Nội dung

Available online http://arthritis-research.com/content/6/6/R492 Research article Open Access Vol 6 No 6 Acoustic stiffness and change in plug cartilage over time after autologous osteochondral grafting: correlation between ultrasound signal intensity and histological score in a rabbit model Hiroshi Kuroki1, Yasuaki Nakagawa2, Koji Mori3, Mao Ohba2, Takashi Suzuki2, Yasuyuki Mizuno2, Keiji Ando2, Makoto Takenaka4, Ken Ikeuchi4 and Takashi Nakamura2 1Department of Physical Therapy, School of Health Sciences, Faculty of Medicine, Kyoto University, Kyoto, Japan of Orthopaedic Surgery, Graduate School of Medicine, Kyoto University, Kyoto, Japan 3Applied Medical Engineering Science, Graduate School of Medicine, Yamaguchi University, Yamaguchi, Japan 4Institute for Frontier Medical Sciences, Kyoto University, Kyoto, Japan 2Department Corresponding author: Yasuaki Nakagawa, ynaka@kuhp.kyoto-u.ac.jp Received: 16 Mar 2004 Revisions requested: 21 Apr 2004 Revisions received: 11 Jun 2004 Accepted: 30 Jun 2004 Published: 14 Sep 2004 Arthritis Res Ther 2004, 6:R492-R504 (DOI 10.1186/ar1219) © 2004 Kuroki et al.; licensee BioMed Central Ltd. This is an Open Access article: verbatim copying and redistribution of this article are permitted in all media for any purpose, provided this notice is preserved along with the article's original URL. http://arthritis-research.com/content/6/6/R492 Abstract We investigated quantitative changes over time in ultrasound signal intensity (an index of stiffness), signal duration (an index of surface irregularity), and interval between signals (an index of thickness) of plug cartilage in an animal model of autologous osteochondral grafting. A full-thickness osteochondral plug was surgically removed and replaced in male Japanese white rabbits (n = 22). Specimens obtained at day 0 and weeks 2, 4, 8, 12 and 24 postoperatively were assessed using an ultrasound system and by macroscopic and histological evaluation (modified Mankin's score). Histology revealed that the plug sank until 2 weeks postoperatively, and that newly formed cartilagelike tissue covered the plug, but at 24 weeks the tissue detached. The plug itself survived well throughout the period of observation. Although the signal intensity at the plug site was same as that in the sham operated contralateral knee at day 0, from 2 to 24 weeks postoperatively it was less than that in the sham knee. At 8 weeks, this difference was significant (P < 0.05). Modified Mankin's score revealed early degenerative changes at the site, but macroscopic examination did not. Signal intensity correlated significantly with score (both at day 0 and at the five postoperative time points [P < 0.05, r = -0.91] and as a whole [P < 0.05, r = -0.36]). Signal intensity also significantly correlated with the individual subscores for 'cartilage structure' (P < 0.05, r = -0.32) and 'cartilage cells' (P < 0.05, r = -0.30) from the modified Mankin's score, but not significantly with subscores for 'staining' and 'tidemark'. Signal duration correlated significantly with total score (as a whole [P < 0.05, r = 0.34]), but not significantly with the score for cartilage structure (P = 0.0557, r = 0.29). The interval between signals reflected well the actual thickness of the plug site. The significant relationships between ultrasound signal intensity and scores suggest that early degenerative changes in plug cartilage and cartilage-like tissue, especially in the superficial layer, are detectable by high-frequency ultrasound assessment. Keywords: articular cartilage, high-frequency ultrasound, histology, osteochondral grafting surgery Introduction High-frequency pulse echo ultrasound techniques, which reveal a number of features of normal and degenerated articular cartilage [1-16], were recently introduced. The ultrasound signal correlates strongly with thickness of cartilage [5,8,17,18]. The signal also provides information about the integrity of the superficial zone of cartilage [14,8] and about progression of osteoarthritis [6,7,9-15]. OCG = autologous osteochondral graft. It is known that, if untreated, full thickness articular cartilage damage will progress to osteoarthritis [19,20] because articular cartilage has limited intrinsic healing ability [21]. Numerous attempts to induce healing of cartilage defects have been made, but each treatment has strengths and weaknesses [15]; therefore, regeneration