SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Objective

To evaluate the rate of progression of cartilage loss in the knee joint using magnetic resonance imaging (MRI) and to evaluate potential risk factors for more rapid cartilage loss.

Methods

We evaluated baseline and followup MRIs of the knees in 43 patients (minimum time interval of 1 year, mean 1.8 years, range 52–285 weeks). Cartilage loss was graded in the anterior, central, and posterior regions of the medial and lateral knee compartments. Knee joints were also evaluated for other pathology. Data were analyzed using analysis of variance models.

Results

Patients who had sustained meniscal tears showed a higher average rate of progression of cartilage loss (22%) than that seen in those who had intact menisci (14.9%) (P ≤ 0.018). Anterior cruciate ligament (ACL) tears had a borderline significant influence (P ≤ 0.06) on the progression of cartilage pathology. Lesions located in the central region of the medial compartment were more likely to progress to more advanced cartilage pathology (progression rate 28%; P ≤ 0.003) than lesions in the anterior (19%; P ≤ 0.564) and posterior (17%; P ≤ 0.957) regions or lesions located in the lateral compartment (average progression rate 15%; P ≤ 0.707). Lesions located in the anterior region of the lateral compartment showed less progression of cartilage degradation (6%; P ≤ 0.001). No specific grade of lesion identified at baseline had a predilection for more rapid cartilage loss (P ≤ 0.93).

Conclusion

MRI can detect interval cartilage loss in patients over a short period (<2 years). The presence of meniscal and ACL tears was associated with more rapid cartilage loss. Cartilage lesions located in the central region of the medial compartment showed more rapid progression of cartilage loss than cartilage lesions in the anterior and posterior portions of the medial compartment. The findings in this study suggest that patients entering clinical trials investigating antiarthritis regimens may need to be randomized based on location of the lesion.

Conventional radiography is widely used to evaluate the long-term progression of osteoarthritis (OA) and is able to clearly depict the established hallmarks of OA, namely, joint space narrowing, subchondral sclerosis, subchondral cyst formation, and osteophytosis. Conventional radiography is limited by its inability to directly visualize articular cartilage, the tissue in which OA is thought to begin (1). Magnetic resonance imaging (MRI) offers the distinct advantage of detecting signal and morphologic changes in articular cartilage and is the most sensitive and accurate test for evaluating the articular cartilage noninvasively; MRI has been used to detect articular cartilage changes such as cartilage swelling (or “blistering”), surface fraying, fissuring, and varying degrees of cartilage thinning (2, 3). Cartilage-sensitive MRI techniques, such as spoiled gradient-echo (GE) or fast spin-echo (FSE) sequences, have been shown to have a significant correlation with arthroscopic grading scores (r = 0.705) and thus can be used in an effective, noninvasive evaluation of the knee joint (4–7).

At this time, however, little is known about how such cartilage lesions develop and what their natural progression is over time. The purpose of our study was to evaluate the rate of progression of cartilage loss associated with OA of the knee joint using MRI and to evaluate potential risk factors for more rapid cartilage loss. Ultimately, we wanted to determine the prognostic significance of cartilage defects and associated articular disorders identified on knee MRIs. To accomplish this, we performed a retrospective review of MRIs in 43 patients who had undergone repeat MRIs of the knee over a relatively short observation period of an average of 1–2 years.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

This retrospective study identified 43 patients within the Radiology Information System database (Stanford University Medical Center) who had undergone repeat MRI of the same knee on 2 occasions, with a minimum time interval of 1 year between studies (mean 1.8 years, range 52–285 weeks). These MRIs were considered a baseline and followup study in the context of our evaluation; all studies had been performed between October 1992 and October 1998. Patients in this study included 21 women and 22 men, ages 17–65 years (mean 51.4 years). Of the 43 patients identified, 21 had sustained a sports-related injury, 10 experienced an accidental fall or motor vehicle accident, 4 were diagnosed as having OA, and 5 had other causes of joint pain (soft tissue tumors [2 patients], recurrent hemarthrosis, rheumatoid arthritis, and pigmented villonodular synovitis). There were 3 patients whose histories were unknown at the time of submission of this report. Among the 43 patients studied, 26 had a tear of either the medial or lateral meniscus as determined by MRI. Nineteen patients had a tear of the anterior cruciate ligament (ACL) or had previously undergone ACL repair.

Knee MRI.

