Department of Orthopaedic Surgery, Tampere University Hospital, Tampere, Finland 3

Scandinavian Journal of Surgery 101: 56 61, 2012 Sensitivity of MRI for articular cartilage lesions of the patellae V. M. Mattila 1, 2, M. Weckström 1, V. Leppänen 1, M. Kiuru 1, H. Pihlajamäki 1, 3 1 Centre for Military Medicine, Lahti, Finland 2 Department of Orthopaedic Surgery, Tampere University Hospital, Tampere, Finland 3 Department of Orthopaedic Surgery, Central Military Hospital, Helsinki, Finland ABSTRACT Background and Aims: Reliable diagnosis of articular cartilage lesions of the patellae is often based on arthroscopy. However, unnecessary arthroscopies should be avoided. The aim of this study was to assess the sensitivity and applicability of MRI to diagnosing articular cartilage lesions of the patellae. Materials and Methods: We identified 74 consecutive males (mean age 21 years, range 18 28) from the medical records of our institute with the sole diagnosis of articular cartilage lesions of the patellae based on arthroscopy. Magnetic resonance imaging was performed with 1.0 Tesla scanner a mean of 4 weeks before arthroscopy. Sensitivity of symptoms, and MRI for the diagnosis was calculated. Results: Based on arthroscopy, 20 (27%) cases of cartilage lesions of the patellae were grade-i, 32 (43%) were grade-ii, and 22 (30%) were grade-iii. MRI revealed cartilage lesions of the patellae in 49 knees (66%), indicating that the sensitivity of MRI was 66% (95% CI: 53% 74%). MRI sensitivity increased with the severity of chondral lesions: all grade III to IV lesions were detected (sensitivity 100%, 95% CI: 85% 100%) by MRI. Grade of articular cartilage lesions of the patellae based on arthroscopy was not associated with clinical symptoms (p = 0.61). Conclusions: The sensitivity of 1.0 Tesla MRI for detecting grade-i lesions was low and could not be used to confirm the diagnosis of articular cartilage lesions of the patellae. For the detection of more severe grade-ii to III lesions, the MRI sensitivity was markedly higher. MRI may thus be considered an accurate diagnostic tool for identifying more severe cases of articular cartilage lesions of the patellae. Key words: Articular cartilage lesions; patellae; diagnostic study; sensitivity; magnetic resonance imaging; anterior knee pain; risk factor Correspondence: Ville Mattila, M.D. Department of Orthopaedics and Traumatology Tampere University Hospital FI - 33250 Tampere, Finland Email: Ville.Mattila@uta.fi

Sensitivity of MRI for articular cartilage lesions of the patellae 57 INTRODUCTION The clinical significance of softened patellar articular cartilage is questionable because the association between the presence of softened articular cartilage and clinical symptoms is controversial (1 5). It has been suggested that articular cartilage lesions of the patellae, previously chondromalacia patellae, is over-diagnosed when the diagnosis has based on symptoms and clinical signs (1, 3). Magnetic resonance imaging (MRI) is widely used for detecting internal derangements of the knee. It is a noninvasive procedure, and has a low risk of complications. Whether MRI can help ensure the correct diagnosis of articular cartilage lesions of the patellae in patients with symptoms of anterior knee pain, however, is not clear. Previous reports indicate that the overall sensitivity for chondral lesions of the patellae ranges between 30% and 100%, with a specificity between 50% and 94% and a diagnostic accuracy between 77% and 90% (6 8). Arthroscopy provides a reliable diagnosis of patellar articular cartilage lesions (9, 10). However, arthroscopical or other surgical procedures of articular cartilage lesions of the patellae is required in less than 10% of patients (9, 11), while the initial treatment comprises conservative rehabilitation (12). Arthroscopy for shaving or removing fibrillated and traumatized areas of articular cartilage is used for deeper patellae lesions grades-ii to IV, despite evidence that the positive treatment outcomes may deteriorate over time (13, 14). If arthroscopic examination does not reveal any arthroscopically treatable lesions, however, it is a costly diagnostic method that unnecessarily consumes limited healthcare resources and thus diagnostic arthroscopies should not be done. Moreover, as an invasive procedure, arthroscopy causes short-term functional disability, pain, and stress, and involves risks related to anesthesia and surgery. We had the opportunity to examine sensitivity of MRI in a homogenous sample