Frontiers in Biomedical Technologies

Strengthening Our Global Scientific Footprint: A New Chapter for FBT in 2026

Mohammad Reza Ay — 2026-01-26

As we launch the first issue of Frontiers in Biomedical Technologies (FBT) in 2026, it is my great pleasure to extend my deepest appreciation to all those who have contributed to the continued growth and advancement of the journal. I would especially like to acknowledge our domestic and international authors, whose trust in FBT and whose high-quality manuscripts form the foundation of our scientific achievements. Their contributions represent a diverse global community of researchers, and their commitment to excellence has significantly strengthened the journal’s impact and visibility. My sincere gratitude extends as well to our dedicated reviewers, whose constructive and rigorous evaluations ensure that every published article meets the highest standards of scientific quality. I also thank our readers for their continued engagement, and the editorial and executive teams for their sustained efforts, professionalism, and teamwork throughout the years.

The journal’s performance metrics clearly reflect a strong upward trend in both activity and visibility. According to our quantitative report, the number of submitted articles rose from 32 submissions in 2019 (when we seriously started) to 219 submissions in 2024, marking a remarkable expansion in the journal’s reach and recognition. As submissions increased, our acceptance rate naturally evolved—from 89% in 2019 to 45% in 2023—demonstrating both higher selectivity and our ongoing efforts to strengthen the scientific rigor of the journal.

FBT has also continued to expand its international academic footprint since being indexed in Scopus beginning in 2021. Recent citation analyses (2020–2024) show encouraging improvements in the journal’s citation performance, international collaboration, and the diversity of institutions and countries contributing to and citing our publications. These developments affirm FBT’s growing role as a trusted venue for impactful research in biomedical technologies.

As we move into 2026, we warmly invite researchers, scientists, clinicians, and industry experts to continue submitting their high-quality manuscripts to FBT. Your contributions are essential for advancing innovation in biomedical engineering and for supporting the journal’s mission to disseminate meaningful and influential research. In return, our editorial and executive teams remain dedicated to further enhancing the publication quality, review processes, visibility, and overall scientific impact of FBT. We look forward to a productive year filled with new ideas, strong collaborations, and continued academic progress.

Artificial Intelligence Assisted Detection of Respiratory Infectious Diseases Signs From Computed Tomography Images

Faezeh Shalbafzadeh — 2025-12-07

Purpose: Respiratory infectious diseases often manifest as Ground-Glass Opacity (GGO) or consolidation signs in the lungs. Artificial Intelligence (AI)-assisted systems utilizing data mining algorithms such as Waikato Environment for Knowledge Analysis (Weka) can be used for the detection and segmentation of these signs. In this study, we propose using Weka as a comprehensive data mining and machine learning tool to develop the most accurate models for detecting lung signs in chest CT images of patients with respiratory infectious diseases.

Materials and Methods: First, we manually selected specific signs from chest Computed Tomography (CT) images from 600 cases using the Graphical User Interface (GUI) Weka plugin. We then trained the random forest algorithm based on different features and presented the best combined model obtained for the automatic detection of the aforementioned signs. Lastly, the model's performance was evaluated with different metrics.

Results: Our findings indicate that the hybrid texture description features, including “Structure”, “Entropy”, “Maximum”, “Anisotropic”, and “Laplacian” available in Weka, demonstrated the lowest Out-of-Bag (OOB) error rate, highest Area Under the ROC Curve (AUC) value of 0.992, and accuracy of 98.1%.

Conclusion: By leveraging the combination of Weka features, we have successfully developed models for the detection and segmentation of lung signs associated with infectious diseases, from chest CT images. These findings contribute to the field of medical image analysis and hold promise for improving the diagnosis and treatment outcomes of patients with respiratory infectious disorders.

Investigating Patient-Specific Absorbed Dose Assessment for Copper-64 PET/CT

Ali Abdulhasan Kadhim — 2026-01-26

Purpose: There is a growing interest in the clinical application of new PET radiopharmaceuticals. This study focuses on using ⁶⁴Cu-DOTA-Trastuzumab for Positron Emission Tomography–Computed Tomography (PET/CT) imaging in gastric cancer patients. It aims to enhance the understanding of its bio-kinetic distribution and absorbed dose for safe and practical application in nuclear medicine.

Materials and Methods: The study was conducted at the Agricultural, Medical, and Industrial Research School (AMIRS), where ⁶⁴Cu was produced and purified. The radiopharmaceutical ⁶⁴Cu-DOTA-Trastuzumab was prepared, and three patients with confirmed Human Epidermal growth factor Receptor 2 (HER2)-positive gastric cancer underwent PET/CT scans at 1, 12 and 48 hours post-injection. Images were gained using a Discovery IQ PET/CT system and analyzed for an SUV. Bio-distribution was modeled using a two-exponential function, and absorbed doses were calculated using IDAC-Dose 2.1 software. CT doses were also evaluated.

