Evaluation of Accuracy of Clinical Examination and MRI on Diagnosing Anterior Cruciate Ligament and Meniscal Tears in Comparison to Diagnostic Arthroscopy among Patients Attending at Muhimbili Orthopedic Institute

Routine use of Magnetic Resonance Imaging (MRI) as screening test after clinical diagnosis for meniscal and/or anterior cruciate ligament (ACL) has a detrimental effect on patients in limited resourced countries. This study was done to compare accuracy of clinical examination and that of (MRI) on diagnosing meniscal and or (ACL) tears. Methodology: A cross-sectional-descriptive study was done on 57 knees of patients. Clinical examination, MRI and then diagnostic arthroscopy, as the gold standard, were done to all the cases. Results were recorded; the accuracies of MRI and clinical examination were evaluated and their results were compared. Results: Median age of patients was 40 years. Clinical examination had sensitivity of 93.62% and specificity of 40% for diagnosing meniscal tears; and sensitivity of 100%; and specificity of 97.67% for diagnosing ACL tear. MRI had sensitivity of 85.11%, and specificity of 40% for meniscal tear diagnosis and 71% and 100% respectively for ACL tear diagnosis. Diagnostic accuracy was 84.21% for meniscal and 98.24% for ACL tears by clinical examination and by MRI was 77.19% niscal and or ACL injuries. MRI use should be reserved when clinical evaluation is inconclusive or cannot be done.


Introduction
Menisci and cruciate ligaments are among the frequently injured structures of the knee hence one of commonest causes of knee pain. These structures have various functions including stabilising the joint, acting as axis for knee rotational movement and neutralization of sudden blows or forces to the knee [1] [2]. Anterior cruciate ligament (ACL) and menisci injuries often occur during sports activities and motor traffic crashes. Other risk factors associated with meniscal and ACL tears include obesity, and female sex [3]- [8]. When these structures are injured, there are detrimental effects to the knee cartilage. Numerous studies have shown that tears of the meniscus lead to development and progression of osteoarthritis of the knee, hence early detection and proper management are of paramount importance [9] [10] [11] [12].
Diagnosis of ACL and meniscal tears can be done clinically, as initial assessment, through thorough history taking and physical examination. The tests done include joint line tenderness, the Apleys test, McMurrays test, Lachman and the anterior drawer test, with various sensitivities and specificities [13]- [18]. It is also done radiologically, which is non invasive, using the magnetic resonance imaging (MRI) scan.
A thorough clinical examination, especially using more than one test, has been shown to have accuracy similar to that of MRI by several authors, on diagnosing ACL and meniscal tears [19] [20] [21] [22] [23]. Hence some studies recommend that the MRI should serve as an additional investigation in doubtful, inconclusive or complex knee injuries only [24] [25] [26] [27].
However, there are other researchers who recommend the use of MRI, as a primary screening tool, to compliment clinical examination of the patients in deriving to diagnosis of meniscal and/or ACL tear, prior to performing arthroscopy [28] [29] [30] [31] [32]. Therefore it is routinely used after clinical assessment of patients, a trend that poses potential risk of lowering emphasis that doctors need to put on clinical examinations of such patients [24] [25] [27]. MRI however is scarcely available in developing countries like ours and it is expensive. This may prevent early and timely treatment of many patients. Hence it is crucial to consider the economic load of doing these MRI scans to all our patients. This also may cause delays in providing proper and timely treatments [27] [33] [34]. In our setting, data on its accuracy on diagnosing ACL and meniscal tears as compared to clinical assessment are limited, hence the study was carried out to find this out.

