Open research questions in Medical Imaging and Analysis

What the literature leaves open

• More importantly, the control and adjudication analyses showed that numeric HU output generation alone is insufficient as evidence of deployment quality.

  VLM-Assisted Routing and HU-Traceable DICOM Adaptation for Lumbar CT HU Measurement: A Deployment-Oriented Pilot Technical Evaluation · 2026 · DOI

• Future studies should focus on prospective rural validation, and these validations should examine not only discrimination and calibration but also workflow integration, clinician acceptance, fairness, safety, and notably its performance in incomplete or fragmented electronic health records.

  Artificial Intelligence for Clinical Decision Support in Rural Spine Care: A Narrative Review · 2026 · DOI

• The potential independent effect of the primary surgical indication was not evaluated and should be investigated in future multicenter studies. Although all patients received an interbody cage and autologous bone combined with BMP, the potential effects of other cage- and graft-related characteristics were not separately evaluated and should be investigated in future studies.

  Development and validation of explainable machine learning models to predict the risk of adjacent segment disease after lumbar fusion surgery · 2026 · DOI

• Therefore, the underlying physiological mechanisms remain unclear, and the observed signal characteristics should be interpreted with caution. Despite these limitations, our exploratory findings in this discovery-phase study – particularly the identification of unique vertebral phenotypes that have until now been understudied – highlight the potential of early and subtle signal changes as indicators of LBP.

  ISSLS Prize in Bioengineering Science 2026: Hidden in Plain Sight: Machine Learning–Assisted MRI Reveals Novel Vertebral Body Biomarkers of Chronic Low Back Pain in Humans · 2026 · DOI

• To address this issue, future work should focus on collecting more training cases with low BMD values or applying data augmentation techniques to better capture the characteristics of this subgroup. Fourthly, a significant limitation is the restriction to female subjects.

  Highly Accurate and Fully Automated Bone Mineral Density Prediction from Spine Radiographs Using Artificial Intelligence · 2026 · DOI

• In particular, the VerSe challenge has become a key reference standard for vertebral labeling and segmentation across multi-center CT datasets, for comparing algorithm providing a performance robustness under stress-testing model anatomical variation and domain shift (130).

  Artificial intelligence in spinal fracture imaging: from clinical application to translational promise · 2026 · DOI

• Although the evaluated framework demonstrated robust performance for musculoskeletal abnormality detection, several directions remain for future investigation. First, future studies should evaluate the evaluated framework using larger, more diverse, and multicenter datasets to further assess its robustness and generalizability across different imaging protocols, patient populations, and clinical environments.

  An Explainable Deep Learning Framework for Musculoskeletal Abnormality Detection · 2026 · DOI

• The conventional diagnostic methods have low accuracy and inadequate OA classification. There is a need for an efficient prediction model to overcome these shortcomings.

  Detection and classification of knee osteoarthritis severity levels using a Puma optimized YOLOv8-GAN framework · 2026 · DOI

• In the case of voting ensemble variants (VCall or VCtop), the VC ensemble strategy is limited to the kNN (the best baseline classifier) independently of using the complete set (all) or the three most promising baseline classifiers.

  Vertebral Column Pathology Diagnosis Using Ensemble Strategies Based on Supervised Machine Learning Techniques · 2024 · DOI

• Second, although the proposed dynamic parameters are biomechanically interpretable, their relationships with established clinical measures–such as pain intensity scales, disability indices, and static imaging findings–have not yet been quantified.

  HumanMoveNet: a dynamic 3D spine reconstruction framework for low back pain screening and rehabilitation assessment · 2026 · DOI

• Future research should explore longitudinal changes in these biomarker candidates and their associations with inflammatory markers and clinical outcomes.

  ISSLS Prize in Bioengineering Science 2026: Hidden in Plain Sight: Machine Learning–Assisted MRI Reveals Novel Vertebral Body Biomarkers of Chronic Low Back Pain in Humans · 2026 · DOI

• Future prospective studies with standardized treatment, extended follow-up, and complementary functional assessments are warranted to further optimize model stability and clinical applicability.

  Interpretable machine learning model integrating MRI-derived paraspinal muscle parameters for predicting new vertebral compression fractures after vertebral augmentation · 2026 · DOI

• Future research should focus on larger multi-center datasets, domain adaptation strategies to improve cross-institutional robustness, and hybrid or end-to-end deep learning frameworks incorporating attention mechanisms to further enhance classification reliability and clinical interpretability.

  Multi-Task Deep Learning Model for Automated Detection and Severity Grading of Lumbar Spinal Stenosis on MRI: Multi-Center External Validation · 2026 · DOI

• Evaluation was restricted to anatomically intact contralateral shoulders; performance on shoulders with bone defects, malunion, severe osteophytes, or glenoid wear remains unknown.

  A multi-task learning-based network for detecting three-dimensional osteometric landmarks of the shoulder joint · 2026 · DOI

• As a result, potential systemic influences on FL thrombosis and the prognostic value of the thrombus score remain uncertain.

  Tracking False Lumen Remodeling with AI: A Variational Autoencoder Approach After Frozen Elephant Trunk Surgery · 2025 · DOI

• Federated Two hard constraints never go away: labeled data are scarce, and privacy matters.

  Artificial intelligence in spinal fracture imaging: from clinical application to translational promise · 2026 · DOI

• The manual segmentation of bone models is time-consuming; therefore, automatic segmentation should be investigated.

