Differential diagnosis of thyroid nodules by DCE-MRI based on compressed sensing volumetric interpolated breath-hold examination: A feasibility study.

OBJECTIVES: To explore the potential and performance of quantitative and semi-quantitative parameters derived from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) based on compressed sensing volumetric interpolated breath-hold (CS-VIBE) examination in the differential diagnosis of thyroid nodules. MATERIALS AND METHODS: A total of 208 patients with 259 thyroid nodules scheduled for surgery operation were prospectively recruited. All participants underwent routine and DCE-MRI. DCE-MRI quantitative parameters [Ktrans, Kep, Ve], semi-quantitative parameters [wash-in, wash-out, time to peak (TTP), arrival time (AT), peak enhancement intensity (PEI), and initial area under curve in 60 s (iAUC)] and time-intensity curve (TIC) types were analyzed. Differential diagnostic performances were assessed using area under the receiver operating characteristic curve (AUC) and compared with the Delong test. RESULTS: Ktrans, Kep, Ve, wash-in, wash-out, PEI and iAUC were statistically significantly different between malignant and benign nodules (P < 0.001). Among these parameters, ROC analysis revealed that Ktrans showed the highest diagnostic performance in the differentiation of benign and malignant nodules, followed by wash-in. ROC analysis also revealed that Ktrans achieved the best diagnostic performance for distinguishing papillary thyroid carcinoma (PTC) from non-PTC, follicular adenoma (FA) from non-FA, nodular goiter (NG) from non-NG, with AUC values of 0.854, 0.895 and 0.609, respectively. Type III curve is frequently observed in benign thyroid nodules, accounting for 77.4% (82/106). While malignant nodules are more common in type II, accounting for 57.5% (88/153). CONCLUSION: Thyroid examination using CS-VIBE based DCE-MRI is a feasible, non-invasive method to identify benign and malignant thyroid nodules and pathological types.

Uncovering brain functional connectivity disruption patterns of lung cancer-related pain.

Pain is a pervasive symptom in lung cancer patients during the onset of the disease. This study aims to investigate the connectivity disruption patterns of the whole-brain functional network in lung cancer patients with cancer pain (CP+). We constructed individual whole-brain, region of interest (ROI)-level functional connectivity (FC) networks for 50 CP+ patients, 34 lung cancer patients without pain-related complaints (CP-), and 31 matched healthy controls (HC). Then, a ROI-based FC analysis was used to determine the disruptions of FC among the three groups. The relationships between aberrant FCs and clinical parameters were also characterized. The ROI-based FC analysis demonstrated that hypo-connectivity was present both in CP+ and CP- patients compared to HC, which were particularly clustered in the somatomotor and ventral attention, frontoparietal control, and default mode modules. Notably, compared to CP- patients, CP+ patients had hyper-connectivity in several brain regions mainly distributed in the somatomotor and visual modules, suggesting these abnormal FC patterns may be significant for cancer pain. Moreover, CP+ patients also showed increased intramodular and intermodular connectivity strength of the functional network, which could be replicated in cancer stage IV and lung adenocarcinoma. Finally, abnormal FCs within the prefrontal cortex and somatomotor cortex were positively correlated with pain intensity and pain duration, respectively. These findings suggested that lung cancer patients with cancer pain had disrupted connectivity in the intrinsic brain functional network, which may be the underlying neuroimaging mechanisms.

Multiparametric MRI model to predict molecular subtypes of breast cancer using Shapley additive explanations interpretability analysis.

PURPOSE: The purpose of this study was to assess the predictive performance of multiparametric magnetic resonance imaging (MRI) for molecular subtypes and interpret features using SHapley Additive exPlanations (SHAP) analysis. MATERIAL AND METHODS: Patients with breast cancer who underwent pre-treatment MRI (including ultrafast dynamic contrast-enhanced MRI, magnetic resonance spectroscopy, diffusion kurtosis imaging and intravoxel incoherent motion) were recruited between February 2019 and January 2022. Thirteen semantic and thirteen multiparametric features were collected and the key features were selected to develop machine-learning models for predicting molecular subtypes of breast cancers (luminal A, luminal B, triple-negative and HER2-enriched) by using stepwise logistic regression. Semantic model and multiparametric model were built and compared based on five machine-learning classifiers. Model decision-making was interpreted using SHAP analysis. RESULTS: A total of 188 women (mean age, 53 ± 11 [standard deviation] years; age range: 25-75 years) were enrolled and further divided into training cohort (131 women) and validation cohort (57 women). XGBoost demonstrated good predictive performance among five machine-learning classifiers. Within the validation cohort, the areas under the receiver operating characteristic curves (AUCs) for the semantic models ranged from 0.693 (95% confidence interval [CI]: 0.478-0.839) for HER2-enriched subtype to 0.764 (95% CI: 0.681-0.908) for luminal A subtype, inferior to multiparametric models that yielded AUCs ranging from 0.771 (95% CI: 0.630-0.888) for HER2-enriched subtype to 0.857 (95% CI: 0.717-0.957) for triple-negative subtype. The AUCs between the semantic and the multiparametric models did not show significant differences (P range: 0.217-0.640). SHAP analysis revealed that lower iAUC, higher kurtosis, lower D*, and lower kurtosis were distinctive features for luminal A, luminal B, triple-negative breast cancer, and HER2-enriched subtypes, respectively. CONCLUSION: Multiparametric MRI is superior to semantic models to effectively predict the molecular subtypes of breast cancer.

