Increase in incidence of neuroendocrine tumors (NETs) has been attributed in part to the availability of sensitive diagnostic modalities, such as 68Ga-DOTA-peptide PET/CT. However, it suffers from problems such as obscurement of tracer-avid lesions by physiological gut activity and collapsed gut lumen. Contrast-enhanced CT and CT enterography (CTE) do not have these drawbacks.
The aim of this study was to compare the diagnostic performances of contrast-enhanced CT + CTE and the 68Ga-DOTA-peptide PET/noncontrast CT in GEP-NETs.
Fifty-six patients (mean age, 57.8 ± 13.3 years [male:female, 1.95:1]), with histopathologically proven gastroenteropancreatic NETs, who had undergone both 68Ga-DOTANOC-PET/NCCT (60 minutes, post–IV injection of 111–185 MBq) and contrast-enhanced CT (CECT) + CTE (using 1.5–2 L isotonic mannitol solution and 1–2 mg/kg of IV contrast), were retrospectively selected. Twenty-three patients had been referred for identification of primary lesions and 33 for staging/restaging. The scans were independently evaluated by 2 blinded physicians, who documented the number and site of lesions, with reporting confidence (3 = high confidence, 2 = equivocal confidence, 1 = low confidence). Reference standard was created using clinical, biochemical, and imaging parameters (ie, uptake and contrast enhancement), along with corroboration from previous or follow-up scans. Finally, PET images coregistered to the CECT + CTE were independently evaluated for any additional benefit.
The numbers of primary lesions detected by CECT + CTE and PET/CT were 69 and 57, respectively. Lesion-wise sensitivities for patients with unknown primary in CECT + CTE and PET/CT were 57.7% (95% confidence interval [CI], 39.0%–74.5%) and 71.4% (95% CI, 52.9%–84.7%), respectively. Corresponding numbers in patients who had come for staging/restaging were 73.2% (95% CI, 58.1%–84.3%) and 73.8% (95% CI, 58.9%–84.7%). Lesions missed in CECT + CTE were gastrointestinal (n = 14), lymph nodes (n = 25), mesenteric (n = 1), and pancreatic (n = 7), whereas corresponding numbers for PET/CT were 14, 5, 3, and 2. Contrast-enhanced CT + CTE showed more false-positives (n = 26) than PET/CT (n = 9). Lesions missed by CECT + CTE were smaller than detected lesions (median, 9.7 mm [interquartile range, 7.5–31.1] vs 17.7 mm [interquartile range, 12.2–30.0]; P = 0.062), and lesions missed by PET had significantly lower tumor/background (liver) SUVmax ratio (median, 1.3 [interquartile range, 0.6–3.8] vs 4.7 [interquartile range, 2.7–10.8]). The ratio of true-positives to false-positives dropped markedly, when reporting confidence in CECT + CTE was low (4/15 [for rating 1 or 2] vs 93/11 [rating 3]). Corresponding numbers for PET/CT were (40/7 [for rating 1 or 2] vs 80/2 [rating 3]). Combination of these 2 modalities would have increased the lesion-wise sensitivities in patients with unknown primaries to 89.7% (95% CI, 73.6%–96.4%) and the confidence rating of soft tissue lesions to predominantly high (134 lesions rated 3, and 10 rated 1 or 2).
PET/CT is a sensitive modality for staging and restaging well-differentiated NETs. Use of CECT + CTE as a complementary modality in patients with uncertain uptake or high clinical suspicion of gastroenteropancreatic NETs should be considered, as it improves the lesion detection and reporting confidence.