Results: The study found that post-injection imaging at 12 hours or more provided superior image quality. The liver exhibited the highest cumulative activity, followed by the spleen and other organs. The effective dose estimates for ⁶⁴Cu-DOTA-Trastuzumab were within acceptable limits. CT dose calculations revealed that sensitive organs received higher doses.

Conclusion: This study successfully assessed the bio-kinetic distribution and absorbed dose of ⁶⁴Cu-DOTA-Trastuzumab in gastric cancer patients, demonstrating its safety and potential for clinical use. The optimal timing for PET/CT imaging and dosimetry data can inform clinical decision-making. Further research is warranted to explore the therapeutic potential of ⁶⁴Cu-DOTA-Trastuzumab and to establish clinical guidelines for its use.

Assessing Dihydrotestosterone-Induced Skin Alterations in C57BL/6 Mice: Implications for Androgenetic Alopecia through High-Resolution Ultrasound Imaging

Sadegh Shurche — 2026-01-26

Purpose: There are different types of hair loss known as alopecia. Various methods for treating Androgenetic Alopecia (AGA) are being investigated in the preclinical stage using C57BL/6 mice affected by this condition.

The purpose of the study was to evaluate the effects of Dihydrotestosterone (DHT) on the skin layers of male C57BL/6 mice, simulating a model of AGA using high-resolution ultrasound imaging.

Materials and Methods: Seven-week-old male C57BL/6 mice were selected for the study. To induce AGA, three of the mice received intraperitoneal injections of DHT at a dosage of 1 mg per day for five consecutive days, a known method for provoking hair loss via androgenic pathways. High-resolution ultrasound imaging at 40 and 75 MHz frequencies allowed detailed observation of skin layer changes due to DHT administration. Shear modulus and Young modulus were extracted using dynamic loading throughout ultrasonography with a 40 MHz frequency. Both control and AGA-affected groups were evaluated through structural imaging and were compared with histopathological results. Tissues were stained with Hematoxylin-Eosin (H&E) and Trichrome Mason.

Results: Ultrasound imaging revealed that the epidermis thickness was 0.22±0.01 mm in the control group compared to 0.31±0.02 mm in the AGA group at 40 MHz. At 75 MHz, these measurements were 0.10±0.05 mm for the control group and 0.20±0.01 mm for the AGA group. The dermis thickness measurements showed 0.30±0.02 mm in the control group and 0.70±0.04 mm in the AGA group at 40 MHz, while at 75 MHz, the thicknesses were 0.40±0.02 mm for the control group and 0.70±0.04 mm for the AGA group. H&E staining results aligned with these ultrasound findings, confirming increased epidermal and dermal thicknesses in the AGA group. Elasticity metrics indicated a shear modulus of 1.19±0.60 kPa for the control group and 6.70±0.33 kPa for the AGA group, while Young modulus demonstrated values of 6.47±0.32 kPa for the control group and 22.69±1.13 kPa for the AGA group. Further corroboration of altered tissue elasticity was provided by Trichrome staining, indicating significant changes in skin structure.

Conclusion: The administration of DHT in the C57BL/6 mice model leads to notable structural changes in skin layers, evidenced by an increased thickness of both the epidermis and dermis, along with diminished mechanical properties of skin elasticity.

Estimating the Relationship Between EMG Signals and EEG Signal Connections Using Convolutional Neural Networks

Elham Samadi — 2026-01-26

Purpose: Understanding the functional relationships between different parts of the human body can enhance the control of Brain-Computer Interface (BCI) systems. The brain, as the decision-making organ, controls all body parts to perform activities. In this study, the main objective is to estimate the activation of hand muscles and the effect of each muscle on another using Electroencephalogram (EEG) signals.

Materials and Methods: To discover the connection of hand muscles through brain signals, brain connections are extracted as influential components, and a convolutional network is utilized to assess the impact of EEG signals on the relationships among hand muscles. Five different connectivity methods were used to analyze the connections between EEG signal channels, such as correlation, coherence, the directed transfer function, Granger causality, and the phase delay index. The relationships between electromyogram (EMG) signal channels are also calculated using Granger causality. Signals are recorded in two phases: rest and activity, and ultimately, the EMG signal activity is estimated solely using EEG signals.

Results: Simulation results estimate the correlation between the estimated and actual patterns for test data to be around 0.949, indicating a high correlation between the estimated outputs and actual values.