Altered gray matter volume and functional connectivity in lung cancer patients with bone metastasis pain.

Bone metastasis pain (BMP) is a severe chronic pain condition. Our previous studies on BMP revealed functional brain abnormalities. However, the potential effect of BMP on brain structure and function, especially gray matter volume (GMV) and related functional networks, have not yet been clearly illustrated. Voxel-based morphometry and functional connectivity (FC) analysis methods were used to investigate GMV and intrinsic FC differences in 45 right-handed lung cancer patients with BMP(+), 37 lung cancer patients without BMP(-), and 45 healthy controls (HCs). Correlation analysis was performed thereafter with all clinical variables by Pearson correlation. Compared to HCs, BMP(+) group exhibited decreased GMV in medial frontal gyrus (MFG) and right middle temporal gyrus (MTG). Compared with BMP(-) group, BMP(+) group exhibited reduced GMV in cerebelum_6_L and left lingual gyrus. However, no regions with significant GMV differences were found between BMP(-) and HCs groups. Receiver operating characteristic analysis indicated the potential classification power of these aberrant regions. Correlation analysis revealed that GMV in the right MTG was positively associated with anxiety in BMP(+) group. Further FC analysis demonstrated enhanced interactions between MFG/right MTG and cerebellum in BMP(+) patients compared with HCs. These results showed that BMP was closely associated with cerebral alterations, which may induce the impairment of pain moderation circuit, deficits in cognitive function, dysfunction of emotional control, and sensorimotor processing. These findings may provide a fresh perspective and further neuroimaging evidence for the possible mechanisms of BMP. Furthermore, the role of the cerebellum in pain processing needs to be further investigated.

Comparing extracellular volume fraction with apparent diffusion coefficient for the characterization of breast tumors.

PURPOSE: To investigate the feasibility of dual-energy CT (DECT)-derived extracellular volume (ECV) fraction for characterization of breast tumors, compared to apparent diffusion coefficient (ADC) and validated against histopathological findings. MATERIAL AND METHODS: The ECV fraction and ADC were prospectively assessed in patients with breast tumors using chest DECT and breast MRI. The diagnostic performance of ECV fraction and ADC was accessed in predicting breast histopathological subtypes and pathological complete response (pCR) status. Histopathological sections were analyzed by digital image analysis. Pearson's correlation analysis was used to correlate between DECT and histopathological ECV fractions. RESULTS: This study included 271 patients, with 314 breast lesions (61 benign and 253 malignant). The ECV fraction and ADC showed comparable area under the curve (AUC) for distinguishing benign from malignant lesions (p = 0.123) and invasive carcinoma from ductal carcinoma in situ (p = 0.115). There were significant differences in ECV fraction between different hormone receptors and Ki67 states (p = 0.001 âˆ¼ 0.014), while ADC values only differed among various Ki67 states (p < 0.001). The ECV fraction was lower (p = 0.007), ADC was higher (p = 0.013) in pCR than in non-pCR group, with an AUC of 0.748 and 0.730 (p = 0.887), respectively. There was a positive correlation between DECT and histopathological ECV fractions (r = 0.615, p < 0.01). CONCLUSIONS: Routine chest DECT-derived ECV fraction is a viable quantitative imaging biomarker for predicting histopathological subtypes and pCR in patient with breast tumors, and correlated well with histopathology finding.

Early Identification of Pathologic Complete Response to Neoadjuvant Chemotherapy Using Multiphase DCE-MRI by Siamese Network in Breast Cancer: A Longitudinal Multicenter Study.

