Demographics:
The mean (SD) age was 66 (11) years. 204 (of 288, 71%) patients were male. Of the 288 tumors, 266 (92%) were hepatocellular carcinoma (HCC), 11 (4%) were intrahepatic cholangiocarcinoma (IHC), and 11 (4%) were biphenotypic tumors. Median (IQR) tumor volume was 65mL (14–268). Mean (SD) total administered Y90 activity was 2.8 (1.5) GBq. The treated area was distributed as follows: lobar 104 (of 288, 70%), segmental 39 (of 288, 14%), and divisional 45 (of 288, 16%). Table 1 highlights baseline demographics and patient characteristics, comparing those with and without ascites before TARE.
Table 1
, baseline characteristics of patient in the cohort, classified by if they developed ascites after TARE.
Baseline characteristics | WITHOUT new or worsening ascites after TARE (n = 167) | WITH New or worsening ascites after TARE (n = 121) | P-value |
|---|
Age – years | 66 (64–68) | 66 (64–68) | 0.435 |
Male:Female (percent male) | 108:59 (65%) | 96:25 (79%) | 0.007 |
Child-Pugh score | 5.5 (5.4–5.7) | 5.7 (5.7–5.9) | 0.971 |
Bilirubin | 0.75 (0.68–0.83) | 0.92 (0.83-1.00) | 0.005 |
Albumin | 3.85 (3.76–3.93) | 3.63 (3.55–3.72) | 0.002 |
ALBI score | -2.59 (-2.67 - -2.50) | -2.33 (-2.42 - -2.25) | 0.001 |
MELD | 9.3 (8.9–9.7) | 9.8 (9.2–10.3) | 0.945 |
Lobar Y90 delivery | 119 (71%) | 84 (69%) | 0.736 |
Mean total activity administered (GBq) | 2.74 (2.48-3.00) | 2.83 (2.59–3.08) | 0.696 |
Baseline portal vein thrombosis (main or branch) | 41 (25%) | 42 (35%) | 0.060 |
Liver Toxicity after TARE:
Following TARE, total bilirubin increased from 0.8 to 2.3 mg/dL (p < 0.001) and alkaline phosphatase increased from159 to 212 U/L (p < 0.001). There was no significant increase in aminotransferase levels (ALT from 51 to 88, p = 0.206; AST from 69 to 180, p = 0.154). Comparing patients with and without pre-existing ascites, there was no significant difference in the rise of bilirubin (0.6 vs 0.3 mg/dL, p = 0.32). In patients without new/worsening ascites after TARE, bilirubin increased from 0.7 to 1.8 mg/dL (paired t-test, p < 0.001), whereas in those with new/worsening ascites after TARE bilirubin increased from 0.9 to 3.0 mg/dL (paired t-test, p < 0.001). The magnitude of increase in bilirubin was greater in patients who developed new/worsening ascites (p = 0.01). Similarly, MELD score increased significantly in all patients who received TARE, but the rise was greater in those who developed new/worsening ascites (2.2 vs. 5.9, p < 0.001).
Ascites before and after TARE:
At baseline, ascites was present in 60 (of 288, 21%) patients, with a majority classified as mild in severity [44 (of 60, 73%) mild ascites; 16 (of 60, 27%) moderate-severe ascites]. After TARE, 121 (of 288, 42%) developed new onset or worsening ascites. Among these, 93 (of 121, 77%) represented new onset ascites, while 28 (of 121, 23%) involved worsening of pre-existing ascites. Of the 121 patients with new/worsening ascites, 41 (34%) were classified as mild and 80 (66%) as moderate-severe.
Illustrated by Fig. 1, there was no significant difference in the incidence of new/worsening ascites after TARE when comparing patients with and without pre-existing ascites. The development of new onset ascites was observed in 93 (of 228, 41%), while worsening of pre-existing ascites was observed in 28 (of 60, 47%) (p = 0.412).
In the patients who developed new/worsening ascites, 21 (of 121, 17%) experienced new onset hepatic encephalopathy, likely suggesting a more severe liver decompensation. Of these, 15 (of 21, 71%) had no ascites at baseline, while 6 (of 21, 29%) had pre-existing ascites.
Risk Factors for Ascites after TARE
Amongst the patients who developed new/worsening ascites, 96 (of 121, 79%) were male, and 25 (of 121, 21%) underwent prior lobar cTACE. Review of the entire cohort revealed PVT in 83 patients (of 288, 29%) before TARE, with branch PVT in 32 (of 288, 11%) and segmental PVT in 51 (of 288, 18%). There was no significant association between PVT and new/worsening ascites [PVT present at baseline versus absent: 42 (35%) and 79 (65%), p = 0.060]. Lobar TARE was performed in 203 cases (70%). Lobar delivery of TARE was not associated with the development of new/worsening ascites after treatment. [37 (of 85, 44%) non-lobar treatment vs. 84 (of 203, 41%) lobar treatment, p = 0.736].
