- Email: [email protected]
Impact of the Interval between Transarterial Chemoembolization Sessions on Survival in Patients with Unresectable Hepatocellular Carcinoma
CLINICAL STUDY
Impact of the Interval between Transarterial Chemoembolization Sessions on Survival in Patients with Unresectable Hepatocellular Carcinoma Hyung-Don Kim, MD, Jihyun An, MD, Jin Hyoung Kim, MD, Dong Il Gwon, MD, Ji Hoon Shin, MD, Gi-Young Ko, MD, Hyun-Ki Yoon, MD, Kyu-Bo Sung, MD, Kang Mo Kim, MD, and Han Chu Lee, MD
ABSTRACT Purpose: To evaluate clinical impact of different intervals between multiple transarterial chemoembolization sessions in patients with unresectable hepatocellular carcinoma (HCC). Materials and Methods: A retrospective cohort study of 305 consecutive patients with HCC who underwent multiple sessions of on-demand transarterial chemoembolization by two independent physicians with different management policies in terms of transarterial chemoembolization interval was performed; 180 patients had intervals between the first and second transarterial chemoembolization session of o 60 days (short-interval group), and 125 patients had transarterial chemoembolization intervals of Z 60 days (conventional-interval group). Results: The short-interval group had more cases of advanced-stage HCC, less favorable response to transarterial chemoembolization, and higher likelihood of having Child-Pugh class A. The short-interval group underwent more transarterial chemoembolization sessions (6.6 vs 5.5, P ¼ .011), although the total number of admissions and total hospital stay were similar to the conventional-interval group. Overall survival was similar in the two groups in the full and the propensity score–matched cohorts. Although the overall survival of patients with Child-Pugh class A was comparable between the two groups in the full and propensity score–matched cohorts, the short-interval group showed inferior survival (P ¼ .005) and a nonsignificant trend toward inferior survival (P ¼ .117) in the full and propensity score–matched cohorts, respectively, for patients with Child-Pugh class B. Conclusions: Transarterial chemoembolization interval did not affect survival outcomes of patients with Child-Pugh class A. A shorter transarterial chemoembolization interval showed a nonsignificant trend of adversely affecting survival for patients with Child-Pugh class B.
ABBREVIATIONS CI = confidence interval, HCC = hepatocellular carcinoma, HR = hazard ratio, INR = international normalization ratio, mRECIST = modified response evaluation criteria in solid tumor
From the Departments of Internal Medicine (H.-D.K., J.A., K.M.K., H.C.L.) and Radiology (J.H.K., D.I.G., J.H.S., G.-Y.K., H.-K.Y., K.-B.S.), Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 138-736, Republic of Korea. Received August 18, 2015; final revision received and accepted December 4, 2015. Address correspondence to K.M.K.; E-mail: [email protected] None of the authors have identified a conflict of interest. & SIR, 2016 J Vasc Interv Radiol 2016; XX:]]]–]]] http://dx.doi.org/10.1016/j.jvir.2015.12.005
Transarterial chemoembolization is the most widely used treatment modality for intermediate-stage hepatocellular carcinoma (HCC) (1–3). However, transarterial chemoembolization using iodized oil has not been standardized, and its application largely depends on the clinical decision of each physician, with institutions showing variations with techniques, such as different chemotherapeutic agents (eg, doxorubicin and cisplatin); embolic materials and doses; and retreatment strategies, including differences in the interval between transarterial chemoembolization sessions (4–7).
2
’
Multiple Transarterial Chemoembolization Sessions and HCC
The interval between consecutive transarterial chemoembolization sessions is recommended to be 2–4 months in the current guidelines, whether it is performed on-demand or not (2,3). However, this recommended transarterial chemoembolization interval is not evidence based, and appropriate studies of the optimal interval between transarterial chemoembolization sessions are unavailable. Furthermore, intervals of Z 2 months between transarterial chemoembolization sessions may be suboptimal in terms of controlling tumor progression (7). In the present study, the clinical impact of the interval between transarterial chemoembolization sessions on overall patient survival and safety was evaluated.
MATERIALS AND METHODS Study Subjects This study was approved by the institutional review board of the institution, and the requirement for informed consent from the patients was waived. The study population was derived from a historical cohort of 555 consecutive patients with HCC who, in conjunction with concomitant and subsequent treatment with radiation therapy for portal vein invasion and sorafenib, underwent repeated transarterial chemoembolization sessions by two independent physicians (K.M.K., H.C.L.) with different management policies in terms of the transarterial chemoembolization schedule between January 2006 and December 2012. As a baseline study, all patients underwent four-phase dynamic computed tomography (CT), and the diagnosis of HCC was made according to algorithmic guidelines (1–3). Patients with a complete response, based on modified Response Evaluation Criteria in Solid Tumor (mRECIST), after the first transarterial chemoembolization treatment were excluded because a complete response can directly affect the interval between transarterial chemoembolization sessions (n = 132) (5,6,8,9). Patients with main portal vein invasion or bilateral involvement of the first branch portal vein (n = 57), extrahepatic metastasis (n = 54), or other concomitant malignancy (n = 7) were also excluded. Finally, 305 patients were included in the analyses and subdivided into two groups according to the interval between the first and second transarterial chemoembolization session: a shortinterval group (first transarterial chemoembolization interval o 60 d) and a conventional-interval group (first transarterial chemoembolization interval Z 60 d).