Knee MRI was performed using a 1.5T unit (LX or LXII, version 4.7–5.2; GE Medical Systems, Milwaukee, WI) with the use of an extremity coil. A standard knee MRI protocol was used, which included cartilage-sensitive sequences; specifically, a sagittal and coronal proton density–weighted FSE (3,500–4,000/15–19 [repetition time msec/echo time msec]) was obtained with 3 mm–thick sections, a 1-mm intersection gap, echo train length (ETL) 6, bandwidth (BW) 20–22 kHz, 2 numbers of excitations (NEX), a 16-cm field of view (FOV), and a 512 × 192 matrix. Other sequences included a sagittal fat-saturated T2-weighted FSE (4,000/72) (3 mm–thick sections, 1-mm intersection gap, ETL 8, BW 20–22 kHz, 2 NEX, 16-cm FOV, and a 256 × 192 matrix), coronal T1-weighted SE (800/18) (4 mm–thick sections, 1-mm intersection gap, BW 20–22 kHz, 2 NEX, 16-cm FOV, and a 512 × 192 matrix), coronal T2-weighted fat-saturated FSE (4,000/54) (4 mm–thick sections, 1-mm intersection gap, ETL 8, BW 20–22 kHz, 2 NEX, 16-cm FOV, and a 512 × 192 matrix), and an axial fat-saturated proton density FSE (4,000/15) (ETL 8, 2 NEX, 14-cm FOV, and a 256 × 192 matrix). All sequences were used in both the baseline and followup studies of the knee. Images obtained with later versions of the scanner software clearly demonstrated an improvement in image quality. However, similar to other studies (4), results of imaging studies obtained in 1992 and 1993 showed sufficient contrast and signal-to-noise ratio in order to assess articular cartilage.

Evaluation of lesions.

Initially, cartilage lesions were graded independently by 3 investigators (SB, PL, GAB) who were experienced in detecting cartilage lesions by MRI. To minimize bias, the investigators were unaware of the order of the studies during the initial readings. Subsequent to these initial, independent readings, a consensus interpretation was formulated the order of the studies was revealed. A conscious effort was made to minimize bias during the consensus reading of the studies and to adhere to our initial impressions. We decided to eventually read the imaging studies side-by-side for 2 main reasons: 1) because the main premise of the work was to study the progression of specific cartilage lesions, we needed to ascertain the exact location and grade of the followup lesion, and 2) the inter- and intrareader reproducibility in the evaluation of cartilage lesions was not known.

Cartilage loss was graded in the anterior, central, and posterior regions of the medial and lateral knee compartments on a scale from 0 to 6 (normal = grade 0; signal heterogeneity [focal or diffuse signal heterogeneity with an intact cartilage surface] = grade 1; superficial fraying = grade 2; fissuring = grade 3; thinning <50% = grade 4; thinning >50% = grade 5; and full-thickness cartilage loss = grade 6). For clarification, “fraying” is characterized by small, superficial cartilage defects that occur only within 1 mm of the cartilage surface, and “fissuring” is seen as cartilage defects that extend ≥1 mm below the cartilage surface and are <1 mm in width. The size of the lesion was also scored. Lesions measuring ≤1 cm2 were grade A, lesions measuring >1 cm2 and seen on contiguous sagittal slices were grade B. On a few occasions, 2 lesions were identified within the same region. For statistical analysis purposes, we assigned a single grade to each region, choosing to use the higher grade lesion determined for that region.

Knee joints were also evaluated for the presence or absence of meniscal or ACL pathology. Menisci were evaluated for intrasubstance degeneration, as well as simple, complex, radial, flap, peripheral vertical, horizontal cleavage, and bucket handle tears. Other meniscal pathology included prior complete or partial meniscectomy as well as a meniscocapsular separation. For statistical analysis, the presence or absence of meniscal pathology was categorized into 2 groups. The first group included normal, intact menisci as well as those with intrasubstance degeneration. The second group included menisci that had demonstrated tears and those that had undergone meniscectomy.

Similarly, patients with ACL pathology according to MRI criteria were placed into one group, and patients without ACL pathology were placed into another group. Those who had partial and complete tears of the ACL as well as those who had undergone an ACL repair during the interval period were considered to have knee pathology present. Among the patients who had experienced partial or complete ACL tears, all had undergone an ACL repair by the time of the followup study with the exception of a single patient. A total of 11 patients had ACL reconstruction surgery between the baseline and followup MRI scans. Based on limited histories and imaging appearance, 4 patients had undergone meniscectomies during the interval period. Some of the patients may have had meniscal repair outside of our institution, but we did not have access to this information and therefore we could not include such information in our analysis. Patients who experienced ACL reconstruction or meniscectomy during the imaging interval were placed in their respective pathology groups.

