Quantitative Imaging in Medicine and Surgery最新文献

Background: The effectiveness of Liver Imaging Reporting and Data System treatment response algorithm version 2024 (LR-TRA v2024) in hepatocellular carcinoma (HCC) patients undergoing locoregional plus systemic combination therapy remains uncertain. We aimed to investigate the performance of LR-TRA v2024 on magnetic resonance imaging (MRI) in detecting residual HCC following transarterial chemoembolization (TACE) plus systemic therapy.

Methods: This single-center retrospective study included consecutive adult patients who received TACE plus systemic therapy for HCC and subsequent surgical resection (July 2019 to November 2023). All contrast-enhanced preoperative MRIs were independently evaluated by three blinded radiologists for LR-TR, Liver Imaging Reporting and Data System treatment response (LR-TR) categories and two ancillary features. Postoperative pathology was used as the reference standard for residual tumors, which was further categorized as any (>0%) or major (>10%) residual tumors. When investigating the performances of LR-TR categories, the LR-TR Equivocal category was grouped into the LR-TR Viable category. The diagnostic performances were evaluated using positive predicting value (PPV) and negative predicting value (NPV).

Results: Fifty-one patients (median age, 56 years; 45 males) with 63 HCCs were included. For the detection of any residual tumor, the per-lesion PPV and NPV of the LR-TR Viable category were 100.0% and 46.9%, respectively; the per-patient PPV and NPV were 100.0% and 45.5%, respectively. For the detection of major residual tumor, the per-lesion PPV and NPV of the LR-TR Viable category were 80.6% and 84.4%, respectively; the per-patient PPV and NPV were 82.8% and 86.4%, respectively.

Conclusions: LR-TRA v2024 was effective in evaluating treatment response and detecting residuals of HCC to TACE plus systemic therapy.

Background: Endometrial endometrioid carcinoma (EEC) tumor grade is a critical prognostic factor, but its accurate preoperative non-invasive assessment remains challenging due to the limitations of conventional imaging and biopsy. Transvaginal ultrasound (TVUS) is the primary imaging modality but offers limited quantitative insights for grading. Deep learning radiomics (DLR), which combines the strengths of deep learning (DL) for automatic feature extraction and radiomics for quantifying tumor heterogeneity, holds promise for uncovering prognostic information from routine ultrasound images. This study aimed to develop and validate a DLR model based on preoperative TVUS images for the non-invasive differentiation of EEC tumor grades.

Methods: A total of 297 EEC cases with confirmed histological grades, including grade 1 (G1), grade 2 (G2), and grade 3 (G3), were selected from 1,258 endometrial cancer patients who underwent hysterectomy across eight centers. Radiomics features were extracted from TVUS images, and a radiomics model was constructed using the extreme gradient boosting (XGBoost) algorithm. Simultaneously, DL features were extracted using ResNet-50 to establish a DL model. A combined DLR model was then developed by integrating both feature sets, employing five-fold cross-validation for internal validation. An external testing cohort comprising 129 cases with corresponding grading data was collected from three independent centers. The performance of the three models in identifying EEC differentiation grade was compared using receiver operating characteristic (ROC) curve analysis to evaluate their diagnostic accuracy.

Results: In differentiating EEC grades, the DLR model outperformed both the single radiomics and DL models. In the identification of G3 and G1/G2, the AUC of the DLR model was 0.871 and 0.843 in the training cohort and the external testing cohort, respectively. The AUC of the identification of G2 and G1 was 0.856 and 0.816 in the training cohort and the external testing cohort, respectively. Decision curve analysis confirmed the clinical utility of the DLR model.

Conclusions: The DLR model based on TVUS images shows potential value for the non-invasive differentiation of EEC tumor grading and provides a useful supplement for non-invasive clinical staging of endometrial carcinoma prior to surgery.

Background: Although patients with stable angina pectoris (SAP) are generally considered to be at lower risk than those with acute coronary syndromes (ACS), their risk of major adverse cardiovascular events (MACEs) remains substantial. Lesion-specific pericoronary adipose tissue attenuation (PCATa-lesion) reflects local coronary inflammation, and the triglyceride-glucose body mass index (TyG-BMI) is a robust surrogate of insulin resistance (IR) and metabolic dysfunction; however, their combined prognostic value remains unclear. This study aimed to evaluate whether incorporating TyG-BMI and PCATa-lesion into conventional clinical and coronary computed tomography angiography (CCTA) models improves MACEs prediction and risk stratification in SAP patients.

