Use of CEUS LI-RADS for the Accurate Diagnosis of Nodules in Patients at Risk for Hepatocellular Carcinoma: A Validation Study
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Purpose To validate the contrast agent–enhanced US Liver Imaging Reporting and Data System (CEUS LI-RADS) algorithm for accurate diagnosis of hepatocellular carcinoma (HCC) and categorization of all nodules encountered in patients at risk for HCC. Materials and Methods A single-center retrospective review of 196 nodules in 184 patients at risk for HCC (consisting of 139 HCCs, 18 non-HCC malignancies, and 39 benign nodules) was performed in a three-reader blinded read format, with the use of the CEUS LI-RADS algorithm. Pathologic confirmation was available for 143 nodules (122 HCCs, 18 non-HCC malignancies, and three benign nodules). Nodule sizes ranged between 1.0 and 16.2 cm. Nodules assessed with contrast-enhanced US were assigned various CEUS LI-RADS categories by three blinded readers. CEUS LI-RADS categorization was then compared against histopathologic findings, concurrent CT, and/or MR images or follow-up imaging to assess diagnostic accuracy of CEUS LI-RADS. In addition, the proportion of HCC in all LI-RADS (LR) categories, univariable and multivariable feature analysis, and interrater agreement using Light κ were determined. Results The LR-5 category, determined through radiologist categorization of nodules using the CEUS LI-RADS criteria, showed sensitivity, specificity, positive predictive value, and negative predictive value of 86% (119 of 139), 96% (55 of 57), 98% (119 of 121), and 73% (55 of 75), respectively, for the diagnosis of HCC. Two false-positive cases of LR-5 included a cholangiocarcinoma and a combined hepatocellular and cholangiocarcinoma. The remainder of the cholangiocarcinomas in the sample (n = 8) were appropriately categorized as LR-M. Multivariable logistic regression analysis showed that washout of greater than 60 seconds was the contrast-enhanced US feature most predictive of HCC diagnosis, whereas washout of less than 60 seconds was the feature most predictive of nonhepatocellular malignancy. The proportion of HCC nodules categorized in the LR-M and LR-4 categories was 35% and 20%, respectively. Light κ agreement between readers for LI-RADS categorization was 90%. Conclusion This study showed excellent specificity for the CEUS LI-RADS LR-5 category, allowing for confident imaging diagnosis of HCC, without necessity for pathologic confirmation. In addition, there was accurate differentiation of HCC from non-HCC malignancies and benign nodules. Only a single cholangiocarcinoma was misdiagnosed as category LR-5, with the remainder of the cholangiocarcinomas in the sample appropriately characterized as category LR-M. Keywords: Abdomen/GI, Evidence Based Medicine, Liver, Neoplasms-Primary, Ultrasound-Contrast © RSNA, 2020
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it