BACKGROUND: Siamese network (SN) using longitudinal DCE-MRI for pathologic complete response (pCR) identification lack a unified approach to phases selection. PURPOSE: To identify pCR in early-stage NAC, using SN with longitudinal DCE-MRI and introducing IPS for phases selection. STUDY TYPE: Multicenter, longitudinal. POPULATION: Center A: 162 female patients (50.63 ± 8.41 years) divided 7:3 into training and internal validation cohorts. Center B: 61 female patients (50.08 ± 7.82 years) were used as an external validation cohort. FIELD STRENGTH/SEQUENCE: Center A: single vendor 3.0 T with a compressed-sensing volume interpolated breath-hold examination sequence. Center B: single vendor 1.5 T with volume interpolated breath-hold examination sequence. ASSESSMENT: Patients underwent DCE-MRI before and after two NAC cycles, with tumor regions of interest (ROI) manually delineated. Histopathology was the reference for pCR identification. Models developed included a clinical one, four SN models based on IPS-selected phases, and integrated models combining clinical and SN features. STATISTICAL TESTS: Model performance was evaluated using the area under the receiver operating characteristic curve (AUC). The DeLong test was used to compare AUCs. Net reclassification improvement and integrated discrimination improvement (IDI) tests were employed for performance comparison. P < 0.05 was considered significant. RESULTS: In internal and external validation cohorts, the clinical model showed AUCs of 0.760 and 0.718. SN and integrated models, with increasing phases via IPS, achieved AUCs ranging from 0.813 to 0.951 and 0.818 to 0.922. Notably, SN-3 and integrated-3 and integrated-4 outperformed the clinical model. However, input phases beyond 20% did not significantly enhance performance (IDI test: SN-4 vs. SN-3, P = 0.314 and 0.630; integrated-4 vs. integrated-3, P = 0.785 and 0.709). DATA CONCLUSION: The longitudinal multiphase DCE-MRI based on the SN demonstrates promise for identifying pCR in breast cancer. EVIDENCE LEVEL: 1 TECHNICAL EFFICACY: Stage 4.

CT-derived extracellular volume and liver volumetry can predict posthepatectomy liver failure in hepatocellular carcinoma.

OBJECTIVES: Posthepatectomy liver failure (PHLF) is a severe complication of liver resection. We aimed to develop and validate a model based on extracellular volume (ECV) and liver volumetry derived from computed tomography (CT) for preoperative predicting PHLF in resectable hepatocellular carcinoma (HCC) patients. METHODS: A total of 393 resectable HCC patients from two hospitals were enrolled and underwent multiphasic contrast-enhanced CT before surgery. A total of 281 patients from our hospital were randomly divided into a training cohort (n = 181) and an internal validation cohort (n = 100), and 112 patients from another hospital formed the external validation cohort. CT-derived ECV was measured on nonenhanced and equilibrium phase images, and liver volumetry was measured on portal phase images. The model is composed of independent predictors of PHLF. The under the receiver operator characteristic curve (AUC) and calibration curve were used to reflect the predictive performance and calibration of the model. Comparison of AUCs used the DeLong test. RESULTS: CT-derived ECV, measured future liver remnant (mFLR) ratio, and serum albumin were independent predictors for PHLF in resectable HCC patients. The AUC of the model was significantly higher than that of the ALBI score in the training cohort, internal validation cohort, and external validation cohort (all p < 0.001). The calibration curve of the model showed good consistency in the training cohort and the internal and external validation cohorts. CONCLUSIONS: The novel model contributes to the preoperative prediction of PHLF in resectable HCC patients. CRITICAL RELEVANCE STATEMENT: The novel model combined CT-derived extracellular volume, measured future liver remnant ratio, and serum albumin outperforms the albumin-bilirubin score for predicting posthepatectomy liver failure in patients with resectable hepatocellular carcinoma. KEY POINTS: • CT-derived ECV correlated well with the fibrosis stage of the background liver. • CT-derived ECV and mFLR ratio were independent predictors for PHLF in HCC. • The AUC of the model was higher than the CT-derived ECV and mFLR ratio. • The model showed a superior predictive performance than that of the ALBI score.

3D multifrequency magnetic resonance elastography in distinguishing endometrial and cervical adenocarcinoma.

OBJECTIVES: To prospectively evaluate the value of tomoelastography in determining the underlying origins of uterine adenocarcinoma. METHODS: This prospective work was approved by our institutional review board, and all patients provided informed consent. 64 patients with histopathologically confirmed adenocarcinomas originated either from the cervix (CAC: cervical adenocarcinoma) or endometrium (EAC: endometrial adenocarcinoma) underwent MRI and tomoelastography examination on a 3.0 T MR scanner. To biomechanically characterize the adenocarcinoma, two MRE-derived parameters maps were provided in the tomoelastography, namely shear wave speed (c, m/s) and loss angle (φ, radian), which represented the stiffness and fluidity, respectively. The MRE-derived parameters were compared by using a two-tailed independent-sample t-test or Mann-Whitney U test. Five morphologic features were also analyzed by using the χ2 test. Logistic regression analysis was used to develop diagnosis models. Delong test was used to compare the receiver operating characteristic curves whith different diagnostic models and evaluate the diagnostic efficiency. RESULTS: CAC were significantly stiffer and behaved more fluid like than EAC (c: 2.58 ± 0.62 m/s vs.2.17 ± 0.72 m/s, p = 0.029, φ, 0.97 ± 0.19 rad vs.0.73 ± 0.26 rad, p < 0.0001). The diagnostic performance for distinguishing CAC from EAC was similar for c (AUC = 0.71) and for φ (AUC = 0.75). For distinguishing CAC from EAC, the AUC of tumor location was the higher than c and φ (AUC = 0.80). A cmobined model consisting of tumor location, c, and φ achieved the best diagnostic performance, with an AUC of 0.88 (77.27% sensitivity and 85.71% specificity). CONCLUSIONS: CAC and EAC displayed their unique biomechanical features. 3D multifrequency MRE provided added value to the conventional morphologic features in distinguishing the two types of diseases.

Predicting treatment response to concurrent chemoradiotherapy in squamous cell carcinoma of the cervix using amide proton transfer imaging and intravoxel incoherent motion imaging.

PURPOSE: The purpose of this study was to investigate whether amide proton transfer (APT) imaging and intravoxel incoherent motion (IVIM) imaging can predict tumor response to concurrent chemoradiotherapy (CCRT) in patients with squamous cell carcinoma of the cervix (SCCC). MATERIAL AND METHODS: Fifty-nine women (mean age, 54 years ± 10 [standard deviation] years; age range: 32-81 years) with pathologically confirmed SCCC underwent magnetic resonance imaging examination of the pelvis including APT and IVIM before concurrent chemoradiotherapy. They were divided into complete remission (CR) and non-CR groups according to therapeutic effect. APT values and IVIM-derived parameters were measured. Intra- and interobserver agreement for IVIM and APT parameters was assessed using intraclass correlation coefficient (ICC) The independent samples t-test was performed to compare the evaluated parameters between the two groups. Predictive performance for treatment response was evaluated by receiver operator characteristic (ROC) curve analysis. RESULTS: There were 38 and 21 patients in the non-CR and CR groups, respectively. Excellent interobserver and intraobserver agreement were obtained for all IVIM and APT parameters, with ICCs ranging from 0.844 to 0.962. Perfusion fraction (f) and APT values were lower in the CR group compared with the non-CR group (both P < 0.05). The combination of f and APT values showed good diagnostic performances in predicting response to concurrent chemoradiotherapy, with an area under the ROC curve of 0.852 (95% CI: 0.744-0.961), 79% sensitivity (95% CI: 63-90%), 90% specificity (95% CI: 70-99%) and 83% accuracy (95% CI: 71-92%). CONCLUSION: APT and IVIM imaging may serve as noninvasive tools for predicting response to concurrent chemoradiotherapy in patients with SCCC.

Altered Functional Connectivity in Pain-Related Brain Regions and Its Correlation with Pain Duration in Bone Metastasis with Cancer Pain.

Bone metastatic pain is thought to be a severe type of cancer pain that has refractory characteristics and a long duration. This study is aimed at exploring the brain functional connectivity (FC) pattern in lung cancer patients with bone metastatic pain. In this study, 27 lung cancer patients with bone metastatic pain (CP+), 27 matched lung cancer patients without pain-related complaints (CP-), and 27 matched healthy controls (HC) were recruited. All participants underwent fMRI data acquisition and clinical assessments. One-way ANOVA or a Mann-Whitney U test was applied to compare clinical data according to data distribution. Seventeen hypothesis-driven pain-related brain regions were selected as regions of interest (ROIs). FC values among pain-related brain regions across the three groups were computed by using ROI-ROI functional connectivity analysis. ANCOVA with a post hoc test was applied to compare FC differences among the three groups. p < 0.05 indicated statistical significance. Correlation analysis was conducted to explore the potential relationship between the FC values and clinical characteristics. Except for years of education, no significant differences were revealed among the three groups in age, gender, or neuropsychological assessment. In the CP+ group, FC alterations were mainly concentrated in the dorsal lateral prefrontal cortex (DLPFC), anterior cingulate cortex (ACC), secondary somatosensory cortex (SII), and amygdala compared to the CP- group. Among these brain regions with statistical differences, FC between the right DLPFC and the right ACC showed a positive correlation with the duration of cancer pain in the CP+ group. In addition, in the CP- group, altered FC was found in the bilateral SII, ACC, and thalamus compared to the HC group. Altered FC in pain-related brain regions may be a brain pattern of bone metastatic pain and may be associated with the long duration of cancer pain.

