KEY TAKEAWAYS
- The study aimed to assess H&E-stained WSI for AI TME evaluation and characterize ADCA and SCCA in fibrotic lungs.
- The EITL-RF CNN model showed potential for TME evaluation in ADCA and SCCA; further study is needed.
The tumor microenvironment (TME) is crucial in lung cancer initiation, growth, invasion, and metastasis. AI methods have the potential to accelerate TME analysis.
Anjali Saqi and the team aimed to evaluate the feasibility of using hematoxylin and eosin (H&E)-stained whole slide images (WSI) to develop an AI model for TME evaluation, and characterize the TME of adenocarcinoma (ADCA) and squamous cell carcinoma (SCCA) in fibrotic and non-fibrotic lung tissue.
The cohort was derived from chest CT scans of patients with lung neoplasms, both with and without background fibrosis. WSI images were generated from slides of 76 available pathology cases, including ADCA (n = 53) and SCCA (n = 23) in fibrotic (n = 47) and non-fibrotic (n = 29) lung tissue.
Detailed ground-truth annotations of stroma (fibrosis, vessels, inflammation), necrosis, and background were performed on WSI and optimized through an expert-in-the-loop (EITL) iterative procedure using a lightweight random forest (RF) classifier.
A convolutional neural network (CNN)-based model was employed to achieve tissue-level multiclass segmentation. The model was trained on 25 annotated WSI from 13 cases of ADCA and SCCA, both with and without fibrosis, and then applied to the 76-case cohort. The TME analysis included tumor stroma ratio (TSR), tumor fibrosis ratio (TFR), tumor inflammation ratio (TIR), tumor vessel ratio (TVR), tumor necrosis ratio (TNR), and tumor background ratio (TBR).
The model achieved a precision of 94%, sensitivity of 90%, and F1-score of 91% in overall classification. Significant differences were observed in the TSR, P= 0.041) and TFR, P= 0.001) between fibrotic and non-fibrotic adenocarcinoma (ADCA).
Within fibrotic lung tissue, ADCA showed statistically significant differences in TFR (P= 0.039),TIR, P= 0.003), TVR, P= 0.041), TNR, P= 0.0003), and TBR, P= 0.020) compared to SCCA.
The study concluded that the combined EITL-RF CNN model utilizing H&E WSI effectively enables multiclass evaluation and quantification of TME. Significant variations were identified in the TME of ADCA and SCCA with or without background fibrosis. Further research is warranted to ascertain the implications of TME findings for prognosis and treatment strategies.
Funding was provided by the Boehringer Ingelheim.
Source: https://pubmed.ncbi.nlm.nih.gov/38978066/
Saqi A, Liu Y, Politis MG, et al. (2024). “Combined expert-in-the-loop-random forest multiclass segmentation U-net based artificial intelligence model: evaluation of non-small cell lung cancer in fibrotic and non-fibrotic microenvironments.” J Transl Med. 2024;22(1):640. Published 2024 Jul 8. doi:10.1186/s12967-024-05394-2
