Musculoskeletal Imaging Original Research

Transcription

1 Musculoskeletal Imaging Original Research Laundre et al. MRI Detection of Meniscal Tears in Patients With CL Injury Musculoskeletal Imaging Original Research ryan J. Laundre 1 Mark S. Collins 1 Jeffrey R. ond 1 Diane L. Dahm 2 Michael J. Stuart 2 Jayawant N. Mandrekar 3 Laundre J, Collins MS, ond JR, Dahm DL, Stuart MJ, Mandrekar JN Keywords: acute anterior cruciate ligament injury, anterior cruciate ligament tear, knee MRI, lateral meniscus tear, posterior horn of the lateral meniscus, sports medicine, trauma DOI: /JR Received November 20, 2008; accepted after revision January 28, Department of Radiology, Mayo Clinic, 200 First St., SW, Rochester, MN ddress correspondence to. J. Laundre. 2 Department of Orthopedic Surgery, Mayo Clinic, Rochester, MN. 3 Division of iostatistics, Mayo Clinic, Rochester, MN. JR 2009; 193: X/09/ merican Roentgen Ray Society MRI ccuracy for Tears of the Posterior Horn of the Lateral Meniscus in Patients With cute nterior Cruciate Ligament Injury and the Clinical Relevance of Missed Tears OJECTIVE. The sensitivity of MRI for the detection of tears of the posterior horn of the lateral meniscus (PHLM) is lowest compared with that for tears at other meniscal locations, and the presence of simultaneous acute anterior cruciate ligament (CL) injury decreases overall MRI sensitivity for meniscal tears. We rereviewed surgically proven cases of missed meniscal tears in knees with an acute CL injury to determine why PHLM tears may be missed. We also investigated whether the missed PHLM tears were clinically significant that is, if these tears required surgical repair or resection. MTERILS ND METHODS. We reviewed the medical records of 120 patients (< 40 years old) who underwent arthroscopic CL reconstruction within 6 weeks after MRI to identify MRI-missed meniscal tears. Missed PHLM tears were categorized as clearly evident, occult, or subtle during MRI rereview. The two-touch-slice rule served as the primary criterion for tear diagnosis. Secondary MRI findings, including an abnormal superior popliteomeniscal fascicle and apparent far lateral extension of the meniscofemoral ligament, and additional clinical, surgical, and initial MRI findings were compared between meniscal tear groups. RESULTS. The majority (19/28) of missed tears involved the PHLM: Five were clearly evident at the time of rereview, all of which required surgical treatment; six were occult, none of which was treated; and eight were subtle, four of which were treated. Knees with PHLM tears were more likely to have abnormal superior popliteomeniscal fascicles (p = 0.002) and apparent far lateral extension of the meniscofemoral ligament (p = 0.003) than knees with normal lateral menisci. CONCLUSION. Clearly evident missed PHLM tears would not have been missed if the two-touch-slice rule had been strictly applied at prospective MRI interpretation. Unavoidably missed PHLM tears were not clinically significant. However, there were no clinical or secondary MRI findings specific to missed PHLM tears in the setting of acute CL injury. M RI of the knee is a valuable and accurate tool for diagnosing ligamentous and meniscal tears [1] especially in patients with sportsrelated knee injuries. MRI serves as a useful screening tool in patients with an acute knee injury, often allowing orthopedic surgeons to identify those who would benefit from surgery [2 5]. However, the sensitivity of MRI for the detection of meniscal tears is not yet 100%, and athletes have occasionally returned to activity with undiagnosed meniscal tears on the basis of a normal MRI examination [6]. The sensitivity of MRI for the detection of a lateral meniscus tear is significantly lower than that for the detection of a medial meniscus tear: 79.3% compared with 93.3%, respectively [7]. The lowest MRI sensitivity has been attributed to tears involving the posterior horn of the lateral meniscus (PHLM) [8 12]. recent study showed that lateral meniscus tears were more frequently missed if the tear involved only one third of the meniscus or if it was located in the posterior horn [8]. The presence of anterior cruciate ligament (CL) tears, which are often associated with longitudinal PHLM tears [13], also decreases the overall sensitivity of MRI for meniscal tear detection [8 10, 14]. To our knowledge, no study has fully investigated why PHLM tears are missed on MRI. Little is known about why more missed tears occur posterolaterally or with concomitant ligamentous injury. Several prior investigations have been performed involving MRI and surgical correlation or identification of possible diagnostic pitfalls [11, 12, 15]. One study of 41 patients with CL tears JR:193, ugust

2 Laundre et al. imaged within 6 weeks of injury reconfirmed that MRI has a low sensitivity for the detection of lateral meniscus tears (57%), with all false-negative cases occurring in the posterior horn [16]. Some investigators have hypothesized that recognition of secondary findings such as subchondral marrow edema [17], posterior pericapsular edema, abnormal superior popliteomeniscal fascicles [18, 19], apparent far lateral extension of a meniscofemoral ligament [20], and meniscal extrusion [21] would help improve MRI detection of lateral meniscal tears. However, despite these efforts and the advancements in imaging technology, including imaging using higher field strengths [1, 22, 23] or using fast spin-echo (FSE) or radial gradient imaging sequences [24 26], the sensitivity of MRI for the detection of meniscal tears has not significantly improved. The objective of our study was to review surgically proven cases of missed meniscal tears in knees with acute CL injuries to determine why PHLM tears were missed on MRI. We specifically tested previously described