Automated classification of renal cell carcinoma subtypes using scale invariant feature transform
S Raza1, Yachna Sharma, Qaiser Chaudry
1Georgia Institute of Technology, Atlanta, GA 30332, USA. sraza3@gatech.edu
Summary
This study introduces an automated system for classifying renal cell carcinoma subtypes from biopsies, achieving 83% accuracy. The computer-assisted diagnosis (CAD) method reduces pathologist variability and speeds up diagnosis.
Area of Science:
- Pathology
- Medical Imaging
- Computer-Aided Diagnosis
Background:
- Pathologist analysis of renal cell carcinoma biopsies is time-consuming and prone to variability.
- Computer-assisted diagnosis (CAD) systems offer potential to improve accuracy and efficiency.
Purpose of the Study:
- To develop an automated computer-assisted diagnostic system for renal cell carcinoma subtype classification.
- To reduce intra- and inter-user variability in biopsy analysis.
Main Methods:
- Utilized scale-invariant features to capture morphological distinctness of renal cell carcinoma subtypes.
- Classified a heterogeneous dataset of renal cell carcinoma biopsy images.
- Employed automated analysis to minimize human intervention and user subjectivity.
Main Results:
- Achieved a classification accuracy of 83% for renal cell carcinoma subtypes.
- The system successfully classified heterogeneous biopsy images.
- The method minimized human intervention and user subjectivity.
Conclusions:
- The proposed automated CAD system effectively classifies renal cell carcinoma subtypes.
- Scale-invariant features are valuable for capturing morphological differences in cancer subtypes.
- Automated analysis can enhance diagnostic consistency and efficiency in pathology.