Transarterial Chemoembolization Procedure In accordance with the conventional transarterial chemoembolization protocol applied in our liver center (10), superior mesenteric arteriography and common hepatic arteriography were performed to assess the overall anatomy, tumor burden, and portal vein patency, based on intrahepatic HCC status assessed by hepatic
Kim et al
’
JVIR
dynamic CT images. Cisplatin at 2 mg/kg body weight (Cisplan; Dong-A Pharm Co, Seosan, Korea) was infused into the lobar hepatic artery for 15 minutes without an injection of embolic particles. After vascular catheterization with a microcatheter placed selectively or superselectively into the distal tumor-feeding artery, an emulsion of 2–20 mL of iodized oil (Lipiodol Ultra-Fluide; Laboratoires Guerbet, Aulnay-sous-Bois, France) and cisplatin in a 1:1 ratio was administered into the target arteries. Embolization of the arterial tumor feeders was then performed using a Gelfoam slurry (Gelfoam; Upjohn, Kalamazoo, Michigan) until arterial flow stasis was achieved. The Gelfoam slurry was made manually by cutting up a 70 mm 50 mm 10 mm Gelfoam sponge. There was no change in the transarterial chemoembolization protocol throughout the study period. Principally, transarterial chemoembolization was performed on-demand (5,6,8) every 4–16 weeks if there was evidence of residual viable tumor on follow-up CT imaging. Transarterial chemoembolization was not implemented when there was no residual tumor. The decision not to perform further transarterial chemoembolization procedures was made using the following criteria: (i) deterioration in liver function, (ii) transarterial chemoembolization was considered ineffective, (iii) ascites worsened, (iv) severe vascular invasion that made additional transarterial chemoembolization impossible, and (v) other technical problems or contraindications for transarterial chemoembolization.
Study Outcomes and Follow-up The primary outcome of the current study was all-cause mortality. The index date was defined as the date of the first transarterial chemoembolization session. Patients were followed up from the index date to death or the last follow-up date (August 31, 2014). Overall survival was compared for the entire study population and the propensity score–matched cohort. Thereafter, subgroup analysis was performed according to the Child-Pugh class, tumor size, and presence or absence of portal vein invasion. Four-phase dynamic CT was performed at the start and 1 month after transarterial chemoembolization, and the tumor response to transarterial chemoembolization was evaluated by mRECIST, which exhibits superior performance for determining tumor response (11,12). Dynamic CT scans subsequently were principally performed 1 month after a given course of transarterial chemoembolization for the patients with remaining viable tumors. In patients with no viable tumor on dynamic CT images after repeated transarterial chemoembolization, follow-up dynamic CT images were taken at intervals of 2–3 months until tumor recurrence was noted. The total number of transarterial chemoembolization sessions performed in each patient was recorded. To evaluate procedure-related mortality, death within 1 month after the last transarterial chemoembolization session was
Volume XX
’
Number X
’
Month
’
2016
3
analyzed. Regular follow-up of liver function was performed and included the international normalization ratio (INR). As a measure of medical cost, the number of hospital admissions and total duration of hospital stay during serial follow-up were recorded, including not only admissions for the transarterial chemoembolization sessions but also any admissions for supportive care of patients with cancer-related symptoms or deteriorated liver function.
by a larger size of the largest HCC nodule (7.3 cm vs 5.4 cm) and higher prevalence of portal vein invasion (31.7% vs 10.4%) and therefore more advanced Barcelona Clinic Liver Cancer stages (0, 1.7% vs 2.4%; A, 28.3% vs 40.0%; B, 38.3% vs 47.2%; C, 31.7% vs 10.4%). Serum αfetoprotein levels were not significantly different between the groups. After propensity score matching, 104 pairs of patients from the two groups were matched and showed comparable baseline variables (Table 1).
Statistical Analysis
Transarterial Chemoembolization Interval and Hospitalization
Student t test or Mann-Whitney U test was used to compare quantitative variables, and χ2 or Fisher exact test was used to compare qualitative variables. Student t test was used when the variables showed a normal distribution according to visual histogram inspection, and MannWhitney U test was used for variables showing a skewed distribution. Multivariate analysis was performed with Cox proportional hazards model analysis to assess factors associated with overall survival. Overall survival was compared using the log-rank test. To minimize any potential selection bias and confounding factors, propensity score matching was performed. Variables used to derive the propensity score included age; sex; number of tumors; diameter of the largest tumor; presence of portal vein invasion; INR; platelet count; Child-Pugh score; and serum levels of aspartate aminotransferase, alanine aminotransferase, albumin, total bilirubin, and α-fetoprotein. The propensity score was estimated nonparametrically using the matchit function in the R package MatchIt. Matching was performed using the nearest neighbor matching method, using a caliper width of 0.2 multiplied by the standard deviation for the linearly transformed propensity scores (logit transformation). All statistical analyses were performed using IBM SPSS Statistics for Windows, version 20 (IBM Corporation, Armonk, New York) or R software version 3.1.1 (R Foundation for Statistical Computing, Vienna, Austria). All reported P values were two-sided, and a P value o .05 was considered significant.
