Lorenzo Ugga

Purpose: Pituitary adenoma consistency significantly influences surgical strategy and outcomes, yet it cannot be reliably predicted using conventional imaging. By converting images into quantitative data, radiomics can identify imaging patterns that may not be distinguishable by standard visual interpretation. This study aims to evaluate the diagnostic performance of radiomics-based MRI models for predicting pituitary adenoma consistency, to determine its potential for presurgical planning. Materials and methods: A systematic search of PubMed, EMBASE, and Scopus was performed on 12/03/2025 following PRISMA-DTA guidelines and a pre-registered protocol (PROSPERO/CRD420251246028). Eligible studies applied radiomics and machine learning/deep learning to predict pituitary adenoma consistency. We performed random-effects meta-analyses to evaluate the area under the receiver operating characteristic (ROC) curve (AUC). A hierarchical summary ROC (HSROC) model estimated sensitivity and specificity. Risk of bias and study quality were assessed with QUADAS-2 and METRICS. Results: Fourteen studies were included. The pooled discriminative performance was good (AUC = 0.86,95% C.I. [0.76;0.92], I ² = 92.45%, k = 14). The HSROC model showed a sensitivity = 0.74,95% C.I. [0.56;0.87] and specificity = 0.79, 95% C.I. [0.73;0.84]. No significant performance differences were observed across algorithms, dimensionality, sequence, although T2w and combined sequence models showed higher AUC trends. Risk of bias was low to moderate, and study quality was good overall. Heterogeneity resulted high and small-study effect emerged in the main analysis. Conclusions: Radiomics-based MRI models demonstrate good diagnostic performance for predicting pituitary adenoma consistency and may support presurgical assessment and planning. However, substantial heterogeneity, evidence of small-study effects, and limited external validation warrant cautious interpretation of the pooled estimates and currently restrict clinical generalizability. Future studies should prioritize multicenter external validation and standardized radiomics workflows.

Background: Quantitative susceptibility mapping (QSM) is an advanced magnetic resonance imaging (MRI) technique that quantifies tissue magnetic susceptibility, offering non-invasive insights into brain microstructure, including iron content and myelination. While extensively applied in adult neuroimaging, its use in pediatric populations is rapidly expanding. Purpose: This systematic review aims to provide a comprehensive overview of QSM applications in pediatric brain imaging, highlighting methodological advancements, diagnostic potential, and current limitations. Methods: A systematic literature search was performed using PubMed and Google Scholar up to April 2025. Inclusion criteria were original research articles written in English, involving only pediatric populations (0–17 years) and employing QSM in brain imaging. Twenty studies met eligibility criteria and were analyzed in terms of acquisition protocols, post-processing methods, study objectives, and main findings. Results: A systematic search on PubMed and Google Scholar found 54 QSM brain studies in children; after exclusions, 20 original research papers qualified for review and were quality-checked using Quality Assessment of Diagnostic Accuracy Studies version-2 (QUADAS-2). Most studies were recent (85% in the last 5 years), in Asia (55%, with China 35%), and used 3-tesla (T) MRI (80%). Typical imaging parameters: 8 echoes (TE=40 ms), slice thickness=2–2.5 mm, matrix often 256×256; Laplacian was the main phase-unwrapping method and variable-kernel sophisticated harmonic artifact reduction for phase data (VSHARP) the dominant background-field removal. Study aims clustered into improved detection, microstructural analysis, normative comparisons, clinical correlations, developmental patterns, and pathology tracking. Conclusion: QSM emerges as a valuable tool in pediatric neuroimaging, offering quantitative biomarkers for brain development, disease monitoring, and potential clinical translation. Despite promising results, challenges remain, including motion artifacts, lack of normative pediatric data, and methodological heterogeneity. Future research should focus on longitudinal designs, standardization of protocols, and integration with complementary imaging modalities. With further refinement, QSM has the potential to become an integral component of pediatric neuroradiological assessment.

The parapharyngeal space is a complex anatomical site in the head and neck which may harbour clinically occult pathology given its deep-seated location. The vast majority of parapharyngeal space lesions are of intermediate or hyperintense signal on T2W sequences. This review focuses on T2 hypointense parapharyngeal space lesions which are rare and may constitute a diagnostic dilemma. We present the differential diagnosis of these lesions, highlighting the pertinent radiological findings and identifying a histological correlation for the low T2 signal. A brief discussion of the physics principles accounting for these imaging features is also included. We propose a diagnostic algorithm to facilitate diagnosis and avoid unnecessary biopsy, whenever possible.

