Grading knee cartilage degeneration and meniscal injury with combined ACS-accelerated T1ρ and T2 mapping

Abstract

Objective

To evaluate the diagnostic performance of artificial intelligence–compressed sensing (ACS)-accelerated T1ρ mapping combined with conventional T2 mapping for quantitative differentiation of articular cartilage degeneration and meniscal injury severity in knee osteoarthritis (KOA).

Methods

We prospectively enrolled 56 KOA patients and 30 age-matched healthy volunteers between August 2025 and February 2026. All underwent 3T knee MRI using ACS T1ρ and conventional T2 mapping. Relaxation times were quantified in predefined cartilage and meniscal regions. Cartilage degeneration was classified by Kellgren–Lawrence (K–L) grade: 0 (normal), I–II (mild), and III–IV (severe). Meniscal injury was graded using the Stoller system: 0 (normal), I–II (mild), and III (severe). Group and subgroup comparisons used appropriate statistical tests.

Results

T1ρ and T2 values were significantly elevated in KOA versus controls across all regions (P < 0.05). Within KOA, relaxation times differed significantly between mild and severe subgroups (P < 0.05). For cartilage degeneration, AUCs were 0.948 (95% CI: 0.889–1.000) for ACS T1ρ, 0.867 (95% CI: 0.760–0.975) for T2, and 0.964 (95% CI: 0.922–1.000) for the combined model. For meniscal injury, AUCs were 0.851 (95% CI: 0.779–0.922) for ACS T1ρ, 0.802 (95% CI: 0.720–0.885) for T2, and 0.855 (95% CI: 0.784–0.926) for the combined approach.

Conclusion

Integrating ACS T1ρ and T2 mapping enables precise, quantitative stratification of cartilage and meniscal pathology across KOA severity levels, offering comprehensive biochemical assessment prior to morphological changes and highlighting its potential for early intervention and stage-based management.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12880-026-02501-1.

Background

Knee osteoarthritis (KOA) is a common chronic degenerative joint disorder characterized primarily by articular cartilage degeneration and secondary osteophyte formation [1, 2]. Its pathogenesis is influenced by multiple factors, including meniscal injury, osteochondral lesions, obesity, and physical activity [3, 4]. Due to their avascular nature and lack of nociceptors, both articular cartilage and the meniscus are often damaged without eliciting symptoms until adjacent tissues become involved [5, 6]. Moreover, no currently available therapy can reliably regenerate cartilage once advanced degeneration has occurred [7].

Conventional MRI sequences can clearly depict morphological alterations in these tissues—such as thickening, tearing, and avulsion—but are unable to provide quantitative assessment that correlates with these structural changes or detect early biochemical pathology [8]. Consequently, specialized quantitative mapping techniques for articular cartilage and the meniscus, namely T1ρ and T2 mapping, have been increasingly adopted in clinical research [9, 10]. Extensive studies have demonstrated that T2 relaxation time is primarily influenced by two factors: intra-cartilaginous water content and the organization of the collagen fiber network. Disruption of the extracellular collagen matrix or increased water content within the cartilage matrix leads to prolonged T2 relaxation times [11, 12]. In contrast, T1ρ (spin-lattice relaxation in the rotating frame) imaging is highly sensitive to interactions between water molecules and macromolecules such as proteoglycans [13], thereby enabling quantitative evaluation of early biochemical degeneration in cartilage. Nevertheless, the knee T1ρ imaging protocol recommended by the Radiological Society of North America (RSNA) typically requires 6 to 12 min of acquisition time, which has limited its routine clinical implementation. Artificial intelligence–compressed sensing (ACS) acceleration has been widely applied to various anatomical MRI sequences [14, 15]; however, its integration with quantitative T1ρ mapping remains limited. In this study, we employed an ACS-accelerated T1ρ mapping sequence in combination with conventional T2 mapping to quantitatively assess meniscal injury and articular cartilage degeneration associated with knee osteoarthritis (KOA). Our aim was to evaluate the clinical utility of this combined approach and provide an objective imaging basis for the early diagnosis and stage-based management of KOA.

