Mohammed Alharbi

Background/Objectives: Non-small cell lung cancer (NSCLC) is the most common type of lung cancer, representing nearly 85% of cases worldwide. Predicting patient response before treatment, however, remains a major clinical challenge. Radiomics enables non-invasive extraction of quantitative imaging data that reflects tumor characteristics and heterogeneity. In this study, we developed and externally validated a CT-derived radiomic signature for predicting outcomes in advanced NSCLC patients treated with first-line platinum-based chemotherapy. Methods: NSCLC across three tertiary care centers were included. Tumor lesions from baseline contrast-enhanced CT scans were semi-automatically outlined using 3D Slicer software (version 5.12.2), followed by the extraction of 851 quantitative imaging biomarkers through PyRadiomics version 3.0. The extracted parameters comprised histogram-based features, morphological measurements, texture-related variables, and wavelet-derived attributes. Reliability of feature extraction was evaluated using intraclass correlation coefficient analysis, whereas LASSO regression together with stability selection was applied to identify the most relevant predictors. Predictive capability was assessed across five machine learning techniques, including logistic regression, random forest, support vector machine, gradient boosting, and multilayer perceptron models. The resulting radiomics-based composite score was then tested in an independent external validation cohort. Response to treatment was determined according to RECIST 1.1 guidelines. Results: After feature reproducibility assessment and stability screening, 421 radiomic parameters were retained for further analysis and 14 feasible parameters were selected by LASSO. There was good predictive power of the integrated model that included both radiomic and clinical features, with AUC values of 0.876, 0.849, and 0.831 in the training set, internal validation set and external validation set, respectively. The radiomic signature successfully differentiated patients according to their probability of therapeutic response. Furthermore, by using decision-curve analysis, the combined radiomics nomogram was more clinically useful than models based on clinical characteristics within clinically relevant decision-threshold ranges with greater net benefit. Conclusions: This pre-treatment CT-based radiomics signature demonstrates potential as a decision-support tool for chemotherapy-response prediction in NSCLC. However, prospective and multi-ethnic validation is required before clinical application. Based on this data, larger multi-ethnic cohorts should be considered for prospective validation.

Background and aim: Chest CT is a major contributor to population radiation exposure. Conventional, pooled diagnostic reference levels (DRLs) do not account for inter-individual variability in body habitus and are typically used retrospectively. We evaluated dose behavior in adult chest CT, derived BMI-stratified local DRLs, and developed models to enable AI-assisted, prescan dose prediction. Methods: Consecutive adult chest CT examinations from a single center were analyzed. Dose indices (CTDIvol, DLP) and patient factors (BMI, weight, height, age, sex; scan length and planned technical parameters where available) were extracted. DRLs were defined as the 75th percentile overall and within BMI categories (underweight, normal, overweight, and obese). Group differences were assessed using non-parametric tests; associations were examined using correlation analysis. Supervised learning (e.g., Random Forest, Gradient Boosting) was trained to predict CTDIvol and DLP from routinely available variables. Results: BMI-stratified DRLs increased monotonically with habitus: underweight 444.95 mGy·cm/9.60 mGy; normal 513.00/11.55; overweight 756.08/14.65; obese 931.60/20.25 (DLP/CTDIvol). Differences across BMI groups were significant for DLP (H = 31.53, p < 0.001) and CTDIvol (H = 33.61, p < 0.001). DLP correlated moderately with weight and BMI (r ≈ 0.54–0.56, p < 0.001), with a weaker association for age; height was not a meaningful predictor. No sex-based differences in CTDIvol or DLP were observed. Predictive models estimated CTDIvol and DLP with high performance (R² up to ~0.79 and ~0.77, respectively), enabling comparison of predicted dose against BMI-matched DRLs before acquisition. Conclusions: Size-aware, BMI-stratified DRLs provide clinically interpretable investigation levels that avoid pitfalls of pooled benchmarks. Coupled with robust prediction of individualized dose from routine variables, this framework supports a shift from retrospective audit to prospective, point-of-care dose governance and protocol optimization in chest CT.

Background: Repeated brain computed tomography (CT) scans in children may result in substantial cumulative radiation exposure, particularly in young children, who are more sensitive to ionizing radiation. The purpose of the study was to assess the dose levels of radiation in patients who receive repeated brain CT during childhood and adherence rates to pediatric imaging protocols. Methods: A retrospective cross-sectional study was conducted among 177 patients aged ≤5 years who underwent two or more brain CT examinations with a total of 514 CT examinations. The information was gathered through the hospital Picture Archiving and Communication System (PACS), which included patient demographics, scan parameters, and scanner-reported dose indicators such as volume-averaged computed tomography dose index (CTDIvol) and dose-length product (DLP). The effective dose (ED) was calculated and compared with estimated doses based on a nominal pediatric CT protocol. Results: The findings indicated a great variation in scan parameters, with CTDIvol values of 8.9 to 51.7 mGy and DLP values of 177 to 1310 mGy.cm. The number of repeated scans showed a great increase in the cumulative ED (p < 0.001). The median doses in patients below the age of one year were greater than those in older children. There was also a closer relation of scanner-reported doses to adult protocols, which suggests a lack of an optimized pediatric setting. Conclusions: Children under 5 who undergo repeated brain CT scans may face excessive radiation exposure. The matter is aggravated by the fact that scans are performed repeatedly without optimization of the dose, which leads to significant cumulative ED.