and repair of articular cartilage remains a clinical and scientific challenge. R492 Arthritis Research & Therapy Vol 6 No 6 Kuroki et al. Autologous osteochondral grafts (OCGs) are used to treat small, isolated articular cartilage defects. Successful clinical treatment with OCGs has been reported, with few complications at the donor site [22-27]. Using both arthroscopic [22,23,28-31] and open surgical techniques [24,25,32,33], this type of graft has been implanted in knee joint and talus, and in end-stage osteochondritis dissecans lesions of the humeral capitellum. A requirement of the treatment is that the cylindrical plug of the autograft includes both articular cartilage and underlying subchondral bone [22-25,27-30] OCGs have several benefits, including reliable bone union, high survival rate for grafted cartilage, and little risk for disease transmission. In human clinical and animal studies it has been shown that osteochondral plugs maintain hyaline cartilage coverage over the subchondral bone, and that the plugs retain their viability and attach to the surrounding bone [22,24,27,34]. Using a high-frequency pulse echo ultrasound technique, we found that the implantation procedure does not adversely affect the stiffness, surface regularity and thickness of OCG plug cartilage immediately after surgery (at day 0) [13]. However, no research has yet focused on ultrasonographical and histological assessment of postoperative changes in plug cartilage. We therefore investigated changes over time in the plug cartilage after OCG in a rabbit model. OCG was performed on the left knee (Fig. 1). A full thickness cylindrical osteochondral plug (5 mm in diameter, 7 mm in depth), which went through the articular surface and into the subchondral bone, was harvested using the Osteochondral Autograft Transfer System (Arthrex, Naples, Florida). The Osteochondral Autograft Transfer System 'Recipient' Tube Harvester, 5 mm in diameter, was positioned on the patellar groove and was then driven into the subchondral bone to a depth of 7 mm. During creation of the hole, the harvester was maintained at a 90° angle to the articular surface in both sagittal and coronal planes. After insertion to 7 mm depth, the harvester was rotated 90° clockwise and then 90° anticlockwise. The harvester was then pulled out of the joint. The articular cartilage around the hole was shaved in a 7 × 7 mm square with a chisel until bleeding was observed from the subchondral bone. This procedure of inducing bleeding is believed to accelerate healing of surrounding cartilage, and some evidence has been reported that the area is filled by newly formed reparative tissue [23,26]. The harvested plug was then returned precisely to its original site. Thus, the recipient hole was repaired with an autologous osteochondral plug that was of exactly the same size as the hole. The chiselled area around the hole was left. The joint capsule and skin incision were closed with 4-0 nylon sutures. Methods Figure 1 Animals Left knee (OCG surgery): This investigation was approved by the Animal Research Committee of the Kyoto University Graduate School of Medicine (approval number Med Kyo 03155). Twenty-four male Japanese white rabbits (Japan Animals Co. Ltd, Osaka, Japan) were used. The rabbits were randomly assigned to one of six groups defined by the time point at which the animals were evaluated (day 0 [immediately after surgery], and 2, 4, 8, 12 and 24 weeks after surgery; n = 4 in each group). The animals were maintained at the Institute of Laboratory Animals, Graduate School of Medicine, Kyoto University for 3 weeks before the start of the experiments. They were housed, each in a separate cage, in a room maintained at 22°C and 50% humidity, with a 14-hour light/ 10-hour dark cycle, and were given food and water ad libitum. 