of young men with arthroscopically proven articular cartilage lesions of the patellae. Thus, the purpose of the present diagnostic study was to assess the sensitivity of MRI for detecting articular cartilage lesions of the patellae. The association between symptoms, clinical signs and grading of articular cartilage lesions of the patellae was also investigated. MATERIALS AND METHODS In Finland, compulsory military service is required for all male citizens. During the study period of 1999 to 2003, the number of recruits starting their service varied between 21 909 and 30 659 annually, and over 80% of the males aged 18 to 29 years (median, 19 years) completed a service period ranging from 6 to 12 months. The study population consisted of 143 343 military recruits nationwide. Combat training, marching, and other physical training activities were performed almost daily throughout the service period. The patients selected for this study comprised of young military trainees who had passed their entry medical examination as healthy and were able to perform physically demanding military training. Physical and combat training includes physical exercise, causing increased loads to the lower limb and especially to knees because kneeling and crawling are often included in the training activities (15). This training is performed almost daily throughout the service period. Thus, anterior knee pain may markedly interfere with the performance of military training and may threaten the completion of military service. The study protocol was approved by the local ethics committee. Eligible, consecutive patients with articular cartilage lesions of the patellae diagnosed with arthroscopy were identified retrospectively from the medical records of the main military hospital, where all surgical operations on recruits performing compulsory military service were conducted between 1999 and 2003. A computer search was performed on records from 1999 through 2003 using the appropriate diagnostic codes of the tenth edition of the International Classification of Diseases (ICD-10). The inclusion criteria were a sole diagnosis of articular cartilage lesions of the patellae diagnosed by arthroscopy and previous MRI of the knee no more than 3 months before arthroscopy. Exclusion criteria were previous fracture of the distal tibia or patella. We identified 77 patients (3 women and 74 men) with sole diagnosis of articular cartilage lesions of the patellae diagnosed by arthroscopy. Due to the small number of women, they were excluded from the study, resulting in a total of 74 patients. In cases of bilateral articular cartilage lesions of the patellae (15 knees) only one knee per person was randomly chosen. Mean age of the patients was 20.6 years (range 18.3 28.2). The original complete medical records, including radiographs, were retrieved and reviewed to confirm the accuracy of the diagnoses, and to systematically collect data for the present study. Patient medical histories were studied and collected information was recorded in a database. Duration of knee symptoms was classified into one of three categories: 1) less than 1 year, 2) 1 to 2 years, and 3) over 2 years. Clinical features were also classified into one of three categories: 1) discomfort and pain in anterior patella without further symptoms or findings, 2) typical symptoms of anterior knee pain (AKP), namely pain after walking stairs and prolonged sitting, without findings in physical examination, and 3) typical clinical symptoms of AKP with crepitus and pain during active movement of the patella. All military conscripts with anterior knee pain symptom were initially treated conservatively with anti-inflammatory medication in their units, and temporarily exempted from all strenuous physical activity and military training. Physiotherapy, including lower extremity muscle strengthening concentrating on the quadriceps, was employed. Patients not responding to conservative treatment and whose anterior knee pain was restricting their training, were referred by the physicians in the basic military units to orthopedic consultation at the main military hospital to assess the need for MRI and arthroscopy. All patients included into the study had both arthroscopy and MRI. In general, arthroscopy was chosen in Central Military hospital, when the response to nonsurgical treatment was insufficient and when a reliable method was needed to verify the suspected diagnosis of articular cartilage lesions of the patellae, or to differentiate between articular cartilage lesions of the patellae and other internal derangements of the knee, such as medial plica. Plain radiographs, including anteroposterior, lateral, and Merchant views, were routinely obtained in all of the patients before physical examination by an orthopedic surgeon. MRI was performed in all patients prior to knee arthroscopy using a 1.0 Tesla scanner (Signa Horizon; GE Medical Systems, Milwaukee, WI) with a standard knee coil and a field of view of 10 to 16 cm. The knee was extended and quadriceps was relaxed during the scanning. Slice thickness was 3 to 4 mm, with a 0.5 or 1.0-mm intersection gap. Sagittal proton density (PD) spin echo sequence images

