Deep-learning-assisted diagnosis for knee magnetic resonance imaging: Development and retrospective validation of MRNet.

BACKGROUND: Magnetic resonance imaging (MRI) of the knee is the preferred method for diagnosing knee injuries. However, interpretation of knee MRI is time-intensive and subject to diagnostic error and variability. An automated system for interpreting knee MRI could prioritize high-risk patients and assist clinicians in making diagnoses. Deep learning methods, in being able to automatically learn layers of features, are well suited for modeling the complex relationships between medical images and their interpretations. In this study we developed a deep learning model for detecting general abnormalities and specific diagnoses (anterior cruciate ligament [ACL] tears and meniscal tears) on knee MRI exams. We then measured the effect of providing the model's predictions to clinical experts during interpretation. METHODS AND FINDINGS: Our dataset consisted of 1,370 knee MRI exams performed at Stanford University Medical Center between January 1, 2001, and December 31, 2012 (mean age 38.0 years; 569 [41.5%] female patients). The majority vote of 3 musculoskeletal radiologists established reference standard labels on an internal validation set of 120 exams. We developed MRNet, a convolutional neural network for classifying MRI series and combined predictions from 3 series per exam using logistic regression. In detecting abnormalities, ACL tears, and meniscal tears, this model achieved area under the receiver operating characteristic curve (AUC) values of 0.937 (95% CI 0.895, 0.980), 0.965 (95% CI 0.938, 0.993), and 0.847 (95% CI 0.780, 0.914), respectively, on the internal validation set. We also obtained a public dataset of 917 exams with sagittal T1-weighted series and labels for ACL injury from Clinical Hospital Centre Rijeka, Croatia. On the external validation set of 183 exams, the MRNet trained on Stanford sagittal T2-weighted series achieved an AUC of 0.824 (95% CI 0.757, 0.892) in the detection of ACL injuries with no additional training, while an MRNet trained on the rest of the external data achieved an AUC of 0.911 (95% CI 0.864, 0.958). We additionally measured the specificity, sensitivity, and accuracy of 9 clinical experts (7 board-certified general radiologists and 2 orthopedic surgeons) on the internal validation set both with and without model assistance. Using a 2-sided Pearson's chi-squared test with adjustment for multiple comparisons, we found no significant differences between the performance of the model and that of unassisted general radiologists in detecting abnormalities. General radiologists achieved significantly higher sensitivity in detecting ACL tears (p-value = 0.002; q-value = 0.019) and significantly higher specificity in detecting meniscal tears (p-value = 0.003; q-value = 0.019). Using a 1-tailed t test on the change in performance metrics, we found that providing model predictions significantly increased clinical experts' specificity in identifying ACL tears (p-value < 0.001; q-value = 0.006). The primary limitations of our study include lack of surgical ground truth and the small size of the panel of clinical experts. CONCLUSIONS: Our deep learning model can rapidly generate accurate clinical pathology classifications of knee MRI exams from both internal and external datasets. Moreover, our results support the assertion that deep learning models can improve the performance of clinical experts during medical imaging interpretation. Further research is needed to validate the model prospectively and to determine its utility in the clinical setting.

Viscoelastic properties of common suture material used for rotator cuff repair and arthroscopic procedures.

PURPOSE: The purpose of this study was to evaluate the viscoelastic properties of 5 suture materials, commonly used in arthroscopic rotator cuff repairs, when subjected to physiological loads. METHODS: We evaluated 5 commercially available No. 2 sutures undergoing both creep and cyclic testing in both dry air and phosphate-buffered saline solution (PBS) maintained at 37°C. The selected sutures were MagnumWire (ArthroCare, Austin, TX), Ethibond (Ethicon, Somerville, NJ), FiberWire (Arthrex, Naples, FL), Orthocord (DePuy, Warsaw, IN), and Force Fiber (Tornier, Bloomington, MN). RESULTS: Regarding creep testing, in the PBS test environment, FiberWire showed the greatest stiffness (71.1 ± 2.1 N/mm), the smallest initial extension at the 60-N load (1.10 ± 0.04 mm), and the smallest amount of creep (0.57 ± 0.01 mm). Orthocord showed the smallest amount of relaxed elongation in PBS (0.73 ± 0.11 mm). Regarding cyclic testing, in the PBS testing environment, Ethibond exhibited the smallest dynamic creep (0.28 ± 0.02 mm), FiberWire displayed the smallest peak-to-peak displacement (0.17 ± 0.00 mm), and Orthocord showed the smallest amount of relaxed elongation after cyclic loading (0.63 ± 0.11 mm). CONCLUSIONS: FiberWire consistently displayed more extreme viscoelastic properties--greater stiffness and less extensibility--than the other suture types studied. Orthocord showed the smallest amount of relaxed elongation in both testing environments. Differences in testing environment affect the behavior of each suture type. Testing in physiologically approximating conditions such as PBS maintained at 37°C is warranted. CLINICAL RELEVANCE: Although many other factors affect the success of rotator cuff repairs, the viscoelastic properties of sutures may be a useful predictor of suture performance.