1) A parapatellar incision 2) Patella was laterally dislocated 3) OCG was performed on the left knee Harvesting a full-thickness cylindrical osteochondral plug Shaving around the harvesting hole after harvesting The plug was returned to its original site 4) Joint capsule and skin incision were closed with 4–0 nylon sutures Right knee (sham operation): 1) A parapatellar incision Osteochondral grafting Intravenous pentobarbital sodium (25 mg/kg body weight) was used to induce and maintain general anaesthesia. The rabbits were placed supine, and the surgery was performed on both knees. The lower limbs were disinfected, and 2 ml of 0.5% lidocaine was injected subcutaneously into the parapatellar region. A parapatellar incision was made to expose the knee joint, and the patella was laterally dislocated. R493 2) Patella was laterally dislocated 3) Exposure to air for almost same minutes with the OCG 4) Joint capsule and skin incision were closed with 4–0 nylon sutures A diagram of surgical procedures. procedures OCG, autologous osteochondral grafts. Available online http://arthritis-research.com/content/6/6/R492 Sham surgery was performed on the right knee as follows: a parapatellar incision was made; the patella was laterally dislocated; and then the joint capsule and skin incision were closed, all over the same period of time as was required for the OGC procedure (Fig. 1). All rabbits were allowed to move freely in their cages after the surgery. Two rabbits were excluded from the series; one exhibited signs of infection 7 days after surgery and in the other the plug was fractured during surgery. At day 0 (immediately after surgery; n = 3), or 2 (n = 3), 4 (n = 4), 8 (n = 4), 12 (n = 4) or 24 (n = 4) weeks after surgery, the rabbits (with a weight [mean ± standard deviation] 3.2 ± 0.19 kg, range 2.8–3.6 kg) were killed by intravenous injection of a fatal dose of sodium pentobarbital. The implanted osteochondral plugs, the articular cartilage of the defect, and the intact region of the patellar groove were evaluated macroscopically. The plug cartilage was assessed using an ultrasound system and then evaluated histologically. Figure 2 (a) e a b g Transducer Specimen c f d Saline (b) Ultrasound assessment The ultrasound assessment system we used provides quantitative information about tissue properties, and was described previously (Fig. 2a) [10,13,14]. Briefly, the system developed by Mori and coworkers [10] consists of a transducer and a pulser/receiver (Panametrics Japan, Tokyo, Japan), a digital oscilloscope and a personal computer (Fig. 2a). The diameter of the transducer was approximately 3 mm. The central frequency of the ultrasound wave was 10 MHz. As the wave passes through interfaces between media of different acoustic impedances, reflections return to the transducer and generate electrical signals in the transducer that are proportional to the intensity [35]. On examining cartilage, two large amplitude groups of reflected waves were observed (Fig. 2b). By using appropriate wavelet transformation for these amplitude groups [10,13,14], three properties of cartilage can be analyzed. The first amplitude group (group N) represents the signal from the surface of the cartilage, and the second (group K) represents that from the subchondral bone (Fig. 2b). The time interval (µs) between the two signals represents the thickness of the cartilage. The duration (µs) of group N represents the irregularity of the cartilage, because diffused reflection waves in a rough surface return to the ultrasound transducer with a time delay. The intensity of group N is proportional to the Young modulus of cartilage. The Young modulus is determined using the following equations [13]: V = (E / p) Z = ρV R = (Z2 - Z1)/(Z2 + Z1) (a) The ultrasonic pulser/receiver, (b)measurement a digital oscilloscope system consists and (c) aofpersonal a transducer, computer (a) a pulser/receiver, (b) a digital oscilloscope and (c) a personal computer. The system can be used with (d, e) arthroscopy, (f) open surgery and (g) saline bath for experimental purposes. The ultrasound wave output from the transducer travels through saline. The reflected waves return to the transducer and generate electrical signals in the transducer that are proportional to the reflected wave intensity. (b) Typical ultrasound echo (lower) and wavelet map (upper). The wavelet map was calculated from the ultrasound echo using wavelet transform. The first (left) of the two large amplitude groups was the echo (t = 2.0 µs: group N) reflected from the cartilage surface, and the second (right; t = 3.9 µs: group K) was reflected from the subchondral bone. The signal intensity (maximum magnitude, as shown by the scale) of group N represents cartilage stiffness. The time interval between groups N and K represents cartilage thickness. The signal duration of group N represents the surface irregularity of the cartilage. Here, E is the Young modulus, V is the speed of sound, and p is the density of a