Irreversible electroporation (IRE) is an ablation procedure in which cell death is induced by ultrashort electrical pulses. In this unicentric retrospective study we investigated the influence of needle positioning on ablation success. 15 IREs with residual tumor after ablation, detected in the first follow-up MRI, were included, and compared with 30 successful ablations. Evaluation of needle geometry revealed significantly higher values for needle divergence (NDiv, 7.0° vs. 3.7°, p = 0.02), tumor-center-to-ablation-center distance (TACD, 11.6 vs. 3.2 mm, p < 0.001), tumor-to-needle distance (4.7 vs. 1.9 mm, p = 0.04), and tumor diameter per needle (7.5 vs. 5.9 mm/needle, p = 0.01) in patients with residual tumor. The average number of needles used was higher in the group without residual tumor after ablation (3.1 vs. 2.4, p = 0.04). In many cases with residual tumor, needle depth was too short (2.1 vs. 6.8 mm tumor overlap beyond the most proximal needle tip, p < 0.01). The use of a stereotactic navigation system in 10 cases resulted in a lower NDiv value (2.1° vs. 5.6°, p < 0.01). Thus, correct needle placement seems to be a crucial factor for success and the assistance of a stereotactic navigation system might be helpful. As most important geometrical parameter TACD could be identified. Main reasons for high TACD were insufficient needle depth and a lesion location out of the needle plane.

Background/Objectives: Despite the growing popularity of large language models (LLMs), there remains a notable lack of research examining their role in workplace management. This study aimed to address this gap by evaluating the performance of ChatGPT-3.5, ChatGPT-4.0, Gemini, and Gemini Advanced as famous LLMs in responding to workplace management questions specific to radiology. Methods: ChatGPT-3.5 and ChatGPT-4.0 (both OpenAI, San Francisco, CA, USA) and Gemini and Gemini Advanced (both Google Deep Mind, Mountain View, CA, USA) generated answers to 31 pre-selected questions on four different areas of workplace management in radiology: (1) patient management, (2) imaging and radiation management, (3) learning and personal development, and (4) administrative and department management. Two readers independently evaluated the answers provided by the LLM chatbots. Three 4-point scores were used to assess the quality of the responses: (1) overall quality score (OQS), (2) understandabilityscore (US), and (3) implementability score (IS). The mean quality score (MQS) was calculated from these three scores. Results: The overall inter-rater reliability (IRR) was good for Gemini Advanced (IRR 79%), Gemini (IRR 78%), and ChatGPT-3.5 (IRR 65%), and moderate for ChatGPT-4.0 (IRR 54%). The overall MQS averaged 3.36 (SD: 0.64) for ChatGPT-3.5, 3.75 (SD: 0.43) for ChatGPT-4.0, 3.29 (SD: 0.64) for Gemini, and 3.51 (SD: 0.53) for Gemini Advanced. The highest OQS, US, IS, and MQS were achieved by ChatGPT-4.0 in all categories, followed by Gemini Advanced. ChatGPT-4.0 was the most consistently superior performer and outperformed all other chatbots (p < 0.001–0.002). Gemini Advanced performed significantly better than Gemini (p = 0.003) and showed a non-significant trend toward outperforming ChatGPT-3.5 (p = 0.056). ChatGPT-4.0 provided superior answers in most cases compared with the other LLM chatbots. None of the answers provided by the chatbots were rated “insufficient”. Conclusions: All four LLM chatbots performed well on workplace management questions in radiology. ChatGPT-4.0 outperformed ChatGPT-3.5, Gemini, and Gemini Advanced. Our study revealed that LLMs have the potential to improve workplace management in radiology by assisting with various tasks, making these processes more efficient without requiring specialized management skills.

Purpose: Evaluation of the influence of intrinsic and extrinsic conditions on ablation zone volumes (AZV) after microwave ablation (MWA). Methods: Retrospective analysis of 38 MWAs of therapy-naïve liver tumours performed with the NeuWave PR probe. Ablations were performed either in the ‘standard mode’ (65 W, 10 min) or in the ‘surgical mode’ (95 W, 1 min, then 65 W, 10 min). AZV measurements were obtained from contrast-enhanced computed tomography immediately post-ablation. Results: AZVs in the ‘standard mode’ were smaller than predicted by the manufacturer (length 3.6 ± 0.6 cm, 23% below 4.7 cm; width 2.7 ± 0.6, 23% below 3.5 cm). Ablation zone past the tip was limited to 6 mm in 28/32 ablations. Differences in AZV between the ‘surgical mode’ and ‘standard mode’ were not significant (15.6 ± 7.8 mL vs. 13.9 ± 8.8 mL, p = 0.6). AZVs were significantly larger in case of hepatocellular carcinomas (HCCs) (n = 19) compared to metastasis (n = 19; 17.8 ± 9.9 mL vs. 10.1 ± 5.1 mL, p = 0.01) and in non-perivascular tumour location (n = 14) compared to perivascular location (n = 24, 18.7 ± 10.4 mL vs. 11.7 ± 6.1 mL, p = 0.012), with both factors remaining significant in two-way analysis of variance (HCC vs. metastasis: p = 0.02; perivascular vs. non-perivascular tumour location: p = 0.044). Conclusion: Larger AZVs can be expected in cases of HCCs compared with metastases and in non-perivascular locations. Using the ‘surgical mode’ does not increase AZV significantly.