  Magnetic Resonance Imaging–Based 3-Dimensional Models of the Pelvis and Hip Using Machine Learning for Automatic Bone Segmentation in a Dynamic Hip Impingement Simulation · 2025 · DOI

• Limited research focuses on anterior–posterior positions of the cervical spine, mostly on segmentation-related studies.

  Cervical Spondylosis Diagnosis Based on Convolutional Neural Network with X-ray Images · 2024 · DOI

• Our study's sample size was small, and the findings need to be validated in a prospective study with a large sample size.

  Three-Dimensional Lumbosacral Reconstruction by An Artificial Intelligence-Based Automated MR Image Segmentation for Selecting the Approach of Percutaneous Endoscopic Lumbar Discectomy · 2024 · DOI

• These findings support synthetic radiographs as both a shareable research asset and a useful pretraining resource for ML applications otherwise constrained by limited data availability.

  Synthetic cervical spine radiographs: expert validation and transfer learning for low-data settings · 2026 · DOI

• Further prospective studies are needed to determine whether the findings of this study translate into clinical benefits such as improved diagnostic completeness, earlier treatment initiation, and better patient outcomes. The study highlights the need for further research into the use of AI-assisted CT interpretation in the detection and treatment of malignant spinal lesions.

  Enabling earlier detection of spinal lesions in CT imaging with artificial intelligence—a case study · 2026 · DOI

• Traditional segmentation methods are unable to handle clinical requirements. The lack of efficient and accurate spinal CT image segmentation methods.

  SpinalSAM-R1: A vision–language multimodal interactive system for spine CT segmentation · 2026 · DOI

• Primary endpoints could include patient-reported outcomes (ODI, VAS) at 12 months, with secondary endpoints addressing complication rates, reoperation, and decision regret. Feasibility considerations would include infrastructure for multimodal data integration, clinician training, and patient consent for continuous monitoring.

  Digital twins and multimodal artificial intelligence in spine care: a scoping review of concepts, evidence, and translational barriers · 2026 · DOI

• The primary constraint remains the relatively small size of the clinical dataset, which, despite augmentation techniques, may limit the model’s generalization to rare or complex cases.

  Hybrid CNN-Swin Transformer Model to Advance the Diagnosis of Maxillary Sinus Abnormalities on CT Images Using Explainable AI · 2025 · DOI

• The study is retrospective and limited to a single AI system and a single radiologist. The requirement for a written informed consent from study participants was waived, and the study was approved by the Ethics Committee of the Bavarian State Chamber of Physicians.

  Enabling earlier detection of spinal lesions in CT imaging with artificial intelligence—a case study · 2026 · DOI

Most-cited papers in Medical Imaging and Analysis

• VerSe: A Vertebrae labelling and segmentation benchmark for multi-detector CT images · Medical Image Analysis · 2021 · 342 citations
• The DISC Predictive Scales (DPS): Efficiently Screening for Diagnoses · Journal of the American Academy of Child & Adolescent Psychiatry · 2001 · 297 citations
• XR (Extended Reality: Virtual Reality, Augmented Reality, Mixed Reality) Technology in Spine Medicine: Status Quo and Quo Vadis · Journal of Clinical Medicine · 2022 · 255 citations
• Beyond Self-Attention: Deformable Large Kernel Attention for Medical Image Segmentation · 2024 · 213 citations
• Automated bone mineral density prediction and fracture risk assessment using plain radiographs via deep learning · Nature Communications · 2021 · 177 citations
• Cervical vertebral maturation assessment on lateral cephalometric radiographs using artificial intelligence: comparison of machine learning classifier models · Dentomaxillofacial Radiology · 2020 · 102 citations
• Artificial Intelligence in Spinal Imaging and Patient Care: A Review of Recent Advances · Neurospine · 2024 · 78 citations
• Lumbar spine segmentation in MR images: a dataset and a public benchmark · Scientific Data · 2024 · 73 citations
• Computer-assisted bone segment navigation · Journal of Cranio-Maxillofacial Surgery · 1998 · 73 citations
• “Black Bone” MRI: a novel imaging technique for 3D printing · Dentomaxillofacial Radiology · 2017 · 61 citations

Most recent work

• MedCTM: A CNN-Transformer-Mamba Hybrid Network for Medical Image Classification · Information Processing & Management · 2026
• Read like a radiologist: Efficient vision-language model for 3D medical imaging interpretation · Medical Image Analysis · 2026
• Reliable uncertainty quantification for 2D/3D anatomical landmark localization using multi-output conformal prediction · Medical Image Analysis · 2026
• Clinical Application of Deep Learning for Spine MRI Interpretation: A Multicenter Evaluation of Artificial-Intelligence-Assisted versus Manual Reading on Diagnostic Agreement with the Reference Standard · Research · 2026
• Anatomy-Aware Text-Visual Fusion with Dual-Perspective Prompts for Fine-Grained Lumbar Spine Segmentation · International Journal of Computer Vision · 2026
• 2D, 2.5D, or 3D? Comparing Dimensional Approaches in Deep Neural Networks for 3D Medical Image Analysis · Journal of Imaging Informatics in Medicine · 2026
• Automated Detection of Cervical Spinal Cord Compression on MRI Using YOLO11 Deep Learning Architecture · Spine · 2026
• Digital twins and multimodal artificial intelligence in spine care: a scoping review of concepts, evidence, and translational barriers · Spine Deformity · 2026
• Deep learning in medical image analysis for bone tumor: A comprehensive survey · Information Fusion · 2026
• Neural architecture search for 3D biomedical image classification · Multimedia Systems · 2026

Related topics in Engineering