Comparative Effects of Acellular Dermal Matrix and Atelocollagen Injections in Intratendinous Partial-Thickness Rotator Cuff Tears: A Retrospective Cohort Study Authors: Jae-Yoon Baek, Eun-Ji Yoon, Kyeong-Eon Kwon, Jong-Ho Kim Published: 2026 ABSTRACT AbstractBackgroundThe purpose was to compare the functional and structural outcomes between injectable acellular dermal matrix (ADM) and atelocollagen in patients with intratendinous partial-thickness rotator cuff tears.MethodsFrom January 2020 to May 2021, 115 consecutive patients with magnetic resonance imaging (MRI)-confirmed intratendinous partial-thickness supraspinatus tears who met the inclusion criteria were retrospectively identified. After applying predefined exclusion criteria, 85 patients remained for final analysis and were categorized into 2 groups: intratendinous injection with 1 mL ADM (group A, n = 40) and 1 mL type I atelocollagen (group B, n = 45). The visual analog scale (VAS) score for pain, American Shoulder and Elbow Surgeons (ASES) score, Constant Shoulder score (CSS), and range of motion (ROM) were evaluated before injection and at 1, 3, 6, and 9 months and the last follow-up after injection. MRI was performed at 6 months after injection to evaluate rotator cuff integrity and healing response.ResultsThe mean follow-up period was 14.1 months. Both groups demonstrated significant improvements in pain, functional outcomes, and ROM from baseline to final follow-up. ASES scores increased from 60.3 to 88.4 in group A and from 67.4 to 90.2 in group B (both p < 0.001). CSS improved from 58.3 to 73.2 in group A and from 61.1 to 71.9 in group B (both p < 0.001). External rotation at the side improved significantly in both groups (group A: 73.7° to 80.5°, p = 0.035; group B: 74.2° to 79.5°, p = 0.011), and other ROM parameters demonstrated similar longitudinal improvement. Activity VAS decreased from 5.1 to 1.5 in group A and from 5.8 to 1.5 in group B. However, no statistically significant differences were observed between the 2 treatment groups in any clinical outcome measures at final follow-up. MRI demonstrated tendon improvement in 32.5% of ADM-treated and 35.6% of atelocollagen-treated patients (p = 0.821).ConclusionsBoth ADM and atelocollagen injections were associated with clinically meaningful improvement in patients with intratendinous partial-thickness rotator cuff tears, with generally maintained MRI-based structural outcomes at follow-up.Keywords: Rotator cuff injuries, Acellular dermis, Collagen type I, Allografts, Injections Rotator cuff tears are a common cause of shoulder pain and dysfunction. Although advances in arthroscopic techniques and augmentation materials have improved surgical outcomes, retear rates remain high: long-term retear rates of 27%–50% have been reported after arthroscopic repair overall, and rates approach 78% in massive tears.1,2,3,4,5) Given these challenges, effective conservative treatments are increasingly emphasized, particularly for partial-thickness rotator cuff tears (PTRCTs), where avoiding surgery can be beneficial. Collagen, a major component of the extracellular matrix, constitutes 60%–85% of the dry weight of tendons, with type I collagen accounting for about 90% of the tendon’s composition. In normal connective tissue, collagen fibers transfer force, release energy, and avert mechanical breakdown.6) Collagen-based biomaterials have been applied across multiple fields, including dermatology (wound healing), dentistry (gingival repair), and neurosurgery (dural grafts). In orthopedics, collagen has been studied for its role in cartilage and ligament regeneration, and more recently as a biologic option for tendon healing.7) Atelocollagen, produced by removing antigenic telopeptides via pepsin digestion, is a highly purified form of collagen with low toxicity and immunogenicity. Because of its biocompatibility, safety, and stability, atelocollagen has been proposed as a favorable injectable scaffold for tendon healing. In contrast, acellular dermal matrix (ADM), derived from decellularized human dermis, preserves the collagen-rich extracellular matrix and acts as a biological scaffold, facilitating fibroblast proliferation, neovascularization, and re-epithelialization.8,9,10,11,12,13) ADM has demonstrated favorable biomechanical properties, including increased suture pullout strength, and has shown promising results in augmentation of large and massive rotator cuff tears.14,15,16) Most previous research has focused on patch-type ADM, whereas evidence on gel-type formulations is scarce. Furthermore, recent studies have questioned the clinical effectiveness of ADM patches, raising interest in whether injectable ADM gel could provide a simpler and less invasive alternative while retaining key advantages.17) Intratendinous partial-thickness tears represent a distinct morphologic subtype characterized by a contained defect within the tendon substance while preserving both the articular and bursal surfaces. This structural feature may allow injected biologic