.0 mm) were identical to the MSME T2 sequence. The bandwidth was 195 Hz/Px, and the data acquisition time for this sequence was 4:44 min for 18 slides. For exact comparability, slices were positioned identically with the same imaging parameters with regard to the localization for both the T2 multiecho spin echo and the T2d sequence. The acquisition time for the respective sequence is highlighted in Table 1.

Table 1. Time requirements for the standard and the hybrid procedure:
Standard morphological sequencessag. PDTSE: 6:11 min.sag. Dual FSE: 6:46 min.cor. TIRM: 2:35 min.Total: 15:32 min.
Standard multi-echo spin-echo T2 mapping   4:09 min.
New hybrid DESS T2d approach   4:44 min.

Data Analysis

The evaluations of the morphological MOCART score, as well as the biochemical T2 values, were performed on a Leonardo Workstation (Siemens Healthcare, Erlangen, Germany). Two readers, an experienced senior musculoskeletal radiologist (25 years of experience) and an orthopedic surgeon (10 years of experience) with a special interest in musculoskeletal MR imaging, analyzed the data sets in consensus while blinded to patient name and postoperative follow-up interval, and were only advised about the localization of the cartilage transplants. The assessment of the MOCART score and T2 assessment were done in random order.

The MOCART scoring was performed on the basis of the morphological standard sequences using the PD-TSE, the dual-FSE, and the TIRM sequences as well as on the basis of the morphological images provided by the new DESS T2d sequence. The following nine variables were assessed according to the point-scoring system of Marlovits and Trattnig: the degree of repair filling (i); the integration of the cartilage repair tissue with the border zone (ii); the structure of the surface (iii); the structure of the entire repair tissue (iv); the signal intensity of the repair tissue compared with native cartilage (v); the constitution of the subchondral lamina (vi); the constitution of the subchondral bone (vii); possible adhesions (viii); and possible effusion (ix). The maximum score achievable in the evaluation of the nine variables is 100. Figure 1 shows an example of a patient scan after MACT of the femoral condyle with respect to the morphological MOCART evaluation.

Figure 1.

High-resolution MRI of a 31-year-old male patient 12 months after matrix-associated chondrocyte transplantation (MACT) (arrows) of the medial femoral condyle. Sagittal DESS T2d sequence (TR/TE 19.9/4.2, 35.6/flip angle = 33°) (a), sagittal PD-TSE (2400/38/flip angle = 160°) (b), sagittal T2-weighteddDual-FSE sequence (5120/9.5, 67;flip angle 140°) (c), and coronal T1-weighted TIRM (7690/41, flip angle 150°) (d).

Biochemical T2 values vary, as assessed on the basis of the standard MSME approach, as well as on the basis of the new DESS T2d approach. T2 and T2d maps were obtained on-line by the built-in MapIt software (Siemens Medical Solutions, Erlangen, Germany). MSME T2 maps were obtained by a pixel-wise, mono-exponential, nonnegative least squares (NNLS) fit analysis. T2d maps were calculated using the ratio of the S− and the S+ signal, as presented by Welsch and co-workers (23). On the basis of the morphological images and the surgical reports, the cartilage repair tissue sites and sites of healthy control cartilage were identified on the T2 and the T2d maps. Control cartilage was defined as normal on the morphological sequences if cartilage thickness was preserved, the surface was intact, and no intrachondral signal alterations were visible. Manual region-of-interest (ROI) analysis was performed on a region of morphologically normal-appearing cartilage (control cartilage) and on the area of cartilage repair (repair tissue). The ROIs divided the full thickness of the control cartilage, as well as the repair tissue, into equal-sized deep and superficial halves (from the subchondral border to the joint surface) and evaluation was performed for global (full-thickness), deep, and superficial cartilage aspects. In animal examinations and, recently, in initial in vivo studies, this “zonal” T2 evaluation was able to characterize the constitution, and possibly, the maturation of cartilage repair tissue (17, 26, 27). All areas of cartilage repair and selected healthy control cartilage sites were located within the weight-bearing zone of the femoral cartilage, defined on the basis of the peripheral margins of the menisci (28). The ROIs were assessed directly on the T2 maps as reported by Mosher et al. (29). ROIs were first drawn on the DESS T2d maps, then transferred, for control reasons, back to the morphological DESS images, and, after that, transferred to the MSME T2 maps. To exclude changes due to possible movement between the DESS T2d and the SE T2 measurements, the ROIs were subsequently controlled site-by-site for their identical location. Analysis was done on two to three consecutive slides, depending on the size of the cartilage repair tissue. Altogether, 1404 ROIs were analyzed and correlated. For the ROI covering the full thickness of cartilage repair tissue, the mean pixel count was 613 ± 368 pixels; for the healthy articular cartilage sites, the mean pixel count was 527 ± 319 pixels. An example of T2 and T2d maps from two patients after MACT of the femoral condyle is presented in Figures 2 and 3. For better visualization, the ROIs are included for the repair tissue (arrows) and the control cartilage.

