Although ZTE sequences have been increasingly applied in musculoskeletal imaging, their systematic use in the evaluation of diabetic foot complications has not been previously investigated. This study demonstrates that the implementation of a ZTE MRI sequence significantly enhances the detection of clinically relevant osseous alterations in diabetic foot-related disease, with high diagnostic performance even under variable image quality conditions.

Our results are consistent with previous literature that has established the superior performance of ZTE for cortical bone imaging and the detection of subtle osseous pathology compared to conventional MRI sequences [11,12,13]. The ability to reconstruct isotropic, high-resolution images in multiple planes was particularly beneficial for the evaluation of joint dislocations, a key complication in Charcot neuro-osteoarthropathy.

In the evaluation of pneumatosis, the typical ZTE artifact of air hyperintensity proved advantageous, allowing better visualization of sinus tracts extending from skin ulcers, showing how technical features generally considered artifacts may facilitate the detection of clinically relevant findings.

The Likert analysis of image quality revealed that most ZTE examinations achieved high scores, indicating excellent depiction of anatomical structures. Even in the minority of cases rated as low quality (Likert ≤ 2), ZTE sequence remained diagnostically valuable, detecting additional bone alterations—such as erosions, bone pneumatosis, and fragmentation—not visible on standard protocol MRI images.

Comparison with CT as a reference standard for the detection of osseous alterations [8, 10, 20,21,22] confirmed the high sensitivity and specificity of ZTE for detecting bone erosions, bone exposure, and periosteal reaction, with diagnostic accuracy reaching 100% for these findings in our cohort. Performance for bone fragmentation and sclerosis was slightly lower, likely reflecting the challenges inherent to detecting minute fragments or chronic sclerotic changes, which could also be related to the thinner slice thickness that was available for CT (0.625 mm vs. 1.25 mm for ZTE).

ZTE’s short acquisition times are especially relevant in diabetic patients, who are often elderly, may have limited mobility and discomfort repositioning for an additional CT scan.

Furthermore, avoiding ionizing radiation exposure may benefit patients undergoing serial follow-up examinations, as repeated CT scans could otherwise result in a clinically significant cumulative radiation exposure over time [23,24,25].

ZTE also demonstrated excellent interobserver agreement, with Cohen’s kappa values indicating substantial to almost perfect concordance among expert readers for all key alterations. Moreover, moderate to excellent concordance was found between the expert radiologist and the radiology resident. This high reproducibility, observed across readers with varying levels of experience, supports the robustness of ZTE assessment in diverse radiological settings.

The study has several limitations, including its retrospective design, single-centre setting, and relatively small sample size. Furthermore, the absence of systematic histological confirmation precludes absolute diagnostic certainty, given it was not routinely performed in our cohort due to its invasive nature and limited feasibility in clinical practice. Instead, diagnoses were established through a composite reference standard integrating MRI findings, clinical assessment, laboratory data, and follow-up, mirroring real-life diagnostic pathways but introducing a degree of uncertainty. Although ZTE is conceptually vendor-agnostic, its clinical implementations are currently proprietary and may require dedicated software packages and compatible system configurations. Equivalent short-TE or zero-TE sequences are available on most major MRI platforms under vendor-specific names; however, acquisition parameters and reconstruction algorithms may differ. This inter-vendor variability may influence reproducibility and represents a potential limitation. Future prospective studies with larger, multicentre cohorts and standardized reference standards, ideally including histopathology, are warranted to confirm and expand upon these findings. Moreover, future developments may include the integration of artificial intelligence tools to optimize ZTE acquisition, potentially reducing scan times and further enhancing image quality and diagnostic performance.

In summary, ZTE MRI significantly improves the detection of osseous alterations in diabetic foot, even under suboptimal imaging conditions. Its consistently high image quality, ability to reveal additional findings, and strong inter-reader concordance support its integration into standard MRI protocols for bone assessment.