Purpose: To evaluate the effect of image registration on the diagnostic performance of transfer learning (TL) using pretrained convolutional neural networks (CNNs) and three-phasic dynamic contrast-enhanced computed tomography (DCE-CT) for primary liver cancers.

Methods: We retrospectively evaluated 215 consecutive patients with histologically proven primary liver cancers, including six early, 58 well-differentiated, 109 moderately differentiated, 29 poorly differentiated hepatocellular carcinomas (HCCs), and 13 non-HCC malignant lesions containing cholangiocellular components. We performed TL using various pretrained CNNs and preoperative three-phasic DCE-CT images. Three-phasic DCE-CT images were manually registered to correct respiratory motion. The registered DCE-CT images were then assigned to the three color channels of an input image for TL: pre-contrast, early phase, and delayed phase images for the blue, red, and green channels, respectively. To evaluate the effects of image registration, the registered input image was intentionally misaligned in the three color channels by pixel shifts, rotations, and skews with various degrees. The diagnostic performances (DP) of the pretrained CNNs after TL in the test set were compared by three general radiologists (GRs) and two experienced abdominal radiologists (ARs). The effects of misalignment in the input image and the type of pretrained CNN on the DP were statistically evaluated.

Results: The mean DPs for histological subtype classification and differentiation in primary malignant liver tumors on DCE-CT for GR and AR were 39.1%, and 47.9%, respectively. The highest mean DPs for CNNs after TL with pixel shifts, rotations, and skew misalignments were 44.1%, 44.2%, and 43.7%, respectively. Two-way analysis of variance revealed that the DP is significantly affected by the type of pretrained CNN (P = 0.0001), but not by misalignments in input images other than skew deformations.

Conclusion: TL using pretrained CNNs is robust against misregistration of multiphasic images and comparable to experienced ARs in classifying primary liver cancers using three-phasic DCE-CT.

Download full-text PDF

Source
http://dx.doi.org/10.1007/s11548-019-01987-1DOI Listing

Publication Analysis

Top Keywords

primary liver
16
liver cancers
12
pretrained cnns
12
three-phasic dce-ct
12
dce-ct images
12
input image
12
dynamic contrast-enhanced
8
contrast-enhanced computed
8
computed tomography
8
transfer learning
8

Similar Publications

Background: Liver fibrosis is caused by chronic toxic or cholestatic liver injury. Fibrosis results from the recruitment of myeloid cells into the injured liver, the release of inflammatory and fibrogenic cytokines, and the activation of myofibroblasts, which secrete extracellular matrix, mostly collagen type I. Hepatic myofibroblasts originate from liver-resident mesenchymal cells, including HSCs and bone marrow-derived CD45+ collagen type I+ expressing fibrocytes.

View Article and Find Full Text PDF

.

Expert Rev Gastroenterol Hepatol

January 2025

Division of Gastroenterology and Hepatology, Department of Internal Medicine, University of California San Diego, San Diego, CA.

Introduction: The gut microbiota has a complex relationship with the human host and is key to maintaining health. Disruption of the healthy diverse gut microbial milieu plays an important role in the pathogenesis of several diseases including infection (CDI), inflammatory bowel disease, irritable bowel syndrome, alcohol-related liver disease and metabolic-dysfunction associated steatotic liver disease (MASLD). Fecal microbiota transplantation (FMT) is highly effective in treating CDI, though its utility in other diseases is still being explored.

View Article and Find Full Text PDF

SARS-CoV-2 infection is accompanied by elevated liver enzymes, and patients with pre-existing liver conditions experience more severe disease. While it was known that SARS-CoV-2 infects human hepatocytes, our study determines the mechanism of infection, demonstrates viral replication and spread, and highlights direct hepatocyte damage. Viral replication was readily detectable upon infection of primary human hepatocytes and hepatoma cells with the ancestral SARS-CoV-2, Delta, and Omicron variants.

View Article and Find Full Text PDF

Aim: We previously reported that abdominal aortic calcification is associated with poor overall and recurrence-free survival after hepatectomy for hepatocellular carcinoma (HCC). However, the effect of abdominal aortic calcification on cancer-specific prognosis in very old patients with several comorbidities remains unknown. This multicenter study aimed to evaluate the impact of abdominal aortic calcification on the cumulative recurrence rate and recurrence-free survival in patients with HCC aged >80 years.

View Article and Find Full Text PDF

The growing burden of metabolic disorders manifested by hypertension, type 2 diabetes mellitus, hyperlipidemia, obesity and non-alcoholic fatty liver disease presents a significant global health challenge by contributing to cardiovascular diseases and high mortality rates. Β-blockers are among the most widely used drugs in the treatment of hypertension and acute cardiovascular events. In addition to blocking the receptor sites for catecholamines, third-generation β-blockers with associated vasodilating properties, such as carvedilol and nebivolol, provide a broad spectrum of metabolic effects, including anti-inflammatory and antioxidant properties and a favorable impact on glucose and lipid metabolism.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!