Materials and methods

This prospective study was approved by the Institutional Review Board of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (Approval No. UH-2025-0802). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Clinical data

Between August 2025 and February 2026, 56 patients diagnosed with knee osteoarthritis (KOA) and 30 age-matched healthy volunteers were consecutively enrolled at this institution. The KOA group included 23 males and 33 females, with an age range of 25–82 years (mean ± standard deviation: 50.0 ± 12.0 years). Body mass index (BMI) ranged from 18.5 to 30.0 kg/m² (mean ± standard deviation: 24.1 ± 2.8 kg/m²). In the control group, there were 15 males and 15 females, with an age range of 30–77 years (mean ± standard deviation: 48.0 ± 11.0 years). BMI ranged from 18.3 to 31.0 kg/m² (mean ± standard deviation: 23.6 ± 3.1 kg/m²) (Table 1).

Table 1.

Baseline characteristics of study participants

Characteristic	Control group (n = 30)	Experimental group (n = 56)	P value
Sex, male/female	15/15	23/33	0.43
Age (years)	48.0 ± 11.0	50.0 ± 12.0	0.40
Height (cm)	166.0 ± 7.6	163.6 ± 7.5	0.17
Weight (kg)	66.3 ± 12.8	64.6 ± 9.4	0.49
BMI (kg/m²)	23.6 ± 3.1	24.1 ± 2.8	0.74

Data are presented as mean ± standard deviation or n/n for sex. P values were calculated using the chi-square test for categorical variables and independent samples t-tests for continuous variables

Inclusion criteria for knee osteoarthritis (KOA) were: (1) age ≥ 35 years; (2) recurrent knee pain within the past month; (3) morning stiffness ≤ 30 min; (4) crepitus during active joint motion; and (5) standing or weight-bearing radiographs showing joint space narrowing, marginal osteophytes, or subchondral cystic changes. Patients meeting any two or more criteria were diagnosed with KOA.

Exclusion criteria for all participants were: (1) history of knee trauma, arthroscopy, or surgery; (2) systemic corticosteroid therapy within the preceding three months; (3) inflammatory or autoimmune arthropathies (e.g., rheumatoid arthritis, ankylosing spondylitis, reactive arthritis, psoriatic arthritis); (4) infectious, tuberculous, hematologic, or metabolic arthropathies; (5) ligamentous injury of any etiology, including post-repair states; and (6) severe image artifacts leading to unreliable or distorted T1ρ/T2 relaxation time measurements.

Imaging protocol and methodology

Imaging equipment and parameters

All knee MRI examinations were performed using a United Imaging uMR 880 3.0T MR scanner equipped with a dedicated 24-channel knee coil. Sagittal T2 mapping was acquired using a multi-echo spin-echo (SE) sequence with five echoes. The imaging parameters were as follows: repetition time (TR) = 2100 ms; echo times (TEs) = 14.5, 29.0, 43.5, 58.0, and 72.5 ms; slice thickness = 3.0 mm; slice gap = 0.3 mm; field of view (FOV) = 16 cm × 16 cm; and reconstruction matrix = 360 × 324. Sagittal T1ρ mapping was based on a fast spin-echo (FSE) sequence accelerated by artificial intelligence–compressed sensing (ACS) with a combined acceleration factor of 3.63. The imaging parameters were: TR = 900 ms; effective TE = 45 ms; slice thickness = 3.0 mm; slice gap = 0.3 mm; FOV = 16 cm × 16 cm; reconstruction matrix = 360 × 324; spin-lock frequency = 500 Hz; and spin-lock times (TSLs) = 0, 30, and 50 ms. Prior to the main study, a pilot validation experiment confirmed that the ACS-accelerated T1ρ mapping yielded quantitative results consistent with those of conventional T1ρ mapping. (Mean difference: 0.8; ICC: 1.000). Full details of the statistical analysis, including the Bland-Altman plot, are available in the Supplementary Material. Therefore, only the accelerated sequence was used in the subsequent patient and control cohorts.