materials to remain localized within the tear site, potentially enhancing intratendinous retention and facilitating a more sustained biological effect. In contrast, articular- or bursal-sided tears may allow dispersion of the injectate into adjacent compartments, potentially limiting its local therapeutic action. Although intratendinous partial-thickness tears may show a relatively more favorable natural course than some articular- or bursal-sided partial tears, they may remain persistently symptomatic and are often associated with intratendinous degenerative change. In addition, PTRCTs are not always static lesions and may progress over time, with some tears enlarging or converting to full-thickness defects during nonoperative follow-up.18,19) In this setting, biologic injection therapy may be clinically relevant not only because the contained morphology of the lesion may enhance local retention of the injectate, but also because such treatment may help modulate the degenerative tendon microenvironment, support symptom relief, and potentially contribute to structural preservation or mitigation of tear progression in selected patients. Accordingly, this study aimed to compare the clinical and structural outcomes of 2 biologic injectable materials—gel-type ADM and type I atelocollagen—used as adjunctive conservative treatment options for magnetic resonance imaging (MRI)-confirmed intratendinous partial-thickness supraspinatus tears. This comparison was conducted under a standardized rehabilitation protocol to evaluate relative treatment responses between the 2 injectables. We hypothesized that clinical and structural outcomes would differ between the 2 treatment groups. METHODS This retrospective study was conducted at a single institution (Yeouido St. Mary’s Hospital) following institutional review board approval (IRB No. SC24RISI0159), and the requirement for informed consent was waived due to its retrospective design. Patient Selection We retrospectively reviewed the medical records and imaging studies of consecutive patients who underwent collagen injection-based conservative treatment for rotator cuff tears between January 2020 and May 2021. During this period, the 2 injectable biologic materials were used sequentially in clinical practice. Type I atelocollagen was used from January 2020 to August 2020, whereas ADM was used from September 2020 to May 2021. Patients were eligible if they met the following inclusion criteria: (1) MRI-confirmed intratendinous supraspinatus partial-thickness tear, (2) treatment with either atelocollagen or ADM injection, (3) complete medical and imaging records available for review, and (4) a minimum follow-up of 1 year. Before biologic injection was considered, all patients had undergone at least 3 months of noninjection conservative treatment, including oral analgesics or nonsteroidal anti-inflammatory drugs (NSAIDs) and rehabilitation-based management such as physical therapy and home exercise, but remained symptomatic. A total of 115 patients met the inclusion criteria during the study period. Exclusion criteria were as follows: (1) articular- or bursal-sided partial-thickness tear, (2) full-thickness rotator cuff tear, (3) large-to-massive tear, (4) prior injection therapy including corticosteroid, prolotherapy, collagen, or polydeoxyribonucleotide, (5) previous trauma involving the affected shoulder, (6) isolated subscapularis tear, (7) previous surgery on the affected shoulder, or (8) concomitant shoulder pathology such as glenohumeral arthritis, labral lesions, fractures, or severe stiffness. After applying the exclusion criteria, 30 patients were excluded, leaving 85 patients for final analysis. The final cohort comprised 40 patients in the ADM group and 45 patients in the atelocollagen group (Fig. 1). Because of the retrospective design, a priori sample size calculation was not performed. The sample size was determined by the number of consecutive eligible patients who met the predefined inclusion and exclusion criteria during the study period. All tears were diagnosed based on preintervention MRI performed within 3 months of treatment initiation and were confirmed by an experienced musculoskeletal radiologist (JEK). Flowchart of patient selection. From an initial 115 eligible patients, 30 were excluded based on predefined criteria, leaving 85 patients divided into the acellular dermal matrix (ADM; n = 40) and atelocollagen (n = 45) groups. MRI: magnetic resonance imaging, PDRN: polydeoxyribonucleotide. Intratendinous tears were preferentially included because their contained morphology permits focal accumulation of the injectate within the tendon substance. In contrast, articular- or bursal-sided tears may allow dispersion of the material into adjacent compartments, potentially reducing intratendinous retention. Injection Technique Intratendinous injection of ADM and atelocollagen into the supraspinatus tendon tear site was performed in groups A and B, respectively. All procedures were conducted by a single senior