RESULTS Clinical Characteristics The present study comprised 305 patients with HCC; 180 belonged to the short-interval group, and 125 belonged to the conventional-interval group. The mean age and sex ratios were similar between the two groups. Baseline biochemical results were comparable between the two groups, including the serum alanine aminotransferase level (45.7 IU/mL vs 45.2 IU/mL), albumin level (3.5 g/ dL vs 3.4 g/dL), total bilirubin level (1.1 mg/dL vs 1.1 mg/dL), and INR (1.10 vs 1.12). However, the platelet count was significantly higher (166.1K/mm3 vs 129.8K/ mm3) and Child-Pugh class A were more frequent in the short-interval group. The short-interval group had tumors of more locally advanced stages, characterized
The median interval between the first and second transarterial chemoembolization session was 51 days (interquartile range, 38–70 d). As shown in Figure 1, the transarterial chemoembolization interval was significantly different between the two physicians because they had different transarterial chemoembolization management policies: physician A, a median of 70 days (interquartile range, 61– 82 d); physician B, a median of 42 days (interquartile range, 36–55 d; P o .001). The mean total number of transarterial chemoembolization sessions was significantly higher in the short-interval group (6.6 vs 5.5, P ¼ .011). However, the total number of hospital admissions (8.7 vs 7.8, P ¼ .066) and total hospital stay (55.6 d vs 54.5 d, P ¼ .767) were comparable between the two groups. The short-interval group showed less partial response (73.9% vs 89.6%) but more stable disease (20.0% vs 8.0%) and progressive disease (6.1% vs 2.4%) after the first transarterial chemoembolization session according to mRECIST (P ¼ .003). After propensity score matching, responses to transarterial chemoembolization were not statistically different between the two groups (P ¼ .157), and other transarterial chemoembolization and hospitalization parameters showed similar patterns in the full and the propensity score–matched cohorts (Table 2).
INR Levels at Specific Time Points after First Transarterial Chemoembolization The mean INR level of the study subjects showed a trend toward aggravation and was similar between the shortinterval and conventional-interval groups at 3, 6, and 12 months after the first transarterial chemoembolization session. The same was true for the propensity score– matched cohort (Table 3).
Factors Associated with Survival Cox regression analyses revealed that tumor diameter 4 5 cm was a significant independent predictor of patient survival (adjusted hazard ratio [HR], 1.60; 95% confidence interval [CI], 1.21–2.12; P ¼ .001), along with partial response based on mRECIST (HR, 0.63; 95% CI, 0.46–0.87; P ¼ .005) and serum α-fetoprotein 4 400 ng/ mL (HR, 1.47; 95% CI, 1.11–1.93; P ¼ .007) (Table 4). Preserved liver function characterized by Child-Pugh class A showed a statistically marginal association with overall survival (HR, 0.72; 95% CI, 0.51–1.03; P ¼ .069).
4
’
Multiple Transarterial Chemoembolization Sessions and HCC
Kim et al
’
JVIR
Table 1 . Baseline Characteristics of Study Patients Full Cohort
Characteristics Age (y)* Male sex
Short-Interval Group (n ¼ 180) 56.1 ⫾ 10.4 156 (86.7%)
Conventional-Interval Group (n ¼ 125) P Value 56.7 ⫾ 9.5 106 (84.8%)
Etiology HBV (%) HCV (%) Others (%) ALT (IU/mL)* Albumin (g/dL)* Total bilirubin (mg/dL)*
Child-Pugh class A (%) B (%)
.654 .645
47 (26.1%) 45.7 ⫾ 28.8 3.5 ⫾ 0.5 1.1 ⫾ 1.0
183 (11–4,555)
94 (75.2%) 8 (6.4%) 23 (18.4%) 45.2 ⫾ 7.2 3.4 ⫾ 0.5
Conventional-Interval Group (n ¼ 104) P Value
57.4 ⫾ 10.9 90 (86.5%)
56.4 ⫾ 9.9 89 (85.6%)
68 (65.4%) 11 (10.6%)
80 (76.9%) 4 (3.8%)
1.1 ⫾ 0.6 1.12 ⫾ 0.11 129.8 ⫾ 66.4 74 (14–684)
.858 .148 o .001 .083
45.1 ⫾ 26.6 3.5 ⫾ 0.5
20 (19.3%) 43.9 ⫾ 24.7 3.5 ⫾ 0.5
.724 .870
1.1 ⫾ 0.6
1.1 ⫾ 0.6
.516
1.10 ⫾ 0.10 138.2 ⫾ 59.3
1.10 ⫾ 0.11 134.2 ⫾ 66.8
.504 .650
136 (11–943)
82 (17–581)
93 (89.4%)
87 (83.7%)
.048 160 (88.9%)
101 (80.8%)
20 (11.1%)
24 (19.2%)
2.4 ⫾ 1.7 7.3 ⫾ 4.3
2.7 ⫾ 2.6 5.4 ⫾ 3.7
.270 o .001
Portal vein invasion (%)
57 (31.7%)
13 (10.4%)
o .001
11 (10.6%)
17 (16.3%)
2.6 ⫾ 2.6 5.4 ⫾ 3.8
2.6 ⫾ 1.8 5.6 ⫾ 3.4
9 (8.7%)
13 (12.5%)
o .001 3 (1.7%)
3 (2.4)%
3 (2.9%)
3 (2.9)%
50 (40.0%)
41 (39.4%)
40 (38.5%)
B (%) C (%)
69 (38.3%) 57 (31.7%)
59 (47.2%) 13 (10.4%)
51 (49.0%) 9 (8.7%)
48 (46.2%) 13 (12.5%)
24.4 (12.7–38.4)
.412
4 .999 .660 .367 .842
51 (28.3%)
23.6 (14.8–34.6)
.958 .223
A (%)
Follow-up (mo)†
.498 .841 .091
25 (24.0%) .883 .196
No. HCC nodules* Size of HCC nodule (cm)* BCLC stage 0 (%)
Short-Interval Group (n ¼ 104)
.165 117 (65%) 16 (8.9%)
1.10 ⫾ 0.11 INR* Platelet count ( 1,000/mm3)* 166.1 ⫾ 98.7 α-Fetoprotein (ng/mL)†
PS-Matched Cohort
25.3 (14.8–34.6)
25.0 (12.5–37.5)
.821
Note–Data are presented as n (%) or mean ⫾ SD or median (interquartile range). ALT ¼ alanine aminotransferase; BCLC ¼ Barcelona Clinic Liver Cancer; HBV ¼ hepatitis B virus; HCC ¼ hepatocellular carcinoma; HCV ¼ hepatitis C virus; INR ¼ international normalized ratio; PS ¼ propensity score. n Student t test was used. † Mann-Whitney U test was used.
0.97–1.61; P ¼ .121) and propensity score–matched (HR, 1.00; 95% CI, 0.74–1.34; P ¼ .992) cohorts.
Survival Analysis
Figure 1. Graphic depiction of the interval between the first and second transarterial chemoembolization sessions according to physician.
A first transarterial chemoembolization interval of o 60 days was not a significant predictor of overall survival in univariate analysis of the full (HR, 1.23; 95% CI,
The median survival times for patients with shortened and conventional intervals were 28.5 months and 28.6 months, respectively. According to Kaplan-Meier analysis, patients with different transarterial chemoembolization intervals showed similar survival results in both the full (P ¼ .121) and the propensity score–matched (P ¼ .993) cohorts (Fig 2a, b). Patients were subdivided into Child-Pugh class A (n ¼ 261) and Child-Pugh class B (n ¼ 44) groups. Although the overall survival of patients with Child-Pugh class A was similar between the two groups (P ¼ .403), patients with Child-Pugh class B in the short-interval group showed inferior survival results (P ¼ .005) (Fig 3a, b). There were no significant survival differences between the two groups regardless of whether there was portal vein invasion or no portal vein invasion (P ¼ .290 or P ¼ .851, respectively) (Fig 3c, d). Patients with Child-Pugh class A showed similar survival outcomes in the propensity
Volume XX
’
Number X
’
Month
’
2016
5
Table 2 . Profiles of Transarterial Chemoembolization Sessions and Hospitalizations Full Cohort Short-Interval Group (n ¼ 180)
Characteristics
PS-Matched Cohort
Conventional-Interval Group (n ¼ 125)
P Value
Response to first transarterial chemoembolization
Short-Interval Group (n ¼ 104)
Conventional-Interval Group (n ¼ 104)
.003
P Value .157
based on mRECIST PR SD
133 (73.9%) 36 (20.0%)
PD
11 (6.1%)
Interval between first and second transarterial
112 (89.6%) 10 (8.0%)
83 (79.8%) 16 (15.4%)
3 (2.4%)
41.1 ⫾ 8.0
78.7 ⫾ 20.5
6.6 ⫾ 3.6
8.7 ⫾ 4.3 55.6 ⫾ 30.0
5 (4.8%)
93 (89.4%) 8 (7.7%) 3 (2.9%)
o .001
41.6 ⫾ 8.3
78.7 ⫾ 20.3
o .001
5.5 ⫾ 3.4
.011
6.4 ⫾ 3.3
5.4 ⫾ 3.3
.043
7.8 ⫾ 4.0 54.5 ⫾ 33.1
.066 .767
8.8 ⫾ 4.3 54.2 ⫾ 27.9
7.9 ⫾ 3.9 56.4 ⫾ 35.1
.092 .614
chemoembolization (d)* Total transarterial chemoembolization sessions* Total admissions* Total hospital duration (d)*
mRECIST ¼ modified Response Evaluation Criteria in Solid Tumor; PD ¼ progressive disease; PR ¼ partial response; PS ¼ propensity score; SD ¼ stable disease. n Student t test was used.
Table 3 . Comparison of INR at Specific Time Points after First Transarterial Chemoembolization Full Cohort
Baseline INR INR at 3 mo from first
PS-Matched Cohort
Short-Interval
Conventional-Interval
Short-Interval
Conventional-Interval
Group (n ¼ 180)
Group (n ¼ 125)
P Value
Group (n ¼ 104)
Group (n ¼ 104)
P Value
1.10 ⫾ 0.11 1.14 ⫾ 0.14
1.12 ⫾ 0.11 1.15 ⫾ 0.25
.148 .364
1.10 ⫾ 0.10 1.13 ⫾ 0.14
1.10 ⫾ 0.11 1.16 ⫾ 0.27
.504 .390
1.17 ⫾ 0.31
1.16 ⫾ 0.19
.818
1.15 ⫾ 0.19
1.16 ⫾ 0.20
.623
1.21 ⫾ 0.35
1.19 ⫾ 0.22
.611
1.17 ⫾ 0.24
1.17 ⫾ 0.20
.958
transarterial chemoembolization INR at 6 mo from first transarterial chemoembolization INR at 12 mo from first transarterial chemoembolization INR ¼ international normalized ratio; PS = propensity score. Note–Student t test was used.
score–matched cohort (P ¼ .628). For the patients with Child-Pugh class B in the matched cohort, there was a nonsignificant trend for inferior survival in the short-interval group (P ¼ .117) (Fig 4a–d). There was no difference in the proportions of patients who died within 30 days after the last transarterial chemoembolization session (1.1% [2/180] in the short-interval group and 1.6% [2/125] in the conventional-interval group, P 4 .99).