Quantitative analysis

Quantitative T1ρ and T2 relaxation times were calculated using a mono-exponential decay model fitted to the acquired data.For T1ρ mapping (Eq. 1), the signal intensity ( S) was fitted as a function of spin-lock time (TSL):

1

For T2 mapping (Eq. 2), the signal intensity was fitted as a function of echo time (TE):

2

Where () represents the initial signal intensity. The fitting algorithm was performed using MMfusion software to generate the parametric maps. This methodology is consistent with established practices in quantitative cartilage imaging [16].

Image post-processing

Two experienced radiologists, blinded to clinical information, independently graded the severity of knee osteoarthritis (KOA) according to the Kellgren–Lawrence (K–L) criteria [17]. Interobserver agreement for quantitative measurements was excellent, with an Intraclass Correlation Coefficient (ICC) of 0.88 for T1ρ and T2 relaxation times, as well as for image grading (Kappa = 0.85). Additionally, intraobserver variability analysis for the primary radiologist demonstrated high consistency (ICC = 0.92). Discrepant assessments were resolved through consensus discussion. Cases in which consensus could not be reached were excluded from the study. The acquired ACS T1ρ and T2 mapping datasets were transferred to a United Imaging uOmnispace post-processing workstation. Using MMfusion software, the parametric maps were co-registered with proton density fat-saturated fast spin-echo (PD FS FSE) images to generate corresponding color-coded maps.Regions of interest (ROIs) of consistent size were manually placed on predefined articular cartilage and meniscal regions. Care was taken to maintain uniform ROI areas across measurements. For cartilage, ROIs were positioned away from the cartilage–synovial fluid interface to minimize partial volume effects. For the menisci, ROIs were placed to avoid adjacent cortical bone, articular cartilage, and joint fluid.The following anatomical regions were evaluated [18]: medial femoral condyle (MFC), medial tibial plateau (MTP), medial facet of the patella, lateral femoral condyle (LFC), lateral tibial plateau (LTP), anterior horn of the medial meniscus (AHM), posterior horn of the medial meniscus (PHM), and corresponding regions of the lateral meniscus (Fig. 1).

Fig. 1.

Region-of-interest (ROI) delineation methodology for articular cartilage and menisci on quantitative MRI maps. a, b: Meniscal ROIs were manually drawn on co-registered ACS T1ρ (a) and T2 (b) parametric maps fused with proton density–weighted (PDw) fast spin-echo images to ensure anatomical accuracy. c, d: Articular cartilage ROIs were similarly delineated on fused T1ρ (c) and T2 (d) maps with PDw images, covering the full thickness of weight-bearing regions in the femoral condyles, tibial plateaus, and patellar facets. All segmentations were performed by a musculoskeletal radiologist blinded to clinical data

Grading of KOA severity

The severity of knee osteoarthritis (KOA) was graded according to the Kellgren–Lawrence (K–L) criteria based on digital radiography (DR) images acquired during the same clinical visit as the MRI examination. For the purpose of statistical analysis, knees were categorized into three groups: Grade 0 (normal), Grades I–II (mild KOA), and Grades III–IV (severe KOA).

Grading of meniscal injury severity

Meniscal injury was graded using the Stoller classification system. The grades were defined as follows:

Grade 0: Normal meniscus, demonstrating uniform low signal intensity and regular morphology on all MRI sequences.

Grade I: A focal, globular, or irregular area of increased intrameniscal signal that does not extend to either articular surface.

Grade II: A linear band of increased intrameniscal signal extending to the capsular margin but not communicating with any articular surface.

Grade III: A linear area of increased signal intensity that communicates with at least one articular surface of the meniscus.