Conclusion: Research indicates that exploring techniques for calculating relationships can be useful in evaluating the synergy and causal connections between EMG and EEG signals. In comparison to alternative graph-based techniques, this approach, utilizing regression analysis, demonstrated notably superior performance. This study could contribute to advancements in rehabilitation techniques and brain-computer interfaces.

Monte Carlo Simulation of Damage in Born Neutron Capture Therapy (BNCT) Converter Materials by High-Energy Proton Beam Spallation

Ali Nouraddini- Shahabadi — 2026-01-26

Purpose: High-energy protons are generally used for neutron production by Pb, W, Li, Be, and Ta targets that are used for the Born Neutron Capture Therapy (BNCT) technique. Neutron production targets are destroyed by proton spallation (evaporation of nuclei). The purpose of this study is the investigation of neutron activation and proton spallation damage of converter targets using the MCNPX code, which is based on the Monte Carlo method.

Materials and Methods: The MCNPX code was used to extract the activation and spallation information of secondary particle production in Pb, W, Li, Be, and Ta targets. The neutron activation and proton spallation damage, including radioactive elements production in converter targets, was extracted from data in the MCNPX output file.

Results: Results showed that the highest probability of radioactive elements production by proton with low-level energy in the Ta target are 180Hf, 179Hf, and 178Hf, and in the Li target is 7Be, respectively. In addition, the most probable radioactive elements produced by 200, 800, and 1200 MeV proton spallation in lead target are 118Tl and 78Pt, and in tungsten target are 98Hf, 110Ta, and 111Ta, respectively. The calculations showed that the production of radioisotopes in reactions with neutrons is lower than the production in reactions with a proton beam, and with increases in the energy of the proton beam, production of the radioactive elements was increased.

Conclusion: The results illustrated that the radioactive elements are produced in W, Pb, Li, Be, and Te targets in the BNCT method, which should be avoided as radiation hazards.

The Relationship between the Number of Segments and Gantry Angle on the Complexity of Head and Neck IMRT Plans

Fatemeh Zahra Nosrati — 2026-01-26

Purpose: This study aims to investigate the relationship between the Modulation Complexity Score (MCS) and the number of Monitor Units (MUs), number of segments, and gantry angles.

Materials and Methods: Treatment planning was performed for 60 patients with head and neck tumors using the step-and-shoot IMRT technique on the RayStation Treatment Planning System (TPS). Treatment plans were divided into two groups, including 30 simple plans (group 1) and 30 complex plans (group 2). Then the relationship between the MCS and the number of Monitor Units (MUs), the number of segments, and the MCS per beam for different gantry angles in the two groups and all plans was investigated.

Results: The Pearson correlation results for both groups and all plans showed a strong relationship between the number of MUs and the MCS (p<0.001). This indication of the strong correlation between MCS and MU in head and neck treatment plans for the first group plans shows a better correlation with the MU. The Pearson correlation results for both groups showed a strong relationship between the number of segments and the MCS (p<0.001). The lowest MCS value or the highest complexity was related to the angles of 161-180 degrees, and the highest MCS value or the lowest level of complexity was for the gantry angles of 281-300 degrees.

Conclusion: The correlation between the number of MU, the number of segments, and the MCS in head and neck plans shows that these items can be used to control complexity and reduce dose uncertainties.

The Structural and Functional Connectivity of the Pineal Gland, and its Age and Gender Associations

Pejman Kiani — 2026-01-26

Purpose: The Pineal Gland (PG) is a structure located in the midline of the brain, and is considered the main part of the epithalamus. There are reports on the role of this area for brain function by hormone secretion, as well as few reports on its role in brain cognition. However, little knowledge is available on the structural and functional connectivity of the PG with other brain regions, as well as its age and gender associations.

Materials and Methods: In this work, we used the diffusion and resting-functional MRI data of 282 individuals, in the age range of 19 to 76 years old. All participants were checked for their medical and mental health by a general practitioner, and the MRI data were collected using a 3 Tesla scanner. The diffusion data were analyzed using the Explore DTI software (version 8.3), and the fMRI data were analyzed using the CONN toolbox (version 18.0).

Results: Two white matter tracts connecting the PG Body to PG Roots and PG to Pons were extracted in this study. The mean FA of the two tracts were 0.26 ± 0.06 and 0.24 ± 0.08, respectively. Neither the FA values of the tracts nor their lengths, showed any associations with age and gender; However, with increasing age, the likelihood of successfully identifying the PG-Pons tract decreased. In functional connectivity analysis, five brain regions showed positive connectivity with the PG, including the superior temporal gyrus, middle temporal gyrus, brain stem, vermis, and the subcallosal cortex, and 25 regions showed negative connectivity. These connectivities did not show an association with gender, but some associations with age were observed.

