Parallel Tile Processing for Cancer Metastasis Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting cancer metastasis in whole-slide images are time-consuming and prone to errors due to the large size and complexity of the images, leading to high computational challenges and inter- and intra-observer variabilities among pathologists, making it difficult to adopt as a daily tool in clinical settings.
Innovation Solution
A system and method that utilizes an asynchronous queue to process tiles of a whole-slide image in parallel using a learning model, such as a Fully Convolutional Neural Network (FCN), which allows for efficient detection of cancer metastasis by dividing the image into overlapping tiles and applying the model to multiple tiles simultaneously, reducing computational burden and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current methods divide WSIs into heavily overlapping patches and analyze each one with powerful CNN models to achieve high accuracy, then detection accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent divides the WSI into a grid of patches, but processes them in an asynchronous pipeline where multiple patches are loaded into GPU memory simultaneously and processed in parallel batches. This segmentation approach maintains detection accuracy while reducing total processing time through parallel computation.
Solution Approach 2:
The system pre-loads multiple patches into GPU memory before processing, and uses an asynchronous queue to manage the pipeline. This preliminary action ensures that data is ready for processing without waiting for sequential operations, significantly reducing processing time while maintaining accuracy.
2Reliability
If pathologists exhaustively examine extremely large WSIs to find potential metastases, then detection completeness is improved, but the process becomes time-consuming and error-prone
Solution Approach 1:
The patent introduces a CNN-based computer-aided diagnosis system as an intermediary that assists pathologists. The system processes the entire WSI automatically, providing detection results that complement pathologist examination, thereby improving detection completeness without requiring pathologists to spend excessive time on exhaustive manual review.
Solution Approach 2:
The system replaces the manual mechanical examination process with an automated computational approach using CNN models. This substitution maintains high detection completeness while dramatically reducing the time required, as the computational system can process images without fatigue or human error.
3Adaptability or versatility
If WSIs are divided into small patches for independent processing to meet CNN input requirements, then model applicability is improved, but computational cost increases and real-time clinical application becomes difficult
Solution Approach 1:
The patent implements a dynamic asynchronous processing pipeline that adapts to the computational requirements of different patches. The system loads multiple patches into GPU memory in batches, processes them in parallel, and manages memory dynamically. This approach maintains model applicability to small patches while significantly improving computational efficiency through parallel processing and optimized memory management.
Data Source
AI summary
Embodiments of the disclosure provide systems and methods for detecting cancer metastasis in a whole-slide image. The system may include a communication interface configured to receive the whole-slide image and a learning model. The whole-slide image is acquired by an image acquisition device. The system may also include a memory configured to store a plurality of tiles derived from the whole-slide image in a queue. The system may further include at least one processor, configured to apply the learning model to at least two tiles stored in the queue in parallel to obtain detection maps each corresponding to a tile, and detect the cancer metastasis based on the detection maps.