surgeon (JHK) under ultrasonographic guidance using a high-resolution ultrasound system (RS85, Samsung Medison) with a 12-MHz linear high-frequency probe. Patients were positioned in the modified Crass position, with the affected arm internally rotated and extended, placing the hand on the ipsilateral buttock. This position facilitated optimal visualization of the supraspinatus tendon and allowed a safe medial-to-lateral needle trajectory while minimizing additional tendon injury. After sterile preparation, a needle was advanced from medial to lateral through the deltoid muscle and subacromial space under real-time ultrasound guidance. The needle tip was carefully positioned within the intratendinous tear site, and accurate placement was confirmed sonographically (Fig. 2). Ultrasonography-guided injection at the intratendinous tear site (arrow). Upon injection, the atelocollagen disperses within the tear gap, filling the space between the bursal and articular layers of the torn rotator cuff. BSSP: bursal layer of supraspinatus tendon, ASSP: articular layer of supraspinatus tendon, GT: greater tuberosity. A commercially available 1-mL gel formulation was used in all cases, and the full 1 mL of the injectate was administered intratendinously. The material was injected slowly to allow dispersion within the tear gap. If increased resistance or patient discomfort occurred during injection, the needle position was adjusted while maintaining intratendinous placement to ensure complete delivery of the injectate. The injection technique was identical between the 2 groups, including patient positioning, ultrasound guidance, needle approach, and injection site. A standardized injection volume of 1 mL was used in all cases, and all procedures were performed by a single surgeon. The only difference between groups was the type of injected material. ADM and Type I Atelocollagen We used gel-type (1 mL) products of type I atelocollagen (RegenSeal, Sewon Cellontech Co.) derived from the porcine dermis. Collagen was treated with pepsin to remove the terminal telopeptides. We also used gel-type (1 mL) products of hydrated human collagen from ADM (CG Reallo Inject, CGBIO Co.). Shoulder ROM Evaluation and Clinical Assessment Visual analog scale (VAS) score was used to measure pain intensity before and after intervention. Shoulder range of motion (ROM), including forward flexion (FF), external rotation with the arm at the side, and internal rotation at the back, was carefully assessed using a goniometer. The inferior margins of the scapula were used as a reference point for the seventh thoracic vertebra (T7), and internal rotation was determined by evaluating the patient’s ability to reach the spinal spinous process with the tip of their thumb while seated: T7–T12 to 12–7 points, L1–L5 to 6–2 points, and buttock to 1 point. Functional outcomes were measured with the American Shoulder and Elbow Surgeons (ASES) score, Constant Shoulder score (CSS), and Korean Shoulder score (KSS), which is a validated score for the Korean population. Both shoulder ROM and functional scores were assessed and recorded before injection and at 1, 3, 6, and 9 months and last follow-up after injection. All patients were followed up for a minimum of 12 months. All clinical assessments and ROM measurements were performed by a single examiner, a musculoskeletal radiologist (JEK), using a standardized protocol to ensure consistency across evaluations. The significance of changes in functional outcome scores reported by patients at each follow-up was evaluated using the method proposed by Norman et al.20) In the absence of more specific criteria, the threshold of discrimination for changes in health-related quality of life is generally approximated to half a standard deviation. Therefore, the minimally important difference (MID) was determined based on half of the standard deviation of the postinjection change in functional outcome scores, such as ASES score, KSS, and CSS. Additional baseline variables were collected to assess comparability between groups, including symptom duration, body mass index (BMI), comorbidities (including diabetes mellitus), and Goutallier classification as a measure of fatty degeneration. Baseline clinical outcome measures, including VAS, ASES, CSS, and ROM, were also recorded and compared between groups. Rehabilitation All patients underwent the same rehabilitation. All enrolled patients received the same home exercise education program. Gentle, progressive stretching exercises including the pendulum, sleeper’s stretch, and posterior capsular stretching were prescribed. When the patient had achieved a comfortable passive ROM in the affected shoulder, strengthening exercises with resistance bands were carried out. Band resistance was increased incrementally. The patients did not change to a higher resistance band until exercises on the current band were easy and painless. NSAIDs were prescribed to patients who had unbearable pain during exercise. Image Evaluation To assess the presence