Artificial intelligence (AI) applications are becoming increasingly common in radiology. However, ensuring reliable operation and expected clinical benefits remains a challenge. A systematic testing process aims to facilitate clinical deployment by confirming software applicability to local patient populations, practises, adherence to regulatory and safety requirements, and compatibility with existing systems. In this work, we present our testing process developed based on practical experience. First, a survey and pre-evaluation is conducted, where information requests are sent for potential products, and the specifications are evaluated against predetermined requirements. In the second phase, data collection, testing, and analysis are conducted. In the retrospective stage, the application undergoes testing with a pre selected dataset and is evaluated against specified key performance indicators (KPIs). In the prospective stage, the application is integrated into the clinical workflow and evaluated with additional process-specific KPIs. In the final phase, the results are evaluated in terms of safety, effectiveness, productivity, and integration. The final report summarises the results and includes a procurement/deployment or rejection recommendation. The process allows termination at any phase if the application fails to meet essential criteria. In addition, we present practical remarks from our experiences in AI testing and provide forms to guide and document the testing process. The established AI testing process facilitates a systematic evaluation and documentation of new technologies ensuring that each application undergoes equal and sufficient validation. Testing with local data is crucial for identifying biases and pitfalls of AI algorithms to improve the quality and safety, ultimately benefiting patient care.

Stereotactic radiosurgery (SRS) is an effective treatment for essential tremor (ET) and Parkinsonian tremor (PT). However, current treatment methods can be time-consuming and expose patients to unnecessary radiation. This study investigates the use of the Varian TrueBeam system with FFF beam modes as a potential solution for these issues.

By modelling patient exposures of interventional procedures, this study compares the reduction of radiation detriment between Digital Variance Angiography (DVA) and Digital Subtraction Angiography (DSA).

This study evaluates the impact of partial volume effect (PVE) correction on [F]fluoromisonidazole (FMISO) PET images, focusing on the conversion of standardized uptake values (SUV) to partial oxygen pressure (pO) and the subsequent determination of hypoxic tumor volume (HTV).

Cosmic rays have the potential to induce significant changes in atmospheric chemical reactions by generating ions, thereby influencing the atmosphere's chemical composition. The use of particle-molecule interaction models that account for the molecular structure of the atmospheric medium can advance our understanding on the role of ions, and enables a quantitative analysis of the impact of ion-molecule reactions on atmospheric modifications. This study marks the initial effort to expand the Geant-DNA toolkit for atmospheric applications.

Targeted alpha therapies show great potential for cancer treatment due to their high linear energy transfer (LET) and low range. At is currently employed in clinical trials. Targeted alpha therapies (TAT) are effective as an adjuvant treatment for cancer or to treat micrometastases and diffuse cancers. A deeper understanding of the induced initial damage is crucial to enhance treatment planning.

To evaluate the performance of pre-trained deep learning schemes (DLS) in hepatic steatosis (HS) grading of Non-Alcoholic Fatty Liver Disease (NAFLD) patients, using as input B-mode US images containing right kidney (RK) cortex and liver parenchyma (LP) areas indicated by an expert radiologist.

Cardiac substructures are critical organs at risk in left-sided breast cancer radiotherapy being often overlooked during treatment planning. The treatment technique plays an important role in diminishing dose to critical structures. This review aims to analyze the impact of treatment- and patient-related factors on heart substructure dosimetry and to identify the gaps in literature regarding dosimetric reporting of cardiac substructures.

The aim of our study is to verify the reliability of Diffusion Kurtosis Imaging (DKI) parameters through correlation with perfusion metrics obtained by Dynamic Contrast Enhanced (DCE)- and Dynamic Susceptibility Contrast (DSC)-MRI techniques in histologic-proven primary central nervous system lymphoma (PCNSL).

The purpose of this study is to investigate the accuracy of few-time-points (FTP) time-integrated activity coefficients (TIACs) in peptide-receptor radionuclide therapy (PRRT) using non-linear mixed-effects (NLME) modeling.

This study aimed to develop a statistical approach for edge spread function (ESF) phase alignment to improve the accuracy of modulation transfer function (MTF) measurements at the edges of computed tomography (CT) images.

Medical Physics departments primarily concentrate on clinical operations and regulatory compliance, which often restricts their ability to improve technical efficiencies. Nonetheless, developing technical capabilities is crucial as the healthcare sector increasingly depends on advanced technologies. A part-time software engineer was successfullyrecruited and integrated into the medical physics team to address operational needs and provide technical solutions. The engineer designed tailored systems, established automated dose tracking to ensure regulatory compliance, and worked alongside clinical staff for effective data management. Furthermore, they created a standardised operating environment for research initiatives, provided computational infrastructure for machine learning endeavours, and optimised workflows through automation. The integration improved workflow efficiency, expanded research capacity, and enhanced system integration, illustrating the significant benefits of incorporating technical expertise within medical physics teams.

This study aimed to determine the effect of model selection on simplified dosimetry for the kidneys using Bayesian fitting (BF) and single-time-point (STP) imaging.