We also evaluated the images for the presence or absence of medial and lateral collateral ligament pathology, graded as normal (grade 0), sprain (grade 1), and partial (grade 2) and complete tears (grade 3). Bone marrow edema was also assessed: none (grade 0), mild (extending <1 cm from the subchondral bone; grade 1), moderate (extending 1–2 cm from the subchondral bone; grade 2), and severe (extending >2 cm from the subchondral bone; grade 3). Additional assessments determined the presence or absence of osteophytes and subchondral sclerosis.

Statistical analysis.

The analyses consisted of chi-square tests for 2 and multiway tables for binary outcomes, and t-tests for quantitative outcomes (based on grade scores). In some cases, a logistic regression was also used, to test and control for the effect of repeated measurements on subjects. Lesions were graded from 0 through 6, and measured at baseline and followup. The scores were used in 2 different ways in the analyses:

  • For each subject, the change in score (numeric difference) between baseline and followup was used as a quantitative measurement. Differences between groups were then based on the difference between mean score change, and a t-test was used to test this difference, for example, to test whether size B lesions tended to show larger mean differences than size A lesions (P < 0.93).

  • An alternative, more robust measurement of progression is the sign of the change in score, i.e., an increase in score or not. This is a binary response, and comparisons among groups were based on chi-square tests for homogeneity of proportions.

This latter binary outcome was used to test for differences in the 6 “compartments” or groupings of regions, resulting in a 2 × 6 table, which showed an overall significance with P < 0.006. The individual differences (each compartment versus the rest) were based on 2 × 2 tables, and a Bonferroni adjustment was used to correct for the 5 comparisons.

We have repeated measurements on subjects, and comparisons among regions should really adjust for a subject effect. We tested for this by replacing the chi-square tests with logistic regression models, e.g., for the overall compartment effect we used

  • equation image

which adjusts for the subject effect. The effect of region is tested by fitting this model with and without the region effect. The likelihood ratio test results in a chi-square statistic adjusted for subject. Although the subject effect was strongly significant, the region effect was hardly changed.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Progression of cartilage lesions.

A total of 146 lesions were identified in the baseline studies of 43 patients. An additional 84 new lesions were identified in the followup studies of the same 43 patients (the baseline grade of these additional lesions was 0). Table 1 depicts the natural progression of these lesions. Grade 1 lesions were the most prevalent of the lesions, accounting for slightly more than half (52%) of the lesions found at baseline and 40% of the new lesions found on followup. Among the 53 grade 1A lesions identified at baseline (focal region of signal heterogeneity measuring 1 cm2 in area), 60% progressed, becoming either larger in size (from size A to B) or higher in grade (grades 2–6) (Figure 1). Eight of the grade 1A lesions (15%) did not change in size or grade, and 25% “reverted” to grade 0 (please see Discussion section for our use of the term “reverted”). When progression was seen, grade 1A lesions had a tendency to progress to grade 3A or 3B fissuring lesions. More specifically, 12 (23%) of the grade 1A lesions progressed to grade 3 fissuring lesions, a pattern of progression that predominantly occurred in the patellofemoral compartment. The remaining 10 grade 1A lesions (19%) progressed to varying degrees of cartilage thinning (more or less than 50%, including full-thickness cartilage loss) and surface fraying. An example of a grade 1A lesion and subsequent cartilage loss is depicted in Figure 2.

Table 1. Natural progression of cartilage lesions*
Baseline gradeFollowup grade
01A1B2A2B3A3B4A4B5A5B6A6B
  • *

    Values shown within the matrix indicate the number of graded lesions found at baseline (vertical axis) and their corresponding followup grade (horizontal axis). For example, a total of 3 lesions that were identified at baseline as grade 1B subsequently became grade 5B lesions on followup. In another example, a total of 6 grade 3A lesions found at baseline did not change upon followup, while a total of 13 grade 1A lesions reverted to grade 0. Grading scale: 0 = normal; 1 = signal heterogeneity (focal or diffuse signal heterogeneity with an intact cartilage surface); 2 = superficial fraying; 3 = fissuring; 4 = thinning <50%; 5 = thinning >50%; 6 = full-thickness cartilage loss. Size of the lesion was also scored. Lesions measuring ≤1 cm2 were grade A, and lesions measuring >1 cm2 and seen on contiguous sagittal slices were grade B. Thus, a grade 3B lesion is a lesion that demonstrates fissuring involving an area >1 cm2.

