Related Experiment Videos
Using spiral intensity profile to quantify head and neck cancer
Koon Y Kong1, Yachna Shanna1, S Hussain Raza2
1Georgia Institute of Technology, Atlanta, GA, 30332 USA.
Proceedings. IEEE International Symposium on Bioinformatics and Bioengineering
|April 11, 2017
Summary
A novel spiral intensity profile algorithm aids in segmenting cancerous tissue in microscopy images, improving biomarker analysis for head and neck cancer. This method enhances reproducibility and reduces pathologist workload.
Area of Science:
- Digital pathology
- Biomarker discovery
- Computational imaging
Background:
- Manual segmentation of regions of interest (ROI) in microscopy images is subjective and time-consuming.
- Accurate identification of cancerous tissue is crucial for biomarker analysis, such as for folic acid receptors in head and neck cancer.
- Existing algorithms often process pixel intensity and spatial information separately, limiting segmentation accuracy.
Purpose of the Study:
- To develop and validate a novel spiral intensity profile algorithm for segmenting cancerous regions in light microscopy images.
- To assess the potential of folic acid receptors as biomarkers in head and neck cancer by enabling accurate quantification within segmented cancerous areas.
- To improve the reproducibility and efficiency of image analysis in digital pathology.
Main Methods:
- A new algorithm utilizing a spiral intensity profile was developed to integrate pixel intensity and spatial information for image segmentation.
- The algorithm performs segmentation at multiple scales, from whole cancer regions to individual nuclei.
- The performance was evaluated by comparing its output to manually segmented images.
Main Results:
- The spiral intensity profile algorithm achieved a specificity of 83.7% and a sensitivity of 61.1% when compared to manual segmentation.
- The method demonstrated the ability to segment cancerous tissue at various scales, including cancer regions, nuclei clusters, and individual nuclei.
- Spiral intensity profiles were identified as a valuable feature for enhancing existing segmentation algorithms.
Conclusions:
- The proposed spiral intensity profile algorithm offers a promising approach for automated segmentation of cancerous tissues in microscopy images.
- This method facilitates accurate quantification of biomarkers like folic acid receptors within specific cancerous regions, improving diagnostic potential.
- The algorithm's ability to segment at different scales enhances its utility for detailed analysis in head and neck cancer research and digital pathology.