Personalized Prior Study Ranking to Reduce Radiology Search Time
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The process of identifying relevant prior studies for a current medical procedure is laborious and inefficient, often leading to time-consuming manual searches and errors, and results in unnecessary imaging studies due to inconsistent radiologist criteria and lack of automated relevance assessment.
Innovation Solution
An optimization server uses a personalized model to automatically identify and rank relevant prior studies based on a radiologist's preferences, incorporating machine learning to determine relevance scores and update models dynamically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual search through prior studies is performed, then radiologist can identify relevant studies, but time consumption increases significantly
Solution Approach 1:
The patent replaces the manual mechanical search process with an automated computer-based system that uses machine learning models to identify relevant prior studies. The system automatically retrieves, ranks, and presents prior studies based on their relevance to the current study, eliminating the need for radiologists to manually search through archives while maintaining high accuracy in identification.
Solution Approach 2:
The system enables prior studies to 'serve themselves' by automatically retrieving and presenting relevant studies without human intervention. The automated retrieval and ranking system allows the computer system to perform the identification task independently, freeing radiologists from time-consuming manual search activities while ensuring consistent and accurate results.
2Productivity
If automated search using semantic parameters is used, then search time is reduced, but search results are limited to identical modality and body part
Solution Approach 1:
The patent implements dynamic search capabilities where the system automatically adjusts search parameters and criteria based on the specific clinical context and study characteristics. Rather than using fixed semantic parameters, the system dynamically determines which parameters are most relevant for each search, allowing it to retrieve studies across different modalities and body parts when clinically appropriate, thus maintaining both speed and versatility.
Solution Approach 2:
The system changes search parameters automatically based on the current study's characteristics and clinical needs. Instead of being constrained to identical modality and body part parameters, the system dynamically modifies search criteria to include complementary modalities and related anatomical regions when relevant, thereby expanding the versatility of automated search while maintaining efficiency.
3Adaptability or versatility
If radiologist uses personal criteria for selecting prior studies, then relevance assessment is customized, but consistency and reproducibility decrease
Solution Approach 1:
The patent incorporates feedback mechanisms where the system learns from radiologist interactions and decisions. The machine learning model is trained on radiologist preferences and feedback, allowing it to customize the retrieval and ranking of prior studies according to individual radiologist workflows while maintaining consistent and reproducible results across different users and time periods. This feedback loop ensures both customization and reliability.
Solution Approach 2:
The system performs preliminary analysis and pre-ranking of prior studies based on learned criteria before presentation to the radiologist. By pre-processing and organizing studies according to relevance metrics derived from training data, the system establishes a consistent foundation that can be customized for individual users while maintaining reproducibility in the core identification process.
4Reliability
If exhaustive manual review of all prior studies is performed, then completeness of assessment is improved, but workflow efficiency decreases
Solution Approach 1:
The patent extracts and prioritizes only the most relevant prior studies from the complete set, presenting a curated subset to the radiologist. Rather than requiring review of all available studies, the system automatically identifies and extracts the top-ranked relevant studies based on multiple criteria including modality, anatomy, timing, and clinical context, ensuring completeness of assessment for critical studies while maintaining workflow efficiency by limiting the number of studies requiring detailed review.
Solution Approach 2:
The system performs a comprehensive automated assessment of all prior studies to determine relevance, then presents only the necessary subset for radiologist review. This partial action approach allows the system to complete the exhaustive analysis computationally while requiring minimal human effort, achieving both completeness in assessment and efficiency in workflow by dividing the task between automated processing and human judgment.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A device, system, and method optimizes usage of prior studies. The method performed by an optimization server includes receiving a request for relevant prior studies for a patient from a practitioner device utilized by a medical professional, the request including a current study for the patient, the relevant prior studies being relevant to the current study. The method includes determining the relevant prior studies from prior studies of the patient based on a personalized model, the personalized model associated with the medical professional, the personalized model indicating a relevance score of the relevant prior studies to the current study. The method includes transmitting the relevant prior studies to the practitioner device.