For analysis, Grade 0 was classified as normal, Grades I–II as mild meniscal injury, and Grade III as severe meniscal injury.

Statistical analysis

Statistical analyses were performed using SPSS software (version 26.0; IBM Corp., Armonk, NY, USA). Categorical data were compared using the chi-square test. Continuous data were expressed as mean ± standard deviation (mean ± SD). The paired-samples t-test was used for within-group comparisons of continuous variables, while the independent-samples t-test was applied for between-group comparisons. For comparisons among multiple groups, one-way analysis of variance (ANOVA) was conducted, followed by post hoc pairwise comparisons using Tamhane’s T2 test. A two-tailed P value < 0.05 was considered statistically significant.

Binary logistic regression incorporating T1ρ and T2 relaxation times was performed to predict KOA severity (mild vs. severe), with the resulting predicted probabilities utilized to construct the combined ROC curve.

Receiver operating characteristic (ROC) curve analysis was performed to evaluate the efficacy of T1ρ values, T2 values, and their combination in differentiating the severity of knee cartilage degeneration and meniscal injury. The area under the ROC curve (AUC) was calculated as a measure of diagnostic accuracy.

Results

Comparison of T1ρ values in articular cartilage between groups

The T1ρ values in the medial femoral condyle (MFC), medial tibial plateau (MTP), medial facet of the patella, lateral femoral condyle (LFC), and lateral tibial plateau (LTP) were significantly higher in the observation group than in the control group (P < 0.05). Within the observation group, the T1ρ values in these five regions were significantly higher in the severe degeneration subgroup than in the mild degeneration subgroup (P < 0.05) (Table 2).

Table 2.

Comparison of T1ρ relaxation times (ms) between control and knee osteoarthritis (KOA) groups across articular cartilage and meniscal regions

Group	n	MTC	MTP	MFP	LFC	LTP	AHM	PHM
Control	30	36.34±4.04	36.26±4.51	38.09±4.79	37.26±4.64	37.22±4.63	17.48±11.55	21.66±11.22
KOA	56	42.50±7.26	40.52±5.00	42.08±7.40	44.83±5.94	41.87±6.73	28.56±10.13	30.75±10.08
P value	-	<0.05	<0.05	<0.05	<0.05	<0.05	<0.05	<0.05
Mild degeneration	17	34.88±4.05	37.81±4.90	35.77±5.40	40.99±4.95	35.87±4.75	-	-
Severe degeneration	39	45.82±5.66	41.70±4.60	44.84±6.47	46.50±5.58	44.48±5.74	-	-
P value	-	<0.05	<0.05	<0.05	<0.05	<0.05	-	-

MFC, Medial Femoral Condyle; MTP, Medial Tibial Plateau; MFP, Medial Facet of the Patella; LFC, Lateral Femoral Condyle; LTP, Lateral Tibial Plateau; AHM, Anterior Horn of the Medial Meniscus; PHM, Posterior Horn of the Medial Meniscus

Comparison of T2 values in articular cartilage between groups

The T2 values in the MFC, MTP, medial facet of the patella, LFC, and LTP were significantly higher in the observation group than in the control group (P < 0.05). Within the observation group, the T2 values in these regions were significantly higher in the severe degeneration subgroup compared to the mild degeneration subgroup (P < 0.05) (Table 3).

Table 3.