primary and secondary MRI findings that have been associated with missed PHLM tears. We also investigated whether the missed PHLM tears were clinically significant that is, either requiring surgical repair or resection or requiring conservative management. Materials and Methods Study Population Institutional review board approval was obtained before the performance of this study, allowing retrospective review of patient records and images without informed consent. ll patients younger than 40 years old who had an MRI-diagnosed CL injury (n = 948) and who underwent primary arthroscopic CL reconstruction at our institution from January 1, 1999, until May 31, 2006 (n = 809), were cross-referenced using a computer search of surgical and radiology databases (Fig. 1). Patients who underwent MRI within 6 weeks of knee arthroscopy (n = 153) were identified to reduce the possibility of an undiagnosed injury occurring during the period between the MR examination and knee arthroscopy [8]. Clinical History and Intraoperative Findings We retrospectively reviewed clinical records and excluded patients who had a chronic CL tear or had undergone prior ipsilateral knee arthroscopy or surgery, identifying 122 knees with acute CL injury in 120 unique patients (median age, 18.5 years) (Fig. 1). The date of injury was also recorded. Knee arthroscopy findings served as Patients with MRI-diagnosed CL injury, < 40 years old (948 patients) Occult (6) Missed PHLM tears (19) Clearly evident (5) (Cross-referenced) Patients with MRI within 6 weeks of arthroscopy (120 patients) Subtle (8) the study s gold standard for determining the distribution (anterior horn, body, or posterior horn; with tears possibly involving one, two, or three segments); configuration (complex, bucket handle, displaced flap, oblique, radial, longitudinal peripheral, or partial thickness or split type); and surgical treatment (meniscal repair or resection or partial meniscectomy), if considered necessary, of any meniscal tears. Patients with primary CL repair, < 40 years old (809 patients) Unmissed Normal lateral meniscus PHLM tears (abnormal medial meniscus) (31) (21) Fig. 1 Flow diagram outlines study patients, all of whom underwent primary anterior cruciate ligament (CL) repair within 6 weeks after diagnostic MRI. Patients are grouped according to classification of tear of posterior horn of lateral meniscus (PHLM): missed PHLM tear, unmissed PHLM tear, or normal PHLM but abnormal medial meniscus at surgery. MRI Findings ll MRI examinations were performed at our institution with a 1.5-T superconducting magnet (Signa, GE Healthcare) and a dedicated knee extremity coil. For most of the examinations highresolution FSE proton density weighted and fat-suppressed FSE T2-weighted paired sequences were performed in the sagittal and coronal planes. Fat-suppressed FSE T2-weighted images were also obtained in the axial plane. FSE proton density weighted sequence parameters were as follows: 2 signals acquired; TR/TE, 2,000/18; echo-train length, 4; receiver bandwidth, 32 khz; matrix, ; field of view, cm; and slice thickness, 3 mm. Fat-suppressed FSE T2- weighted sequence parameters were 2 signals acquired; 4,000/50; echo-train length, 8; receiver bandwidth, 20 khz; matrix, ; field of view, cm; and slice thickness, 3 mm. Several of the older studies were performed using conventional spin-echo (SE) proton density weighted and conventional SE T2-weighted sequences in the sagittal and coronal planes in lieu of the FSE sequences. For those studies, the following parameters were used: 2 signals acquired; TR/ TE range, 2,000/20 60; matrix, ; field of view, cm; and slice thickness, 3 mm. Initial MRI findings The original (prospective) radiologist interpretation was analyzed and compared with the intraoperative findings specifically to confirm the presence, distribution, and configuration of any meniscal tears, most notably those involving the PHLM. The following findings were also noted: injuries to multiple knee ligaments; posterior cruciate ligament (PCL), medial collateral ligament (MCL), and lateral collateral ligament (LCL) abnormalities; and any bone contusions including, but not limited to, the lateral femoral condyle and lateral tibial plateau. MRI rereview findings If a PHLM tear was missed at the initial MRI interpretation, the intraoperative findings were used to more closely scrutinize the anatomic area of concern, and a consensus retrospective reinterpretation of the MRI study was performed on a dedicated, high-resolution interactive workstation (dvantage, GE Healthcare) by two experienced musculoskeletal radiologists with 10 and 17 years of experience, respectively. The two-slice-touch rule, previously described by De Smet and Tuite [27], served as the primary finding for a meniscal tear diagnosis: tear is highly likely present if two or more MR images show meniscal distortion or abnormal signal extending to an articular surface [28]. Each missed tear was categorized on the basis of its MR appearance at the time of rereview as clearly evident, occult (unavoidably missed), or subtly evident. Secondary findings were also assessed. The superior popliteomeniscal fascicles, considered abnormal if disrupted or absent on sagittal imaging, were classified as normal, probably normal, equivocal, probably abnormal, or abnormal. pparent far lateral extension of the meniscofemoral ligament insertion onto the lateral meniscus was diagnosed if the meniscofemoral ligament extended 14 mm lateral to the PCL [20]. Posterior pericapsular edema was considered present if abnormal T2 signal was seen in or around the posterior capsule of the 516 JR:193, ugust 2009

