Inter-tubercle angle measured on MRIs may serve as a predictor for degenerative subscapularis tears in patients diagnosed with rotator cuff tears under arthroscopy: a case-control study

Abstract

Background

Subscapularis (SSC) tears are very common, but the mechanism of the tear remains unclear. Previous studies have revealed that SSC tears have associations with the reduction of coracohumeral distance and increase of coracoid overlap. However, existing parameters cannot explain the predominance of articular-side tears. Thus, new signs need to be found.

Methods

A total of 51 patients who were diagnosed as SSC tears were retrospectively included in the study from 2023 to 2024. 42 rotator cuff tear patients without SSC tears were also enrolled as the control group. With the arm in neutral rotation, the coracohumeral distance (CHD), coracoid overlap (CO), intertubercular angle (ITA), and humerus head-glenoid cavity ratio (HCR) were measured on MRI images of the coracoid disappearance plane. Independent samples t-test was used for analyzing the difference.

Results

The mean ITA in patients with SSC tears was 90.2 ± 11.2° (range, 62.4-126.3°) and was significantly higher (P = 0.0002) than that in control group (79.5 ± 10.8°; range, 55.7-109.3°). The HCR between both groups showed no significant difference (P = 0.193). CHD and CO measurements were consistent with previous studies.

Conclusion

Our study confirmed the association between ITA and SSC tears. Patients with SSC tears tended to have a bigger ITA. However, the tendency of SSC tears was independent of HCR. ITA may serve as a potential predictor for SSC tears.

Key terms

rotator cuff tear; coracohumeral distance; coracoid overlap; inter-tubercle angle; humerus head-glenoid cavity ratio; subscapularis tears.

Background

Rotator cuff tears represent a common age-related pathology, with a prevalence exceeding 30% among adults aged 60 years and older [1, 2]. Notably, subscapularis tears have garnered increased clinical attention due to their functional implications. Epidemiological studies confirm that older populations exhibit a disproportionately higher prevalence of subscapularis tendon injuries [3].

The subscapularis (SSC), the largest component of the rotator cuff complex, originates from the scapula’s subscapular fossa and inserts on the lesser tubercle of the humerus. Functionally, it internally rotates the humerus and stabilizes the glenohumeral joint, preventing anterior humeral head displacement [4–6]. It forms the rotator cuff with the supraspinatus, infraspinatus, and teres minor muscles, enabling multiplanar shoulder motion [4].

Clinically, subscapularis tears present with shoulder pain, internal rotation weakness, and limited glenohumeral motion. If conservative treatment fails, arthroscopic surgery may be required. These injuries cause chronic pain and add socioeconomic burden to healthcare systems and families [3]. The precise etiopathogenesis of subscapularis tears remains unclear. Therefore, more research on tear mechanism is essential to guide targeted therapies.

In elderly populations, subscapularis tears are predominantly associated with age-related degenerative changes. Among proposed etiological theories, subcoracoid impingement represents the predominant explanatory framework. In 2005, Burkhart et al. first linked a reduced coracohumeral distance (CHD) to subscapularis pathology [5, 7]. Subsequent studies have consistently identified CHD as a critical biomechanical parameter. A study demonstrate that CHD decreases during active humeral internal rotation (LaPrade et al.) [8]. It also shows significant asymmetry in bilateral shoulders of individuals with unilateral subscapularis tears (Chen et al.) [9]. Unlike degenerative tears, traumatic subscapularis injuries fail to demonstrate significant CHD reduction, suggesting distinct etiopathogenetic pathways.

Beyond CHD, several morphological predictors of subscapularis pathology have been identified [10]. Significant research focus has centered on coracoid process morphology. Kawamata et al. demonstrated that a rounded or beak-shaped coracoid process correlates with increased susceptibility to subscapularis tears [11]. Yoo et al. further identified the coraco-lesser tuberosity angle as a significant variable, reporting greater values in affected patients [12]. Complementing these findings, Leite et al. (2019) established that increased coracoid overlap (CO) is associated with subscapularis tears [10].