Comparison of the Differential Diagnostic Performance of Intravoxel Incoherent Motion Imaging and Diffusion Kurtosis Imaging in Malignant and Benign Thyroid Nodules.

Objective: This study aimed to compare the diagnostic capacity between IVIM and DKI in differentiating malignant from benign thyroid nodules. Material and Methods: This study is based on magnetic resonance imaging data of the thyroid with histopathology as the reference standard. Spearman analysis was used to assess the relationship of IVIM-derived parameters D, f, D* and the DKI-derived parameters Dapp and Kapp. The parameters of IVIM and DKI were compared between the malignant and benign groups. Binary logistic regression analysis was performed to establish the diagnostic model, and receiver operating characteristic (ROC) curve analysis was subsequently performed. The DeLong test was used to compare the diagnostic effectiveness of different prediction models. Spearman analysis was used to assess the relationship of Ki-67 expression and parameters of IVIM and DKI. Results: Among the 93 nodules, 46 nodules were malignant, and 47 nodules were benign. The Dapp of DKI-derived parameter was related to the D (P < 0.001, r = 0.863) of IVIM-derived parameter. The Kapp of DKI-derived parameter was related to the D (P < 0.001, r = -0.831) of IVIM-derived parameters. The malignant group had a significantly lower D value (P < 0.001) and f value (P = 0.013) than the benign group. The malignant group had significantly higher Kapp and lower Dapp values (all P < 0.001). The D+f had an area under the curve (AUC) of 0.951. The Dapp+Kapp had an AUC of 0.943. The D+f+Dapp+Kapp had an AUC of 0.954. The DeLong test showed no statistical significance among there prediction models. The D (P = 0.007) of IVIM-derived parameters and Dapp (P = 0.045) of DKI-derived parameter were correlated to the Ki-67 expression. Conclusions: IVIM and DKI were alternative for each other in in differentiating malignant from benign thyroid nodules.

3D convolutional neural network model from contrast-enhanced CT to predict spread through air spaces in non-small cell lung cancer.

PURPOSE: The purpose of this study was to compare the efficacy of five non-invasive models, including three-dimensional (3D) convolutional neural network (CNN) model, to predict the spread through air spaces (STAS) status of non-small cell lung cancer (NSCLC), and to obtain the best prediction model to provide a basis for clinical surgery planning. MATERIALS AND METHODS: A total of 203 patients (112 men, 91 women; mean age, 60 years; age range 22-80 years) with NSCLC were retrospectively included. Of these, 153 were used for training cohort and 50 for validation cohort. According to the image biomarker standardization initiative reference manual, the image processing and feature extraction were standardized using PyRadiomics. The logistic regression classifier was used to build the model. Five models (clinicopathological/CT model, conventional radiomics model, computer vision (CV) model, 3D CNN model and combined model) were constructed to predict STAS by NSCLC. Area under the receiver operating characteristic curves (AUC) were used to validate the capability of the five models to predict STAS. RESULTS: For predicting STAS, the 3D CNN model was superior to the clinicopathological/CT model, conventional radiomics model, CV model and combined model and achieved satisfactory discrimination performance, with an AUC of 0.93 (95% CI: 0.70-0.82) in the training cohort and 0.80 (95% CI: 0.65-0.86) in the validation cohort. Decision curve analysis indicated that, when the probability of the threshold was over 10%, the 3D CNN model was beneficial for predicting STAS status compared to either treating all or treating none of the patients within certain ranges of risk threshold CONCLUSION: The 3D CNN model can be used for the preoperative prediction of STAS in patients with NSCLC, and was superior to the other four models in predicting patients' risk of developing STAS.

Early alterations in cortical morphology after neoadjuvant chemotherapy in breast cancer patients: A longitudinal magnetic resonance imaging study.

There is growing evidence that chemotherapy may have a significant impact on the brains of breast cancer patients, causing changes in cortical morphology. However, early morphological alterations induced by chemotherapy in breast cancer patients are unclear. To investigate the patterns of those alterations, we compared female breast cancer patients (n = 45) longitudinally before (time point 0, TP0) and after (time point 1, TP1) the first cycle of neoadjuvant chemotherapy, using voxel-based morphometry (VBM) and surface-based morphometry (SBM). VBM and SBM alteration data underwent correlation analysis. We also compared cognition-related neuropsychological tests in the breast cancer patients between TP0 and TP1. Reductions in gray matter volume, cortical thickness, sulcal depth, and gyrification index were found in most brain areas, while increments were found to be mainly concentrated in and around the hippocampus. Reductions of fractal dimension mainly occurred in the limbic and occipital lobes, while increments mainly occurred in the anterior and posterior central gyrus. Significant correlations were found between altered VBM and altered SBM mainly in the bilateral superior frontal gyrus. We found no significant differences in the cognition-related neuropsychological tests before and after chemotherapy. The altered brain regions are in line with those associated with impaired cognitive domains in previous studies. We conclude that breast cancer patients showed widespread morphological alterations soon after neoadjuvant chemotherapy, despite an absence of cognitive impairments. The affected brain regions may indicate major targets of early brain damage after chemotherapy.

