To utilize deep learning for the automated segmentation of peritoneal cancer index (PCI) areas in CT scans to enhance assessment accuracy and minimize interobserver variability.
Approach:
Data Collection: Retrospective imaging and clinical data from patients diagnosed with gastric, ovarian, or colorectal cancer were collected, focusing on CT scans and subsequent diagnostic laparoscopy (DLS).
Segmentation Methodology: Deep learning methods, specifically nnU-Net and Swin UNETR, were employed to segment the abdomen into 13 regions defined by the radiological PCI (rPCI) scoring system.
Interobserver Agreement Evaluation: Interobserver agreement was quantified by having multiple annotators independently annotate a separate set of scans, with discrepancies resolved by an expert radiologist.
Key Findings:
This study introduces the first deep learning method for segmenting rPCI regions.
Automated segmentation is designed to reduce interobserver variability and decrease assessment time.
The method incorporates geometric boundary definitions of rPCI through deterministic post-processing.
Interpretation:
The introduction of automated segmentation for rPCI regions may enhance the standardization and accuracy of PM assessments in clinical settings.
Limitations:
The study is based on a dataset of 62 CT scans, which may limit the generalizability of the findings.
Annotations were conducted by a limited number of clinical researchers, potentially affecting the robustness of the results.
Conclusion:
Automated segmentation of rPCI regions using deep learning represents a significant advancement in the noninvasive assessment of peritoneal metastases.
by Pieter C. Gort, Lotte J. S. Fleurkens-Ewals, Lenah D. Kampmeijer, Anna F. van Herwijnen, Marion W. Tops-Welten, Cris H. B. Claessens, Joost Nederend, Ignace H. J. T. De Hingh, Max J. Lahaye, Misha D. P. Luyer, Fons van der Sommen
Pancreatic adenocarcinoma, the most common form of pancreatic cancer, accounts for roughly 90 to 95 percent of cases. It's notoriously difficult to catch early, because symptoms often don't appear until the disease has already spread