Methods: In this retrospective study, patients with SAP who underwent CCTA from January 2017 to December 2020 were included. Clinical and imaging data were collected, including PCATa-lesion, TyG-BMI, plaque characteristics, and coronary artery calcium score (CACS). Statistical analyses included Cox proportional hazards regression to estimate hazard ratios (HRs) and 95% confidence intervals (CIs), time-dependent receiver operating characteristic curve, Kaplan-Meier analysis and decision curve analysis (DCA).

Results: A total of 212 patients were enrolled, with 43 MACEs occurring over a median follow-up of 36 months. Multivariable Cox regression analysis identified age (HR =1.052, 95% CI: 0.999-1.108; P=0.049), degree of stenosis (DS) (HR =1.079, 95% CI: 1.047-1.112; P=0.031), TyG-BMI (HR =2.198, 95% CI: 1.091-4.426; P=0.027) and PCATa-lesion (HR =1.117, 95% CI: 1.067-1.169, P<0.001) as independent predictors of MACEs. Kaplan-Meier curve demonstrated that patients in the highest tertile of PCATa-lesion and those with elevated TyG-BMI had a significantly increased risk of MACEs (P<0.001). Higher PCATa-lesion values were also significantly associated with increased incidence of high-risk plaques (HRP) (P=0.014). Subgroup analysis revealed a significant difference in PCATa-lesion between SAP patients with and without comorbid diabetes mellitus (DM) (P=0.016); importantly, elevated PCATa-lesion levels were associated with a substantially higher risk of adverse events in DM patients compared to non-DM individuals. Furthermore, the integrated model incorporating PCATa-lesion and TyG-BMI demonstrated superior goodness-of-fit, discriminatory ability, and net clinical benefit across a range of risk thresholds compared to the conventional model (age and DS only).

Conclusions: PCATa-lesion is an independent prognostic factor for MACEs in patients with SAP. The combination of PCATa-lesion and TyG-BMI provides incremental predictive value for assessing MACEs risk in SAP patients.

Background: Atherosclerotic internal carotid artery occlusion (AICAO) is associated with a high risk of stroke recurrence despite standard medical therapy. This study aimed to evaluate the predictive value of the neutrophil-lymphocyte ratio (NLR), an accessible marker of systemic inflammation, for identifying patients at higher risk of recurrent stroke.

Methods: This retrospective study enrolled 136 patients with AICAO, whose NLR data were collected. Recurrent stroke was evaluated via clinical and vascular imaging follow-up. Receiver operating characteristic (ROC) analysis was performed to determine the optimal NLR cutoff value. The value of NLR in predicting stroke recurrence was determined via a Cox regression model.

Results: Of the 281 initially screened patients, 136 met the study's inclusion criteria (age 62±10 years; 68% male). Among the patients, 17 (12.5%) experienced ipsilateral stroke (1-year rate, 8.8%; 2-year rate, 12.5%). The median baseline NLR was higher in patients with recurrence [3.38, interquartile range (IQR), 2.20-4.95] than in those without recurrence (2.39, IQR, 1.82-3.02) (Mann-Whitney P=0.007). The ROC analysis indicated an optimal NLR cutoff of 3.36 [area under the curve (AUC) =0.703; 95% confidence interval (CI): 0.559-0.847; P=0.007; sensitivity =0.59; specificity =0.82]. Patients were stratified into NLR-high (>3.36; n=31) and NLR-low (≤3.36; n=105) groups, with the 1-year stroke rates being 22.5% (7/31) and 5.7% (6/105), respectively, with an absolute risk difference of 16.8% (95% CI: 3.4-30.2%). In the Kaplan-Meier analysis, the log-rank P value was 0.039. In the univariate Cox analysis, an NLR >3.36 yielded a hazard ratio (HR) of 3.83 for stroke recurrence (95% CI: 1.48-9.94; P=0.006). In the multivariable Cox analysis, an NLR >3.36 remained independently associated with recurrence (HR 4.17, 95% CI: 1.59-10.91; P=0.004). When NLR was modelled as a continuous log-transformed variable, each 1-unit increase yielded an HR of 1.99 (95% CI: 1.49-2.66; P<0.001).

