Related Experiment Videos
Quantitative nuclear texture features analysis confirms WHO classification 2004 for lung carcinomas.
Katharina Schmid1, Nina Angerstein, Silvana Geleff
1Department of Clinical Pathology, Medical University of Vienna, Vienna, Austria. katharina.schmid@meduniwien.ac.at
Summary
Nuclear chromatin texture analysis can distinguish small-cell from non-small-cell lung carcinoma (93%) and carcinoids from non-small-cell lung carcinomas (92%). However, it cannot differentiate neuroendocrine tumor subtypes, aligning with the 2004 WHO classification focusing on mitotic counts and necrosis over morphology.
Area of Science:
- Pulmonary Pathology
- Computational Pathology
- Cancer Diagnostics
Background:
- The 2004 World Health Organization (WHO) classification introduced new criteria for lung carcinoma subtypes.
- Accurate discrimination of lung carcinoma subtypes is crucial for appropriate treatment and prognosis.
- Morphological features alone can be insufficient for classifying certain neuroendocrine lung tumors.
Purpose of the Study:
- To evaluate the efficacy of nuclear chromatin texture feature analysis in discriminating lung carcinoma subtypes according to the 2004 WHO classification.
- To determine if image analysis of nuclear chromatin can replace or supplement traditional morphological assessments.
Main Methods:
- A cohort of 185 lung carcinoma cases was analyzed, including carcinoids, small-cell carcinomas, large-cell neuroendocrine carcinomas, adenocarcinomas, and squamous cell carcinomas.
- Feulgen staining and automated high-resolution image analysis were employed to measure nuclear chromatin texture features.
- Classification rules based on chromatin granularity and compactness were developed and applied to discriminate tumor groups.
Main Results:
- Granularity-based analysis correctly discriminated small-cell from non-small-cell lung carcinoma in 93% of cases.
- Compactness-based analysis distinguished carcinoids from non-small-cell lung carcinomas with 92% accuracy.
- No significant discrimination was achieved between different subtypes of large-cell carcinomas or between neuroendocrine tumors.
Conclusions:
- Nuclear chromatin texture analysis shows potential for distinguishing major lung carcinoma categories but is insufficient for differentiating neuroendocrine tumor subtypes.
- The findings support the 2004 WHO classification's shift towards mitotic counts and necrosis for neuroendocrine tumor classification, deeming morphological discrimination via image analysis inappropriate.
- Further research may explore combined features for improved classification accuracy.