Academic Radiology最新文献

Rationale and objectives: With the emergence of disease-modifying therapies, precise staging of dementia is urgent. This study aimed to develop a machine learning model integrating multimodal data to achieve objective staging of dementia severity in patients with cognitive impairment.

Materials and methods: A total of 149 patients (100 with Alzheimer's disease) were recruited. Demographic data, neuropsychological scores, and multimodal PET features were collected. Subjects were randomly split (7:3) into training and validation cohorts. PET features were screened using Boruta and LASSO to generate composite SUVR scores, while key demographic and neuropsychological predictors were identified through univariate and multivariate logistic regression analyses. Seven machine learning algorithms (logistic regression, decision tree, random forest, XGBoost, LightGBM, support vector machine, and artificial neural network) were trained using grid search and fivefold cross-validation. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration plots, and decision curve analysis (DCA), with SHAP analysis employed for interpretability.

Results: The cohort comprised 80 very mild-to-mild (CDR 0.5-1) and 69 moderate-to-severe (CDR 2-3) dementia cases. Key predictors included years of education, MMSE, and composite amyloid and FDG SUVR scores. The XGBoost model demonstrated robust performance, achieving an AUC of 0.888 (95% CI: 0.777-0.967) in the independent validation cohort. SHAP analysis identified MMSE and composite FDG SUVR scores as the most significant contributors to disease staging.

Conclusion: This study constructed and internally validated an interpretable multimodal model for dementia severity staging. While the results are promising, the developed web-based tool currently serves as a proof-of-concept to demonstrate how such models could potentially assist in optimizing patient management and screening candidates for novel therapies, pending further external validation.

Rationale and objectives: Repeated mild traumatic brain injury (mTBI), particularly from concussions, impairs cerebral perfusion and brain waste-clearance pathways, leading to lasting neurological deficits and elevated risk of neurodegeneration. Conventional pharmacological treatments targeting single pathways have shown limited efficacy in clinical trials. Photobiomodulation (PBM) has emerged as a promising noninvasive approach with the potential to improve both vascular function and clearance.

Aim: To determine whether transcranial PBM at 1267 nm, administered during the acute phase after repetitive concussion, improves cortical perfusion, oxygenation, intracranial compliance, meningeal lymphatic drainage, and neurological function in mice.

Materials and methods: Male C57BL/6 mice (n=20) were randomized to PBM and sham groups and subjected to three repeated consecutive closed-head concussive impacts at 1.5-hour intervals to model repetitive mTBI. Transcranial PBM (1267 nm, 10 mW/cm², 5 mm diameter spot) was applied 4 h after the last impact for 45 min (three 10-minute sessions separated by 5-minute intervals). The 1267 nm wavelength lies within a biological transparency window that supports deeper transcranial penetration than shorter near-infrared wavelengths. Outcomes included cortical microcirculation, tissue oxygenation, intracranial compliance, meningeal lymphatic drainage, and neurological severity score. Statistical analyses were performed using two-way analysis of variance for multiple comparisons, with p < 0.05 considered significant.

Results: Repetitive concussion produced stepwise declines in cortical perfusion and oxygenation, reduced cerebral compliance, impaired lymphatic clearance, and worse neurological scores. PBM partially reversed these deficits compared with sham, improving microcirculation and oxygenation toward baseline levels, increasing cerebral compliance, restoring meningeal lymphatic drainage, and lowering neurological severity scores.

Conclusion: Acute transcranial PBM at 1267 nm mitigates cerebrovascular, biomechanical, and meningeal lymphatic dysfunction after repetitive concussion, with associated functional benefit. By concurrently improving perfusion, oxygen delivery, intracranial compliance, and lymphatic drainage, PBM represents a mechanistically grounded, noninvasive candidate therapy for early adjunct intervention after mTBI.

Rationale and objectives: Intra-procedural patient motion is common in clinical MRI and degrades image quality. We evaluated whether certain clinical characteristics and patient sociodemographic factors are associated with patient motion during MRI.

Materials and methods: We retrospectively reviewed consecutive MRI reports from 2022 at two U.S. urban health systems (university and safety net). Exams were acquired using standardized protocols and reported by a single academic radiology group. Motion was defined by any mention of "motion" in the report body or impression. Clinical and sociodemographic characteristics were extracted from the electronic health record. Multivariable logistic regression assessed associations between patient characteristics and motion, adjusting for care setting.

Results: Among 68,517 MRIs, emergency and inpatient studies had higher odds (OR) [95%CI] of motion compared with outpatient exams (emergency OR 2.55; [2.38,2.74]; inpatient OR 3.26; [3.06,3.48]). After adjustment, older age remained associated with motion, (≥85 years OR 1.91; [1.64,2.23], vs. 19-34 years). Male gender had higher odds than female (OR 1.14; [1.08,1.20]). Black race had higher odds than White (OR 1.18; [1.10,1.27]). Increasing obesity was associated with greater odds (class II OR 1.11; [1.01,1.22]; class III OR 1.48; [1.34,1.64]). Ethnicity, preferred language, and health system were not significantly associated with motion.

Conclusion: Motion on MRI disproportionately affects elderly patients, men, individuals with obesity, and Black patients, independent of care setting. Accounting for motion-risk factors into scheduling, pre-scan counseling, and positioning protocols may reduce motion-limited studies. Addressing these patterns through workflow design is important for promoting equitable, high-quality MRI across diverse patient populations.

Rationale and objectives: To compare the efficacy of automatic breast ultrasound (ABUS) with handheld ultrasound (HHUS) as second-look US techniques in the identification of lesions that are detected on breast MRI, and to determine factors that affect lesion detection on second-look US examinations.

Material and methods: This single-center prospective study included 54 patients with 66 MRI-detected breast lesions referred for MRI-guided biopsy. All patients underwent second-look evaluation with both HHUS and ABUS. Histopathology or imaging follow-up served as the reference standard.US-guided biopsy or preoperative surgical localization was performed for lesions detected by US; while the rest of the patients underwent MRI-guided biopsy or localization procedures. Patients who refused to undergo MRI-guided procedures were followed-up with MRI for at least 2 year. ABUS and HHUS examinations were performed by different radiologists and evaluated based on BI-RADS lexicon.

Results: Out of the 66 lesions detected on MRI examinations, 30 (45.5%) were benign, 16 (24.2%) were high-risk lesions, and 20 (30.3%) were malignant. HHUS demonstrated 56/66 (84.9%), while ABUS demonstrated 46/66 (69.7%) of them; and the difference was statistically significant (p=0.010). Two out of 13 (15.4%) lesions detected only by HHUS, and 2/3 (66.7%) lesions detected only by ABUS were malignant. None of the 7 lesions (10.6%), that could not be detected by either method, were malignant. There was no statistically significant difference between the number of lesions detected on HHUS and ABUS in terms of lesion size, depth, lesion type (mass/non-mass), lesion localization, parenchymal density, kinetic features or morphological findings (p>0.05). Sensitivity was similar for HHUS and ABUS (90% for both). However, the positive predictive value (PPV) was higher for ABUS (39.1%) than for HHUS (32.1%) (notably due to a higher proportion of malignant lesions among ABUS-only detected findings).

Conclusion: HHUS was superior to ABUS in detecting lesions in second-look US evaluation. However, cancer detection rates were similar, resulting in a higher positive predictive value for ABUS. Our results show that the two methods are complementary to each other and have the potential to increase lesion detection rate when used together in clinics where both methods are available.

Rationale and objectives: To evaluate the diagnostic performance of preoperative computed tomography (CT) and magnetic resonance imaging (MRI)-based radiomics models in detecting liver metastases in patients with colorectal cancer (CRC).

Materials and methods: Following PRISMA 2020 guidelines, we systematically searched major databases up to July 2025. Study selection, data extraction, and quality assessment (Radiomics Quality Score and QUADAS-2) were performed independently. Separate bivariate random-effects meta-analyses were conducted for prognostic (metachronous) and diagnostic (synchronous) predictions.