3 MRI Detection of Meniscal Tears in Patients With CL Injury TLE 1: Demographic Data and Clinical History of Patients With an cute nterior Cruciate Ligament Injury and Tear Involving the Posterior Horn of the Lateral Meniscus (PHLM) Knee Category No. of Knees Median Patient ge (y) lateral compartment on sagittal or axial imaging [18, 19]. lateral meniscus posterior root tear was diagnosed if MRI showed distortion and increased signal intensity within the meniscal root that extended to the articular surface. Lateral meniscus extrusion was diagnosed if the distance between the peripheral margin of the meniscal body and the peripheral margin of the tibial articular cartilage was less than 1 mm [21]. Similar image reinterpretation and imaging analysis were performed on two other groups of patients: those with prospectively diagnosed PHLM tears and those with surgically normal lateral menisci but surgically proven torn medial menisci. Median Interval (wk) etween Injury and MRI etween MRI and Surgery Prior Knee Injury ll knees Knees with PHLM tear Missed on MRI Not missed on MRI Statistical nalysis nalyses were performed on three groups of patients: those with missed PHLM tears (n = 19); those with unmissed PHLM tears (n = 31); and those with a normal lateral meniscus but a torn medial meniscus at arthroscopy (n = 21). Continuous variables were summarized using median values, and comparisons between the groups of interest were made using Wilcoxon s rank sum test. Categoric variables were summarized using frequencies and were compared using the chisquare or Fisher s exact test as appropriate. ll the tests were two-sided, and p values < 0.05 were considered statistically significant. nalyses were performed using SS software (version 9.0, SS Institute). Comparison of missed and unmissed PHLM tears served as an internal control to determine variables attributable to missed tears. Comparison of missed PHLM tears with respect to their appearance at the time of rereview (clearly evident, not evident, or subtly evident) was then performed to determine whether any MR findings contributed to missed opportunities in detecting PHLM tears. Finally, knees with PHLM tears and knees with a normal lateral meniscus but torn medial meniscus were compared to determine whether any MRI findings were specific to the knees with PHLM tears. Results One hundred twenty-two knees with an acute CL injury were identified in 120 unique patients, with a median age of 18.5 years (Table 1). Patients were imaged on average 1.3 weeks from the date of injury, and all underwent MRI within 6 weeks before arthroscopy (median = 1.3 weeks). The patients had experienced a variety of noncontact and contact knee injuries including twisting pivoting, hyperextension, deceleration, valgus varus stress, or jumping landing mechanisms of injury. Of the 122 knees with an acute CL injury, 102 meniscal tears were identified at surgery: 59 lateral meniscus tears (in 53 knees) and 43 medial meniscus tears (in 42 knees). Twenty-eight discrete meniscal tears were missed, 23 of which were in the lateral meniscus (Table 2). Twenty-eight meniscal tears diagnosed on the initial MRI interpretation were not present during arthroscopy (overcalled tears): 10 in the lateral meniscus and 18 in the medial meniscus. Thus, MRI sensitivity was 58.5% and 88.1% and MRI specificity was 85.5% and 77.5% for lateral meniscus and medial meniscus tears, respectively. The majority of missed meniscal tears occurred in the posterior meniscal horns, with 19 involving the PHLM (Table 3 and Fig. 1). Rereview of PHLM Tears Retrospective rereview of the 19 missed PHLM tears showed that five were clearly evident on rereview using the two-slice-touch rule. ll five of these posterior horn tears were surgically treated: Two were débrided and repaired with bioabsorbable suture and three were resected. Six were occult (unavoidably missed), and none of these tears was treated. Eight posterior horn tears were subtly evident. Four of these tears were surgically treated: Two tears were débrided and repaired with bioabsorbable suture, and two were resected (Table 3 and Fig. 1; p = 0.002). ll six of the PHLM tears that were not evident during rereview were surgically classified as partial-thickness or split-type tears compared with only three of the eight subtle PHLM tears (p = 0.015). Of those PHLM tears that were clearly evident, 60% were displaced and none was a partial-thickness or split-type tear. Last, three of the subtle and three of the clearly evident PHLM tears had abnormal or probably abnormal superior popliteomeniscal fascicles, whereas only two of the subtle PHLM tears had abnormal far lateral extension of the meniscofemoral ligament. The occult (unavoidably missed) tears did not have either of these findings (Table 3). There was no statistically significant difference between these subgroups when compared to the remaining variables including surgical findings, initial MRI findings, and rereview MRI findings. Missed PHLM Tears Versus Unmissed PHLM Tears Missed PHLM tears (n = 19) were less likely to have abnormal superior popliteomeniscal fascicles (p = 0.001) or abnormal meniscofemoral ligament extension (p = 0.030; Table 3) than unmissed PHLM tears TLE 2: Sensitivity of MRI for the Detection of Meniscal Tears in Knees With an cute nterior Cruciate Ligament Injury Location of Tear Total No. of Menisci No. of Intact Menisci No. of Torn Menisci No. of Missed Torn Menisci MRI Sensitivity for Meniscal Tear (%) Total No. of Discrete Tears No. of Missed Discrete Tears Discrete Tear Sensitivity (%) Lateral meniscus Medial meniscus Total JR:193, ugust

