Radiologia Medica最新文献

Objective: To perform a systematic review assessing the role of subspecialty radiologist in a multidisciplinary team (MDT) focused on pancreatic adenocarcinoma (PDAC) and the impact on patient outcomes.

Methods and materials: Ovid MEDLINE, Ovid Embase, Cochrane Database of Systematic Reviews (Ovid), Cochrane Central Register of Controlled Trials (Ovid), and Web of Science searched until August 31, 2023, for studies including radiology review in PDAC MDT. Data extraction, risk of bias, and applicability assessment were performed by two authors. Random-effects meta-analysis was performed.

Results: Ten studies (four retrospective, six prospective, and four multi-center vs. six single center), including 2,067 patients, met inclusion criteria in a pool of 13,043 studies. Three studies including 977 patients demonstrated succinct changes in patient management (average estimate 0.53 (95% CI 0.50-0.56)) after a radiology review at the MDT. A random-effects meta-analysis of these studies show that approximately 44% of patients had their treatment plans changed after a meeting at the MDT that included a subspecialty trained radiologist, but there is a large range of uncertainty (18%-74%) due to differences in the studies' groups, what they considered a "change" and local practices. Two studies highlighted an important nuance: while MDT input is valuable, the interpretations can vary between teams or even within the same team over time, reflecting the complex nature of PDAC. None of the studies showed a statistically significant survival advantage solely from MDT involvement, likely because overall survival depends on many downstream factors.

Conclusions: The presence of radiology review in the setting of multidisciplinary team meetings provides crucial information, leading to management changes. Future studies could explore advanced imaging modalities such as PET-MRI and establish the impact of intra/interobserver variability within the multidisciplinary team.

Clinical relevance/application: This work highlights the ongoing involvement of subspecialty trained radiologists in the multidisciplinary team for PDAC, and it significantly influences clinical decisions. Applicability concerns exist due to a potential lack of resources for robust MDT initiatives outside of tertiary centers.

Pancreatic ductal adenocarcinoma (PDAC) remains one of the most aggressive malignancies, with limited therapeutic options and poor prognosis. Dynamic contrast-enhanced imaging provides valuable non-invasive information on tumor biology, and rim enhancement (RE) on computed tomography (CT) or magnetic resonance imaging (MRI) has emerged as a potential biomarker of aggressive disease. To clarify its clinical significance, a systematic review and meta-analysis of studies published up to May 31st, 2025, in Medline, Scopus, Web of Science, and the Cochrane Library was conducted. Twelve studies (10 retrospective, 2 prospective) including 2207 patients were analyzed. The pooled prevalence of RE was 36.3%, with a good inter-observer agreement (κ = 0.808). RE was consistently associated with reduced resectability (odds ratio [OR] 3.35, 95% confidence interval [CI] 2.19-5.12, p < 0.001), poorer histological differentiation (OR 4.23, 95% CI 1.05-17.07, p = 0.043), and increased frequency of SMAD4 (OR 1.78, 95% CI 1.22-2.60, p = 0.003) and KRAS mutations (OR 2.55, 95% CI 1.37-4.75, p = 0.003). Patients with RE also experienced shorter progression-free, disease-free, and overall survival after both surgical and non-surgical treatments (all p < 0.001). These findings indicate that RE is a reproducible imaging marker of aggressive tumor biology in PDAC, reflecting unfavorable pathological and molecular features and serving as a predictor of resectability and survival.

Objective: To evaluate the performance of intravoxel incoherent motion (IVIM) and diffusion kurtosis imaging (DKI) in identifying the macrotrabecular-massive (MTM) subtype of hepatocellular carcinoma (HCC).

Materials and methods: This prospective study enrolled 76 consecutive patients who underwent surgical resection for HCC. Preoperative MRI examination, including multiple b value diffusion-weighted imaging (DWI), was performed. The apparent diffusion coefficient (ADC), true diffusion coefficient (D), pseudo-diffusion coefficient (D*), perfusion fraction (f), kurtosis-corrected ADC (AD), and apparent diffusion kurtosis (AK) were calculated for the whole tumor (including central and peripheral parts) and within two peritumor 3-mm-thick layers (i.e., inner and outer layers). The presence of the MTM subtype was histopathologically assessed. Based on the quantitative imaging parameters between the two groups, logistic regression analysis was used to determine predictors of the MTM subtype, and receiver operating characteristics (ROC) curves were used to evaluate the diagnostic performance.

