The main result emerging from the obtained data is that, indeed, there is a correlation between the presence of specific anatomical variants of the intraarticular portion of the long head biceps tendon and the SLAP lesion. Until now, many of these conditions have been considered insignificant and not responsible for shoulder pathology.
Only a few authors have looked for a relationship between anatomic variants and upper shoulder pathologies in adults [22,23,24]. Studies in the literature show an association between distinct types of long head biceps tendon anchorage and shoulder instability secondary to post-traumatic biceps anchor injury . In particular, an association between posterior anchorage and SLAP lesion is demonstrated. 
Many authors have described pathologies and anatomic variants of the superior region of the shoulder [1, 3, 7, 14, 21, 25,26,27,28,29,30,31,32,33,34]. However, few authors classified or analyzed the variants systematically with a large group of participants [21, 25]. Some studies have classified or analyzed the anatomical variants of long head biceps tendon , while others have examined the diagnostic accuracy of MRI in diagnosing pathological patterns . However, these studies were mainly conducted in small groups of patients and not systematically performed.
Congenital variants of the relationship of the long head biceps tendon with the rotator cuff have been reported in the literature; some have described absence of long head biceps tendon [22, 30,31,32,33], others the intracapsular and extracapsular course of the tendon [22, 23, 36]. Anatomical variants of long head biceps tendon origin are reported to occur with a frequency of 1.9–7.4%.
Our data show that split tendon is a rare finding, found in 0.9% of cases, while the variant characterized by adherent long head biceps tendon with the supraspinatus was seen in 12.3% of cases, followed by the variant with the mesotenon, in 9.5% of cases.
Many studies have shown that these conditions are inherited and are the result of partial detachment from the mesothelium or synovial fusion with the inferior surface of the capsule [37,38,39].
With respect to the association between certain anatomic variants and the presence of SLAP lesions, in our case series, it appears that the long head biceps tendon mesotenon (MESO-PA) variant may predispose to superficial shoulder pathology; specifically, we found a higher-than-expected frequency of SLAP 1 in the group of patients with MESO-PA anatomical variant. The OR for this variant, though, could not be calculated.
To the best of our knowledge, there are no studies in the literature that can confirm or disprove this finding. The reason for this apparent association could be found in a decreased range of movement of the intraarticular portion of the long head biceps tendon.
In concordance with the literature, we found significant correlation between the adherent long head biceps tendon group and the free long head biceps tendon group for SLAP lesion [12, 13, 40, 41]. In fact, compared with expectation, we found a higher frequency of type II SLAP in patients with adherent (ADH-PM) long head biceps tendon.
This result could also be in line with the association of type I SLAPs in the presence of mesotenon-type anatomic variants. In this case, because there is an additional limitation to the range of motion of the long head biceps tendon intraarticular portion, the association is with a type II SLAP. As the degree of fixation increases, the degree of injury to the glenoid fibrocartilage seems to be higher.
Finally, we found an increased risk of developing SLAP type IV in those patients who presented with a SPL-long head biceps tendon as an anatomical variant. In this case, the presence of this anatomical variant seems to act as a risk factor itself in developing a type IV SLAP lesion.
Again, to the best of our knowledge, there are no studies in the literature that can confirm or debunk this observation. The reasons for this result would appear to belong to a different order than the previously described variants and are likely to be due mainly to biomechanical reasons similar to those leading to a greater association of SLAP in cases of posterior CLB tendon insertion described by Jakanani et al. .
In our case series, we did not find any cases of long head biceps tendon agenesis. Four anatomical variants, from hypoplasia of the biceps to its complete absence, have been described in the literature . Other authors have described three variants of the proximal extracapsular portion of the long head biceps tendon . Cases of split tendon and complete fusion with the inferior surface of the rotator cuff have been described in three cases [24, 34, 37].
The strengths of our study are the large initial sample size, the systematic use of MR arthrography, and the single-center design of the study. Other authors evaluated the sensitivity and specificity of MRI versus MR arthrography in 150 shoulders. The sensitivity and specificity values of MRI without contrast medium were respectively 83% and 100% for anterior labrum lesions, 85% and 100% for SLAP lesions, and 92% and 100% for supraspinatus tendon lesions.
Compared to arthroscopy, MR arthrography reaches sensitivity and specificity values respectively of 98% and 100% for anterior labral lesions, specifically 98% and 99% for SLAP lesions and 100% sensitivity and specificity for cuff lesions [7, 13].
A limitation of our study is the absence of systematic comparison between MR-arthrography and arthroscopy, but data in the literature consider MR arthrography the gold standard in the evaluation of the upper portion of the shoulder [5, 43].