Abstract
Background:
Small-bore needle arthroscopy (SNA) has emerged as a minimally invasive diagnostic tool in outpatient orthopaedics, allowing direct intra-articular visualization and serving as a potential alternative to magnetic resonance imaging (MRI).
Purpose:
To compare the diagnostic accuracy of a 2.0-mm needle arthroscope with MRI and determine which more accurately predicts appropriate surgical intervention in patients with glenohumeral pathology.
Study Design:
Cohort study (Diagnosis); Level of evidence, 2.
Methods:
In this institutional review board–approved, prospective, single-center, blinded study, 30 consecutive patients scheduled for shoulder arthroscopy between May and July 2020 were enrolled. All patients underwent preoperative MRI followed by 2 diagnostic arthroscopies using a 2.0-mm SNA and a standard 4.0-mm arthroscope. Deidentified arthroscopic recordings and MRIs were reviewed by 2 blinded, fellowship-trained orthopaedic sports medicine surgeons. The standard 4.0-mm arthroscope served as the gold standard. Diagnostic agreement and appropriate treatment recommendations—reflected by Current Procedural Terminology (CPT) codes—were assessed for both SNA and MRI.
Results:
The 2.0-mm diameter arthroscope demonstrated superior diagnostic accuracy compared with MRI for most intra-articular shoulder pathologies, with the exception of humeral head chondromalacia. SNA showed higher sensitivity, specificity, positive predictive value (PPV), and negative predictive value in evaluating the long head of the biceps (LHB), superior labrum, infraspinatus, subscapularis, and glenoid cartilage. The strongest diagnostic agreement with the standard 4.0-mm arthroscope was observed for LHB lesions (κ = 0.73), subscapularis tears (κ = 0.66), and articular-sided supraspinatus tears (κ = 0.62). For CPT 29827 (arthroscopic rotator cuff repair), SNA yielded a higher PPV (100% vs 83.3%).
Conclusion:
Small-bore needle arthroscopy using a 2.0-mm platform demonstrated high diagnostic performance in identifying glenohumeral pathology, particularly involving the LHB, subscapularis, and articular-sided supraspinatus. SNA outperformed MRI in predicting the appropriate surgical procedure and may serve as a reliable diagnostic alternative in the outpatient setting.
Keywords
Before undergoing shoulder surgery, patients often ask their surgeon about the length of time their arm is immobilized in a sling, the length of time during which arm motion is restricted, and most importantly the length of overall recovery and return to preinjury function. The answer to these questions is predicated on an accurate preoperative assessment of shoulder pathology. For example, partial-thickness versus full-thickness tears of the rotator cuff tendon and pathology of the long head of the biceps (LHB) can significantly alter recovery expectations based on the extent of pathology (eg, debridement vs repair, or tenodesis). The treating surgeon relies on a comprehensive physical examination and often utilizes advanced imaging such as magnetic resonance imaging (MRI) to ascertain an accurate diagnosis and predict surgical recovery time.
Presently, the standard method of evaluating soft tissue pathology in the shoulder is with MRI. However, the limited diagnostic accuracy of noncontrast MRI in differentiating partial- from full-thickness tears of the supraspinatus as well as the subscapularis tendons has been described.3,6,11 The value of noncontrast MRI in diagnosing intra-articular shoulder pathology such as glenohumeral cartilage lesions, superior labral tears, or pathology of the LHB tendon has also come under question.2,6,10,12
The use of MRI may also be limited by cost, contraindication because of metal implants or pacemakers, and delay of definitive treatment due to insurance authorization and imaging center accessibility.
Over the past decade, small-bore needle arthroscopy (SNA) has become more widespread as the technology, image resolution, and unit costs have improved. Voigt et al 13 estimated that the cost savings associated with SNA for shoulder pathology alone could save a mean of $59 million annually compared with MRI.