Results: There were 58 MTM-negative tumors and 18 MTM-positive tumors. The ADC, D, and AD values for the whole tumor as well as the central and peripheral parts were significantly lower, while the D* value for the tumor peripheral part and D* value for the peritumor inner layer were higher in MTM-positive tumors compared to MTM-negative tumors (all P < 0.05). Multivariable analysis revealed that the AD value for the tumor peripheral part (odds ratio [OR], 0.498; P = 0.027) and the D* value for the peritumor inner layer (OR, 1.919; P = 0.038) were independent predictors of the MTM subtype. Incorporating these two parameters into a conventional model based on MRI features significantly improved diagnostic performance (area under the curve [AUC], 0.871 vs. 0.752; P = 0.020).

Conclusion: Our proof-of-concept study demonstrates that advanced DWI-derived parameters of the invasive tumor front, located between the tumor core and liver parenchyma, provide incremental value over conventional MRI features for evaluating the MTM subtype of HCC.

Background/objectives: This study aimed to develop an integrated model based on cone-beam breast computed tomography (CBBCT) and hematological indicators to predict the prognosis of preoperative stage I-III breast cancer.

Methods: A retrospective analysis was performed on 243 patients with pathologically confirmed stage I-III breast cancer. A novel machine learning framework for feature selection was employed, which integrates 14 distinct algorithms and explores 101 possible combinations, enhancing the ability to identify the most relevant features in high-dimensional medical imaging datasets. After feature selection, a patient risk score was calculated to construct a nomogram model for breast cancer prognosis. The nomogram model was evaluated using receiver operating characteristic (ROC) curve analysis and calibration curves. Univariate and multivariate regression analyses were conducted to validate the screened features and determine independent risk factors.

Results: A machine learning computational framework based on 101 combinations selected 12 prognostic indicators for overall survival (OS) and 18 for disease-free survival (DFS) from 37 CBBCT and hematological features. The model incorporating clinical and imaging indicators achieved an average area under the curve (AUC) value of 0.832 in both the training and validation datasets, demonstrating superior overall survival (OS) prediction performance compared to the clinical model without CBBCT indicators (AUC = 0.777). Similarly, the AUC values for DFS prediction in the training and validation sets were 0.996 and 0.732, respectively. Molecular typing, enhancement curve types, and morphology were independent risk factors for OS in the clinical prediction model. Calcification was an independent risk factor associated with DFS. A nomogram model was established combining the above features.

Conclusions: Our study successfully screened prognostic-related CBBCT and hematological features. The developed nomogram showed satisfactory preoperative predictive efficacy for stage I-III breast cancer.

Purpose: This study aimed to develop a deep learning model, capable of extracting both spatial and temporal features from contrast-enhanced ultrasound (CEUS) data and integrating with patient clinical parameters, for the differential diagnosis between hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC).

Materials and methods: We retrospectively analyzed the CEUS data (ultrasound contrast agent: SonoVue®-sulfur hexafluoride microbubbles) from 165 ICC patients and 140 date-matched HCC patients. A deep learning model, namely CEUS-CD-Net, was developed to extract spatial-temporal features from dynamic CEUS data and integrate them with patient clinical parameters for the differential diagnosis between HCC and ICC. The performance of CEUS-CD-Net was evaluated using the area under the receiver operating characteristic curve (AUC), with comparisons against other methods including the single-source data-based models (CEUS-Net and CD-Net, based merely on dynamic CEUS or patient clinical data), CEUS static image-based model (sCEUS-Net), time-intensity curve-based model (TIC-Model), and the assessment by radiologists.

