Volumetric CT-based segmentation of NSCLC using 3D-Slicer

Area of Science:

• Medical Imaging
• Oncology
• Computational Biology

Background:

• Accurate tumor volume assessment is crucial for non-small cell lung cancer (NSCLC) treatment planning.
• Manual segmentation of tumors on computed tomography (CT) scans is time-consuming and can be variable.

Purpose of the Study:

• To evaluate the clinical relevance and accuracy of a semiautomatic CT-based segmentation method using 3D-Slicer for NSCLC.
• To compare the performance of 3D-Slicer segmentation against manual delineations and pathological measurements.

Main Methods:

• A competitive region-growing algorithm in 3D-Slicer was used for semiautomatic tumor segmentation in 20 NSCLC patients.
• Three observers performed 3D-Slicer segmentations twice, compared to manual slice-by-slice delineations by five physicians.
• Tumor contours were validated against macroscopic tumor diameter from pathology as the gold standard.

Main Results:

• 3D-Slicer segmentations showed high agreement (overlap fractions > 0.90) with manual delineations.
• The semiautomatic method resulted in lower volume variability (p=0.0003) and smaller uncertainty areas (p=0.0002).
• 3D-Slicer segmentations strongly correlated with pathology (r=0.89), indicating high accuracy.

Conclusions:

• Semiautomatic 3D-Slicer segmentation provides accurate and stable tumor contouring for NSCLC, outperforming manual methods.
• 3D-Slicer is a valuable tool for treatment decisions and high-throughput research like Radiomics, overcoming manual segmentation bottlenecks.