1.1 Rehabilitation challenges after rotator cuff repair: Limitations of traditional time‑driven protocols

Rehabilitation after rotator cuff repair must balance two conflicting needs: sufficient immobilization to protect the repair interface and promote tendon-to-bone healing, and early mobilization to prevent joint stiffness, muscle atrophy, and adhesion formation. Postoperative shoulder stiffness is a common complication, closely associated with insufficient rehabilitation training and excessive immobilization [5]. Meanwhile, the reported retear rate after rotator cuff repair ranges from 15% to 21%, and can reach as high as 94% in cases of massive tears [6]. A number of systematic reviews and meta-analyses have shown that although retears have a negative effect on functional recovery, the effect size is modest and does not reach the threshold for clinical significance. This means that some patients may still achieve acceptable functional outcomes despite a retear — a finding that suggests postoperative rehabilitation decisions should not be informed solely by the goal of “absolute prevention of retears.” This observation is supported by the evidence from Holtedahl et al. [7]. In addition, age, tear size, fatty infiltration, muscle atrophy, and the appropriateness of the postoperative rehabilitation protocol have all been identified as significant risk factors for retear [8]. Traditional “one-size-fits-all” protocols that rely on fixed time points neglect the substantial inter-individual variability in tear characteristics, tissue quality, and healing trajectory. There is a clear need to shift toward a precision rehabilitation model based on the patient’s individual tissue healing status.

1.2 Technical advantages of musculoskeletal ultrasound: non‑invasive, bedside, dynamic, and quantifiable assessment of tendon healing and muscle morphology

Musculoskeletal ultrasound is a non-invasive, real-time, and reproducible imaging modality that has recently demonstrated distinct advantages in the postoperative assessment of rotator cuff repair [9,10]. Advances in high-frequency transducers have enabled clearer visualization of fine tendon structures, panoramic imaging allows larger-scale structural assessment, and the introduction of elastography makes it possible to quantitatively evaluate the biomechanical properties of tendons and muscles [11]. Compared with magnetic resonance imaging, musculoskeletal ultrasound offers several practical benefits, including bedside applicability, dynamic imaging capability, and repeatable follow-up [12]. It also avoids artifacts caused by metal implants, making it particularly suitable for postoperative rehabilitation monitoring [13]. Musculoskeletal ultrasound can be used not only for the differential diagnosis of rotator cuff tears but also for assessing surgical outcomes and supporting rehabilitation monitoring at different postoperative stages [11]. In recent years, ultrasound elastography has also seen important progress in the preoperative evaluation of rotator cuff tears. Zhang et al. (2024) analyzed 106 patients who underwent arthroscopic rotator cuff repair and found that preoperative shear wave velocity (SWV) and elastic modulus were positively correlated with the Constant score at one year postoperatively. The combination of these two parameters predicted postoperative retear with an area under the curve (AUC) of 0.95 and a sensitivity of 91.70%, indicating that preoperative ultrasound parameters can effectively predict long-term functional outcomes and retear risk [14].

1.3 Research gap: existing reviews focus on diagnosis or treatment; none have systematically synthesized the evidence linking ultrasound monitoring with rehabilitation-related decisions

In recent years, the application of musculoskeletal ultrasound in rehabilitation after rotator cuff repair has gradually shifted from “structural assessment” to “functional monitoring.” Existing reviews have systematically summarized the normal sonographic appearances of the postoperative rotator cuff and common complications [15], and others have compared the use of ultrasound versus MRI in the postoperative setting [13]. In addition, clinical studies have confirmed that postoperative strain elastography scores are significantly correlated with muscle strength and function [16], that musculoskeletal ultrasound combined with shear wave elastography can assess tendon healing status after surgery for different tear types [17], and that changes in ultrasound parameters are positively correlated with shoulder function recovery [18]. However, most of these studies remain at the level of monitoring; they have not systematically summarized how ultrasound parameters can be directly translated into rehabilitation decisions (e.g., when to initiate active movement, how to adjust load intensity, or how to predict prognosis). In other words, although a large number of studies have used musculoskeletal ultrasound to monitor postoperative rehabilitation, a systematic review specifically focusing on how musculoskeletal ultrasound monitoring may inform rehabilitation decision-making after rotator cuff repair is still lacking.