Whether focusing on CHD, CO, or other radiographic parameters, attention has been directed toward the friction damage between the subscapularis tendon and the coracoid process. Such damage would theoretically be concentrated on the surface of the SSC that faces the coracoid process. However, in our clinical observations, we found that subscapularis injuries also occur on the articular side. This anatomical specificity necessitates identifying alternative mechanisms governing tendon failure.

Consequently, this article investigates the biomechanical relationship between intrinsic tendon tension and anatomical torque, proposing two novel predictive parameters: intertubercular angle (ITA) and humerus head-glenoid cavity ratio (HCR). We hypothesize that elevated ITA and HCR values are associated with increased risk of subscapularis tendon failure and will elucidate their underlying pathomechanical mechanisms.

Materials and methods

Patient selection and surgical operation

This study received approval from the Ethics Committee of Peking University Third Hospital. Approximately 200 patients who underwent shoulder arthroscopy at our institution between 2023 and 2024 were retrospectively analyzed. Patients with glenohumeral osteoarthritis, intra-articular tumors, or inflammatory joint diseases were excluded. After also excluding those with significant prior ipsilateral humeral or scapular trauma, previous surgery on the affected shoulder, or inadequate radiologic image quality, 93 patients with rotator cuff tears were identified. From this population, 42 patients without subscapularis tears were assigned to the control group (Fig. 1). Among the 51 cases in the SSC tear group, we identified 17 special cases. Their subscapularis was torn but not clearly visible on T2-weighted MRI images, whereas arthroscopy revealed the lesion. We compared them with the control group to further validate the association between ITA, HCR and SSC tear. Demographic data including age, sex, and body mass index (BMI) were systematically extracted from all participants’ medical records for subsequent analysis.

Fig. 1.

Study flow diagram

Subscapularis tears were confirmed by arthroscopic observation. Preoperative MRI evaluates tear size, location, and tendon integrity. After brachial plexus block and general anesthesia, three to five 5-mm arthroscopic portals are established. Diagnostic arthroscopy identifies tendon discontinuity and degeneration. Repair consists of anatomical reattachment of the torn tendon to the lesser tubercle using suture anchors. The procedure ends with portal closure and sterile dressing.

MRI measurement protocols

All measurements were conducted using a GE MRI system by two trained musculoskeletal radiologists. They separately conducted two measurements at the interval of two months to ensure the accuracy. Scans were acquired with the humerus in neutral rotation. The axial plane at the coracoid tip disappearance level served as the primary measurement plane; when osseous landmarks were obscured by high-signal artifacts, adjacent superior/inferior planes were referenced. The lesser tubercle (subscapularis insertion), greater tubercle (infraspinatus insertion) and humeral head’s geometric center were identified. If the infraspinatus and supraspinatus muscles cannot be clearly distinguished on this plane, sagittal images can be used for localization. In the triangle formed by these three points, the angle with its vertex at the geometric center is the ITA. On this plane, we measure the length of the most lateral part of the glenoid and the diameter of the humerus; the ratio of these two is the HCR (Fig. 2). CHD was measured as the minimal coracoid-to-humeral-head distance per Burkhart et al. [5]. CO was quantified as the distance from the coracoid’s most anterolateral point to the glenoid axis, consistent with Leite et al. [10]. Final values represented the mean of both examiners’ measurements after confirming intra- and inter-observer reliability with intraclass correlation coefficients (ICCs) not less than 0.75.

Fig. 2.

The T2-weighted magnetic resonance image showing the measurement

The humeral head’s geometric center was the ITA vertex, with the angle formed by lines connecting the vertex to each tubercle. DHH is diameter of humeral head. LGC is length of glenoid cavity. HCR is LGC divided by DHH. CHD was defined as the minimal coracoid-to-humeral-head distance. CO was defined as the distance from the coracoid’s most anterolateral point to the glenoid axis.