Figure 2.

Sagittal DESS T2d-map (19.9/4.2, 35.6/flip angle = 33°) (a) and multi-echo spin-echo T2-map (1200/13.8, 27.6, 41.4, 55.2, 69, 82.8/ flip angle = 180°) (b) of the same patient (Fig. 1) with the area of MACT marked by arrows. The comparable behavior of the T2d and the T2d values becomes visible.

Figure 3.

Sagittal DESS T2d-map (19.9/4.2, 35.6/flip angle = 33°) (a) and multi-echo spin-echo T2-map (1200/13.8, 27.6, 41.4, 55.2, 69, 82.8/ flip angle = 180°) (b) of a patient 24 months after MACT of the medial femoral condyle.

Further evaluation was prepared to classify image quality and possible artifacts for the sequences used for the morphological MOCART scoring (performance to guarantee a sufficient analysis of the cartilage repair tissue and of the adjacent cartilage) and the quantitative T2 evaluation (subjective quality of the quantitative T2 map to perform the ROI measurements). Hence, standard sequences (PD-TSE, dual FSE, TIRM and MSME) were assessed, as well as the new DESS T2d approach. The musculoskeletal radiologist and the orthopedic surgeon with a special interest in musculoskeletal MRI again graded image quality and artifacts in consensus. For image quality, a four-level scale was used, in which a score of 4 indicated excellent image quality; a score of 3, good image quality; a score of 2, acceptable image quality; and a score of 1, poor image quality (30). Artifacts, including motion, susceptibility, metal/postsurgical, banding, etc,. were subjectively graded as absent (iv), mild (iii), moderate (ii), or severe (i) (31).

Morphological and biochemical analysis was provided by both the standard and the new DESS T2d sequences. The standard approach was correlated to the new DESS T2d approach. In addition, the image quality and artifacts were described for the standard and the new DESS T2d approach. With regard to morphological MRI, the MOCART score was assessed and the overall score, as well as the different variables, were correlated between the two methods. Biochemical T2 evaluation was also provided for global (full-thickness) as well as zonal (deep and superficial) cartilage aspects of the control cartilage and the repair tissue. Possible differences between the control cartilage and the repair tissue were assessed, a possible significant zonal increase was evaluated as a sign of hyaline or hyaline-like cartilage structure (17, 27), and, finally, the standard MSME approach was correlated to the new DESS T2d approach. The morphological as well as the biochemical analysis was performed for all patients together and in terms of the respective follow-up interval for group I (1–6 months) (n = 13), group II (12 months) (n = 10), group III (24 months) (n = 15), and group IV (36–96 months) (n = 12) to focus on possible cartilage repair tissue maturation (18, 26).