Nomogram based on CT-derived extracellular volume for the prediction of post-hepatectomy liver failure in patients with resectable hepatocellular carcinoma.

OBJECTIVES: This study aimed to develop and validate a nomogram based on extracellular volume (ECV) derived from computed tomography (CT) for predicting post-hepatectomy liver failure (PHLF) in patients with resectable hepatocellular carcinoma (HCC). METHODS: A total of 202 patients with resectable HCC from two hospitals were enrolled and underwent multiphasic contrast-enhanced CT before surgery. One hundred twenty-one patients from our hospital and 81 patients from another hospital were assigned to the training cohort and the validation cohort, respectively. CT-derived ECV was measured using nonenhanced and equilibrium-phase-enhanced CT images. The nomogram was developed with independent predictors of PHLF. Predictive performance and calibration were assessed by receiver operator characteristic (ROC) analysis and Hosmer-Lemeshow test, respectively. The Delong test was used to compare the areas under the curve (AUCs). RESULTS: CT-derived ECV had a strong correlation with the postoperative pathological fibrosis stage of the background liver (p < 0.001, r = 0.591). The nomogram combining CT-derived ECV, serum albumin (Alb), and serum total bilirubin (Tbil) obtained higher AUCs than the albumin-bilirubin (ALBI) score for predicting PHLF in both the training cohort (0.828 vs. 0.708; p = 0.004) and the validation cohort (0.821 vs. 0.630; p < 0.001). The nomogram showed satisfactory goodness of fit for PHLF prediction in the training and validation cohorts (p = 0.621 and 0.697, respectively). CONCLUSIONS: The nomogram contributes to the preoperative prediction of PHLF in patients with resectable HCC. KEY POINTS: • CT-derived ECV had a strong correlation with the postoperative pathological fibrosis stage of the background liver. • CT-derived ECV was an independent predictor of PHLF in patients with resectable HCC. • The nomogram based on CT-derived ECV showed a superior prediction efficacy than that of clinical models (including Child-Pugh stage, MELD score, and ALBI score).

Myocardial extracellular volume derived from contrast-enhanced chest computed tomography for longitudinal evaluation of cardiotoxicity in patients with breast cancer treated with anthracyclines.

OBJECTIVES: To assess the value of myocardial extracellular volume (ECV) derived from contrast-enhanced chest computed tomography (CT) for longitudinal evaluation of cardiotoxicity in patients with breast cancer (BC) treated with anthracycline (AC). MATERIALS AND METHODS: A total of 1151 patients with BC treated with anthracyclines, who underwent at least baseline, and first follow-up contrast-enhanced chest CT were evaluated. ECV and left ventricular ejection fraction (LVEF) were measured before (ECV0, LVEF0), during ((ECV1, LVEF1) and (ECV2, LVEF2)), and after (ECV3, LVEF3) AC treatment. ECV values were evaluated at the middle of left ventricular septum on venous phase images. Cancer therapy-related cardiac dysfunction (CTRCD) was recorded. RESULTS: Mean baseline LVEF values were 65.85% ± 2.72% and 102 patients developed CTRCD. The mean ECV0 was 26.76% ± 3.03% (N0 = 1151). ECV1, ECV2, and ECV3 (median interval: 61 (IQR, 46-75), 180 (IQR, 170-190), 350 (IQR, 341-360) days from baseline) were 31.32% ± 3.10%, 29.60% ± 3.24%, and 32.05% ± 3.58% (N1 = 1151, N2 = 841, N3 = 511). ECV1, ECV2, and ECV3 were significantly higher than ECV0 (p < 0.001). ECV0 and ECV1 showed no difference between CTRCD (+) and CTRCD (-) group (p1 = 0.150; p2 = 0.216). However, ECV2 and ECV3 showed significant differences between the two groups (p3 < 0.001; p4 < 0.001). CONCLUSION: CT-derived ECV is a potential biomarker for dynamic monitoring AC cardiotoxicity in patients with BC.