Conclusions: In symptomatic nonacute patients with AICAO, NLR is a predictor for recurrent stroke under standard medical treatment. Moreover, an NLR >3.36 is associated with a higher risk of stroke recurrence, and intensive surveillance in high-risk patients with this marker may be necessary.

Background: Contrast-enhanced T2 fluid-attenuated inversion recovery (CE-T2 FLAIR) and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) provide complementary information on lesion enhancement and vascular permeability. This study aimed to assess the correlation between CE-T2 FLAIR enhancement and DCE-MRI-derived permeability parameters in brain metastases.

Methods: This single-center retrospective study included 43 patients with 80 brain metastases confirmed by pathology or follow-up between January 2018 and July 2024. All patients underwent T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), T2 FLAIR, DCE-MRI, CE-T2 FLAIR, and contrast-enhanced T1WI (CE-T1WI) examinations. Quantitative DCE-MRI parameters were evaluated for all lesions, including the volume transfer constant (K^trans) and the reverse volume transfer constant (K_ep). CE-T2 FLAIR enhancement was assessed using contrast ratio (CR) and percentage increase (PI). Lesions were grouped by enhancement level on CE-T2 FLAIR relative to CE-T1WI: hyperenhancement (Group A), similar enhancement (Group B), and hypoenhancement (Group C). Group differences were assessed using the Kruskal-Wallis test, followed by Bonferroni-adjusted Mann-Whitney U tests for pairwise comparisons; associations between CR/PI and K^trans/K_ep were examined using Spearman's rank correlation.

Results: Groups A, B, and C included 17, 45, and 18 lesions, respectively. Group A showed significantly higher CR and PI and lower K^trans and K_ep compared with Groups B and C (all P<0.05). Group B also demonstrated significantly higher CR and PI and lower permeability values than Group C (P<0.05). Lesions in Groups A and B were significantly smaller than those in Group C (P<0.05). CR was negatively correlated with K^trans (r=-0.467, P<0.001) and K_ep (r=-0.526, P<0.001). PI was negatively correlated with K^trans (r=-0.658, P<0.001) and K_ep (r=-0.716, P<0.001).

Conclusions: Vascular permeability of brain metastases is a key factor contributing to the differential enhancement observed between CE-T2 FLAIR and CE-T1WI, with CE-T2 FLAIR demonstrating superior sensitivity in detecting metastases with low vascular permeability.

Background: Total kidney volume (TKV)-based indices are central to imaging classification in autosomal dominant polycystic kidney disease (ADPKD) but primarily reflect cumulative changes, limiting their ability to detect current disease activity. We evaluated quantitative dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) as a complementary tool for assessing renal perfusion in mild and rapidly progressive ADPKD.

Methods: Five healthy subjects and twenty patients were enrolled, 10 with mild ADPKD [estimated glomerular filtration rate (eGFR) ≥60 mL/min/1.73 m² and htTKV ≤750 mL/m] and 10 with severe ADPKD (eGFR <60 mL/min/1.73 m² or htTKV >750 mL/m). Healthy subjects underwent three DCE-MRI scans over three different scanners within a week, while patients underwent two DCE-MRI scans in a single scanner, 4.2±2.5 days apart, with P4 phantoms for scanner-specific error correction. Eleven pharmacokinetic (PK) parameters from the extended Tofts model (ETM), Tofts model (TM), and Shutter Speed Model (SSM) were measured before and after correction. Reproducibility or repeatability was evaluated via within-subject coefficient of variation (wCV), group differences via analysis of variance (ANOVA), and correlations via Pearson correlation coefficient (r).

Results: The P4-based error correction strategy improved the reproducibility of ETM-derived volume transfer constant (K^trans ) in healthy subjects by nearly 5-fold and in ADPKD patients by almost 3-fold. The ETM-derived K^trans was significantly higher in mild vs. severe ADPKD (0.17±0.04 vs. 0.09±0.02 min^-1; P<0.001) and significantly correlated with htTKV (r=-0.79, P<0.001), TCV (r=-0.77, P<0.001), and eGFR (r=0.68, P<0.001). The ETM-derived K^trans achieved 95% accuracy in distinguishing mild from severe ADPKD, the highest among all PK parameters after correction, whereas it was only 50% before correction.

Conclusions: Intrarenal K^trans demonstrated strong correlations with both functional and anatomical indicators of ADPKD severity, highlighting its potential as an imaging biomarker and a viable complement to htTKV.