Univariable and multivariable logistic regression analyses of baseline characteristics and the development of new/worsening ascites after TARE is presented in Table 2. Factors such as PVT type, tumor volume, extent of Y90 infusion (segmental, divisional, lobar), total activity administered, Child-Pugh and MELD scores, baseline creatinine, INR, and alkaline phosphatase were not significantly associated with the development of new/worsening ascites after TARE. Multivariable analysis showed that pre-existing ascites was not associated with worsening ascites after TARE. However, male gender (OR = 2.1, p = 0.01), baseline albumin (OR = 0.4, p = 0.003), and ALBI score (OR = 2.89, p < 0.001) were all significantly associated with new/worsening ascites after TARE.
Table 2
, Univariable and multivariable logistic regression analysis evaluating association between baseline characteristics and new/worsening ascites after TARE.
| | HR (95%CI) | P-value |
|---|
Univariable analysis |
|---|
Age – years | 0.99 (0.98-1.00) | 0.807 |
Sex (0 = male, 1 = female) | 0.46 | 0.005 |
Child-Pugh score | 1.24 | 0.088 |
Albumin | 0.43 | 0.001 |
Bilirubin | 1.8 | 0.017 |
ALBI score | 2.6 | < 0.001 |
MELD | 1.06 | 0.143 |
Pre-existing ascites | 1.26 | 0.410 |
Lobar Y90 delivery | 0.92 | 0.765 |
Total Y90 activity (GBq) | 1.03 | 0.888 |
Baseline portal vein thrombosis (0 = none, 1 = main portal, 2 = branch portal) | 1.03 | 0.768 |
Total tumor volume | 0.999 | 0.198 |
Multivariable analysis |
Age – years | 0.99 (0.97–1.02) | 0.770 |
Sex (0 = male, 1 = female) | 0.54 (0.3–0.9) | 0.032 |
ALBI score | 3.7 (1.7-8) | < 0.001 |
Pre-existing ascites | 1.4 (0.58–3.4) | 0.464 |
| TARE = transarterial radioembolization, ALBI score = Albumin-Bilirubin score. |
Child-Pugh and ALBI Score Stratification:
A majority (234, 86%) of patients had a baseline Child-Pugh A classification, while the remaining (38, 14%) were Child-Pugh B.
At baseline, 118 (of 288, 41%) patients were classified as ALBI Grade I, while 161 (of 288, 56%) were ALBI Grade II (9 patients had missing values that precluded ALBI calculation). Patients who developed new/worsening ascites after TARE had significantly higher ALBI scores (-2.59 vs. -2.34, p = 0.0001). Among those with new/worsening ascites, 34 (of 121, 28%) were baseline ALBI grade 1, whereas 85 (of 121, 70%) were ALBI grade 2 (p < 0.001).
Among patients with pre-existing ascites, 14 (of 60, 23%) classified as ALBI grade 1, while 45 (of 60, 75%) were ALBI grade 2 (p = 0.001). Within this group with pre-existing ascites, patients who experienced progression of ascites had significantly higher baseline ALBI grades [3 out of 14 (21%) in ALBI grade 1, versus 25 out of 45 (56%) in ALBI grade 2, p = 0.026].
Liver Transplant and Overall Survival
Patients with pre-existing ascites had a median overall survival of 11 months (IQR: 5–33), compared to 24 months (IQR: 14-not reached) for those without pre-existing ascites (Logrank p < 0.001). There was no statistically significant difference in median survival between patients with baseline mild versus moderate-severe ascites (15 months vs. 7 months, Logrank, p = 0.247). (Fig. 2)
Median overall survival was 27 months (IQR: 12-not-reached) for patients with baseline ALBI Grade 1 and 17 months (IQR: 5–46 months) for those with baseline ALBI Grade 2 (Logrank p = 0.001). In the subset of patients with pre-existing ascites, median overall survival was 27 months with baseline ALBI Grade I and 7 months with baseline ALBI Grade II (Logrank p = 0.001)
Cox regression showed a decreased survival in patients who developed new/worsening ascites after TARE (HR = 2.1, p < 0.001) (Fig. 3). Multivariable Cox regression showed that Child-Pugh class (HR = 2.3, p < 0.001), pre-existing ascites (HR = 2, p < 0.001), and ALBI score (HR = 1.5, p < 0.011) were significantly associated with higher mortality.
Following TARE, 41 (of 237, 17%) patients underwent liver transplant. This included 4 (10%) patients who had pre-existing ascites and 12 (29%) patients who experienced new/worsening ascites after TARE.