Muscle injuries are one of the most prevalent ailments in any sport that requires explosive movements. In the case of football in particular, 92-97% of such injuries occur in the lower limbs, the most common being those affecting the hamstrings (28-37%), quadriceps (19-32%), adductors (19-23%) and calf muscles (12-13%). The wide range of recovery patterns and time they may present poses a challenge in deciding the ideal time for a player to return to competition. Rushing into this decision could potentially lead to reoccurrence and, consequently, to a longer final recovery time than it should have been in the first place.

Despite preventive strategies, many athletes (12-43%) suffer from a reoccurrence of their muscle injuries during the rehabilitation period. The reasons for this are diverse: the lack of consensus on the ideal time to return to competition in the case of hamstring injuries (HMI), the wide range of injury and recovery types, the high physical demands during matches or the different criteria in rehabilitation protocols. Not even the most sophisticated imaging methods, such as magnetic resonance imaging, have proven to be a reliable tool for predicting the exact moment at which an injured football player should return to competition.

Based on this premise, Dr. Xavier Valle, a member of the medical staff at F.C. Barcelona, led a study intended to find a classification system capable of assessing the degree of injury, offering a prognosis of the time of return to competition and identifying the injuries with the highest reocurrence rate – Return to Play Prediction Accuracy of the MLG‐R Classification System for Hamstring Injuries in Football Players: A Machine Learning Approach. The classification system selected was the MLG-R, based on magnetic resonance imaging. Its acronym refers to the mechanism of skeletal muscle injury (M), its location (L), the degree of severity (G) and the number of muscle re-injury (R). Three different statistical and machine learning approaches (linear regression, random forest, and eXtreme Gradient Boosting) were used to assess the importance of each factor of the MLG-R classification system in determining the return to play, as well as to offer a prediction of the expected return to play.

Relevant Variables to Assess the Time for Return to Play

Between 2010 and 2020, 76 hamstring injuries met the inclusion criteria for the study which included FC Barcelona, FC Barcelona B, Juvenil A and Juvenil B players. Regarding the results, the most frequently affected muscle was the biceps femoris long portion with 51 cases (71.1%), the most frequent location was the proximal third of the thigh around the proximal myotendinous junction with 33 cases (43.4%), and the most common grade was the 3rd with 50 cases (65.8%). The average duration of all injuries was 29.11 days.

The comprehensive approach based on the three statistical approach models and machine learning showed that the degree of injury (3rd) and location (free tendon) were the main determinants for return to play. When assessing differences in return to play according to degree of severity, grade 3 injuries showed a waiting time for return to play of 37 days, while the period for the other grades ranged from 12-14 days.

Another relevant indicator for assessing return to play is the location of the injury at the myotendinous junction. When the injury affects the proximal myotendinous junction, the average recovery period is 31.7 days, while if it affects the distal myotendinous junction, the average recovery period drops to 23.9 days.

The highest recovery time was detected in grade 3 injuries of the biceps femoris long head located in the free tendon, with a mean time of 56 days. The time is reduced drastically when this type of injury is located in the medial tendon, with a mean time of 24 days.

The Extracellular Matrix, a Key Factor in Prognosis

One of the main contributions of the study was to focus attention in the evaluation of hamstring injuries to the role of the extracellular matrix and connective tissue structures between the muscle and tendon. For years, muscle contraction has been analysed in a linear, unidimensional and simplistic way, just like the extracellular matrix (ECM). However, muscle contraction occurs in three dimensions, so it is necessary to assess the muscle as a whole to understand its structure, function, mechanism and pathology.

The ECM is a three-dimensional structure distributed asymmetrically and from higher to lower structural density. Therefore, a thorough understanding of the muscles anatomy and specially of the myotendinous junctions is basic to the proper understanding of muscle injuries.

The implementation of the classification system introduced in the present study provides an objective description of the injury at a topographical (where), chronological (how many times) and structural (degree of injury) level, thus minimising the subjectivity of the description and enabling the return to play to be predicted more accurately than with other publications.

It must be pointed out that the main factor that causes a longer return to play after a hamstring injury is the injury affecting the connective tissue structures of the hamstring. Therefore, the ECM structure and its role in force generation and transmission is a key factor in identifying the signs, symptoms and prognosis of muscle injuries which is why the proposed classification was designed with the aim of assessing and quantifying ECM injury.