The study, published in BMC Medical Imaging by a multidisciplinary group from Xinjiang Medical University and its Sixth Affiliated Hospital, introduces a model called WEM-Mamba, short for Wavelet-Enhanced MedMamba. The work sits at the intersection of two of the hottest trends in medical deep learning: state space models, the efficient new architecture family that includes Mamba, and frequency domain analysis, which extracts information about textures and edges that spatial-domain models often overlook. According to the authors, most existing approaches, whether convolutional neural networks, vision transformers, or earlier Mamba variants, concentrate almost exclusively on modeling features in the spatial domain, leaving the rich frequency content of medical images underexploited. For a task like rotator cuff classification, where subtle textural signatures in the radiograph may hint at tendon damage, that omission could mean losing exactly the evidence that matters most.

To understand why frequency matters, it helps to think about what an X-ray actually encodes. Radiographs are dense grids of intensity values, and while clinicians read them as pictures of bones and soft tissue shadows, mathematically they can be decomposed into components that vary at different spatial scales and orientations. Fine, rapidly changing patterns correspond to high-frequency content such as edges and fine texture; smooth, slowly varying patterns correspond to low-frequency content such as broad anatomical structures. The Haar wavelet transform, a classical mathematical tool that dates back decades, provides a principled way to split an image into these frequency bands while keeping the spatial location of each feature intact. WEM-Mamba exploits this by embedding a Haar wavelet decomposition step directly inside its core processing module, so the network can examine the image simultaneously through spatial and frequency lenses.

Technically, the framework builds on MedMamba, a medical imaging variant of the Mamba architecture. Mamba and its relatives are built on state space models, which process sequences, including sequences of image patches, using recurrent-style computations that scale linearly with input length rather than quadratically as attention mechanisms in transformers do. That efficiency makes state space models attractive for medical imaging, where high-resolution images and limited training data are the norm. The researchers’ key innovation is the Wavelet-Enhanced SS-Conv-SSM module, or WESS. In this module, an input feature map is decomposed by the Haar wavelet transform into multiple frequency sub-bands. These sub-bands are processed and then synergistically fused with the global context modeling that the state space model provides, before an inverse wavelet transform can reconstruct the enhanced representation. The result is a hybrid design in which convolutional operations capture local detail, the state space model captures long-range dependencies across the image, and the wavelet pathway injects multi-band frequency features that neither of the other two components would extract on their own.

The team evaluated WEM-Mamba on a dataset of shoulder X-ray images collected retrospectively from routine clinical examinations at the Sixth Affiliated Hospital of Xinjiang Medical University, with ethics approval and anonymized patient data. They benchmarked their model against fifteen representative competitors spanning the major architecture families: convolutional neural networks, transformers, multi-layer perceptrons, and Mamba variants. WEM-Mamba came out on top, achieving an accuracy of 0.8950, an F1-score of 0.9309, a precision of 0.9078, a recall of 0.9552, and an area under the curve of 0.9116. Notably, the recall figure, which measures how many true tears the model successfully flags, is the highest of its headline metrics, a property that matters greatly in screening contexts where missed diagnoses carry the greatest cost. Equally important for real-world deployment, the model is compact: it contains 14.92 million parameters and requires only 2.04 gigafloating-point operations per inference, a footprint modest enough to run on ordinary clinical hardware rather than specialized computing clusters.

The performance-to-cost ratio is one of the study’s most compelling aspects. Deep learning models in medical imaging have often chased accuracy at the expense of size and speed, producing architectures too heavy for the very hospitals that need them most. By combining the linear-scaling efficiency of state space models with the lightweight Haar wavelet transform, the authors demonstrate that careful architectural design, rather than brute-force scale, can deliver strong classification performance on modest computational budgets. For clinics in underserved regions, where shoulder X-rays are already the first-line examination but specialist radiology expertise may be scarce, a compact model that augments the initial radiograph with a tear probability estimate could meaningfully accelerate referrals for MRI and definitive treatment.

The research also carries a broader message for the machine learning community. Wavelet-based feature extraction has a long pedigree in signal processing and image compression, but it has been somewhat eclipsed in the deep learning era by purely learned representations. The success of WEM-Mamba suggests that classical mathematical transforms and modern sequence models are not rivals but complements. The wavelet decomposition imposes a physically meaningful inductive bias, telling the network explicitly where the fine textures and edges live, while the state space model supplies the flexible global reasoning that fixed transforms cannot. In domains where training data are limited, as is nearly always the case in medicine, building such structural knowledge into the architecture may prove more data-efficient than expecting a generic network to rediscover it from scratch.