Statistical Analysis

Statistical analysis was performed by analyses of variance using a three-way analysis of variance with random factors, considering the fact of different measurements within each patient. For the trend in between the cartilage layers (deep to superficial), a three-way analysis of variance, with random effects and two repeated measure factors, was performed. For correlation between the standard and the new DESS T2d approaches, a bivariate correlation using the Pearson coefficient was used. Quality and artifacts were compared using a Student's t-test. The SPSS version 16.0 software (SPSS Institute, Chicago, IL) for Mac (Apple, Cupertino, CA) was used, and a P value less than 0.05 was considered statistically significant.


Morphological MOCART Scoring

The MOCART score for all patients was 68.8 ± 13.2 assessed with the standard morphological sequences, and 68.7 ± 12.6 assessed with the morphological images of the DESS T2d sequence, which yielded a highly significant (P < 0.001) Pearson correlation coefficient of 0.945. When looking at the different variables of the MOCART score, the Pearson correlation coefficient for the (i) defect fill was 0.992, which shows a highly significant correlation (P < 0.001). The (ii) cartilage interface with cartilage integration to the border zone showed a Pearson coefficient of 0.968 (P < 0.001). The (iii) constitution of the surface of the repair tissue correlated also highly significantly, with a correlation coefficient of 0.842 (P < 0.001) between the standard sequences and the DESS T2d sequence. The (iv) structure of the repair tissue (Pearson: 0.690; P < 0.001) and the (v) signal intensity of the repair tissue (Pearson: 0.881; P < 0.001) correlated significantly, as did the (vi) subchondral lamina (Pearson: 0.848; P < 0.001) and the (vii) subchondral bone (Pearson: 0.725; P < 0.001) when comparing the standard and the hybrid approach. The diagnosis of possible adhesions (Pearson: 1.000; P < 0.001) and joint effusion (Pearson: 1.000; P < 0.001) showed similar results.

For the evaluation based on the postoperative follow-up period, the MOCART score showed a clear increase between group I (1–6 months) and group II (12 months) for both methods. Subsequently, group II, group III, and group IV showed only a slight increase; however, again, visible for both the standard and the new approach (Fig. 4; Table 2).

Figure 4.

Bar-plot with error bars (standard deviation) to depict the mean MOCART values of the patients based on their respective postoperative follow-up interval. A slight increase in the MOCART values over time becomes visible using the standard sequences, as well as the DESS T2d approach.

Table 2. Morphological Magnetic Resonance of Cartilage Repair Tissue (MOCART) scoring as performed by the respective sequence
Follow-up intervalMOCART standardMOCART DESS T2d
1–6 monthsMean60,462,3
Std. Deviation13,512,7
12 monthsMean70.568.5
Std. Deviation9,310,6
24 monthsMean71,371,7
Std. Deviation14,314,1
36–96 monthsMean73,372,1
Std. Deviation11,510,5

Biochemical T2 Mapping

T2 and T2d relaxation times are given in milliseconds (ms) for mean ± standard deviation. Mean values for all patients were comparable in sites of control cartilage and in sites of cartilage repair. Quantitative T2 values for the standard method were 52.5 ± 11.4 for control cartilage and 54.4 ± 11.4 for cartilage repair tissue (P = 0.157) and quantitative T2d values for the hybrid approach were 46.6 ± 10.3 for control cartilage and 47.5 ± 13.0 for cartilage repair tissue (P = 0.589). The zonal evaluation revealed a significant increase from deep to superficial cartilage sites for the control cartilage with the standard method—T2: deep:50.3 ± 9.7; superficial:54.8 ± 9.6 (P < 0.001); and with the hybrid approach—T2d: deep:42.2 ± 9.6; superficial:51.0 ± 12.2 (P < 0.001), as well as for the cartilage repair tissue (standard method—T2: deep:52.8 ± 11.2; superficial:56.0 ± 12.0 (P < 0.001); and hybrid method—T2d: deep:44.2 ± 12.3; superficial:50.7 ± 14.7) (P < 0.001).