0241353157361511
1A138103175122010
1B51110011020300
2A0001020000100
2B0010401011000
3A01001610101001
3B0000002000200
4A0000010100000
4B0000000013201
5A0000000000101
5B0000000011903
6A0000000000103
6B0000000000004
thumbnail image

Figure 1. Graded lesions grouped according to percentage progressed (increased either in size or grade), remained stable (no change), or changed to a lower grade.

Download figure to PowerPoint

thumbnail image

Figure 2. Grade 1A lesion and progressive cartilage loss. A, Baseline study, demonstrating a small focus of signal heterogeneity (arrows) involving an area of <1 cm2 (grade 1A), which is located in the central region of the lateral tibiofemoral compartment. Note the adjacent foci of increased (solid arrow) and decreased (dashed arrow) signal intensity. Articular cartilage in the anterior and posterior regions appeared intact and without focal signal abnormalities. Both the anterior and posterior articular cartilage are normal (grade 0). B, Followup study performed >12 months later, showing that the central lesion identified at baseline had progressed to full-thickness cartilage loss (grade 6A) (solid arrows). Note the new foci of signal heterogeneity (grade 1A) in the central region (low signal intensity; dashed arrow). The articular cartilage of the anterior and posterior regions remained normal in appearance (grade 0).

Download figure to PowerPoint

In comparison, 21% of the grade 1B lesions had a tendency to progress to varying degrees of cartilage thinning (more or less than 50% of the original cartilage thickness) instead of fissuring, which occurred in only 2 cases (8%). Grade 1B lesions also appeared to be more resistant to change; this pattern of change was observed in nearly half of these lesions. Similar to grade 1A, one-fourth of the grade 1B lesions reverted to a smaller size lesion or to grade 0 (Figure 1).

Of the 21 grade 3A lesions identified, 62% progressed in size or increased in grade, 29% did not change, and 9% reverted to a lower grade. Nearly one-half (n = 10) of the grade 3A lesions increased from size A to B.

Interestingly, despite the tendencies noted above, statistical analysis of the data revealed that no specific grade of lesion identified at baseline had a predilection for more rapid cartilage loss (P ≤ 0.93).

Analysis by compartments.

A statistical analysis between the medial and lateral tibiofemoral compartments was performed to compare rates of cartilage loss. Furthermore, to determine whether location of the lesion within each of these compartments influenced cartilage loss, the medial and lateral tibiofemoral compartments were further subdivided into anterior, central, and posterior regions, as shown in Figure 3 and described as follows. To determine the anterior, central, and posterior segments of the medial tibiofemoral compartment, we took the sagittal slice of the medial femoral condyle that had the largest anteroposterior dimension (typically, the center slice of the medial femoral condyle). Once we identified the appropriate slice, we took equidistant one-thirds along the curve of the chondral surface to define the anterior, central, and posterior segments of the divided curve. These divisions were extended in the lateral and medial directions to further define the medial and lateral aspects of the medial tibiofemoral compartment. A similar method was used with the lateral tibiofemoral compartment. For example, a cartilage lesion identified in the anterior medial tibiofemoral compartment would be one identified in the anterior one-third of either the femoral or tibial articular cartilage.

thumbnail image

Figure 3. Subdivision of medial and lateral tibiofemoral compartments into regions. Analysis was performed by dividing the compartments into anterior (gray line), central (white line), and posterior (dotted line) regions.

Download figure to PowerPoint

The statistical relationship between the anterior, central, and posterior regions of the medial and lateral tibiofemoral compartments, excluding the patellofemoral compartment, is depicted in Table 2. Since patellofemoral pathology is related to patellofemoral tracking abnormalities, presence of a medial patella plica, and infrapatellar impingement syndrome, and because these entities were not directly evaluated, the data from the patellofemoral compartment were removed from the statistical analysis. The study was designed to evaluate the risk factors for progressive cartilage loss in the tibiofemoral compartments. Factors that are thought to contribute to tibiofemoral disease include meniscal and cruciate ligament pathology, so specific attention was placed on these structures and they were included in the analysis. Special attention to the location of the lesion within the tibiofemoral compartment helped determine whether the rate of cartilage loss was influenced by location.