DISCUSSION Although there has been no appropriate study of the interval between consecutive sessions of transarterial chemoembolization, these sessions have conventionally been performed
with a median interval of 2 months (7); guidelines recommend transarterial chemoembolization intervals of 3–4 months (2) or 2–3 months (3). From the oncologic point of view, transarterial chemoembolization intervals of 4 2 months may be insufficient, particularly given that systemic chemotherapy is delivered every 3 weeks, because local progression or metastasis of viable tumors will almost always occur in that time (7). However, there has been concern that transarterial chemoembolization applied with an interval o 2 months might result in further deterioration of liver function and lead to poor survival outcomes as a result of liver failure. In the present study, the survival outcomes and safety profiles of patients with HCC undergoing repeated
6
’
Multiple Transarterial Chemoembolization Sessions and HCC
Kim et al
’
JVIR
Table 4 . Multivariate Regression Analysis of Overall Survival in Full Cohort Overall Mortality Univariate Analysis
Multivariable Analysis
Variables
HR
95% CI
P Value
HR
95% CI
P Value
Age (y) Male sex
0.99 0.95
0.98–1.00 0.65–1.40
.122 .807
— —
— —
— —
Etiology (HBV)
1.09
0.81–1.44
.603
—
—
—
Child-Pugh class A Multiple tumors (vs single)
0.74 1.08
0.52–1.04 0.83–1.42
.083 .565
0.72 —
0.51–1.03 —
.069 —
Tumor diameter Z 5 cm
1.85
1.41–2.42
o .001
1.60
1.21–2.12
.001
Presence of portal vein invasion α-fetoprotein 4 400 ng/mL
1.64 1.70
1.21–2.22 1.31–2.22
.002 o .001
— 1.47
— 1.11–1.93
— .007
PR based on mRECIST
0.60
0.44–0.82
.001
0.63
0.46–0.87
.005
Transarterial chemoembolization interval o 60 d
1.23
0.97–1.61
0.121
—
—
—
Note–Cox proportional hazards model was used for all analyses. CI ¼ confidence interval; HBV ¼ hepatitis B virus; HR ¼ hazard ratio; mRECIST ¼ modified response evaluation criteria in solid tumor; PR ¼ partial response.
Figure 2. Overall survival rate of the full cohort (a) and the propensity score–matched cohort (b).
on-demand transarterial chemoembolization on the basis of different intervals between transarterial chemoembolization sessions were evaluated. The study showed that there was no significant difference in overall survival between patients with transarterial chemoembolization intervals o 60 days and patients with transarterial chemoembolization intervals of 60 days or longer. The short-interval resulted in inferior survival outcomes only in the subgroup of patients with Child-Pugh class B in the full cohort, whereas patients with Child-Pugh class B in the propensity score–matched cohort showed only a trend for inferior survival. Also, INR levels at specific time points after transarterial chemoembolization were not different between the two interval groups. Despite higher numbers of transarterial chemoembolization sessions in the short-interval group than in the conventional-interval group, no increases in the total number of admissions and total hospital duration
were observed in the full and propensity score–matched cohorts. In addition, in the present study, aggressive transarterial chemoembolization with a shortened interval tended to be applied in patients with advanced HCC, which is characterized by larger nodules, more frequent invasion of major vessels, and a more unfavorable response to transarterial chemoembolization, although the liver function of these patients was relatively preserved. It is conceivable that short-interval transarterial chemoembolization might be better at controlling and curbing tumor progression, owing to the finding that overall survival was not inferior in the short-interval group with unfavorable tumor parameters. However, this effect did not translate into better survival in the multivariate and propensity score–matched analyses. In conjunction with the absence of concrete evidence that short-interval transarterial chemoembolization better
Volume XX
’
Number X
’
Month
’
2016
7
Figure 3. Overall survival rate in patients with Child-Pugh class A (a), with Child-Pugh class B (b), with portal vein invasion (c), and with no portal vein invasion (d) in the full cohort.
prevented tumor progression, there would be no strong rationale for recommending short-interval transarterial chemoembolization. In particular, the shortened transarterial chemoembolization interval led to poor survival outcomes in patients with Child-Pugh class B, suggesting that the risk of deterioration of liver function caused by a shortened transarterial chemoembolization interval outweighs the potential benefit of aggressive tumor control with frequent transarterial chemoembolization sessions in patients with Child-Pugh class B. Although a statistically significant difference was not reached in patients with Child-Pugh B in the propensity score– matched cohort, possibly because of the small number of patients included in the analysis, application of shortinterval transarterial chemoembolization in patients with Child-Pugh B still seems to be hazardous considering the trend for inferior survival outcomes. In terms of safety, serial INR levels after the first transarterial chemoembolization session and treatmentrelated mortality—described as death within 30 days of
undergoing transarterial chemoembolization—did not differ between the two study groups. In addition, the finding that short-interval transarterial chemoembolization involved more sessions of this treatment without significantly increasing the number of hospitalizations or duration of hospital stay may indicate the overall safety and potential clinical efficacy of a short-interval approach. We assume that better tumor control in the short-interval group may have decreased the number of admissions and duration of hospital stay for supportive care related to cancer progression. In contrast to previous concerns, these results imply that the shortened transarterial chemoembolization interval does not cause a further decline in liver function and increased patient mortality, especially in cases with preserved liver function. To the best of our knowledge, the current findings are novel because no relevant study has previously directly compared clinical outcomes according to the interval between transarterial chemoembolization procedures. A
8
’
Multiple Transarterial Chemoembolization Sessions and HCC
Kim et al
’
JVIR
Figure 4. Overall survival rate in patients with Child-Pugh class A (a), Child-Pugh class B (b), portal vein invasion (c), and no portal vein invasion (d) in the propensity score–matched cohort.