The correlation between the MSME T2 evaluation and the new DESS T2d evaluation was highly significant (P < 0.001) for the mean values (Pearson coefficient: 0.429) as well as for the deep (Pearson: 0.350) and the superficial (Pearson:0.468) values. When looking at the control cartilage (Pearson coefficients: mean: 0.476; deep:0.389; superficial:0.508), the correlation was slightly higher compared with the repair tissue (Pearson: mean:0.353; deep:0.277; superficial: 0.412). Again, all correlations were shown to be highly significant (P < 0.001). The correlation graphs of both approaches (MSME T2 versus DESS T2d) are visualized in Figure 5 for the mean values (in terms of control cartilage and repair tissue) and in Figure 6 for the zonal (deep and superficial) values).

Figure 5.

Correlation plot for mean T2 and T2d values (single values and regression line with its slope [R Sq linear]) of the repair tissue and the control cartilage. The control cartilage shows slightly better correlation compared with the repair tissue.

Figure 6.

Correlation plots, with 95% confidence interval and a regression line with its slope (R Sq linear), of the multi-echo spin-echo T2 and DESS T2d analysis. The assessment of the deep (a) and the superficial (b) aspects of articular cartilage are visualized.

The evaluation over time, with regard to the different follow-up groups, showed comparable results for both quantitative (T2 and T2d) approaches. Whereas the T2 as well as the T2d values of the control cartilage showed stable results throughout groups I–IV (P ≥ 0.05), the cartilage repair tissue showed significantly higher results in group I and group II, compared with group III and IV (P < 0.05). When comparing the control cartilage and the repair tissue, again for both sequences, the relaxation times of the repair tissue in group I and II were longer than those of the control cartilage (P < 0.05), whereas the relaxation times in groups III and IV were comparable (P ≥ 0.05). For the zonal evaluation, a clear zonal increase could be assessed for the control cartilage in all groups (P < 0.001). The repair tissue, however, showed no zonal increase from deep to superficial in group I (P ≥ 0.05), but a significant zonal increase in groups II–IV (P < 0.05). These zonal results were again comparable in the T2 and the T2d evaluations. The results concerning the different follow-up groups are provided in Figure 7 and Table 3.

Figure 7.

Mean T2 (black) and DESS T2d (gray) values are displayed for the repair tissue (dotted line) and the control cartilage (continuous line) of the patients based on their postoperative follow-up interval. Although the DESS T2d values are lower compared with the multi-echo spin-echo T2 values, their behavior over time seems to be comparable.

Table 3. Quantitative T2 values (milliseconds; ms) of the repair tissue and the surrounding control cartilage with respect to the post-operative interval
 Follow-up Interval T2 deepT2 supT2 meanT2d deepT2d supT2d mean
Repair Tissue1–6 monthsMean (ms)61,662,762,153,055,954,5
Std. Deviation10,712,311,213,518,315,6
12 monthsMean (ms)54,959,357,146,255,450,8
Std. Deviation9,710,19,711,110,49,9
24 monthsMean (ms)48,151,449,640,146,543,3
Std. Deviation8,89,99,212,113,912,6
36–96 monthsMean (ms)50,554,352,441,148,444,8
Std. Deviation11,713,011,98,613,610,7
Control Cartilage1–6 monthsMean (ms)49,355,152,242,654,048,3
Std. Deviation10,011,69,612,019,615,3
12 monthsMean (ms)48,851,$50,342,050,446,2
Std. Deviation11,410,610,68,19,48,0
24 monthsMean (ms)50,855,153,044,151,047,6
Std. Deviation10,08,48,69,010,29,3
36–96 monthsMean (ms)51,356,253,739,849,344,5
Std. Deviation8,28,68,09,18,78,4

