A rare variant of oromandibular limb hypogenesis syndrome: a case report of glossopalatal ankylosis

Frontiers in Dental Medicine · 2026-05-01

Introduction: Oromandibular limb hypogenesis syndrome (OLHS) Type IIIA is an exceptionally rare congenital anomaly, with only a few cases documented in the literature. It is characterized by developmental hypoplasia of the tongue, mandible, and limbs, manifesting as hypoglossia, micrognathia, and hypomelia. Case presentation: We report the case of a 9-year-old patient presenting with restricted mouth opening and feeding difficulties. Clinical examination revealed fusion of the tongue to the palate with synechia. Surgical intervention involved the release of the dense fibrous band between the tongue and palate, followed by palatoplasty to correct the cleft palate. The cleft palate was successfully repaired using the Von Langenbeck technique, restoring oral functionality. Conclusion: This case highlights the clinical manifestations, differential diagnosis, and surgical management of OLHS Type IIIA. It underscores the importance of timely intervention to improve functional outcomes in patients with this rare syndrome.

Real-time prostate cancer risk stratification and scan tailoring using deep learning on abbreviated prostate MRI: A prospective evaluation

Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

Motivation: MRI is valuable for detecting and managing Prostate Cancer (PCa), but its use is limited by long scan times. While DCE helps with staging and biopsy guidance, its value only applies if PCa is present. Goal(s): We aim to develop a DL model that identifies csPCa from bpMRI scans in real-time, determining whether DCE is needed. Approach: We trained a DL model using bpMRI to provide feedback directly at the MRI scanner, guiding the need for further imaging. Results: In a prospective test, the model achieved an AUC of 0.86 for PI-RADS ≥ 3. Sensitivity and specificity for csPCa were 0.92 and 0.47. Impact: This study demonstrates that a DL model can guide the selective use of mpMRI based on bpMRI, optimizing resources. This approach could streamline PCa screening, improve patient care, and inspire further research into adaptive and personalized MRI protocols.

Harnessing Side Information for Highly Accelerated MRI

Lecture notes in computer science · 2025-09-19

A Trust-Guided Approach to MR Image Reconstruction With Side Information

IEEE Transactions on Medical Imaging · 2025-07-31 · 6 citations

Reducing MRI scan times can improve patient care and lower healthcare costs. Many acceleration methods are designed to reconstruct diagnostic-quality images from sparse k-space data, via an ill-posed or ill-conditioned linear inverse problem (LIP). To address the resulting ambiguities, it is crucial to incorporate prior knowledge into the optimization problem, e.g., in the form of regularization. Another form of prior knowledge less commonly used in medical imaging is the readily available auxiliary data (a.k.a. side information) obtained from sources other than the current acquisition. In this paper, we present the Trust-Guided Variational Network (TGVN), an end-to-end deep learning framework that effectively and reliably integrates side information into LIPs. We demonstrate its effectiveness in multi-coil, multi-contrast MRI reconstruction, where incomplete or low-SNR measurements from one contrast are used as side information to reconstruct high-quality images of another contrast from heavily under-sampled data. TGVN is robust across different contrasts, anatomies, and field strengths. Compared to baselines utilizing side information, TGVN achieves superior image quality while preserving subtle pathological features even at challenging acceleration levels, drastically speeding up acquisition while minimizing hallucinations. Source code and dataset splits are available on github.com/sodicksonlab/TGVN.

Resolving ambiguous space: Leveraging side information with deep learning to extend the limits of MR image reconstruction

Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

Motivation: Reconstruction quality sharply declines beyond certain acceleration levels, resulting in non-diagnostic images. Leveraging diverse sources of readily available side information offers a promising solution to this challenge, improving disambiguation during reconstruction and enabling higher acceleration rates while preserving diagnostic image quality. Goal(s): To reliably incorporate additional contextual information (relevant side information) into the MR image reconstruction. Approach: Eliminate undesirable solutions from the ambiguous space of the forward operator, while remaining faithful to the acquired data. Results: Compared to a set of baselines that also use side information, our method reconstructs high-quality knee MR images in the presence of heretofore challenging levels of under-sampling. Impact: By leveraging readily available sources of information which may not generally be used for image reconstruction, our approach reduces ambiguities, enabling more accurate solutions even with highly-sparse measurements.