1.4 Study objective

Accordingly, the aim of this systematic review is to systematically search and synthesize the available literature to answer the following core question: which ultrasound parameters, at what postoperative time points, and in what manner have been used to inform adjustments of rehabilitation protocols after rotator cuff repair? By systematically mapping the evidence linking ultrasound parameters to rehabilitation decisions, this review seeks to provide clinicians with evidence‑based support for ultrasound‑informed precision rehabilitation, to identify key gaps in the current evidence base, and to offer directions for future research.

2.1 Registered

The protocol for this systematic review was registered with the PROSPERO International Prospective Register of Systematic Reviews (registration number: CRD420261418165).

2.2 Search strategy

Searches were conducted in three electronic databases: PubMed, China National Knowledge Infrastructure (CNKI), and Wanfang Data Knowledge Service Platform. The search period covered January 2020 to April 2026, and the language was restricted to English and Chinese.

For PubMed, the following search string was used: (“rotator cuff repair” AND (ultrasound OR elastography) AND (rehabilitation OR decision)).

For the Chinese databases, the search strings were adapted accordingly: CNKI: SU=(’肩袖’ + ‘冈上肌’) * (’修复’ + ‘修补’ + ‘关节镜’) * (’超声’ + ‘肌骨超声’ + ‘弹性成像’) * (’康复’ + ‘术后’ + ‘功能恢复’ + ‘训练’); Wanfang Data: (主题:(肩袖) OR 主题:(冈上肌)) AND (主题:(修复) OR 主题:(修补) OR 主题:(关节镜)) AND (主题:(超声) OR 主题:(肌骨超声) OR 主题:(弹性成像)) AND (主题:(康复) OR 主题:(功能恢复) OR 主题:(术后) OR 主题:(训练)). The complete electronic search strategies are provided in S2 File.

2.3 Inclusion and exclusion criteria

The inclusion and exclusion criteria were defined according to the PICOS framework as follows.

2.4 Literature screening and data extraction

Literature screening was performed following the process recommended by the PRISMA 2020 guideline. Duplicate records were first removed using reference management software (NoteExpress). Two investigators independently screened the titles and abstracts against the inclusion criteria and excluded records that clearly did not meet the criteria. Full texts of the potentially eligible articles were then retrieved and independently assessed by the two investigators to make the final decision on inclusion.

Data extraction was also performed independently by the two investigators using a pre‑designed data extraction form. The extracted information included (1) basic study characteristics (first author, year of publication, study design); (2) participant characteristics (sample size, tear type and size, postoperative follow‑up time points); (3) musculoskeletal ultrasound parameters (ultrasound device, parameter types: cross‑sectional area, echo intensity, shear wave velocity, strain ratio, etc.); (4) rehabilitation decision‑related information; and (5) level of evidence. After extraction, the two investigators cross‑checked the data to ensure accuracy. Disagreements between reviewers were resolved through discussion, with a third reviewer consulted when necessary. All screening decisions were reached by consensus; however, formal inter-rater agreement statistics (e.g., Cohen's kappa) were not calculated. The characteristics of included studies are summarized in Table 1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Methodological characteristics of included studies. LOE, level of evidence, graded according to the Oxford Centre for Evidence-Based Medicine (CEBM) levels of evidence.

