Large language models for structured reporting in radiology: performance of GPT-4, ChatGPT-3.5, Perplexity and Bing

Hepatic encephalopathy post-TIPS: Current status and prospects in predictive assessment

Transjugular intrahepatic portosystemic shunt (TIPS) is an essential procedure for the treatment of portal hypertension but can result in hepatic encephalopathy (HE), a serious complication that worsens patient outcomes. Investigating predictors of HE after TIPS is essential to improve prognosis. This review analyzes risk factors and compares predictive models, weighing traditional scores such as Child-Pugh, Model for End-Stage Liver Disease (MELD), and albumin-bilirubin (ALBI) against emerging artificial intelligence (AI) techniques. While traditional scores provide initial insights into HE risk, they have limitations in dealing with clinical complexity. Advances in machine learning (ML), particularly when integrated with imaging and clinical data, offer refined assessments. These innovations suggest the potential for AI to significantly improve the prediction of post-TIPS HE. The study provides clinicians with a comprehensive overview of current prediction methods, while advocating for the integration of AI to increase the accuracy of post-TIPS HE assessments. By harnessing the power of AI, clinicians can better manage the risks associated with TIPS and tailor interventions to individual patient needs. Future research should therefore prioritize the development of advanced AI frameworks that can assimilate diverse data streams to support clinical decision-making. The goal is not only to more accurately predict HE, but also to improve overall patient care and quality of life.

ChatGPT and assistive AI in structured radiology reporting: A systematic review

INTRODUCTION

The rise of transformer-based large language models (LLMs), such as ChatGPT, has captured global attention with recent advancements in artificial intelligence (AI). ChatGPT demonstrates growing potential in structured radiology reporting-a field where AI has traditionally focused on image analysis.

METHODS

A comprehensive search of MEDLINE and Embase was conducted from inception through May 2024, and primary studies discussing ChatGPT's role in structured radiology reporting were selected based on their content.

RESULTS

Of the 268 articles screened, eight were ultimately included in this review. These articles explored various applications of ChatGPT, such as generating structured reports from unstructured reports, extracting data from free text, generating impressions from radiology findings and creating structured reports from imaging data. All studies demonstrated optimism regarding ChatGPT's potential to aid radiologists, though common critiques included data privacy concerns, reliability, medical errors, and lack of medical-specific training.

CONCLUSION

ChatGPT and assistive AI have significant potential to transform radiology reporting, enhancing accuracy and standardization while optimizing healthcare resources. Future developments may involve integrating dynamic few-shot prompting, ChatGPT, and Retrieval Augmented Generation (RAG) into diagnostic workflows. Continued research, development, and ethical oversight are crucial to fully realize AI's potential in radiology.

ChatGPT in radiology: A systematic review of performance, pitfalls, and future perspectives

PURPOSE

The purpose of this study was to systematically review the reported performances of ChatGPT, identify potential limitations, and explore future directions for its integration, optimization, and ethical considerations in radiology applications.

MATERIALS AND METHODS

After a comprehensive review of PubMed, Web of Science, Embase, and Google Scholar databases, a cohort of published studies was identified up to January 1, 2024, utilizing ChatGPT for clinical radiology applications.

RESULTS

Out of 861 studies derived, 44 studies evaluated the performance of ChatGPT; among these, 37 (37/44; 84.1%) demonstrated high performance, and seven (7/44; 15.9%) indicated it had a lower performance in providing information on diagnosis and clinical decision support (6/44; 13.6%) and patient communication and educational content (1/44; 2.3%). Twenty-four (24/44; 54.5%) studies reported the proportion of ChatGPT's performance. Among these, 19 (19/24; 79.2%) studies recorded a median accuracy of 70.5%, and in five (5/24; 20.8%) studies, there was a median agreement of 83.6% between ChatGPT outcomes and reference standards [radiologists' decision or guidelines], generally confirming ChatGPT's high accuracy in these studies. Eleven studies compared two recent ChatGPT versions, and in ten (10/11; 90.9%), ChatGPTv4 outperformed v3.5, showing notable enhancements in addressing higher-order thinking questions, better comprehension of radiology terms, and improved accuracy in describing images. Risks and concerns about using ChatGPT included biased responses, limited originality, and the potential for inaccurate information leading to misinformation, hallucinations, improper citations and fake references, cybersecurity vulnerabilities, and patient privacy risks.

CONCLUSION

Although ChatGPT's effectiveness has been shown in 84.1% of radiology studies, there are still multiple pitfalls and limitations to address. It is too soon to confirm its complete proficiency and accuracy, and more extensive multicenter studies utilizing diverse datasets and pre-training techniques are required to verify ChatGPT's role in radiology.

An evaluation of the capabilities of language models and nurses in providing neonatal clinical decision support

AIM

To assess the clinical reasoning capabilities of two large language models, ChatGPT-4 and Claude-2.0, compared to those of neonatal nurses during neonatal care scenarios.

