Development and Clinical Translation of RPBM for Quantitative Assessment of Myofascial Pain

# Development and Clinical Translation of RPBM for Quantitative Assessment of Myofascial Pain ## Overview This dataset contains multimodal quantitative MRI (qMRI) data acquired to characterize masticatory myofascial pain syndrome (MMPS). MMPS is a highly prevalent pain condition involving the masticatory muscles that currently lacks validated, objective imaging biomarkers for clinical diagnosis and monitoring. ## Study design and participants A total of 90 participants (ages 18-65) were enrolled and divided into two groups: a control (asymptomatic) group and an MMPS (combined latent/active myofascial pain) group. Patients were recruited in part from the Division of Oral and Maxillofacial Surgery at Weill Cornell Medicine (WCM), supplemented by community recruitment near WCM. The study was approved by the Institutional Review Board of Weill Cornell Medicine (IRB# 24-07027694). All participants provided written informed consent prior to enrollment. **Inclusion criteria:** - Men and women 18-65 years of age - Ability to provide informed consent - Symptoms meeting criteria for the control or MMPS (latent or active) group - For the MMPS group: myofascial pain of the masticatory muscles per DC/TMD criteria -- present or historical pain in the jaw, temple, ear, or preauricular region, modified by jaw movement, function, or parafunction **Exclusion criteria:** - Standard MRI contraindications (cardiac pacemaker/pacing wires, intracranial clips, metal implants, external metallic devices within 10 mm of the head, suspected or confirmed intraocular metal, severe obesity or inability to lie supine, claustrophobia) - Pregnancy or breastfeeding - Traumatic injury of the masticatory muscles or temporomandibular joint within the prior 12 months - Mandibular fracture within the prior 12 months - Initiation of additional myofascial pain treatment within the prior month - Masseter botulinum toxin injection within the prior 4 months **Data completeness:** 6 enrolled participants (sub-2010, sub-2027, sub-2028, sub-2049, sub-2055, sub-2062) were excluded due to no-show or failed scan and are not included in this release -- participant ID numbering is therefore non-contiguous by design, not a data error. ## MMPS classification and trigger-point identification Participants were classified into the control or MMPS group by a board-certified oral and maxillofacial surgeon prior to MRI scanning, based on the Visual Analog Scale (VAS) of pain, the Jaw Functional Limitation Scale (JFLS), and the presence of myofascial trigger points on clinical exam. Enrolled participants completed a symptom questionnaire, then underwent a clinical exam by the surgeon to assess for trigger points. Where a trigger point was identified, a fiducial marker was placed on the skin at that location prior to scanning so that its position could be localized in the MR images. ## Imaging protocol All data were acquired on 3T Siemens Prisma scanners at Weill Cornell Medicine (WCM). All sequences were acquired in the coronal plane, with the exception of the T1w MPRAGE, which was originally acquired as a full-FOV sagittal 3D volume. The T1w file included in this release has been reformatted to the coronal plane and cropped to a reduced slice range matching the DWI coverage. Whole-muscle regions of interest were delineated in the masticatory muscles -- right and left masseter, temporalis, medial pterygoid, and lateral pterygoid -- using deep-learning-based segmentation (nnU-Net). **No ROI/segmentation files are included in this data release** -- only the source imaging data described below. Sequences included in this release: - **T1-weighted MPRAGE** -- anatomical reference and ROI registration target. - **Time-dependent diffusion tensor imaging (DTI)** -- diffusion times of 22-300 ms, used to estimate RPBM parameters (muscle fiber diameter, sarcolemma permeability, sarcomere length). The 22 ms diffusion time was acquired with pulsed-gradient spin-echo (PGSE); the remaining diffusion times (42-300 ms) were acquired with stimulated echo acquisition mode (STEAM). - **Multi-b-value diffusion-weighted imaging (IVIM)** -- three orthogonal directions, b = 0, 40, 70, 110, 200, 300, 600 s/mm^2, acquired with both flow-compensated and monopolar diffusion encoding, used to estimate tissue diffusivity (D), pseudo-diffusivity (D*), and perfusion fraction (f_p). - **T2 mapping** -- 16-echo CPMG acquisition (TE = 7.7-123.2 ms), stored as raw multi-echo data (BIDS `MESE` suffix) rather than a pre-computed T2 map. - **T2-star/proton density fat fraction** -- 6-echo spiral UTE sequence (TE = 0.05-7.85 ms), magnitude and phase, stored as raw multi-echo data (BIDS `MEGRE` suffix, `part-mag`/`part-phase`) rather than pre-computed maps. Diffusion acquisitions (DTI and IVIM) and used a head-foot (H>F) phase-encoding direction. Not every subject has every modality -- a missing `T2`, `UTE`, `dwi` diffusion-time series, or IVIM acquisition for a given subject reflects that sequence failing or not being acquired for that subject, not a processing error. ## De-identification All data were anonymized by IDEAL. To ensure complete de-identification of imaging data for public repository submission, voxels containing potentially identifiable facial features -- including the nasal cavity, teeth, and ears -- were removed (zeroed out) from the imaging volume prior to data sharing. As an additional de-identification measure for the T2 (CPMG) and UTE sequences, the slice range was cropped to only the slices spanning masticatory muscle ROI (masseter, temporalis, medial pterygoid, or lateral pterygoid), further limiting any facial slice coverage and excess data beyond what the study required. Because of this crop, SliceTiming metadata from the original acquisition no longer corresponds to the retained slices and has been omitted from the T2/UTE sidecars. ## Funding NIH Grant R61AT012270.