To develop and validate a radiomics model to distinguish tumor marker-negative sellar germ cell tumors (GCTs) from Langerhans cell histiocytosis (LCH), highlighting the clinical significance of accurate differentiation.
Approach:
Key Findings:
The combined model of radiomics with clinical and imaging semantic features achieved the best performance with an AUC of 0.81, indicating strong diagnostic capability.
Statistically significant differences in AUC were confirmed by the DeLong test (p < 0.05), underscoring the model's reliability.
Interpretation:
Radiomics-based machine learning offers a promising non-invasive method for differentiating between tumor marker-negative sellar GCTs and LCH, aiding in treatment decision-making.
Limitations:
The study is retrospective and conducted at a single institution, which may limit generalizability and applicability in diverse clinical settings.
Potential biases in feature extraction and model validation processes could affect the robustness of the findings.
Conclusion:
Radiomics and machine learning can effectively distinguish between sellar GCTs and LCH, potentially improving clinical outcomes through better diagnostic accuracy and informed treatment decisions.
