Image Classification and Automated Machine Learning to Classify Lung Pathologies in Deceased Feedlot Cattle

Abstract

Bovine respiratory disease (BRD) and acute interstitial pneumonia (AIP) are the main reported respiratory syndromes (RSs) causing significant morbidity and mortality in feedlot cattle. Recently, bronchopneumonia with an interstitial pattern (BIP) was described as a concerning emerging feedlot lung disease. Necropsies are imperative to assist lung disease diagnosis and pinpoint feedlot management sectors that require improvement. However, necropsies can be logistically challenging due to location and veterinarians’ time constraints. Technology advances allow image collection for veterinarians’ asynchronous evaluation, thereby reducing challenges. This study’s goal was to develop image classification models using machine learning to determine RS diagnostic accuracy in right lateral necropsied feedlot cattle lungs. Unaltered and cropped lung images were labeled using gross and histopathology diagnoses generating four datasets: unaltered lung images labeled with gross diagnoses, unaltered lung images labeled with histopathological diagnoses, cropped images labeled with gross diagnoses, and cropped images labeled with histopathological diagnoses. Datasets were exported to create image classification models, and a best trial was selected for each model based on accuracy. Gross diagnoses accuracies ranged from 39 to 41% for unaltered and cropped images. Labeling images with histopathology diagnoses did not improve average accuracies; 34–38% for unaltered and cropped images. Moderately high sensitivities were attained for BIP (60–100%) and BRD (20–69%) compared to AIP (0–23%). The models developed still require fine-tuning; however, they are the first step towards assisting veterinarians’ lung diseases diagnostics in field necropsies.

Strategies for Bovine Respiratory Disease (BRD) Diagnosis and Prognosis: A Comprehensive Overview

“…These models are trained on images of necropsied right lateral lungs annotated with gross diagnoses. Although they have demonstrated moderate sensitivity in detecting BRD, there is a need for further refinement to enhance their diagnostic accuracy [123].…”

Epidemiological Factors Associated with Gross Diagnosis of Pulmonary Pathology in Feedyard Mortalities

“…An animal could also have been treated at least one time for AIP or BP and at least one time for OTHER diagnoses (BP or AIP + OTHER). At treatment, 13.2% (47) of the animals were diagnosed with AIP, 34.5% (123) with BP, 5% (18) with AIP + BIP, 10.6% (38) with BP or AIP + OTHER, 9.5% (34) with OTHER, and 27.2% (97) were not diagnosed or treated while alive. When comparing the diagnoses found at necropsy with the feedyard diagnoses, AIP had an agreement of 22.5% and a Cohen's kappa of 0.098 (Figure 1).…”

“…Recently, researchers reported the simultaneous presence of AIP and BP in deceased feedyard cattle. A new classification has been suggested as bronchopneumonia with an interstitial pneumonia (BIP) due to the presence of lesions compatible with AIP in caudodorsal lobes and BP on cranioventral lobes [ 16 , 17 , 18 , 19 ]. In previous studies when lesions consistent with BIP were found during necropsy, the final diagnosis or cause of death was limited to AIP or BP only, and it was dependent on the pathology that affected the majority of the lung [ 14 , 20 , 21 ].…”

Automated Lung Disease Detection and Classification Using Quantum Glowworm Swarm Optimizer with Quasi Recurrent Neural Network on Chest X-Ray Images

“…The utmost generally used image techniques are CT, and CXR imaging. These are analytical tools in allowing physicians to look at the internal structure of the human body by not involving surgery [5]. Discovering pulmonary ailment is presently accomplished by a professional radiologist through CXR image inspection because of its non-invasive valuation and comfort for total discoveries of the chest conditions in short.…”