Methodology
A descriptive cross sectional study was done in 57 knees of patients with menisci and/or ACL tears. Clinical examination, MRI and then diagnostic arthroscopy were done to all the candidates at Muhimbili Orthopedic Institute (MOI) between May 2018 to December 2018. MOI is the largest Orthopaedic and trauma referral center in Tanzania. Knee arthroscopic surgeries generally are performed by Arthroscopy and Sports Medicine Fellowship trained Orthopaedic Surgeons, and few other Orthopaedic Specialists. Patients were enrolled from clinics. Convenient sampling technique was used to obtain the sample for the study. Sample size was calculated using the average occurrence through MOI medical records. The pilot study, that was done for patients treated for ACL and meniscal injuries in the past period from January 2017 to June 2017 which revealed prevalence of 3.3%.
Inclusion criteria for the study were consented patients, of 18 and above years; patients with knee symptoms clinically diagnosed to have meniscus tear and or anterior cruciate ligament tear; patients who were also suitable to do MRI and clinically fit for arthroscopic procedure. The exclusion criteria included patients with previous history of surgical interventions of the ipsilateral knee; with ipsilateral other joint conditions such as neoplasm, inflammatory diseases, infections or grade three and above of degenerative changes seen on plain radiography; patients with acute tear or injury preventing proper physical examination; with associated periarticular or intra articular fractures and with contraindication to do MRI like having intracerebral aneurysmal clips, cardiac pacemaker, metallic foreign body or metal implants which are not made of titanium.
Two Arthroscopy and Sports medicine Fellowship trained Orthopaedic Surgeons and the principal investigator were involved. Thorough history taking and physical examinations were done. McMurray test, joint line tenderness and Apley's test were done to assess for meniscal tears. Anterior drawer test and Lachman test were done to assess for presence of ACL tear. The combined results of physical tests were used to draw the conclusion for presence or absence of meniscal and ACL tears respectively. Findings were recorded. Standard plain xrays were then done and evaluated to ensure the inclusion criteria are met.
MRI scanning of all knees was done before doing diagnostic arthroscopy. This MRI scanning process was done in different diagnostic centers. All the images were collected/uploaded and stored in the compact discs and films. These MRI had similar magnetic field intensity. The discs and films were studied and reported by a Consultant Radiologist at MOI. Results were then recorded. These patients was lastly scheduled for diagnostic arthroscopy, where the findings were recorded and used as gold standard for determining accuracy of MRI and accuracy of clinical results. The time intervals between the performance of the MRI scan and the arthroscopy were recorded. All arthroscopic procedures were performed by the Arthroscopy and Sports medicine Fellowship trained Orthopaedic surgeons. These are the same steps that are applied in our routine practices at the hospital, hence the study represented realistic scenario of how it is carried out when patients attend at the outpatient clinics.
Information about the patients were collected through interviews and filled in the structured questionnaires; physical examination, MRI scanning and arthroscopic findings were also recorded.
At the end of data collection, three groups of patients were established, that is, those with isolated injury of meniscus or ACL; those with both meniscal and ACL injury and those with normal MRI or clinical examination but with knee symptoms and positive results from diagnostic arthroscopy.
Data analysis was done using Stata version 15.1 (StataCorp, College Station, Texas) where the final analyses were performed. Five main parameters for finding validity of these diagnostic methods were calculated using arthroscopic diagnostic results as reference points/gold standard. Sensitivity, Specificity, Negative Predictive value and Positive Predictive Value and overall Accuracy of both MRI and clinical examination were calculated. The comparison of diagnostic accuracy of MRI and clinical examination was done to find out which had higher accuracy on diagnosing meniscal and or ACL tear where Receiver Operating Characteristic (ROC) curve analysis was also used.
Statistical significance of variables was found using Wilcoxon rank-sum (Mann-Whitney) test.

Socio-Demographics
Among patients enrolled in the study, female were the predominant group with 59.65% (n = 34) of participants. Overall, the median age was 40 years with inter quartile range (IQR) of 28 -44 years. Females with knee injuries were predominantly older than male patients with median of 42 years (IQR 37 -45 years) as compared to 37 years (IQR 24 -42 years) respectively. This difference was statistically significant (p < 0.05). Majority of participants had secondary education. Statistical significance of variables was found using Wilcoxon rank-sum (Mann-Whitney) test. Table 1 elaborates on the socio demographic features of the participants, explaining on their age, level of education and occupation.