Conclusion: This study is novel in estimating the functional and structural connectivity of the PG with other brain areas, and also in assessing the association of these connections with age and gender, which could help to increase our knowledge on the functional neuroanatomy of the pineal gland.

Evaluation of Micro-Shear Bond Strength of Composite Resin Repairs with Different Adhesives and Surface Treatments

Negar Soltanian — 2026-01-26

Purpose: In dental operations, repairing old restorations is a typical clinical procedure. This study aimed to evaluate the micro-Shear Bond Strength (µSBS) of composite resin repairs with different adhesives and surface treatments.

Materials and Methods: After preparation, ninety resin composite discs were divided into three groups of thirty at random: no surface preparation, diamond milling roughness, and sandblasting. After 5000 heat cycles, each group was randomly divided into three subgroups of Single Bond (3M), Composite Primer (GC), and Schotch Bond Universal (3M) (n = 10). One-hundred eighty composite cylinders of the new composite were prepared by squeezing the composite into a silicon tube. The samples were then subjected to 5000 heat cycles. After thermocycling, µSBS tests were done at a cross head speed of 0.5 mm/min. Tukey tests and two-way ANOVA were employed to analyze the data.

Results: In the unprepared group, the universal bond and composite primer micro-shear bond strength were significantly higher than the single bond group (p < 0.05). In the milling group, the universal bond micro-shear bond strength was significantly higher than the composite primer and single bond group (p < 0.05). In the sandblasted group there were no significant differences in μSBS among adhesives. In single bond adhesive, the micro-shear bond strength of milling was significantly greater than the sandblasted and unprepared groups (p < 0.05). In the universal adhesive group, the micro-shear bond strength of the milling group was significantly higher than the sandblasted and unprepared groups (p < 0.05).

Conclusion: The type of adhesive and the method of surface preparation have an impact on micro-shear bond strength. The greatest micro-shear bond strength was demonstrated by universal bond application combined with milling roughening.

Chest Wall Motion Tracking By Contactless Optical Single Camera-Based Method Using Virtual Markers, a Feasibility Study

Mohammad Ali Bijari — 2026-01-27

Purpose: Real-time motion tracking of the thorax region of patient's body is a main issue in various parts of medical fields, such as radiotherapy. Several strategies were proposed by using different monitoring hardware. In this work, a contactless method is proposed using an optical camera to trace breathing motion by implementing virtual markers defined on the chest area. A detailed algorithm has been developed to analyze the video frames and track each virtual point in real real-time fashion.

Materials and Methods: In this work, Python program and its OpenCV library have been utilized for breathing motion, two-dimensionally. Database utilized in this work is motion data taken from the breathing motion of a real volunteer. The motion data was captured using a cellphone optical camera, and the gathered data was transferred to the in-room computer system by means of WiFi. It’s worth mentioning that 15 virtual test points were determined using Artificial Intelligence concept of Python inside the chest area.

Results: Final results show that the performance accuracy of the monitoring proposed idea is acceptable. The chest area is determined automatically and will be variable for each patient, uniquely. Various normal and deep breathings were tested in real time at different respiration frequencies. For example, two-dimensional motion displacements of a test point are 4.75 and 7.15 mm for normal and deep breathing, respectively.

Conclusion: The main robustness of the proposed motion tracking method is simplicity, contactless, and using virtual markers determination, while real infra-red markers are currently used clinically by being located on patient's chest skin.

A Straightforward Approach to fNIRS Channel Selection for Distinguishing Mental States from Resting States: Effective in both Subject-Dependent and Subject-Independent Classification Models

Ateke Goshvarpour — 2026-01-27

Purpose: Functional Near-Infrared Spectroscopy (fNIRS) is a relatively novel tool that measures local hemodynamic changes, including oxygenated hemoglobin [Oxy-Hb], deoxygenated hemoglobin [Deoxy-Hb], and total hemoglobin [Tot-Hb]. Its safety, portability, non-invasiveness, and cost-effectiveness make it a preferred technique for designing Brain-Computer Interfaces (BCIs). This study aims to develop an accurate fNIRS-based BCI module for classifying mental tasks and the resting state.

Materials and Methods: Rather than relying on conventional statistical features, our approach utilizes nonlinear indices derived from a 2D Poincaré plot. These measures are computationally efficient and capable of revealing the underlying dynamics of the system. Our primary innovation lies in the development of a novel feature and selection method. We assessed mental task recognition in both subject-dependent and subject-independent classification modes.