MRI-based radiomics to compare the survival benefit of induction chemotherapy plus concurrent chemoradiotherapy versus concurrent chemoradiotherapy plus adjuvant chemotherapy in locoregionally advanced nasopharyngeal carcinoma: A multicenter study.

BACKGROUND AND PURPOSE: It remains uncertain whether induction chemotherapy (IC) plus concurrent chemoradiotherapy (CCRT) or CCRT plus adjuvant chemotherapy (AC) is more effective in patients with locoregionally advanced nasopharyngeal carcinoma (LA-NPC). This study aimed to develop and validate a joint radiomic and clinical signature (RCS) for the prognostic stratification of LA-NPCs and to identify patients who might benefit more from IC + CCRT or CCRT + AC. MATERIALS AND METHODS: Overall, 893 LA-NPC patients who received IC + CCRT or CCRT + AC were enrolled from four hospitals. RCS based on pretreatment magnetic resonance images and clinical data was constructed for predicting 5-year progression-free survival (PFS). The predictive ability of the RCS and TNM staging system for 5-year PFS, locoregional relapse-free survival (LRRFS), distant metastasis-free survival (DMFS) and overall survival (OS) were compared by Harrell's concordance indices (C-indices). Patients were divided into high- and low-risk subgroups based on RCS scores. The survival benefit of IC + CCRT vs. CCRT + AC in different subgroups was compared by Kaplan-Meier survival curves. RESULTS: The RCS combining the radiomic signature, TNM stage and EBV DNA demonstrated significantly higher C-indices than TNM stage for predicting 5-year PFS, LRRFS, DMFS and OS in the training and validation cohorts. In the high-risk group (RCS score ≥ 0.25), CCRT + AC achieved significantly better PFS, LRRFS, DMFS and OS than IC + CCRT. In the low-risk group (RCS score < 0.25), IC + CCRT yielded significantly better outcomes than CCRT + AC. CONCLUSION: The RCS provides a noninvasive way to predict the outcomes of LA-NPC and helps identify patients who may benefit more from IC + CCRT vs. CCRT + AC.

Dual-source dual-energy computed tomography-derived quantitative parameters combined with machine learning for the differential diagnosis of benign and malignant thyroid nodules.

BACKGROUND: This study aimed to investigate the ability of quantitative parameter-derived dual-source dual-energy computed tomography (DS-DECT) combined with machine learning to distinguish between benign and malignant thyroid nodules. METHODS: Patients with thyroid nodules and pathological surgical results who underwent preoperative DS-DECT were selected. Quantitative parameter-derived DS-DECT was applied to classify benign and malignant nodules. Then, machine learning and binary logistic regression analysis models were constructed using the DS-DECT quantitative parameters to distinguish between benign and malignant nodules. The receiver operating characteristic curve was used to assess the diagnostic performance. The DeLong test was used to compare the diagnostic efficacy. RESULTS: One hundred and thirty patients with 139 confirmed thyroid nodules were involved in the study. The malignant group had a significantly higher iodine concentrationnodule (arterial phase) (P=0.001), normalized iodine concentration (arterial phase) (P=0.002), iodine concentration difference (P<0.001), spectral curve slope (nonenhancement) (P=0.007), spectral curve slope (arterial phase) (P=0.001), effective atomic number (nonenhancement) (P<0.001), and effective atomic number (arterial phase) (P=0.039) than the benign group. The binary logistic regression analysis model had an AUC (area under the curve) of 0.76, a sensitivity of 0.821, and a specificity of 0.667. The machine learning model had an AUC of 0.86, a sensitivity of 0.822, specificity of 0.791 in the training cohort, an AUC of 0.84, a sensitivity of 0.727, and specificity of 0.750 in the testing cohort. CONCLUSIONS: Multiple quantitative parameters of DS-DECT combined with machine learning could differentiate between benign and malignant thyroid nodules.

Application of machine learning with multiparametric dual-energy computed tomography of the breast to differentiate between benign and malignant lesions.