Novel drug product to improve joint motion and function and reduce pain after arthroscopic anterior cruciate ligament reconstruction.

PURPOSE: OMS103HP, an investigational drug product containing ketoprofen, amitriptyline, and oxymetazoline, is added to arthroscopic irrigation solution. OMS103HP was evaluated in patients undergoing arthroscopic anterior cruciate ligament reconstruction to assess the drug's safety and ability to improve postoperative knee function and motion, reduce postoperative pain, and allow earlier return to work. METHODS: This was a prospective, double-blind, vehicle-controlled, parallel-group, randomized study. Allograft anterior cruciate ligament reconstruction patients in both treatment and vehicle control groups were monitored for safety and efficacy (e.g., measurements of knee function and motion, pain, and return to work) over a 30-day postoperative period. The efficacy endpoints of primary interest were assessed by use of both responder and time-to-event analyses. RESULTS: There were statistically significant differences (P < or = .05) between the OMS103HP and vehicle control groups in the endpoints of knee function (knee function composite and straight-leg raise component of knee function composite), range of motion (median number of days to maximum passive flexion of 90 degrees or greater without pain and time to discontinuation of continuous passive motion), quadriceps and hamstring muscle strength, successful pain management, and return to work. The overall incidences of adverse events and abnormal clinical laboratory values for both OMS103HP- and vehicle-treated subjects were similar, and none was attributed to OMS103HP. CONCLUSIONS: The clinical benefits of OMS103HP in this study were reduced postoperative pain; improved postoperative knee motion, quadriceps and hamstring muscle strength, and knee function; and earlier return to work as measured by surgeons, nurses, and physical therapists across repeated clinic visits and rehabilitation sessions and recorded by patients in daily diaries. The drug was well tolerated.

Acute compartment syndrome of the thigh in a football athlete: a case report and the role of the vacuum-assisted wound closure dressing.

We present a case of compartment syndrome of the thigh due to blunt injury in a Division I American football player managed with fasciotomy and vacuum-assisted wound closure. This case report discusses the vacuum-assisted wound closure dressing as an alternative to more traditional closure techniques such as suture retention devices and split-thickness skin grafting. We feel that any surgeon involved in performing fasciotomies should be familiar with this increasingly used closure device and its potential complications.

Controversial topics in shoulder arthroscopy.

This is a review of the symposium entitled Controversial Topics in Shoulder Arthroscopy 2001, presented at the Arthroscopy Association of North America 20th Annual Meeting in Seattle, Washington in April of 2001. The covered topics include micro-instability, partial thickness rotator cuff tears, pathology of the biceps tendon, and thermal capsulorrhaphy.

Electrothermal assisted shoulder capsulorraphy--monopolar.

Monopolar electrothermal stabilization of the shoulder shows considerable promise as a treatment alternative in athletes and patients with recurrent instability. Range of motion is preserved, recovery is faster than with open procedures, there is little disruption or alteration of inherent anatomy, and, most importantly, results at 2 years appear comparable to other surgical procedures in high-demand populations. The procedure is technically easy to perform, and the complication rate is low. Success depends on surgeon's understanding of the applications of RF energy, the use of proper surgical technique, careful patient selection, attention to the rehabilitation program, and the patient's compliance with postoperative protocol. Long-term follow-up is necessary to determine if results for this procedure deteriorate over time, especially in patients with multidirectional instability.