Results: Our findings demonstrated a maximum accuracy of 93.75% for subject-specific style and 91.67% for subject-independent style.

Conclusion: In summary, the simplicity and high performance of the proposed framework suggest promising future directions for designing online fNIRS-based BCI systems.

Improving Reinforcement Learning Algorithm Based on Non-Negative Matrix Factorization Method for Controlling an Arm Model

Elham Farzaneh Bahalgerdy — 2026-01-27

Purpose: Reinforcement Learning (RL) is attracting great interest because it enables systems to learn by interacting with the environment. This study aims to enhance the RL algorithm to become more similar to human motor control by combining it with the Non-negative matrix factorization (NMF) method.

Materials and Methods: In the study, the signals recorded from six muscles involved in arm-reaching movement without carryinga certain weight.were pre-processed, and the optimal number of synergy patterns was extracted using NMF and the Variance Account For (VAF) methods. This, in turn, contributes to reducing the calculations. Subsequently, the robustness of the two-link arm model with six muscles was evaluated under various noise levels applied to the action coefficient matrix. Finally, the average synergy pattern was done on the mentioned arm model, and the RL algorithm controlled it by producing the action coefficient matrix.

Results: The average VAF% was 97.25±0.45%, and the number of synergies was four. The tip-of-the-arm model was able to reach the target after an average of 100 episodes.

Conclusion: The results indicated that the similarity in the extracted synergy patterns helps to model a system that is more similar to motor control. Additionally, the results of the synergistic patterns revealed that the two-link arm model with six muscles was suitable for the model. While controlling the model with the RL algorithm, the desired end-point position and path were achieved.

Detection and Classification of Automated Brain Stroke Lesion with Optimized Dual Stage Deep Stacked Auto-Encoder

Sunil Babu Melingi — 2026-01-27

Purpose: The brain is the main organ of the human body. Stroke, also known as cerebral thrombosis, is a medical condition in which a rupture occurs in the blood vessels in the brain, resulting in brain damage. Symptoms may appear when the brain's blood flow and other nutrients are disrupted. A stroke is a medical emergency that can lead to long-term neurological impairment, resulting in complications and, in some cases, fatalities.

Materials and Methods: According to the World Health Organization, stroke is the primary determinant of mortality and disability globally. The early identification of various cardiovascular warning signs can reduce the impact of a stroke. The brain stroke dataset is used in existing methods. Delimitation of classifying stroke is difficult because the complication of lesion shapes and acquiring a ground truth is problematic, as it requires high clinical expertise and anatomical knowledge. In this manuscript, a Detection and Classification of Automated Brain Stroke Lesion with optimized Dual Stage Deep Stacked Auto-Encoder is proposed to detect brain stroke at an early stage with great accuracy.

Results: The input image is taken from slice-level Non-Contrast CT pictures dataset. The collected images are pre-processed and enhanced by removing skull regions, and then the rotations are performed by midline symmetry process. The preprocessed ROI region is fed to feature extraction, and the features are extracted using the wavelet domain. Next, the extracted features are given to classification and classified using Dual stage deep stacked auto encoder (DS-DSAE) optimized with Evolved Gradient Descent Optimization (EGDO) to effectively classify acute infarct, chronic infarct and ischemic stroke, haemorrhagic stroke, and normal.

Conclusion: The goal is to reduce computing complexity and enhance accuracy. The performance of the proposed wavelet-DS-DSAE-EGDO method achieves High accuracy 30.56%, 12.32%, 15.6%, 16.6%, 25.6%, 32.2%; High Precision 28.74%, 32.2%, 14.5%, 16.55%, 17.8%, 23.4% is comparing with the existing methods

Novel Method to Estimate Kinetic Microparameters from Dynamic Whole-Body Imaging in Regular-Axial Field-of-View PET Scanners

Kyung-Nam Lee — 2026-01-27

Purpose: For Whole-Body (WB) kinetic modeling based on a typical PET scanner, a multi-pass multi-bed scanning protocol is necessary given the limited axial field-of-view. Such a protocol introduces loss of early-dynamics in Time-Activity Curves (TACs) and sparsity in TAC measurements, inducing uncertainty in parameter estimation when using Least-Squares Estimation (LSE) (i.e., common standard), especially for kinetic microparameters. We present a method to reliably estimate microparameters, enabling accurate parametric imaging, on regular-axial field-of-view PET scanners