Table 2. Analysis of lesions by compartment*
CompartmentNo. of lesions at baseline (% of total)No. of lesions at followup (% of total)No. progressed (%)P
  • *

    AMTFC = anterior medial tibiofemoral compartment; CMTFC = central medial tibiofemoral compartment; PMTFC = posterior medial tibiofemoral compartment; ALTFC = anterior lateral tibiofemoral compartment; CLTFC = central lateral tibiofemoral compartment; PLTFC = posterior lateral tibiofemoral compartment.

  • Significance of the difference in rate of progression.

AMTFC18 (12)29 (13)16 (19)≤0.564
CMTFC27 (18)37 (16)24 (28)≤0.003
PMTFC19 (13)24 (10)14 (17)≤0.957
ALTFC9 (6)10 (4)5 (6)≤0.001
CLTFC8 (5)18 (8)13 (15)≤0.707
PLTFC9 (6)16 (7)13 (15)≤0.707

The rates of progression of lesions located in the anterior, central, and posterior regions of the medial tibiofemoral compartment were 19%, 28%, and 17%, respectively. Lesions located in the central region of the medial compartment were significantly (P ≤ 0.003) more likely to progress to more advanced cartilage pathology than lesions in the anterior (P ≤ 0.564) and posterior (P ≤ 0.957) regions of the medial tibiofemoral compartment or lesions located in the lateral compartment. In contrast, lesions located in the anterior region of the lateral compartment were significantly less likely to progress to cartilage degradation (P ≤ 0.001). An example of progressive cartilage loss in the medial tibiofemoral compartment is shown in Figure 4.

thumbnail image

Figure 4. Progressive cartilage loss in a tibiofemoral compartment. A, Baseline study. Sagittal fast spin-echo proton density image (3,500/15 [repetition time msec/echo time msec]) of the medial femoral condyle, showing an area of cartilage thinning involving <50% of the original cartilage thickness and extending over a surface area of >1 cm2 (grade 4B lesion). B, Followup scan, showing that the lesion had progressed to full-thickness cartilage loss (involving an area >1 cm2), corresponding to a grade 6B lesion (solid arrows). A new lesion (full-thickness cartilage loss ≤1 cm2, grade 6A; dashed arrow) has evolved in the anterior portion of the medial femoral condyle, where, in the previous study, no lesion was found.

Download figure to PowerPoint

Effect of meniscal tears on progression of articular cartilage pathology.

Among the 43 patients studied, 26 (60%) had a meniscal tear involving either the medial or lateral meniscus. Fifteen patients (35%) had isolated medial meniscal tears, 6 (14%) had isolated lateral meniscal tears, and 5 (12%) had both medial and lateral meniscal tears. Seventeen patients (40%) had no meniscal tears in either the baseline or the followup study. Twenty-two percent of cartilage lesions that were observed in the presence of a meniscal tear demonstrated an increase in grade of the cartilage lesion when compared with the baseline grade. In contrast, cartilage lesions that were observed in the absence of a meniscal tear increased to a higher grade in 14.9% of the cases. The difference between the 2 groups in terms of progression to a higher grade was statistically significant (P ≤ 0.018).

The size of the cartilage lesion, categorized as either size A (area ≤1 cm2) or size B (area >1 cm2), was not influenced by the presence or absence of meniscal tears. Among the cartilage lesions identified in the setting of a meniscal tear, 45% progressed to a greater size B (area >1 cm2), which was not significantly different (P ≤ 0.799) from the rate of progression of cartilage lesions identified in the absence of a meniscal tear (39%).

Effect of ACL tears on progression of articular cartilage pathology.

Nineteen of the 43 patients studied demonstrated a tear of the ACL or had previously undergone ACL repair. In patients with an intact ACL, 33% of the cartilage lesions progressed to a higher grade. Although the rate of progression of articular cartilage lesions to a higher grade was greater (49%) in patients with ACL pathology, the difference was only of borderline significance (P ≤ 0.06). Furthermore, the original size (size A or B) of the lesion was not significantly influenced by ACL pathology (P ≤ 0.89).

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

The current study indicates that MRI is capable of detecting cartilage loss in a relatively short time. MRI can predict areas of articular cartilage that are prone to more rapid disease progression and can identify early lesions that may progress to further cartilage loss.

