Breast Cancer, Digital Pathology, Lymph Node Metastasis, Virtual Staining
Conditions
Brief summary
The goal of this observational study is to develop an artificial intelligence model to transform unstained lymph node tissue slice images directly into stained images. The main questions it aims to answer are: Can the virtual staining model generate hematoxylin and eosin (H&E) and immunohistochemistry (IHC) images suitable for clinical diagnosis from unstained paraffin-embedded lymph node slice images, including those from breast axillary lymph nodes and other tumor lymph nodes? Can the virtual staining model generate H&E and IHC images suitable for clinical diagnosis from unstained frozen sentinel lymph node slice images from breast cancer patients? Researchers will retrospectively collect paraffin-embedded lymph node slices from tumor patients and prospectively collect frozen sentinel lymph node slices from breast cancer patients.
Interventions
None listed
Sponsors
Affiliated Cancer Hospital of Shantou University Medical College
Yunnan Cancer Hospital
Study design
Observational model
OTHER
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to 75 Years
Healthy volunteers
No
Inclusion criteria
Part 1: Female patients aged 18-75 with breast cancer; Undergoing surgical excision of breast cancer and sentinel lymph node biopsy/axillary lymph node dissection; Lymph nodes with clear postoperative paraffin pathological results. Part 2: Patients aged 18-75 with one of the following cancers: thyroid, lung, esophagus, stomach, colorectal, prostate, bladder, or cervix; Undergoing surgical resection of lymph nodes; Lymph nodes with clear postoperative paraffin pathological results. Part 3: Female patients aged 18-75 with breast cancer; Undergoing surgical excision of breast cancer and sentinel lymph node biopsy; Sentinel lymph nodes with clear postoperative paraffin pathological results.
Exclusion criteria
Part 1 / Part 2: Lymph node diagnosis is missing; Absence of lymph node component in the slice. Part 3: Sentinel lymph node diagnosis is missing; Absence of lymph node component in the frozen slice.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Lymph node metastasis status | 2024-2025 | Lymph node metastasis status: metastasis or non-metastasis |
| Accuracy, Sensitivity, Specificity，Area under the curve， | 2024-2025 | The performance of pathologists diagnosing the lymph node metastasis status by virtual and real staining whole slide images |
| Positive predictive value，Negative predictive value | 2024-2025 | The performance of pathologists diagnosing the lymph node metastasis status by virtual and real staining whole slide images |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Peak Signal-to-Noise Ratio(PSNR) | 2024-2025 | Scores of the similarity between virtual and real staining of lymph nodes, with values ranging from 0 to infinity, the higher scores mean the better outcomes |
| Multi-Scale Structural Similarity (MS-SSIM) | 2024-2025 | Scores of the similarity between virtual and real staining of lymph nodes, with values ranging from 0 to 1, the higher scores mean the better outcomes |
| Pearson correlation coefficient | 2024-2025 | Scores of the similarity between virtual and real staining of lymph nodes, with values ranging from 0 to 1, the higher scores mean the better outcomes |
Outcome results
None listed