Statistical methods

Two-way mixed-effects model was used to evaluate intra-observer reliability, and two-way random-effects was used to evaluate inter-observer reliability [13]. Continuous demographic and radiographic parameters were compared between subscapularis tear and control groups using independent samples t-tests. The comparison of HCR used the Mann–Whitney U test. Categorical gender distribution was analyzed using Pearson’s chi-square test to determine association with tear status. All analyses were performed in IBM SPSS Statistics (Version 26.0), with statistical significance defined as p < 0.05 (two-tailed).

Results

Table 1 shows the ICCs between the two examiners (A and B) and the two examines they each took(A₁, A₂, B₁, B₂). The value of all the four imaging data is greater than 0.75, which indicates the excellent agreement of measurement.

Table 1.

ICCs of the radiologic parameters between the two examiners

Radiologic parameters	A1 vs. A2	B1 vs. B2	A1 vs. B1	A2 vs. B2
ITA	0.90	0.97	0.88	0.83
HCR	0.91	0.88	0.84	0.77
CHD	0.85	0.91	0.95	0.80
CO	0.96	0.97	0.95	0.87

This study comprised a cohort of 93 patients, stratified into a subscapularis tear group (n = 51) and a control group without tears (n = 42). No significant intergroup differences existed in gender distribution, height, weight, or body mass index. In the SSC tear group, the mean age is 57.67 ± 7.94, compared to 53.50 ± 12.56 in the control group (p = 0.055). This p value approaches statistical significance. To exclude its potential influence, a multivariable logistic regression model was constructed with group assignment as the dependent variable and all aforementioned demographic and imaging parameters as independent variables. The resultant analysis yielded an age-related significance level of 0.832, indicating that age does not constitute a determinant of disease status. These demographic data are presented in Table 2.

Table 2.

Demographic data of the two groups

	Subscapularis tear group	Control group	p-value
	(n = 51)	(n = 42)
Age (mean ± SD)	57.67 ± 7.94	53.50 ± 12.56	0.055
Sex (female/male)	27/24	19/23	0.707
Weight (kg, mean ± SD)	70.88 ± 11.23	72.60 ± 11.65	0.473
Height (cm, mean ± SD)	165.14 ± 8.56	167.67 ± 9.86	0.195
BMI (mean ± SD)	26.04 ± 2.70	25.74 ± 3.02	0.614

The radiologic parameters of the two groups are shown in Table 3. In the SSC tear group, the mean ITA was 90.2 ± 11.2°, compared to 79.5 ± 10.8° in the control group. The SSC tear group had a significantly larger ITA (p < 0.001, Table 3). The mean HCR was 0.56 ± 0.07 in the SSC tear group and 0.54 ± 0.07 in the control group, with no significant difference (p = 0.193, Table 3). The mean CHD was 8.06 ± 1.47 mm in the SSC tear group and 11.10 ± 2.11 mm in the control group, significantly lower in the SSC tear group (p < 0.001, Table 3). The mean CO was 17.63 ± 3.69 mm in the SSC tear group and 11.45 ± 2.85 mm in the control group, significantly larger in the SSC tear group (p < 0.001, Table 3).

Table 3.

Radiologic parameters of the two groups

	Subscapularis tear group	Control group	p-value
	(n = 51)	(n = 42)
ITA (°, mean ± SD)	90.2 ± 11.2	79.5 ± 10.8	< 0.001*
HCR (median, lower quartile, upper quartile)	0.55 (0.51, 0.62)	0.55 (0.50, 0.58)	0.354
CHD (mm, mean ± SD)	8.06 ± 1.47	11.10 ± 2.11	< 0.001*
CO (mm, mean ± SD)	17.63 ± 3.69	11.45 ± 2.85	< 0.001*

In order to test and verify the specificity and sensitivity of ITA, we constructed a receiver operating characteristic (ROC) curve and area under the ROC curve (AUC) was 0.77, which was shown in Fig. 3. When the value was set at 85.7°, the sensitivity was 0.73, and the specificity was 0.76.