Comparison of the anterior, central, and posterior regions of the medial and lateral compartments demonstrated that changes in cartilage integrity are significantly more likely to occur in the central region of the medial tibiofemoral compartment; the central portion of the medial tibiofemoral compartment is prone to more rapid progression of cartilage loss. This may be the result of higher biomechanical loads in this compartment.

Studies of walking mechanics have shown that the knee is loaded asymmetrically. The medial tibiofemoral compartment of the knee is subject to significantly higher loads relative to the lateral compartment. In addition, the outcome of treatment for medial compartment OA has been related to the magnitude of the load on the medial compartment during walking (8–11). Harman et al have found that knees in a range of various alignments demonstrate articular cartilage wear patterns in the central to anterior regions of the medial tibial plateau (12). Furthermore, in conditions resembling the stance phase of a normal gait cycle, a computer model of the weight-bearing surfaces of the total knee found the central portion of the medial tibial plateau and the central/posterior portions of the lateral tibial plateau to be the initial contact points during loading of the knee (13). These studies corroborate our findings that cartilage lesions located in the central region of the medial compartment are prone to more rapid progression of cartilage loss than cartilage lesions in the anterior and posterior portions of the medial compartment.

There are several cartilage lesions identified by both arthroscopists and radiologists, with one lesion in particular generating much interdisciplinary (and intradisciplinary) debate. This is the MR grade 1 lesion, and many believe this lesion is difficult to diagnose with either MRI or arthroscopy. The results of our study, which demonstrate a relatively high rate of persistence or progression of grade 1 lesions, underscore the need for future clinical studies, advanced MRI, and histologic/arthroscopic correlation to better characterize these findings. There have already been substantial efforts made to better characterize this lesion, as described below.

The grade 1 lesion is a focus of signal heterogeneity within an intact cartilage surface, as defined within our MRI classification system and by others using similar classification systems (3, 7). The arthroscopic correlate to this lesion has been suggested to be cartilage “softening,” “swelling,” or “blistering,” which is also characterized as a grade 1 lesion in the arthroscopic grading scheme (14, 15). Histologically, grade 1 lesions are thought to represent anatomic aberrations, especially in the early changes of OA. Donohue et al showed that indirect blunt trauma to canine articular cartilage is able to produce ultrastructural changes, namely, the disruption of the collagenous network within the extracellular matrix, while the articular surface remains intact (16). More specifically, some authors believe that focal alterations in the network of collagen fibrils result in increased local hydration of cartilage, which is seen as aberrant signal intensity and disruption of the normal, laminar appearance on MRI (17, 18). They may represent microfissures that are below the spatial resolution of the standard MRI pulse sequences. They may also represent gross changes in composition in the extracellular matrix (e.g., breakdown of collagen, change in proteoglycan composition, or change in water content). In contrast, focal proteoglycan depletion in the articular cartilage of rat knees has been shown to be a less likely reason for grade 1 lesions (19).

Regardless of the etiology of these lesions, our study suggests that they are important given that 29 (38%) of 77 areas progressed to higher grade lesions (grades 2–6). Grade 1 signal heterogeneity lesions may represent actual articular cartilage derangement and may serve as predictors of future articular cartilage degeneration. Thus, detection of grade 1 lesions, using even the current MRI sequences, should warn of potential cartilage loss, and close attention should be directed to the affected articular cartilage on the followup studies. It is hoped that further development of high-resolution MRI sequences will elucidate the nature of these lesions, and that sodium MRI or gadolinium-enhanced MRI may characterize the biochemical changes (20). Recent work by Mosher et al, using a 3.0T magnet, demonstrated a significant increase in T2 relaxation of the transitional zone of patellar articular cartilage in symptomatic patients when compared with asymptomatic controls (21). Akella et al were able to correlate the loss of proteoglycan from the cartilage matrix with MRI signal alterations using a 4.0T magnet (22). Studies such as these are highly encouraging.

Despite the substantial proportion of grade 1 signal heterogeneity lesions that progressed, one-fourth of the lesions reverted to grade 0. The somewhat capricious nature of grade 1 lesions and the phenomenon of reversion may be explained as follows. Lesions were not seen on the followup study because of partial volume averaging. Grade 1 lesions seen at baseline represent artifacts of partial volume averaging or the magic angle phenomenon. To minimize this possibility, we confirmed the lesion presence on 2 contiguous slices and/or in an alternative imaging plane. Furthermore, the magic angle phenomenon is a less likely possibility since it is a homogeneous signal artifact (23, 24). Reversion to a smaller or lower grade lesion may, in fact, also represent repair and healing.