Previous studies have validated the accuracy of SNA compared with standard arthroscopy and MRI in the knee and shoulder.4,5,8,14,15 We believe that this is the first study to compare SNA with MRI in accurately predicting which specific procedure is most appropriate for patients undergoing shoulder surgery. Our study had 2 main objectives. The first was to determine the diagnostic accuracy of SNA as compared with standard MRI for intra-articular glenohumeral pathology. The second objective was to determine which modality would better predict the appropriate surgical procedure (Current Procedural Terminology [CPT] code) for each operative shoulder.
Methods
After obtaining institutional review board approval, a prospective, blinded, clinical trial was performed on 30 consecutive patients who were scheduled to undergo operative arthroscopy for intra-articular shoulder pathology. An a priori 2-tailed power analysis was performed using an open-source statistical power application (G*Power) and determined that for a 10% difference between 2 means in correctly diagnosing a pathology (90% vs 80%), a standard deviation of 10, an alpha of .05, and a power of 10, there needed to be 17 patients per group. To mitigate the effect of potential measurement error and data loss, 30 patients were enrolled, each undergoing MRI, SNA, and 4.0-mm arthroscopy. The study period was May 2020 to July 2020. Inclusion criteria were patients >18 years old with suspected pathology of a single rotator cuff muscle or tendon, biceps tendinopathy, or other intra-articular pathology, as identified through clinical history of physical examination. Exclusion criteria were the presence of a 2-tendon rotator cuff tear on MRI, moderate or severe glenohumeral arthritis, patients <18 years old, or the diagnosis of infection, inflammatory arthropathy, adhesive capsulitis, or glenohumeral instability. The decision to perform rotator cuff repair was based on tear characteristics, tissue quality, and clinical relevance. Full-thickness, symptomatic, and mobile tears to the native footprint were considered repairable. High-grade, partial-thickness tears involving >50% of tendon thickness were also repaired. Each patient underwent a comprehensive history and physical examination as well as a shoulder noncontrast MRI (1.5 T) and was then determined to meet the criteria for arthroscopic surgical treatment.
Surgical Technique
After general anesthesia was induced and an interscalene nerve block performed by the anesthesiologist, an examination under anesthesia of both shoulders was performed by the senior author (T.S.D.) and the patient was positioned in the lateral decubitus position with the arm suspended in 30° of abduction and forward flexion using 4.5 kilograms of distal traction. The shoulder was then prepared and draped in the usual sterile fashion and the standard "time out" protocol was followed by the surgical team. A standard posterior viewing portal for both the SNA and the standard arthroscopy was placed 2 cm inferior and medial to the posterolateral acromial edge. An intra-articular diagnostic arthroscopy was performed and recorded using the 2.0-mm SNA device first. The following diagnostic progression was performed: LHB along with the biceps anchor and superior labral complex, LHB from origin to its exit from the joint at the bicipital groove, undersurface of the supraspinatus and infraspinatus, posterior humeral head (bare area), axillary pouch and overall inferior capsular volume, posterior labrum (from inferior to superior), subscapularis tendon, anterior capsule and glenohumeral ligaments (superior, middle, then inferior), and articular cartilage of the glenoid and humerus. The recording was then stopped and the 2.0-mm arthroscope removed and replaced with a standard 4.0-mm arthroscope. The same progressive diagnostic examination was performed and recorded using the 4.0-mm arthroscope. Both independent recordings were saved to a memory device and the surgery was then completed as indicated. The examination under anesthesia, 2.0-mm recording, 4.0-mm recording, and surgical procedure were all completed by a single surgeon (T.S.D.). After each procedure, surgical dictation and coding using the American Medical Association's standard CPT code set was also completed by the same surgeon. Each patient's MRIs, 2.0-mm video recording, and 4.0-mm video recording were then deidentified and electronically saved for later review. All patients were followed for ≥3 months after the procedure to evaluate for any complications.
Equipment
The SNA system used in this study was approved by the US Food and Drug Administration for diagnostic evaluation of intra-articular pathology and is an alternative to a rod-lensed arthroscope (Nanoscope; Arthrex Inc.). The technology is unique in that it utilizes a disposable, chip-on-tip 2.0-mm arthroscope that connects to a tablet-like, medical-grade control unit and delivers a 400 × 400–pixel image (Figure 1).
