From Z-spectra to tissue maps: machine learning-based brain tissue classification in CEST MRI

Chemical Exchange Saturation Transfer (CEST) MRI enables noninvasive molecular imaging by selectively saturating exchangeable protons. Conventional tissue-specific CEST analysis often relies on projecting segmentations derived from structural MRI onto the CEST plane, introducing dependence on anatomical priors and additional registration steps. This proof-of-concept study investigates whether voxel-wise brain tissue identity can be inferred directly from CEST Z-spectral features, allowing tissue maps to be reconstructed in the native CEST acquisition space without anatomical input during inference. Eight classifiers spanning classical machine-learning methods and deep-learning architectures were trained and evaluated using 45-point voxel-wise spectra derived from B 0 -corrected, normalized, and spline-interpolated two-dimensional CEST data. Data were collected from six healthy subjects with 72 scans. Five subjects were used for subject-wise leave-one-subject-out cross-validation, while the sixth was reserved for independent held-out testing. Reference tissue labels for white matter, gray matter, and cerebrospinal fluid were generated from T 1 -weighted segmentations using FSL FAST and SPM12 and registered to the CEST plane. The evaluated classifiers comprised a one-dimensional Transformer and seven comparator models: SVM, Random Forest, FCNN, CNN, TCN-LSTM, ResNet, and Inception. On the held-out subject, the Transformer achieved a mean 88.48% accuracy against FSL labels and a mean 81.85% accuracy against SPM labels, with maximum per-scan accuracy of 96.88%. ROC analysis showed consistently high separability for cerebrospinal fluid and white matter across the evaluated classifiers, with AUCs of 0.97–0.98 and 0.95–0.96, respectively. Gray matter showed lower separability overall; nevertheless, the Transformer achieved the highest gray matter AUC of 0.92. In an unseen post-hemispherectomy case, the resection cavity was predominantly classified as cerebrospinal fluid, providing preliminary qualitative support for spectral-only voxel-wise tissue classification in substantially altered anatomy. These findings establish the feasibility of reconstructing voxel-wise WM, GM, and CSF maps directly from CEST Z-spectra and support further validation across larger, more diverse cohorts and acquisition settings toward reducing structural-to-CEST tissue-label transfer during inference.