Assessing Tumour Budding in Lung Squamous Cell Carcinoma: A Comparative Analysis of Digital Whole-Slide Imaging and Light Microscopy.

INTRODUCTION: Tumour budding (TB) is an emerging prognostic marker in solid cancers and has recently been endorsed by the IASLC as a component of grading in lung squamous cell carcinoma (LUSC). With the growing adoption of digital pathology, the reproducibility of TB assessment on whole-slide images (WSIs) warrants evaluation. METHODS: We assessed the utility of WSI for TB evaluation in a retrospective cohort of 204 resected LUSC cases. Peritumoral budding (PTB) and intratumoural budding (ITB) were independently scored by two pathologists, applying the International Tumour Budding Consensus Conference (ITBCC) criteria established for colorectal cancer and endorsed by the IASLC for use in LUSC. A total of 816 TB hotspot regions were analysed. RESULTS: Substantial intraobserver reliability was observed, with strong correlations between absolute bud counts in WSI and conventional light microscopy (CLM; R = 0.845, p < 0.001; R = 0.880, p < 0.001). Interobserver agreement was moderate for both PTB and ITB across platforms (Cohen's κ: CLM 0.582 [PTB], 0.570 [ITB]; WSI 0.593 [PTB], 0.423 [ITB]). Discordance mainly occurred in lower TB categories due to differences in hotspot selection and morphological mimics. When applying the recently proposed IASLC two-tier cutoff (0-9 buds vs. ≥10 buds), interobserver agreement for PTB improved modestly (κ = 0.49 for CLM; κ = 0.53 for WSI), with overall agreement exceeding 93%. CONCLUSION: These findings support the feasibility and reproducibility of TB assessment using WSI and highlight its potential for standardizing LUSC TB evaluation in digital pathology workflows. WSI also demonstrated practical advantages, including easier field selection and annotation.