Results: Twenty studies (3765 patients) were included in the systematic review. Twenty studies were included in the systematic review. Of these, 18 studies were included in the quantitative meta-analysis. For predicting metachronous metastases (13 studies), the pooled AUC was 0.83 (95% CI: 0.73-0.90), although significant publication bias suggested that this estimate may be optimistically inflated. For the detection of synchronous metastases (five studies), the pooled AUC was 0.85 (95% CI: 0.76-0.91). Heterogeneity was moderate to substantial. However, significant publication bias was detected for prognostic models (Deeks' test, P < 0.001), suggesting that these pooled estimates may be optimistically inflated.

Conclusion: Radiomics has the potential to predict metachronous and detect synchronous liver metastases in CRC. However, methodological weaknesses (mean Radiomics Quality Score ∼48%), geographic bias, and publication bias limit this evidence. Multinational validation is required before clinical application of the findings.

Rationale and objectives: Inflammation plays a crucial role in the pathophysiology of intracranial aneurysms (IAs). Our previous research demonstrated that blood inflammatory indices can serve as predictors of aneurysmal wall enhancement (AWE), which signifies the presence of inflammation within the aneurysmal wall and functions as an imaging biomarker for IA instability. Here, we aimed to further examine the relationship between blood inflammatory indices, AWE, and long-term clinical outcomes in patients with intracranial fusiform aneurysms (IFAs).

Materials and methods: We reviewed patients with IFAs who underwent high-resolution magnetic resonance imaging and blood laboratory tests, as recorded in our maintained database. Initially, a cross-sectional study was conducted to identify potential blood inflammatory indices that could predict the average value of aneurysmal wall enhancement in three dimensions (3D-AWE_avg). Subsequently, a follow-up study was performed to further elucidate the potential predictors associated with overall poor outcomes (CAO) in the same cohort.

Results: A total of 92 patients were included in the cross-sectional study. Both the systemic inflammation response index (SIRI), the systemic immune-inflammation index (SII), and the neutrophil-to-lymphocyte ratio (NLR) were associated with 3D-AWE_avg in univariate analysis; however, only SIRI was found to independently predict 3D-AWE_avg (P = 1.1 × 10^-5). In the follow-up study, 64 patients were included, with a mean follow-up period of 29.27 months. SIRI (13.725 [2.467-76.349], P = .003) and 3D-AWE_avg (5.387 [1.320-21.988], P = .019) were identified as the predictors of CAO in patients with IFAs. Furthermore, patients with a high SIRI value (≥0.725 × 10⁹/L, log-rank = 0.002) or a 3D-AWE_avg ≥ 0.604 (log-rank = 3 × 10^-5) had significantly higher risk of CAO.

Conclusion: SIRI predicts both aneurysmal wall enhancement and long-term adverse outcomes in patients with IFAs, supporting its potential role as a novel biomarker for risk stratification and clinical decision-making in this population.

Rationale and objectives: This study aims to compare the breast magnetic resonance imaging (MRI) features and clinicopathological characteristics of invasive breast cancer patients with different enhancement patterns, and to investigate the relationship between enhancement patterns and lymphovascular invasion (LVI).

Materials and methods: This retrospective study consecutively enrolled 1185 female patients with invasive breast cancer who underwent dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) at our institution. Propensity score matching (PSM) was employed to match patients between mass enhancement (ME) group and non-mass enhancement (NME) group. With the occurrence of LVI as the clinical endpoint, covariates with a standardized mean difference (SMD) greater than 0.1 and the enhancement patterns were incorporated into a Firth's bias-reduced logistic regression analysis to further evaluate the relationship between enhancement patterns and LVI.

Results: Compared to the ME group, lesions in the NME group demonstrated significantly higher rates of axillary lymph node positivity and LVI (both P<0.001). After PSM, differences in ADC values and the distribution of the triple-negative subtype persisted between the two groups. Firth regression analysis identified NME as a risk factor for LVI (P<0.001). Compared to the Luminal A subtype, the Luminal B (P<0.001), HER2-positive (P<0.001), and triple-negative (P=0.001) subtypes were all associated with a significantly increased risk of LVI. ADC value did not demonstrate a significant association with LVI (P=0.537).

Conclusion: NME was identified as a significant independent risk factor for LVI. Compared to the Luminal A subtype, the Luminal B, HER2-positive, and triple-negative subtypes were all associated with a significantly increased risk of LVI. The ADC value did not demonstrate a significant association with LVI.