DESIGN

A cross-sectional study with a comparative evaluation using a survey instrument that included six neonatal intensive care unit clinical scenarios.

PARTICIPANTS

32 neonatal intensive care nurses with 5-10 years of experience working in the neonatal intensive care units of three medical centers.

METHODS

Participants responded to 6 written clinical scenarios. Simultaneously, we asked ChatGPT-4 and Claude-2.0 to provide initial assessments and treatment recommendations for the same scenarios. The responses from ChatGPT-4 and Claude-2.0 were then scored by certified neonatal nurse practitioners for accuracy, completeness, and response time.

RESULTS

Both models demonstrated capabilities in clinical reasoning for neonatal care, with Claude-2.0 significantly outperforming ChatGPT-4 in clinical accuracy and speed. However, limitations were identified across the cases in diagnostic precision, treatment specificity, and response lag.

CONCLUSIONS

While showing promise, current limitations reinforce the need for deep refinement before ChatGPT-4 and Claude-2.0 can be considered for integration into clinical practice. Additional validation of these tools is important to safely leverage this Artificial Intelligence technology for enhancing clinical decision-making.

IMPACT

The study provides an understanding of the reasoning accuracy of new Artificial Intelligence models in neonatal clinical care. The current accuracy gaps of ChatGPT-4 and Claude-2.0 need to be addressed prior to clinical usage.

Large Language Models for Automated Synoptic Reports and Resectability Categorization in Pancreatic Cancer

Background Structured radiology reports for pancreatic ductal adenocarcinoma (PDAC) improve surgical decision-making over free-text reports, but radiologist adoption is variable. Resectability criteria are applied inconsistently. Purpose To evaluate the performance of large language models (LLMs) in automatically creating PDAC synoptic reports from original reports and to explore performance in categorizing tumor resectability. Materials and Methods In this institutional review board-approved retrospective study, 180 consecutive PDAC staging CT reports on patients referred to the authors' European Society for Medical Oncology-designated cancer center from January to December 2018 were included. Reports were reviewed by two radiologists to establish the reference standard for 14 key findings and National Comprehensive Cancer Network (NCCN) resectability category. GPT-3.5 and GPT-4 (accessed September 18-29, 2023) were prompted to create synoptic reports from original reports with the same 14 features, and their performance was evaluated (recall, precision, F1 score). To categorize resectability, three prompting strategies (default knowledge, in-context knowledge, chain-of-thought) were used for both LLMs. Hepatopancreaticobiliary surgeons reviewed original and artificial intelligence (AI)-generated reports to determine resectability, with accuracy and review time compared. The McNemar test, t test, Wilcoxon signed-rank test, and mixed effects logistic regression models were used where appropriate. Results GPT-4 outperformed GPT-3.5 in the creation of synoptic reports (F1 score: 0.997 vs 0.967, respectively). Compared with GPT-3.5, GPT-4 achieved equal or higher F1 scores for all 14 extracted features. GPT-4 had higher precision than GPT-3.5 for extracting superior mesenteric artery involvement (100% vs 88.8%, respectively). For categorizing resectability, GPT-4 outperformed GPT-3.5 for each prompting strategy. For GPT-4, chain-of-thought prompting was most accurate, outperforming in-context knowledge prompting (92% vs 83%, respectively; P = .002), which outperformed the default knowledge strategy (83% vs 67%, P < .001). Surgeons were more accurate in categorizing resectability using AI-generated reports than original reports (83% vs 76%, respectively; P = .03), while spending less time on each report (58%; 95% CI: 0.53, 0.62). Conclusion GPT-4 created near-perfect PDAC synoptic reports from original reports. GPT-4 with chain-of-thought achieved high accuracy in categorizing resectability. Surgeons were more accurate and efficient using AI-generated reports. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Chang in this issue.

Comparing the Diagnostic Performance of GPT-4-based ChatGPT, GPT-4V-based ChatGPT, and Radiologists in Challenging Neuroradiology Cases

PURPOSE

To compare the diagnostic performance among Generative Pre-trained Transformer (GPT)-4-based ChatGPT, GPT‑4 with vision (GPT-4V) based ChatGPT, and radiologists in challenging neuroradiology cases.

METHODS

We collected 32 consecutive "Freiburg Neuropathology Case Conference" cases from the journal Clinical Neuroradiology between March 2016 and December 2023. We input the medical history and imaging findings into GPT-4-based ChatGPT and the medical history and images into GPT-4V-based ChatGPT, then both generated a diagnosis for each case. Six radiologists (three radiology residents and three board-certified radiologists) independently reviewed all cases and provided diagnoses. ChatGPT and radiologists' diagnostic accuracy rates were evaluated based on the published ground truth. Chi-square tests were performed to compare the diagnostic accuracy of GPT-4-based ChatGPT, GPT-4V-based ChatGPT, and radiologists.