4 Laundre et al. TLE 3: Surgical, Initial MRI, and Rereview MRI Findings in Patients With Tears Involving the Posterior Horn of the Lateral Meniscus (PHLM) and Patients With a Normal Lateral Meniscus Surgical findings Findings (n = 31). There was also a correlating significant difference between the median distances of lateral meniscofemoral ligament extension (p = 0.005). No statistically significant difference between PHLM tears missed on prospective MRI and those detected on prospective MRI was detected for the following variables: clinical history data including patient age, injury-to-mri interval, MRI-to-arthroscopy interval, mechanism of injury, or prior ipsilateral knee injury (Table 1); surgical findings including the total number and locations ppearance of PHLM Tears Initially Missed on MRI at MRI Rereview Clearly Evident Occult Subtle of lateral meniscus or medial meniscus tears or PHLM tear configuration; MR sequence type; initial MRI findings including injuries to multiple knee ligaments and bone contusions; or MRI rereview findings including posterior pericapsular edema, lateral meniscus root tear, or lateral meniscus extrusion (Table 3). Missed on Initial MRI PHLM Tear Not Missed on Initial MRI Normal Lateral Meniscus No. of knees No. of medial meniscus tears No. of lateral meniscus tears No. of PHLM tears No. of surgically treated PHLM tears PHLM tear configuration Complex tear ucket-handle tear Displaced flap Oblique tear Radial tear Longitudinal peripheral tear Partial-thickness or split-type tear MRI findings Multiligamentous Injury one contusions MRI rereview secondary findings Superior popliteomeniscal fascicles Normal Probably normal Equivocal Probably abnormal bnormal Far lateral meniscofemoral ligament insertion bnormal extension ( 14 mm) Median distance of extension (mm) Posterior pericapsular edema Tear of lateral meniscus root Extrusion of lateral meniscus PHLM Tears Versus Surgically Confirmed Normal Lateral Menisci t arthroscopy, all knees in the normal lateral meniscus group (n = 21) had a medial meniscus tear involving the posterior horn; none had a lateral meniscus tear. Three lateral meniscus tears (14.3%) were diagnosed on the initial MR examination, but they were not present during arthroscopy (i.e., overcalled tears or false-positives). Compared with knees with normal lateral menisci (n = 21), both missed PHLM tears (n = 19) and all PHLM tears (n = 50) were more likely to have an abnormal superior popliteomeniscal fascicle (p = and 0.002, respectively; Table 3). None of the knees with a normal lateral meniscus had an 518 JR:193, ugust 2009

5 MRI Detection of Meniscal Tears in Patients With CL Injury abnormally classified superior popliteomeniscal fascicle. oth the presence of an abnormal meniscofemoral ligament extension and, if present, the median distance of that extension were not statistically significant when comparing knees with missed PHLM tears and knees with a normal lateral meniscus (p = and 0.231, respectively). However, when comparing all knees with PHLM tears and knees with normal lateral menisci, only the presence or absence of an abnormal lateral meniscofemoral ligament extension was significant (p = 0.003). The median distances of lateral meniscofemoral ligament extension for these two groups did not differ significantly (p = 0.294) (Table 3). There was no statistical significance between either the missed PHLM tears or all the PHLM tears and the knees with a normal lateral meniscus for the remaining variables including surgical findings, initial MRI findings, and rereviewed findings, as listed earlier. Discussion Clinical Relevance The sensitivity of MRI for the detection of lateral meniscus tears in the setting of acute CL injury in our study (58.5%) is comparable to sensitivities observed in previous studies that showed MRI has a lower sensitivity in the presence of CL tears (57 69%) [8 10, 14, 16]. The sensitivity of MRI for lateral meniscus tears in our study was lower than the pooled weighted sensitivity for the diagnosis of all lateral meniscus tears (79.3%) [7]. Furthermore, the sensitivity of MRI in our study for the detection of medial meniscus tears in knees with acute CL injury (88.1%) is comparable to that reported in a previous study (88%) [9] and is lower than the pooled weighted sensitivity for all medial meniscus tears (93.3%) [7]. Similar to the results of previous studies, we found that most missed meniscal tears with an associated acute CL injury were located in the posterior horns of the menisci. Of the 19 missed PHLM tears, all five tears that were clearly evident at the time of re-review were unstable at arthroscopy and required surgical treatment. The majority were classified as displaced flap-type tears (Fig. 2). lso, three of the tears were located very close to the lateral meniscus root. ll of these missed tears were clearly evident at rereview using the two-slice-touch rule that is, all showed abnormal meniscal signal or distortion on two or more MR images (Fig. 3). Fig year-old woman with noncontact, twisting knee injury and acute anterior cruciate ligament tear. Tear of posterior horn of lateral meniscus (PHLM) missed at initial MRI was clearly evident at MRI rereview. rthroscopy showed displaced-flap PHLM tear with flap fragment displaced into intercondylar notch. Displaced tear fragment was excised. During MRI rereview, displaced-flap PHLM tear was clearly evident. ll clearly evident PHLM tears were surgically repaired or resected. and, Consecutive coronal fast spin-echo proton density weighted images arranged posterior () to anterior () show displaced-flap tear of PHLM with meniscal flap fragment displaced into intercondylar notch (arrows), clearly depicting tear noted at arthroscopy. Note truncated PHLM (arrowhead, ). Fig year-old woman with noncontact, twisting knee injury and acute anterior cruciate ligament tear. Tear of posterior horn of lateral meniscus (PHLM) missed at initial MRI was clearly evident at MRI rereview. rthroscopy showed unstable oblique tear through PHLM, which was débrided and repaired with absorbable suture. During MRI rereview, PHLM