Automated detection of regions of interest for tissue microarray experiments: an image texture analysis
2007

Automated Detection of Cancerous Regions in Breast Tissue Images

Sample size: 2395 publication 10 minutes Evidence: moderate

Author Information

Author(s): Karaçali Bilge, Tözeren Aydin

Primary Institution: Drexel University

Hypothesis

Can automated image analysis effectively identify regions of interest in breast tissue for molecular profiling?

Conclusion

The study demonstrates a highly efficient automated method for identifying cancer-specific regions in histology slides.

Supporting Evidence

  • The automated method identified cancer-specific regions that were previously marked by pathologists.
  • Both grayscale and color segmentation techniques were used to analyze the same dataset.
  • The study achieved better than 95% specificity in identifying normal and cancer-specific clusters.

Takeaway

This study shows a computer program can help doctors find cancer areas in breast tissue images, making it easier to study and treat cancer.

Methodology

The study used image texture analysis and statistical learning algorithms to classify breast tissue images into normal, cancerous, and non-specific categories.

Potential Biases

Potential operator dependency in initial evaluations by pathologists.

Limitations

The method's validity needs to be tested on a larger set of images from clinical trials.

Participant Demographics

Breast tissue samples from 6 specimens, including both normal and cancerous tissues.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2342-7-2

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication