Table 1
Summary of the study population
| Characteristics | ||
|---|---|---|
| Patients | ||
| Age (years) | 59.9 ± 7.9 | |
| Gender | Female | 12 (40.0%) |
| Male | 18 (60.0%) | |
| HCV | 13 (43.3%) | |
| HCV + Alcohol | 7 (23.3%) | |
| Liver disease etiology | Alcohol | 4 (13.3%) |
| HBV | 1 (3.3%) | |
| Other | 5 (16.7%) | |
| Child-Pugh class | A | 10 (33.3%) |
| B | 20 (66.7%) | |
| BCLC stage | A | 17 (56.7%) |
| B | 13 (43.3%) | |
| Tumors | ||
| Baseline tumor size (mm) | 20.3 ± 10.1 | |
| Lobar | 16 (34.8%) | |
| Treatment selectivity | Segmental | 22 (47.8%) |
| Sub-segmental | 8 (17.4%) | |
| Baseline Lipiodol uptake (HU) | 328.3 ± 185.2 | |
| Lipiodol uptake at final evaluation (HU) | 205.0 ± 219.5 | |
| Washout (HU) | 136.9 ± 127.6 | |
| Follow-up time (months) | 5.6 ± 6.2 | |
| Tumor recurrence | Yes | 19 (41.3%) |
| No | 27 (58.7%) |
The numerical data were summarized as mean ± standard deviation and the categorical data were shown as frequency (percentage). BCLC = Barcelona clinic liver cancer; HBV = hepatitis B virus; HCV = hepatitis C virus; HU = Hounsfield unit
Table 2
Results of cox regression analysis to predict factors for tumor recurrence
| Predicting factor | Univariate model | Multivariate model* | ||
|---|---|---|---|---|
| Hazard ratio (CI) | p Value | Hazard ratio (CI) | p Value | |
| Lesion size (mm) | 0.70 (0.24-2.09) | 0.527 | - | - |
| Treatment selectivity (Segmental or subsegmental vs. lobar) | 0.19 (0.07–0.51) | 0.001 | 3.05 (0.62–15.05) | 0.171 |
| Baseline Lipiodol uptake (HU) | 0.37 (0.19–0.70) | 0.002 | 0.002 (0.0–0.087) | 0.001 |
| Presence of Lipiodol washout | NA** | NA | - | - |
| Washout rate (HU/month) | 3.37 (1.78–6.36) | < 0.0001 | 149.03 (11.20–1983.54) | < 0.0001 |
CI = confidence interval; HU = Hounsfield unit
Figure 1
Baseline Lipiodol uptake after TACE for HCC. Significantly higher Lipiodol uptake was noted in tumors without recurrence and in tumors treated selectively with TACE.
Figure 2
Baseline Lipiodol uptake and washout rate in relation to early tumor recurrence. Early recurring tumors demonstrated significantly lower baseline Lipiodol uptake and higher Lipiodol washout rates relative to tumors recurring after the median recurrence free interval (= 3 months) and tumors with no recurrence.
Figure 3
ROC curve showing the area under curve and optimal cutoff value of baseline Lipiodol uptake after TACE to predict tumor recurrence.
Figure 4
Comparison of the cumulative hazards of tumor recurrence over the follow-up period between baseline Lipiodol uptake above and below the identified threshold. Tumors with Lipiodol uptake < 270.2 HU demonstrated significantly higher hazards of recurrence after TACE.
Figure 5
ROC curve showing the area under curve and optimal cutoff value of Lipiodol washout rate after TACE to predict early tumor recurrence.
Figure 6
Comparison of the cumulative hazards of tumor recurrence over the follow-up period between Lipiodol washout rates above and below the identified threshold. Tumors with Lipiodol washout rates ≥ 37.8 HU/month demonstrated significantly higher hazards of recurrence after TACE.