Current problems in diagnostic radiology最新文献

The diagnostic algorithm for histoplasmosis highlights the importance of imaging and emphasizes the role of the radiologist in the diagnostic workup. Here we describe a case series of patients with a novel sign of lung involvement in histoplasmosis which we have coined the Pseudo-Pulmonary Arteriovenous Malformation (PAVM) sign, the usage of which would help in the imaging diagnosis of histoplasmosis aid by distinguishing it from PAVMs. PAVMs carry risk for serious complications such as systemic emboli and may require treatment; whereas, histoplasmomas do not. Differentiation of histoplasmosis from other diagnoses can be made with laboratory studies, but may require bronchoscopy, biopsy, or both. Meanwhile, PAVMs should not be biopsied due to risk of bleeding. For these reasons, distinguishing PAVMs and histoplasmosis radiologically therefore greatly impacts clinical management, and it is important for radiologists to be aware of this appearance of histoplasmosis to avoid misinterpretation as PAVM and effectively inform clinical care.

Gallbladder pathologies caused by gallstones are commonly encountered in clinical practice, making accurate diagnosis critical for effective patient management. Radiologists play a key role in differentiating these conditions through imaging interpretation, ensuring that appropriate treatment is initiated. The imaging features of gallstone associated diseases are classified into various categories, such as inflammatory conditions, benign lesions, malignant tumors, and associated complications. A comprehensive understanding of these categories and their radiologic manifestations is essential for accurate diagnosis and management of gallbladder pathology. By integrating clinical knowledge with radiologic findings, clinicians and radiologists will be equipped with practical tools to identify and distinguish between different gallstone causing conditions.

In academic and research settings, computer-aided nodule detection software has been shown to increase accuracy, efficiency, and throughput. However, radiologists need to be familiar with the spectrum of errors that can occur when these algorithms are employed in routine clinical settings. We review the spectrum of errors that may result from computer-aided nodule detection. In our clinical practice, we have seen errors in nodule detection, nodule localization, and nodule characterization. Each of these categories are demonstrated with illustrative cases. Through these illustrative cases, readers can be more familiar with nuances and pitfalls generated by computer-aided detection software. Although computer-aided nodule detection software is rapidly advancing, radiologists still need to thoroughly review images with mindfulness of some of the errors that can be generated by AI platforms for nodule detection.

The need for emergent, contrast-enhanced neuroimaging in stroke patients with a history of severe reaction to iodinated contrast represents a unique dilemma in emergency departments. There is currently a lack of evidence-based management protocols for these cases. We describe a protocol established at our institution, based off American College of Radiology (ACR) guidelines and institutional experience, to guide decision-making in these scenarios.

Unless radiologist performance assessment is sufficiently deep, comprehensive, and balanced, it may tend to omit, obscure, or distort key aspects of the important contributions that radiologists make, with adverse consequences for employers, radiologists themselves, and above all, the patients they serve. Here we present a model of performance assessment that includes eight key dimensions, which can be tailored as appropriate to the needs of particular programs and radiologists.

Arachnoid granulations have been known for centuries yet remain incompletely understood. While traditionally associated with cerebrospinal fluid transport, the precise mechanism remains uncertain. This manuscript reviews the literature on the anatomy, histology, and imaging findings of arachnoid granulations and their mimickers and anomalous variations. We highlight variations in incidence, size, and characteristics of arachnoid granulations on imaging, and hypothesize that these variations may be explained by arachnoid granulations being dynamic secondary to varying functionality. We review the pathophysiologic role of arachnoid granulations in pathologies related to hydrocephalus, neurodegenerative disorders, and intracranial hypertension and hypotension. A further understanding of arachnoid granulations, their mechanism in cerebrospinal fluid transport, and change over time may provide a basis for future imaging markers and therapies.

Background: Radiologists increasingly use artificial intelligence (AI) to enhance diagnostic accuracy and optimize workflows. However, many lack the technical skills to effectively apply machine learning (ML) and deep learning (DL) algorithms, limiting the accessibility of these methods to radiology researchers who could otherwise benefit from them. Large language models (LLMs), such as GPT-4o, may serve as virtual advisors, offering tailored algorithm recommendations for specific research needs. This study evaluates GPT-4o's effectiveness as a recommender system to enhance radiologists' understanding and implementation of AI in research.

Intervention: GPT-4o was used to recommend ML and DL algorithms based on specific details provided by researchers, including dataset characteristics, modality types, data sizes, and research objectives. The model acted as a virtual advisor, guiding researchers in selecting the most appropriate models for their studies.

