Artificial intelligence-based recurrence prediction outperforms classical histopathological methods in pulmonary adenocarcinoma biopsies.

INTRODUCTION: Between 10 and 50% of early-stage lung adenocarcinoma patients experience local or distant recurrence. Histological parameters such as a solid or micropapillary growth pattern are well-described risk factors for recurrence. However, not every patient presenting with such a pattern will develop recurrence. Designing a model which can more accurately predict recurrence on small biopsy samples can aid the stratification of patients for surgery, (neo-)adjuvant therapy, and follow-up. MATERIAL AND METHODS: In this study, a statistical model on biopsies fed with histological data from early and advanced-stage lung adenocarcinomas was developed to predict recurrence after surgical resection. Additionally, a convolutional neural network (CNN)-based artificial intelligence (AI) classification model, named AI-based Lung Adenocarcinoma Recurrence Predictor (AILARP), was trained to predict recurrence, with an ImageNet pre-trained EfficientNet that was fine-tuned on lung adenocarcinoma biopsies using transfer learning. Both models were validated using the same biopsy dataset to ensure that an accurate comparison was demonstrated. RESULTS: The statistical model had an accuracy of 0.49 for all patients when using histology data only. The AI classification model yielded a test accuracy of 0.70 and 0.82 and an area under the curve (AUC) of 0.74 and 0.87 on patch-wise and patient-wise hematoxylin and eosin (H&E) stained whole slide images (WSIs), respectively. CONCLUSION: AI classification outperformed the traditional clinical approach for recurrence prediction on biopsies by a fair margin. The AI classifier may stratify patients according to their recurrence risk, based only on small biopsies. This model warrants validation in a larger lung biopsy cohort.

The value of prognostic and predictive parameters in early-stage lung adenocarcinomas: A comparison between biopsies and resections.

INTRODUCTION: Since lung adenocarcinoma (LUAD) biopsies are usually small, it is questionable if their prognostic and predictive information is comparable to what is offered by large resection specimens. This study compares LUAD biopsies and resection specimens for their ability to provide prognostic and predictive parameters. METHODS: We selected 187 biopsy specimens with stage I and II LUAD. In 123 cases, subsequent resection specimens were also available. All specimens were evaluated for growth pattern, nuclear grade, fibrosis, inflammation, and genomic alterations. Findings were compared using non-parametric testing for categorical variables. Model performance was assessed using the area under the curve for both biopsies and resection specimens, and overall (OS) and disease-free survival (DFS) was calculated. RESULTS: The overall growth pattern concordance between biopsies and resections was 73.9%. The dominant growth pattern correlated with OS and DFS in resected adenocarcinomas and for high-grade growth pattern in biopsies. Multivariate analysis of biopsy specimens revealed that T2-tumors, N1-status, KRAS mutations and a lack of other driver mutations were associated with poorer survival. Model performance using clinical, histological and genetic data from biopsy specimens for predicting OS and DSF demonstrated an AUC of 0.72 and 0.69, respectively. CONCLUSIONS: Our data demonstrated the prognostic relevance of a high-grade growth pattern in biopsy specimens of LUAD. Combining clinical, histological and genetic information in one model demonstrated a suboptimal performance for DFS prediction and good performance for OS prediction. However, for daily practice, more robust (bio)markers are required to predict prognosis and stratify patients for therapy and follow-up.