Mechanisms of Injury
The mechanisms that led to the tears of menisci and ACL were grouped into four categories. Sports and unknown mechanism, as some patients could not recalled exactly how they sustained the injuries, were the leading causes of tears among the patients.
The categories frequencies of mechanism of injury were statistically significant (p < 0.05). Figure 1 showing percentage contributions of injury mechanisms as grouped into four categories.  However, in the unknown mechanism group of patients, 79% associated their knee pain with their daily activities like jogging, running and squatting that they were regularly doing at home. Other patients reported that the pain was felt when they were frequently walking long distances, climbing stairs at their work places (>30 stairs) and standing for long time (approximately above 4 hours) at their work places. All of these patients had isolated meniscal tears and more than 88% were females of whom majority of them appeared obese.
Most of females were observed to delay (>2 yrs) to seek medical help as compared to males counterpart (about 6 months) (p < 0.05). Time interval between clinical and MRI diagnosis and doing diagnostic arthroscopy ranged from two to three weeks and none of participants sustained new injury during that time interval.

Arthroscopic Findings
Diagnostic arthroscopy was eventually done to all the cases. There was higher number of females with meniscal tears than males. On the contrary, more males had ACL tears than females. However in both scenarios the differences seen were not statistically significant (p > 0.12 and p > 0.22 respectively). Table 3 elaborates results of diagnostic arthroscopy that were used as confirmatory results for the true diagnosis of patients.
Therefore the sensitivities, specificities, negative and positive predictive values of MRI and clinical examination were calculated, with diagnostic arthroscopic as the gold standard. Finally accuracieof both were found and compared. Table 4 shows Clinical diagnostic results for ACL.      Table 7 summarises the sensitivity, specificity, negative and positive predictive values: and overall accuracy of the MRI results. Table 8 shows Clinical diagnostic results for meniscal tear. Table 9 summarises the sensitivity, specificity, negative and positive predictive values: and overall accuracy of the clinical examination results for meniscal tear. Table 10 shows MRI results for meniscal tear. Table 11 summarises the sensitivity, specificity, negative and positive predictive values: and overall accuracy of the MRI meniscal tear results.

Evaluation of Diagnostic Accuracy of MRI as Compared to That of Clinical Examination
The diagnostic accuracy of MRI and that of clinical examination were also found through ROC (Receiver Operating Characteristic) curve area to best illustrate comparison of diagnostic performances between clinical examination and MRI scan in each segment of concern. Open Journal of Orthopedics    Accuracy of diagnosing anterior cruciate ligament tear by clinical examination was 98.24% and that of MRI scan was 92.98%. Therefore, clinical examination had a higher accuracy as compared to MRI scan in diagnosing ACL tears and the difference seen was statistically significant (p < 0.05).