of an intratendinous partial-thickness tear of the supraspinatus tendon and to measure changes after injection in both the anteroposterior and mediolateral directions, coronal oblique and sagittal oblique images from T2-weighted MRI were used. All patients underwent MRI at 6 months after injection. All MRI examinations were performed using a 3.0-T scanner (Magnetom Verio; Siemens Healthineers) at Yeouido St. Mary’s Hospital. Coronal oblique, sagittal oblique, and axial fast spin-echo T2-weighted images (repetition time/echo time, 3,600–5,180/64–71 ms) were used. The definition of a tear was a hyperintense signal with a filling defect within the tendon on T2-weighted images. Specifically, an intratendinous tear was defined as a defect within the tendon substance while maintaining the integrity of both the bursal and articular layers. Additionally, tears with preserved continuity at the tendon insertion site were classified as partial-thickness tears. Tear size of the supraspinatus tendon was measured on 2 imaging planes (sagittal oblique and coronal oblique). In this study, since the indication was set for intratendinous partial-thickness tears, it was challenging to adopt commonly used tendon integrity grading systems such as the Sugaya classification. Therefore, the outcomes were categorized into 3 groups with reference to the classification approaches described by Kim et al.21) and Maman et al.22) Given that intratendinous partial-thickness tears are typically small and subtle, a threshold of 2 mm was adopted to improve sensitivity in detecting clinically relevant changes while still considering potential measurement variability. Accordingly, tendon integrity was categorized as improved (> 2 mm decrease in tear size on at least 1 imaging plane, category 1), unchanged (< 2 mm difference, category 2), or progression (> 2 mm increase in tear size, category 3).21,22) We classified the tear as category 1 or 3 when there was a change of more than 2 mm in at least 1 image between the sagittal oblique and coronal oblique images. Fatty infiltration of the rotator cuff muscles was assessed on sagittal oblique T1-weighted MRI images at the level of the scapular Y-view using the Goutallier classification, adapted for MRI. This grading system ranges from 0 to 4, where grade 0 indicates no visible fat, grade 1 represents minimal fatty streaks, grade 2 reflects a predominance of muscle over fat, grade 3 indicates approximately equal amounts of muscle and fat, and grade 4 represents greater fat than muscle. MRI-based changes in tear size were independently assessed by 2 orthopedic surgeons who were blinded to group allocation: 1 with over 14 years of orthopedic experience (JHK) and 1 with over 6 years of orthopedic experience (EJY). Interobserver reliability was evaluated using intraclass correlation coefficients (ICCs).23) To assess intraobserver reliability, 1 examiner repeated the measurements after a 1-month interval. Statistical Analysis Statistical analyses were conducted using the IBM SPSS Statistics version 19 (IBM Corp.). Longitudinal outcomes measured at multiple time points (preinjection, 1, 3, 6, and 9 months, and 1 year) were analyzed using repeatedmeasures analysis with group (ADM vs. atelocollagen) as a between-subject factor and time as a within-subject factor. Main effects of group and time and the group-by-time interaction were evaluated. When significant interactions were identified, post hoc between-group comparisons at individual time points were performed with Bonferroni adjustment for multiple comparisons. Sphericity was assessed using Mauchly’s test, and Greenhouse-Geisser correction was applied when sphericity was violated. VAS and internal rotation (vertebral level score), although ordinal in nature, were analyzed as continuous variables to facilitate longitudinal modeling, consistent with common practice in clinical orthopedic research. Categorical variables, including MRI-based healing status categories and MID achievement rates, were compared between groups using the chi-square test or Fisher’s exact test as appropriate. Interobserver reliability was assessed using ICCs based on a 2-way mixed-effects model assuming single measurements and absolute agreement. RESULTS Patient Demographics A total of 85 patients were included in this study, comprising 40 in group A (ADM injection) and 45 in group B (atelocollagen injection). The mean age was 61.6 ± 12.0 years in group A and 59.0 ± 10.9 years in group B, showing no statistically significant difference (p = 0.315). The mean follow-up imaging duration was 6.5 ± 1.5 months in group A and 7.1 ± 2.5 months in group B (p = 0.181). Sex distribution was also similar between the 2 groups, with the proportion of male patients being 52.5% (21 / 40) in group A and 42.2% (19 / 45) in group B (p = 0.465) (Table 1). The mean anterior–posterior tear size was 0.6 ± 0.2 cm in group A and 0.6 ± 0.3 cm in group B (p = 0.664). The medial–lateral tear size was 0.8 ± 0.3 cm in group A and 0.9 ± 0.3 cm in group B (p = 