Comparison of T2 relaxation times (ms) between control and knee osteoarthritis (KOA) groups across articular cartilage and meniscal regions

Group	n	MTC	MTP	MFP	LFC	LTP	AHM	PHM
Control	30	35.07±2.80	34.46±2.92	33.08±3.61	35.95±3.39	31.77±3.69	7.89±7.52	7.30±6.20
KOA	56	41.91±7.87	39.40±8.67	38.79±8.43	42.06±8.78	37.73±8.58	16.66±9.14	16.35±9.27
P value	-	<0.05	<0.05	<0.05	<0.05	<0.05	<0.05	<0.05
Mild degeneration	17	36.77±5.55	33.75±6.00	33.34±5.39	35.76±4.07	33.22±6.13	-	-
Severe degeneration	39	44.15±7.73	41.86±8.55	41.17±8.46	44.81±8.89	39.69±8.81	-	-
P value	-	<0.05	<0.05	<0.05	<0.05	<0.05	-	-

Comparison of T1ρ and T2 values in the meniscus between groups

The T1ρ and T2 values in both the anterior and posterior horns of the meniscus were significantly higher in the observation group than in the control group (P < 0.05). Within the observation group, a total of 112 menisci were classified according to the aforementioned criteria, resulting in 56 menisci with mild injury and 56 with severe injury. Both T1ρ and T2 values were significantly higher in the severe injury subgroup than in the mild injury subgroup (P < 0.05) (Table 4).

Table 4.

Comparison of T1ρ and T2 relaxation times (ms) in menisci between mild and severe injury groups

Group	n (menisci)	T1ρ (ms)	T2 (ms)
Mild injury	56	23.78±0.67	11.70±8.51
Severe injury	56	35.54±6.54	21.30±7.07
P value	-	<0.05	<0.05

Diagnostic efficacy of the two sequences for cartilage degeneration and meniscal injury severity

In the assessment of cartilage degeneration severity, ACS T1ρ mapping demonstrated an AUC of 0.948 (95% CI: 0.889–1.000) with a sensitivity of 92.3% and specificity of 88.2%, outperforming T2 mapping (AUC = 0.867, 95% CI: 0.760–0.975; sensitivity = 79.5%; specificity = 94.1%). The combination of both sequences yielded an AUC of 0.964 (95% CI: 0.922–1.000), indicating superior diagnostic performance (Fig. 2).

Fig. 2.

Receiver operating characteristic (ROC) curves for the diagnostic performance of ACS T1ρ mapping, T2 mapping, and their combined model in assessing the severity of articular cartilage degeneration in patients with knee osteoarthritis (KOA). AUCs were 0.948 (95% CI: 0.889–1.000) for ACS T1ρ, 0.867 (95% CI: 0.760–0.975) for T2 mapping, and 0.964 (95% CI: 0.922–1.000) for the combined model. Shaded areas represent 95% confidence intervals.

For meniscal injury grading, ACS T1ρ and T2 mapping showed comparable sensitivities (both 91.1%) but differed in specificity: ACS T1ρ had a specificity of 69.6%, slightly higher than that of T2 mapping (62.5%). Their respective AUCs were 0.851 (95% CI: 0.779–0.922) for ACS T1ρ and 0.802 (95% CI: 0.720–0.885) for T2 mapping. The combined model further improved diagnostic accuracy, achieving an AUC of 0.855 (95% CI: 0.784–0.926) (Fig. 3).

Fig. 3.

Receiver operating characteristic (ROC) curves comparing the diagnostic accuracy of ACS T1ρ mapping, T2 mapping, and the combined model for grading meniscal injury severity in KOA patients. AUC values were 0.851 (95% CI: 0.779–0.922) for ACS T1ρ, 0.802 (95% CI: 0.720–0.885) for T2 mapping, and 0.855 (95% CI: 0.784–0.926) for the combined model. Shaded areas represent 95% confidence intervals

Discussion

Our study demonstrates that the combined application of ACS-accelerated T1ρ and conventional T2 mapping provides a robust method for quantitatively differentiating the severity of knee osteoarthritis (KOA) [19, 20]. We found significantly elevated T1ρ and T2 relaxation times in both articular cartilage and menisci of KOA patients compared to healthy volunteers, with values showing a progressive increase corresponding to the worsening of cartilage degeneration and meniscal injury. Importantly, the integrated diagnostic model leveraging both T1ρ and T2 parameters exhibited the highest accuracy, outperforming either sequence alone in distinguishing between mild and severe pathological stages [21, 22].

