Background: This study investigated alternative, non-invasive methods for human papillomavirus (HPV) detection in head and neck cancers (HNCs). We compared two approaches: analyzing computed tomography (CT) scans with a Deep Learning (DL) model and using radiomic features extracted from CT images with machine learning (ML) models. Methods: Fifty patients with histologically confirmed HNC were included. We first trained a modified ResNet-18 DL model on CT data to predict HPV status. Next, radiomic features were extracted from manually segmented regions of interest near the oropharynx and used to train four ML models (K-Nearest Neighbors, logistic regression, decision tree, random forest) for the same purpose. Results: The CT-based model achieved the highest accuracy (90%) in classifying HPV status. Among the ML models, K-Nearest Neighbors performed best (80% accuracy). Weighted Ensemble methods combining the CT-based model with each ML model resulted in moderate accuracy improvements (70–90%). Conclusions: Our findings suggest that CT scans analyzed by DL models hold promise for non-invasive HPV detection in HNC. Radiomic features, while less accurate in this study, offer a complementary approach. Future research should explore larger datasets and investigate the potential of combining DL and radiomic techniques.

Background: This study aimed to develop a novel human cell geometry for the Geant4-DNA simulation toolkit that explicitly incorporates all 23 chromosome pairs of the human cell. This approach contrasts with the existing, default human cell, geometrical model, which utilizes a continuous Hilbert curve. Methods: A Python-based tool named “complexDNA” was developed to facilitate the design of both simple and complex DNA geometries. This tool was employed to construct a human cell geometry with individual pairs of chromosomes. Subsequently, the performance of this chromosomal model was compared to the standard human cell model provided in the “molecularDNA” Geant4-DNA example. Results: Simulations using the new chromosomal model revealed minimal discrepancies in DNA damage yield and fragment size distribution compared to the default human cell model. Notably, the chromosomal model demonstrated significant computational efficiency, requiring approximately three times less simulation time to achieve equivalent results. Conclusions: This work highlights the importance of incorporating chromosomal structure into human cell models for radiation biology research. The “complexDNA” tool offers a valuable resource for creating intricate DNA structures for future studies. Further refinements, such as implementing smaller voxels for euchromatin regions, are proposed to enhance the model's accuracy.

In recent years, the use of Monte Carlo (MC) simulations in the domain of Medical Physics has become a state-of-the-art technology that consumes lots of computational resources for the accurate prediction of particle interactions. The use of generative adversarial network (GAN) has been recently proposed as an alternative to improve the efficiency and extending the applications of computational tools in both medical imaging and therapeutic applications. This study introduces a new approach to simulate positron paths originating from Fluorine 18 (18F) isotopes through the utilization of GANs. The proposed methodology developed a pure conditional transformer least squares (LS)-GAN model, designed to generate positron paths, and to track their interaction within the surrounding material. Conditioning factors include the predetermined number of interactions, and the initial momentum of the emitted positrons, as derived from the emission spectrum of 18F. By leveraging these conditions, the model aims to quickly and accurately simulate electromagnetic interactions of positron paths. Results were compared to the outcome produced with Geant4 Application for Tomography Emission (GATE) MC simulations toolkit. Less than 10 % of difference was observed in the calculation of the mean and maximum length of the path and the 1-D point spread function (PSF) for three different materials (Water, Bone, Lung).

This study aims to validate GATE and GGEMS simulation toolkits for brachytherapy applications and to provide accurate models for six commercial brachytherapy seeds, which will be freely available for research purposes. The AAPM TG-43 guidelines were used for the validation of two Low Dose Rate (LDR), three High Dose Rate (HDR), and one Pulsed Dose Rate (PDR) brachytherapy seeds. Each seed was represented as a 3D model and then simulated in GATE to produce one single Phase-Space (PHSP) per seed. To test the validity of the simulations’ outcome, referenced data (provided by the TG-43) was compared with GATE results. Next, validation of the GGEMS toolkit was achieved by comparing its outcome with the GATE MC simulations, incorporating clinical data. The simulation outcomes on the radial dose function (RDF), anisotropy function (AF), and dose rate constant (DRC) for the six commercial seeds were compared with TG-43 values. The statistical uncertainty was limited to 1% for RDF, to 6% (maximum) for AF, and to 2.7% (maximum) for the DRC. GGEMS provided a good agreement with GATE when compared in different situations: (a) Homogeneous water sphere, (b) heterogeneous CT phantom, and (c) a realistic clinical case. In addition, GGEMS has the advantage of very fast simulations. For the clinical case, where TG-186 guidelines were considered, GATE required 1 h for the simulation while GGEMS needed 162 s to reach the same statistical uncertainty. This study produced accurate models and simulations of their emitted spectrum of commonly used commercial brachytherapy seeds which are freely available to the scientific community. Furthermore, GGEMS was validated as an MC GPU based tool for brachytherapy. More research is deemed necessary for the expansion of brachytherapy seed modeling.

Pediatric patient-specific dosimetry of ionizing radiation is of great scientific and social interest. Children provide a higher relative cancer-risk from exposure to ionizing radiation compared to adults. The proposed study reviews the recent techniques applied in pediatric imaging and therapy applications for dosimetry purposes. Modern medicine makes use of advance computational tools for the personalization of internal and external dosimetry, especially in the sensitive group of children. Several groups of pediatric computational models have been developed which are combined with Monte Carlo (MC) simulations, machine learning (ML) techniques, and image processing algorithms for accurate dosimetry assessment. More specifically, this paper reviews the dosimetry applications in pediatric diagnostic procedures, including computed tomography and nuclear medicine applications. Right afterward, the most recent applications in therapeutic brachytherapy protocols are presented, which is a rather sensitive procedure in pediatrics. Finally, modern tools for dosimetry optimization are discussed, reviewing the most indicative applications with: 1) MC simulations for pediatric dosimetry assessment; 2) pediatric computational models, widely used in medical applications; and 3) ML techniques that provide an alternative method for estimating individualized absorbed doses.