NanoScope (Arthrex) is a single-use, chip-on-tip 2.0 mm–diameter scope with integrated digital imaging that delivers a 400 × 400–pixel high-resolution visualization of intra-articular structures. The arthroscope connects to a portable touchscreen monitor that serves as an interface for image capture, video recording, and system control.
Data Analysis
Two fellowship-trained orthopaedic sports surgeons (J.A.G., A.M.F.) who were uninvolved in the care of the study participants reviewed all the deidentified MRIs, along with the 2.0-mm and the 4.0-mm arthroscope video recordings in a blinded and randomized fashion, and completed a diagnostic data collection form after reviewing each of the diagnostic tools. Thirty days later, the same surgeons reviewed the same data set to measure interobserver reliability of the collected information. With the 4.0-mm data used as the control, the NSP data were compared with the MRI data to determine which was more accurate in diagnosing intra-articular shoulder pathology. In addition, the authors were interested in determining if the information collected from the SNA recording was more accurate at predicting the best surgical procedure for the patient when compared with the information collected from the MRIs. To distinguish between the presence of pathology and pathology warranting surgical intervention, the reviewing surgeon listed the recommended CPT codes based solely on the stand-alone information available from each SNA and MRI video. Using standard 4.0-mm arthroscopy as the reference, we calculated sensitivity and positive predictive value (PPV) of both SNA and MRI for predicting whether an arthroscopic rotator cuff repair (CPT 29827) was ultimately performed.
Statistical Analysis
Statistical analysis was performed by an independent statistician using SAS Version 9.4 (SAS Institute Inc). Evaluation of the data was made using several methods: Cohen kappa coefficient, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).
Results
A total of 30 patients were included in this study. All patients underwent preoperative MRI, diagnostic 2.0-mm SNA, and standard 4.0-mm diagnostic arthroscopy (Figure 2). Independent evaluation of imaging and arthroscopic video was performed on 2 separate occasions, 30 days apart, by 2 fellowship-trained sports medicine surgeons blinded to the surgical procedures performed and uninvolved with patient care.
(A) Coronal T2-weighted magnetic resonance imaging; (B) 4-mm arthroscopic view; and (C) 2.0-mm needle arthroscopic view of the same patient demonstrating an articular-sided supraspinatus tear in the left shoulder.
Diagnostic Performance of 2.0-mm SNA Versus MRI for Intra-articular Shoulder Pathology
Across all intra-articular pathologies except humeral head chondromalacia, SNA demonstrated a higher percentage agreement and PPV compared with MRI. Specifically, SNA yielded higher sensitivity, specificity, PPV, and NPV when evaluating the LHB, superior labrum, infraspinatus, subscapularis, and glenoid cartilage (Table 1). SNA showed the highest diagnostic agreement for pathology of the intra-articular portion of the LHB, with a kappa value of 0.73. In comparison, MRI demonstrated a lower agreement (κ = 0.52). Similar trends were observed for superior labral pathology, where SNA outperformed MRI in diagnostic consistency (Table 1).
Diagnostic Agreement and Performance of Anatomy Evaluation as Compared With 4.0-mm Arthroscopy a
Significant findings (P < .05) are shown in bold. LHB, long head of the biceps; MRI, magnetic resonance imaging; NPV, negative predictive value; 2.0-mm needle arthroscopy; PPV, positive predictive value. Diagnostic performance comparison between the 2.0-mm arthoscopy and MRI across intra-articular shoulder pathologies. The 2.0-mm arthroscopy demonstrated higher percentage agreement and PPV than MRI for all evaluated structures except humeral head chondromalacia. Sensitivity, specificity, PPV, and NPV were greater with the 2.0-mm arthroscopy for the LHB, superior labrum, infraspinatus, subscapularis, and glenoid cartilage.