Fig. 3.

Receiver operating characteristic curve of the intertubercular angle(ITA)

Considering all these factors, larger ITA and CO are associated with an increased likelihood of SSC tears, and smaller CHD may be related to them as well. However, no significant impact of HCR, BMI or other factors on SSC tears has been found so far.

Table 4 shows the comparison between the 17 special group mentioned before and the control group. The mean ITA is 87.1 ± 11.4°, larger than 79.5 ± 10.8° of the control group (p = 0.012). The HCR is 0.59 (0.53, 0.63), larger than 0.55 (0.50, 0.58) of the control group (p = 0.030). In the meantime, CHD is smaller than control group (p < 0.001), and CO larger than the control group (p < 0.001). This indicates that ITA is a probable radiologic predictor for SSC tears.

Table 4.

Demographic data and radiologic parameters of the special case group

	Special case group	Control group	p-value
	(n = 17)	(n = 42)
Age (mean ± SD)	59.69 ± 7.50	53.50 ± 12.56	0.069
Sex (female/male)	11/6	19/23	0.176
BMI (mean ± SD)	26.18 ± 2.01	25.74 ± 3.02	0.593
ITA (°, mean ± SD)	87.1 ± 11.4	79.5 ± 10.8	0.012*
HCR (median, lower quartile, upper quartile)	0.59 (0.53, 0.63)	0.55 (0.50, 0.58)	0.030*
CHD (mm, mean ± SD)	8.01 ± 1.12	11.10 ± 2.11	< 0.001*
CO (mm, mean ± SD)	16.22 ± 3.32	11.45 ± 2.85	< 0.001*

Discussion

To our knowledge, this is the first study to reveal the association between ITA, HCR and SSC tears. In our study, patients with SSC tears exhibited a significantly higher ITA compared to the control group. Additionally, the measurements of CHD and CO were consistent with our expectations. There was no difference in HCR between overall subscapularis tear group and the control group. But the special cases’ HCR was larger than in the control group. The reason for the difference in HCR may be that the sample size of the special group was too small, which led to statistical significance. Further validation with more special cases is needed. The meaning of the comparison between the special case group and the control group lies in that if the MRI images don’t show the erosion of SSC, ITA can probably be a imaging predictor for SSC tear to prepare for the surgery.

Although the precise etiology of SSC tears remains unclear, primary research focuses on intrinsic tendon degeneration, subacromial impingement (SAI), and trauma. In 2003, Lo and Burkhart described the “roller-wringer effect”: during shoulder inter rotation (IR), the coracoid compresses the superficial SSC tendon, causing degeneration. Thus, a lower CHD increases impingement risk [5]. In 2016, Oh et al. used dynamic ultrasound to show that minimal CHD occurs during shoulder IR with adduction, supporting this mechanism [14].

Since the coracoid impingement model’s inception, consistent evidence has established CHD as a key biomarker for SSC integrity. Hekimoglu et al. (2012) used MRI to show that CHD positively correlates with Acromio-humeral interval (AHI) and Coraco-clavicular interval (CCI) (p < 0.01) [15]. Brunkhorst et al. (2013) subsequently used dynamic US to prove active glenohumeral-IR significantly reduces CHD [8], directly increasing subcoracoid impingement potential. Balke et al. (2016) confirmed diminished CHD specifically associates with degenerative SSC lesions, unlike traumatic tears where CHD remains normal versus controls [7]. Zhu et al. (2021) further established pathological-side CHD reduction versus asymptomatic contralateral shoulders [9], confirming inter-limb asymmetry as a diagnostic indicator.