Results: CEUS-CD-Net achieved an AUC of 0.884 (95% CI, 0.794-0.938) on the test cohort, significantly outperforming the single-source data-based models of CEUS-Net (0.827 [0.730-0.896]) and CD-Net (0.812 [0.718-0.887]), as well as sCEUS-Net (0.772 [0.669-0.851]) and TIC-Model (0.731 [0.633-0.823]). In the subset of determinate cases, CEUS-CD-Net achieved an AUC of 0.893 [0.806-0.950], which was better than the one obtained by radiologists' assessment (0.790 [0.683-0.868]). Model visualization results revealed that CEUS-CD-Net surpassed radiologists in discerning subtle patterns reflected by CEUS.

Conclusion: The integration of spatial and temporal features of dynamic CEUS data, coupled with clinical parameters of patients in CEUS-CD-Net, significantly improved the differential diagnosis between HCC and ICC.

Computed tomography (CT) is one of the most widely used diagnostic imaging modalities and currently represents the leading source of radiation exposure in the general population. Consequently, regulatory measures have been introduced to reduce unnecessary patient irradiation. To align with these requirements and to promote consistency across institutions in Italy-ensuring comparable radiation exposure for the same type of examination-the Italian Society of Medical and Interventional Radiology (SIRM), through its dedicated subcommittee, has developed this position paper.

Purpose: This multicentric retrospective study aimed to evaluate the diagnostic outcomes, adherence to guideline-based recommendations, and sustainability implications of 280 initial chest CT scans performed without contrast. The scans were conducted at the Radiology Unit of the Foundation IRCCS Ca' Granda Ospedale Maggiore Policlinico of Milan and the Radiology Department of Treviglio-Caravaggio Hospital-ASST Bergamo Ovest. The study focused on optimizing patient selection and radiological recommendations to align with evidence-based guidelines.

Materials and methods: This retrospective study included 280 patients (mean age 61.68 years; 51.07% women, 48.93% men) who underwent their first chest CT scan without contrast. Diagnostic outcomes were analyzed across different clinical queries. Incidental findings and radiologists' recommendations were assessed for alignment with Fleischner Society and ACR guidelines. Deviations were categorized as unnecessary imaging suggestions or missed indications. Radiation exposure was quantified using the dose-length product (DLP).

Results: Diagnostic outcomes were positive in 54.64% of cases. Incidental findings occurred in 34.64% of cases, with guideline adherence at 81.43%. Deviations included unnecessary imaging in 14.29% and missed follow-up indications in 4.28% of cases. The median DLP was 237.5 mGy·cm (IQR 171.8-320.1).

Conclusion: This study highlights significant opportunities to refine patient selection and radiological recommendations through adherence to Fleischner Society and ACR guidelines. By integrating evidence-based practices into routine workflows, the findings advocate for reduced unnecessary imaging, enhanced diagnostic pathways, and sustainable healthcare practices in chest CT imaging.

Purpose: To evaluate the prognostic value of the Neck Imaging Reporting and Data Systems (NI-RADS) in early post-operative imaging for predicting recurrence and survival outcomes in high-risk oral cavity squamous cell carcinoma (SCC) patients undergoing post-operative radiotherapy (PORT).

Methods: This retrospective study included 84 patients with high-risk oral cavity SCC who were scheduled for PORT after radical surgery between January 2013 and May 2024. Early imaging with contrast-enhanced CT or MRI was performed within 12 weeks post-surgery and scored using the NI-RADS system. Associations between NI-RADS scores, recurrence, and survival outcomes were analyzed using Kaplan-Meier and Cox proportional hazards models.

Results: Although NI-RADS was originally designed for post-treatment surveillance, we applied it to early post-operative imaging as an exploratory risk-stratification tool. NI-RADS scores significantly predicted regional disease-free survival (DFS) and overall survival (OS). Patients with higher NI-RADS T and N scores had poorer outcomes. Multivariable analysis confirmed early NI-RADS T as an independent predictor of OS (p = 0.01). Interobserver agreement for NI-RADS classifications was strong (Weighted Kappa: T = 0.837, N = 0.855). Although higher radiotherapy doses were administered to patients with NI-RADS 2-3 scores, these patients demonstrated worse outcomes, reflecting aggressive disease.

