Radiology. Imaging cancer最新文献

Purpose To evaluate the reproducibility of apparent diffusion coefficient (ADC) measurements in head and neck squamous cell carcinoma (HNSCC) using a 1.5-T MR-linac (MRI-guided linear accelerator) system. Materials and Methods In this retrospective study, conducted between February 2021 and January 2024, patients with HNSCC lesions underwent echo-planar imaging diffusion-weighted MRI on a 1.5-T MR-linac system at two time points before the start of radiation therapy. Mean and median ADC values and volumes were measured for each lesion. Absolute and percent reproducibility coefficients (RCs) were calculated. Linear regression analyses and F tests were performed to determine whether the lesion volume or time between scans impacted reproducibility. Results The final cohort consisted of 37 patients (36 male, one female; median age, 63 years [IQR, 47-80 years]), with 34 primary tumors and 55 lymph nodes included in the analysis. For primary tumors and lymph nodes, the median of the mean ADC, median ADC, and volume were 1.19 × 10^-3 mm²/sec (IQR, 1.01-1.34 × 10^-3 mm²/sec) and 1.02 × 10^-3 mm²/sec (IQR, 0.88-1.33 × 10^-3 mm²/sec), 1.16 × 10^-3 mm²/sec (IQR, 0.99-1.36 × 10^-3 mm²/sec) and 1.03 × 10^-3 mm²/sec (IQR, 0.90-1.35 × 10^-3 mm²/sec), and 2.9 cm³ (IQR, 1.5-8.9 cm³) and 3.6 cm³ (IQR, 1.7-8.3 cm³), respectively. The respective RC values of mean ADC were 0.365 × 10^-3 mm²/sec and 0.355 × 10^-3 mm²/sec for tumors and lymph nodes, and the respective percent RC values were 29.9% and 31.1%; similar values were obtained for the median ADC. Reproducibility did not correlate with either the lesion volume or scan interval, but a trend toward poorer reproducibility with smaller volumes was observed. Conclusion This MR-linac sequence demonstrated acceptable reproducibility for detecting larger ADC changes but might still miss some clinically significant changes. Keywords: Radiation Therapy, MR-Diffusion Weighted Imaging, Radiation Therapy/Oncology, Head/Neck © RSNA 2026 Supplemental material is available for this article.

Purpose To develop and validate a risk score model for predicting aggressive recurrence (AR) of hepatocellular carcinoma (HCC) after microwave ablation (MWA). Materials and Methods This retrospective study included consecutive patients with solitary HCC (≤3 cm) who underwent preoperative contrast-enhanced MRI and MWA from June 2014 to October 2019. In the training set, multivariable Cox regression analysis identified clinical, laboratory, and MWA parameters and MRI features associated with AR. The resulting predictive model, named Aggressive Recurrence after Microwave Ablation (ARMA), was evaluated in the test set using the concordance index (C-index) and compared with the Barcelona Clinic Liver Cancer staging system and known risk factors. Kaplan-Meier survival curves were compared using the log-rank test. Results A total of 393 eligible patients (mean age, 57.7 years ± 11.2 [SD]; 299 [76.1%] male) were included. AR occurred in 79 of 275 (28.7%) patients in the training set and 37 of 118 (31.4%) patients in the test set. Independent predictors of AR-free survival (ARFS) were periportal location (hazard ratio [HR], 1.72; P < .001), ablation duration (HR, 1.00; P = .02), and arterial phase hyperenhancement proportion (APHE) at MRI (HR, 1.77; P = .002). The ARMA model showed better predictive performance for ARFS compared with other staging systems and known risk factors (C-indexes, 0.72 vs 0.58-0.63; P value range, <.001-.003). High-risk patients (ARMA score ≥ 0.700) showed worse ARFS (P < .001) and recurrence-free survival (P = .002) than low-risk patients (ARMA score < 0.700). Conclusion The ARMA score, integrating periportal location, ablation duration, and APHE, predicted ARFS and identified patients at high risk for AR. Keywords: Model Training, Model Validation, Ablation Techniques, Liver, Outcomes Analysis, Aggressive Recurrrence, Microwave Ablation, Periportal Hepatocellular, Carcinoma, Arterial Phase Hyperenhancement Proportion, Ablation Duration Supplemental material is available for this article. © RSNA, 2026.

