MR Diagnosis Of Articular Cartilage Injury In The Knee: A Clinical Study And An Experimental Study

Purpose: MR is a powerful technique for evaluation of articular cartilage injury. 1. To assess the efficacy of FS-3D-FISP, FS-2D-FLASH and SE-T1WI sequences in the detection of articular cartilage injury of the knee.2. To assess the accuracy of three-dimensional reconstruction based on 3D-FISP-FS sequence magnetic resonance images of articular cartilage lesions, using actual cartilage lesions as reference standards. Material and Methods:1. Clinical study:During 2002.12 to 2003.6, 34 consecutive patients (18 men and 16 women, age range 14-65 years, mean 41.2 years) with persistent symptoms of knee pain who were scheduled for arthroscopy underwent MR examination of the knee on a 1.5 T MR unit (Symphony; Siemens Medical System) prior to arthroscopy. With use of a flex coil, the hyaline articular cartilage was imaged with the following sequence: FS-3D-FISP, FS-2D-FLASH, SE-T1WI sequence .All the patients underwent arthroscopy with a Stryker X6000 arthroscope within tow weeks after MR examination. The grading of the articular cartilage on MR images and arthroscopy were used the 5-category grading scale which referenced the Outerbridge system and the ICRS hyaline cartilage injury classification system. The 2D MR images were transferred by local area network to a workstation (Advantage 3.1; General Electric Medical System) and processed tow-dimensional and three-dimensional reconstruction. In addition , 10 knees of 10 healthy volunteers underwent MR examination . 2. Experimental study Five freshly slaughtered porcine knees was used in this articular cartilage study. Articular cartilage injuries of grade-2 , grade-3 and grade-4 were created on articular surfaces of the five porcine knee joints . The grade-2 and grade-3 lesions were created by shaving off thin slices of articular cartilage with a scalpel. The grade-4 lesions were made by incision with a scalpel, and then the cartilage in the lesions'center was removed by tweezers. The articular surfaces were washed with the saline after operating, a little amount saline was infused into the articular capsule, and then the articular capsule was sutured. The five porcine knees were scanned using a Siemens Symphony magnetic resonance imaging system and a flex coil. All MR scanning are finished within 6 hours after slaughtered. The porcine knees were imaged with FS-3D-FISP sequence in the sagital plane. After the porcine knees were scanned, they were debrided to the joint to expose the lesions. By using calipers and a ruler calibrated to 0.05 mm, lesion dimensions were determined. Thelengths were defined as the biggest measurement of the articular cartilage injury area. And then on the perpendicular direction, the biggest widths were measured. For reducing the error of measurement, each part of lesions was measured three times, then take its average value finally. Using an Advantage Workstation 3.1 of General Electric Medical System, the MR images of FS-3D-FISP sequence were reconstructed into three dimensional images Results:1. Part 1: 1.1 FS-3D-FISP, FS-2D-FLASH and SE-T1WI sequences can show the normal shape of articular cartilage.10 knees of volunteers were scaned. The MR images of FS-3D-FISP sequence show the articular cartilage as high signal band with the laminar appearance. The MR images of FS-2D-FLASH sequence are similar to the images of FS-3D-FISP,but the laminar appearance is not obviously. The articular cartilage is middle signal on T1-weighted images. 1.2 The Kappa values of MR sequences were compared respectively. The diagnosis accuracy of FS-3D-FISP sequence was obviously better than FS-2D-FLASH(u=5.25, P <0.01) and SE-T1WI(u=5.14, P <0.01) sequences. In diagnosing articular cartilage injury, FS-3D-FISP sequence and FS-2D-FLASH sequence or SE-T1WI sequence have obviously statistic difference. FS-2D-FLASH sequence and SE-T1WI sequence have no statistic difference(u=0.199, P>0.05). 1.3 For the detection of articular cartilage injury withFS-3D-FISP sequence, sensitivity was 91.4%, specificity was 97%, Kappa was 0.818. So the correlation of detection of articular cartilage injury is good between FS-3D-FISP sequence and arthroscopy. 1.4 It is an important indirect sign of articular cartilage injury that the abnormal signal is founded in the subchondral bone and marrow. In the cases of our study, 77.6% lesions were founded long T1 signal in the region of the subchondral bone and marrow near the lesions, and in FS-3D-FISP and FS-2D-FLASH sequences the region were high signal . The abnormal signal represent the abnormal structure of the subchondral bone and marrow. So imaging and therapy of articular cartilage injury should include the cartilage and the subchondral bone and marrow to improve the effect of it. 1.5 Three-dimensional reconstruction of articular cartilage injury is helpful for the clinical doctor to make sure the locations of the lesions. By far this technique can only be used as a complement method of tow-dimensional MR imaging. 2. Part 2: 2.1 The percent error of the measurement on three-dimensional reconstruction image were related to the MR scan thickness. When MR scan thickness is 1mm, the error of the measurement is the smallest. According to the MR scan thickness, all the specimens were divided into three groups , and Friedman Test was performed ( χ2=74.45,P<0.01). 2.2 Friedman Test was performed to evaluate the differencebetween SSD and MIP ,and showed a statistical difference(χ2=8.65,P=0.013<0.05) . With the SSD reconstruction image the lesions can be measured more accurately in case of higher SNR. 2.3 The percent error of the measurement on SSD reconstruction image were related to the lesions dimensions. Analysis of variance was performed to three groups , and showed a significant statistical difference(F=16.601,P <0.01). 2.4 The percent error of the measurement on SSD reconstruction images were related to the lesions degrees,and the error of grade 4 was the smallest. Mann-Whitney U test was performed to tow groups of articular cartilage grade 3 and 4, and showed a statistical difference(Z=-2.24,P=0.025<0.05). 2.5 This study demonstrates that three-dimensional reconstruction of cartilage from thin-section MR imaging can show cartilage surface irregularities. The direction of measurement was not limited by the MR scan plane, and the distance of any two points can be measured. So using three-dimensional reconstruction imagesof 1mm to measure cartilage lesion dimensions is more accurately than using tow-dimensional MR images. Friedman Test was performed (χ2=8.65,P=0.013<0.05). Conclusion:1. Comparing with arthroscopy, the diagnosis accuracy of FS-3D-FISP sequence was obviously better than FS-2D-FLASH and SE-T1WI sequences. And the correlation of detection of articular cartilage injury is good between FS-3D-FISP sequence and arthroscopy.2. It is an important indirect sign of articular cartilage injury that the abnormal signal is founded in the subchondral bone and marrow. 3. Three-dimensional reconstruction of articular cartilage injury is helpful for the clinical doctor to make sure the locations of the lesions. 4.The percent error of the measurement on three-dimensional reconstruction image were related to the MR scan thickness, the method of three-dimensional reconstruction, the lesions dimensions and the lesions degrees. 5. Three-dimensional reconstruction of cartilage from MR imaging of 1mm thickness can show cartilage surface irregularities.