Recent evidence links CO to SSC pathology. Leite et al. (2020) found greater CO in SSC tear patients than controls [10]. Consistently, Zhu et al. similarly reported increased CO in affected shoulders versus asymptomatic contralaterals. This fits the coracoid impingement model: higher CO means greater distal coracoid prominence, raising mechanical conflict risk during glenohumeral motion [9]. Thus, CO enlargement indicates elevated SSC tear susceptibility.

Beyond CHD and CO, three other radiographic predictors of SSC pathology have been identified. First, the coraco-lesser tuberosity angle (CLA) is significantly larger in SSC tear patients than in controls (41.4 ± 4.2° vs. 38.7 ± 4.0°, p < 0.001). However, the mechanism behind this finding remains unclear [12]. Second, coracoid morphology affects tear susceptibility. The rates are: flat 21.8%, round 64.7%, beak 46.3% (p < 0.001). A more prominent coracoid tip increases mechanical injury to the SSC tendon, consistent with the roller-wringer mechanism [11]. Third, muscular architecture studies show pathologic SSC remodeling [16–18]. Compared to intact tendons, the muscle belly is 10% shorter, while the tendon is 32% longer in torn tendons [18].

Subscapularis tendon tears often initiate at the insertion site on the lesser tuberosity and then extend medially as they worsen. Yoo classified this process into five types, ranging from mild to severe [19]. In our clinical practice, we observed that subscapularis tendon tears usually begin on the articular side of the tendon, and the roller-wringer effect does not adequately explain this phenomenon. Therefore, a new mechanism needs to be identified.

We hypothesized that the intrinsic subscapularis tension mechanism may be involved in tear development, and therefore we used ITA and HCR for validation. The results confirmed that in SSC tears, the ITA is larger. In the humeral traction system composed of the infraspinatus and SSC, a larger ITA indicates a more medial insertion of the SSC on the lesser tuberosity. This leads to a smaller moment arm of the SSC tendon. We speculate that to maintain the same traction effect (torque), the SSC requires a greater tensile force—that is, increased intrinsic tension of the SSC—which increases its susceptibility to injury. This requires confirmation by subsequent finite element analysis and biomechanical experiments, which may be conducted in future studies to explain the issue more clearly.

Study limitations include its single-center design, precluding multicenter validation. This is a retrospective study, so it cannot definitively establish a causal relationship between ITA and SSC tears; it can only suggest that an association may exist between the two. In terms of measurement, there may be ambiguity in identifying the insertion of the infraspinatus tendon, which could affect the applicability of the conclusions to some patients. Additionally, the small sample size of the tear group may limit the generalizability of the findings. Meanwhile, sample size constraints prevented age-matching between SSC tear and control groups, introducing potential confounding. Demographic data collection was restricted to basic parameters (age/sex/BMI); unmeasured confounders like smoking history or repetitive overhead activities may influence outcomes.

Conclusion

Our study demonstrated that the ITA in patients with SSC tears was markedly higher than in those without, suggesting it may be associated with such tears.

Author contributions

YS and XC were responsible for conceptualization of the experimental framework, design of the methodology, and review and revision of the initial manuscript. FY was responsible for verification of the experimental design and oversight of the research progress. YW and XF participated in the initial literature search. ZS and ZL jointly contributed to the collection of patients’ MRI scans, parameter measurements, data analysis, chart production, and the drafting of the initial manuscript.

Funding

This research received the funding of National Natural Science Foundation of China (grant number 82372438), Beijing Natural Science Foundation (grant number 7252163), Peking University Clinical Scientist Training Program, supported by “the Fundamental Research Funds for the Central Universities” (grant number BMU2025PYJH013), and Key clinical project of Peking University Third Hospital (grant number BYSY2022049).

Data availability

The data that support the findings of this study are available from Peking University Third Hospital but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Peking University Third Hospital.

Declarations

Ethics approval and consent to participate

This study received approval from the Ethics Committee of Peking University Third Hospital (application number M20260220).

Consent for publication

The publication of personal data in this article has been consented to by the participants.

Competing Interests

The authors declare no competing interests.