Purpose To determine whether combining the arterial phase onset time to washout onset time interval (AWTI) with washout onset time improves the accuracy of contrast-enhanced US (CEUS) Liver Imaging Reporting and Data System (LI-RADS) and reduces the incidence of hepatocellular carcinoma (HCC) being misclassified as LR-M. Materials and Methods CEUS and clinical data from patients with focal liver lesions (FLLs), collected between January 2019 and October 2024, were retrospectively analyzed. The AWTI was calculated for all FLLs with washout < 60 seconds. A revised classification was proposed: (a) washout < 45 seconds + AWTI < 21 seconds for LR-M; and (b) washout ≥ 45 seconds + AWTI ≥ 21 seconds for LR-5. Diagnostic performance of the revised classification was compared with that of the standard LI-RADS with washout < 60 seconds (and ≥60 seconds). Results The study included 352 patients (median age, 56 years [IQR, 46-65]; 284 male), each with one FLL. Among HCCs, 75.9% exhibited washout ≥ 60 seconds. In contrast, 75.0% of intrahepatic cholangiocarcinomas, 52.0% of metastatic liver carcinomas, and 66.7% of other malignancies demonstrated washout < 45 seconds. Of benign FLLs, 12.8% showed washout ≥ 60 seconds. Among FLLs classified as LR-5, 92.6% were HCCs; among FLLs classified as LR-M, 41.8% were HCCs. The optimal AWTI cutoff to distinguish LR-M from LR-5 was 21 seconds. The revised LR-M (washout < 45 seconds + AWTI < 21 seconds) significantly increased the positive predictive value to 92.6% (P < .05). The revised LR-5 (washout ≥ 45 seconds + AWTI ≥ 21 seconds) significantly improved the sensitivity and negative predictive value to 89.0% and 87.0%, respectively (both P < .05). The diagnostic accuracy and area under the receiver operating characteristic curve were 88.4% and 0.88, respectively, despite slight decreases in the specificity and positive predictive value. Conclusion Combining washout onset time with AWTI as an alternative for standard washout onset time (<60 sec/≥60 sec) reduced misclassification of HCC and significantly improved the diagnostic performance of CEUS LI-RADS for LR-M and LR-5. Keywords: Ultrasound, Reconstruction Algorithms, Diagnosis, Classification, Ultrasound-Contrast, Abdomen/GI, Liver, Contrast Agents-Intravenous, Efficacy Studies, Hepatocellular Carcinoma, Contrast-enhanced Ultrasound, Liver Imaging Reporting and Data System, Washout Time Supplemental material is available for this article. © RSNA, 2025 See also commentary by Hui and Chiang in this issue.

Purpose To evaluate the association of two minimum ablation margin (MAM) estimation methods: manual three-dimensional segmentation and biomechanical deformable image registration (DIR) with local tumor progression-free survival (LTPFS) after percutaneous lung cryoablation of colorectal pulmonary metastases. Materials and Methods This retrospective single-institution study included patients who underwent percutaneous lung cryoablation between May 2012 and March 2023 (median follow-up, 3.6 years; range, 0.8-9.6 years). Manual three-dimensional MAM was calculated for all tumors, and DIR-based MAM was estimated in a subset with intraprocedural full-thorax imaging. Cox proportional hazards regression assessed associations between MAM and LTPFS. The primary outcome was 2-year LTPFS, with a secondary outcome of whether DIR-based MAM was associated with LTPFS. Results Thirty-eight patients (76 colorectal pulmonary metastases in 63 procedures; mean age, 58 years ± 12 [SD]; 19 female patients) were included. Technical success was 100% (63 of 63). Per-lesion LTPFS at 1 and 2 years was 87% and 84%. Five tumors (6.6%; five of 76) had no margin (0 mm), 55 tumors (72%; 55 of 76) had a MAM 0-5 mm, and 16 (21%; 16 of 76) had an MAM greater than 5 mm. Median LTPFS was 6 months for the 0-mm group and was not reached for the MAM 0-5-mm and the greater than 5-mm groups (P < .001). Receiver operating characteristic analysis for 2-year local progression yielded an area under the receiver operating characteristic curve (AUC) of 0.81 (95% CI: 0.62, 0.95). Among 36 tumors (22 patients) with full-lung imaging available, the DIR-based MAM had an AUC of 0.76 (95% CI: 0.55, 0.98), and the manual method had an AUC of 0.79 (95% CI: 0.52, 1.00). At multivariable analysis, MAM greater than 0-5 mm (hazard ratio [HR], 0.05; P < .001), MAM greater than 5 mm (hazard ratio, 0.03; P = .005), and ablation volume (hazard ratio, 0.93; P < .001) were independently associated with reduced LTP. Conclusion Quantitative MAMs are strongly associated with LTPFS, underscoring the importance of achieving adequate margins in percutaneous lung cryoablation. Keywords: Pulmonary Cryoablation, Deformable Image Registration, Colorectal Cancer, Pulmonary Metastasis, Minimum Ablation Margin © RSNA, 2026.