Background: Papillary thyroid carcinoma (PTC) is the most common malignant tumor of the thyroid gland, and its aggressiveness determines distinct therapeutic and management strategies. This study aimed to evaluate the diagnostic performance of quantitative speed-of-sound (QSOS), elastography, and B-mode imaging in assessing the aggressiveness of PTC.

Methods: A total of 157 patients with solitary PTC were enrolled and classified into two groups based on pathological findings: an invasive group (n=118) and a non-invasive group (n=39). All patients underwent B-mode imaging, QSOS imaging, and elastography. Speed-of-sound (SOS)1, SOS2, and SOS3 represent the SOS measurements at different regions of the nodule, and the difference refers to the value of SOS1 minus SOS3.

Results: Our results demonstrated that the maximum diameter of PTC in the invasive group was significantly larger than that in the non-invasive group {9.5 [interquartile range (IQR), 7.1-12.3] vs. 6.2 (IQR, 5.2-8.5) mm}. Additionally, the invasive group exhibited a lower SOS2 compared to the non-invasive group [1,570.1 (IQR, 1,565.3-1,576.5) vs. 1,591.1 (IQR, 1,582.8-1,594.4) m/s]. Furthermore, maximum elasticity (Emax) was higher in the invasive group [52.7 (IQR, 41.1-66.0) vs. 45.9 (IQR, 35.7-58.1) kPa], whereas the difference was smaller in the invasive group (6.5±13.2 vs. 11.6±11.9 m/s). Binary logistic regression analysis identified SOS2, maximum diameter, and the difference as independent predictors of PTC invasiveness. Their odds ratios (ORs) were 0.857, 1.274, and 0.945, respectively (all P<0.05). A multiple logistic regression model was developed based on these three variables, which as a combined predictor demonstrated high accuracy in diagnosing the aggressiveness of PTC, with an area under the curve (AUC) of 0.91, a sensitivity of 0.924, and a specificity of 0.816.

Conclusions: Our study is the first to examine the invasiveness of PTC using QSOS technology. To predict PTC invasiveness, a combined predictor integrating SOS, elastography, and B-mode imaging parameters of thyroid nodules was developed. This indicator provides physicians with valuable guidance for the early diagnosis and treatment of PTC due to its excellent accuracy, sensitivity, and specificity.

Background: Early differentiation between invasive adenocarcinoma (IAC) and non-IAC pulmonary nodules is crucial for guiding clinical decision-making. Therefore, this study aimed to distinguish IAC from non-IAC pulmonary nodules using intra-tumor radiomics signatures, habitat radiomics analysis, and a combined nomogram by integrating generative adversarial network (GAN) based super-resolution reconstruction.

Methods: In this multi-center retrospective study, 858 patients [mean ± standard deviation (SD): 57.635±12.978 years] were enrolled as the training set (Center 1, 501 non-IAC cases vs. 357 IAC cases) and 272 external testing patients (Centers 2 and 3, 183 non-IAC cases vs. 89 IAC cases; mean ± SD: 57.037±11.683 years) were included. Univariate and multivariate analyses were conducted to explore clinical characteristics. Radiomics features were extracted from intra-tumor regions and sub-regions. After feature selection, machine learning models, namely the Intra-Model and Habitat-Model, were developed. A combined nomogram integrating significant clinical factors, intra-tumor radiomics and habitat radiomics was constructed and evaluated using area under receiver operator characteristics curve (AUC), decision curve analysis (DCA), and other quantified metrics.

Results: The Habitat-Model outperformed Intra-Model (training AUC: 0.893 vs. 0.853; testing AUC: 0.882 vs. 0.875) in predicting IAC invasiveness. The combined nomogram demonstrated an incremental advancement in IAC stratification [training AUC: 0.907 (95% CI: 0.887-0.927); testing AUC: 0.895 (95% CI: 0.849-0.941)], with DCA confirming 28-34% net benefit improvement over single-modality approaches at critical thresholds (10-25% risk). Age (P<0.001) and nodule diameter (P<0.001), along with intra-tumor and habitat radiomics, were identified as key contributing factors.

Conclusions: The spatially resolved habitat radiomics model exhibited higher discriminative accuracy than the classical intra-tumor radiomics model. The combined nomogram framework, which integrated intra-tumor radiomics, habitat radiomics, and significant clinical biomarkers, achieved state-of-the-art performance in IAC stratification. This framework provides a robust tool for precision therapeutic decision-making in pulmonary nodule management.