Multiple Instance Learning for the Detection of Lymph Node and Omental Metastases in Carcinoma of the Ovaries, Fallopian Tubes and Peritoneum
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.Artificial intelligence (AI) models can accurately detect ovarian cancer metastases in lymph nodes and omentum. This AI tool shows potential for pre-screening pathology slides, saving time and aiding histopathology workflows.
Area Of Science
- Computational pathology
- Digital pathology
- Oncology
Background
- Surgical pathology for tubo-ovarian and peritoneal cancers involves extensive tissue assessment for metastases.
- Lymph nodes and omentum are routinely examined, contributing significantly to diagnostic workload.
- AI solutions for ovarian cancer metastasis detection in these tissues are underdeveloped.
Purpose Of The Study
- To develop and evaluate an AI model for detecting ovarian carcinoma metastases in lymph nodes and omentum.
- To assess the utility of AI in streamlining histopathology workflows for ovarian cancer diagnosis.
Main Methods
- Utilized an attention-based multiple-instance learning (ABMIL) approach with a vision-transformer foundation model.
- Classified whole-slide images (WSIs) for the presence or absence of ovarian carcinoma metastases.
- Trained and validated the model on 855 WSIs from 404 patients.
Main Results
- Achieved an AUROC of 0.998 and 100% balanced accuracy for lymph node metastasis detection.
- Reached an AUROC of 0.963 and 98.0% balanced accuracy for omental metastasis detection.
- Demonstrated high performance in identifying ovarian cancer spread in surgical resection specimens.
Conclusions
- The AI model shows significant potential for identifying ovarian carcinoma nodal and omental metastases.
- This tool can pre-screen WSIs, offering substantial time savings for histopathologists.
- The AI can help alleviate staffing shortages and improve efficiency in diagnostic pathology.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.