Diagnostic Performance and Surgical Agreement of SNA Versus MRI for Intra-articular Shoulder Pathologies
The diagnostic performance and treatment concordance of the 2.0-mm SNA and MRI were evaluated across a spectrum of intra-articular shoulder pathologies, using standard 4.0-mm arthroscopy as the reference standard. Overall, NSP outperformed MRI in both diagnostic accuracy and alignment with surgical decision-making. In particular, SNA demonstrated significantly higher sensitivity, specificity, and interrater agreement (kappa values) for the detection of supraspinatus tears, subscapularis pathology, and abnormalities of the LHB (Table 2). These findings were statistically significant and indicated a strong correlation between SNA findings and final surgical management. In contrast, while MRI often achieved high sensitivity, it was frequently limited by lower specificity and inconsistent agreement with surgical treatment, most notably in the evaluation of labral and infraspinatus pathology (Table 2). These results suggest that in the diagnostic workup of shoulder pathology, SNA may offer a more accurate and reliable tool to inform appropriate operative intervention compared with preoperative MRI.
Diagnostic Performance and Agreement of Treatment Responses as Compared With 4.0-mm Arthroscopy a
Significant findings (P < .05) are shown in bold. LHB, long head biceps; MRI, magnetic resonance imaging; NPV, negative predictive value; PPV, positive predictive value. Diagnostic accuracy and treatment agreement of 2.0-mm arthroscopy and MRI across intra-articular shoulder pathologies, referenced to standard 4.0-mm arthroscopy.
Concordance of SNA and MRI With Operative CPT Coding
When assessing the ability to predict appropriate CPT codes, SNA demonstrated greater concordance with operative findings compared with MRI (Table 3). The kappa statistic was significantly higher for the 2.0-mm arthroscope compared with MRI (κ = 0.87 vs 0.56). NSP demonstrated substantial agreement with the gold standard 4.0-mm arthroscopy (κ = 0.87; 95% CI, 0.71-1.00; P < .0001), whereas MRI showed only moderate agreement (κ = 0.56; 95% CI, 0.31-0.81; P = .0003) (Table 3). These findings support NSP reliability in guiding procedural decision-making across a range of shoulder pathologies.
Agreement as Compared With 4.0-mm Scope for CPT Codes a
CPT, Current Procedural Terminology; MRI, magnetic resonance imaging.
Diagnostic Accuracy in Predicting Rotator Cuff Repair (CPT 29827)
Cohen kappa coefficient was used to assess inter- and intrarater reliability, with values closer to 1 indicating stronger agreement and values closer to 0 indicating weaker agreement; values <0.4 are generally considered poor, 0.4 to 0.6 moderate, and >0.6 substantial agreement. Using this metric, SNA demonstrated superior diagnostic accuracy for articular-sided rotator cuff tears compared with MRI (70.0% [95% CI, 34.8 to 93.3] vs 30.0% [95% CI, 6.7 to 65.3]) with stronger agreement with surgical findings (κ = 0.62 [95% CI, 0.31 to 0.92] vs κ = 0.30 [95% CI, –0.04 to 0.63]; P = .001) (Table 1). Subscapularis and supraspinatus tears showed moderate-to-substantial agreement with operative findings, with kappa values of 0.66 and 0.62, respectively (Table 1). In contrast, MRI yielded lower agreement, with kappa values of 0.35 for subscapularis and 0.29 for supraspinatus tears (Table 1). When evaluating rotator cuff tears requiring surgical repair (CPT 29827), the SNA demonstrated comparable sensitivity with MRI (93.3% vs 93.8%) but showed a notably higher PPV (100.0% vs 83.3%) (Table 4). This suggests that while both modalities are similarly effective at detecting rotator cuff pathology, SNA more reliably confirms pathology that warrants operative intervention, potentially reducing false positives and improving surgical planning.