tear was clearly evident on sagittal images near posterior lateral meniscal root after applying two-slice-touch rule [27]. ll clearly evident PHLM tears were surgically repaired or resected. and, Consecutive sagittal fast spin-echo proton density weighted images arranged medial () to lateral () show abnormal meniscal signal (arrows) extending to PHLM articular surfaces on two separate images, clearly depicting tear seen at arthroscopy. On the other hand, six of the missed PHLM tears were not evident even on retrospective rereview (occult tears) (Fig. 4). ll were stable to probing at arthroscopy and were neither surgically repaired nor surgically resected; thus, we conclude that these MRI-occult tears were not clinically significant. Furthermore, all were surgically classified as either partial-thickness or split-type tears. Last, eight PHLM tears were considered subtle at rereview; only four were surgically repaired (2/8) or resected (2/8). Three tears were classified as partial thickness or split type at arthroscopy. Three tears might have been prospectively identified if the abnormal superior popliteomeniscal fascicle had been recognized (Fig. 5; normal superior popliteomeniscal fascicle, Fig. 4). Two of these tears could have been prospectively identified if the abnormal meniscofemoral ligament attachment had been identified (Fig. 6). Overall, if all subtle and clearly evident PHLM tears had not been missed on MRI and only the occult PHLM tears from the retrospective review were analyzed, the sensitivity of MRI for the detection of lateral meniscus tears would have increased from 58.5% to 81.1%. Such a corrected sensitivity is comparable to the pooled weighted sensitivity of 79.3% for all lateral meniscus tears regardless of CL injury. Significant Secondary Findings: bnormal Superior Popliteomeniscal Fascicle and Far Lateral Meniscofemoral Ligament Extension In our retrospective rereview of the PHLM tears, we found that the presence or absence of two secondary findings an abnormal superior popliteomeniscal fascicle and apparent JR:193, ugust

6 Laundre et al. far lateral extension of the meniscofemoral ligament was statistically significant when comparing the missed PHLM tear and unmissed PHLM tear groups. When compared with knees with a normal lateral meniscus, knees with PHLM tears (both missed and combined missed and unmissed PHLM tears) were more likely to have abnormal superior popliteomeniscal fascicles. However, missed PHLM tears were more likely to have normal superior popliteomeniscal fascicles compared with unmissed tears (Table 3). natomically, the popliteomeniscal fascicles are thin synovium-lined struts that attach the PHLM to the joint capsule [29]. In the past, failure to visualize these fascicles on arthrography served as an indirect sign for lateral meniscus tear [30]. The presence of an abnormal superior popliteomeniscal fascicle on MRI has been shown to be highly associated with a tear of the lateral meniscus, with a positive predictive value and specificity previously reported as 79% and 96%, respectively; thus, it serves as a useful clue for Fig year-old man with noncontact, jumping landing knee injury and acute anterior cruciate ligament tear. Tear of posterior horn of lateral meniscus (PHLM) missed at initial MRI was occult at MRI rereview. rthroscopy revealed small, superficial, partial-thickness tear that coursed through approximately 20% of superior surface of far posterior horn; however, no tear was evident on MRI even during rereview. Incidentally, MRI also showed normal superior popliteomeniscal fascicles. None of MRI-occult (unavoidably missed) PHLM tears was surgically repaired or resected. and, Sagittal fast spin-echo (FSE) proton density weighted () and FSE T2-weighted () images show normal PHLM and normal superior popliteomeniscal fascicles (arrows). Fig year-old boy with noncontact, twisting knee injury and acute anterior cruciate ligament tear. Tear of posterior horn of lateral meniscus (PHLM) missed at initial MRI was subtle at MRI rereview. rthroscopy revealed stable, partial-thickness, oblique tear involving PHLM. MRI rereview depicted subtle signal abnormality in PHLM (not shown) and abnormal superior popliteomeniscal fascicle. and, Sagittal fast spin-echo (FSE) proton density weighted () and FSE T2-weighted () images show abnormal superior popliteomeniscal fascicle, with mostly disrupted fascicle seen near its attachment to PHLM (arrows). Very subtle signal abnormality was noted in PHLM (not shown), most likely representing tear identified at arthroscopy. the detection of subtle lateral meniscal abnormalities. However, this finding is not sensitive for lateral meniscus tears, reported at 31% [18, 19, 31]. Within our focused, three-group comparison of knees with PHLM tears and knees with a normal lateral meniscus, 18 of the 21 intact lateral menisci had normal or probably normal superior fascicles, yielding 86% specificity, whereas 30 of the 50 torn PHLM had abnormal or probably abnormal superior fascicles, yielding 60% sensitivity. Of note, four knees with missed PHLM tears and three knees with intact lateral menisci had superior popliteomeniscal fascicles classified as equivocal. Our data also reflect a high association between the presence of an abnormal superior popliteomeniscal fascicle and a PHLM tear, with a 100% positive predictive value. This finding lends credence to the suspicion that superior fascicle abnormalities are often associated with lateral meniscus tears, possibly because the mechanisms of injury and resulting biomechanical forces that tear the meniscus also affect the adjacent fascicle [18]. We noted that knees with PHLM tears were more likely to have abnormal superior popliteomeniscal fascicles than knees with a normal lateral meniscus. However, knees with