Variable Resolution DBT Reconstruction for Lesion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital breast tomosynthesis reconstruction methods face challenges in efficiently processing and storing large data volumes due to uniform slice thickness and resolution, which can lead to overlooked diagnostically relevant information and increased reading time, as they fail to account for non-uniform lesion distribution and varying anatomical structures.
Innovation Solution
A method for reconstructing digital breast tomosynthesis volumes with variable slice thickness and resolution, where different reconstruction algorithms are applied to regions based on diagnostically relevant information density, allowing for high-resolution reconstruction in areas with high relevance and low-resolution reconstruction in areas with low relevance, thereby optimizing data processing and storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If uniform slice thickness and resolution are used for entire breast volume reconstruction, then storage space and reading time are reduced, but diagnostically relevant information may be overlooked due to lack of high-resolution detail in lesion-prone areas
Solution Approach 1:
The patent applies different reconstruction algorithms and slice thicknesses to different regions of the breast volume based on lesion probability. High-resolution reconstruction is applied to regions with high lesion probability, while lower-resolution reconstruction is applied to regions with low lesion probability, thereby optimizing the balance between data volume and image resolution.
Solution Approach 2:
The breast volume is segmented into multiple regions based on lesion probability distribution. Each region is then processed separately with appropriate reconstruction parameters, allowing selective high-resolution reconstruction in critical areas while using lower resolution in less critical areas.
2Measurement precision
If high-resolution reconstruction is applied to entire breast volume, then diagnostic accuracy is improved, but storage space requirements and reading time increase significantly
Solution Approach 1:
Instead of applying uniform high-resolution reconstruction throughout the entire breast volume, the patent selectively applies high-resolution reconstruction only to regions with high lesion probability. This localized approach maintains diagnostic accuracy in critical areas while significantly reducing overall data volume and storage requirements.
Solution Approach 2:
The patent applies partial high-resolution reconstruction only where needed (in regions with high lesion probability) rather than excessive high-resolution reconstruction throughout the entire volume. This partial action approach optimizes the balance between diagnostic accuracy and data management.
3Loss of time
If thick slab reconstruction is used, then reading time is reduced and storage space is saved, but diagnostically relevant information in high-probability regions may be missed due to insufficient resolution
Solution Approach 1:
The patent applies different slab thicknesses to different regions based on lesion probability. Thinner slabs with higher resolution are used in regions with high lesion probability, while thicker slabs are used in regions with low lesion probability, optimizing both reading efficiency and diagnostic accuracy.
Solution Approach 2:
The slab thickness is made variable rather than fixed, allowing dynamic adjustment based on the specific requirements of different breast regions. This dynamic slab thickness approach enables optimized reading time and resolution balance across the entire volume.
Data Source
AI summary
A method for compressing digital breast tomosynthesis data, a system and a control unit for image reconstruction of three-dimensional digital breast tomosynthesis volumes (DBT). The volume to be reconstructed is analyzed in order to identify clusters of regions in the volume with a low and high degree of diagnostically relevant information. Depending on the affiliation or belonging to a certain cluster, a specific reconstruction algorithm and a specific slab thickness are determined in order to be used for reconstruction of the cluster. Thus, different clusters are reconstructed differently.