0.297). There were no statistically significant differences in tear size between the 2 groups. Additional baseline variables were analyzed to further assess comparability between groups. There were no significant differences in BMI (23.6 ± 2.6 vs. 23.8 ± 2.9, p = 0.737), baseline Goutallier grade (1.40 ± 0.50 vs. 1.60 ± 0.50, p = 0.067), postintervention Goutallier grade at the final follow-up (1.27 ± 0.45 vs. 1.27 ± 0.45, p = 0.932), symptom duration (11.2 ± 9.7 vs. 8.6 ± 4.8 months, p = 0.127), or diabetes mellitus prevalence (20.0% vs. 26.7%, p = 0.480) between groups A and B. Baseline clinical parameters, including VAS pain scores, functional outcome scores (ASES, CSS, and KSS), and ROM, were also comparable between the 2 groups, with no statistically significant differences observed. Table 1Baseline Demographics of the Acellular Dermal Matrix and Atelocollagen GroupsVariableGroup AGroup Bp-valueAge (yr)61.6 ± 12.059.0 ± 10.90.315BMI (kg/m2)23.6 ± 2.623.8 ± 2.90.737Sex (male)21 / 40 (52.5)19 / 45 (42.2)0.465Diabetes mellitus8 / 40 (20.0)12 / 45 (26.7)0.480Follow-up imaging duration (mo)6.5 ± 1.57.1 ± 2.50.181Symptom duration (mo)11.2 ± 9.78.6 ± 4.80.127PreinterventionVAS pain score (activity)5.1 ± 2.05.8 ± 2.60.195ASES score60.3 ± 16.667.4 ± 18.60.066CSS58.3 ± 14.461.2 ± 13.10.352KSS71.0 ± 16.476.2 ± 14.80.132FF (°)161.1 ± 21.1153.7 ± 18.40.093ERside (°)73.7 ± 14.874.2 ± 14.10.880ERabd (°)81.0 ± 17.880.9 ± 15.50.975Add (°)36.5 ± 6.437.3 ± 4.80.513IR score10.7 ± 2.110.9 ± 1.90.646Tear sizeAP (cm)0.6 ± 0.20.6 ± 0.30.664ML (cm)0.8 ± 0.30.9 ± 0.30.297Goutallier grade (SSP)Preintervention1.40 ± 0.501.60 ± 0.500.067Postintervention1.27 ± 0.451.27 ± 0.500.932Values are presented as mean ± standard deviation or number (%). Group A: acellular dermal matrix injection, Group B: atelocollagen injection.BMI: body mass index, VAS: visual analog scale, ASES: American Shoulder and Elbow Surgeons, CSS: Constant Shoulder score, KSS: Korean Shoulder score, FF: forward flexion, ERside: external rotation at side, ERabd: external rotation at abduction, Add: adduction, IR: internal rotation, AP: anterior to posterior, ML: medial to lateral, SSP: supraspinatus. Baseline Demographics of the Acellular Dermal Matrix and Atelocollagen Groups Values are presented as mean ± standard deviation or number (%). Group A: acellular dermal matrix injection, Group B: atelocollagen injection. BMI: body mass index, VAS: visual analog scale, ASES: American Shoulder and Elbow Surgeons, CSS: Constant Shoulder score, KSS: Korean Shoulder score, FF: forward flexion, ERside: external rotation at side, ERabd: external rotation at abduction, Add: adduction, IR: internal rotation, AP: anterior to posterior, ML: medial to lateral, SSP: supraspinatus. Healing Status Interobserver reliability for MRI-based assessment of tear size changes was excellent (ICC = 0.91; 95% CI, 0.86–0.94). Regarding healing status assessed by MRI at the final follow-up, the distribution of healing categories was comparable between the 2 treatment groups (Table 2). In group A (ADM injection), 13 of 40 patients (32.5%) demonstrated improvement (category 1), 23 patients (57.5%) showed no measurable change (category 2), and 4 patients (10.0%) exhibited progression (category 3). In group B (atelocollagen injection), 16 of 45 patients (35.6%) showed improvement, 23 patients (51.1%) had no change, and 6 patients (13.3%) exhibited progression. When analyzing improvement (category 1) alone, there was no statistically significant difference between the 2 groups (32.5% vs. 35.6%, p = 0.821). Similarly, the proportion of patients showing no change (category 2) did not differ significantly (p = 0.664), and progression (category 3) was also comparable between groups (p = 0.743). The absolute between-group difference in MRI-based improvement rate was −3.1 percentage points (95% CI, −23.2 to 17.1). Overall, MRI-based evaluation revealed no significant difference in healing distribution between the ADM and atelocollagen groups. Table 2Changes in Tear Size on Follow-up Magnetic Resonance ImagingHealing*Group AGroup Bp-value113 (32.5)16 (35.6)0.821223 (57.5)23 (51.1)0.66434 (10.0)6 (13.3)0.743Values are presented as number (%). Group A: acellular dermal matrix injection, Group B: atelocollagen injection.*1: improved, 2: unchanged, 3: progression. Changes in Tear Size on Follow-up Magnetic Resonance Imaging Values are presented as number (%). Group A: acellular dermal matrix injection, Group B: atelocollagen injection. *1: improved, 2: unchanged, 3: progression. Range of Motion Repeated-measures ANOVA demonstrated a significant main effect of time for all ROM parameters (all p < 0.001), indicating progressive improvement over time in both groups. No significant group-by-time interaction was observed for external rotation at the side, adduction, external rotation in abduction, or internal rotation (all p > 0.05), suggesting comparable recovery trajectories between ADM and atelocollagen injections. For FF, a significant group-by-time interaction was