Discussion
In this study the diagnostic accuracy of physical examination was higher than that of MRI; with p < 0.05 for ACL tear and p > 0.09 for meniscal tears.
In the study, majority (45%) of patients who did not remember the actual mechanism of their injuries, had meniscal tears. Similar phenomenon was seen in the study done by Cerabona and colleagues where they observed that greater number of their participants could not recall well the mechanism of injury nor they could not correlate their injury mechanisms with specific pattern of meniscal tears they sustained. [35] A study done by Barbara reported that work related  walking > 2 miles per day, standing > 2 hrs per day and stair climbing of >30 flights per day had 1.65, 1.63 and 2.28 times risks of sustaining meniscal tears respectively [36]. This positive history of work related walking, stair climbing, routine home based physical exercises and overweight may have a role in the unexplained meniscal tears, as these were also seen as some of risk factors for meniscal tears by researchers [4] [36]. The physical examination was found out to have higher sensitivity as compared to MRI on diagnosing ACL injuries (100% vs 71%). The study done by Gaisgow also revealed that physical examination had 100%sensitivity on diagnosing ACL tear while MRI had 94% [37]. Similarly, Nilton et al. showed that physical examination has a higher sensitivity than MRI on diagnosing ACL tear [19].
De Havens, [38] who was comparing physical examination with arthroscopic findings showed that Lachman test alone has an accuracy of 100% when patient is under anasthesia. However, Specificity of MRI in this study was slightly higher than that of clinical examination (100% vs 97.67%). This was also observed in the study done by De Havens and the study by Muthuuri [38] [39].
Decreased sensitivity of MRI may be attributed by difficulties of displaying the ligament fully on sagittal plane due to its anatomical obliquity passage across the joint, as depicted by some studies [40]. MRI results are operator dependent hence the more the skilled and experienced the operator is, the better capability of depicting of the lesion, as elaborated by Geijer [41]. Magnetic Frequency Intensity of the MRI scanner also may affect the accuracy of MRI to pick up the torn ACL, although the effect has been shown by studies to be insignificant [42] [43].
For diagnosing meniscal tears, sensitivity of physical examination was shown Rayan et al. [44] evaluated 131 patients and found that physical examination had better sensitivity and diagnostic accuracy for meniscal tears than MRI (86% vs 76%). The same finding was seen in a study done by Ercin et al. [24] where clinical examination had higher accuracy in diagnosing medial meniscal tears than MRI (93% vs 83%) and marginal higher accuracy for lateral meniscal tears. However, study done by Leonardo showed that sensitivity of MRI in detecting meniscal injury is higher than that of physical examination (100% vs 85% for medial meniscus and 89% vs 70% for lateral meniscus) [29]. Similar results was also shown in the study by Yan, [45] where MRI had higher sensitivity and accuracy when compared with physical examination in diagnosing meniscal tears.
The differences observed in these studies may be attributed due to differences in skills and experiences of the involved Radiologists in interpreting the images and Medical Personnels who performed the clinical examinations as these have an impact on finding the correct results as elaborated in studies. Differences in MRI scanning protocols in various centres may interfere with accuracy of results. The protocol specifics vary depending on centre's MRI machine software and hardware, radiologist involved, time constraints and patient factors [26] [40].
The study revealed that the overall diagnostic accuracy of physical examination in detecting anterior cruciate ligament and meniscal tears is higher as compared to that of MRI. Similar findings were seen in the study done by Navali et al. where he reported that accuracy of physical examination to be 95.8% and that of MRI to be 92.8% [46]. Nilton et al. evaluated 120 patients and found that diagnostic accuracy of physical examination was better than that of MRI (90.27% vs 83.33% respectively) [19]. The same results of higher accuracy of clinical examination than of MRI in diagnosing ACL and meniscal tears were reported by Kocabey [26]. Thomas et al. found out that the diagnostic accuracy of MRI is lower hence unnecessary MRI scanning increases financial burden and delays treatments to patients [47].
However, Noha et al. [48] was evaluating reliability and value of MRI in ACL and meniscal tear diagnosis. He found MRI to have higher overall accuracy in diagnosing them as compared to physical examination.
The differences observed above may be affected by the person carrying out the task. Accuracy of physical examination for meniscal and ACL tears depends on the person performing the physical tests. Usually, when the tests are being performed by experienced or skilled medical personnels, results are better than when tests are performed by inexperienced medical personnels [19] [28]. This may be the reason for higher accurate results in clinical examination. However, in this study, all the MRI results were reported by Consultant Radiologist to minimise bias effect.

Study Limitations
The MRI scan was done in various centers hence some cd reports of different operating systems could not open at Muhimbili Orthopedic Institute hence led to loss of some participants. However these MRIs were of similar magnetic field intensity.
Convenient sampling technique was applied due to unavoidable circumstances hence this study has shed some light on the subject and calls for bigger randomized studies in the future Time constraint, that limited the number of participants hence i urgently encourage further study with bigger sample size to better represent the population

Conclusion
Physical examination had higher accuracy than MRI (98.24% vs 92.98% for ACL (p < 0.05) and 84.21% vs 77.19% for meniscal tears (p > 0.09)). When clinical signs and symptoms are well evaluated and results indicate ACL and or meniscal tear, doing an MRI prior to performing diagnostic and interventional arthroscopy is unlikely to be of significance. It is therefore not recommended. Hence MRI use in these diagnoses should be reserved for specific situations like in acute phase of injury where physical examinations cannot be done, in inconclusive physical examination results or in more complex knee injuries. This will foster early and appropriate treatment; and avoid unnecessary costs of MRI to patients.

Recommendations
All Doctors are urged to utilise thoroughly history taking and physical examination skills on patients with knee complaints. This may lower significantly the unnecessary MRI done and allow patients with meniscal or ACL tears early access to appropriate interventions, at lower costs.

Data Availability
Data can be made available on request from the corresponding author.

Funding Statement
This study was self-funded, by the principal author Open Journal of Orthopedics