Sensitivity and PPV as Compared With 4.0-mm Scope for CPT Code 29827 a
CPT, Current Procedural Terminology; MRI, magnetic resonance imaging; PPV, positive predictive value; RCR, rotator cuff repair. In predicting the need for arthroscopic RCR (CPT 29827), 2.0-mm arthoscopy and MRI demonstrated similar sensitivity (93.3% and 93.8%, respectively); however, 2.0-mm arthoscopy showed superior PPV (100.0% vs 83.3%), indicating greater reliability in preoperative surgical planning.
Complications
No complications related to the use of SNA were observed during the study period. Although injury to the intra-articular anatomy of the shoulder from a preoperative SNA is theoretically possible, there were no cases in which there was clear indication that the pathology noted at surgery was either acute or related to the trauma of a recent, needle puncture–type injury. All patients were monitored for ≥3 months after the diagnostic procedures, with no reported cases of infection, neurovascular injury, synovitis, persistent joint effusion, or need for additional intervention attributable to the SNA procedure. No patients reported increased postoperative pain or functional limitation that could be associated with the use of the NSP device. These findings support the short-term safety profile of SNA when used in the evaluation of intra-articular glenohumeral pathology.
Discussion
This study demonstrated that SNA had a higher sensitivity, specificity, PPV, and NPV when evaluating the LHB, superior labrum, infraspinatus, subscapularis, and glenoid cartilage. These results are similar to several other studies that evaluated the efficacy of SNA.12,14 A recent study by Wagner et al 14 employing a similar design reported comparable findings. The authors demonstrated that relative to standard 4.0-mm shoulder arthroscopy, SNA exhibited diagnostic accuracy comparable with that of MRI in identifying articular cartilage, labral, rotator cuff, and biceps pathology. When evaluating all intra-articular shoulder pathology collectively, SNA demonstrated higher sensitivity and PPV, but lower specificity and NPV, compared with MRI. A prospective, blinded, multicenter study by Gill et al 5 evaluated the use of SNA in the knee. The authors concluded that in-office SNA was statistically equivalent to operative diagnostic arthroscopy in the evaluation of intra-articular, nonligamentous knee pathology. While this investigation focused specifically on the knee joint, the findings support the utility of SNA as a safe and effective modality for the diagnosis of intra-articular joint pathology.
The ability to identify intra-articular pathology in patients with shoulder pain can be challenging. MRI is commonly used preoperatively to evaluate patients who have shoulder pain and are contemplating surgery. However, multiple studies have shown the difficulties associated with this type of imaging modality.2,3,6,7,10-12 Brockmeyer et al 3 looked at 334 consecutive shoulder arthroscopies for rotator cuff pathologies and compared the arthroscopic recordings with preoperative MRI. They found that the sensitivity for MRI to identify partial-thickness tears was 51.6%, specificity 77.2%, PPV 41.3%, and NPV 83.7% and concluded that the diagnostic accuracy of MRI in detecting partial-thickness tears is limited. Regarding the LHB and labral injuries, diagnosis with MRI has also been challenging. Malavolta et al 7 looked at 90 shoulders comparing MRI to intraoperative arthroscopy. They found that the sensitivity and specificity for identifying LHB tears weas 67% and 98%, respectively. Ruess et al 10 evaluated the diagnostic accuracy of preoperative MRI in 83 patients with arthroscopically confirmed type 2 superior labrum anterior to posterior (SLAP) lesions. The cohort included 37 noncontrast MRIs and 46 magnetic resonance arthrograms. Community radiologists correctly identified the SLAP lesion in only 51% of cases. Sensitivity improved from 27% with noncontrast MRI to 70% with magnetic resonance arthrography. Furthermore, diagnostic accuracy increased when studies were interpreted by fellowship-trained musculoskeletal radiologists.