Background: Radiomics offers quantitative, high-dimensional data from conventional imaging and holds promise for improving diagnosis and treatment of pituitary adenomas (PAs). This systematic review aimed to synthesize current clinical applications of radiomics in PAs, focusing on diagnostic, predictive, and prognostic modeling. Methods: This review followed the PRISMA 2020 guidelines. A systematic search was performed in PubMed, Scopus, and Web of Science on 10 January 2024, and updated on 5 March 2024, using predefined keywords and MeSH terms. Studies were included if they evaluated radiomics-based models using MRI for diagnosis, classification, consistency, invasiveness, treatment response, or recurrence in human PA populations. Data extraction included study design, sample size, MRI sequences, feature types, machine learning algorithms, and model performance metrics. Study quality was assessed via the Newcastle-Ottawa Scale. Descriptive statistics summarized study characteristics; no meta-analysis was performed due to heterogeneity. Results: Out of 341 identified articles, 49 studies met inclusion criteria, encompassing a total of more than 9350 patients. The majority were retrospective (43 studies, 88%). MRI sequences used included T2-weighted imaging (35 studies, 71%), contrast-enhanced T1WI (34 studies, 69%), and T1WI (21 studies, 43%). PyRadiomics was the most common feature extraction tool (20 studies, 41%). Machine learning was employed in 43 studies (88%), predominantly support vector machines (16 studies, 33%), random forests (9 studies, 18%), and logistic regression (9 studies, 18%). Deep learning methods were applied in 17 studies (35%). Regarding diagnostic performance, 22 studies (45%) reported an (AUC) ≥0.85 in test datasets. External validation was performed in only 6 studies (12%). Radiomics applications included histological subtype prediction (14 studies, 29%), surgical outcome prediction (13 studies, 27%), invasiveness assessment (7 studies, 15%), tumor consistency evaluation (8 studies, 16%), and response to medical or radiotherapy treatments (3 studies, 6%). One study (2%) addressed automated segmentation and volumetry. Conclusions: Radiomics enables high-performance, noninvasive prediction of PA subtypes, consistency, invasiveness, treatment response, and recurrence, with 22 studies (45%) reporting AUC ≥0.85. Despite promising results, clinical translation remains limited by methodological heterogeneity, low external validation (6 studies, 12%), and lack of standardization.

BACKGROUND AND PURPOSE: Identifying patients with multiple sclerosis (pwMS) at higher risk of clinical progression is essential to inform clinical management. We aimed to build prognostic models using machine learning (ML) algorithms predicting long-term clinical outcomes based on a systematic mapping of volumetric, radiomic, and macrostructural disconnection features from routine brain MRI scans of pwMS. MATERIALS AND METHODS: In this longitudinal monocentric study, 3T structural MRI scans of pwMS were retrospectively analyzed. Based on a ten-year clinical follow-up (average duration=9.4±1.1 years), patients were classified according to confirmed disability progression (CDP) and cognitive impairment (CI) as assessed through the Expanded Disability Status Scale (EDSS) and the Brief International Cognitive Assessment of Multiple Sclerosis (BICAMS) battery, respectively. 3D-T1w and FLAIR images were automatically segmented to obtain volumes, disconnection scores (estimated based on lesion masks and normative tractography data), and radiomic features from 116 gray matter regions defined according to the Automated Anatomical Labelling (AAL) atlas. Three ML algorithms (Extra Trees, Logistic Regression, and Support Vector Machine) were used to build models predicting long-term CDP and CI based on MRI-derived features. Feature selection was performed on the training set with a multi-step process, and models were validated with a holdout approach, randomly splitting the patients into training (75%) and test (25%) sets. RESULTS: We studied 177 pwMS (M/F = 51/126; mean±SD age: 35.2±8.7 years). Long-term CDP and CI were observed in 71 and 55 patients, respectively. Regarding the CDP class prediction analysis, the feature selection identified 13-, 12-, and 10-feature subsets obtaining an accuracy on the test set of 0.71, 0.69, and 0.67 for the Extra Trees, Logistic Regression, and Support Vector Machine classifiers, respectively. Similarly, for the CI prediction, subsets of 16, 17, and 19 features were selected, with 0.69, 0.64, and 0.62 accuracy values on the test set, respectively. There were no significant differences in accuracy between ML models for CDP (p=0.65) or CI (p=0.31). CONCLUSIONS: Building on quantitative features derived from conventional MRI scans, we obtained long-term prognostic models, potentially informing patients’ stratification and clinical decision-making. ABBREVIATIONS: MS, multiple sclerosis; pwMS, people with MS; HC, healthy controls; ML, machine learning; DD, disease duration; EDSS, Expanded Disability Status Scale; TLV, total lesion volume; CDP, confirmed disability progression; CI, cognitive impairment; BICAMS, Brief International Cognitive Assessment of Multiple Sclerosis.