Caveats remain, and the authors are candid about them. The study was conducted on a single-center dataset, and all conclusions are explicitly framed as valid under the data and experimental settings of this particular investigation. The researchers state that further validation on larger-scale, multi-center, and external datasets is required before the approach can be considered robust, and that the clinical application value of the framework needs continued assessment in subsequent research. Single-center datasets can encode site-specific scanner settings, patient demographics, and imaging protocols, all of which can inflate apparent performance when a model is tested on data from the same source. Multi-center trials with external test sets are the standard next step for any diagnostic AI aspiring to clinical use, and this work is no exception.

Nevertheless, the study represents a thoughtful and timely contribution to computer-aided diagnosis of musculoskeletal injuries. It addresses a genuine clinical bottleneck, the gap between cheap first-line X-rays and definitive but costly MRI, with a method that is technically novel, computationally lean, and empirically competitive. The work was supported by a range of Chinese research funding programs, including the National Natural Science Foundation of China and multiple Xinjiang regional initiatives, and the article is published open access, making the full technical details available to researchers worldwide. If the promised multi-center validation bears out the single-center results, wavelet-enhanced state space models like WEM-Mamba could become a standard tool in the radiology toolbox, helping ensure that a patient’s first, cheapest image is not the last word on their shoulder pain.

Subject of Research: A wavelet-enhanced Mamba-based deep learning framework for classifying rotator cuff tears in shoulder X-ray images

Article Title: WEM-Mamba: a Wavelet-Enhanced MedMamba framework for rotator cuff tear classification in X-ray images

Article References: Wang, Z., Liu, Z., Li, J., Aboduaini, Y., Abulaiti, A., Li, S., Yakufu, M., Halike, A., Xv, L., & Li, G. (2026). WEM-Mamba: a Wavelet-Enhanced MedMamba framework for rotator cuff tear classification in X-ray images. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02740-2

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02740-2

Keywords: rotator cuff tear, deep learning, Mamba, state space model, wavelet transform, frequency domain features, X-ray imaging, medical image classification, computer-aided diagnosis, shoulder pain, MedMamba, BMC Medical Imaging

The research, published in BMC Medical Imaging, was led by Kyu-Chong Lee and Kyung-Sik Ahn of Korea University Anam Hospital in Seoul, together with colleagues from several Korean medical centers. The team set out to answer a deceptively simple question: when a shoulder MRI shows abnormalities of the joint capsule, what do those abnormalities actually mean? Adhesive capsulitis has long been associated with characteristic changes in two specific regions of the shoulder joint, the rotator interval and the axillary pouch. The rotator interval is a triangular space between the supraspinatus and subscapularis tendons through which the long head of the biceps tendon passes, while the axillary pouch is the redundant fold of capsule at the lower part of the joint that allows the arm to lift overhead. In classic frozen shoulder, both regions typically show thickening, edema, and inflammation.

To investigate, the researchers conducted a retrospective analysis of 483 adults who underwent shoulder MRI between January and December 2020. Two musculoskeletal radiologists independently reviewed each scan, assessing abnormalities in the rotator interval and the axillary pouch, as well as tears of four key tendons: the supraspinatus, the infraspinatus, the subscapularis, and the long head of the biceps tendon. The team defined radiologic capsular abnormality, abbreviated RCA, as the presence of abnormalities in both the rotator interval and the axillary pouch simultaneously. They then used univariable and multivariable logistic regression analyses to identify which clinical and imaging factors were associated with RCA, and performed subgroup analyses comparing four patient groups: those with no capsular abnormality, those with isolated axillary pouch abnormality, those with isolated rotator interval abnormality, and those with full RCA, adjusting for age where appropriate.

The headline finding was striking. Of the 483 patients studied, 234, or 48.45 percent, had radiologic capsular abnormality. In other words, nearly one in every two shoulders scanned showed the capsular changes that many radiologists and clinicians associate with frozen shoulder. Yet clinically diagnosed adhesive capsulitis was present in only 45 patients, or 9.32 percent of the cohort. Of those 45 patients, 43 did indeed have RCA, confirming that the imaging finding is sensitive for the disease. But the reverse is emphatically not true: the overwhelming majority of patients with RCA, more than four out of five, did not have a clinical diagnosis of adhesive capsulitis. The imaging finding, in isolation, is simply not specific enough to serve as a diagnosis.

The regression analyses revealed a web of associations that complicate the picture further. Older age was independently associated with higher odds of RCA, with each additional year of age raising the odds by roughly five percent, an odds ratio of 1.05 with a 95 percent confidence interval of 1.03 to 1.07 and a p-value below 0.001. This makes biological sense, since the joint capsule, like every other soft tissue in the body, undergoes degenerative change over time. A capsule that looks thickened or edematous in a 65-year-old may simply reflect the accumulated wear of decades rather than an active inflammatory disorder. Any diagnostic framework that treats capsular abnormality as synonymous with frozen shoulder must therefore account for the age of the patient, or risk labeling normal age-related change as pathology.