observed (p = 0.014). At 1 month, group A demonstrated higher FF compared with group B (168.3° ± 5.2° vs. 161.0° ± 17.8°), indicating more rapid early recovery. However, this difference was not maintained at final follow-up. At 1 year, FF was 165.8° ± 14.7° in group A and 167.7° ± 7.3° in group B, reflecting comparable long-term outcomes. From baseline to final follow-up, FF changed from 161.1° ± 21.1° to 165.8° ± 14.7° in group A and from 153.7° ± 18.3° to 167.7° ± 7.3° in group B. External rotation at the side improved from 73.7° ± 14.8° to 80.5° ± 13.9° in group A and from 74.2° ± 14.0° to 79.5° ± 2.0° in group B at final follow-up. Adduction improved from 36.5° ± 6.4° to 38.7° ± 5.1° in group A and from 37.3° ± 4.8° to 40.0° ± 0.0° in group B. Internal rotation improved from 10.7 ± 2.1 to 11.3 ± 1.8 in group A and from 10.9 ± 1.9 to 11.5 ± 1.2 in group B. External rotation in abduction demonstrated a similar gradual improvement pattern in both groups. No significant between-group differences were observed at final follow-up for any ROM parameter (Fig. 3). Longitudinal changes in range of motion parameters, including forward flexion (FF; A), external rotation at abduction (ERabd; B), external rotation at side (ERside; C), adduction (Add; D), and internal rotation (IR; E). Group A (acellular dermal matrix) is represented by a solid line with upward triangle markers (▲), and group B (atelocollagen) by a dashed line with square markers (■). Error bars denote the 95% CI of the mean. The x-axis represents follow-up time points (preinjection, 1 month, 3 months, 6 months, 9 months, and 1 year). Clinical Outcomes Repeated-measures ANOVA revealed a significant main effect of time for resting VAS, activity VAS, ASES, CSS, and KSS scores (all p < 0.001), indicating substantial improvement over time in both treatment groups. No significant group-by-time interaction was observed for any patient-reported outcome measure (all p > 0.05), suggesting comparable recovery trajectories between ADM and atelocollagen injections (Table 3). Table 3Repeated-Measures Analysis of Variance Results for Clinical OutcomesVariableTime effect (p-value)Group effect (p-value)Group × time interaction (p-value)ASES score< 0.0010.273> 0.05CSS< 0.0010.548> 0.05KSS< 0.001> 0.05> 0.05VAS score (activity)< 0.001> 0.05> 0.05VAS score (resting)< 0.001> 0.05> 0.05Group A: acellular dermal matrix injection, Group B: atelocollagen injection.ASES: American Shoulder and Elbow Surgeons, CSS: Constant Shoulder score, KSS: Korean Shoulder score, VAS: visual analog scale. Repeated-Measures Analysis of Variance Results for Clinical Outcomes Group A: acellular dermal matrix injection, Group B: atelocollagen injection. ASES: American Shoulder and Elbow Surgeons, CSS: Constant Shoulder score, KSS: Korean Shoulder score, VAS: visual analog scale. Resting VAS decreased from 4.3 ± 2.5 to 1.5 ± 2.0 in group A and from 5.2 ± 3.3 to 0.9 ± 1.5 in group B at final follow-up. Activity-related VAS improved from 5.1 ± 2.0 to 1.5 ± 2.0 in group A and from 5.8 ± 2.6 to 1.5 ± 1.4 in group B. The mean reductions in both resting and activity VAS exceeded the MID (1.5–1.8), indicating clinically meaningful pain improvement in both groups (Fig. 4). Longitudinal changes in visual analog scale (VAS) pain scores during activity (A) and at rest (B). Group A (acellular dermal matrix) is represented by a solid line with upward triangle markers (▲), and group B (atelocollagen) by a dashed line with square markers (■). Error bars denote the 95% CI of the mean. The x-axis represents follow-up time points (preinjection, 1 month, 3 months, 6 months, 9 months, and 1 year). Functional outcome measures also demonstrated marked improvement. The ASES score increased from 60.3 ± 16.6 to 88.4 ± 13.8 in group A and from 67.4 ± 18.6 to 90.2 ± 12.4 in group B. The CSS improved from 58.3 ± 14.3 to 73.2 ± 7.4 in group A and from 61.1 ± 13.0 to 71.9 ± 6.5 in group B. The KSS increased from 71.0 ± 16.3 to 92.3 ± 9.3 in group A and from 76.1 ± 14.8 to 91.9 ± 8.4 in group B (Fig. 5). Longitudinal changes in American Shoulder and Elbow Surgeons (ASES; A), Constant Shoulder score (CSS; B), and Korean Shoulder score (KSS; C) scores. Group A (acellular dermal matrix) is represented by a solid line with upward triangle markers (▲), and group B (atelocollagen) by a dashed line with square markers (■). Error bars denote the 95% CI of the mean. The x-axis represents follow-up time points (preinjection, 1 month, 3 months, 6 months, 9 months, and 1 year). The proportion of patients achieving the MID was comparable between groups. For ASES, 87.5% of patients in the ADM group and 68.9% in the atelocollagen group achieved MID at final follow-up (p = 0.073). A post hoc power analysis for this comparison yielded an estimated power of 56%. For the CSS, MID was achieved by 57.5% and 62.2% of patients, respectively (p = 0.824). Similarly, 82.5% of ADM-treated patients and 64.4% of atelocollagen-treated patients exceeded the MID for KSS (p = 0.104). There were no statistically