Quality and Artifacts

The subjective image quality to measure the performance of the standard sequences was graded excellent in 44 cases, good in 5 cases, and acceptable in 1 case. For the new DESS T2d approach, the performance was excellent in 40 cases, good in 9 cases, and acceptable in 1 case. A poor image quality was not seen in either category. This resulted in a comparable diagnostic quality for the standard sequences (3.86 ± 0.41) compared with the new hybrid approach (3.78 ± 0.46) (P = 0.361). The subjective measurements of the visible artifacts in the cartilage repair tissue or in the adjacent cartilage revealed slightly better results for the standard sequences, where artifacts were absent in 48 cases, mild in two cases, and with no cases of moderate or severe artifacts, compared with the DESS T2d sequence, where artifacts were absent in 41 cases, mild in eight cases, moderate in one case, and with no cases of severe artifacts (P = 0.024).


The results of the present study show that the hybrid DESS T2d sequence provides the possibility of combining morphological and biochemical MRI in one fast 3D-sequence. In addition, compared with standard morphological sequences and standard quantitative T2 evaluation, comparable information could be achieved at enormous time-savings, with ∼ 20 min for the standard sequences and ∼ 5 min for the hybrid DESS T2d sequence. Although the accuracy, when comparing and correlating the quantitative T2 evaluation using the MSME T2 approach and the DESS T2d approach, is still limited, the presented hybrid technique is valuable in that it is innovative and decreases scan time.

Overall, in patients after cartilage repair, it is of the utmost importance to have a noninvasive tool—such as MRI—to follow-up on the outcome of these surgical procedures. As it is typically young patients with a severe cartilage injury who are treated, cartilage repair procedures are possibly the only therapeutic option to prevent osteoarthritis. In the recent literature, in addition to morphological MR evaluations, biochemical approaches provided promising results (4, 5, 17, 32, 33). Nevertheless, the combination of a high-resolution morphological MR protocol, together with new biochemical approaches, is (scan−) time consuming, which detracts from clinical use. This is even more relevant when patients are examined on a 1.0T or 1.5T MR scanner and not on a 3.0T MR system. Hence, the presented approach allows the inclusion of a new biochemical MR parameter (T2d) into a single sequence that is part of a clinical MR protocol, with no loss of time. This might open the possibility, in other cartilage pathologies, as well, such as osteoarthritis, for the more widespread clinical use of additional biochemical MRI.

The main goal of an MRI measurement in the follow-up after cartilage repair is to provide a high quality morphological description of the cartilage repair tissue and the surrounding structures. In the present study, the established MOCART score (1, 4, 5, 11, 34–37) revealed results comparable to standard morphological sequences or to the morphological part of the new DESS-T2d approach. Not only did the overall MOCART score show a highly significant correlation, but also the single variables showed this correlation, indicating that both the prepared standard MR sequences, as well as the DESS sequence (21, 38), with regard to morphological imaging, were able to depict the constitution of the repair tissue and the adjacent structures. The results of the MOCART score are very much comparable to an existing longitudinal study by Trattnig et al (39). Although the present study is based on a cross-sectional approach, the applied standard sequences and the DESS T2d sequence revealed comparable results independent of the follow-up time point.