Objectives: In this retrospective study, we performed a volumetric analysis of paranasal cavity pneumatization in a population of adult patients with cystic fibrosis compared to healthy controls, providing parcel evaluation of each sinus, and analyzing the prevalence of major anatomical sinonasal variants in the two groups. Methods: We compared paranasal sinus volumes of 89 adult patients with cystic fibrosis and 144 healthy controls who underwent paranasal sinus computed tomography. Volumes were segmented and extracted on tomographic images using the freely available software MRIcron 2019, then compared using a t-test; the z-score test was used to determine whether the two groups differ significantly in terms of major anatomical variants prevalence. Results: Overall sinus volumes in patients with cystic fibrosis patients differ significantly as compared to the healthy population (p < 0.00001). Furthermore, with the only exception of ethmoid sinus pneumatization, which was similar in both populations, all the other sinuses were statistically different. No significant difference emerged concerning anatomical variants’ prevalence. Conclusions: Our results further stress the impact of cystic fibrosis on sinus structure in adult patients, better revealing the consequences of the disease on upper airways and in optimizing the management of patients with sinonasal manifestations.

Artificial intelligence (AI) is transforming radiology by improving diagnostic accuracy, streamlining workflows, and enhancing operational efficiency. However, these advancements come with significant sustainability challenges across environmental, economic, and social dimensions. AI systems, particularly deep learning models, require substantial computational resources, leading to high energy consumption, increased carbon emissions, and hardware waste. Data storage and cloud computing further exacerbate the environmental impact. Economically, the high costs of implementing AI tools often outweigh the demonstrated clinical benefits, raising concerns about their long-term viability and equity in healthcare systems. Socially, AI risks perpetuating healthcare disparities through biases in algorithms and unequal access to technology. On the other hand, AI has the potential to improve sustainability in healthcare by reducing low-value imaging, optimizing resource allocation, and improving energy efficiency in radiology departments. This review addresses the sustainability paradox of AI from a radiological perspective, exploring its environmental footprint, economic feasibility, and social implications. Strategies to mitigate these challenges are also discussed, alongside a call for action and directions for future research. Critical relevance statement: By adopting an informed and holistic approach, the radiology community can ensure that AI’s benefits are realized responsibly, balancing innovation with sustainability. This effort is essential to align technological advancements with environmental preservation, economic sustainability, and social equity. Key Points: AI has an ambivalent potential, capable of both exacerbating global sustainability issues and offering increased productivity and accessibility. Addressing AI sustainability requires a broad perspective accounting for environmental impact, economic feasibility, and social implications. By embracing the duality of AI, the radiology community can adopt informed strategies at individual, institutional, and collective levels to maximize its benefits while minimizing negative impacts. Graphical

Objectives: To investigate the intra- and inter-rater reliability of the total methodological radiomics score (METRICS) and its items through a multi-reader analysis. Materials and methods: A total of 12 raters with different backgrounds and experience levels were recruited for the study. Based on their level of expertise, raters were randomly assigned to the following groups: two inter-rater reliability groups, and two intra-rater reliability groups, where each group included one group with and one group without a preliminary training session on the use of METRICS. Inter-rater reliability groups assessed all 34 papers, while intra-rater reliability groups completed the assessment of 17 papers twice within 21 days each time, and a “wash out” period of 60 days in between. Results: Inter-rater reliability was poor to moderate between raters of group 1 (without training; ICC = 0.393; 95% CI = 0.115–0.630; p = 0.002), and between raters of group 2 (with training; ICC = 0.433; 95% CI = 0.127–0.671; p = 0.002). The intra-rater analysis was excellent for raters 9 and 12, good to excellent for raters 8 and 10, moderate to excellent for rater 7, and poor to good for rater 11. Conclusion: The intra-rater reliability of the METRICS score was relatively good, while the inter-rater reliability was relatively low. This highlights the need for further efforts to achieve a common understanding of METRICS items, as well as resources consisting of explanations, elaborations, and examples to improve reproducibility and enhance their usability and robustness. Key Points: Questions Guidelines and scoring tools are necessary to improve the quality of radiomics research; however, the application of these tools is challenging for less experienced raters. Findings Intra-rater reliability was high across all raters regardless of experience level or previous training, and inter-rater reliability was generally poor to moderate across raters. Clinical relevance Guidelines and scoring tools are necessary for proper reporting in radiomics research and for closing the gap between research and clinical implementation. There is a need for further resources offering explanations, elaborations, and examples to enhance the usability and robustness of these guidelines.