Diagnostic Accuracy of 18F-FDG PET/CT in Detection of Local Recurrence in Rectal Cancer and the Added Value of Dual Time Point Scanning

Abstract

Objectives: To assess diagnostic accuracyof FDG PET/CT and the added value ofdual time point PET/CT (DTP) in detectionof local recurrence (LR) in patients withrectal cancer (RC). Methods: Patients (n =50, 41 males and 9 female, mean age 52 ±11 years). All patients underwent resection± chemotherapy and/or radiotherapy. 37patients were suspicious for LR on contrastenhanced CT (ce CT). All patientsunderwent whole body FDG PET/CT scan.In 18 patients 2 hours delayed pelvicPET/CT images were done. SUVmax cutoff of 3.0 was set to differentiate benignfrom malignant lesions based on ROCanalysis. Suspicious pelvic lesions werecorrelated with biopsies in 28 patients(56%) and with clinical and/or imagingfollow-up (FDG PET/CT, CT or MRI) in22 patients (44%). Sensitivity, specificity,positive and negative predictive values, andaccuracy in detection of LR using ce CTdata and following PET/CT werecalculated. Results: Nine patients had LR(18%). SUV max was higher in all patientswith LR. Sensitivity specificity, PPV,NPV, and accuracy for detecting recurrentlesions were significantly higher forPET/CT and PET/CT with tumor markersversus ce CT (p 0) in 4/18 patients withconfirmed LR (true positive) and revealedincrease in delayed SUV max (ΔSUV max>0) 4/18 with no evidences of LR (falsepositive) while 10/18 showed decrease inSUV max (ΔSUV max ≤0) in delayedimages with confirmed no LR (truenegative). The combined early SUVmaxand delayed increase in SUVmax revealedimprovement in overall accuracy comparedto either parameter alone. Conclusions:PET/CT has an excellent sensitivity and ahigher overall accuracy for detection oflocal rectal cancer recurrence whencompared to ce CT. Delayed PET/CT whenperformed is capable of improving thespecificity, PPV and accuracy of thePET/CT study.