LIRA

LIRA analyzes histopathology images using convolutional neural networks to quantify pulmonary lesion types and pathology severity in tuberculosis animal models.

Key Features:

• Automated Histopathological Analysis: Automates analysis of histopathology images from tuberculosis-infected animal models to reduce subjective scoring variability.
• Quantitative Scoring System: Generates rapid quantitative pathology scores as an alternative to semi-quantitative, pathologist-dependent methods.
• Convolutional Neural Networks (CNNs): Employs CNNs trained on images from the C3HeB/FeJ tuberculosis mouse model to classify seven distinct pathology features and three lesion types.
• Broad Applicability: Methodology can be adapted to other disease models and tissue types for digital histopathology analysis.

Scientific Applications:

• Preclinical evaluation of tuberculosis therapies: Provides standardized quantitative measures of pulmonary pathology to compare treatment efficacy across cohorts in preclinical studies.
• Disease progression assessment and statistical analysis: Supports objective assessment of disease progression and supplies quantitative data for statistical evaluation of therapeutic interventions.

Methodology:

Convolutional neural networks trained on histopathology images from the C3HeB/FeJ tuberculosis mouse model classify seven pathology features and three lesion types to produce quantitative pathology scores.