1 January 2021 in Korean Journal of Radiology
Kim HS, et al. Korean J Radiol. 2021 Apr;22:e47. https://doi.org/10.3348/kjr.2020.0837
14 April 2022 in Ultrasonography
Purpose: This study investigated the ultrasound (US) features of malignancy in patients with Hürthle cell neoplasms (HCNs) of the thyroid gland.Methods: The present study included 139 HCNs that had undergone surgical excision at a single institution from 1996 to 2020 and had preoperative US images. The sonographic characteristics of HCNs were correlated with their pathological results. The US findings associated with malignancy were explored using logistic regression analysis, and the diagnostic performance and cutoff were assessed using receiver operating characteristic analysis.Results: The most common US findings of HCNs were a solid content (76.3%), oval to round shape (100%), hypoechogenicity (70.5%), a smooth margin (95.0%), the halo sign (90.6%), and no calcifications (93.5%). HCNs were commonly smaller in pathologic measurements than in US measurements (smaller, same, and greater than US measurements in 60.4%, 21.6%, and 18.0% of HCNs, respectively; P<0.001). On US, malignant nodules were significantly larger than benign nodules (3.4±1.6 cm vs. 2.2±1.2 cm, P3.35 cm might be used as a criterion to suggest malignancy. The size of HCNs often showed discrepancies between US and pathologic measurements.
1 February 2024 in Journal of Magnetic Resonance Imaging
1 January 2025 in Journal of the Korean Society of Radiology
김민제, et al. J Korean Soc Radiol. 2025 Nov;86(6):830-846. https://doi.org/10.3348/jksr.2025.0076
1 October 2020 in Ultrasonography
Excision of metastatic lesions is an important treatment strategy in patients with malignant melanoma, both at the initial diagnosis and upon recurrence. Since nonpalpable lesions cannot be easily visualized in the surgical field, we evaluated the effectiveness of ultrasound (US)-guided tattooing using a charcoal suspension for the localization of nonpalpable metastatic lesions of malignant melanoma. Between November 2009 and June 2019, we retrospectively reviewed 65 nonpalpable lesions in 29 patients with malignant melanoma who underwent preoperative US-guided tattooing using a charcoal suspension for histologically confirmed or suspected metastases. The characteristics of the tattooed lesions were analyzed. The effectiveness of the procedure was evaluated based on the detection rate in the surgical field and the presence or absence of residua on postoperative follow-up US. Procedure-related complications were also analyzed. Of 65 lesions, 33 (50.8%) were histologically confirmed as metastases before the tattooing procedure, while the other 32 were suspected of being metastases based on imaging studies. The mean lesion size was 9.8 mm (range, 1.3 to 24.4 mm). The final pathology revealed metastases in 59 lesions (90.8%), including lymph node (n=51), muscle (n=5), and in-transit (n=3) metastases. Sixty-one lesions (93.8%) were successfully detected intraoperatively and removed without residua on follow-up US. Four residual lesions were removed after repeated localization (n=2) or by intraoperative US (n=2). No relevant complications were noted. Preoperative US-guided tattooing localization can safely and effectively delineate nonpalpable metastatic melanoma lesions to aid in successful surgical excision.
1 January 2026 in Korean Journal of Radiology
To evaluate whether a custom agentic artificial intelligence (AI) pipeline can overcome the limitations of general-purpose large language model tools, when compared with a generic commercial tool (Google NotebookLM [NBLM]), for generating podcast-style summaries of radiology research articles. Twenty-two PDF-format original research articles published in the April 2025 issue of Radiology were processed using our Programmable, Phoneme-Aware PDF-to-Podcast Pipeline (P5) and NBLM to generate 44 audio episodes. P5 utilizes a multi-agent workflow for script generation, quality assurance, pronunciation enhancement, and audio synthesis. Four radiologists from a pool of 25 (7 generalists and 18 specialists) were randomly assigned to evaluate each blinded audio episode, yielding 176 total evaluations. The primary outcomes were the number of hallucinations (factual errors) per episode and the percentage of hallucination-free episodes. Secondary outcomes included the number of inappropriate statements, mispronunciations, and flow disruptions; the composite quality score (Quality Assessment of Educational Podcasts [QAEP]); the key results coverage score; and overall listener preference. Data were analyzed using generalized linear mixed models. The P5 method produced significantly fewer hallucinations per episode compared with NBLM (mean, 0.32 vs. 0.93; P < 0.001) and a higher proportion of hallucination-free episodes (71.6% [63/88] vs. 56.8% [50/88]; P = 0.013), consistently across generalists and specialists. P5 demonstrated significantly fewer mispronunciations (mean, 0.11 vs. 1.62; P < 0.001) and flow disruptions (mean, 0.19 vs. 1.06; P < 0.001) per episode, as well as higher mean QAEP composite scores (4.62 vs. 4.28; P < 0.001), compared with NBLM, consistently across generalists and specialists. Raters preferred P5 over NBLM in 72.7% (64/88) of comparisons (P = 0.003). Our custom agentic AI pipeline generated podcast-style summaries of radiology research articles with significantly higher quality and greater listener preference than the generic commercial tool.