missed PHLM tears were more likely to have normal superior popliteomeniscal fascicles than those with unmissed tears. Perhaps the knees with missed PHLM tears sustained enough injury to damage only the lateral meniscus but not the fascicles, thus leading to a less extensive tear that is more subtle or more difficult to visualize on MRI. On the other hand, forces that injure both the fascicles and lateral meniscus may be great enough to produce a more conspicuous meniscal tear. Overall, six knees with missed PHLM tears had abnormal or probably abnormal superior popliteomeniscal fascicles: three with subtle and three with clearly evident PHLM tears on rereview (Table 3). Prospective recognition of this secondary finding and subsequent identification of these PHLM tears would have increased MRI sensitivity for the detection of lateral meniscus tears from 58.5% to 67.9%. dditionally, many investigators have suggested that the close anatomic relationship of the PHLM and the meniscofemoral ligament leads to poor MRI sensitivity for meniscal tears [9, 16, 19, 20]. In fact, the meniscofemoral ligament attachment site has been described as a pseudotear [32]. Investigators have recently reported that abnormal far lateral insertion of the meniscofemoral ligament onto the PHLM that is, greater than or equal to four images (slice thickness of 3 mm and gap of 0.5 mm = effectively 14 mm) should be considered as a probable longitudinal tear of the PHLM [20]. Within our focused three-group comparison, 18 of the 21 normal lateral menisci had normal meniscofemoral ligament extension, yielding 86% specificity; 16 of the 50 torn PHLMs had abnormal far lateral meniscus insertion, yielding 32% sensitivity. Of note, the meniscofemoral ligament was not discretely identified in 12 cases. Last, abnormal extension of the meniscofemoral ligament was highly associated with a PHLM tear, with an 84% positive predictive value. Knees with PHLM tears were more likely overall to have an abnormal insertion of the meniscofemoral ligament than knees with a normal lateral meniscus. However, knees with missed PHLM tears were more likely to have a normal insertion of the meniscofemoral ligament on the lateral meniscus than knees with unmissed tears (Table 3). 520 JR:193, ugust 2009

7 MRI Detection of Meniscal Tears in Patients With CL Injury Once more, we hypothesize that if an injury that produces enough biomechanical force to disrupt the meniscofemoral ligament attachment occurs, it may occasionally result in a more conspicuous or extensive PHLM tear. Of the knees with a missed PHLM tear, two had an abnormal far lateral extension of the meniscofemoral ligament: oth were not prospectively identified and both were categorized as subtle PHLM tears on rereview. These tears were described as either longitudinal peripheral or partial-thickness or splittype tears of the far PHLM at surgery, which is compatible with the rereview MRI findings. Prospective recognition of this secondary finding and subsequent identification of these two PHLM tears would have increased MRI sensitivity for the detection of lateral meniscus tears from 58.5% to 62.3%. C E G Insignificant Variables In reviewing clinical history data, we did not find a statistically significant associa- H tion between missed PHLM tears and patient age, although the median age of patients with missed PHLM tears (17 years) was slightly less than that of patients with MRI-detected tears (20 years; p = 0.053). The time between injury and MRI and the time between MRI and arthroscopy were considerably low, and there was no statistical evidence that either MRI delay or arthroscopic delay played a role in PHLM tears being missed on MRI. Investigators have previously suggested that some tears may be missed as a result of unreported interval knee injuries in an CL-deficient or unstable knee if the interval between imaging and surgery is too great. Similar to the findings reported by De Smet and Graf [9], we found no association between missed tears and arthroscopic delay. nalysis of the initial MRI findings also revealed no statistical association between missed PHLM tears and the total number of meniscal tears or the presence of a concurrent medial meniscus tear. The majority of D F Fig. 6 Side-by-side comparison of knee with apparent far lateral extension of meniscofemoral ligament ( D) and knee with normal insertion of meniscofemoral ligament onto posterior horn of lateral meniscus (PHLM) (E H). D, Sagittal conventional spin-echo proton density weighted images (3.0-mm slice thickness, 1.5-mm skip) show abnormal persistence of meniscofemoral ligament, which is discretely identified posterior to PHLM on five images lateral to posterior cruciate ligament (PCL) (corresponding to 22.5 mm of lateral extension) in 15-year-old girl with contact, valgus stress knee injury and acute anterior cruciate ligament tear. PHLM tear was subtle at MRI rereview. rthroscopy showed stable, partial-thickness, peripheral tear of PHLM. MRI rereview revealed apparent far lateral insertion of meniscofemoral ligament of Humphrey onto lateral meniscus. First () image lateral to PCL shows normal meniscofemoral ligament (arrow, ). Third () and fifth (C) images lateral to PCL show apparent meniscofemoral ligament (arrows, and C) that remains separate from PHLM, likely representing peripheral tear seen at arthroscopy. Sixth (D) image lateral to PCL shows that apparent meniscofemoral ligament has mostly inserted onto PHLM. E H, Sagittal conventional spin-echo proton density weighted images show normal meniscofemoral ligament of Humphrey with normal insertion onto PHLM in 27-year-old woman with noncontact, twisting knee injury and acute anterior cruciate ligament tear. PHLM tear was occult at MRI rereview. rthroscopy revealed very superficial, partialthickness