Results
Over the 20-year study period, 501 adult HCC recipients with pretransplant LRT underwent LT with an overall median follow-up time of 27.3 months (15.7 months for recipients who died and 39.3 months for recipients alive at last follow-up).
Recipient, tumor, and treatment characteristics are shown in Table 1. The median age was 60 years (IQR: 55–65), and 74% were male. Hepatitis C was the most common underlying diagnosis (63%), followed by Hepatitis B (18%), alcoholic liver disease (8%), and nonalcoholic steatohepatitis (4%). The median laboratory MELD score was 12 (IQR: 9–15), BMI was 26 (IQR: 24–30), total cholesterol was 145 (IQR: 121–174), immediate pretransplant AFP was 12 (IQR: 5–45), maximum pretransplant AFP was 27 (IQR: 8–139), and NLR was 2.5 (IQR: 1.7–4.1).
The majority of recipients (71%) had 1 radiographic lesion, with 20%, 8%, and 2% having 2, 3, and 4 or more lesions. The median maximum and cumulative tumor diameters were 3 (IQR: 2.4–4) and 3.7 (IQR: 2.6–5); 80% of recipients were within MC, 13% were beyond MC and within UCSF criteria, and 7% were beyond UCSF criteria. Of recipients beyond MC but within UCSF criteria, 95% were successfully downstaged to MC, compared with only 30% of recipients outside UCSF criteria.
Of the 501 recipients, 272 (54%), 148 (29%), 53 (11%), and 28 (6%) underwent 1, 2, 3, and 4 or more pretransplant LRTs, including TACE (63%), TA (51%), and percutaneous ethanol ablation (5%). Recipients were categorized into 4 mutually exclusive LRT groups including TACE and not TA (48%), TA and not TACE (35%), both TACE and TA (15%), and percutaneous ethanol injection alone (2%). Pathological examination revealed that 35, 340, and 126 recipients demonstrated no tumor necrosis, partial tumor necrosis, or complete tumor necrosis (cPR), respectively.
Comparison of Recipients With and Without cPR
Demographic and laboratory variables of recipients with and without cPR are shown in Table 2. Compared with recipients without cPR, patients with HCC demonstrating cPR had significantly lower MELD scores (10.7 vs 12.0, P = 0.002), immediate (7 vs 14, P < 0.001) and maximum (15 vs 31, P = 0.034) pretransplant AFP, and were less likely to have post-LRT AFP equal to maximum AFP (3.0% vs 13.9%, P = 0.003). There were no significant differences in recipient age, sex, diagnosis, BMI, diabetes, hypertension, cholesterol, or NLR among the groups.
Radiographic and treatment factors are shown in Table 2. Compared with patients without cPR, recipients demonstrating cPR were significantly more likely to have only 1 lesion (84% vs 67%, P = 0.002), to have tumors within MC (89% vs 77%, P = 0.004) and UCSF (97% vs 91%, P = 0.037) criteria, lesser cumulative tumor diameters (3.5 vs 3.9 cm, P = 0.002), and LRT that included ablation (62% vs 47%, P = 0.022); and less likely to have LRT with only TACE (37% vs 52%, P = 0.022). On radiological assessment before LT, recipients without cPR were significantly less likely to have no viable tumor (43% vs 73%, P < 0.001) and more likely to have probable (26% vs 20%) or definite viable tumor (31% vs 6%, P < 0.001). There were no significant differences in the maximum tumor diameter or number of LRT among the groups.
Survival Outcomes
The overall, recurrence-free, and disease-specific survival for the entire cohort at 1, 3, and 5 years was 86%, 71%, and 63%; 84%, 67%, and 60%; and 97%, 90%, and 87% (Fig. 1). Recipients with cPR had significantly better 1-, 3-, and 5-year recurrence-free (92%, 79%, 73%, P = 0.008; Fig. 2A) and disease-specific (100%, 100%, 98%, P < 0.001; Fig. 2B) survival than patients with partial pathologic response (RFS 80%, 62%, and 55%, and disease-specific survival (DSS) 95%, 86%, and 83%) and no pathologic response (RFS 91%, 75%, and 65%, and DSS 100%, 90%, and 85%). The recurrence-free and disease-specific survival of recipients with and without cPR are shown in Figure 3. Compared with recipients without cPR (no/partial response), patients with cPR had significantly superior 1-, 3-, and 5-year recurrence-free (92%, 79%, 73% vs 81%, 63%, 56%, P = 0.006; Fig. 3A) and disease-specific (100%, 100%, 98.6% vs 96%, 89%, 86%, P < 0.001) survival (Fig. 3B).