Furthermore, studies have demonstrated a strong association between meniscal injury and KOA [23]. The meniscus shares a similar biochemical composition with articular cartilage, differing mainly in the relative proportions of its components [24, 25]. Research by Krishna et al. [26]. showed that during knee joint degeneration, the extracellular matrix in both the meniscus and articular cartilage undergoes comparable changes—particularly in the degeneration and depletion of proteoglycans. Critically, these biochemical alterations do not produce detectable morphological changes in either the cartilage or meniscus during the early stages of disease and are therefore invisible on conventional MRI [27, 28]. Even in intermediate and advanced stages, radiographic or MRI-based grading remains subjective and imprecise, heavily dependent on the interpreter’s experience. Consequently, there is a clear clinical need for precise, quantitative biomarkers to support objective assessment and guide individualized treatment strategies [29, 30].

Quantitative assessment of articular cartilage and the meniscus can be achieved using T1ρ and T2 mapping techniques. Measurement of T1ρ and T2 relaxation times provides a comprehensive reflection of the tissue status of knee cartilage and meniscus. Consistent with this, our findings demonstrated statistically significant differences in T1ρ and T2 values between patients with knee osteoarthritis (KOA) and healthy controls [22]. Furthermore, within the KOA group, T1ρ and T2 values progressively increased with the severity of cartilage degeneration and meniscal injury, showing significant differences between mild and severe injury subgroups—results that align with previous studies [31].The pathophysiological basis for these observations lies in the biochemical alterations characteristic of KOA: a reduction in proteoglycan content leads to elevated T1ρ values, while disorganization of the collagen network increases the mobility of water protons, thereby raising T2 values. Because these processes occur concurrently, the simultaneous application of T1ρ and T2 mapping enables a multi-dimensional evaluation of cartilage and meniscal integrity, facilitating more precise grading of KOA severity [32, 33]. We acknowledge the absence of direct histopathological correlation; however, the link between T1ρ/T2 values and specific cartilage alterations is firmly grounded in previous in vitro and animal studies. As invasive confirmation is impractical in routine clinical settings, our work bridges this gap by applying these validated principles to establish a robust, non-invasive imaging tool. conventional T1ρ mapping is limited by its prolonged acquisition time, which stems from the need for multiple spin-lock time (TSL) acquisitions [34, 35]. This limitation motivated our use of accelerated compressed sensing (ACS) technology. The integration of artificial intelligence (AI) substantially facilitates the simultaneous quantitative acquisition of both sequences.In our study, the ACS-accelerated T1ρ sequence required approximately 3 min for acquisition—representing at least a 50% reduction. Critically, the reconstruction is performed in under a minute on the scanner’s built-in workstation, where fused images can be directly viewed, delineated, and measured within the scanner’s View software, adding no extra post-processing time and demonstrating high feasibility for routine clinical integration. Beyond this efficiency, ACS offers distinct technical advantages. By integrating an AI module trained on extensive datasets as a reconstruction constraint, it effectively suppresses noise and artifacts common in traditional compressed sensing. Crucially, its normalization process preserves pixel value integrity during iterative fusion, ensuring high quantitative fidelity for T1ρ measurements without compromising diagnostic accuracy.

