Song Z, et al. Korean J Radiol. 2020 Jul;21:e160. https://doi.org/10.3348/kjr.2020.0254

Objective: To derive and validate a location-specific radiomics score (Rad-score) based on noncontrast computed tomography for predicting poor deep and lobar spontaneous intracerebral hemorrhage (SICH) outcome.Methods: In total, 494 SICH patients from multiple centers were retrospectively reviewed. Poor outcome was considered mRS 3–6 at 6 months. The Rad-score was derived using optimal radiomics features. The optimal location-specific Rad-score cut-offs for poor deep and lobar SICH outcomes were identified using receiver operating characteristic curve analysis. Univariable and multivariable analyses were used to determine independent poor outcome predictors. The combined models for deep and lobar SICH were constructed using independent predictors of poor outcomes, including dichotomized Rad-score in the derivation cohort, which was validated in the validation cohort.Results: Of 494 SICH patients, 392 (79%) had deep SICH, and 373 (76%) had poor outcomes. The Glasgow Coma Scale score, haematoma enlargement, haematoma location, haematoma volume and Rad-score were independent predictors of poor outcomes (all P < 0.05). Cut-offs of Rad-score, 82.90 (AUC = 0.794) in deep SICH and 80.77 (AUC = 0.823) in lobar SICH, were identified for predicting poor outcomes. For deep SICH, the AUCs of the combined model were 0.856 and 0.831 in the derivation and validation cohorts, respectively. For lobar SICH, the combined model AUCs were 0.866 and 0.843 in the derivation and validation cohorts, respectively.Conclusion: Location-specific Rad-scores and combined models can identify subjects at high risk of poor deep and lobar SICH outcomes, which could improve clinical trial design by screening target patients.

Objectives: The BRAF^V600E is the most common oncogene in thyroid cancer and is associated with the aggressiveness of papillary thyroid carcinoma (PTC). The aim of this study was to investigate the effectiveness of the arterial enhancement fraction (AEF) and dual-layer detector spectral computed tomography (DLCT) parameters for predicting the BRAF^V600E mutation in PTC. Methods: A total of 237 patients with PTC who underwent DLCT and BRAF^V600E mutation detection (mutant group: n = 187; wild group: n = 50) were retrospectively reviewed. The receiver operating characteristic curves evaluated the effectiveness of the prediction models based on the significantly different variables using logistic regression analysis. The nomogram of the prediction model with the highest AUC in the validation cohort was constructed. Results: The AUCs of the DLCT+ Hashimoto’s thyroiditis (HT) and AEF + DLCT + HT prediction models were 0.901 and 0.896, respectively, in the training cohort and 0.801 and 0.853 in the validation cohort. The calibration curve revealed the good agreement between the prediction results and the actual observations using the AEF + DLCT + HT model. The DCA demonstrated that the model can provide net benefit for all threshold probabilities. Conclusions: As an effective and visually noninvasive prediction tool, the AEF + DLCT + HT-based nomogram presented satisfactory effectiveness in preoperatively predicting the BRAF^V600E mutation in PTC.

Background/Objectives: Accurately distinguishing malignancy in thyroid micronodules (≤10 mm) is crucial for clinical management, yet it is challenging due to the limitations of conventional ultrasonography-guided biopsy. This study aims to evaluate the predictive value of dual-layer spectral computed tomography (DSCT)-derived arterial enhancement fraction (AEF) in diagnosing thyroid microcarcinomas. Methods: In the study, 321 pathologically confirmed thyroid micronodules (benign = 131, malignant = 190) from Chongqing General Hospital underwent preoperative DSCT. Quantitative parameters of DSCT, including the normalized iodine concentration (NIC), normalized effective atomic number (NZ_eff), and slope of the spectral Hounsfield unit curve (λHU_(40–100)), were assessed. Both single-energy CT (SECT)-derived AEF (AEF_S) and DSCT-derived AEF (AEF_D) were calculated. Conventional image features included microcalcifications and enhancement blurring. Correlation between AEF_D and AEF_S was determined using Spearman’s correlation coefficient. Diagnostic performance was evaluated by calculating the area under the curve (AUC) using receiver operating characteristic (ROC) analysis. Results: Malignant micronodules exhibited significantly lower AEF_D (0.958 vs. 1.259, p < 0.001) and AEF_S (0.964 vs. 1.436, p < 0.001) versus benign nodules. Arterial phase parameters—APλHU_(40–100), APNIC, APNZ_eff—differed significantly between groups (all p < 0.001), whereas venous phase parameters (VPλHU_(40–100), VPNIC, VPNZ_eff) showed no differences (all p > 0.05). Multivariate analysis revealed that λHU_(40–100) as an independent predictor of malignancy, with an odds ratio (OR) of 0.600 (95% confidence interval (CI): 0.437–0.823; p = 0.002) and an AUC of 0.752 (95% CI: 0.698–0.806). A significant positive correlation was identified between AEF_D and AEF_S (r = 0.710; p < 0.001). For diagnosing malignancy, AEF_D demonstrated superior overall performance (AUC: 0.794; sensitivity: 70.5%; specificity: 81.7%; accuracy: 75.1%) to AEF_S (0.753; 71.1%; 74.0%; 72.3%), APλHU_(40–100) (0.752; 68.9%; 75.6%; 71.7%), and calcification (0.573; 21.6%; 92.4%; 50.5%). Clinically, AEF_D reduced the unnecessary biopsy rate to 18.3%, preventing 107 procedures in our cohort. Conclusions: AEF_D and AEF_S demonstrated strong correlation and comparable diagnostic performance in the evaluation of thyroid micronodules. Furthermore, AEF_D showed favorable diagnostic efficacy compared to both spectral parameters and conventional imaging feature. More importantly, the application of AEF_D significantly reduced unnecessary biopsy rates, highlighting its clinical value in optimizing patient management.

Objective: Solitary pulmonary nodules (SPNs) have become increasingly common. This study aims to improve lung cancer screening and management by enhancing malignant risk prediction through the integration of the classical Brock, Mayo, and PKUPH models. Methods: Patients diagnosed with SPNs were regarded as subjects. Prospective and retrospective follow-up were employed to determine the outcomes of SPNs. TheA weighted average method was applied to calculate the weight coefficient and risk scores of the combined SPNs model. The receiver operating characteristic (ROC) curve and precision-recall curve (PRC) were used to evaluate the discrimination and classification performance of the combined SPNs model and individual models. Decision curve analysis, Hosmer–Lemeshow test, and calibration curves were used to assess the clinical utility and calibration of these models. Results: A total of 662 patients were included, and the malignant rate was 13.90%. The combined SPNs model demonstrated the best discrimination and classification ability, with an area under the ROC curve (AUROC) of 0.87 and an area under the PR curve (AUPRC) of 0.62 in the training set. In the validation set, the AUROC was 0.86 and the AUPRC was 0.51. The combined SPNs model exhibited greater net benefit compared with each individual model. The combined model stratified SPNs into high-, medium-, and low-risk categories, with malignancy probabilities of 3.49%, 16.67%, and 62.50%, respectively. All models showed satisfactory calibration, as indicated by the Hosmer–Lemeshow test ( p > 0.1). Conclusion: The combined SPNs model improved the accuracy and generalizability of malignant risk prediction for SPNs, offering a feasible alternative to address recalibration and out-of-distribution issues when applying existing predictive models to Chinese lung cancer screening population in routine clinical practice.