58 V. M. Mattila, M. Weckström, V. Leppänen, M. Kiuru, H. Pihlajamäki Figure 1B A B Fig. 1. A 22-year-old male patient with patellofemoral pain without crepitation for 2 years in both knees. His symptoms increased during military service. Clinical examination revealed no abnormal findings. B: Axial fat-suppressed T1-weighted magnetic resonance image of A: Axial fat-suppressed T1-weighted magnetic resonance image of the same patient s left knee showing grade-ii chondromalacia in the right knee showing grade-ii chondromalacia in the medial the medial and lateral facets of the patella (white arrows in mefacet of the patella (white arrows). Arthroscopy showed chondial facet). Subsequent arthroscopy showed changes of grade-iii dromalacia lesion (2 2.5 cm); the cartilage was fissured and fragchondromalacia in a 2 2 cm. mented, but lesions did not reach the subchondral bone. Figure 2A with fat suppression or sagittal T1-weighted spin echo sequence images were obtained. Axial T1- and T2-weighted sequences with fat suppression were obtained, as well as T2-weighted coronal images. The images were initially evaluated by a musculoskeletal radiologist on duty using a standard MRI interpretation protocol. For the purpose of this study, another musculoskeletal radiologist (M.K.) then reevaluated the images blinded to the previous MRI results and arthroscopy find21 ings. In cases in which these reevaluations differed from the original (3 cases) (inter-observer reliability, kappa = 0.9), a decision was reached by consensus. MRI grading of articu lar cartilage lesions of the patellae was made based on the arthroscopic grading system by Shahriaree: (12) 0, normal; 1, softening or blister by high signal intensity and swelling of the cartilage on magnetic resonance images; 2, fissuring; 3, fragmentation and fissuring, and 4, full thickness fissure and exposed bone. Arthroscopy was performed after MRI examination (mean time between MRI and arthroscopic examination was 3.8 weeks [range 0 12 weeks]. One experienced orthopedic surgeon performing the arthroscopy evaluated the articular surfaces blinded to the MRI results. The knees were examined systematically using a probe inserted into the knee joint. Under direct vision and with careful probing, special attention was paid to the patellar surface and the degree of fibrillation, fragmentation, and possible softening of the cartilage surface based on the arthroscopic grading system by Shahriaree (12). To determine the diagnostic sensitivity of the 1.0 Tesla MRI and symptoms and physical examination, the MRI findings, symptoms, and clinical findings were then compared with the arthroscopic findings, the arthroscopic results serving as the gold standard. We used a single table analysis when calculating the sensitivity of MRI and clinical symptoms. A chi-square test was used in two-way tables 22 Fig. 2. A 19-year-old male patient with strain-related swelling and patellofemoral pain since the start of the military service. Clinical examination detected patellofemoral crepitus. A: Axial fat-suppressed T1-weighted magnetic resonance image showed grade-i chondromalacia, softening, and swelling of the cartilage in the patellar facets. In arthroscopy,the cartilage was smooth but in probe examination it was softened. and the level of significance was set to p = 0.05. Ninety-five percent confidence intervals were calculated by Wilson s method. When calculating results for grade I articular car-

