Compositional MRI of cartilage typically exploits two principal biophysical mechanisms, dipolar cross-relaxation and chemical exchange 1, 2, to assess the integrity of the extracellular matrix (ECM). A two-pool model is often assumed to differentiate water compartments within the collagen-proteoglycan matrix 3. One component is the “bound” pool, referring to water molecules that interact strongly with ECM components, particularly collagens and proteoglycans, yielding short T2 values (< 10 ms). The other pool represents more mobile, bulk-like free water with T2 values typically in a range of 20–60 ms 4, 5. Dipolar cross-relaxation arises from through-space dipole–dipole interactions between water and immobile macromolecular protons and is prominent in collagen-rich regions with strong residual dipolar couplings. Chemical exchange reflects proton transfer between labile sites (e.g., OH, NH, and NH2 groups) and bulk water. In cartilage, exchangeable protons are abundant in the OH groups of glycosaminoglycans within proteoglycans. Collagen also contains exchangeable protons (OH from hydroxyproline and hydroxylysine, backbone amide NH, and side-chain amine NH2), but these are generally less accessible for chemical exchange-based imaging than the highly abundant hydroxyl groups of proteoglycans 6, 7. While dipolar cross-relaxation and chemical exchange always coexist and cannot be strictly separated mechanistically, MRI sequence design can selectively weight their relative contributions. Conventional off-resonance magnetization transfer (MT) imaging predominantly probes dipolar cross-relaxation, as the saturation pulse targets the broad spectral line of macromolecular protons, with collagen as the primary contributor in cartilage. In contrast, techniques such as R1ρ dispersion and chemical exchange saturation transfer (CEST) are designed to enhance sensitivity to chemical exchange, making them more responsive to changes in proteoglycan content. Although cartilage exhibits biexponential T2 (or T2*) relaxation, conventional MT imaging using spin-echo or gradient-echo sequences (TE ~1.5–20 ms) is often dominated by the long-T2 (or T2*) water component, reflecting superficial zone hydration or averaged bulk water content, while signal from tightly bound water in deep cartilage, calcified cartilage, or collagen-dense regions is markedly lost due to rapid T2 (or T2*) decay. In comparison, ultrashort echo time (UTE) sequences overcome this limitation by reducing TE to tens of microseconds, thereby capturing signal from both short- and long-T2* components 8. When combined with MT preparation (UTE-MT), this approach enables comprehensive interrogation of the macromolecular-bound proton pool throughout cartilage depth. In this issue of the Journal of Magnetic Resonance Imaging, Hu et al. applied UTE-MT and UTE-T2* imaging to tibiotalar cartilage before, immediately after, and 4 weeks following marathon running 9. The UTE-MT-derived magnetization transfer ratio (MTR) revealed a significant postmarathon decrease across most cartilage subregions, with partial recovery at 4 weeks, whereas UTE-T2* changes were limited and regionally inconsistent. This divergence underscores a fundamental biophysical distinction: UTE-T2* primarily reflects alterations in water content, collagen orientation, and susceptibility-related dephasing, while UTE-MT probes the macromolecular proton pool through the magnetization transfer pathway and is more directly sensitive to ECM integrity. The acute reduction in MTR observed shortly after marathon running likely reflects mechanically induced ECM perturbation, including transient collagen microstructural disorganization and/or proteoglycan redistribution, accompanied by an increased free water fraction. This shift may dilute the relative contribution of the macromolecular-bound water pool, thereby lowering the observed MTR through a reduction in the bound-to-free proton pool ratio. The partial recovery observed in selected subregions at 4 weeks aligns with the capacity for early, reversible cartilage adaptation to mechanical loading in the absence of overt structural damage. In this study by Hu et al., both UTE-MT and UTE-T2* demonstrated excellent repeatability and inter-rater reliability, with UTE-MT showing slightly superior consistency. Conventional T2 and T2*-based metrics are sensitive to residual dipolar couplings from highly organized collagen fibrils, producing orientation dependence relative to the main magnetic field (the “magic angle” effect). This orientation sensitivity complicates interpretation across subregions. While MT measurements are not entirely free of orientation effects, UTE-MT reduces signal loss from T2* decay and orientation-dependent dephasing, thereby mitigating magic angle-related confounds. By capturing short-T2 components and enabling more accurate modeling of magnetization transfer, UTE-MT provides greater reliability 10. The long-T2* component, arising from bulk water, dominates UTE-T2* contrast but contributes only indirectly to UTE-MT, which is primarily sensitive to the short-T2* macromolecular-bound proton pool. This biophysical distinction largely accounts for the observed methodological divergence. This feasibility study highlights UTE-MT as a promising biomarker for detecting reversible, premorphological cartilage stress responses that may precede structural degeneration detectable by conventional imaging. Its macromolecular sensitivity, reduced orientation dependence, and enhanced short-T2* capture enable UTE-MT to characterize early degeneration and monitor adaptive versus maladaptive responses to mechanical loading. Future longitudinal studies will be essential to determine whether such transient UTE-MT alterations predict long-term cartilage health. The authors declare no conflicts of interest.
Wang et al. (Fri,) studied this question.