With regard to biochemical MRI, the new compositional parameter called DESS T2d was used and compared with a standard MSME T2 sequence. This parameter was calculated from the two contrasts of the DESS sequence (23, 40), which results in a gradient-echo sequence optimized for cartilage imaging that produces an averaged image from these two echoes (41). The assessed T2d values are nevertheless noticeably lower than the standard MSME T2 values, due to an existing T1 dependency, a clear dependency to the applied flip angle and other sources of errors discussed by Welsch and co-workers (23). In addition, the standard MSME T2 mapping technique for the evaluation of articular cartilage, as presented by Mosher and Dardzinski (14, 42), uses 4–12 echoes to compute quantitative T2 values based on a fitted T2 curve, whereas the presented DESS T2d approach calculates its quantitative T2d values based on an equation mainly taking the two contrasts (FISP and PSIF) of the DESS sequence into consideration (23), which could be another possible source of error. The results of this study are nevertheless promising. Comparable to the MSME T2 sequence, with the new DESS T2d approach, the very important zonal variation (43) is clearly visible. This significant increase from deep to superficial cartilage portions is evident in the native control cartilage and can be assessed using the classical MSME T2 approach as well as with the new DESS T2d approach. For the cartilage repair tissue, the stratification of the repair tissue is slightly less pronounced; however, again, visible using both methodologies. Histologically validated animal studies have reported this zonal variation in T2 values as an indicator of hyaline cartilage (26, 27). Recent human in vivo studies have discussed the formation of this zonal variation over time in the follow-up of comparable matrix-associated autologous cartilage repair procedures, as a sign of repair tissue maturation (18, 37), based on the findings of Watrin-Pinzano et al (26). In addition to the zonal variation, the decrease in the T2 values over time, and the adaptation to the T2 values of the control cartilage, is seen to depict the reorganization of cartilage structure after MACT (16, 18). In the present study, both signs of repair tissue maturation could be assessed with both the MSME T2 approach and the DESS T2d approach. The significant correlation of both sequences revealed nevertheless only moderate correlation values, with Pearson coefficients between 0.277 and 0.508. These findings indicate that visualized components of articular cartilage are not similar. However both sequences are able to quantify the zonal structure of articular cartilage providing comparable results in the differentiation of native cartilage and cartilage repair tissue. Therefore, the presented results indicate that the DESS T2d approach may be of use in the clinical follow-up of patients after cartilage repair procedures, and, possibly, after other joint-preserving therapies, adding a usable compositional value to a stable morphological evaluation. Particularly when looking at the results of the subjectively assessed image quality and possible artifacts, the DESS T2d approach seems to offer a large amount of valuable information in a relatively short scan time. A further potential benefit of this DESS T2d sequence is the possibility of isotropic 3D imaging. Whereas morphological isotropic DESS imaging provides information on the status of the knee joint with the possibility of cartilage morphometry, an isotropic T2d map might open new possibilities in biochemical MRI.

The limitations of the present study are the cross-sectional character and the lack of histological validation. Nevertheless, the relatively high number of included MR scans seems to offer a good basis for the performed comparison of standard sequences and the DESS T2d sequence. The validity of the evaluation due to the postoperative follow-up, nevertheless, might be limited. Our study used a flip angle of 33° for the DESS T2d approach, which has been reported to show the best agreement between morphological and functional imaging (23) This has to be validated in future studies, especially when considering very recent findings of substantial signal improvements by increasing the flip angle of the DESS sequence to 90° (44). Similar to the present study, the goal of upcoming approaches should be to ensure the optimal morphological information with the additional benefit of an extra compositional parameter. A further limitation of the study might be the use of non–fat-suppressed sequences for both the DESS T2d sequence and the MSME T2 sequence, which was done to ensure comparability due to a comparable depiction of the interface line between cartilage and bone and chemical shift artifacts.

In conclusion, the preliminary data from this initial study demonstrate that, using the DESS T2d approach, the morphological description and the biochemical T2 assessment of the repair tissue and the surrounding control cartilage is possible in patients after MACT using only one hybrid sequence. The morphological results of the DESS T2d approach, as evaluated by the MOCART score, are comparable to the results of standard morphological MR sequences. Comparably, however, with lower correlation values, the quantitative T2 values (mean T2 values as well as deep and superficial T2 values for the zonal T2 evaluation) of the DESS T2d sequence could be correlated to the quantitative T2 values as assessed by the standard MSME-T2 sequence. Furthermore, the important zonal stratification, demonstrated as an increase in T2 values from deep to superficial, could be comparably assessed using both, the standard T2 and the new T2d methodology. Although upcoming studies on larger patient groups and other pathologies are required to confirm the clinical use of the DESS T2d sequence, the presented hybrid approach provides the possibility to combine morphological and biochemical MRI in one fast 3D sequence, and thus, may make this an attractive incentive for the clinical use of biochemical MRI.