The economic implications of evaluating shoulder pathology using SNA are substantial. Voigt et al 13 conducted a cost analysis comparing in-office SNA with MRI and reported that substituting SNA for MRI in the evaluation of 540,000 knee procedures could yield an estimated annual savings of $151 million. Similarly, for 165,000 shoulder procedures, the projected annual savings was $59 million. McMillan et al 9 further analyzed per-case cost savings and found that replacing noncontrast MRI with SNA resulted in savings of $418 in independent facilities and $961.08 in hospital-based settings for the knee. For the shoulder, the corresponding savings were $554.62 and $1097.62, respectively. A separate cost-effectiveness study comparing SNA with MRI for the knee concluded that SNA was less costly and yielded equivalent or improved clinical outcomes. 1 Although our study did not focus on a cost analysis perspective, McMillan et al 9 reported that SNA is reimbursable, with mean in-office reimbursement of $628.92 (range, $340.00-$1391.00) for diagnostic knee procedures and $492.38 (range, $471.00-$593.00) for diagnostic shoulder procedures.
The findings of our study demonstrated that SNA can be more accurate for diagnosing a variety of intra-articular shoulder pathology compared with MRI. More specifically, our findings demonstrated that SNA was able to more accurately predict which rotator cuff tears would require surgical repair (CPT 29827). It is not uncommon that patients will be informed on the possibility that certain procedures may or may not be performed and the decision to perform a certain procedure relies on an intraoperative diagnostic arthroscopy. The use of a SNA can give the surgeon and patient accurate, real-time, cooperative decision-making capability and manage patient rehabilitation expectations before incision. Collectively, these efforts will give patients a more complete understanding of their shoulder pathology and treatment options, which may correlate with an improved overall health care experience and satisfaction. 9
Limitations
Our study is not without limitations. First, all MRIs were performed with a 1.5-T magnet. The accuracy of the preoperative MRI may have been improved if a higher field strength magnet was used. Another limitation involves the applicability of our findings across different operative settings with distinct consideration regarding patient positioning and anesthesia. While both the SNA and subsequent surgical procedures in this study were performed in an anesthetized patient placed in the lateral decubitus position, we recognize that this represents a different setup from the in-office scenario, particularly in an awake patient in the beach-chair position. As such, our findings may not be directly reproducible in these alternative settings, and this difference should be acknowledged before extending our results to broader clinical recommendations. Future studies are warranted to determine whether these findings are translatable to procedures performed in the beach-chair position or other operative environments.
Conclusion
This study showed that SNA is a safe and effective method of diagnosing intra-articular glenohumeral pathology. Our study supports greater accuracy in identifying intra-articular shoulder pathology and predicting the appropriate CPT to address this pathology in anesthetized patients undergoing SNA in a lateral position as compared with awake patients undergoing an MRI. Although these diagnostic tools are not often interchangeable, we believe that our findings highlight the potential benefits of SNA as an alternative to MRI and encourage further research on the utility and benefits of in-office SNA. When diagnosing pathology within the shoulder, the SNA device had a higher PPV than MRI when assessing all anatomic pathology except for humeral head chondromalacia. In particular, SNA is highly accurate in diagnosing lesions of the intra-articular LHB, tears of the subscapularis, and articular-sided tears of the supraspinatus. When compared with MRI, the SNA was able to more accurately predict which rotator cuff tears would require surgical repair (CPT 29827). When considering all CPT codes for the shoulder in this series, the SNA was more accurate at predicting the appropriate surgical procedure than the MRI. Additionally, the potential cost savings may help reduce the overall economic impact of evaluating and treating shoulder pain. Further studies are needed, however, to evaluate the efficacy of this 2.0-mm arthroscopy of the shoulder in wide awake patients in an office setting.
Footnotes
Final revision submitted January 14, 2026; accepted March 3, 2026.
One or more of the authors has declared the following potential conflict of interest or source of funding: Support was provided by Arthrex Inc (grant No. US-19075). T.S.D. is a board/committee member for the AOSSM and the California Orthopedic Association and is a paid consultant, presenter, and speaker for Arthrex Inc.
Ethical approval for this study was obtained from Brany Institutional Review Board (File # 20-08-030-713).
Data Accessibility Statement
The data supporting the findings in this study are available from the corresponding author upon reasonable request and with Synergy Orthopedics.