Sensitivity of MRI for articular cartilage lesions of the patellae 59 Table 1 Sensitivity of symptoms and clinical examination and MRI in chondromalacia patellae in 74 male military conscripts. Symptoms and clinical examination True positive False negative Sensitivity (95% CI) Discomfort and pain in anterior patellae without further symptoms or findings 21 53 28% (19 40%) Pain after walking stairs and prolonged sitting, without findings in clinical examination 24 50 32% (22 43%) Pain after walking stairs and prolonged sitting, and crepitus and pain during active movement of the patella. 29 45 39% (28 49%) MRI findings Any chondromalacia patellae visible in MRI 49 27 66% (53 74%) Grade I chondromalacia patellae in MRI 04 16 20% 0(8 42%) Grade II chondromalacia patellae in MRI 23 09 72% (55 84%) Grade III to IV chondromalacia patellae in MRI 22 00 100% (85 100%) tilage lesions of the patellae, only arthroscopically detected grade I lesions were considered positive (gold standard) for cartilage lesions and compared to MRI results in which grade 0 was considered negative and grades I to IV positive. Results for grades II IV were calculated respectively, arthroscopically detected grade II to IV lesions were considered positive and compared to MRI results in which grade 0 was considered negative and grades I to IV positive. SPSS 17.0 for Windows was used for the statistical analysis. RESULTS Of all the male conscripts performing their compulsory military service during the 5-year period, we identified 74 consecutive patients with a sole diagnosis of articular cartilage lesions of the patellae based on arthroscopy. Twenty (27%) cases of the articular cartilage lesions of the patellae based on arthroscopy were grade-i, 32 (43%) were grade-ii, and 22 (30%) were grade-iii. Grade-IV lesions were not found. The grading of articular cartilage lesions of the patellae based on arthroscopy was not associated with the clinical symptoms of AKP (p = 0.61) or the duration of the symptoms before arthroscopy (p = 0.75). Thirtyone patients (42%) had suffered symptoms of anterior knee pain for less than 1 year, 9 (12%) for 1 to 2 years, and 34 (46%) for more than 2 years. The sensitivities of MRI, symptoms, and physical examination are shown in Table 1. Due to the study setting, all patients had articular cartilage lesions of the patellae and thus the specificity of the diagnostic methods could not be assessed. MRI detected articular cartilage lesions of the patellae in 49 knees, resulting in a sensitivity of 66% (95% CI: 53% 74%; Table 1). MRI scans detected grade I articular cartilage lesions of the patellae with a sensitivity of 20% (95% CI: 8 42%). The higher the grade, the better the detection, and grade III articular cartilage lesions of the patellae was always detected in MRI (sensitivity 100%, 95% CI: 85% 100%). In 19 knees (26%) MRI was normal. Other MRI diagnoses included one medial plica, one medial meniscal tear, and one partial tear in the patellar tendon. Only 39% (95% CI: 28% 49%) of the patients with proven articular cartilage lesions of the patellae had typical symptoms and physical findings of AKP, namely pain after walking stairs and prolonged sitting, and crepitus and pain during active movement of the patella. In 28% of the patients, the only symptom was discomfort and pain in the anterior aspect of the knee without findings in physical examination. The results of MRI and arthroscopy for the different grades of articular cartilage lesions of the patellae presented in Table 2. Deeper chondral lesions were better detected; 16/25 grade I articular cartilage lesions of the patellae were not visible in MRI. Table 3 Table 2. Association of arthroscopic and MRI grading in 74 knees with chondromalacia patellae. Arthroscopic grade* MRI grade* I II III Total 0 16 09 00 25 I 03 02 02 07 II 00 19 09 28 III 01 02 09 12 IV 00 00 02 02 Total 20 32 22 74 * Shahriaree (12) Table 3 Association of symptoms, physical examination, and arthroscopic grading in 74 knees with chondromalacia patellae. Arthroscopic grade* Symptoms and physical examination I II III Total Discomfort and pain in anterior patellae without further symptoms or findings 03 11 07 21 Pain after walking stairs and prolonged sitting, without findings in clinical examination 08 10 06 24 Pain after walking stairs and prolonged sitting, and crepitus and pain during active movement of the patella 09 11 09 29 Total 20 32 22 74 * Shahriaree (12)