RESULTS

GPT‑4 and GPT-4V-based ChatGPTs achieved accuracy rates of 22% (7/32) and 16% (5/32), respectively. Radiologists achieved the following accuracy rates: three radiology residents 28% (9/32), 31% (10/32), and 28% (9/32); and three board-certified radiologists 38% (12/32), 47% (15/32), and 44% (14/32). GPT-4-based ChatGPT's diagnostic accuracy was lower than each radiologist, although not significantly (all p > 0.07). GPT-4V-based ChatGPT's diagnostic accuracy was also lower than each radiologist and significantly lower than two board-certified radiologists (p = 0.02 and 0.03) (not significant for radiology residents and one board-certified radiologist [all p > 0.09]).

CONCLUSION

While GPT-4-based ChatGPT demonstrated relatively higher diagnostic performance than GPT-4V-based ChatGPT, the diagnostic performance of GPT‑4 and GPT-4V-based ChatGPTs did not reach the performance level of either radiology residents or board-certified radiologists in challenging neuroradiology cases.

BI-RADS Category Assignments by GPT-3.5, GPT-4, and Google Bard: A Multilanguage Study

Background The performance of publicly available large language models (LLMs) remains unclear for complex clinical tasks. Purpose To evaluate the agreement between human readers and LLMs for Breast Imaging Reporting and Data System (BI-RADS) categories assigned based on breast imaging reports written in three languages and to assess the impact of discordant category assignments on clinical management. Materials and Methods This retrospective study included reports for women who underwent MRI, mammography, and/or US for breast cancer screening or diagnostic purposes at three referral centers. Reports with findings categorized as BI-RADS 1-5 and written in Italian, English, or Dutch were collected between January 2000 and October 2023. Board-certified breast radiologists and the LLMs GPT-3.5 and GPT-4 (OpenAI) and Bard, now called Gemini (Google), assigned BI-RADS categories using only the findings described by the original radiologists. Agreement between human readers and LLMs for BI-RADS categories was assessed using the Gwet agreement coefficient (AC1 value). Frequencies were calculated for changes in BI-RADS category assignments that would affect clinical management (ie, BI-RADS 0 vs BI-RADS 1 or 2 vs BI-RADS 3 vs BI-RADS 4 or 5) and compared using the McNemar test. Results Across 2400 reports, agreement between the original and reviewing radiologists was almost perfect (AC1 = 0.91), while agreement between the original radiologists and GPT-4, GPT-3.5, and Bard was moderate (AC1 = 0.52, 0.48, and 0.42, respectively). Across human readers and LLMs, differences were observed in the frequency of BI-RADS category upgrades or downgrades that would result in changed clinical management (118 of 2400 [4.9%] for human readers, 611 of 2400 [25.5%] for Bard, 573 of 2400 [23.9%] for GPT-3.5, and 435 of 2400 [18.1%] for GPT-4; P < .001) and that would negatively impact clinical management (37 of 2400 [1.5%] for human readers, 435 of 2400 [18.1%] for Bard, 344 of 2400 [14.3%] for GPT-3.5, and 255 of 2400 [10.6%] for GPT-4; P < .001). Conclusion LLMs achieved moderate agreement with human reader-assigned BI-RADS categories across reports written in three languages but also yielded a high percentage of discordant BI-RADS categories that would negatively impact clinical management. © RSNA, 2024 Supplemental material is available for this article.

Feasibility of GPT-3 and GPT-4 for in-Depth Patient Education Prior to Interventional Radiological Procedures: A Comparative Analysis

PURPOSE

This study explores the utility of the large language models, GPT-3 and GPT-4, for in-depth patient education prior to interventional radiology procedures. Further, differences in answer accuracy between the models were assessed.

MATERIALS AND METHODS

A total of 133 questions related to three specific interventional radiology procedures (Port implantation, PTA and TACE) covering general information as well as preparation details, risks and complications and post procedural aftercare were compiled. Responses of GPT-3 and GPT-4 were assessed for their accuracy by two board-certified radiologists using a 5-point Likert scale. The performance difference between GPT-3 and GPT-4 was analyzed.

RESULTS

Both GPT-3 and GPT-4 responded with (5) "completely correct" (4) "very good" answers for the majority of questions ((5) 30.8% + (4) 48.1% for GPT-3 and (5) 35.3% + (4) 47.4% for GPT-4). GPT-3 and GPT-4 provided (3) "acceptable" responses 15.8% and 15.0% of the time, respectively. GPT-3 provided (2) "mostly incorrect" responses in 5.3% of instances, while GPT-4 had a lower rate of such occurrences, at just 2.3%. No response was identified as potentially harmful. GPT-4 was found to give significantly more accurate responses than GPT-3 (p = 0.043).

CONCLUSION

GPT-3 and GPT-4 emerge as relatively safe and accurate tools for patient education in interventional radiology. GPT-4 showed a slightly better performance. The feasibility and accuracy of these models suggest their promising role in revolutionizing patient care. Still, users need to be aware of possible limitations.