Background: Epidural procedures benefit from a pre-procedural informed estimation of epidural depth, as anticipating the approximate distance can support safer needle placement and reduce technical difficulties during analgesia or anesthesia procedures. The influence of ethnicity has been established across different populations worldwide; however, there is a lack of Saudi-specific MRI data on epidural depth among the adult population. Aim of this Study: To measure the skin to epidural space distance (SED) at the lumbar interspaces L3–L4 and L4–L5 in the Saudi adult population using magnetic resonance imaging (MRI) and to examine its correlations with age, sex, height, weight, and body mass index (BMI). Methods: In this retrospective cross-sectional study, sagittal T1-weighted lumbar MRI images of the spine of 169 adult Saudi patients were studied. The age group ranged from 20 to 70 years, with an equal distribution of males and females. The skin to epidural space distance (SED) was measured at the L3–L4 and L4–L5 interspaces, and its correlations with age, sex, height, weight, and BMI were analyzed. Results: The average measurement of skin to epidural space distance (SED) was 59.08 mm in L3–L4, and 63.21 in L4–L5. BMI and weight showed strong positive correlations with SED across both levels. Female sex was associated with longer SED values at L4–L5. There was no significant correlation between SED and age or height of the patients. Conclusions: MRI-based assessment of SED revealed strong correlations with weight and BMI, but no correlation with height, age, and sex. These findings support the individualized estimation of epidural depth and needle length selection to enhance procedural safety and reduce complications.

The majority of data collected and obtained from various sources over a patient’s lifetime can be assumed to comprise pertinent information for delivering the best possible treatment. Medical data, such as radiographic and histopathology images, electrocardiograms, and medical records, all guide a physician’s diagnostic approach. Nevertheless, most machine learning techniques in the healthcare field emphasize data analysis from a single modality, which is insufficiently reliable. This is especially evident in radiology, which has long been an essential topic of machine learning in healthcare because of its high data density, availability, and interpretation capability. In the future, computer-assisted diagnostic systems must be intelligent to process a variety of data simultaneously, similar to how doctors examine various resources while diagnosing patients. By extracting novel characteristics from diverse medical data sources, advanced identification techniques known as multimodal learning may be applied, enabling algorithms to analyze data from various sources and eliminating the need to train each modality. This approach enhances the flexibility of algorithms by incorporating diverse data. A growing quantity of current research has focused on the exploration of extracting data from multiple sources and constructing precise multimodal machine/deep learning models for medical examinations. A comprehensive analysis and synthesis of recent publications focusing on multimodal machine learning in detecting diseases is provided. Potential future research directions are also identified. This review presents an overview of multimodal machine learning (MMML) in radiology, a field at the cutting edge of integrating artificial intelligence into medical imaging. As radiological practices continue to evolve, the combination of various imaging and non-imaging data modalities is gaining increasing significance. This paper analyzes current methodologies, applications, and trends in MMML while outlining challenges and predicting upcoming research directions. Beginning with an overview of the different data modalities involved in radiology, namely, imaging, text, and structured medical data, this review explains the processes of modality fusion, representation learning, and modality translation, showing how they boost diagnosis efficacy and improve patient care. Additionally, this review discusses key datasets that have been instrumental in advancing MMML research. This review may help clinicians and researchers comprehend the spatial distribution of the field, outline the current level of advancement, and identify areas of research that need to be explored regarding MMML in radiology.

Background: Spectral detector computed tomography (SDCT) has demonstrated superior diagnostic performance and image quality in liver disease assessment compared with traditional CT. Selecting the right reconstruction algorithm and tube voltage is essential to avoid increased noise and diagnostic errors. Objectives: This study evaluated improvements in image quality achieved using various virtual tube voltages and reconstruction algorithms for diagnosing common liver diseases with spectral CT. Methods: This retrospective study involved forty-seven patients who underwent spectral CT scans for liver conditions, including fatty liver, hemangiomas, and metastatic lesions. The assessment utilized signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR), with images reconstructed using various algorithms (IMR, iDose) at different levels and virtual tube voltages. Three experienced radiologists analyzed the reconstructed images to identify the best reconstruction methods and tube voltage combinations for diagnosing these liver pathologies. Results: The signal-to-noise ratio (SNR) was highest for spectral CT images using the IMR3 algorithm in metastatic, hemangioma, and fatty liver cases. A strong positive correlation was found between IMR3 at 120 keV and 70 keV (p-value = 0.000). In contrast, iDOSE2 at 120 keV and 70 keV showed a low correlation of 0.291 (p-value = 0.045). Evaluators noted that IMR1 at 70 keV provided the best visibility for liver lesions (mean = 3.58), while IMR3 at 120 keV had the lowest image quality (mean = 2.65). Conclusions: Improvements in image quality were noted with SDCT, especially in SNR values for liver tissues at low radiation doses and a specific IMR level. The IMR1 algorithm reduced noise, enhancing the visibility of liver lesions for better diagnosis.

Background: Magnetic resonance imaging (MRI) exams may cause patients to feel anxious before or during the scan, which affects the scanning outcome and leads to motion artifacts. Adequate preparation can effectively alleviate patients’ anxiety before the scan. We aimed to assess the effect of different preparation methods on MRI-induced anxiety: We conducted a prospective randomized study on MRI patients between March and May 2022. We divided 30 patients into two groups: the control group, which received routine preparation (RP), and the experimental group, which received video preparation (VP). We used the State-Trait Anxiety Inventory (STAI) to measure anxiety levels before and after the interventions. We assessed patients’ self-satisfaction after the scan: After preparation, VP (STAI mean = 10.7500) and RP (STAI mean = 12.7857), we observed a significant association between the pre- and post-STAI results in VP (p = 0.025). The effects of both methods in decreasing anxiety were more significant for first-timers (p = 0.009 in RP/0.014 in VP). We noted high satisfaction levels for both forms of preparation. The VP technique was superior in reducing patient anxiety, especially in first-time MRI patients. Hence, VP techniques can be used in different clinical settings to reduce anxiety and facilitate patients’ understanding of the instructions given.