Ultra-high-dose-rate (UHDR) radiation therapy has demonstrated promising potential in reducing toxicity to organs-at-risk (OARs). Proton therapy is uniquely positioned to deliver UHDR by leveraging the Bragg peak in conjunction with patient-specific range modulators (PSRMs) to generate a spread-out Bragg peak (SOBP). Existing proton FLASH (pFLASH) treatment planning workflows typically follow a two-step process: 1) generating a multi-energy intensity-modulated proton therapy (IMPT) plan to determine spot weights and 2) subsequently converting this plan into a single-energy pFLASH delivery using PSRM optimization. However, the intrinsic coupling between spot weight distribution and PSRM design has not been fully investigated, which may limit the achievable dosimetric and radiobiological advantages of pFLASH therapy.

Conventional four-dimensional cone beam computed tomography (4D-CBCT) usually requires long scan time and high radiation dose. Fast and low-dose 4D-CBCT is preferred but its quality is compromised by the reduced number of projections used for reconstruction.

Medical image segmentation is vital for clinical diagnosis, yet deep learning models face domain shift challenges when test data distributions differ from training data. Single-source domain generalization (DG) has emerged as a solution to these limitations of deep learning models by training models on a single source domain to generalize to unseen target domains. Current single-source DG methods employ domain randomization (e.g., input/feature-space perturbations) to simulate unseen target domains, but face two key limitations: (1) Texture bias in CNN-based architectures due to their local processing nature, leading to overfitting and poor generalization; (2) Restricted augmented style diversity caused by solely source-dependent feature perturbations and input- or feature-only augmentation, resulting in insufficient coverage of target domain distributions and degraded model's generalization ability.

Diffusing alpha-emitter radiation therapy (DaRT) is currently being evaluated in several clinical trials as an interstitial temporary brachytherapy source. However, determining source strength solely based on gamma spectrometry measurements of Ra activity has limitations. This method does not account for changes in the desorption of Ra progeny from the seed's surface or the therapeutically relevant alpha emissions. Variations in the source construction, especially Ra depth in the seed, significantly impact the Ra progeny desorption probabilities which leads to changes in the dose deposited in the tumor. Therefore, an improved source strength specification for DaRT seeds is needed, one that accounts for variations in source construction that affects the absorbed dose delivered to the tumor.

Vestibular schwannomas (VS), the third most common nonmalignant brain tumor, pose a significant challenge for geometric segmentation due to their irregular shape. The consistent anatomical origin of VS constitutes a domain-invariant prior capable of attenuating heterogeneity; however, contemporary studies have yet to exploit this characteristic.

Intensity-modulated radiation therapy (IMRT) treatment planning technology has made personalized cancer care a reality for millions of patients in recent decades. The complexity of these tools, however, often leads to intractability and suboptimal mitigation strategies in otherwise well-optimized plans. An example of such a strategy is reducing data granularity through voxel sampling.

Healthy tissue toxicity is the main limitation for the treatment of brain tumors using radiotherapy (RT). Megavoltage X-ray Minibeam Radiation Therapy (MBRT) and proton Minibeam Radiation Therapy (pMBRT) are novel therapeutic approaches demonstrated to reduce healthy tissue toxicity, enhancing the therapeutic possibilities of challenging cases. Owing to the different radiation types delivering the dose, their dosimetric properties may make one approach more suitable than the other depending on tumor type and location. While both techniques have been extensively studied in small animal models, canine clinical studies provide a closer approximation to human conditions.

Cardiac magnetic resonance imaging (CMRI) is a non-invasive medical examination method that provides a comprehensive evaluation of the anatomy, function, blood flow, and histology for cardiovascular diseases. Accurately segmentation of the left and right ventricles and myocardium from CMR images can significantly aid doctors in diagnosing cardiovascular diseases. However, due to the variable shapes of the right ventricle, narrow and tubular myocardial structures, and unclear boundaries caused by small grayscale differences between cardiac substructures, CMR image segmentation remains a challenging task.

Liver diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) measurement has proven valuable in diagnosing liver diseases. In certain patient populations, a free-breathing (FB) liver DWI approach is desirable to improve patient comfort and broaden clinical applicability. However, maintaining high image quality under FB conditions and achieving satisfactory ADC repeatability can be challenging when using routine diffusion-weighted single-shot echo-planar imaging (DW-ss-EPI).

With the development of photon-counting detectors (PCD), spectral CT has gained greater flexibility in utilizing spectral information, making it a powerful tool for material decomposition. Traditional decomposition methods rely on predefined physical models to estimate effective atomic numbers and density. However, these models may fail to maintain accuracy under flexible scanning protocols or for a broader range of materials.