Purpose: To investigate whether the iodized oil (Lipiodol, Guerbet Group, Villepinte, France) retention pattern influences the treatment efficacy of combined transarterial Lipiodol injection (TLI) and thermal ablation in patients with hepatocellular carcinoma (HCC). Methods: Data of 198 patients (280 HCC lesions), who underwent TLI plus computed tomography (CT)-guided thermal ablation at three separate medical institutions between June 2014 and September 2020, were reviewed and analyzed. The Lipiodol retention pattern was classified as complete or incomplete based on non-enhanced CT at the time of ablation. The primary outcome was local recurrence-free survival (LRFS) for lesions; the secondary outcome was overall survival (OS) for patients. Propensity score matching (PSM) was performed using a caliper width of 0.1 between the two groups. Differences in LRFS and OS between the two groups were compared using the log-rank test. Results: A total of 133 lesions exhibited a complete Lipiodol retention pattern, while 147 exhibited an incomplete pattern. After PSM analysis of baseline characteristics of the lesions, 121 pairs of lesions were matched. LRFS was significantly longer for lesions exhibiting complete retention than for those exhibiting incomplete retention (P = 0.030). After PSM analysis of patient baseline characteristics, 74 pairs of patients were matched. There was no significant difference in OS between the two groups (P = 0.456). Conclusion: Lipiodol retention patterns may influence the treatment efficacy of combined TLI and thermal ablation for HCC lesions. However, a survival benefit for the Lipiodol retention pattern among HCC patients was not observed and needs further confirmation.

Background: Accurate prediction of tumor response after drug-eluting beads transarterial chemoembolization (DEB-TACE) remains challenging in hepatocellular carcinoma (HCC), given tumor heterogeneity and dynamic changes over time. Existing prediction models based on single timepoint imaging do not capture dynamic treatment-induced changes. This study aims to develop and validate a predictive model that integrates deep learning and machine learning algorithms on longitudinal contrast-enhanced MRI (CE-MRI) to predict treatment response in HCC patients undergoing DEB-TACE. Methods: This retrospective study included 202 HCC patients treated with DEB-TACE from 2004 to 2023, divided into a training cohort ( n = 141) and validation cohort ( n = 61). Radiomics and deep learning features were extracted from standardized longitudinal CE-MRI to capture dynamic tumor changes. Feature selection involved correlation analysis, minimum redundancy maximum relevance, and least absolute shrinkage and selection operator regression. The patients were categorized into two groups: the objective response group ( n = 123, 60.9%; complete response = 35, 28.5%; partial response = 88, 71.5%) and the non-response group ( n = 79, 39.1%; stable disease = 62, 78.5%; progressive disease = 17, 21.5%). Predictive models were constructed using radiomics, deep learning, and integrated features. The area under the receiver operating characteristic curve (AUC) was used to evaluate the performance of the models. Results: We retrospectively evaluated 202 patients (62.67 ± 9.25 years old) with HCC treated after DEB-TACE. A total of 7,182 radiomics features and 4,096 deep learning features were extracted from the longitudinal CE-MRI images. The integrated model was developed using 13 quantitative radiomics features and 4 deep learning features and demonstrated acceptable and robust performance with an receiver operating characteristic curve (AUC) of 0.941 (95%CI: 0.893–0.989) in the training cohort, and AUC of 0.925 (95%CI: 0.850–0.998) with accuracy of 86.9%, sensitivity of 83.7%, as well as specificity of 94.4% in the validation set. Conclusions: This study presents a predictive model based on longitudinal CE-MRI data to estimate tumor response to DEB-TACE in HCC patients. By capturing tumor dynamics and integrating radiomics features with deep learning features, the model has the potential to guide individualized treatment strategies and inform clinical decision-making regarding patient management.