https://doi.org/10.1371/journal.pone.0357622.t001

2.5 Risk of bias assessment

The methodological quality (risk of bias) of the included studies was independently assessed by two reviewers. The assessment tool was selected according to the study design. For the single randomized controlled trial, the Cochrane Risk of Bias tool version 2 (RoB 2) was used, evaluating bias arising from the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selective reporting of results. For the nine cohort studies, the Risk of Bias in Non-randomized Studies of Interventions tool version 1 (ROBINS-I) was applied, assessing bias due to confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selective reporting of results. For the one case series, the JBI Critical Appraisal Checklist for Case Series was used. Each study was rated as having either “low risk of bias,” “moderate risk of bias,” “high risk of bias,” or “some concerns.” The two reviewers independently performed the assessment and then cross-checked their results. Disagreements were resolved through discussion, with a third reviewer consulted when necessary. For the single randomized controlled trial included in this review, the Cochrane Risk of Bias tool version 2 (RoB 2) was used; the trial (He et al., 2021) was judged to have low risk of bias across all domains. For the case series included in this review (Liu et al., 2022), the JBI Critical Appraisal Checklist for Case Series was used; the study was judged to have moderate risk of bias.

3.1 Study selection

The study selection process was reported following the PRISMA 2020 guidelines. The initial search yielded 278 records: 126 from PubMed, 52 from CNKI, and 100 from Wanfang Data. After removing duplicates using NoteExpress, 268 records remained. Screening by title and abstract excluded 221 records, leaving 47 articles for full‑text assessment. Based on the inclusion and exclusion criteria, 36 articles were excluded after full‑text review. Consequently, 11 studies were included in this systematic review. The screening process and its results are presented in a PRISMA flow diagram Figure 1.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA 2020 flow diagram of study selection process.

https://doi.org/10.1371/journal.pone.0357622.g001

3.2 Characteristics of included studies

The 11 included studies were published between January 2020 and April 2026. In terms of study design, there were six prospective cohort studies, three retrospective cohort studies, one case series, and one randomized controlled trial. Sample sizes ranged from 39 to 95 patients, with a total of 673 participants across the included studies. Postoperative assessment time points ranged from one week to 12 months, with one month, three months, and six months being the most common evaluation points.

In terms of ultrasound parameters, six studies focused on tendon parameters (mainly SWV), four studies examined muscle parameters (CSA and EI), and one study reported both types of parameters. Regarding the association with rehabilitation decisions, three studies explored the relationship between ultrasound parameters and the initiation or progression of active movement, seven investigated retear risk warning, four involved prognostic prediction of return to sport, seven addressed load intensity adjustment, and three covered rehabilitation bottleneck identification.

3.3 Summary of ultrasound parameters

The included studies primarily reported two categories of ultrasound parameters: muscle morphology parameters (e.g., cross‑sectional area [CSA] of the supraspinatus muscle, used to assess muscle atrophy) and tendon structural parameters (e.g., tendon thickness [TH], shear wave velocity [SWV], used to assess tendon healing quality). A detailed summary of ultrasound equipment and technical specifications is provided in S1 Table.

3.4 Summary of how ultrasound parameters inform rehabilitation decisions

3.4.2 Decision thresholds.

Quantifiable decision threshold information is currently limited. Shimizu et al. established age-stratified reference ranges for the strain ratio. At 0° of abduction, the range was 0.72 to 4.17 for patients younger than 50 years and 0.98 to 4.50 for those 50 years or older. A strain ratio exceeding the upper limit suggests abnormally increased passive stiffness of the tendon and calls for adjustment of the rehabilitation protocol [24]. Solari et al. reported that tendon SWV increased progressively after surgery, from 6.20 m/s at one week to 7.50 m/s at six months (P < 0.001). Based on this normal healing trajectory, the authors speculated that a SWV of around 7.50 m/s has been proposed as a potential reference value reflecting postoperative tendon recovery; however, this value has not been validated as a clinical decision threshold [20]. However, this speculation has not yet been validated by prospective studies, and whether 7.50 m/s can serve as a universal reference value remains to be confirmed by further research. Figure 2

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Key ultrasound parameters for rehabilitation decision‑making after rotator cuff repair.