Septic arthritis in postoperative anterior cruciate ligament reconstruction.

A review of postoperative infected anterior cruciate ligament reconstructions was done on 3500 consecutive arthroscopic procedures. The purpose was to assess incidence, diagnosis, treatment, and outcome factors. Six postoperative intraarticular infections were detected. Average followup was 3 years (range, 2-8 years). The rate of infection was 0.14%. Five men and one woman with a median age of 32.5 years (range, 20-51 years) comprised the study group. The average interval from the onset of symptoms to the initial arthroscopic intervention was 7.5 days (range, 2-20 days). Staphylococcus aureus was present in three knees, Staphylococcus epidermidis in two, and Streptococcus nonhemolytic in one. All patients had initial arthroscopic debridement and lavage followed by 6 weeks of intravenous antibiotics. Two grafts were removed: one patient had delayed ligament reconstruction and the other had total knee arthroplasty. The remaining patients had full range of motion. In the group with the best result, two patients had Staphylococcus epidermidis and one had Staphylococcus aureus, which was treated 2 days after clinical symptoms began. The other two patients infected with Staphylococcus aureus had unsatisfactory results. Anterior cruciate ligament infection is rare, but diagnosable. When treated early with appropriate antibiotic therapy and arthroscopic debridement, four of six grafts were retained. If the infection does not respond rapidly to early therapy, then graft removal is an option.

Monopolar thermal treatment of symptomatic anterior cruciate ligament instability.

Patients with anterior cruciate ligament instability resulting from incomplete tears or elongation in continuity without ligament detachment historically have been treated conservatively or by graft replacement. The literature is sparse regarding alternative treatments. The current study presents experience using monopolar thermal repair on 28 consecutive knees with partial anterior cruciate ligament tears all symptomatically unstable. All lesions were less than 6 months old (average, 77 days; range, 7-180 days) and with a difference of 6 mm or more (average, 9 mm; range, 6-13 mm) when comparing both knees using KT-1000 evaluation. Incomplete tears of the anterior cruciate ligament were seen at arthroscopic evaluation. The rehabilitation protocol included use of a brace for at least 6 weeks and progressive weightbearing. A 2-year minimum followup (range, 24-35 months) was done in all patients following the International Knee Documentation Committee guidelines. The overall outcome was normal or nearly normal in 96% of the patients. One failure occurred at 8 weeks. Twenty-six knees had a KT-1000 difference between 0 and 2 mm (average, 1.9 mm). Because thermal application causes death to some of the cells directly treated, it should be taken into account in selection and application. Immediately after thermal use, the anterior cruciate ligament, although thicker and tighter, is at first weaker than normal. Rehabilitation and compliance are critical during early ligament healing. This procedure seems to be a reasonable alternative to anterior cruciate ligament grafting in selected patients.

Anterior cruciate ligament reconstruction using cryopreserved allografts.

Primary ACL reconstruction historically has been done using autograft tissues whereas allografts have been limited to revision cases and patients who are older or with lower physical demands because some animal studies suggested a slower biologic incorporation rate. The purpose of the current study was to evaluate the effectiveness of the cryopreserved Achilles tendon allograft in primary ACL reconstruction in a consecutive series of athletes. Fifty consecutive patients with a strenuous or moderate preinjury activity level, as defined by the International Knee Documentation Committee (IKDC), had ACL reconstruction using cryopreserved Achilles tendon allografts secured with bioabsorbable interference screws. Five patients were professional athletes. The average age of the patients was 36 years (range, 17-50 years). A 3- to 5-year followup study was done in all of the patients using the IKDC form. Tunnel widening was measured in the lateral radiographs at the widest level. The overall outcome was normal or nearly normal in 94% of the patients. No failures were reported in this series. Forty-six patients (92%) returned to their same preinjury sport activity level. The average KT-1000 side-to-side difference was 2.3 mm. Average tibial tunnel widening was 2.7 mm (range, 0-6 mm); no significant correlation was observed between increased tunnel size and a fair or poor clinical outcome. This experience shows that favorable results can be obtained with cryopreserved Achilles tendon allografts in athletes in whom avoiding donor site morbidity may be an issue in terms of a prompt return to sport.