Skin dose measurements for the inline MRI-linac show a highly concentrated skin dose hot spot at the central axis of the beam. This is because the electron contamination is focused in the presence of the inline magnetic field, causing a significant entry surface (or skin) dose hot spot. The future clinical treatment will employ a 2 cm water electron filter to reduce electron contamination from the X-ray beam; however, this approach affects the overall dose distribution, and the skin dose remains significantly high.

Recent advancements in cone-beam CT (CBCT) technology have significantly improved image quality leading to increased adoption of CBCT-guided online adaptive radiotherapy (ART). However, the impact of subject motion on these novel fast acquisition approaches on ART is not fully understood.

Deep Learning (DL) techniques have enabled up to 6-fold acceleration in musculoskeletal magnetic resonance imaging (MRI) while preserving diagnostic image quality. Further, improvements in acceleration and generalization require novel approaches. We propose a DL framework that integrates Simultaneous Multislice (SMS) imaging with Parallel Imaging (PI) to enhance current DL-based reconstruction.

Amyloid PET imaging has been extensively employed in the noninvasive assessment of amyloid-beta accumulation in Alzheimer's disease. Various amyloid radiotracers are commonly used in clinical settings; however, the limited interchangeability among these radiotracers hinders the feasibility of long-term clinical trials and multicenter comparisons. The Centiloid method was proposed for standardization, though providing a single score per image; voxel-wise translation remains a formidable task.

3D lattice radiation therapy (3D-LRT) is an effective treatment solution that can offer excellent local tumor control with limited morbidity. Pencil beam scanning (PBS) proton therapy achieves excellent dose conformity and eliminates exit doses to normal tissue, making it a great candidate to implement 3D-LRT. But it may also introduce high entrance doses to normal tissues. Ultra-high dose rate (UHDR) beams may offset this disadvantage by triggering the FLASH normal tissue protection effect.

Swallowing-induced anatomical motion can compromise the accuracy of head-and-neck radiotherapy, underscoring the need for proactive motion management. Because tongue motion precedes laryngeal displacement, it may serve as a predictive signal for swallowing onset.

Radiation therapy is a cornerstone of cancer management, utilized in over half of all cancer treatments. Treatment precision is critical for ensuring the efficacy of radiation therapy. Current radiotherapy modalities-x-ray photon, proton, and electron therapies-each offer distinct advantages and challenges depending on tumor characteristics and treatment goals. Despite technological advancements in treatment planning and image guidance, challenges such as intra-fractional and inter-fractional tumor motion and the lack of in vivo dose verification persist. Radiacoustic imaging (RAI) has emerged as a promising solution for real-time in vivo dose verification and adaptive radiotherapy. By detecting acoustic waves generated during radiation dose deposition, RAI offers the potential to enhance treatment precision and outcomes. This review explores the development, performance, and clinical applications of RAI, particularly its utility across photon, proton, and electron therapies, as well as its potential in FLASH radiotherapy. We also discuss its future prospects, projecting its integration into clinical workflows and its role in transforming cancer treatment.

The accurate segmentation of lesions in breast ultrasound images directly affects the assessment of lesion size, location, and morphological characteristics, and plays a crucial role in clinical diagnosis and treatment. However, due to multiple inherent difficulties such as noise interference, low contrast between the lesion and the surrounding tissues, and the complexity and diversity of the lesion, this task remains highly challenging.

Two-dimensional to three-dimensional (2D/3D) registration is critical in image-guided interventions, particularly in vascular procedures such as endovascular therapy (EVT), where accurate alignment between preoperative 3D images and intraoperative 2D x-ray angiography can improve procedural safety and precision. However, achieving both real-time and precise 2D/3D registration remains challenging due to high computational demands.

The increase in public medical imaging datasets has raised concerns about potential patient reidentification from head CT scans. However, existing defacing algorithms, which help protect patient confidentiality, fail to preserve critical radiotherapy structures, including organs at risk (OARs) and planning target volumes (PTVs) in head and neck cancer (HNC) patients. Furthermore, current algorithms do not address the defacing of DICOM-RT structure set and dose data, which also contain information for facial surface rendering.

Compared to photon therapy (XT), proton therapy (PT) can often reduce normal tissue toxicity for head and neck (HN) cancer patients, despite being a limited resource. On the other hand, clinical decision-making process to select between PT and XT (e.g., treatment planning and then plan evaluation for comparing normal tissue complication probabilities (NTCP) between XT and PT) is time-consuming and resource demanding.