significant differences between groups in the proportion of patients achieving clinically meaningful functional improvement. DISCUSSION In this retrospective comparative study, both gel-type ADM and type I atelocollagen injections were associated with significant improvements in pain, functional outcomes, and ROM in patients with intratendinous PTRCTs. No statistically significant differences were observed between the 2 treatment groups at final follow-up. The magnitude of improvement in patient-reported outcomes exceeded the MID, indicating that these changes were not only statistically significant but also clinically meaningful. Multiple baseline variables known to influence rotator cuff outcomes—including tear size, fatty degeneration (Goutallier classification), BMI, symptom duration, comorbidities, and baseline functional scores—were comparable between groups, reducing the likelihood that the observed findings were driven by baseline differences, although residual confounding cannot be completely excluded. Regarding structural healing assessed by MRI at 6 months after injection, 32.5% of ADM-treated patients and 35.6% of atelocollagen-treated patients demonstrated tendon improvement, and most patients in both groups showed no progression of tear size. Although the ADM group showed faster early recovery of FF at 1 month, this difference was not sustained over time and its clinical significance remains uncertain. One possible explanation is that the preserved extracellular matrix structure of ADM may have provided a more favorable early biologic scaffold for modulation of the local tendon environment, potentially facilitating pain reduction or early restoration of motion. In addition, although the proportion of patients achieving the MID for the ASES score was numerically higher in the ADM group, this difference did not reach statistical significance. Given the limited post hoc power for this comparison, this finding should be interpreted cautiously, and the possibility of a type II error due to limited sample size cannot be excluded. Intratendinous partial-thickness tears represent a distinct pathologic entity characterized not only by focal fiber disruption but also by underlying degenerative changes within the tendon matrix. Histologic and clinical studies have demonstrated that rotator cuff tendinopathy involves collagen disorganization, altered cellularity, and matrix degeneration, rather than purely mechanical fiber discontinuity.6,14,24) Previous longitudinal imaging studies have shown that PTRCTs do not invariably progress and that structural progression is variable over time, with many lesions demonstrating relative stability on serial MRI.18,19,22,25) Furthermore, clinical outcomes and structural integrity are not always strongly correlated in rotator cuff pathology; patients may experience symptomatic improvement despite persistent or even recurrent structural defects.1,26,27) Nevertheless, the relative structural stability of some intratendinous partial-thickness tears does not imply that intervention is unnecessary in all patients. The patients included in the present study were symptomatic and underwent injection as an adjunctive conservative treatment option in routine clinical practice. In this context, the rationale for injectable treatment was not limited to achieving measurable structural healing on MRI, but also to modulating the biologic milieu of the degenerative intratendinous lesion, supporting symptom relief, and potentially reducing the risk of further tear progression. Thus, injection therapy may be considered in selected symptomatic patients even when the natural history of these lesions is not uniformly progressive. In this context, injectable biologic materials such as atelocollagen and gel-type ADM may function less as mechanical reinforcements and more as biologically active scaffolds. Collagen-based matrices have been shown to support cellular infiltration and matrix remodeling and to provide a structural framework that may facilitate tendon healing processes.6,10,13,28) Biologic injection strategies have also been investigated in the nonoperative management of partial-thickness tears, with randomized and meta-analytic data suggesting that healing-factor–based injections such as platelet-rich plasma may provide clinically meaningful symptomatic improvement, although imaging-defined structural changes remain inconsistent.29,30) Unlike patch augmentation techniques designed to reinforce tendon-to-bone constructs, injectable formulations are intended to modulate the intratendinous microenvironment in a minimally invasive manner. Although both atelocollagen and ADM are collagen-based biomaterials, their biological mechanisms of action may differ. Atelocollagen, a highly purified type I collagen with reduced antigenicity, has been shown to promote tendon healing by enhancing cellular proliferation and regulating inflammatory responses within the tendon microenvironment. Experimental studies have demonstrated increased tenocyte