@article{YIN2026101930,
title = {Femtosecond Label-free Imaging: A Rapid and Reliable Alternative for Intraoperative Pathological Assessment in Thoracic Oncology},
journal = {JTCVS Open},
pages = {101930},
year = {2026},
issn = {2666-2736},
doi = {https://doi.org/10.1016/j.xjon.2026.101930},
url = {https://www.sciencedirect.com/science/article/pii/S2666273626003530},
author = {Hao Yin and Fangyi Liu and Wendi Zhu and Wenlong Yu and Kunbo Zhang and Rongkui Luo and Fenghao Sun and Yunlong Wang and Boxue Zhang and Bingwei Xu and Xin Zhu and Mingxiang Feng and Lijie Tan and Ming Li},
keywords = {Lung cancer, Esophageal cancer, Femtosecond label-free imaging, Multimodal optical imaging, Frozen section, Intraoperative diagnosis},
abstract = {Objective
To evaluate femtosecond label-free imaging (FLI) combined with artificial intelligence as a rapid alternative to frozen section pathology for intraoperative assessment in thoracic oncologic surgery.
Methods
This prospective study enrolled 144 patients undergoing resection for thoracic tumors at Shanghai Zhongshan Hospital (June-December 2025). Fresh specimens (259 lung, 96 esophageal) underwent FLI scanning using ultrashort laser pulses to generate multiple nonlinear optical signals (third harmonic generation, second harmonic generation, two-photon and three-photon fluorescence) without sectioning or staining. Deep learning algorithms were developed for lung tumor classification and exploratory high-risk feature prediction in ESCC. To avoid data leakage, specimen-level splits were adopted for lung tumor classification and patient-level splits for ESCC high-risk feature prediction. The models were constructed based on the pretrained UNI v1 foundation model combined with an attention-based multiple-instance learning (ABMIL) framework, with data augmentation (flipping and rotation) applied during training. Performance was evaluated against conventional histopathology using receiver operating characteristic (ROC) curve analysis, with data partitioned into training, validation, and test sets at a 7:1:2 ratio.
Results
FLI demonstrated significant time advantage over frozen section analysis (median 5.4 vs 36.3 minutes, P<0.001). Morphological concordance with hematoxylin and eosin histology was achieved across tissue types. AI-based tumor classification achieved mean area under the curve of 0.953 for lung specimens. Depth scanning at 3-μm intervals (up to 90 μm) enabled volumetric margin assessment, revealing depth-dependent tumor distribution variations not captured by single-plane frozen sections. Exploratory models predicted high-risk features in esophageal squamous cell carcinoma with mean area under the curve of 0.653-0.766 for lymph node metastasis, perineural invasion, and lymphovascular invasion.
Conclusions
Femtosecond label-free imaging provides rapid, morphologically concordant pathological assessment with significant workflow advantages over conventional frozen section analysis. Depth-scanning capability addresses sampling limitations inherent to single-plane evaluation. Exploratory AI models demonstrate feasibility for intraoperative prediction of high-risk pathological features in esophageal cancer; however, these results are preliminary and require validation in larger, multicenter cohorts before clinical deployment.}
}