previous study by Ernst et al (8) compared survival outcomes between on-demand transarterial chemoembolization and transarterial chemoembolization performed every 2 months with a fixed schedule. In that study, transarterial chemoembolization was more frequently performed in the fixed-schedule group and showed inferior survival. However, the relatively narrower interval of the fixed-schedule group was predominantly induced by a longer interval between sessions for the on-demand transarterial chemoembolization caused by delaying the next transarterial chemoembolization session in patients with a complete response. Such patients were excluded from the current study. Furthermore, this relatively narrow interval was 2 months, which was used as a cutoff to discriminate between the short-interval and conventional-interval groups. Although no clinical benefit of short-interval transarterial chemoembolization in prolonging patient survival was demonstrated, the current study is clinically important because it provides evidence that frequent transarterial chemoembolization could be a contraindication in
patients with Child-Pugh class B and that more aggressive forms of transarterial chemoembolization could be considered in patients with preserved liver function, together with the finding that a short transarterial chemoembolization interval did not increase the number of hospitalizations and duration of hospital stay. Furthermore, the study results provide clinical evidence that the shortest time required for the recovery of liver function to perform the next transarterial chemoembolization session safely is o 60 days. Based on the current results, a future prospective study should try to confirm the efficacy and safety of short-interval transarterial chemoembolization regarding tumor progression and survival outcomes in selected patients with preserved liver function who need aggressive tumor control because of large tumor burden. In addition to the retrospective nature of the present study, there are several other limitations. First, the possibility of a selection bias remained. Short-interval transarterial chemoembolization tended to be applied to patients with local progression of tumors and preserved liver function, which might have directly affected study
Volume XX
’
Number X
’
Month
’
2016
outcomes. To overcome the effect of a selection bias and evaluate the effect of the transarterial chemoembolization interval on survival, statistical approaches, including propensity score–matched analysis, were implemented. After propensity score matching, all the baseline characteristics, including tumor parameters, were comparable, and the responses to transarterial chemoembolization were not statistically different between the two groups, providing a sound background to the comparison of the two groups in a nonrandomized setting. Second, the first transarterial chemoembolization interval may not fully represent the entire HCC management policy. Although the two physicians who treated study patients had clearly different management policies, multiple factors, including patient tolerability, liver function after transarterial chemoembolization sessions, and extent of remnant viable tumor, might have also played a role in determining when the subsequent transarterial chemoembolization session was performed. Furthermore, a mean value of the total intervals between transarterial chemoembolization could not be incorporated because transarterial chemoembolization was performed on-demand. However, considering the complexities of the application of transarterial chemoembolization for the treatment of HCC (4–7) and the current finding that the total number of transarterial chemoembolization sessions significantly differed according to the first interval between the transarterial chemoembolization sessions, the first transarterial chemoembolization interval may arguably be the only realistic parameter that reflects the management strategy regarding the interval between transarterial chemoembolization sessions. Third, patients with portal vein invasion were included in the analyses, even though portal vein invasion is a contraindication for transarterial chemoembolization under Western guidelines (1,2). However, evidence is accumulating regarding the efficacy of transarterial chemoembolization in patients with portal vein invasion (10,13–15). Furthermore, it was shown that there was no survival difference regarding the transarterial chemoembolization interval according to whether the patient had portal vein invasion. Fourth, as for the transarterial chemoembolization protocol, infusion of chemotherapeutic material through the lobar artery, which may have an additional antitumor effect because of possibly wider delivery of chemotherapeutic material, is not standard and requires scientific verification in future studies. However, it is unlikely that chemotherapy infusion at the lobar artery level affected the study results because this procedure was identically performed in both study groups. Although the follow-up principle of performing dynamic CT scanning 1 month after transarterial chemoembolization sessions was identical for both physicians, the follow-up intervals might not have been the same because of the differences in the frequency of transarterial chemoembolization. However, because unequal follow-up imaging was the consequence
9
rather than the cause of the different management policy, it is unlikely that follow-up imaging variations would affect the clinical outcome. Lastly, the number of study subjects with Child-Pugh class B was possibly too small to draw a conclusive inference. However, we believe that our data are still valuable in that we at least showed that short-interval transarterial chemoembolization in patients with Child-Pugh class B requires extreme caution and could be hazardous. In addition, no other data are currently available regarding this issue. In conclusion, patients with HCC undergoing repeated on-demand transarterial chemoembolization with an interval of o 60 days show similar overall survival rates without a significant increase in the number of hospitalizations and duration of hospital stay compared with patients with transarterial chemoembolization intervals of Z 60 days. Subgroup analysis showed that a transarterial chemoembolization interval of o 60 days led to a nonsignificant trend toward inferior survival outcomes for patients with Child-Pugh class B.
ACKNOWLEDGMENT We thank Dr. Seungbong Han (Department of Applied Statistics, Gachon University) for full support of the statistical analysis.