Conclusion: Early application of NI-RADS in post-operative imaging provides valuable prognostic insights, enabling risk stratification and tailored management in high-risk oral cavity SCC patients. Streamlining imaging workflows and exploring alternative therapeutic strategies for high-risk groups may further optimize outcomes.

Background: RAS mutational status is a critical prognostic biomarker in colorectal liver metastases (CRLM), traditionally assessed via tissue biopsy. This study evaluates the potential of radiomic features extracted from CT and MRI to non-invasively predict RAS mutations using machine learning algorithms.

Methods: In this study, 77 CRLM metastases (mean size 34.9; range 17-56 mm) with known RAS mutational status were analyzed. Radiomic features were extracted from hepatobiliary-phase MRI and portal venous-phase CT. After removing highly correlated features (Pearson |r| > 0.7) and applying z-score normalization, LASSO logistic regression with repeated tenfold cross-validation was used for feature selection. A total of 41 predictive features were identified. The dataset was split into training (70%) and testing (30%), ensuring that all lesions from a given patient were assigned exclusively to either the training or testing set. To address class imbalance in the training data, the Random Over Sampling Examples (ROSE) algorithm was applied exclusively to the training set. Six classification models (Stepwise Logistic Regression, LASSO, Random Forest, GBM, Neural Network, and CART) were trained and evaluated using ROC/AUC and other diagnostic metrics. DeLong's test was applied for pairwise AUC comparisons.

Results: MRI-derived features, particularly from wavelet-transformed gldm and first-order matrices, showed strong predictive power, with several achieving > 0.75 AUC individually. The gradient boosting machine (GBM) outperformed all other models with an AUC of 0.998 and an accuracy of 95.6%. Random forest and CART also demonstrated high discriminative performance (AUCs of 0.990 and 0.914, respectively). Nine features were consistently ranked among the top 20 predictors across all models, suggesting robust modality-independent imaging biomarkers. DeLong's test confirmed statistically significant AUC differences between GBM and logistic regression models (p < 0.05).

Conclusions: The results of this pilot study suggest that radiomic analysis combining CT and MRI modalities, particularly when processed through ensemble machine learning methods, holds the potential to accurately predict RAS mutational status in CRLM. While promising, these findings should be interpreted with caution, considering the study's limitations, including the small patient cohort and its design. These factors highlight the need for prospective validation in larger, multicenter cohorts to confirm the generalizability of the models. Nevertheless, these preliminary results support the use of multiparametric radiomics as a potential non-invasive tool for preoperative molecular stratification.

Forensic radiology training significantly enhances the diagnostic accuracy of postmortem computed tomography (PMCT) in firearm-related deaths, reducing inter-reader variability and improving injury detection. This study examines the impact of forensic expertise on PMCT interpretation, evaluating whether specialized training supersedes clinical radiology experience (non-forensic). A retrospective observational analysis was conducted at the University of Palermo, examining 10 firearm fatalities (homicides or suicides) between 2021 and 2024. The sample included individuals aged 25 to 66, with injuries from both short- and long-barrel firearms. Four radiologists with varying forensic experience analyzed the PMCT scans: an experienced forensic radiologist, an experienced clinical radiologist without forensic training, a radiology resident with forensic training, and a radiology resident without forensic expertise. Findings were compared against autopsy results as the gold standard. A lesion-based analysis was carried out in performance metrics considering the total number of findings (n = 960) and the number of findings in each subgroup (ranging from 40 up to 230 lesions). Inter-rater agreement was assessed using Fleiss' kappa and Cohen's kappa, while diagnostic performance was evaluated with ROC curve analysis. Results showed significantly higher diagnostic accuracy among radiologists with forensic training, particularly in detecting entrance and exit wounds, as well as organ injuries. These findings underscore the critical role of forensic radiology training in enhancing PMCT reliability, particularly for firearm-related injuries. Standardized reporting protocols and structured training programs are crucial for strengthening medicolegal investigations, thereby ensuring accurate and reproducible forensic imaging assessments. Future research should explore advanced imaging techniques, including radiomics and AI-driven analysis, to optimize forensic radiology practices.