MSK, musculoskeletal; RCT, randomized controlled trial; SWV, shear wave velocity; CSA, cross‑sectional area; EI, echo intensity. Created by the authors.

https://doi.org/10.1371/journal.pone.0357622.g002

3.5 Risk of bias assessment results

Among the 11 included studies, five were rated as having low risk of bias (He et al., 2021; Maloof et al., 2023; Zhu et al., 2024; Yan et al., 2025; Maloof et al., 2025), and six were rated as having moderate risk of bias (Itoigawa et al., 2020; Liu et al., 2022; Kim et al., 2023; Ye et al., 2024; Solari et al., 2024; Wang et al., 2025). No study was found to have high risk of bias. The main sources of moderate risk of bias included retrospective study design, inadequate adjustment for confounding factors, and failure to report inter-observer agreement for ultrasound measurements. Fig 3, Fig 4.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Risk of bias across ROBINS-I domains for the included cohort studies.

https://doi.org/10.1371/journal.pone.0357622.g003

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Risk of bias assessment of individual cohort studies using the ROBINS-I tool.

https://doi.org/10.1371/journal.pone.0357622.g004

4.1 Summary of main findings

This systematic review is the first to focus on the question of how musculoskeletal ultrasound monitoring can inform rehabilitation decision-making after rotator cuff repair. We systematically synthesized evidence from studies published between January 2020 and April 2026. The main findings can be summarized into three points.

SWV appears to be the most promising quantitative indicator for rehabilitation decision-making. Multiple studies included in this review consistently showed that changes in SWV correlate well with pain relief and functional recovery. A 2024 Chinese study further reported that the AUC of postoperative SWV for distinguishing rehabilitation outcomes using a Constant-Murley score of 70 as the cutoff was 0.781 (95% CI not reported), whereas the AUCs of tendon thickness and width were only 0.531 and 0.466, respectively [25].

One month and 12 weeks postoperatively are two key decision windows, corresponding respectively to retear risk warning and prognostic prediction of return to sport, and can also serve as a basis for adjusting load intensity. A systematic review by Seth et al. (16 studies, 520 patients) reported that SWV at one month postoperatively was significantly higher in the retear group than in the healed group (P < 0.05), suggesting that SWE measurement at one month may have potential as a warning indicator for retear risk, although predictive modeling was not performed in the primary studies [26]. Ishikawa et al. (2021) provided parallel evidence from the perspective of muscle activity. Using real-time tissue elastography to follow 20 patients after rotator cuff repair, they found that supraspinatus muscle activity increased markedly between six weeks and three months postoperatively, and by three months it had recovered to a level not significantly different from that of healthy controls. This finding independently validates, from the dimension of muscle activity recovery, the reliability of around 12 weeks (i.e., three months) as a key turning point for functional recovery [27], which is consistent with the study by Maloof et al.

Muscle morphology parameters (CSA and EI) can serve as quantitative indicators of muscle atrophy and fatty infiltration and can be used to identify patients with a “rehabilitation bottleneck.” Zhu et al. (2024) found that improvements in CSA and EI at six months postoperatively were significantly positively correlated with the Constant score (r = 0.633 and 0.693, respectively), and the combination of the two parameters predicted functional recovery with an AUC of 0.922 (95% CI: 0.848 to 0.967) [18]; however, Umehara et al. (2023) showed that even in the absence of obvious degenerative changes in muscle morphology and composition, the active and passive shear moduli of the supraspinatus muscle remained impaired, suggesting that mechanical properties may be more sensitive rehabilitation monitoring indicators than morphological parameters [28].

It should be noted, however, that the lack of correlation between tendon thickness and functional scores reported by Kim et al. (n = 42) should be interpreted with caution. With a sample size of 42, this study had limited statistical power to detect correlations below approximately r = 0.42 (assuming 80% power at α = 0.05, two-sided). Thus, the null finding cannot be taken as evidence of no association, and larger studies are needed to clarify the clinical relevance of morphological parameters.

