Table 2. ¹⁸F-FDG-PET for predicting outcome after curative liver resection for HCC.

Authors	n	Stratification of subsets	Major study results
Torizuka et al.59	17	--------------------	Pre-resection SUV was 6.89 ± 3.39 in low grade and 3.21 ± 0.58 in high grade tumors (p < 0.005).
Kobayashi et al.60	60	High (≥3.2) vs. low (<3.2) SUVmax	Sensitivity and specificity of SUVmax ≥3.2 for predicting MVI were 77.8% and 74%. It increased to 88.9% and 82.4% by combining SUVmax with lens culinaris agglutinin a-reactive AFP.
Baek et al.61	54	Low (<6.36) vs. high (≥6.36) TMR	TMR ratio on pre-resection 18F-FDG PET correlated with MVI (p = 0.005) and tumor differentiation (p = 0.002). TMR ≥6.36 almost reached statistical significance in multivariate analysis for predicting HCC relapse (p = 0.061).
Ochi et al.62	89	High (≥8.8) vs. low (<8.8) SUVmax	SUVmax correlated significantly with tumor distance to microsatellite lesion pattern (R = 0.57; p < 0.0001). SUVmax was identified as an independent predictor of microsatellite distance >1 cm (HR = 1.60; 95%CI:1.23–2.26; p = 0.002) and extrahepatic HCC recurrence (HR = 1.24; 95%CI 1.01–1.55; p = 0.033).
Hatano et al.63	31	High (>2) vs. low (<2) SUV ratio	Overall 5-year survival rate was 63% in the high and 29% in the low SUV ratio subsets (p = 0.006). SUV ratio correlated significantly with tumor-related mortality (p = 0.001), tumor number (p = 0.002), tumor size (p = 0.001), vascular invasion (p = 0.005) and capsule invasion (p = 0.001). It did not remain as an independent prognostic factor for overall survival in multivariable analysis.
Seo et al.64	70	Low (<5) vs. high (≥5) SUV Low (<2) vs. high (≥2) TNR	Overall and recurrence-survival rates were significantly lower in the high than in the low FDG uptake groups (p = 0.002; p = 0.0005 for SUV; p = 0.001; p = 0.0002 for TNR). TNR but not SUV was identified as an independent predictor of postoperative recurrence (HR = 1.3; 95%CI 1.03–1.62; p = 0.03) and overall survival (HR = 1.6; 95%CI 1.07–2.38; p = 0.02).
Han et al.65	298	Low (<3.5) vs. high (>3.5) SUV	Preoperative SUV >3.5 was identified as an independent predictor of high grade tumor (HR = 3.305; 95%CI: 1.214–8.996; p = 0.019), tumor recurrence (HR = 2.025; 95%CI: 1.046–3.921; p = 0.036), and overall survival (HR = 7.331; 95%CI: 2.182–24.630; p = 0.001).
Ahn et al.66	93	Low (<4) vs. high (≥4) SUVmax Low (<2) vs. high (≥2) TNR	SUVmax and TNR correlated significantly (p < 0.001) with poor tumor differentiation. SUVmax ≥4 and TNR ≥2 were significant predictors for early recurrence-free survival (p = 0.026; p = 0.015) and overall survival (p = 0.005; p = 0.013). However, PET was no independent prognostic factor.
Kitamura et al.67	63	Low (<2) vs. high (≥2) TNR	TNR ≥2 was identified as an independent predictor for time interval to HCC recurrence. It was significantly lower in patients with recurrence beyond 1 year (4.4 ± 1.6; p < 0.05) or no recurrence (3.8 ± 1.5; p < 0.01) compared to those with early (within 1 year) tumor relapse (8.4 ± 6.3). Apart from that, TNR was identified as an independent prognostic factor for recurrence patterns according to the MC. It was significantly lower in patients developing tumor relapse within the MC (1.9 ± 1.6; p < 0.05) or no recurrence (1.3 ± 0.5; p < 0.01) compared to patients with tumor recurrence exceeding the MC (2.9 ± 2.6).

Abbreviations: ¹⁸F-FDG, ¹⁸F-fludeoxyglucose; CI, confidence interval; HCC, hepatocellular carcinoma; HR, hazard ratio; MC, Milan criteria; MVI, microvascular invasion; PET, positron emission tomography; SUV, standard uptake value; TMR, tumor-to-muscle ratio; TNR, tumor-to-nontumor uptake ratio.