REFERENCES 1. Bruix J, Sherman M. Management of hepatocellular carcinoma: an update. Hepatology 2011; 53:1020–1022. 2. European Association for the Study of the Liver; European Organisation for Research and Treatment of Cancer. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012; 56:908–943. 3. Omata M, Lesmana LA, Tateishi R, et al. Asian Pacific Association for the Study of the Liver consensus recommendations on hepatocellular carcinoma. Hepatol Int 2010; 4:439–474. 4. Vogl TJ, Naguib NN, Nour-Eldin NE, et al. Review on transarterial chemoembolization in hepatocellular carcinoma: palliative, combined, neoadjuvant, bridging, and symptomatic indications. Eur J Radiol 2009; 72:505–516. 5. Cheng AL, Amarapurkar D, Chao Y, et al. Re-evaluating transarterial chemoembolization for the treatment of hepatocellular carcinoma: consensus recommendations and review by an international expert panel. Liver Int 2014; 34:174–183. 6. Terzi E, Golfieri R, Piscaglia F, et al. Response rate and clinical outcome of HCC after first and repeated cTACE performed “on demand”. J Hepatol 2012; 57:1258–1267. 7. Marelli L, Stigliano R, Triantos C, et al. Transarterial therapy for hepatocellular carcinoma: which technique is more effective? A systematic review of cohort and randomized studies. Cardiovasc Intervent Radiol 2007; 30:6–25. 8. Ernst O, Sergent G, Mizrahi D, Delemazure O, Paris JC, L’Hermine C. Treatment of hepatocellular carcinoma by transcatheter arterial chemoembolization: comparison of planned periodic chemoembolization and chemoembolization based on tumor response. AJR Am J Roentgenol 1999; 172:59–64. 9. Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment for hepatocellular carcinoma. Semin Liver Dis 2010; 30:52–60. 10. Kim GA, Shim JH, Yoon SM, et al. Comparison of chemoembolization with and without radiation therapy and sorafenib for advanced hepatocellular carcinoma with portal vein tumor thrombosis: a propensity score analysis. J Vasc Interv Radiol 2015; 26:320–329. e326. 11. Shim JH, Lee HC, Kim SO, et al. Which response criteria best help predict survival of patients with hepatocellular carcinoma following
10
’
Multiple Transarterial Chemoembolization Sessions and HCC
chemoembolization? A validation study of old and new models. Radiology 2012; 262:708–718. 12. Jung ES, Kim JH, Yoon EL, et al. Comparison of the methods for tumor response assessment in patients with hepatocellular carcinoma undergoing transarterial chemoembolization. J Hepatol 2013; 58:1181–1187. 13. Tawada A, Chiba T, Ooka Y, et al. Efficacy of transarterial chemoembolization targeting portal vein tumor thrombus in patients with hepatocellular carcinoma. Anticancer Res 2014; 34:4231–4237.
Kim et al
’
JVIR
14. Yoon SM, Lim YS, Won HJ, et al. Radiotherapy plus transarterial chemoembolization for hepatocellular carcinoma invading the portal vein: long-term patient outcomes. Int J Radiat Oncol Biol Phys 2012; 82: 2004–2011. 15. Kim KM, Kim JH, Park IS, et al. Reappraisal of repeated transarterial chemoembolization in the treatment of hepatocellular carcinoma with portal vein invasion. J Gastroenterol Hepatol 2009; 24:806–814.
Impact of the Interval between Transarterial Chemoembolization Sessions on Survival in Patients with Unresectable Hepatocellular Carcinoma Hyung-Don Kim, MD, Jihyun An, MD, Jin Hyoung Kim, MD, Dong Il Gwon, MD, Ji Hoon Shin, MD, Gi-Young Ko, MD, Hyun-Ki Yoon, MD, Kyu-Bo Sung, MD, Kang Mo Kim, MD, and Han Chu Lee, MD
ABSTRACT Purpose: To evaluate clinical impact of different intervals between multiple transarterial chemoembolization sessions in patients with unresectable hepatocellular carcinoma (HCC). Materials and Methods: A retrospective cohort study of 305 consecutive patients with HCC who underwent multiple sessions of on-demand transarterial chemoembolization by two independent physicians with different management policies in terms of transarterial chemoembolization interval was performed; 180 patients had intervals between the first and second transarterial chemoembolization session of o 60 days (short-interval group), and 125 patients had transarterial chemoembolization intervals of Z 60 days (conventional-interval group). Results: The short-interval group had more cases of advanced-stage HCC, less favorable response to transarterial chemoembolization, and higher likelihood of having Child-Pugh class A. The short-interval group underwent more transarterial chemoembolization sessions (6.6 vs 5.5, P ¼ .011), although the total number of admissions and total hospital stay were similar to the conventional-interval group. Overall survival was similar in the two groups in the full and the propensity score–matched cohorts. Although the overall survival of patients with Child-Pugh class A was comparable between the two groups in the full and propensity score–matched cohorts, the short-interval group showed inferior survival (P ¼ .005) and a nonsignificant trend toward inferior survival (P ¼ .117) in the full and propensity score–matched cohorts, respectively, for patients with Child-Pugh class B. Conclusions: Transarterial chemoembolization interval did not affect survival outcomes of patients with Child-Pugh class A. A shorter transarterial chemoembolization interval showed a nonsignificant trend of adversely affecting survival for patients with Child-Pugh class B.