tear at posterior horn meniscosynovial junction; however, no tear was evident on MRI even during rereview. First (E) image lateral to PCL shows normal meniscofemoral ligament (arrow, E). Third (F), fifth (G), and sixth (H) images lateral to PCL show normal PHLM with no discrete apparent meniscofemoral ligament. missed tears occurred in the posterior horns of the menisci, but there was no difference in the tear distribution between the missed and unmissed PHLM tear cases. Moreover, injuries to multiple knee ligaments (MCL, LCL, or PCL), which are often associated with more significant knee injury [33] and a lower MRI sensitivity for meniscal tears [14], were not associated with missed PHLM tears. lthough bone marrow edema patterns serve as useful indirect signs of meniscal injury [17], bone contusions involving the lateral femoral condyle and lateral tibial plateau or bone contusions seen elsewhere were also not significantly associated with missed tears. We found no relationship between MR sequence type (i.e., fat-suppressed FSE vs fatsuppressed conventional SE) and missed PHLM tears. lthough the T2-weighted FSE sequences were performed with fat suppression, our results support previously reported evidence showing no significant difference in MRI accuracy for meniscal tears between JR:193, ugust

8 Laundre et al. FSE and conventional SE sequences without fat saturation [8, 27], despite a previous study that showed a 10% lower sensitivity for fatsuppressed FSE imaging compared with fatsuppressed conventional SE imaging [24]. In light of the MRI rereview findings, there was no significant association between missed PHLM tears and the presence of posterior pericapsular edema, similar to findings reported in a previous study that confirmed a lack of association between posterolateral pericapsular edema and lateral meniscus tears [18]. lthough a previous investigation showed an increased prevalence of lateral meniscus root tears in knees with CL tears and lateral extrusion of the lateral meniscus [21], we found no significant relationship between these two findings and missed PHLM tears. Future Directions and Conclusions Some of the occult partial-thickness or split-type PHLM tears in our study may not be adequately depicted using current MRI techniques. It may be possible to improve MRI techniques to increase the sensitivity of MRI for meniscal tears. Increased signalto-noise ratio afforded by high-field-strength imaging (3 T), shown to have 96% sensitivity [22], could be used to improve the spatial resolution of menisci imaging, possibly revealing the small tears that are not evident by current MR techniques. ecause these MRI-occult partial-thickness or split-type meniscal tears do not extend through the entire surface of the meniscus, we hypothesize that they may simply have nondisplaced, apposed tear edges without any intervening fluid at the time of MRI. Such tears should easily be identified at arthroscopy when the meniscus is mechanically probed. Moreover, if these tears were well apposed and the knees were adequately immobilized, these small tears may have partially healed before surgery, which could also contribute to their stability at the time of arthroscopy. In such cases, dynamic knee imaging [34] may prove useful in identifying torn menisci by applying stress on the menisci, possibly by flexion extension positioning of the knee or by applying a weight-bearing axial load on the joint. However, based on our data, such provocative maneuvers may not be clinically justifiable in the evaluation of MRI-occult PHLM tears because none of these unavoidably missed tears in our study required surgical treatment. nother future goal of our research includes clinical follow-up of patients with untreated meniscal tears to see whether their tears healed successfully. Correlating clinical improvement with the missed or unmissed tears may allow us to verify that the PHLM tears not evident on MRI were truly clinically insignificant with regard to surgical management. Limitations of our study include its relatively small sample size, which reduced statistical power especially during the rereview of the 19 missed PHLM tears. Furthermore, because of the focus of our study design, there was no control group of knees with an intact CL. Such a comparison would have allowed us to directly compare differences in knees with and knees without CL injury and to verify a statistically significant lower MRI sensitivity for lateral meniscus tears in the setting of a concurrent CL tear [9]. dditionally, the sensitivity, specificity, and positive predictive values derived from the analysis of secondary MRI findings specifically, abnormal superior popliteomeniscal fascicles and abnormal far lateral insertion of the meniscofemoral ligament are limited because the focused analysis compared only knees with a torn PHLM and knees with a normal lateral meniscus but abnormal medial meniscus. Therefore, these statistics cannot be generalized to include knees with lateral meniscus tears not involving the PHLM or knees in which the lateral meniscus and medial meniscus are both normal, groups that were not specifically included in our analyses. Finally, our study excluded those who did not undergo surgery at our institution and those who were only conservatively managed. In conclusion, the results of our study showed that missed PHLM tears that were clearly evident on MRI rereview would not have been prospectively missed on MRI if the two-slice-touch rule had been strictly applied. ll of these missed PHLM tears were unstable at surgery and, thus, required surgical repair or resection. On the other hand, all of the missed PHLM tears that were not evident on rereview using the two-slice-touch rule were not clinically significant: They did not require surgical