Materials and Methods: Our method, denoted Parameter Combination-Driven Estimation (PCDE), relies on the generation of reference truth TAC database, and subsequently selected, the best parameter combination as the one arriving at TAC with the highest Total Similarity Score (TSS), focusing on the general image quality, overall visibility, and tumor detectability metrics. Our technique has two distinctive characteristics: 1) improved probability of having one-on-one mapping between early and late dynamics in TACs (the former missing from typical protocols), and 2) use of multiple aspects of TACs in the selection of best fits. To evaluate our method against conventional LSE, we plotted trade-off curves for noise and bias. In addition, the overall Signal-to-Noise Ratio (SNR) and spatial noise were calculated and compared. Furthermore, the Contrast-to-Noise Ratio (CNR) and Tumor-to-Background Ratio (TBR) were also calculated. We also tested our proposed method on patient data (¹⁸F-DCFPyL PSMA PET/CT scans) to further verify clinical applicability.

Results: Significantly improved general image quality performance was verified in microparametric images (e.g. noise-bias trade-off performance). The overall visibility and tumor detectability were also improved. Finally, for our patient studies, improved overall visibility and tumor detectability were demonstrated in mico parametric images, compared to the use of conventional parameter estimation.

Conclusion: The proposed method provides improved microkinetic parametric images compared to the common standard in terms of general image quality, overall visibility, and tumor detectability.

Perceptual Decisions Recognition in Healthy Individuals Using Electroencephalogram Signals

Ali Barzegar Khanghah — 2026-01-27

Purpose: Making a decision based on available sensory information is called “Perceptual Decision-Making”. Since the uncertainty and difficulty in individuals' perceptual decision-making can create many adverse effects in their personal and social lives, research in this field seems necessary to achieve a more comprehensive understanding of the brain during perceptual decision-making. Despite numerous studies in this field, no robust system can objectively recognize people's perceptual decisions. This study investigates healthy individuals' Electroencephalogram (EEG) signals during a perceptual decision-making task to fill this research gap.

Materials and Methods: The research employs an online EEG dataset based on visual stimuli, including faces and cars, obtained from 16 participants. After preprocessing the EEG signals, 26 features were extracted from the signals to explore the impact of coherence and spatial prioritization of stimulus on the decision-making process using Friedman’s non-parametric statistical analysis. Then, a Fuzzy Radial Basis Function (FRBF) network with the extracted features from TP9 and TP10 channels as input was utilized to classify the data based on the uncertainty of the processes in the brain.

Results: The statistical analysis revealed that differences in the coherence of the stimulus representations have a significant (P-value < 0.05) greater impact on an individual's decision-making process than spatial prioritization. Also, the FRBF network classifier achieved an accuracy of 90.3% in classifying the test data as either a "Face" or "Car.

Conclusion: The classification accuracy results showed that the proposed method is an effective procedure for recognizing human decisions.

Using Transfer Learning Approach for Down Syndrome Features Extraction and Data Augmentation for Data Expansion

Farah F. Alkhalid — 2026-01-27

Purpose: People with Down Syndrome must be served specially because they have an intellectual disability with abnormality in memory and learning, so creating a model for DS recognition may provide safe services to them, using the transfer learning technique can improve high metrics with a small dataset, depending on previous knowledge, there is no available Down syndrome dataset, one can use to train.

Materials and Methods: A new dataset is created by gathering images, two classes (Down=209 images, non-Down=214 images), and then expanding this dataset using Augmentation to create the final dataset of 892 images (Down=415 images, Non-Down=477 images). Finally, using a suitable training model, in this work, Xception and Resnet models are used, and the pre-trained models are trained on Imagenet dataset, which consists of (1000) classes.

Results: By using the Xception model and the Resnet model, it is concluded that when using the Resnet model the accuracy is 95.93% and the loss function is 0.16, while by using the Xception model, the accuracy is 96.57% and the loss function is 0.12.

Conclusion: Transfer learning is used to overcome the suitability of dataset size and minimize the cost of training, and time processing the accuracy and loss function is good when using the Xception model, in addition, the Xception metrics are the best compared with the previous studies.

Dental X–Ray Images for Automated Detection of Caries Classes Using Deep Learning Techniques

Sindu Divakaran — 2026-01-27

Purpose: Dental caries can emerge anywhere in the mouth, particularly in the interior of the cheeks and the gums. Some of the indications are patches on the inner lining of the mouth, along with bleeding, toothache, numbness, and an unusual red and white staining. Hence, it is important to predict the presence of a cavity at an early stage. The currently available manual method is inefficient and hence we provide an advanced method by using the deep learning concepts.

Materials and Methods: In this work, different types of algorithms such as Res Net, Deeper Google Net, and mini VGG Net are to be used to predict the class of cavity at an early stage.