The normal brain is an important dose-limiting organ for brain cancer patients undergoing radiotherapy. This study aims to develop a model to calculate photon isoeffective doses (D) to normal brain that can explain the incidence of grade 2 or higher somnolence syndrome (SS⩾2) after Boron Neutron Capture Therapy (BNCT).

This study aims to compare the dosimetric impact of incorporating systematic and random setup uncertainties in the robust optimization of head and neck cancer (HNC) Intensity Modulated Proton Therapy (IMPT) plans.

Commercial radiation therapy treatment planning systems (TPS) provide the three-dimensional time-integrated planned dose distribution. A four-dimensional (4D) dose calculation is essential to minimise dose-rate effects on pacemaker or in total body irradiation treatment or for time-dependent patient-specific quality assurance. We introduce the BeamSplitter, an algorithm in a commercial TPS generating 4D dose calculation.

This study aims to validate the Light-Ion Quantum Molecular Dynamics (LIQMD) model, an advanced version of the QMD model for more accurate simulations in hadron therapy, incorporated into Geant4 (release 11.2).

Inflammatory breast cancer (IBC) is a rare and aggressive type of breast cancer, as many physicians may not be aware of it in terms of symptoms and diagnosis. Mammography is the first choice in breast screenings and diagnosis. Because of a lack of expertise and imaging datasets, IBC portrayal and machine learning-based diagnosis systems have not yet been studied thoroughly. Developing scanning and diagnosis tools can close the knowledge gap and barriers to a timely IBC diagnosis.

Brain cancer poses a significant global health challenge, with mortality rates showing a concerning surge over recent decades. The incidence of brain cancer-related mortality has risen from 140,000 to 250,000, accompanied by a doubling in new diagnoses from 175,000 to 350,000. In response, magnetic resonance imaging (MRI) has emerged as a pivotal diagnostic tool, facilitating early detection and treatment planning. However, the translation of deep learning approaches to brain cancer diagnosis faces a critical obstacle: the scarcity of public clinical datasets reflecting real-world complexities. This study aims to bridge this gap through a comprehensive exploration and augmentation of training data. Initially, a battery of pre-trained deep models undergoes evaluation on a main brain cancer MRI "BT-MRI" dataset, yielding remarkable performance metrics, including 100% accuracy, precision, recall, and F1-Score, substantiated by the Score-CAM methodology. This initial success underscores the potential of deep learning in brain cancer diagnosis. Subsequently, the model's efficacy undergoes further scrutiny using a supplementary brain cancer MRI "BCD-MRI" dataset, affirming its robustness and applicability across diverse datasets. However, the ultimate litmus test lies in confronting the model with synthetic testing datasets crafted to emulate real-world scenarios. The synthetic testing datasets, a BCD-MRI testing sub-dataset enriched with noise, blur, and simulated patient motion, reveal a sobering reality: the model's performance plummets, exposing inherent limitations in generalization. To address this issue, a diverse set of optimization strategies and augmentation techniques, ranging from diverse optimizers to sophisticated data augmentation methods, are exhaustively explored. Despite these efforts, the problem of generalization persists. The breakthrough emerges with the integration of noise and blur as augmentation techniques during the training process. Leveraging Gaussian noise and Gaussian blur kernels, the model undergoes a transformative evolution, exhibiting newfound robustness and resilience. Retesting the refined model against the challenging synthetic datasets reveals a remarkable transformation, with performance metrics witnessing a notable ascent. This achievement underscores the important role of correct selection of data augmentation in fortifying the generalization of deep learning models for brain cancer diagnosis. This study not only advances the frontiers of diagnostic precision in brain cancer but also underscores the paramount importance of methodological rigor and innovation in confronting the complexities of real-world clinical scenarios.

In the adjuvant right breast radiation therapy, after breast-conserving surgery, we wanted to look for a parameter that would help in the choice between the 3D-CRT or VMAT techniques, considering the risk of pneumonia to the ipsilateral lung (IL) linked to high doses. We also investigated the risk of second tumors in the IL related to the VMAT low doses.

The use of three-dimensional (3D) printing in medical applications is quickly becoming mainstream. There have been an increasing number of publications discussing its implementation in radiotherapy, and the technology has become more affordable. The objective of this study was to establish how widely 3D printing is currently being utilised and what has been done to validate the processes and outcomes. A survey was sent to the UK and Ireland medical physics mailing lists. The questions were designed to establish how many centres were using 3D printers, how 3D printers were being utilized, the type of printing technologies being used, and how risk was being addressed. A total of 60 radiotherapy centres responded to the survey, with 38 % of the respondents currently using 3D printing. The majority (85 %) of the remaining respondents said they may or would have a 3D printer in the next 3 years. The majority of users were using FDM-type printers. The main variability among 3D printer users was how risk management and QA were addressed. This survey has demonstrated that there is an increased appetite for 3D printers in radiotherapy even beyond phantoms and bolus. Yet, despite this, guidance on implementation, compliance with the medical device directive and risk management remains sparse. As a consequence, centres have adopted a variety of approaches to risk management and QA.