Electrical Engineering and Systems Science > Image and Video Processing
[Submitted on 23 Nov 2023]
Title:Reducing Histopathology Slide Magnification Improves the Accuracy and Speed of Ovarian Cancer Subtyping
View PDFAbstract:Artificial intelligence has found increasing use for ovarian cancer morphological subtyping from histopathology slides, but the optimal magnification for computational interpretation is unclear. Higher magnifications offer abundant cytological information, whereas lower magnifications give a broader histoarchitectural overview. Using attention-based multiple instance learning, we performed the most extensive analysis of ovarian cancer tissue magnifications to date, with data at six magnifications subjected to the same preprocessing, hyperparameter tuning, cross-validation and hold-out testing procedures. The lowest magnifications (1.25x and 2.5x) performed best in cross-validation, and intermediate magnifications (5x and 10x) performed best in hold-out testing (62% and 61% accuracy, respectively). Lower magnification models were also significantly faster, with the 5x model taking 5% as long to train and 31% as long to evaluate slides compared to 40x. This indicates that the standard usage of high magnifications for computational ovarian cancer subtyping may be unnecessary, with lower magnifications giving faster, more accurate alternatives.
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.