The development of large-scale pathological image analysis algorithms is essential for enhancing patient outcomes and propelling advancements in healthcare. With the burgeoning availability of big data in computational pathology, there exists a significant opportunity to craft more precise and individualized diagnostic and treatment strategies. However, the complexity and diversity of digital pathological images and their varying modalities present substantial challenges in devising scalable image processing frameworks capable of managing inter-subject variations and delivering robust, timely medical image analysis. At the same time, ethical implementation of the underlying tools is of paramount interest, to ensure addressing model bias, data imbalance, fairness, and transparency, requiring usability and co-design approaches to be integrated with computational pipeline development.

To overcome these hurdles, it is imperative to foster interdisciplinary collaborations among engineers, scientists, clinicians, medical AI ethicists, and human factor researchers. By uniting the expertise of these professionals, innovative methods for data analysis, management, and sharing can be developed to optimize the utilization of big data in computational pathology. Moreover, advancements in large-scale data processing and sophisticated foundation models can assist in discerning patterns and relationships between imaging characteristics and clinical outcomes, thus enhancing the accuracy of diagnoses and tailoring treatment plans more effectively to empower various stakeholders, including clinicians and patients. 

While the directions for computational pathology are shaping up, the development of large-scale pathological image analysis algorithms has not kept pace with the increasing sophistication and complexity of image modalities. It is crucial, therefore, to prioritize the creation of patient-centered clinical image analysis frameworks that incorporate large-scale data, models, and infrastructures. These systems must be robust yet capable of rapid processing to support timely decision-making in clinical settings. There is a pressing need for the development of fast, efficient algorithms that can handle extensive volumes of digital pathologic data. Similarly, the tools developed need to be codesigned with important stakeholders, including patients and clinicians, ensuring providing critical support in clinical decision making and ultimately improving the speed and accuracy of patient care interventions.

This JMI special section welcomes original research papers that develop new methods for computational pathology with large-scale data, models, and infrastructures. We also welcome high-quality submissions from work presented at top conferences, such as MICCAI, IPMI, MIDL, CVPR, ICLR, ICML, and others, that propose new methods and techniques for leveraging patient-centered clinical image analysis with large-scale data, models, and infrastructures. Topics of interest include, but are not limited to, the following:

Computational pathology algorithms

• Patient-centered clinical image analysis with image and non-image data analysis
• Domain adaptation techniques for transferring models across different clinical domains
• Deep learning for pathological image analysis using large datasets
• The development of fast and efficient algorithms that can process large amounts of pathological imaging data in point-of-care decision support
• Multi-omics and texture analysis using digital pathology
• Personalized medicine using digital pathology data
• Big Data analytics for image-guided therapies
• Explainable AI techniques for clinical decision-making using foundation models
• Privacy-preserving techniques for using foundation models in clinical image analysis
• Quality assurance and quality control in digital pathology studies
• Ethical AI development in computational pathology
• Usability and co-design study in to enhance computational pathology pipeline

Clinical digital pathology image analysis with large-scale deep learning models

• Developing models for clinical outcomes using large-scale digital pathology datasets
• Efficient fine-tuning methods for clinical applications using foundation models
• Evaluating and benchmarking foundation models for digital pathology image analysis applications
• Developing foundation models for real-time image analysis in clinical settings
• Adapting foundation models for rare disease diagnosis and treatment planning

Computational Pathology with advanced computational infrastructures

• Developing algorithms for data standardization and harmonization in multi-site studies
• Developing federated learning and distributed computing frameworks for large-scale digital pathology imaging datasets
• Federated learning methods for computational pathology
• The development of fast and efficient infrastructures that can process large amounts of pathological image data in point-of-care decision support
• Developing automated computational pathology pipelines for large-scale medical imaging studies and deployment

Manuscripts should conform to the JMI author guidelines:  http://spie.org/JMIauthorinfo. Prospective authors should submit an electronic copy of their manuscript through the online submission system at https://jmi.msubmit.net. Please indicate in your cover letter that the submission is for this special section.

Each manuscript will be reviewed by at least two independent reviewers. Peer review will commence immediately upon manuscript submission, with a goal of making a first decision within six weeks. Special sections are opened for publication once a minimum of four papers have been accepted; each paper is published as soon as the copyedited and typeset proofs are approved by the author.

Theranostics, the combination of the terms therapeutics and diagnostics, is an emerging field of medicine that uses a radiopharmaceutical to identify and locate disease based on specific targets or receptors (diagnosis), followed by a second radiopharmaceutical to deliver therapeutic levels of radiation absorbed dose to target tissue (therapy).  Building upon its foundation in nuclear medicine, more recently the field has gained additional attention because of demonstrations of its ability to improve patient outcomes with low side effects. Successful clinical applications of theranostics include treatments for neuroendocrine tumors and prostate cancer.