4.2 Comparison with existing reviews

This review differs from the existing literature in three aspects. First, previous reviews have largely focused on the technical application of ultrasound elastography—for example, a 2023 review summarized the potential of SWE in the diagnosis and monitoring of rotator cuff tears [26], and a 2024 systematic review summarized the role of ultrasound elastography in assessing postoperative retear risk [29]. However, none of these reviews took the quantitative link between ultrasound parameters and rehabilitation decision-making as their core research question. Although Sciarretta et al. (2023) emphasized that “rehabilitation protocols should always be individualized,” they did not address how ultrasound imaging could be used to achieve such individualization [30].

Second, this review is the first to systematically map the complete evidence chain from “ultrasound parameter acquisition” to “rehabilitation decision output,” covering both the types of decisions (initiation and progression of active movement, load intensity adjustment, prognostic prediction of return to sport, retear risk warning and rehabilitation downstaging, and rehabilitation bottleneck identification) and decision thresholds, thereby filling the decision-oriented gap in existing reviews. Men et al. (2025) explicitly stated that “musculoskeletal ultrasound can not only be used for the differential diagnosis of rotator cuff tears but also for evaluating surgical outcomes and guiding rehabilitation at different postoperative stages” [11], which is highly consistent with the perspective of the present review. However, that article is a narrative review rather than a systematic review and did not systematically search for or synthesize evidence specifically related to decision-making.

Third, a 2024 Chinese review on “Advances in rehabilitation after rotator cuff surgery” also pointed out that “rehabilitation protocols should be individualized according to the patient's injury and surgical conditions” [31]. However, that review did not address the specific application path of using musculoskeletal ultrasound as a tool for individualization. By systematically synthesizing decision-related examples from primary studies, the present review provides a preliminary evidence framework for translating the concept of “individualized rehabilitation” into the practice of “image-informed precision rehabilitation.”

4.3 Clinical implications: feasibility of constructing an “ultrasound‑informed rehabilitation framework”

The value of SWE in informing postoperative rehabilitation advancement has been further substantiated by recent evidence. A systematic review by Lin et al. synthesized data from 11 studies and confirmed that postoperative changes in muscle stiffness, particularly during the 1- to 6-month window, are associated with retear risk and functional recovery, highlighting the potential of sonoelastography to support postoperative monitoring strategies [29]. Complementing this, a prospective cohort study by Yan et al. involving 73 patients demonstrated that SWV at 3 months postoperatively correlated positively with Constant-Murley scores (r = 0.618 to 0.643, P < 0.001), providing direct evidence that quantitative ultrasound parameters can serve as an objective imaging basis for tailoring rehabilitation exercise protocols [17]. Together, these findings provide preliminary support for the concept that ultrasound monitoring may provide objective information to support individualized rehabilitation decision-making.

Specifically, the findings of this review support the construction of a precision rehabilitation pathway at three levels. First, at the evidence level. In addition to the dynamic SWV trajectory reported by Solari et al. [20], this judgment has been directly validated clinically. A follow-up study using high-frequency ultrasound and elastography in patients after supraspinatus tendon repair confirmed that SWV has reference value for evaluating pain relief (VAS) and determining rehabilitation outcomes, and may provide clinicians and rehabilitation physicians with additional objective information when developing individualized rehabilitation plans [32].

Second, at the technical level. Musculoskeletal ultrasound is non-invasive, real-time, and reproducible, making it well suited for bedside rehabilitation decision-making. Men et al. (2025) clearly stated that “ultrasound can monitor the repair process of injured tissues and rehabilitation progress in real time during rehabilitation, and by comparing serial images, an individualized rehabilitation protocol can be developed for each patient” [11]. The utility of elastography for assessing tissue biomechanical properties has also been demonstrated in other musculoskeletal conditions, such as plantar fascia assessment in diabetic patients [33].

Third, at the practical level. One study has reported the effect of isokinetic training under musculoskeletal ultrasound monitoring in patients after arthroscopic rotator cuff repair [34], suggesting that ultrasound monitoring can be embedded into routine rehabilitation workflows as an adjunctive tool.

