Explainable Cryobiopsy AI Model, CRAI, to Predict Disease Progression for Transbronchial Lung Cryobiopsies with Interstitial Pneumonia
Uegami, W.; Okoshi, E. N.; Lami, K.; Nei, Y.; Ozasa, M.; Kataoka, K.; Kitamura, Y.; Kohashi, Y.; Cooper, L. A. D.; Sakanashi, H.; Saito, Y.; Kondoh, Y.; the study group on CRYOSOLUTION, ; Fukuoka, J.
Show abstract
BackgroundInterstitial lung disease (ILD) encompasses diverse pulmonary disorders with varied prognoses. Current pathological diagnoses suffer from inter-observer variability,necessitating more standardized approaches. We developed an ensemble model AI for cryobiopsy, CRAI, an artificial intelligence model to analyze transbronchial lung cryobiopsy (TBLC) specimens and predict patient outcomes. MethodsWe developed an explainable AI model, CRAI, to analyze TBLC. CRAI comprises seven modules for detecting histological features, generating 19 pathologically significant findings. A downstream XGBoost classifier was developed to predict disease progression using these findings. The models performance was evaluated using respiratory function changes and survival analysis in cross-validation and external test cohorts. FindingsIn the internal cross-validation (135 cases), the model predicted 105 cases without disease progression and 30 with disease progression. The annual {Delta}%FVC was -1.293 in the non-progressive group versus -5.198 in the progressive group, outperforming most pathologists diagnoses. In the external test cohort (48 cases), the model predicted 38 non-progressive and 10 progressive cases. Survival analysis demonstrated significantly shorter survival times in the progressive group (p=0.034). InterpretationCRAI provides a comprehensive, interpretable approach to analyzing TBLC specimens, offering potential for standardizing ILD diagnosis and predicting disease progression. The model could facilitate early identification of progressive cases and guide personalized therapeutic interventions. FundingNew Energy and Industrial Technology Development Organization (NEDO) and Japanese Ministry of Health, Labor, and Welfare.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Transformer-based deep learning model for the diagnosis of suspected lung cancer in primary care based on electronic health record data 91%
- Distinct lung cell signatures define the temporal evolution of diffuse alveolar damage in fatal COVID-19 91%
- Redefining mesothelioma types as a continuum uncovers the immune and vascular systems as key players in the diagnosis and prognosis of this disease 90%
Similar papers in this journal
- An ML prediction model based on clinical parameters and automated CT scan features for COVID-19 patients 94%
- Automated processing of thermal imaging to detect COVID-19 94%
- Immune, metabolic, anatomical, and functional features of people after successful tuberculosis treatment: an exploratory analysis 94%
Similar papers in this journal
- MIXTURE of human expertise and deep learning—Developing an explainable model for predicting pathological diagnosis and survival in patients with interstitial lung disease 96%
- Genomic Characterization of Lung Cancer in Never-Smokers Using Deep Learning 92%
- Artificial intelligence-driven morphology-based enrichment of malignant cells from body fluid 90%
Similar papers in this journal
- Graph perceiver network for lung tumor and bronchial premalignant lesion stratification from histopathology 91%
- Detection of Colorectal Adenocarcinoma and Grading Dysplasia on Histopathologic Slides Using Deep Learning 91%
- Fatal COVID-19 and non-COVID-19 Acute Respiratory Distress Syndrome is Associated with Incomplete Alveolar Type 1 Epithelial Cell Differentiation from the Transitional State Without Fibrosis 89%
Similar papers in this journal
- Deep learning models for poorly differentiated colorectal adenocarcinoma classification in whole slide images using transfer learning 91%
- Volumetric lung cancer screening reduces unnecessary low-dose computed tomography scans: results from a single-centre prospective trial on 4,119 subjects 89%
- Detection, Isolation and Quantification of Myocardial Infarct with Four Different Histological Staining Techniques 89%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.