BACKGROUND: Multiparametric dual-energy computed tomography (mpDECT) is widely used to differentiate various kinds of tumors; however, the data regarding its diagnostic performance with machine learning to diagnose breast tumors is limited. We evaluated univariate analysis and machine learning performance with mpDECT to distinguish between benign and malignant breast lesions. METHODS: In total, 172 patients with 214 breast lesions (55 benign and 159 malignant) who underwent preoperative dual-phase contrast-enhanced DECT were included in this retrospective study. Twelve quantitative features were extracted for each lesion, including CT attenuation (precontrast, arterial, and venous phases), the arterial-venous phase difference in normalized effective atomic number (nZeff), normalized iodine concentration (NIC), and slope of the spectral Hounsfield unit (HU) curve (λHu). Predictive models were developed using univariate analysis and eight machine learning methods [logistic regression, extreme gradient boosting (XGBoost), stochastic gradient descent (SGD), linear discriminant analysis (LDA), adaptive boosting (AdaBoost), random forest (RF), decision tree, and linear support vector machine (SVM)]. Classification performances were assessed based on the area under the receiver operating characteristic curve (AUROC). The best performances of the conventional univariate analysis and machine learning methods were compared using the Delong test. RESULTS: The univariate analysis showed that the venous phase λHu had the highest AUROC (0.88). Machine learning with mpDECT achieved an excellent and stable diagnostic performance, as shown by the mean classification performances in the training dataset (AUROC, 0.88-0.99) and testing (AUROC, 0.83-0.96) datasets. The performance of the AdaBoost model based on mpDECT was more stable than the other machine learning models and superior to the univariate analysis (AUROC, 0.96 vs. 0.88; P<0.001). CONCLUSIONS: The performance of the AdaBoost classifier based on mpDECT data achieved the highest mean accuracy compared to the other machine learning models and univariate analysis in differentiating between benign and malignant breast lesions.

Development and Validation of a CT-Based Signature for the Prediction of Distant Metastasis Before Treatment of Non-Small Cell Lung Cancer.

RATIONALE AND OBJECTIVES: To develop and validate a radiomics model, a clinical-semantic model and a combined model by using standard methods for the pretreatment prediction of distant metastasis (DM) in patients with non-small-cell lung cancer (NSCLC) and to explore whether the combined model provides added value compared to the individual models. MATERIALS AND METHODS: This retrospective study involved 356 patients with NSCLC. According to the image biomarker standardization initiative reference manual, we standardized the image processing and feature extraction using in-house software. Finally, 6692 radiomics features were extracted from each lesion based on contrast-enhanced chest CT images. The least absolute shrinkage selection operator and the recursive feature elimination algorithm were used to select features. The logistic regression classifier was used to build the model. Three models (radiomics model, clinical-semantic model and combined model) were constructed to predict DM in NSCLC. Area under the receiver operating characteristic curves were used to validate the ability of the three models to predict DM. A visual nomogram based on the combined model was developed for DM risk assessment in each patient. RESULTS: The receiver operating characteristic curve showed predictive performance for DM of the radiomics model (area under the curve [AUC] values for training and validation were 0.76 [95% CI, 0.704 - 0.820] and 0.76 [95% CI, 0.653 - 0.858], respectively). The combined model had AUCs of 0.78 (95% CI, 0.723 - 0.835) and 0.77 (95% CI, 0.673 - 0.870) in the training and validation cohorts, respectively. Both the radiomics model and combined model performed better than the clinical-semantic model (0.70 [95% CI, 0.634 - 0.760] and 0.67 [95% CI, 0.554 - 0.787] in the training and validation cohorts, respectively). CONCLUSION: The radiomics model and combined model may be useful for the prediction of DM in patients with NSCLC.

Development and validation of a radiogenomics model to predict axillary lymph node metastasis in breast cancer integrating MRI with transcriptome data: A multicohort study.

Introduction: To develop and validate a radiogenomics model for predicting axillary lymph node metastasis (ALNM) in breast cancer compared to a genomics and radiomics model. Methods: This retrospective study integrated transcriptomic data from The Cancer Genome Atlas with matched MRI data from The Cancer Imaging Archive for the same set of 111 patients with breast cancer, which were used as the training and testing groups. Fifteen patients from one hospital were enrolled as the external validation group. Radiomics features were extracted from dynamic contrast-enhanced (DCE)-MRI of breast cancer, and genomics features were derived from differentially expressed gene analysis of transcriptome data. Boruta was used for genomics and radiomics data dimension reduction and feature selection. Logistic regression was applied to develop genomics, radiomics, and radiogenomics models to predict ALNM. The performance of the three models was assessed by receiver operating characteristic curves and compared by the Delong test. Results: The genomics model was established by nine genomics features, and the radiomics model was established by three radiomics features. The two models showed good discrimination performance in predicting ALNM in breast cancer, with areas under the curves (AUCs) of 0.80, 0.67, and 0.52 for the genomics model and 0.72, 0.68, and 0.71 for the radiomics model in the training, testing and external validation groups, respectively. The radiogenomics model integrated with five genomics features and three radiomics features had a better performance, with AUCs of 0.84, 0.75, and 0.82 in the three groups, respectively, which was higher than the AUC of the radiomics model in the training group and the genomics model in the external validation group (both P < 0.05). Conclusion: The radiogenomics model combining radiomics features and genomics features improved the performance to predict ALNM in breast cancer.