Regarding diagnostic performance, our results showed that, for assessing cartilage degeneration severity, the T1ρ sequence yielded a significantly higher area under the curve (AUC) and sensitivity than T2 mapping, although T2 mapping exhibited higher specificity. The combination of both sequences further improved diagnostic accuracy, as reflected by an even higher AUC. In evaluating meniscal injury severity, both sequences achieved AUCs greater than 0.8, with T1ρ showing a slightly higher AUC than T2 mapping; again, their combination resulted in the highest AUC. These findings can be explained by the distinct pathological sensitivities of the two techniques. In the natural history of KOA, proteoglycan loss precedes alterations in water distribution and collagen architecture. Consequently, T1ρ mapping can detect the earliest biochemical changes of osteoarthritis, which likely accounts for its higher sensitivity in our study [36]. In contrast, the high specificity of T2 mapping may be attributed to its strong dependence on the relatively stable high water content in cartilage and meniscus. Moreover, structural heterogeneity across different load-bearing regions of cartilage and meniscus in KOA patients leads to regional variations in collagen and proteoglycan content and distribution. As a result, T1ρ and T2 values often exhibit distinct spatial patterns. Relying on a single sequence may therefore overlook critical biochemical alterations, potentially leading to inaccurate assessment. By contrast, combined T1ρ and T2 mapping provides complementary information, enabling a more robust and accurate quantitative evaluation.

Thus, the simultaneous application of both sequences for quantitative assessment in early-stage KOA offers superior diagnostic accuracy compared to either sequence alone—a conclusion fully supported by our findings. Multi-dimensional biochemical profiling allows for earlier and more precise detection of disease progression. Furthermore, timely therapeutic intervention targeting both articular cartilage and the meniscus may delay the onset and progression of KOA, prevent further structural damage, and ultimately contribute to greater pain relief and improved quality of life for patients.

Limitations and future directions

This study has several limitations. First, We recognize that excluding patients with comorbidities such as ligament injuries, synovitis, and subchondral lesions ensures dataset homogeneity but limits the applicability of our results to real-world KOA populations with mixed pathologies. Second, the sample size was relatively small, particularly in the subgroup analyses comparing mild versus severe degeneration. This limited sample size, combined with the absence of an external validation cohort, may pose a risk of overestimating the diagnostic performance. Third, regions of interest (ROIs) were delineated across the full thickness of cartilage without stratification into the superficial and deep zones. Layer-specific analysis may offer greater sensitivity to early biochemical changes and thus enable a more precise assessment of knee osteoarthritis. Future studies should involve larger, multi-center cohorts to validate the current findings. Moreover, systematic investigations are needed to evaluate how different ACS acceleration factors and varying ROI delineation protocols influence T1ρ and T2 relaxation values, as well as their diagnostic accuracy in grading cartilage degeneration and meniscal injury. Such work would establish a more robust, standardized quantitative imaging framework for clinical application.

Conclusion

In conclusion, this study demonstrates that the combined use of ACS-accelerated T1ρ and T2 mapping sequences enables quantitative differentiation of the severity of both articular cartilage degeneration and meniscal injury in patients with knee osteoarthritis (KOA). Each sequence exhibits complementary strengths: T1ρ mapping shows higher sensitivity, while T2 mapping offers greater specificity. Their integration yields superior diagnostic efficacy compared to either sequence alone. Therefore, the simultaneous application of ACS T1ρ and T2 mapping provides intuitive, non-invasive quantitative biomarkers that support early diagnosis and stratified management of KOA, highlighting its considerable potential for clinical translation.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 ^{(35.6KB, docx)}

Abbreviations

OA

Osteoarthritis

KOA

Knee osteoarthritis

BMI

Body mass index

MRI

Magnetic resonance imaging

ROI

Region of interest

ACS

Artificial intelligence–compressed sensing

K-L

Kellgren–Lawrence grading system

MFC

Medial femoral condyle

MTP

Medial tibial plateau

Author contributions

SY and WF conceived and designed the study. SY, and ZL performed the data collection and MRI scanning. SY and ZL conducted the image post-processing and statistical analysis. SY drafted the manuscript. WF critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript.

Funding

This work was supported by the Natural Science Foundation of Hubei Province (2026AFB659,2026AFC1403).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was reviewed and approved by the Institutional Review Board of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (Approval No. 2025 − 0802). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to their inclusion in the study.

Consent for publication

The consent for publication has been obtained from all participants.

Competing interests

The authors declare no competing interests.