material. Z and R are the acoustic impedance of a material and reflectance, respectively. In the present study, Z2 is the acoustic impedance of articular cartilage and Z1 is that of saline. Z1 is a constant (1.48 × R494 Arthritis Research & Therapy Vol 6 No 6 Kuroki et al. 106 kg/m2S at room temperature). From the three equations above, the Young modulus of cartilage, E, is given by the following: 2 Z12  R +1  E= ×  ρ  R −1  The reflectance of cartilage in saline, R, is nearly 0.10. In the case, E is proportional to R, if the density of cartilage, P, is constant. The intensity of group N is directly proportional to R. Consequently, the intensity of group N may be used as an index of cartilage stiffness. An indentation test demonstrated a significant relationship between the intensity of group N and the aggregate modulus [10], which is one of the indices of mechanical properties of articular cartilage [36,37]. As described above, in theory it is reasonable to regard the intensity of the wave as an index of stiffness of cartilage. The intensity was represented as relative values because the change in this index was calculated from the three equations given above. Therefore, we used signal intensity as an index of stiffness, signal duration as an index of surface irregularity, and interval between signals as an index of thickness. The plug cartilage and the corresponding site on the sham side were evaluated at three different sites: the center of the plug, 0.25 mm distal to the center, and 0.25 mm proximal to the center. Histological evaluation For histological examination, the specimens were fixed in 10% neutral buffered formalin for 7 days, decalcified with 0.25 mol/l EDTA in phosphate-buffered saline (pH 7.4), dehydrated in graded ethanol, and embedded in paraffin wax. Sagittal sections (6 µm thick) were then cut, stained with safranin-O/fast green and haematoxylin and eosin, and examined microscopically. All sections were observed and evaluated by three authors. Histological evaluation of plug cartilage was performed using the modified Mankin's score [38] (original score proposed by Mankin and coworkers [39]). The grading system was composed of four categories – cartilage structure (6 points), cartilage cells (3 points), staining (4 points) and tidemark integrity (2 points) – with a highest score of 14 points; normal cartilage scored 0 (Table 1) [38]. When we observed newly formed tissue that covered the plug cartilage, the finding was counted as 'pannus and surface irregularities' (2 points). Statistical analysis Data for histological scores were analyzed statistically using the nonparametric Kruskal-Wallis test and the posthoc Scheffe's F-test (for comparison between weeks postoperatively), and using the Mann-Whitney U-test (for R495 comparison between grafted and sham sides). Ultrasound data were analyzed using parametric repeated measures analysis of variance and the post-hoc Scheffe's F-test. The relationships between ultrasound data (mean of the three measurements) and the score were analyzed using nonparametric Spearman's rank-order correlation. Results Macroscopic findings Day 0 and postoperative week 2 At day 0 (Fig. 3a) the plug was intact and the margins around the plug and the shaved square were clearly apparent. At 2 weeks postoperatively (Fig. 3b) the margin around the square could be clearly detected. Although the articular surface of the plug was smooth and regular, the plug had subsided a little. Postoperative weeks 4 and 8 At 4 weeks (Fig. 3c) and 8 weeks postoperatively (Fig. 3d), the site around and over the plug was filled with newly formed reparative tissue. The margin around the plug and the shaved square could be easily detected at 4 weeks. At 8 weeks postoperatively the margin was a little faint but still detectable. Although the surfaces of the plug and the defect looked irregular at 4 and 8 weeks, the plug was glossy. Postoperative weeks 12 and 24 The plug survived well and no osteoarthritic changes such as osteophyte formation were observed at 12 weeks (Fig. 3e) and 24 weeks postoperatively (Fig. 3f). At 12 weeks the margin around the plug and the shaved square was a little faint but it was still detectable. At 24 weeks postoperatively, however, the margin was very faint. At 12 and 24 weeks, the surface of the plug cartilage was as smooth as that of the adjacent intact cartilage. Although the plug and intact cartilage