Return to Play Prediction Accuracy of the MLG-R Classification System for Hamstring Injuries in Football Players: A Machine Learning Approach

Abstract

BackgroundMuscle injuries are one of the main daily problems in sports medicine, football in particular. However, we do not have reliable means to predict the outcome, i.e. return to play (RTP) from severe injury. The aim of the present study was to evaluate the capability of the MLG-R classification system to grade hamstring muscle injuries by severity, offer a prognosis for the RTP, and identify injuries with higher risk of reinjury. And to assess the consistency of our proposed system by investigating its intra-and inter-observer reliability. MethodsAll male professional football players from FC Barcelona (FCB)-senior A and B and the two U-19 teams-with injuries occurred between February 2010 and February 2020 were reviewed. Only players with clinical presentation of hamstring muscle injury, with complete clinic information and magnetic resonance images (MRI), were included. Three different statistical and machine learning approaches (linear regression, random forest, and XGBoost) were used to assess the importance of each factor of the MLG-R classification system in determining the RTP as well as to offer a prediction of the expected RTP. We used the Cohen's kappa and the intraclass correlation coefficient (ICC) to assess the intra-and interobserver reliability. ResultsBetween 2010 and 2020, 76 hamstring injuries corresponding to 42 different players were identified, of which 50 (65.8%) were Grade 3 r , 54 (71.1%) affected the biceps femoris long head (BFlh), and 33 of the 76 (43.4%) were located at the proximal myotendinous junction (MTJ). The mean RTP for Grade 2, 3, and 3 r injuries were 14.3, 12.4, and 37 days, respectively. Injuries affecting the proximal MTJ had a mean RTP of 31.7 days while those affecting the distal part of the MTJ had a mean RTP of 23.9 days. The analysis of the grade 3r BFlh injuries located at the FT showed a median RTP time of 56 days while the injuries located at the central tendon had a shorter RTP of 24 days (p=0.038). The statistical analysis showed an excellent predictive power of the MLG-R classification system with a mean absolute error of 9.8 days and an R-squared of 0.48. The most important factors to determine the RTP were if the injury was at the free tendon (FT) of the BFlh or if it was a Grade 3 r injury. For all the items of the MLG-R classification the intra-and inter-observer reliability was excellent (k > 0.93) except for fibers blurring (κ = 0.68). ConclusionThe main determinant for long RTP after hamstring injury is the injury affecting the connective tissue structures of the hamstring.We developed a reliable hamstring muscle injury classification system based on MRI that showed excellent results in terms of reliability, prognosis capability and objectivity. It is easy to use in clinical 2 daily practice, and can be further adapted to future knowledge. The adoption of this system by the medical community would allow to uniform diagnosis leading to better injury management.

Association Between Magnetic Resonance Imaging Findings and Time to Return to Sport After Muscle Injuries in High-Level Youth Athletes

“…First, the involvement of connective tissue seems to be a risk factor for longer time loss (18.5 days for muscle, 31 days for MTJ, and 62 days for tendon). This had already been reported in athletics studies by Pollock et al 6,20 and in elite football players by Valle et al 43 and Shamji et al 7 Their findings suggest that it is not only the connective tissue involvement, but also the extent of the damage and its specific location (eg, endomysium, perimysium, and epimysium), which could be the main determinant of injury severity. Second, proximal injuries showed extended RTS compared with distal and mid-third injuries.…”

“…Second, proximal injuries showed extended RTS compared with distal and mid-third injuries. The literature is conflicting in this regard, with some studies indicating that proximal injuries are prognostic of extended time loss, 43,45 whereas others do not report any correlation. 20,39 It is possible that proximal location is not a prognostic factor by itself, but a consequence of the anatomical location of some of the connective structures (eg, proximal MTJ, biceps femoris cojoined tendon, and free tendon) that have proved to cause longer time loss when they are involved.…”

MRI findings prior to return to play as predictors of reinjury in professional athletes: a novel decision-making tool

“…Despite the fact that in recent years certain publications have downplayed or questioned the role of MRI in monitoring muscle injuries [ 21 , 26 ], the current results demonstrate that certain MRI findings are predictive of reinjuries and may be a useful tool in the multifactorial decision-making process of recovery and RTP [ 35 – 37 ].…”

Indirect Rectus Femoris Injury Mechanisms in Professional Soccer Players: Video Analysis and Magnetic Resonance Imaging Findings

“…[1][2][3][4][5] In football, thigh muscle injuries are the most common injuries, rectus femoris (RF) being the most frequent location in the quadriceps 4,[6][7][8][9] When deciding the best treatment for a muscle injury, a correct diagnosis and injury classification are the first steps. 10 A full comprehension of the muscle belly affected, the injury location, and an evaluation of the connective tissue damage are the basis for designing the treatment. 10,11 The most common treatment approach for RF injuries is nonoperative, but there are certain indications in which surgical treatment should be considered.…”

“…10 A full comprehension of the muscle belly affected, the injury location, and an evaluation of the connective tissue damage are the basis for designing the treatment. 10,11 The most common treatment approach for RF injuries is nonoperative, but there are certain indications in which surgical treatment should be considered. 5,[12][13][14][15] Understanding newly discovered anatomical structures of the RF can aid in diagnosis and treatment.…”