Accurate and Rapid Detection of Peritoneal Metastasis from Gastric Cancer by AI-assisted Stimulated Raman Cytology

Abstract

Peritoneal metastasis (PM) is the most common form of distant metastasis and one of the leading causes of death in gastric cancer (GC). For locally advanced GC, clinical guidelines recommend peritoneal lavage cytology for intraoperative PM detection. Unfortunately, current peritoneal lavage cytology is limited by low sensitivity (<60%). Here we established the stimulated Raman cytology (SRC), a chemical microscopy-based intelligent cytology. By taking advantages of stimulated Raman scattering in label-free, high-speed, and high-resolution chemical imaging, we firstly imaged 53951 exfoliated cells in ascites obtained from 80 GC patients (27 PM positive, 53 PM negative), at the Raman bands corresponding to DNA, protein, and lipid, respectively. Then, we revealed 12 single cell features of morphology and composition that were significantly different between PM positive and negative specimens, including cellular area, lipid protein ratio, etc. Importantly, we developed a single cell phenotyping algorithm to further transform the above raw features to feature matrix. Such matrix was crucial to identify the significant marker cell cluster, the divergence of which was finally used to differentiate the PM positive and negative. Compared with histopathology, the gold standard of PM detection, our SRC method assisted by machine learning classifiers could reach 81.5% sensitivity, 84.9% specificity, and the area under receiver operating characteristic curve of 0.85, within 20 minutes for each patient. Such remarkable improvement in detection accuracy is largely owing to incorporation of the single-cell composition features in SRC. Together, our SRC method shows great potential for accurate and rapid detection of PM from GC.

Artificial intelligence-assisted stimulated Raman molecular cytology enables accurate and rapid intraoperative assessment of the surgical margin in pancreatic ductal adenocarcinoma

“…The comparison between PNM assessment with different feature sets reflects the predominant significance of componential information. In comparison, its relative contribution was relatively smaller in the SRS microscopy-based peritoneal metastasis detection as we studied before 26 . Such difference may be attributed to a more distinct abnormality in the lipid metabolism of PDAC cells 27 .…”

“…3b . This was similar to the K-PCA in our previous study 26 , but different in that the dimensionality reduction and clustering were performed on all N and T samples and then applied to each M sample instead of directly on each M sample because the exfoliated cells collected from a PNM sample were much less. The methods used here were named K-N-PCA or K-N-ISOMAP for distinguishment ( Fig.…”

“…However, the sensitivity of conventional cytology, which only exploits the morphological information of cells, is typically lower than 60% 23-25 . This problem can be significantly improved by detecting the altered metabolism of cancer cells, as demonstrated by cytology based on spontaneous Raman spectroscopy 51 or SRS microscopy 26 . Note that the former required several hours for spectra acquisition and thus cannot be used intraoperatively, while the latter can be finished within 30 min.…”

Label-free virtual peritoneal lavage cytology via deep-learning-assisted single-color stimulated Raman scattering microscopy

“…In contrast, smaller coverage is required for cytology because free malignant cells can be captured more easily when mixed evenly with benign cells. As demonstrated in our previous study 27 , a 0.72 mm² SRS image costing only 1.5 min and containing ~600 cells can be sufficient for the diagnosis of each patient. After SRS microscopy, the slides were stained with hematoxylin and eosin.…”

“…In comparison, SRS has a higher potential for cellular imaging with subcellular spatial resolution and chemical-bond-level molecular specificity. In our previous work 27 , with SRS microscopy, we established a label-free PM detection method that matched well with the gold standard, laparoscopy-guided biopsy, and according to clinical guidelines, a corresponding SRS-based peritoneal lavage cytology with virtual staining is worth exploring. Currently, SRS microscopy can provide histological contrast in tissue sections as hematoxylin-eosin (H&E) stain does 28–30 .…”