were glossy in all of the specimens obtained from 0 to 24 weeks, the appearance of the shaved area was not so. In the shaved area, no reparative tissue was observed throughout the 24 weeks. All sham-operated cartilage was grossly normal, and there was no evidence of articular damage. Histological findings At day 0 and postoperative week 2 Histological examination at day 0 revealed that the plug had been inserted flush with the surrounding articular surface. The site of the defect (the shaved area) was clearly recognizable because the cartilage around the plug had been shaved until it bled (Fig. 4a,4b). At 2 weeks postoperatively the plug had subsided a little and the newly formed tissue covered half of the plug. No tissue was observed at the site of the defect (Fig. 4c,4d). The actual thickness of cartilage increased a little with the overlying tissue. Available online http://arthritis-research.com/content/6/6/R492 Figure 3 Macroscopic findings. findings A full-thickness osteochondral plug of 5 mm in diameter and 7 mm in depth was harvested from the patellar groove. Articular cartilage around the hole was shaved in a 7 × 7 mm square until bleeding from the subchondral bone was observed. The harvested plug was then returned precisely to its original site. (a) At day 0, the plug was intact and the margin around the plug and the shaved square was clearly recognizable. (b) At 2 weeks postoperatively, the margin around the square could be clearly detected. Although the articular surface of the plug was smooth and regular, the plug had subsided a little. (c) The margin around the plug and the shaved square could be easily detected at 4 weeks. (d) At 8 weeks postoperatively the margin was a little faint but could still be detected. At 4 (panel c) and 8 weeks (panel d), the site around and over the plug was filled with newly formed reparative tissue. Although the surface of the plug and the defect looked irregular, the plug was glossy. (e) The margin around the plug and the shaved square was a little faint but could still be detected at 12 weeks. (f) At 24 weeks postoperatively, however, the margin was very faint. The plug survived well and osteoarthritic changes such as osteophyte formation were not observed at 12 and 24 weeks postoperatively (panels e and f). The surface of the plug cartilage was as smooth as that of the adjacent intact cartilage. R496 Arthritis Research & Therapy Vol 6 No 6 Kuroki et al. Table 1 Figure 4 Modified Mankin's histological scores Subscore Details Cartilage structure Normal 0 Surface irregularities 1 Pannus and surface irregularities 2 Clefts to transitional zone 3 Clefts to radial zone 4 Clefts to calcified zone 5 Complete disorganization 6 Cartilage cells Normal 0 Pyknosis, lipid degeneration hypercellularity 1 Clusters 2 Hypocellularity 3 Safranin-O, thionine, Alcian blue Normal 0 Slight reduction 1 Moderate reduction 2 Severe reduction 3 No staining 4 Tidemark integrity Intact 0 Destroyed 1 Postoperative weeks 4 and 8 By 4 weeks postoperatively the implanted osteochondral plug had united with the area of the subchondral bone (Fig. 4e,4f). Newly formed reparative tissue, which stained faintly with safranin-O, covered the plug cartilage, and the plug cartilage was well stained (Fig. 4e). At 8 weeks (Fig. 5a) the plug cartilage was extremely well stained and was covered with the newly formed tissue, which was partly stained. The surface of the tissue was irregular. At 4 and 8 weeks postoperatively, the actual thickness of cartilage increased with the tissue. No reparative tissue was observed at the site of the defect at 4 or 8 weeks postoperatively. Postoperative weeks 12 and 24 Although the newly formed reparative tissue over the plug cartilage was not distinctive on macroscopic observation at 12 weeks postoperatively (Fig. 3e), histological observation revealed that reparative tissue covered the R497 Safranin-O/fast (original magnification green staining 10×) of plug cartilage and the shaved area (original magnification 10×). (a, b) No abnormalities were observed at day 0. The plug (between the arrows) was inserted flush with the surrounding articular surface and the site of the shaved area (between the triangles) was clearly recognizable. There was a space between the plug and the surrounding tissue. (c, d) At 2 weeks postoperatively, the plug (between the arrows) had subsided a little and newly formed tissue covered