60 V. M. Mattila, M. Weckström, V. Leppänen, M. Kiuru, H. Pihlajamäki shows the association between symptoms and physical findings and arthroscopically identified grading of the articular cartilage lesions of the patellae. The association was not significant (p = 0.61). Plain radiographs of all 74 knees showed normal findings in 70 cases. Lateralization of the patella and a flat sulcus angle (< 138 degrees) was found in three knees, and Osgood-Schlatter disease was found in one knee. DISCUSSION The primary aim of this study was to assess the sensitivity of 1.0 T MRI in arthroscopically confirmed articular cartilage lesions of the patellae. The sensitivity of MRI for detecting grade-i lesions was low and could not be used to confirm the diagnosis of articular cartilage lesions of the patellae. For the detection of more severe grade-ii to III lesions, the MRI sensitivity was markedly higher. MRI may thus be considered an accurate diagnostic tool for identifying more severe cases of articular cartilage lesions of the patellae. The present study also showed no association between articular cartilage lesions of the patellae and symptoms of anterior knee pain. Thus, symptoms or findings in clinical examination suspected to be related to articular cartilage lesions of the patellae should not be used as an indication for knee arthroscopy. Overall sensitivity of MRI in articular cartilage lesions of the patellae was poor, although sensitivity increased with an increase in the severity of cartilage lesions. Although some studies exist concerning the sensitivity of MRI in detecting articular cartilage lesions of the patellae, the populations have been heterogenous (6, 7, 16, 17). Thus our study with large homogenous sample adds to the literature. A previous study reported that the sensitivity of MRI for detecting all grades of articular cartilage lesions of the patellae is 57% (7), which is similar to the sensitivity of 66% in our study. McCauley and coworkers demonstrated a slightly higher sensitivity (72%) (6) and interestingly, also clearly higher values (up to 90% to 100%) have been published (16, 17). Another study that used the same articular cartilage lesion grading presented by Shahriaree (12) as in the present study reported a sensitivity of 0% for grade-i lesions (18). In grade-ii and III, the imaging sensitivity was considerably higher (72% and 100%). There is evidence from the previous literature, that the sensitivity depends highly on the sequences used. In a study by Waldschmidt and co-workers, the sensitivity of grade II lesions was between 13% (PD-weighted MRI) and 47% (T2-weighted MRI), while the corresponding figure for grade-iv lesions was 50% (spoiled gradientrecalled imaging) and 75% (T1-, T2-, and PD-weighted MRI) (18). Although there were no patients with grade-iv cartilage lesions in our sample, our results clearly show that sensitivity was increased with deeper lesions. One of the secondary findings of the present study was the poor correlation between cartilage lesions identified in arthroscopy and the symptoms of anterior knee pain. This finding was somewhat expected because previous literature demonstrated the absence of pathognomonic symptoms (1, 5, 19). In our sample, only one-third of patients with diffuse AKP as the sole symptom had articular cartilage lesions of the patellae, while in patients with symptoms of AKP and patellofemoral crepitus the sensitivity was 39%. Our results were even lower than that of the previous study that showed that 60% of the patients with clinically diagnosed anterior knee pain with retropatellar crepitus had articular cartilage lesions of the patellae (1). Based on the findings of the present study, it seems evident that cartilage lesions of the patellae cannot be distinguished by symptoms and signs in physical examination (5). Taking into account that surgical treatment of cartilage lesions of patellae is required in less than 10% of patients based on the literature (11) and that the effectiveness of surgery even in these cases is controversial (13, 14), arthroscopy is not recommended in cases of suspicion of articular cartilage lesions of the patellae. The present study has several limitations. Due to the retrospective selection of patients with all having articular cartilage lesions of the patellae identified by arthroscopy, we were unable to calculate the specificity of the diagnostic tests. The MRI field strength used in the present study (1.0 Tesla) may be considered a limitation, although it is not studied whether a higher field strength might have resulted in a more reliable diagnosis of articular cartilage lesions of the patellae. A higher field strength is not always routinely available. However, the present study has several strengths too. Our study sample consisted of young military trainees between 18 and 28 years of age who all passed their military entry medical examination as healthy. After completing the military basic training, all were expected