Results: A comparison between the accuracy of three different algorithms is given in this paper. Thus, by using efficient deep learning algorithms, it will be able to predict the presence of the cavity and the class of the cavity at an early stage and take the necessary steps to overcome it.

Conclusion: In this work, a comparison between three different algorithms is given and proved that the efficient algorithm is the inception algorithm among the other algorithms that achieves an accuracy of about 98%, which is suitable for use in hospitals.

A Regional Effective Dose, Risk of Exposure-Induced Death, and Annual Per Capita Dose in Diagnostic Radiology Procedures

Mansoreh Zarei — 2026-01-27

Purpose: Ionizing radiation exposure doses during radiological procedures may increase the patient dose; therefore, dose assessment is an important subject. The current study aimed to estimate the Effective Dose (ED), Risk of Exposure-Induced Death (REID), as well as Annual Per Capita Dose (APCD) in routine radiography procedures in Yazd province (Iran).

Materials and Methods: The data related to the exposure parameters and Entrance Surface Air Kerma (ESAK) of 9 public high-patient-load radiography centers (11 radiology devices) were collected from 783 patients. Five routine planar radiological examinations were included: lumbar spine, pelvis, abdomen, chest, and skull. The ED and REID values for each device and examination were obtained using a personal computer-based Monte Carlo (PCXMC, v. 2.0) software. The APCD was estimated by dividing the annual collective effective dose (ACED) to the Yazd population.

Results: The estimated mean ESAK values ranged from 0.26±0.11 mGy (chest examination) to 8.45±5.3 mGy (lumbar examination). The lumbar spine examination had the highest ED value (1.02 ± 0.75 mSv). The highest REID value for abdominal, chest, lumbar, pelvic, and skull examinations is associated with stomach (6.58±7.72), lung (2.36±2.79), stomach (7.03±6.11), colon (3.31±5.49), and other cancers (0.58±0.56). The ACED value due to the radiology examinations was obtained at 45.782 man-Sv.

Conclusion: Our results demonstrated that the dose variations among the patients were remarkably high. Choosing appropriate imaging parameters, reducing the frequency of unnecessary radiology examinations, and performing quality control procedures of radiology machines could reduce the patients' doses.

Establishing Pediatric Thoracic Radiography Diagnostic Reference Levels Using CALDOSE_X: A Data-Driven Approach to Optimize Radiation Safety

Ikbal Bouadel — 2026-01-27

Purpose: Disparities exist in adherence to national radiation safety standards in Morocco, particularly in pediatric conventional radiology. This cross-sectional study aims to establish Moroccan diagnostic reference levels (DRLs) for pediatric thorax radiography.

Materials and Methods: Thorax radiographs of 208 pediatric patients (newborns to 18 years old) from four Moroccan public hospitals were analyzed. Patient demographics (age, gender, weight) and scan parameters were recorded to calculate radiation doses using CALDOSE_X 5.0 software, focusing on entrance surface air kerma (ESAK, mGy) and kerma-area product (KAP, mGy·cm²). Patients were categorized into five age groups (<1 month, 1 month ≤ age < 4 years, 4 years ≤ age < 10 years, 10 years ≤ age < 14 years, and 14 years ≤ age < 18 years). The third quartile (P75) of ESAK and KAP were determined as DRLs. Statistical analyses were performed using SPSS v.21, with p < 0.05 indicating significance.

Results: The P75 values of ESAK and KAP across age groups were 0.61, 0.69, 0.68, 0.82, and 1.29 for ESAK and 350.25, 566.07, 499.14, 950.62, and 1816.06 for KAP. The regional DRLs exceeded those reported in some European countries, likely due to differences in imaging protocols, patient positioning, and exposure parameters. Additionally, irradiated surface area significantly influenced dose variation in patients under 10 years (p < 0.01).

Conclusion: Establishing Moroccan pediatric DRLs highlights the need for dose optimization in pediatric radiography. Optimizing irradiated surfaces and exposure parameters while ensuring adherence to international DRL recommendations is essential to enhance radiation safety in pediatric imaging.

Explainable Artificial Intelligence in Nuclear Medicine: Advancing Transparency in PET and SPECT Imaging and Radiation Therapy

Hossein Arabi — 2026-01-27

The integration of Artificial Intelligence (AI) into nuclear medicine has transformed diagnostic and therapeutic processes, yet the opaque nature of many AI models hinders clinical adoption and trust. This narrative review aims to synthesize the current landscape of explainable AI (XAI) in nuclear medicine, emphasizing its role in enhancing transparency, bias mitigation, and regulatory compliance for robust clinical integration. Key chapters cover the fundamentals of XAI in nuclear medicine; XAI applications in PET and SPECT instrumentation and acquisition; image reconstruction; quantitative imaging and corrections; post-reconstruction processing and analysis; and radiotherapy. The review concludes with a discussion of challenges, limitations, and future directions, advocating for interdisciplinary advancements to bridge AI innovation with practical utility in patient care.