half of the plug. No tissue was observed at the site of the shaved area (between the triangles). There was a slight space between the plug and the surrounding tissue. (e, f) The implanted osteochondral plug (between the arrows) had united in the subchondral bone area by 4 weeks postoperatively. The newly formed reparative tissue, which stained faintly with safranin-O, covered the plug cartilage and the plug cartilage was well stained. No tissue was observed at the site of the shaved area (between the triangles). Bony union was observed between the plug and the host. plug cartilage (Fig. 5b). Actual thickness of cartilage increased with the tissue. The tissue was faintly stained and the plug was well stained (Fig. 5b). At 24 weeks postoperatively, no reparative tissue was observed over the plug (Fig. 5c). Therefore, the actual thickness of cartilage decreased at 24 weeks. The plug cartilage was stained with the safranin-O but slightly less intensely than the intact cartilage. No reparative tissue was observed at the site of the defect at 12 or 24 weeks postoperatively. Modified Mankin's score The mean scores for the plug cartilage at day 0 and weeks 2, 4, 8, 12 and 24 postoperatively were 0.33, 1.67, 2.00, 2.75, 3.50 and 2.50, respectively. Those of the corresponding sham-operated site were 0.00, 0.67, 0.00, 1.00, Available online http://arthritis-research.com/content/6/6/R492 Figure 5 10×) Safranin-O/fast green staining of plug cartilage (original magnification 10×). (a) At 8 weeks the plug cartilage (between the arrows) was extremely well stained and was covered with newly formed tissue, which was stained partly and the surface of the tissue was irregular. (b) At 12 weeks reparative tissue covered the plug cartilage (between the arrows) and the tissue was faintly stained and the plug was well stained. (c) At 24 weeks postoperatively no reparative tissue covered the plug (between the arrows) and the plug cartilage was stained with safranin-O, but slightly less strongly than the plug cartilage at 8 or 12 weeks. 0.50 and 0.75, respectively. No differences between scores existed on the sham side, but on the grafted side the scores differed significantly between day 0 and week 12 (Fig. 6; P < 0.05). Mann-Whitney U-test revealed that scores between the right and left knees at 4, 12 and 24 weeks postoperatively were significantly different (P < 0.05). Ultrasound data on cartilage Signal intensity at day 0 was the same in the grafted and sham sides at 1.5 (relative value). At 2 weeks postoperatively the signal intensity on both sides had decreased. The decrease was not significant on the sham side, but it was significant on the grafted side (P < 0.001; Fig. 7a and Table 2). At 8 weeks, although signal intensity on both sides had decreased, the difference was not significant on the sham side but it was significant on the grafted side (P < 0.001; Fig. 7a and Table 2). Between 2 and 24 weeks postoperatively, the intensity on the grafted side was lower than that on the sham side (P < 0.05 at 8 weeks; Fig. 7a). At 24 weeks the values were 0.6 on the grafted side and 1.0 on the sham side. Signal duration on day 0 was 0.6 µs on the grafted side and 0.5 µs on the sham side. At 2 and 8 weeks postoperatively, Figure 6 respectively at (24W) The day mean 0postoperatively (0D) modified and atMankin's weeks were20.33, scores (2W),1.67, 4of(4W), the 2.00, plug 8 2.75, (8W), cartilage 3.50 12 (12W) (open and 2.50, circles) and 24 at day 0 (0D) and at weeks 2 (2W), 4 (4W), 8 (8W), 12 (12W) and 24 (24W) postoperatively were 0.33, 1.67, 2.00, 2.75, 3.50 and 2.50, respectively. Those of the corresponding sham site (closed circles) were 0.00, 0.67, 0.00, 1.00, 0.50 and 0.50 points, respectively. The Kruskal–Wallis test and Scheffe's F-test revealed that, on the grafted side, the score differed significantly between day 0 and week 12 (* P < 0.05). Mann–Whitney U-test revealed that the score differed significantly (P < 0.05) between the sham and experimental sites at 4, 12 and 24 weeks postoperatively (not indicated in the graph). the differences in signal duration between the sides were 0.2 and 0.3 µs, respectively (P < 0.05, Fig. 7b). At 12 weeks the difference in signal duration between the sides was 0.1 µs. At 24 weeks postoperatively, the signal duration on both sides was approximately the same (0.5 µs; Fig. 7b). At day 0 the interval between signals on both sides was the