to have attained a physical level enabling them to march or ski a distance of 15 km on two consecutive days, carrying a full military pack weighing 25 kg and a rifle, while maintaining fitness for battle. Due to the compulsory nature of the military service, the study participants accurately reflect the young adult male population of the country. Moreover, the age group of our study is optimal, because this is the age frame in which degenerative chondral changes, such as osteoarthritis, do not normally exist. This relatively large group of skeletally mature young adults underwent uniform examination and treatment for anterior knee pain according to the policy of a single institution. Standard methods and protocols of the institution were used for the evaluation of clinical symptoms, plain radiographs, magnetic resonance images, and arthroscopy findings. In conclusion, the sensitivity of MRI for detecting grade-i lesions was low and MRI findings could not be used to confirm a diagnosis of articular cartilage lesions of the patellae. For the detection of more severe grade-ii to III lesions, the sensitivity of MRI was markedly higher. MRI may therefore be considered a useful diagnostic tool for identifying more severe cases of articular cartilage lesions of the patellae. In addition, the sensitivity of symptoms and physical examination for articular cartilage lesions of the pa-

Sensitivity of MRI for articular cartilage lesions of the patellae 61 tellae is poor and diagnosis cannot be based on symptoms alone. ACKNOWLEDGEMENTS The study was financially supported by the Medical Research Fund of Tampere University Hospital., Tampere, Finland. REFERENCES 01. Leslie IJ, Bentley G: Arthroscopy in the diagnosis of chondromalacia patellae. Ann Rheum Dis 1978;37:540 547 02. Llopis E, Padron M: Anterior knee pain. Eur J Radiol 2007; 62:27 43 03. Teitge RA: Patellofemoral syndrome a paradigm for current surgical strategies. Orthop Clin North Am 2008;39:287 311, v 04. Wittstein JR, O Brien SD, Vinson EN et al: MRI evaluation of anterior knee pain: predicting response to nonoperative treatment. Skeletal Radiol 2009;38:895 901 05. Dye SF: The pathophysiology of patellofemoral pain: a tissue homeostasis perspective. Clin Orthop Relat Res 2005;100 110 06. McCauley TR, Kier R, Lynch KJ et al: Chondromalacia patellae: diagnosis with MR imaging. AJR Am J Roentgenol 1992;158: 101 105 07. Lee SH, Suh JS, Cho J et al: Evaluation of chondromalacia of the patella with axial inversion recovery-fast spin-echo imaging. J Magn Reson Imaging 2001;13:412 416 08. Macarini L, Perrone A, Murrone M et al: Evaluation of patellar chondromalacia with MR: comparison between T2-weighted FSE SPIR and GE MTC. Radiol Med 2004;108:159 171 09. Kruger T, Wohlrab D, Birke A et al: Results of arthroscopic joint debridement in different stages of chondromalacia of the knee joint. Arch Orthop Trauma Surg 2000;120:338 342 10. Casscells SW: Arthroscopy of the knee joint. A review of 150 cases. J Bone Joint Surg Am 1971;53:287 298 11. Dehaven KE, Dolan WA, Mayer PJ: Chondromalacia patellae in athletes. Clinical presentation and conservative management. Am J Sports Med 1979;7:5 11 12. Shahriaree H: Chondromalacia. Contemporary Orthopaedics 1985;27 39 13. Federico DJ, Reider B: Results of isolated patellar debridement for patellofemoral pain in patients with normal patellar alignment. Am J Sports Med 1997;25:663 669 14. Stoller D: Magnetic resonance imaging in orthopaedics and sports medicine. 3rd ed. 2007, Baltimore: Lippincott Williams and Wilkins 15. Mattila VM, Kuronen P, Pihlajamaki H: Nature and risk factors of injury hospitalization in young adults: a follow-up of 135,987 military conscripts. Scand J Public Health 2007;35:418 423 16. Suh JS, Cho JH, Shin KH et al: Chondromalacia of the knee: evaluation with a fat-suppression three-dimensional SPGR imaging after intravenous contrast injection. J Magn Reson Imaging 1996;6:884 888 17. Handelberg F, Shahabpour M, Casteleyn PP: Chondral lesions of the patella evaluated with computed tomography, magnetic resonance imaging, and arthroscopy. Arthroscopy 1990;6: 24 29 18. Gagliardi JA, Chung EM, Chandnani VP et al: Detection and staging of chondromalacia patellae: relative efficacies of conventional MR imaging, MR arthrography, and CT arthrography. AJR Am J Roentgenol 1994;163:629 636 19. Grelsamer R, Moss G, Ee G et al: The patellofemoral syndrome; the same problem as the Loch Ness Monster? Knee 2009;16:301 302 Received: February 2, 2011 Accepted: May 24, 2011