Terahertz Computed Computed Tomography and Imaging Challenges

Masoumeh Mohamadian — 2026-01-27

Purpose: Terahertz imaging has emerged as a promising technique for non-destructive evaluation and imaging applications, offering unique advantages over traditional imaging modalities. This paper presents an overview of the current state of Terahertz Computed Tomography and highlights the challenges faced in its implementation.

THz-CT utilizes electromagnetic waves in the terahertz frequency range to reconstruct three-dimensional images of objects with high resolution and penetration capabilities. The ability to visualize internal structures without the use of ionizing radiation has significant implications for various fields.

Materials and Methods: Despite its potential, it faces several challenges that need to be addressed for its widespread adoption. Firstly, the limited availability and complexity of THz sources and detectors hinder the practical implementation of THz-CT systems. Efforts are being made to develop compact, efficient, and cost-effective THz sources and detectors to overcome these limitations.

Secondly, THz waves are highly susceptible to scattering and absorption by various materials. This poses challenges in achieving accurate and artifact-free reconstructions, especially in applications involving biological samples.

Results: Furthermore, the relatively long acquisition times required for THz-CT imaging limit its real-time applications. Efforts are underway to develop faster acquisition methods to reduce acquisition times while maintaining image quality.

Lastly, the lack of standardized protocols and benchmarks for THz-CT imaging hinders the comparison and reproducibility of results across different systems and studies.

Also, in addition to its various applications, terahertz medical imaging and medical microbiological detection play a significant role in the diagnosis of several types of cancers.

Conclusion: In conclusion, THz-CT holds great promise for various imaging applications, but several challenges need to be overcome for its widespread adoption. Addressing the limitations associated with THz sources, scattering and absorption effects, acquisition times, and standardization will pave the way for the realization of the full potential of THz-CT in the future.

Integration of Multimodal Large Language Models in Medical Imaging and Omics Data: A Comprehensive Review

Raja Vavekanand — 2026-01-27

Purpose: This review focuses on how Multimodal Large Language Models (MLLMs) and multimodal AI models are advancing healthcare by integrating medical imaging and omics data. By integrating imaging techniques such as MRI, CT, and PET with genomics, transcriptomics, and proteomics, these models offer a comprehensive understanding of diseases, particularly in areas like cancer diagnosis and treatment. The study also highlights the challenges of managing complex datasets and ensuring effective feature selection.

Materials and Methods: Analysed studies leveraging advanced AI models, such as Convolutional Neural Networks (CNNs) and Multimodal Neural Networks (MM-Nets), to integrate diverse data sources. These models enhance medical imaging with omics data to improve disease prediction and management. Applications reviewed include cancer subtype classification, survival outcome prediction, and precision medicine, with a particular focus on non-invasive diagnostic tools.

Results: The findings underscore the transformative potential of multimodal healthcare. They significantly improve the identification of biomarkers and enable personalized treatment approaches. For instance, models like VGG19-CNN and PAGE-Net demonstrated higher accuracy in predicting cancer-specific outcomes and integrating genomic and imaging data. Moreover, the applications to single-cell analysis and radiomics showcased their ability to uncover molecular-level insights, advancing precision medicine.

Conclusion: represents a breakthrough in healthcare, combining diverse data types to deliver actionable insights for disease management. While challenges such as handling complex datasets and ensuring model transparency remain, ongoing advancements in AI technologies are paving the way for their wider adoption. These models hold immense promise for improving diagnostics, guiding treatment strategies, and enhancing patient outcomes, marking a significant step toward the era of personalized medicine.

The Future of Surgery: Embracing Robotic Systems in Surgical Practice

Mehrdad Ghaderi — 2026-01-27

Robotic surgery has transitioned from a phenomenon to the norm in the medical field, particularly in minimally invasive surgery. Robotic-assisted surgery offers greater precision, quicker recovery, and better patient outcomes, but issues like astronomical costs, technical issues, and ethical issues prevent its adoption. Robotic surgery's advantages—greater precision, less invasive procedures, and better clinical outcomes—are outlined here while addressing issues to its adoption. New technologies like AI integration, autonomous technology, and tele-surgery are revolutionary but will have to be accompanied by strong regulatory frameworks. Technologists', clinicians', and policy makers' collaboration is important to patient safety and equitable access as robot surgery advances.