Liver Imaging Reporting Data System Research Articles

Evaluating Large Language Models for Automated Reporting and Data Systems Categorization: Cross-Sectional Study.

Large language models show promise for improving radiology workflows, but their performance on structured radiological tasks such as Reporting and Data Systems (RADS) categorization remains unexplored. This study aims to evaluate 3 large language model chatbots-Claude-2, GPT-3.5, and GPT-4-on assigning RADS categories to radiology reports and assess the impact of different prompting strategies. This cross-sectional study compared 3 chatbots using 30 radiology reports (10 per RADS criteria), using a 3-level prompting strategy: zero-shot, few-shot, and guideline PDF-informed prompts. The cases were grounded in Liver Imaging Reporting & Data System (LI-RADS) version 2018, Lung CT (computed tomography) Screening Reporting & Data System (Lung-RADS) version 2022, and Ovarian-Adnexal Reporting & Data System (O-RADS) magnetic resonance imaging, meticulously prepared by board-certified radiologists. Each report underwent 6 assessments. Two blinded reviewers assessed the chatbots' response at patient-level RADS categorization and overall ratings. The agreement across repetitions was assessed using Fleiss κ. Claude-2 achieved the highest accuracy in overall ratings with few-shot prompts and guideline PDFs (prompt-2), attaining 57% (17/30) average accuracy over 6 runs and 50% (15/30) accuracy with k-pass voting. Without prompt engineering, all chatbots performed poorly. The introduction of a structured exemplar prompt (prompt-1) increased the accuracy of overall ratings for all chatbots. Providing prompt-2 further improved Claude-2's performance, an enhancement not replicated by GPT-4. The interrun agreement was substantial for Claude-2 (k=0.66 for overall rating and k=0.69 for RADS categorization), fair for GPT-4 (k=0.39 for both), and fair for GPT-3.5 (k=0.21 for overall rating and k=0.39 for RADS categorization). All chatbots showed significantly higher accuracy with LI-RADS version 2018 than with Lung-RADS version 2022 and O-RADS (P<.05); with prompt-2, Claude-2 achieved the highest overall rating accuracy of 75% (45/60) in LI-RADS version 2018. When equipped with structured prompts and guideline PDFs, Claude-2 demonstrated potential in assigning RADS categories to radiology cases according to established criteria such as LI-RADS version 2018. However, the current generation of chatbots lags in accurately categorizing cases based on more recent RADS criteria.

Contrast-enhanced ultrasound (CEUS) in HCC diagnosis and assessment of tumor response to locoregional therapies.

Hepatocellular carcinoma (HCC) is a global problem constituting the second leading cause of cancer deaths worldwide, thereby necessitating an accurate and cost-effective solution for managing care. Ultrasound is well poised to address this need due to its low cost, portability, safety, and excellent temporal resolution. The role of ultrasound for HCC screening has been well established and supported by multiple international guidelines. Similarly, contrast-enhanced ultrasound (CEUS) can be used for the characterization of focal liver lesions in high-risk populations, and standardized criteria for CEUS have been established by the American College of Radiology Liver Imaging Reporting & Data System (LI-RADS). Following HCC identification, CEUS can also be highly beneficial in treatment planning, delivery, and monitoring HCC response to locoregional therapies. Specific advantages of CEUS include providing real-time treatment guidance and improved diagnostic performance for the detection of residual tumor viability or recurrence, thereby identifying patients in need of retreatment substantially earlier than contrast-enhanced CT and MRI. This review provides a primer on ultrasound and CEUS for the screening and characterization of HCC, with an emphasis on assessing tumor response to locoregional therapies.

LI-RADS 4 or 5 categorization may not be clinically relevant for decision-making processes: A prospective cohort study

Introduction and objectivesThe liver imaging reporting data system (LI-RADS) for hepatocellular carcinoma (HCC) was proposed to standardize and enhance consensus of reporting. However, clinical utility of LI-RADS has not been evaluated in Latin America. We therefore sought to compare LI-RADS categories with histopathology findings in liver transplant (LT) explants in a regional center. Materials and methodsProspective cohort study conducted between 2012 and 2018 in a single center from Argentina including patients with HCC listed for LT. LI-RADS definitions were applied to magnetic resonance images (MRI) or computed tomography (CT) abdominal scans at time of listing and at final pre-LT reassessment and compared to explant pathology findings; specifically, major nodule (NOD1). ResultsOf 130 patients with HCC listed for LT (96.1% with cirrhosis and 35.6% with hepatitis C virus infection), 72 underwent LT. Overall, 65% had imaging HCC diagnosis based on MRI (n = 84), 26% with CT (n = 34) and 9% (n = 12) with both methods. Among LT patients with pre-transplant imaging at our institution (n = 42/72), 69% of the NOD1 were LR-5, 21% LR-4 and 10% LR-3. Definite HCC diagnosis was 50% in LR-3 NOD1 (CI 18–90); none presented microvascular invasion. In LR-4 NOD1, HCC was confirmed in 89% (CI 59–98), of which 11% showed microvascular invasion; whereas in LR-5 NOD1 77% (CI 64–87) had confirmed HCC, 17% with microvascular invasion. ConclusionsLI-RADS was useful to standardize reports; however, no significant differences were observed between LR-4 and LR-5 HCC probability when compared to explant pathology.

Magnetic Resonance Imaging Evaluation of Hepatocellular Carcinoma Treated With Stereotactic Body Radiation Therapy: Long Term Imaging Follow-Up

To determine the natural history of imaging findings seen on magnetic resonance imaging (MRI) of hepatocellular carcinoma (HCC) treated with stereotactic body radiation therapy (SBRT). Although arterial hyperenhancement is a key feature of untreated HCC, our clinical experience suggested that tumors that never progressed could still show hyperenhancement. Therefore, we undertook a systematic study to test the hypothesis that persistent arterial phase hyperenhancement (APHE) after SBRT is an expected finding that does not suggest failure of treatment. One hundred forty-six patients undergoing SBRT for HCC between January 1, 2007, and December 31, 2015, were screened retrospectively using an institutional review board-approved prospectively maintained registry. Inclusion criteria were (1) HCC treated with SBRT, (2) multiphasic MRI ≤3months before SBRT, (3) up to 1year of follow-up MRI post-SBRT, and (4) cirrhosis. The exclusion criterion was ≤3months of locoregional therapy to the liver segment containing the SBRT-treated HCC. Pre- and post-SBRT MRI from up to 3 years were analyzed in consensus by independent pairs of subspecialty-trained radiologists to determine the temporal evolution of major features for HCC and imaging findings in off-target parenchyma. Sixty-two patients with 67 HCCs (Organ Procurement and Transplantation Network imaging criteria [OPTN] 5a [n=26], OPTN 5b [n=28], OPTN 5x [n=7]; Liver Imaging Reporting Data System [LI-RAD]-M [n=4] and LiRADs-4 [n=2]) were studied. Tumor size either decreased (66% [44 of 67]) or remained unchanged (34% [23 of 67]) within the first 12months. Post-SBRT APHE was common (58% [39 of 67]). When graded using modified Response Evaluation Criteria in Solid Tumors at 3 to 6months, 25% (17 of 67) met criteria for complete response and 75% (50 of 67) met criteria for stable disease. SBRT is an effective locoregional treatment option for HCC. Persistent APHE is common and does not necessarily indicate viable neoplasm; thus, standard response assessment such as modified Response Evaluation Criteria should be used with caution, particularly in the early phases after SBRT therapy.