same (0.5 µs). Between 2 and 8 weeks postoperatively, the difference between them was 0.1 or 0.2 µs. At 12 weeks the difference increased (0.3 µs), but at 24 weeks postoperatively no difference was observed between the two sides (for both the interval was 0.6 µs; Fig. 7c). Relationship between ultrasound data and modified Mankin's score Spearman's rank-order correlation revealed that the modified Mankin's score was significantly correlated with signal intensity (P = 0.0176, r= -0.36; Fig. 8a) and with signal duration (P = 0.0269, r = 0.34; Fig. 8b). The signal intensity was also significantly correlated with the score for category 'cartilage structure' in the modified Mankin's score (P = 0.0343, r = -0.32; Fig. 8c) and that for 'cartilage cells' (P = 0.0499, r = -0.30; Fig. 8d), but the correlation was not significant for 'staining' or 'tidemark'. The correlations between signal duration and the scores for structure (P = 0.0557, r = 0.29) and cartilage cells (P = 0.4630, r = R498 Arthritis Research & Therapy Vol 6 No 6 Kuroki et al. (a) Intensity (index of stiffness)(relative value) Figure 7 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 * ** 0D(N=3) ** 2W (N=3) 4W(N=4) 8W(N=4) 12W (N=4) 24W(N=4) Duration time (index of surface irregularity) (µsec) (b) 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 * * 0D(N=3) 2W (N=3) 4W(N=4) 8W(N=4) 12W (N=4) 24W(N=4) 4W(N=4) 8W(N=4) 12W (N=4) 24W(N=4) Interval (index of thickness) (µsec) (c) 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0D(N=3) 2W (N=3) (a) Change over time in ultrasound signal intensity (an index of stiffness) of the plug site (open circles) and the sham side (closed circles). circles) Intensity is represented as relative values. Repeated measures analysis of variance and post-hoc test (Scheffe's F-test) revealed that the signal intensity of the two sides differed significantly at 8 weeks (* P < 0.05) and that the signal intensity of the grafted side differed significantly among the groups (P values are presented in Table 2). Especially between day 0 (D0) and week 2 (2 W) and between weeks 4 (4 W) and 8 (8 W), the values differed significantly (** P < 0.001), but there was no difference between 2 W and 4 W or among the groups 8 W, 12 W and week 24 (24 W). (b) Change over time in signal duration (an index of surface irregularity) of the plug site and the sham side. Values are presented as µs. Signal duration of the two sides differed significantly at 2 W and 8 W (* P < 0.05). (c) Changes over time in interval between signals (an index of thickness) of the plug site and the sham side. Values are represented as µs. Values on both sides were the same (0.5 µs) at D0. After 2–8 weeks postoperatively, the values almost paralleled each other and the difference between them was 0.1 or 0.2 µs. At 12 weeks the difference increased to 0.3 µs, but by 24 weeks postoperatively no difference was observed, at which time the value for both sides was 0.6 µs. R499 Available online http://arthritis-research.com/content/6/6/R492 Table 2 Level of significance (P values) in signal intensity of plug cartilage 24 weeks Day 0 P < 0.001 12 weeks P < 0.001 8 weeks 4 weeks P < 0.001 P < 0.001 2 weeks P < 0.001 2 weeks P < 0.001 P < 0.05 P < 0.05 NS - 4 weeks P < 0.001 P < 0.001 P < 0.001 - - 8 weeks NS NS - - - 12 weeks NS - - - - The P value was not significant (NS) between weeks 2 and 4, or among weeks 8, 12 and 24. Values for signal intensity are presented in Fig. 7a. 0.11) were not significant. When mean values for signal intensity of the plug site were calculated among each of the six postoperative groups, the mean signal intensity was significantly and strongly correlated with the score (P = 0.0130, r = -0.91; Fig. 8e). Discussion Although the concept of ultrasound assessment is over a decade old [1,2,35,40], new techniques continue to be developed and reported. These techniques include measurement of surface fibrillation [3] and tissue thickness [6,8,9], comparison of the speed of sound [7], ultrasound backscatter [11,41], needle probe [5], use of high frequency (50 MHz) [4,41], real-time analysis [16], ultrasound indentation [17] and mechano-acoustic diagnosis [12], among others. It has been suggested that ultrasound examination is a sensitive method for evaluating structural properties [12,41], surface roughness [1-3] and cartilage thickness [6,8,9,17,40]. In our system [10,13,14], wavelet