Diagnostic performance and quantitative characterization of PET/computed tomography in peritoneal deposits assessment.

AIM: The study aimed to analyze the diagnostic utility of fluoro-deoxy glucose [FDG PET/computed tomography (CT)] in assessing peritoneal deposits and to establish optimal quantitative metabolic thresholds for differentiating malignant from benign involvement, comparing it with conventional CT.

METHOD: A total of 202 patients who had been identified with various types of malignant tumors and who were suspected or confirmed to have peritoneal deposits underwent whole-body FDG PET/CT. Qualitative assessment and quantitative metabolic parameters, including standardized uptake value (SUVmax, SUVmean, and SUVpeak), metabolic tumor volume, and total lesion glycolysis, were evaluated. Results were validated against histopathology or clinical/radiological follow-up. Receiver operating characteristic curve analysis determined the discriminative power and optimal cut-off values.

RESULT: FDG PET/CT demonstrated superior diagnostic accuracy, identifying 181 true positive cases with an overall sensitivity of 98.4%. Only three false negative cases were reported in the primary PET/CT analysis; however, further histological correlation of all cases revealed that lower FDG avidity was notably associated with specific subtypes, such as mucinous and signet ring cell carcinomas. In contrast, conventional CT reported only 126 true positive cases and a high rate of false negative (28.7%, n = 58) with a specificity of 46.6% and sensitivity of 69.5%. Among metabolic metrics, SUVmax, SUVmean, and SUVpeak showed nearly the same diagnostic performance, with an area under the curve of (0.915, 0.916, and 0.910), respectively (P = 0.0001). These values indicate 'excellent' power in distinguishing malignant from benign lesions.

CONCLUSION

FDG PET/CT performs noticeably better than traditional CT in the identification and mapping of peritoneal deposits, especially when anatomical alterations are minor. The high discriminative power of SUV-based metrics, especially with a threshold of 1, provides an excellent discriminative power in distinguishing malignant from nonmalignant lesions.