Tendon tears added another layer of complexity. Tears of the subscapularis tendon, one of the four muscles that make up the rotator cuff, were positively associated with RCA. Patients with Lafosse type I subscapularis tears, the mildest grade, had 1.82 times the odds of RCA compared with those without such tears, while those with type II or higher tears had 2.15 times the odds, both statistically significant results. The long head of the biceps tendon, which runs directly through the rotator interval, also emerged as relevant: in subgroup analyses, isolated rotator interval abnormality was associated with a higher prevalence of both subscapularis and biceps tendon tears than no capsular abnormality, even after adjusting for age. This suggests that inflammation or degeneration in the rotator interval often reflects pathology of the tendons that traverse it, rather than a primary capsular disease.

Perhaps most intriguingly, tears of the supraspinatus and infraspinatus tendons, the two most commonly torn rotator cuff tendons, showed the opposite relationship: they were negatively associated with RCA. Patients with these tears were less likely to show capsular abnormalities than patients without them. The authors interpret this finding as consistent with the traditional concept of adhesive capsulitis as a diagnosis of exclusion, a condition diagnosed only after other shoulder pathologies have been ruled out. In practical terms, a shoulder dominated by a large supraspinatus or infraspinatus tear appears less likely to harbor the capsular changes of frozen shoulder, perhaps because the two conditions represent distinct biological processes that do not readily coexist, or because the mechanical picture of a cuff tear overshadows and excludes the capsular one.

The clinical implications of the study are considerable. If nearly half of shoulder MRI scans show capsular abnormalities but only about nine percent of patients have frozen shoulder, then any radiology report that flags capsular abnormality without careful qualification invites overdiagnosis. A patient told that their MRI shows findings consistent with adhesive capsulitis may undergo months of stretching programs, corticosteroid injections, or even capsular distension procedures that address the wrong problem, while the true culprit, perhaps a subscapularis tear or biceps tendon pathology, goes untreated. The authors argue that the term radiologic capsular abnormality itself should be used in reports, precisely because it describes an imaging observation without asserting a clinical diagnosis, and that particular caution is warranted in older patients and in those with subscapularis or biceps tendon pathology.

The study does have limitations inherent to its design. It was retrospective, drawing on patients who had already been referred for shoulder MRI at a single Korean institution, which may limit generalizability to other populations. The clinical diagnosis of adhesive capsulitis was extracted from existing records rather than standardized prospectively, and the researchers note that the article was published initially as an early citable version subject to further edits. Still, the size of the cohort, the independent double reading by two musculoskeletal radiologists, the multivariable statistical framework, and the age-adjusted subgroup analyses give the findings substantial weight. The work was supported by a grant from the National Research Foundation of Korea, and the authors declared no competing relationships with companies whose products might relate to the subject matter.

For patients and clinicians alike, the message is ultimately one of humility about what imaging can and cannot tell us. MRI is a magnificent tool for revealing the anatomy of the shoulder in exquisite detail, but anatomy is not destiny, and a thickened capsule on a scan is a finding, not a verdict. Frozen shoulder remains a clinical diagnosis, rooted in a patient’s history of progressive pain and profound restriction of both active and passive motion, and it should be confirmed by examination rather than inferred from a radiology report. As this study makes clear, the capsule that lights up on MRI may be telling a story about age, about a torn subscapularis tendon, or about an inflamed biceps pulley, and only careful clinical correlation can determine which story is true. In an era when imaging findings increasingly drive diagnostic decisions, the humble conclusion that a scan must never replace the physical examination has rarely been more vividly illustrated.

Subject of Research: Radiologic capsular abnormalities on shoulder MRI and their association with rotator cuff and biceps tendon tears in the diagnosis of adhesive capsulitis

Article Title: Radiologic capsular abnormalities on shoulder MRI: associations with rotator cuff and biceps tendon tears and implications for the diagnosis of adhesive capsulitis

Article References: Radiologic capsular abnormalities on shoulder MRI: associations with rotator cuff and biceps tendon tears and implications for the diagnosis of adhesive capsulitis. (n.d.). https://doi.org/10.1186/s12880-026-02895-y

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02895-y

Keywords: shoulder MRI, adhesive capsulitis, frozen shoulder, rotator cuff tear, subscapularis tendon, biceps tendon, rotator interval, axillary pouch, radiologic capsular abnormality, musculoskeletal radiology, overdiagnosis, BMC Medical Imaging