Indeed, a review of the literature indicates that detection of grade 1 lesions has been challenging at best. There appears to be a controversy between the use of MRI and arthroscopy, as to which is the optimal study for detection of grade 1 lesions. Some authors believe arthroscopy is best for this, while others believe MRI has a unique ability to detect subsurface lesions. For example, in a retrospective analysis of 63 patients, the sensitivity of MRI evaluation of grade 1 lesions (when compared with arthroscopy) was 14.3% despite the use of MRI machines with identical magnetic field strength (1.5T magnet) and similar cartilage-sensitive sequences (proton density FSE) (14). Similarly, a study of 320 patients who had experienced acute trauma, using MRI and arthroscopy, showed that MRI diagnosed only 14% of the grade 1 lesions (25). In contrast, MRI was able to detect between 75% and 94% of lesions characterized as cartilage erosions and 100% of full-thickness cartilage loss.

The discrepancy between MRI and arthroscopy seen with grade 1 lesions can be attributed to the fact that lesions without surface irregularities are inherently difficult to diagnose arthroscopically and, in fact, the authors of such studies have found that MRI can “overgrade” intracartilaginous lesions relative to arthroscopy (26). Furthermore, those investigators concede that the arthroscopic correlate to an MRI grade 1 lesion, which is referred to as cartilage softening or swelling, can be appreciated only by gentle palpation of the articular surface. This is an arthroscopic technique not normally used on the entire cartilage surface, resulting in a high rate of false-negative results. We also can presume that this palpation technique is prone to significant subjectivity. We do not know whether intracartilaginous signal abnormalities seen with MRI necessarily translate into altered mechanical properties. It is conceivable that early cartilage lesions may have stiffness similar to the adjacent normal cartilage, thus making them difficult to detect with the arthroscopic technique. MRI is likely to prove to be an invaluable tool, more sensitive than arthroscopy for evaluation of early cartilage changes and subsurface lesions.

The natural course of cartilage loss appears to be accelerated in the presence of meniscal tears. In our study, we found a strong relation between meniscal tears and lesions that progressed more rapidly. Dandy and Jackson, as well as Frankel et al, have demonstrated that meniscal abnormalities have led to enhanced chondromalacia as a result of abnormal articular forces created by such an alteration (27, 28). Photoelastic studies by Radin et al showed that the meniscus serves to protect articular cartilage by distributing load throughout the articular surface and preventing focal stress concentrations (29). The borderline influence of ACL tears on progression of cartilage derangements may be partially explained by the fact that nearly all torn ACLs were replaced, thus preventing rapid cartilage loss.

Limitations of this study include its retrospective design, limited clinical information, small sample size, varying ages of the patients, and population bias toward patients with chronic knee pain and dysfunction. Furthermore, the MRI readers were ultimately not blinded to the order of the studies, despite the best intentions (see Patients and Methods). Knowledge of the order may introduce bias in image interpretation as well as analysis of disease progression. Also, the sorting of lesion size into 2 different groups was perhaps too crude. Perhaps a more specific size measurement is needed, because we expect that cartilage lesion size would increase more rapidly in patients with a meniscal tear or ACL injuries. However, despite the small sample size and the absence of proper stratification of patient groups, we showed that MRI was able to detect progression of cartilage loss within a short observation period of 1–2 years. Furthermore, patients with arthritis who are enrolled in clinical trials may have to be randomized according to lesion location and the presence of meniscal or cruciate ligament tears.

Unlike conventional radiography, MRI can detect cartilage loss in patients with OA within a short observation period. Cartilage lesions such as focal signal heterogeneity, surface fraying, fissuring, thinning, and full-thickness cartilage loss can be identified and followed longitudinally. No specific lesion has a predilection for more rapid cartilage loss. Grade 1 lesions progress to a higher grade (grades 2–6) in ∼40% of the cases over the observation period, enhancing the importance of this early lesion. The cartilage lesions identified in the central region of the medial compartment are prone to more rapid progression of cartilage loss when compared with those in the anterior and posterior regions of the medial tibiofemoral compartment or the lateral tibiofemoral compartment. Lesions located in the anterior region of the lateral compartment showed significantly less cartilage degradation. Furthermore, meniscal (and possibly ACL) tears predispose knee articular cartilage to more rapid cartilage loss.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES