This study included 78 patients, determined using Cohen's Kappa sample size formula, aged >40 years, collected through consecutive sampling. The inclusion criteria were patients from outpatient clinics who presented with pain at the top of the shoulder and symptoms of a painful arc. Patients with a history of fracture, limited passive range of motion, shoulder surgery, signs of infection, and/or claustrophobia were excluded from the study.

Data collection began with recording demographic and clinical information, including age, sex, height, weight, symptom duration, pain level (assessed using the numeric rating scale [NRS]), dominant arm, and affected side, and were sourced from medical records and patient history. Primary data collection commenced with clinical tests for a painful arc during abduction (90°–140°), followed by Neer sign, Hawkins-Kennedy, empty can, and drop arm tests.

Subsequently, ultrasonography was performed to directly examine the affected shoulder by a certified sonographer with over four years of experience in shoulder imaging. Examination utilized a Konica Minolta Sonimage HS2 device equipped with a high-frequency linear transducer (L18-3, up to 18 MHz). Participants were instructed to position their shoulder in the crass or modified crass position. The supraspinatus tendon was scanned along both long and short axes (axial and sagittal planes).

A low-volume injection of 1% lidocaine (1–2 mL) was then administered into the subacromial space under ultrasonographic guidance using the in-plane technique. Five minutes post-injection, the clinical tests were repeated to assess improvements, and the findings were documented. Within the same week, participants underwent MRI of the affected shoulder using a GE SIGNA Pioneer 3 Tesla MRI machine and shoulder coil.

The results of the measurements, analyzed using descriptive statistics, were presented as categorical data in numerical form. The chi-square test was employed to assess the correlation between the affected side and arm dominance. For clinical tests, test results before and after injection were combined and used for parallel testing. Sensitivity, specificity, positive predictive value, negative predictive value and accuracy were computed for various parameters. Cohen’s kappa coefficient was used to determine intermodality agreement between two variables. All calculations were performed using the statistical software SPSS version 26.0 (IBM Corp.) and Microsoft Excel.

In evaluating suspected pathological processes in the supraspinatus tendon (whether tendinopathy, partial, or full tear), Neer sign, Hawkins-Kennedy test, and empty can test demonstrated higher sensitivity (68.3%, 73%, and 63.5%, respectively) and lower specificity (53.3%, 20%, and 20%, respectively). With subacromial lidocaine injection, the specificity of these tests increased, but this was accompanied by a decrease in sensitivity. In contrast to the other tests, the drop arm test showed lower sensitivity (31.7%) compared to higher specificity (93.3%), and the sensitivity further decreased with subacromial lidocaine injection (9.5%) (Table 2).

In evaluating full tears, Neer sign, Hawkins-Kennedy test, and empty can test also exhibited higher sensitivity (62.1%, 75.9%, and 75.9%, respectively) and lower specificity (34.7%, 26.5%, and 38.8%). Subacromial lidocaine injection increased specificity but was accompanied by a decrease in sensitivity, similar to the detection of supraspinatus tendon pathological processes. The drop arm test also exhibited consistent findings of lower sensitivity (41.4%) and higher specificity (81.6%), with sensitivity decreasing (17.2%) and specificity increasing (95.9%) following subacromial lidocaine injection (Table 3).

On ultrasound imaging for identifying pathological processes in the supraspinatus tendon, 27 of 29 full tears identified on MRI were correctly detected, yielding a sensitivity of 93.1%, specificity of 95.9%, and an accuracy of 94.87%. For normal supraspinatus tendon cases, ultrasound successfully identified 14 of 15 patients, resulting in a sensitivity of 93.3% and specificity of 96.8%, with an accuracy of 96.15%. For partial tears, the sensitivity was 78.9% and specificity was 94.9%, while for tendinopathy, the sensitivity was 80% and specificity was 93.7%, with an accuracy of 91.03% for both types of patients. Ultrasound was able to identify 12 of 15 tendinopathy patients and 15 of 19 partial supraspinatus tendon tear patients (Table 4).

Ultrasound imaging of the supraspinatus tendon in normal and full tear (total tear) cases had a kappa value >0.8, indicating a very strong strength of agreement. For tendinopathy and partial tear cases, the kappa value ranged from 0.7 to 0.8, indicating a substantial strength of agreement. Overall, the data presented in Table 4 yielded a kappa value of 0.825, indicating a very strong level of agreement.

In this study, the average age of the sample aligns with systematic review research showing that shoulder pain is most common in individuals over 50 years old. This study focuses on degenerative rotator cuff tears, prevalent in those over 40, so younger participants were excluded. While this affects the sample's average age, it ensures relevance to the target population. The prevalence of shoulder pain and rotator cuff tears increases with age [12]. The majority of the patients reporting shoulder pain were women, consistent with previous studies [13]. In this study, 80% of participants exhibited pathological processes associated with shoulder pain, and 61% had supraspinatus tendon tears, either partial or full. These findings are consistent with epidemiological data showing that 30%–70% of patients with shoulder pain have rotator cuff tears [14]. Epidemiologically, the prevalence of rotator cuff tears is higher in men than in women (2:1) in cases of trauma or injury [3,14]. However, in degenerative cases, women outnumber men at a ratio of approximately 3:1 [3,15]. This aligns with the results of this study, as the majority of patients visiting the clinic were degenerative cases.

Population-based studies show that the risk of shoulder pain is higher in individuals who are overweight or obese based on BMI, and risk also increases significantly with a larger waist circumference [16]. Despite the average BMI of 22.87kg/m², which falls at the higher end of the normal weight category, this study did not measure waist circumference. A strong relationship between arm dominance and shoulder pain side is supported by previous studies [17]. This suggests that intrinsic and extrinsic factors collaboratively influence the occurrence of pathological processes in the rotator cuff tendon, with the incidence increasing with age (intrinsic factor) and arm dominance in activities (extrinsic factor) [17].

Most patients had already undergone alternative treatments and sought care from other specialists, such as medical rehabilitation and neurology. All patients received treatment in primary healthcare facilities, followed by secondary and eventually tertiary care, where they were recruited for this study. The time spent in each facility and the referral process contributed to the prolonged duration of shoulder pain. The majority of the sample reported moderate pain intensity according to the NRS. No patients reported severe pain. This may be related to the duration of pain experienced before visiting the clinic. Chronic cases tend to have lower pain intensity than acute cases because acute bursitis inflammation plays a more significant role in patient complaints [18]. Additionally, the patients were patients visiting outpatient clinics, meaning most patients presented with tolerable pain, whereas those with severe pain typically sought emergency care. Shoulder pain, mostly caused by rotator cuff tears, tends to be chronic and dull.

Patients with rotator cuff tears can present with shoulder dysfunction, where the affected shoulder cannot be lifted or abducted due to weakness secondary to pain, either as a protective measure or a pain avoidance mechanism [14]. Physical examination of the painful arc was used as an inclusion criterion because it is considered the most sensitive physical examination (up to 96%) for all subacromial pathological processes, including the supraspinatus tendon [18]. Subacromial pain accounts for 85% of all shoulder problems, ranging from subacromial bursitis and tendinosis to cuff tears [19].

The Hawkins-Kennedy clinical test had the highest sensitivity among the four clinical examinations performed in this study for detecting supraspinatus tendon abnormalities. For detecting full-thickness supraspinatus tears, the Hawkins-Kennedy test and the empty can test had the highest sensitivity. In terms of specificity, the drop arm test had the highest specificity for detecting both tendon abnormalities and full-thickness tears of the supraspinatus. These findings are consistent with previous research [18]. With subacromial lidocaine injection, the specificity of all clinical examinations for detecting rotator cuff abnormalities or full-thickness supraspinatus tears increased, but this was accompanied by a decrease in sensitivity. The accuracy of these tests decreased for detecting supraspinatus tendon abnormalities but increased for identifying full-thickness tears.

We hypothesize that the inability to perform certain movements or maintain certain positions against resistance in these clinical examinations is due to the disruption of tendon integrity, and subacromial lidocaine injection can eliminate acute pain syndrome. The results of this study suggest that inflammation of the bursa contributes to positive results in these clinical examinations, not just tendon pathology or tendon integrity.

The decrease in sensitivity and accuracy for identifying tendon pathology after lidocaine injection indicates that positive results were predominantly caused by pain and bursal inflammation. In identifying full-thickness tears specifically, pain and inflammation also played a role in positive sensitivity results, as evidenced by the reduction in sensitivity but increased specificity and accuracy after subacromial lidocaine injection. The pain observed in these tests is more closely associated with the level of substance P in the subacromial bursa rather than the structural pathology of the rotator cuff tendon itself [20]. In fact, a study by Gschwend et al. [21] found that pain in tendinopathy and partial tears (especially bursal-side tears) was more severe than in full-thickness tears. This proves that the degree of pain does not correlate with the severity or size of tendon tears, and the pain response from physical examination is thus independent of the mechanical impact of the tear [22].

The low sensitivity of all examinations performed, particularly the drop arm test, both before and after subacromial lidocaine injection, indicates the physical examinations’ inability to identify specific anatomical structures or pathological processes. This is likely due to the complexity of the anatomical structure of the shoulder and the functional interrelation between these structures [19]. The involvement of tear thickness, whether one-third, two-thirds, or complete tears, as well as tear size from small to large, affects the cuff's function. The pathological process of detachment, defect, or retraction also plays a role. The effectiveness of muscle force transmission by the rotator cuff is also influenced by the integrity of the rotator cable [23,24]. A study by Halder et al. [24] showed that cuff retraction contributed more to shoulder weakness and loss of strength than the occurrence of cuff tears.

Additionally, joint stability is influenced by the force couple mechanism, where a supraspinatus tear that reduces stability in the coronal plane alongside the deltoid muscle can be compensated for by the subscapularis and infraspinatus muscles. These two muscles are capable of generating compressive joint reaction forces on the axial plane sufficient to maintain joint stability against mass moment inertia and compression into the glenoid socket, forming a functional cuff tear [25,26]. Therefore, clinical test can provide misleading results influenced by pain, concurrent pathological processes, or the patient's ability to adapt to the pathological process [22].

From the 78 patients evaluated in this study, the average accuracy reached over 90%, with a kappa value above 0.7 compared to MRI. This result indicates that the diagnostic accuracy of ultrasonography using a high-frequency linear transducer up to 18 MHz is acceptable for detecting rotator cuff abnormalities. However, the accuracy for detecting full-thickness tears is better compared to partial tears or tendinopathy. Most previous studies used 7.5-MHz linear transducers, whereas this study used a transducer capable of up to 18 MHz, which supports an increase in the reliability of the imaging results [27]. Increasing the transducer frequency enhances spatial resolution but is accompanied by reduced penetration and image depth [27]. The impact of BMI also affects the results, with the average BMI of the sample falling within the normal weight category in this study.

A limitation of this study is that physical examinations and ultrasonography were performed directly by the researchers, which could introduce potential bias as each test might influence the likelihood of findings (workup bias). However, this approach replicates daily clinical practice, integrating clinical history, physical examination, and ultrasonographic imaging [28]. The high number of normal results on imaging is a drawback of using ultrasonography alone, as imaging was based only on shoulder pain complaints, not on suspicion of rotator cuff tears. Additionally, a high rate of pathological findings (up to 50%) are observed on imaging of asymptomatic shoulders [29]. Therefore, a thorough clinical examination remains crucial for patients with shoulder complaints [29].

Physical examination and ultrasonography have variable inter- and intraobserver reliability, particularly ultrasonography, which requires a lengthy learning curve. A meta-analysis has shown that musculoskeletal radiologists have higher diagnostic accuracy compared to general radiologists or ultrasonographers. Interestingly, orthopedic surgeons with a shoulder specialization can achieve results nearly comparable to those of musculoskeletal radiologists, reflecting their understanding of shoulder anatomy and its abnormalities [27]. Another limitation is that this study focused on a specific structure, the supraspinatus tendon. Given the complexity of anatomical structures and their functional interrelationships, accuracy may improve with comprehensive examination of both physical and ultrasonographic imaging across the entire rotator cuff.

This study demonstrates a very high level of agreement between ultrasonographic imaging and MRI. This is supported by previous research, which concluded that both MRI and ultrasonography can be used equivalently for detecting full-thickness rotator cuff tears, although ultrasonography is more cost-effective. In addition to diagnostic value, several factors must be considered for diagnostic testing and its clinical implications, such as safety, cost, availability, and the impact of the results. Patients with pacemakers, claustrophobia, or other contraindications for MRI would benefit from having ultrasonography as an alternative imaging option [29]. Incorporating diagnostic ultrasonography as a routine follow-up after physical examination for patients with shoulder pain can provide a better clinical understanding of the underlying shoulder issues, allowing orthopedic surgeons to perform targeted examinations based on clinical findings [28]. Although the increased use of ultrasonography by orthopedic surgeons, sports medicine doctors, and physiotherapists calls for further learning and training to enhance skills in performing and interpreting ultrasonography results. Such training also improves therapeutic efficiency through direct interventions like injections, which have been proven more effective when guided by ultrasonographic imaging [11].

However, ultrasonography is only recommended for screening rotator cuff cases. Information regarding tear size and location, fat infiltration, degree of atrophy, and tendon retraction still requires MRI imaging, although some aspects can be assessed with ultrasonography. Additionally, shoulder pain due to instability, cartilage defects, or labral tears still needs evaluation, with MRI playing an irreplaceable role in supporting diagnosis [29].