transformation was used to assess three indices, namely signal intensity (an index of stiffness), signal duration (an index of surface irregularity) and interval between signals (an index of thickness) [10,13]. These three indices may also be used in combination with arthroscopy, open surgery and saline bath for experimental purposes (Fig. 2a). In the present study we used the saline bath method. After surgery, the signal intensity of the sham cartilage dropped at 2 weeks and recovered 4 weeks postoperatively. At 8 and 12 weeks the intensity dropped again and remained at this level until 24 weeks (Fig. 7a). The sham cartilage was exposed to air for almost the same period of time as was the OCG. Therefore, some effects of this exposure might be present in both sham and plug cartilage. A study reported that 30 min drying of cartilage resulted in patchy necrosis [42]. Another study revealed that ultrastructural changes occurred in chondrocytes after arthrotomy with 1 hour exposure to air [43]. However, 6 weeks after the arthrotomy the chondrocytes had fully recovered from the changes that were noted immediately after exposure to air. Although we observed neither patchy necrosis nor changes in chondrocytes in the sham knees, exposure to air might have some harmful effects on cartilage stiffness. In the plug site on the OCG side a postoperative drop in intensity was also observed (Fig. 7a), but there were some differences from the sham side. At 2 weeks postoperatively the drop in intensity at the plug site was greater than that at the sham site, and recovery 4 weeks postoperatively was limited (Fig. 7a). Although the intensity in both the sham and plug sites was reduced at 8 weeks, the difference between sites was significant (P < 0.05). The intensity at the plug site (about half that at day 0) was then maintained until 12 weeks, but had again dropped a little at 24 weeks. The OCG might have been responsible for these differences. Macroscopic observation revealed that surface of the plug site survived well and underwent repair. At 24 weeks postoperatively in particular, the surface was glossy and as smooth as that of the adjacent, intact cartilage (Fig. 3f). Histological observation, on the other hand, revealed that the plug had sunk or tilted a little at 2 weeks postoperatively or earlier, and that newly formed tissue covered the plug cartilage (Fig. 4c). This histological observation and the drop in signal intensity might be related to each other. The intensity of the newly formed tissue that covered the plug cartilage was lower than that of the sham cartilage (Fig. 7a). At 8 and 12 weeks postoperatively in particular, the intensity was about half that at day 0. Because of this low intensity the tissue could have become detached from the plug cartilage before 24 weeks postoperatively (Fig. 5c). The signal intensity may indicate that the tissue was not sufficiently stiff for weight bearing during these 24 weeks. Use of arthroscopy in this system may detect such weak tissue or cartilage in vivo, before it detaches from the host tissue. Also, we observed no tissue regrowth in the shaved area. Based on the histological findings, we speculate that early weight bearing induced slight plug sinking or tilting, and that the shaved area came into direct contact with the patella. The pressure from the patella and its movement might have prevented the growth of new tissue in the shaved area. R500 Arthritis Research & Therapy Vol 6 No 6 Kuroki et al. Figure 8 (a) sham Signal side,intensity closed circles; (an index n= of44, stiffness) P = 0.0176, correlated r = -0.36) significantly with the modified Mankin's score (plug site, open circles; corresponding site on the sham side, closed circles; n = 44, P = 0.0176, r = -0.36). (b) Signal duration (an index of surface irregularity) correlated significantly with the score (n = 44, P = 0.0269, r = 0.34). Signal intensity also correlated with the score categories (c) 'cartilage structure' (n = 44, P = 0.0343, r = -0.32) and (d) 'cartilage cells' (n = 44, P = 0.0499, r = - 0.30) of the modified Mankin's score. (e) When mean values of signal intensity of the plug site were calculated among each of the six groups, the mean was significantly and strongly correlated with the total modified Mankin's score (n = 6, P = 0.0130, r= -0.91). D0, day 0; 2-w, 2 weeks; 4-w, 4 weeks; 8-w, 8 weeks; 12-w, 12 weeks; 24-w, 24 weeks. R501