Medical imaging plays a vital role in theranostics. In addition to the critical role imaging plays in disease diagnosis, staging, monitoring, and response evaluation, accurate visualization of in vivo radiopharmaceutical biodistribution provides critical information on patient selection for theranostics. Theranostics can incorporate ultrasound, magnetic resonance imaging (MRI), computed tomography (CT), positron-emission tomography (PET), single-photon emission computed tomography (SPECT), and many others. Further, rapid expansion of artificial intelligence (AI) that has revolutionized several aspects of medical imaging is also poised to accelerate advances in theranostics.

This JMI special section welcomes original research papers on theranostics agents, medical imaging methods for theranostics, computational paradigms for dosimetry, dose-response studies, clinical outcomes, applications of artificial intelligence to theranostics, and more. We also welcome high-quality submissions from work presented at top conferences, such as SNMMI, EANM, AAPM, MICCAI, SPIE MI, ISBI, and others.

Topics of interest include, but are not limited to, the following:

• Novel radionuclides in theranostics
• Emerging imaging technologies in theranostics
• Novel theranostics schemas under development
• Imaging biomarkers for tumor characterization
• Methods to identify therapeutic or diagnostic agents
• Prediction of tumor grade and metastatic potential
• Pre-treatment and post-treatment dosimetry in theranostics
• Theranostics methods for cancer prognosis and monitoring
• Multi-modal data modeling methods using imaging, lab data, and health record data
• The role of theranostics in multimodality oncologic care including sequencing and combining therapies
• Theranostics practical aspects in nuclear medicine clinic

Manuscripts should conform to the JMI author guidelines:  http://spie.org/JMIauthorinfo. Prospective authors should submit an electronic copy of their manuscript through the online submission system at https://jmi.msubmit.net. Please indicate in your cover letter that the submission is for this special section.

Each manuscript will be reviewed by at least two independent reviewers. Peer review will commence immediately upon manuscript submission, with a goal of making a first decision within six weeks. Special sections are opened for publication once a minimum of four papers have been accepted; each paper is published as soon as the copyedited and typeset proofs are approved by the author.

Digital tomosynthesis (DT) is a limited angle tomographic imaging modality that has been primarily developed in the past two decades. It is currently being used for a variety of clinical tasks, focused mostly on imaging of the breast and the chest.

Digital breast tomosynthesis (DBT) was first discussed in the landmark article by Niklason et al. in 1997, and was first approved for commercial use by the FDA for radiographic applications in 2005 and for mammography in 2014. It is well known that screening of asymptomatic women for breast cancer with mammography has contributed to a 30-40% decrease in breast cancer mortality, however, the interpretation of a mammogram is still challenging, especially for the dense breast. Digital breast tomosynthesis allows for visualization of the pseudo 3D breast, thereby reducing the supposition effect (overlapping normal breast structure) that greatly contributes to the difficulty of reading a mammogram. In the past 10 years, retrospective and prospective clinical trials have shown that DBT can improve the cancer detection rate while also reducing the rate of recall. Many improvements in DBT technology have been and are continuing to be developed promising even better performance of this relatively new technology in the future.

Radiographic tomosynthesis is a limited angle tomographic method that provides sectional images through anatomy using a low-dose technique. Studies have shown that detection of pathology is substantially improved compared with conventional radiography. In addition, radiographic tomosynthesis has a number of advantages of conventional CT imaging of the chest, including lower dose, lower financial cost, and higher patient throughput.

This special section of the Journal of Medical Imaging (JMI) seeks contributions in the form of research articles about digital tomosynthesis that highlight a wide spectrum of research areas including, but not limited to:

• History of digital tomosynthesis (DT): instrumentation
• History of DT: clinical use
• Novel hardware developments including detectors, and x-ray sources
• Development of new imaging acquisition strategies, including alternative geometries
• Optimization of DT acquisition parameters
• Novel DT image reconstruction methods including the use of deep learning
• Development of new digital phantoms for modeling of DT systems
• In silico computational models for simulating DT systems
• Model observers for assessing DT image quality
• Virtual Clinical Trials of DT
• Development of new quality control (QC) or acceptance testing methods for assessing system performance of DT
• Clinical aspects of DT
• CAD/AI solutions for DT
• New applications of DT, for example, contrast-enhanced DT, multi-modality systems, phase-contrast DT, new synthetic mammography (SM) algorithms, etc.

Manuscripts should conform to the JMI author guidelines:  http://spie.org/JMIauthorinfo. Prospective authors should submit an electronic copy of their manuscript through the online submission system at https://jmi.msubmit.net. Please indicate in your cover letter that the submission is for this special section.

Each manuscript will be reviewed by at least two independent reviewers. Peer review will commence immediately upon manuscript submission, with a goal of making a first decision within six weeks. Special sections are opened for publication once a minimum of four papers have been accepted; each paper is published as soon as the copyedited and typeset proofs are approved by the author.