intervention. Similar to the results of previous studies, we found that knees with PHLM tears were more likely to have abnormal superior popliteomeniscal fascicles and abnormal far lateral insertion of the meniscofemoral ligament than knees with surgically normal menisci. However, in most cases, missed PHLM tears did not have any specific secondary MRI findings that could aid in detection; our study failed to identify a specific and identifiable cause as to why more PHLM tears remain undiagnosed by MRI in the setting of acute CL injury. cknowledgments We thank Linda M. Greene and Christine. Kenyon for their assistance with manuscript preparation and Vicki C. Schmidt for her assistance with the surgical and MRI database searches. References 1. arnett MJ. MR diagnosis of internal derangements of the knee: effect of field strength on efficacy. JR 1993; 161: McNally EG, Nasser KN, Dawson S, Goh L. Role of magnetic resonance imaging in the clinical management of the acutely locked knee. Skeletal Radiol 2002; 31: Feller J, Webster KE. Clinical value of magnetic resonance imaging of the knee. NZ J Surg 2001; 71: Elvenes J, Jerome CP, Reikerås O, Johansen O. Magnetic resonance imaging as a screening procedure to avoid arthroscopy for meniscal tears. rch Orthop Trauma Surg 2000; 120: Weinstabl R, Muellner T, Vécsei V, Kainberger F, Kramer M. Economic considerations for the diagnosis and therapy of meniscal lesions: can magnetic resonance imaging help reduce the expense? World J Surg 1997; 21: Makdissi M, Eriksson KO, Morris HG, Young D. MRI-negative bucket-handle tears of the lateral meniscus in athletes: a case series. Knee Surg Sports Traumatol rthrosc 2006; 14: Oei EHG, Nikken JJ, Verstijnen CM, Ginai Z, Hunink MGM. MR imaging of the menisci and cruciate ligaments: a systematic review. Radiology 2003; 226: De Smet, Mukherjee R. Clinical, MRI, and arthroscopic findings associated with failure to diagnose a lateral meniscal tear on knee MRI. JR 2008; 190: De Smet, Graf K. Meniscal tears missed on MR imaging: relationship to meniscal tear patterns and anterior cruciate ligament tears. JR 1994; 162: Justice WW, Quinn SF. Error patterns in the MR imaging evaluation of menisci of the knee. Radiology 1995; 196: Quinn SF, rown TF. Meniscal tears diagnosed with MR imaging versus arthroscopy: how reliable a standard is arthroscopy? Radiology 1991; 181: Mesgarzadeh M, Moyer R, Leder DS, et al. MR imaging of the knee: expanded classification and pitfalls to interpretation of meniscal tears. Radio- 522 JR:193, ugust 2009

9 MRI Detection of Meniscal Tears in Patients With CL Injury Graphics 1993; 13: Kalume-rigido M, Wojtys E. Posterior horn lat- 27. De Smet, Tuite MJ. Use of the two-slice- 13. Nikolić DK. Lateral meniscal tears and their evo- eral meniscal tears simulating meniscofemoral touch rule for the MRI diagnosis of meniscal lution in acute injuries of the anterior cruciate ligament attachment in the setting of CL tear: tears. JR 2006; 187: ligament of the knee: arthroscopic analysis. Knee MRI findings. Skeletal Radiol 2007; 36: De Smet, Norris M, Yandow DR, Quintana Surg Sports Traumatol rthrosc 1998; 6: rody JM, Lin HM, Hulstyn MJ, Tung G. Lat- F, Graf K, Keene JS. MR diagnosis of menis- 14. Rubin D, Kettering JM, Towers JD, ritton C. MR imaging of knees having isolated and combined ligament injuries. JR 1998; 170: De Smet, Tuite MJ, Norris M, Swan JS. MR diagnosis of meniscal tears: analysis of causes of errors. JR 1994; 163: Jee WH, McCauley TR, Kim JM. Magnetic resonance diagnosis of meniscal tears in patients with acute anterior cruciate ligament tears. J Comput ssist Tomogr 2004; 28: ergin D, Hochberg H, Zoga C, Qazi N, Parker L, Morrison W. Indirect soft-tissue and osseous signs on knee MRI of surgically proven meniscal tears. JR 2008; 191: lankenbaker DG, De Smet, Smith JD. Usefulness of two indirect MR imaging signs to diagnose lateral meniscal tears. JR 2002; 178: De Smet, singer D, Johnson RL. bnormal superior popliteomeniscal fascicle and posterior pericapsular edema: indirect MR imaging signs of a lateral meniscal tear. JR 2001; 176: Park LS, Jacobson J, Jamadar D, Caoili E, eral meniscus root tear and meniscus extrusion with anterior cruciate ligament tear. Radiology 2006; 239: Magee T, Williams D. 3.0-T MRI of meniscal tears. JR 2006; 187: Cotten, Delfaut E, Demondion X, et al. MR imaging of the knee at 0.2 and 1.5 T: correlation with surgery. JR 2000; 174: lackmon G, Major NM, Helms C. Comparison of fast spin-echo versus conventional spinecho MRI for evaluating meniscal tears. JR 2005; 184: Kowalchuk RM, Kneeland J, Dalinka MK, Siegelman ES, Dockery WD. MRI of the knee: value of short echo time fast spin-echo using high performance gradients versus conventional spinecho imaging for the detection of meniscal tears. Skeletal Radiol 2000; 29: Quinn SF, rown TR, Szumowski J. Menisci of the knee: radial MR imaging correlated with arthroscopy in 259 patients. Radiology 1992; 185: cal tears of the knee: importance of high signal in the meniscus that extends to the surface. JR 1993; 161: Jelaso DV. The fascicles of the lateral meniscus: an anatomic arthrographic correlation. Radiology 1975; 114: Pavlov H, Goldman. The popliteus bursa: an indicator of subtle pathology. JR 1980; 134: Johnson RL, De Smet. MR visualization of the popliteomeniscal fascicles. Skeletal Radiol 1999; 28: Vahey TN, ennett HT, rrington LE, Shelbourne KD, Ng J. MR imaging of the knee: pseudotear of the lateral meniscus caused by the meniscofemoral ligament. JR 1990; 154: ikkina RS, Tujo C, Schraner, Major NM. The floating meniscus: MRI in knee trauma and implications for surgery. JR 2005; 184: Vedi V, Williams, Tennant SJ, Spouse E, Hunt DM, Gedroyc WM. Meniscal movement: an invivo study using dynamic MRI. J one Joint Surg r 1999; 81:37 41 JR:193, ugust