Red Staining of the LHB and the Rotator Interval on Arthroscopy-A Valid Indicator of Symptomatic Inflammation? A Machine Learning Assisted Investigation.

Background: Tendinopathies of the long head of the biceps (LHB) are a common issue in shoulder surgery.

Purpose/hypothesis: The purpose of this study is to determine whether the red discoloration observed in the LHB and synovial rotator cuff interval during arthroscopy invariably indicates significant inflammation, which is frequently used to guide therapeutic decisions about LHB tenotomy or tenodesis. Red coloration of the LHB and the surrounding rotator cuff interval would not necessarily be related to the clinical symptoms of the patient.

Study design: Cross-sectional study; Level of evidence, 3.

Methods: Patients undergoing arthroscopy for symptomatic LHB pathology were included in the study if they met the inclusion criteria. Images of patients who were undergoing arthroscopy for other shoulder pathologies, but who did not exhibit LHB symptoms, were designated as the control group. Standardized imaging and video techniques were employed to identify the LHB and to calculate an objective measure of the total red component of the respective LHB. Machine learning (ML) was utilized to enumerate the respective red component in pixel count for each LHB, along with the redness of the synovial rotator cuff interval. The extent of effusion within the bicipital groove (as evident from axial slides of preoperative magnetic resonance imaging) was noted and subsequently correlated. The ranges into which the red hue of the LHB and the surrounding rotator cuff interval were divided included >60%, 60% to 31%, 30% to 11%, and <10%.

Results: A total of 124 patients (men = 84, women = 40) were included, with a collective sum of 36,941 arthroscopy images (both standalone images and video frames). A total of 31 superior labral lesions, from anterior to posterior lesions, and 10 pulley lesions were detected. Moreover, 46 lesions were subjectively assessed by the surgeon as having LHB tendinitis based on LHB red "inflammation" staining intraoperatively. ML-based quantification revealed >60% LHB redness in 17 instances, between 60% to 31% in 22, between 30% to 11% in 33, and <10% in 52 instances. However, it could not be established that a significant statistical correlation exists between the intensity and extent of LHB redness, or the associated redness in the rotator interval, and their corresponding clinical symptoms related to LHB pathologies. In addition, there was no positive correlation between the effusion area and the LHB red staining, the surrounding sulcus bicipitalis area, or indications of pain.

Conclusion: No positive correlation was found between the redness/intensity of the LHB on arthroscopy and the clinical symptoms within the LHB tests. The assessment of visually detectable "inflammation" of the LHB within arthroscopy as a therapeutic criterion must be questioned.