Multi-coil multi-contrast joint reconstruction with protection from hallucination: Application to low-field MRI

Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

Motivation: In clinical MR imaging workflows, multiple contrast weightings of the same anatomy —sharing significant mutual information —are acquired sequentially. Harnessing this mutual information effectively while avoiding hallucination would enable substantial acceleration in MR acquisition without compromising image quality. Goal(s): To evaluate the efficacy of a projection-guidance in multi-coil, multi-contrast MR image reconstruction. Approach: Refinement block is projected onto the subspace spanned by the trailing right singular vectors of the forward operator, allowing disambiguation of potential solutions without undermining data consistency. Results: Joint reconstruction of three contrasts reduces the clinical workflow from 13 minutes to just 1 minute, maintaining high image quality and outperforming relevant baselines. Impact: MR images from different contrast weightings share substantial information. When this shared information is rigorously leveraged, high acceleration levels can be achieved, even in low-field settings, enabling efficient workflows and facilitating broader adoption.

DIMCIM: A Quantitative Evaluation Framework for Default-Mode Diversity and Generalization in Text-to-Image Generative Models

2025-10-19

Recent advances in text-to-image (T2I) models have achieved impressive quality and consistency. However, this has come at the cost of representation diversity. While automatic evaluation methods exist for benchmarking model diversity, they either require reference image datasets or lack specificity about the kind of diversity measured, limiting their adaptability and interpretability. To address this gap, we introduce the Does-it/Can-it framework, DIM-CIM, a reference-free measurement of default-mode diversity ("Does" the model generate images with expected attributes?) and generalization capacity ("Can" the model generate diverse attributes for a particular concept?). We construct the COCO-DIMCIM benchmark, which is seeded with COCO concepts and captions and augmented by a large language model. With COCO-DIMCIM, we find that widely-used models improve in generalization at the cost of default-mode diversity when scaling from 1.5B to 8.1B parameters. DIMCIM also identifies fine-grained failure cases, such as attributes that are generated with generic prompts but are rarely generated when explicitly requested. Finally, we use DIMCIM to evaluate the training data of a T2I model and observe a correlation of 0.85 between diversity in training images and default-mode diversity. Our work provides a flexible and interpretable framework for assessing T2I model diversity and generalization, enabling a more comprehensive understanding of model performance.

A Trust-Guided Approach to MR Image Reconstruction with Side Information

arXiv (Cornell University) · 2025-01-06

Reducing MRI scan times can improve patient care and lower healthcare costs. Many acceleration methods are designed to reconstruct diagnostic-quality images from sparse k-space data, via an ill-posed or ill-conditioned linear inverse problem (LIP). To address the resulting ambiguities, it is crucial to incorporate prior knowledge into the optimization problem, e.g., in the form of regularization. Another form of prior knowledge less commonly used in medical imaging is the readily available auxiliary data (a.k.a. side information) obtained from sources other than the current acquisition. In this paper, we present the Trust- Guided Variational Network (TGVN), an end-to-end deep learning framework that effectively and reliably integrates side information into LIPs. We demonstrate its effectiveness in multi-coil, multi-contrast MRI reconstruction, where incomplete or low-SNR measurements from one contrast are used as side information to reconstruct high-quality images of another contrast from heavily under-sampled data. TGVN is robust across different contrasts, anatomies, and field strengths. Compared to baselines utilizing side information, TGVN achieves superior image quality while preserving subtle pathological features even at challenging acceleration levels, drastically speeding up acquisition while minimizing hallucinations. Source code and dataset splits are available on github.com/sodicksonlab/TGVN.

Fluid Resuscitation in Oral & Maxillofacial Trauma

Journal of Maxillofacial and Oral Surgery · 2025-05-27

Temporal Generalization: A Reality Check

ArXiv.org · 2025-09-27

Machine learning (ML) models often struggle to maintain performance under distribution shifts, leading to inaccurate predictions on unseen future data. In this work, we investigate whether and under what conditions models can achieve such a generalization when relying solely on past data. We explore two primary approaches: convex combinations of past model parameters (\emph{parameter interpolation}) and explicit extrapolation beyond the convex hull of past parameters (\emph{parameter extrapolation}). We benchmark several methods within these categories on a diverse set of temporal tasks, including language modeling, news summarization, news tag prediction, academic paper categorization, satellite image-based land use classification over time, and historical yearbook photo gender prediction. Our empirical findings show that none of the evaluated methods consistently outperforms the simple baseline of using the latest available model parameters in all scenarios. In the absence of access to future data or robust assumptions about the underlying data-generating process, these results underscore the inherent difficulties of generalizing and extrapolating to future data and warrant caution when evaluating claims of such generalization.