(Enlarge Image)
Figure 1.
Kaplan-Meier overall survival, recurrence-free survival, and disease-specific survival of 501 HCC recipients undergoing liver transplantation after pretransplant locoregional tumor treatment.
(Enlarge Image)
Figure 2.
Kaplan-Meier posttransplant recurrence-free survival (A) and disease-specific survival (B) comparing HCC recipients with complete pathologic response, partial pathologic response, and no pathologic response to pretransplant locoregional therapy.
(Enlarge Image)
Figure 3.
Kaplan-Meier posttransplant recurrence-free survival (A), disease-specific survival (B), and adjusted HCC recurrence (C) comparing recipients with and without a complete pathologic response. LT indicates liver transplantation.
HCC Recurrence
Post-LT recurrence developed in 57 of 375 patients (15.2%) without cPR compared with only 3 of 126 patients (2.4%) with cPR (P < 0.001). When adjusting for the risk of non–HCC-related mortality, recipients without cPR (no/partial response) had a 5-year estimated risk of recurrence of 18% compared with only 2% in recipients with cPR (Fig. 3C). Of the 3 cPR recipients with HCC recurrence, there was only 1 disease-specific death at 41.2 months after LT in a patient who developed portocaval metastases 18.5 months after transplant. One patient with cPR developed lung metastases at 15.2 months after LT and is alive at 22 months (last follow-up). The third cPR patient developed an isolated left lower lung metastasis 105 months after transplant, underwent resection, and is now disease-free for an additional 102 months.
Multivariate Analysis
Multivariate modeling identified 5 independent predictors of cPR (Table 3). These include the final radiologic assessment before LT, showing possible viable tumor (OR: 2.8, 95% CI: 1.4–4.2, P = 0.003) and no viable tumor (OR: 5.0, 95% CI: 2.4–10, P < 0.001) compared with definite viable tumor, post-LRT AFP less than pretransplant maximum AFP and/or 10 or less (OR: 3.03, 95% CI: 1.3–7.1, P = 0.012), days from last LRT to LT (OR: 1.56 per SD increase, 95% CI: 1.2–2.0, P < 0.001), lab MELD (OR: 1.31 per SD decrease, 95% CI: 1.01–1.7, P = 0.045), and radiological maximum tumor diameter (OR: 1.25 per log SD decrease, 95% CI: 0.98–1.6, P = 0.069). The model had an excellent ability to predict cPR with a C statistic of 0.75 (Fig. 4).
The effect of the number and type of LRT in achieving cPR was evaluated in a separate univariate and multivariate analysis that controlled for recipient MELD, pretransplant AFP, and radiological cumulative tumor diameter (Table 4). When controlling for LRT group (TACE not TA, TA not TACE, TACE and TA), an increasing number of treatments tended to decrease the likelihood of achieving cPR, as demonstrated by significantly lower OR with increasing treatments (trend P = 0.024). Furthermore, analysis of 8 unique groups comprising LRT number and modality revealed a general trend for decreased odds of achieving cPR with increasing number of treatments, in particular, TACE, even after controlling for cumulative tumor diameter. cPR was most likely in recipients receiving 2 treatments comprising ablation and TACE (univariate OR: 3.47, multivariate: 3.19) or 1 TA (univariate OR 2.59, multivariate: 1.98), and least likely in recipients receiving 3 or more TACE without TA. A subgroup analysis of recipients with only 1 lesion (n = 355) yielded very similar results, with the highest likelihood of cPR in recipients receiving a combination of TACE and TA (1 TACE and 1 TA, OR: 3.96, P = 0.056) even after controlling for recipient AFP, MELD, and cumulative tumor diameter.