That said, it should be noted that the construction of a precision rehabilitation pathway is still in its early stages. A preliminary attempt has been made to integrate ultrasound monitoring into structured rehabilitation workflows, as illustrated by a study on isokinetic training following arthroscopic repair [34]. Nevertheless, most current studies still follow an “observation-correlation” model (i.e., finding statistical associations between ultrasound parameters and functional outcomes) rather than a “measurement-decision” model (i.e., establishing a direct mapping from ultrasound parameters to rehabilitation instructions). In other words, most research to date has remained at the level of “being able to monitor,” and there is still a considerable methodological gap before ultrasound parameters can be prospectively validated as decision-support tools.

The integration of ultrasound monitoring into rehabilitation decision-making aligns with broader interdisciplinary approaches in sport and health science, which emphasize the synthesis of physiology, biomechanics, and data analytics to develop personalized rehabilitation strategies [35].

4.4 Research gaps and future directions

4.5 Limitations of this review

This review has several limitations. First, the included studies were mainly observational, with a lack of large-scale randomized controlled trials, resulting in an overall low level of evidence. Moreover, some of the included studies were case series, which cannot provide dynamic change data, potentially compromising the completeness of the evidence. Second, there was considerable heterogeneity across studies in ultrasound parameter definitions, measurement methods, and follow-up time points, which precluded a meta-analysis and limited the synthesis to a narrative summary. Of note, in the measurement of muscle morphology parameters, the choice of region of interest (ROI) is an important source of heterogeneity. Yuri et al. (2021) specifically addressed this issue by comparing the sensitivity of three different measurement ranges (anterior region, anterior-middle posterior region, and whole cross-sectional area) of the supraspinatus muscle using real-time tissue elastography. They found that the whole cross-sectional area had the highest sensitivity (81.5%) and the best intraclass correlation coefficient (ICC, 0.92 to 0.99) [42]. This quantitative evidence provides a specific operational pathway to resolve the “non-standardized measurement range” issue identified in this review: future studies reporting muscle morphology parameters such as CSA should preferentially adopt the whole cross-sectional area as the standardized measurement range. In addition, although the screening and data extraction were performed independently by two reviewers, formal inter-rater agreement statistics were not calculated, which may limit the reproducibility assessment of the selection process. Third, in terms of sample size, the included studies had sample sizes ranging from 39 to 95 (total 673 participants), with considerable variation between studies, which may have some impact on the comparability of results. Fourth, gray literature and unpublished studies were not included, so there may be potential publication bias. It is worth noting that during the literature screening process, one relevant master's thesis, titled “Application of musculoskeletal ultrasound after rotator cuff repair,” met the inclusion criteria but was excluded because the publication type was restricted to journal articles [43,44]. That thesis pointed out that the retear rate after rotator cuff repair is high, but clinical diagnosis is difficult; although MRI is the gold standard, it is expensive, and therefore musculoskeletal ultrasound can serve as a low-cost, efficient adjunctive examination method to provide a basis for patients’ next-step treatment. Excluding such gray literature ensures a quality benchmark based on peer-reviewed publications, but it may also lead to the omission of high-quality negative results or exploratory findings. Therefore, caution should be exercised when generalizing the conclusions of this review. Fifth, despite optimization of the search strategy, some relevant studies may still have been missed, particularly those in languages other than English or Chinese. Sixth, the risk of bias assessment revealed that among the 11 included studies, six were rated as having moderate risk of bias and five as having low risk of bias. Although no study was judged to have high risk of bias, the proportion of studies with moderate risk of bias was relatively high. Their limitations mainly involved retrospective design, inadequate control of confounding factors, and failure to report measurement consistency in some studies. These factors may have some impact on the overall strength of the evidence, and therefore the conclusions of this review should be interpreted with caution. Despite these limitations, this review is the first to systematically focus on the integration of evidence on “musculoskeletal ultrasound monitoring to inform rehabilitation decision-making after rotator cuff repair,” and it provides a valuable reference framework for future research directions and clinical practice translation in this field.