Comparative analysis of imaging diagnostic models for tubular basophilia and mineralization of kidney.

Lab Anim Res

Department of Pharmaceutical Engineering, College of Natural Science, Hoseo University, Beabang-eup Hoseo-ro 79-20, Asan-si, 31499, Chungcheongnam-do, Korea.

Published: September 2022

Background: Now that it is possible to efficiently classify and save tissue images of laboratory animals using whole-slide imaging, many diagnostic models are being developed through transfer learning with Convolutional Neural Network (CNN). In this study, transfer learning was performed to gain toxicopathological knowledge using CNN models such as InceptionV3 and Xception. For the classification of tubular basophilia and mineralization, two representative background lesions that commonly occur in toxicological studies, accuracies of diagnosis were compared using MobileNetV2, Xception and InceptionV3. For the simultaneous detection of the two lesions, the accuracy was analysed using You Only Look Once version 4 (YOLOv4).

Results: The accuracy of the classification models was as follows: MobileNetV2 (epoch 50, accuracy: 98.57%) > Xception (epoch 70, accuracy: 97.47%) > InceptionV3 (epoch 70, accuracy: 89.62%). In the case of object detection, the accuracy of YOLOv4 was 98.62% at epoch 3000.

Conclusions: Among the classification models, MobileNetV2 had the best accuracy despite applying a lower epoch than InceptionV3 and Xception. The object detection model, YOLOv4, accurately and simultaneously diagnosed tubular basophilia and mineralization, with an accuracy of 98.62% at epoch 3000.

Download full-text PDF

Source
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC9476251PMC
http://dx.doi.org/10.1186/s42826-022-00139-yDOI Listing

Publication Analysis

Top Keywords

tubular basophilia
12
basophilia mineralization
12
epoch accuracy
12
imaging diagnostic
8
diagnostic models
8
transfer learning
8
inceptionv3 xception
8
accuracy
8
classification models
8
models mobilenetv2
8

Similar Publications

With the development of artificial intelligence (AI), technologies based on machines and deep learning are being used in many academic fields. In toxicopathology, research is actively underway to analyze whole slide image (WSI)-level images using AI deep-learning models. However, few studies have been conducted on models for diagnosing complex lesions comprising multiple lesions.

View Article and Find Full Text PDF

N-methylformamide induces multiple organ toxicity in Fischer 344 rats.

Toxicol Res

April 2023

Department of Veterinary Pathology, College of Veterinary Medicine, Chungbuk National University, Chungdae-ro 1, Seowon-gu, Cheongju, Chungbuk 28644 Republic of Korea.

N-Methylformamide (NMF) is a widely used chemical (CAS No.: 123-39-7) in several industries and its usage is continuously increasing. However, studies for NMF have been focused on hepatotoxicity from now.

View Article and Find Full Text PDF

Multiple animal models have been developed to investigate the pathogenesis of colorectal cancer and to evaluate potential treatments. One model system uses azoxymethane, a metabolite of cycasin, alone and in conjunction with dextran sodium sulfate to induce colon cancer in rodents. Azoxymethane is metabolized by hepatic P450 enzymes and can also be eliminated through the kidneys.

View Article and Find Full Text PDF

Comparative analysis of imaging diagnostic models for tubular basophilia and mineralization of kidney.

Lab Anim Res

September 2022

Department of Pharmaceutical Engineering, College of Natural Science, Hoseo University, Beabang-eup Hoseo-ro 79-20, Asan-si, 31499, Chungcheongnam-do, Korea.

Background: Now that it is possible to efficiently classify and save tissue images of laboratory animals using whole-slide imaging, many diagnostic models are being developed through transfer learning with Convolutional Neural Network (CNN). In this study, transfer learning was performed to gain toxicopathological knowledge using CNN models such as InceptionV3 and Xception. For the classification of tubular basophilia and mineralization, two representative background lesions that commonly occur in toxicological studies, accuracies of diagnosis were compared using MobileNetV2, Xception and InceptionV3.

View Article and Find Full Text PDF

When diagnosing endometrial carcinoma cases, we encountered histological features that strikingly resembled uterine mesonephric-like adenocarcinoma (MLA), but the differential diagnosis remained challenging after performing immunostaining. Considering the aggressive biological behavior and poor prognosis of uterine MLA, we believe that the accurate recognition of mesonephric-like differentiation (MLD) is important in the diagnosis of endometrial carcinoma. We aimed to investigate the clinicopathological and molecular characteristics of such cases and compared them with those of uterine MLAs.

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!