Journal of the Korean Society of Radiology最新文献

Large language models (LLMs) have revolutionized the global landscape of technology beyond the field of natural language processing. Owing to their extensive pre-training using vast datasets, contemporary LLMs can handle tasks ranging from general functionalities to domain-specific areas, such as radiology, without the need for additional fine-tuning. Importantly, LLMs are on a trajectory of rapid evolution, addressing challenges such as hallucination, bias in training data, high training costs, performance drift, and privacy issues, along with the inclusion of multimodal inputs. The concept of small, on-premise open source LLMs has garnered growing interest, as fine-tuning to medical domain knowledge, addressing efficiency and privacy issues, and managing performance drift can be effectively and simultaneously achieved. This review provides conceptual knowledge, actionable guidance, and an overview of the current technological landscape and future directions in LLMs for radiologists.

We present a case of iatrogenic iliac vein rupture with venous thrombosis during total hip arthroplasty that was endovascularly treated with the insertion of covered and uncovered stents, aspiration, and balloon maceration of the thrombus. We describe in detail, with a schematic diagram, an endovascular treatment method that can be performed under local anesthesia alone and explain how to overcome the situation when the diameter of the covered stent is inadequate to cover the iliac vein.

Pancreatic cancer is usually detected through contrast-enhanced CT and MRI. However, pancreatic cancer is occasionally overlooked because of its small size or is misdiagnosed as other conditions due to atypical imaging features that present diagnostic challenges. Considering the rapid growth and poor prognosis associated with pancreatic cancer, the ability to accurately detect and differentiate pancreatic lesions is crucial for appropriate surgical intervention. Reviewing diverse challenging cases of pancreatic cancer at an early stage and other mimicking lesions may help us accurately interpret the imaging features of pancreatic cancer on CT and MRI scans. Therefore, we aimed to illustrate various imaging features of pancreatic cancer and its mimicking lesions and provide valuable insights for differential diagnosis.

Various diseases can affect the esophagus. Endoscopic ultrasound (EUS), which provides precise information about the layers of the esophageal wall, is the primary approach used to investigate esophageal diseases. However, CT is one of the most important imaging modalities for diagnosing esophageal diseases as it can elucidate mediastinal involvement, adjacent lymphadenopathy, and distant disease spread. These two modalities complement each other in the diagnosis of esophageal diseases. Although radiologists may be unfamiliar with EUS procedures and their interpretation, understanding them aids in the differential diagnosis of esophageal conditions. This pictorial essay illustrates the EUS and CT findings of various esophageal diseases originating in the esophageal wall.

Immunoglobulin G4-related disease (IgG4-RD) is an immune-mediated condition characterized by mass-forming inflammation with a sclerosing pattern that can affect nearly any organ. However, involvement of the diaphragm in IgG4-RD is exceptionally rare. We present the case of a 62-year-old male patient with chest radiographic abnormalities. Further investigation with CT revealed an infiltrative mass in the right hemidiaphragm. This mass, composed of engorged feeding vessels, an atypical manifestation of IgG4-RD, was also associated with lymphadenopathy. Surgical excision confirmed the presence of IgG4-positive cell infiltration, solidifying the diagnosis of IgG4-RD. Notably, the patient remained asymptomatic and did not require any treatment postoperatively. This case highlights the uncommon presentation of IgG4-RD as an infiltrative diaphragmatic mass.

The recent advent of large language models (LLMs), such as ChatGPT, has drawn attention to generative artificial intelligence (AI) in a number of fields. Generative AI can produce different types of data including text, images, and voice, depending on the training methods and datasets used. Additionally, recent advancements in multimodal techniques, which can simultaneously process multiple data types like text and images, have expanded the potential of using multimodal generative AI in the medical environment where various types of clinical and imaging information are used together. This review summarizes the concepts and types of LLMs, image generative AI, and multimodal AI, and it examines the status and future possibilities of generative AI in the field of radiology.

Artificial intelligence (AI) is transforming radiology with improved diagnostic accuracy and efficiency, but prediction uncertainty remains a critical challenge. This review examines key sources of uncertainty-out-of-distribution, aleatoric, and model uncertainties-and highlights the importance of independent confidence metrics and explainable AI for safe integration. Independent confidence metrics assess the reliability of AI predictions, while explainable AI provides transparency, enhancing collaboration between AI and radiologists. The development of zero-error tolerance models, designed to minimize errors, sets new standards for safety. Addressing these challenges will enable AI to become a trusted partner in radiology, advancing care standards and patient outcomes.

Necrotizing sarcoid granulomatosis (NSG), now referred to as "sarcoidosis with NSG pattern," is an uncommon variant of sarcoidosis. NSG is characterized by a trio of features: sarcoid granulomas, vasculitis, and extensive areas of necrosis. Symptoms can include cough, fever, chest pain, and dyspnea, typically presenting as either solitary or multiple lung nodules or masses. In this report, we describe a case of NSG accompanied by a persistent low-grade fever. Unlike the dominant presentation of NSG with single or multiple nodules, our case demonstrated diffuse micronodules with combined perilymphatic and random distribution on CT. Histological examination revealed widespread necrotizing granulomas surrounded by anthracotic pigmentation, alongside necrosis and vasculitis, diverging from the classic presentation of sarcoidosis. The diagnosis of NSG was established through a multidisciplinary discussion. The patient was administered oral prednisolone that led to noticeable clinical and radiological improvement within three months.

Epithelioid hemangioma (EH) is an uncommon, benign vascular tumor of mesenchymal origin. It mainly presents as a tumor with a lytic appearance and septations. However, no case reports have documented the predominantly sclerotic appearance of EH. Here, we present the imaging findings, including X-ray, CT, MRI, and histopathological findings of a 24-year-old female with EH.