activity and improved tendon healing following atelocollagen application, while recent molecular studies suggest that atelocollagen may also influence collagen synthesis and inflammatory pathways.31,32,33,34) In contrast, ADM primarily functions as a preserved extracellular matrix scaffold that maintains native collagen architecture and provides a structural framework for cellular infiltration, neovascularization, and tissue remodeling. Previous studies have shown that ADM supports tenocyte attachment and matrix organization, and histologic analyses have demonstrated integration of ADM with host tissue in tendon repair settings.13,35) These differences suggest that although both materials may facilitate tendon healing through biologically active scaffold effects, their interactions with the tendon microenvironment may not be identical. This distinction is important when interpreting the comparable clinical outcomes observed in the present study. The relatively modest rate of MRI-defined structural improvement observed in the present study, contrasted with the substantial proportion of patients exceeding the MID in pain and functional scores, may therefore reflect the known dissociation between imaging findings and clinical symptoms in rotator cuff disease.1,24,26,27) These findings suggest that biologic injection therapy may influence symptomatic and functional recovery even when measurable reduction in tear dimensions is limited. In an arthroscopic setting, augmentation of rotator cuff repair can be technically demanding, often increasing operative time, preparation requirements, and overall cost. These factors may limit its widespread clinical applicability. In contrast, the present study evaluated a minimally invasive, injection-based approach for intratendinous partial-thickness tears and demonstrated that gel-type ADM yielded clinical and structural outcomes comparable to those of atelocollagen. This suggests that injectable biologic scaffolds may serve as a feasible nonoperative treatment alternative in selected patients. This study has several limitations. First, the retrospective, single-institution design may have introduced selection and information bias. Because the 2 treatments were used sequentially during different time periods rather than concurrently, residual confounding related to temporal differences cannot be excluded despite comparable baseline characteristics. In addition, a priori sample size calculation was not performed, and the relatively wide confidence interval for the between-group difference in MRI-based improvement rate indicates that small to moderate differences cannot be excluded. Therefore, the absence of statistically significant differences should not be interpreted as evidence of equivalence or noninferiority. Second, a sham injection or noninjection conservative control group was not included; therefore, the incremental benefit of biologic injection over standard conservative management cannot be determined, and causal interpretation of treatment efficacy is limited. Third, the findings are limited to intratendinous partial-thickness supraspinatus tears and may not be generalizable to other rotator cuff tear patterns. Fourth, the rate of MRI-defined structural improvement was modest, and longer follow-up with larger cohorts is needed to clarify the long-term structural behavior of these tears. Finally, structural outcomes were assessed using MRI rather than arthroscopic visualization; therefore, it remains uncertain whether MRI signal changes or reductions in tear size represent true tendon regeneration or reparative scar formation. Many studies have compared clinical outcomes following collagen application in patients who underwent rotator cuff tear surgery. However, there have been few studies investigating collagen as a nonsurgical treatment option for intratendinous partial tears. This study is particularly unique and significant because it compares 2 formulations—ADM and atelocollagen—which share low immunogenicity and favorable biocompatibility, while also focusing on the gel-type ADM rather than the more common patch-type. Additionally, by evaluating a specific indication, namely intratendinous partial tears, this study provides orthopedic surgeons with an expanded range of nonsurgical treatment options, further emphasizing its clinical value. Both gel-type ADM and type I atelocollagen injections showed significant clinical improvement in patients with intratendinous partial-thickness supraspinatus tears. MRI-based structural outcomes were generally maintained, and no statistically significant differences were observed between the 2 treatment groups at final follow-up. ACKNOWLEDGEMENTS The authors thank Ji-Eun Kim, MD, from the Department of Radiology, Yeouido St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, for her dedicated assistance with musculoskeletal magnetic resonance imaging evaluation. Footnotes CONFLICT OF INTEREST: No potential conflict of interest relevant to this article was reported. References