Methods: The study systematically evaluated GPT-4o's recommendations for clarity, task alignment, model diversity, and baseline selection. Responses were graded to assess the model's ability to meet the needs of radiology researchers.

Results: GPT-4o effectively recommended appropriate ML and DL algorithms for various radiology tasks, including segmentation, classification, and regression in medical imaging. The model suggested a diverse range of established and innovative algorithms, such as U-Net, Random Forest, Attention U-Net, and EfficientNet, aligning well with accepted practices.

Conclusion: GPT-4o shows promise as a valuable tool for radiologists and early career researchers by providing clear and relevant AI and ML algorithm recommendations. Its ability to bridge the knowledge gap in AI implementation could democratize access to advanced technologies, fostering innovation and improving radiology research quality. Further studies should explore integrating LLMs into routine workflows and their role in ongoing professional development.

Background: As a result of the 21st Century Cures Act, radiology reports are immediately released to patients. However, these reports are often too complex for the lay patient, potentially leading to stress and anxiety. While solutions such as patient portals or providing radiologist contact information have been proposed in the past, new generative artificial intelligence technologies like ChatGPT and Google Gemini may provide the most accessible and scalable method of simplifying radiology reports for patients. Here, we gather the opinions of radiologists regarding this possibility.

Methods: An eight-question survey was sent out to all diagnostic/interventional radiology attendings and clinical fellows at our large academic medical center.

Results: From our survey (N = 52), 52.8 % of respondents agreed/strongly agreed that patients should have immediate access to their radiology reports. Only 9.61 % agreed that radiology reports are understandable by the lay patient. Regarding potential avenues to improve patient comprehension of their radiology reports, using artificial intelligence to simplify reports with a manual check by radiologists garnered the most support/strong support (46.2 %). Support of artificial intelligence generated simplifications dropped to (23.1 %) without a manual check.

Conclusion: Patients are increasingly gaining access to their radiology reports, but reports may be too complex for the lay patient. Eventually, artificial intelligence systems may help simplify radiology reports for patients, but there is currently limited support from radiologists.

Purpose: This study aimed to evaluate the utility of tumor mutational burden (TMB) as a marker for radiologic response to immune checkpoint inhibitor (ICI) therapy at a single tertiary cancer center.

Materials and methods: In this retrospective study, out of 1044 patients treated with ICIs between January 2010 and November 2018, 75 patients (38 males and 37 females) with a mean age of 62 (range 22-87) years, who had information about TMB and adequate imaging, were included. Imaging response was determined according to iRECIST criteria. Predictors of objective response were analysed using non-parametric tests, and progression-free survival and overall survival were analysed using log-rank test.

Results: Median TMB was 7.2 mutations/mb [interquartile range: 4-13.5]. The objective radiologic response rate according to iRECIST was 26.7 % (20 patients) and the median time to best response was 61 days [IQR: 47-88 days]. Median TMB in responders (12.5 [IQR: 5-18] muts/mb) was significantly higher than in non-responders (6 [IQR: 3-12] muts/mb) (p = 0.0293). Median TMB was higher in responders in the subgroup of patients treated with Nivolumab (20 vs 4 muts/mb, P = .0043), but not significantly in those treated with Pembrolizumab (9 vs 6 muts/mb, P = .211). There was no difference in PFS (p = 0.37, Log-Rank) or OS (p = 0.053, Log-Rank) between TMB low and high groups.

Conclusion: Higher TMB was associated with objective response to ICI, however, TMB was an imperfect biomarker for PFS and OS in our study.

Interpreting imaging examinations of the pancreas can be a challenge. Several different entities can mimic or mask pancreatic neoplasms, including normal anatomic variants, non-pancreatic lesions, and both acute and chronic pancreatitis. It is important to distinguish these entities from pancreatic neoplasms, as the management and prognosis of a pancreatic neoplasm, particularly adenocarcinoma, have considerable impact on patients. Normal pancreatic variants that mimic a focal lesion include focal fatty atrophy, annular pancreas, and ectopic pancreas. Extra-pancreatic lesions that can mimic a primary pancreatic neoplasm include vascular lesions, such as arteriovenous malformations and pseudoaneurysms, duodenal diverticula, and intra-pancreatic accessory spleen. Both acute and chronic pancreatitis can mimic or mask a pancreatic neoplasm and are also associated with pancreatic ductal adenocarcinoma. Awareness of these entities and their imaging features will enable the radiologist to narrow the differential diagnosis, provide recommendations that expedite diagnosis and avoid unnecessary work-up or delays in patient care.