Brain Alterations Detected by Multiplex MR Imaging in Nasopharyngeal Carcinoma Patients After Treatment

Nasopharyngeal carcinoma (NPC) is one of the most common malignancies in the head and neck region. Radiotherapy is the primary treatment for NPC patients, particularly those with T3 and T4 stages. However, radiotherapy often induces brain changes during the acute phase, early delayed phase, or late delayed phase, which may even lead to temporal lobe brain injury and subsequent cognitive impairment. Conventional magnetic resonance imaging (MRI) typically fails to detect corresponding radiological alterations in the brain regions affected by these delayed reactions, even when cognitive dysfunction manifests in patients as early as 6 months post-radiotherapy, so as the other treatment. Previous structural MRI studies (e.g., diffusion tensor imaging [DTI], voxel-based morphometry [VBM]) and resting-state functional MRI analyses (e.g., regional homogeneity [ReHo], functional connectivity [FC], graph theory-based methods) have demonstrated that subtle structural changes and functional abnormalities can be observed prior to apparent radiological changes in NPC patients post-radiotherapy. Additionally, radiomics models based on machine learning have shown promise in predicting early radiation-induced brain injury. However, conventional MRI sequences provide limited imaging information, involve cumbersome and complex data processing, and cannot simultaneously integrate multiple imaging modalities or analyze diverse parameters in a single scan. Therefore, developing a novel MRI imaging technique capable of rapidly, simply, and efficiently detecting induction/radiation-induced brain injury or necrosis in NPC patients during the early stages remains a critical challenge for improving survival outcomes and prognosis in clinical practice.

Multiparametric quantitative magnetic resonance imaging (MulTipLex, MTP) is an emerging technique that enables simultaneous acquisition of multiple contrast mechanisms and quantitative physicochemical parameters in a single scan. This technology has been widely applied in the diagnosis and quantitative analysis of neurological disorders such as Alzheimer's disease, Parkinson's disease, temporal lobe epilepsy, and tuberous sclerosis. It has also been utilized for prognostic prediction of cerebral infarction and integrated with multimodal imaging (e.g., PET-MRI, nuclear medicine) to assist in the staging of nasopharyngeal carcinoma, prostate cancer, and cervical cancer. Notably, MTP offers rapid imaging acquisition, mature technical workflows for brain segmentation and quantitative analysis, and comprehensive data fusion capabilities. To date, no studies have investigated the application of MTP in evaluating induction/radiation-induced cognitive dysfunction following NPC before, during, and/or after radiotherapy.

All patients underwent MTP imaging and Montreal Cognitive Assessment (MoCA) tests between different treatment courses.All MRI data were acquired using a 3.0 T MRI scanner (uMRI 790; United Imaging Healthcare, Shanghai, China) equipped with a 32-channel phased-array head coil. A single MTP scan allowing for the acquisition of multiple contrasts, including aT1W, SWI, cPDW, cT1w, T2Star, T1Map, PDMap, R2Star, and quantitative susceptibility mapping (QSM), was collected for each subjec.The MTP data were analyzed using BrainTool software. Then a single-group paired t-test, Pearson's correlations,a two-sample t-test, Receiver operating characteristic curve (ROC) analyses were utilized to analyse those MTP data and clinical data.

We hypothesis the MTP MR sequence can effectively evaluate brain alterations and cognitive impairment in NPC patients following their teratment.