ABBREVIATIONS CI = confidence interval, HCC = hepatocellular carcinoma, HR = hazard ratio, INR = international normalization ratio, mRECIST = modified response evaluation criteria in solid tumor
From the Departments of Internal Medicine (H.-D.K., J.A., K.M.K., H.C.L.) and Radiology (J.H.K., D.I.G., J.H.S., G.-Y.K., H.-K.Y., K.-B.S.), Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 138-736, Republic of Korea. Received August 18, 2015; final revision received and accepted December 4, 2015. Address correspondence to K.M.K.; E-mail: [email protected] None of the authors have identified a conflict of interest. & SIR, 2016 J Vasc Interv Radiol 2016; XX:]]]–]]] http://dx.doi.org/10.1016/j.jvir.2015.12.005
Transarterial chemoembolization is the most widely used treatment modality for intermediate-stage hepatocellular carcinoma (HCC) (1–3). However, transarterial chemoembolization using iodized oil has not been standardized, and its application largely depends on the clinical decision of each physician, with institutions showing variations with techniques, such as different chemotherapeutic agents (eg, doxorubicin and cisplatin); embolic materials and doses; and retreatment strategies, including differences in the interval between transarterial chemoembolization sessions (4–7).
2
’
Multiple Transarterial Chemoembolization Sessions and HCC
The interval between consecutive transarterial chemoembolization sessions is recommended to be 2–4 months in the current guidelines, whether it is performed on-demand or not (2,3). However, this recommended transarterial chemoembolization interval is not evidence based, and appropriate studies of the optimal interval between transarterial chemoembolization sessions are unavailable. Furthermore, intervals of Z 2 months between transarterial chemoembolization sessions may be suboptimal in terms of controlling tumor progression (7). In the present study, the clinical impact of the interval between transarterial chemoembolization sessions on overall patient survival and safety was evaluated.
MATERIALS AND METHODS Study Subjects This study was approved by the institutional review board of the institution, and the requirement for informed consent from the patients was waived. The study population was derived from a historical cohort of 555 consecutive patients with HCC who, in conjunction with concomitant and subsequent treatment with radiation therapy for portal vein invasion and sorafenib, underwent repeated transarterial chemoembolization sessions by two independent physicians (K.M.K., H.C.L.) with different management policies in terms of the transarterial chemoembolization schedule between January 2006 and December 2012. As a baseline study, all patients underwent four-phase dynamic computed tomography (CT), and the diagnosis of HCC was made according to algorithmic guidelines (1–3). Patients with a complete response, based on modified Response Evaluation Criteria in Solid Tumor (mRECIST), after the first transarterial chemoembolization treatment were excluded because a complete response can directly affect the interval between transarterial chemoembolization sessions (n = 132) (5,6,8,9). Patients with main portal vein invasion or bilateral involvement of the first branch portal vein (n = 57), extrahepatic metastasis (n = 54), or other concomitant malignancy (n = 7) were also excluded. Finally, 305 patients were included in the analyses and subdivided into two groups according to the interval between the first and second transarterial chemoembolization session: a shortinterval group (first transarterial chemoembolization interval o 60 d) and a conventional-interval group (first transarterial chemoembolization interval Z 60 d).
Transarterial Chemoembolization Procedure In accordance with the conventional transarterial chemoembolization protocol applied in our liver center (10), superior mesenteric arteriography and common hepatic arteriography were performed to assess the overall anatomy, tumor burden, and portal vein patency, based on intrahepatic HCC status assessed by hepatic
Kim et al
’
JVIR
dynamic CT images. Cisplatin at 2 mg/kg body weight (Cisplan; Dong-A Pharm Co, Seosan, Korea) was infused into the lobar hepatic artery for 15 minutes without an injection of embolic particles. After vascular catheterization with a microcatheter placed selectively or superselectively into the distal tumor-feeding artery, an emulsion of 2–20 mL of iodized oil (Lipiodol Ultra-Fluide; Laboratoires Guerbet, Aulnay-sous-Bois, France) and cisplatin in a 1:1 ratio was administered into the target arteries. Embolization of the arterial tumor feeders was then performed using a Gelfoam slurry (Gelfoam; Upjohn, Kalamazoo, Michigan) until arterial flow stasis was achieved. The Gelfoam slurry was made manually by cutting up a 70 mm 50 mm 10 mm Gelfoam sponge. There was no change in the transarterial chemoembolization protocol throughout the study period. Principally, transarterial chemoembolization was performed on-demand (5,6,8) every 4–16 weeks if there was evidence of residual viable tumor on follow-up CT imaging. Transarterial chemoembolization was not implemented when there was no residual tumor. The decision not to perform further transarterial chemoembolization procedures was made using the following criteria: (i) deterioration in liver function, (ii) transarterial chemoembolization was considered ineffective, (iii) ascites worsened, (iv) severe vascular invasion that made additional transarterial chemoembolization impossible, and (v) other technical problems or contraindications for transarterial chemoembolization.
Study Outcomes and Follow-up The primary outcome of the current study was all-cause mortality. The index date was defined as the date of the first transarterial chemoembolization session. Patients were followed up from the index date to death or the last follow-up date (August 31, 2014). Overall survival was compared for the entire study population and the propensity score–matched cohort. Thereafter, subgroup analysis was performed according to the Child-Pugh class, tumor size, and presence or absence of portal vein invasion. Four-phase dynamic CT was performed at the start and 1 month after transarterial chemoembolization, and the tumor response to transarterial chemoembolization was evaluated by mRECIST, which exhibits superior performance for determining tumor response (11,12). Dynamic CT scans subsequently were principally performed 1 month after a given course of transarterial chemoembolization for the patients with remaining viable tumors. In patients with no viable tumor on dynamic CT images after repeated transarterial chemoembolization, follow-up dynamic CT images were taken at intervals of 2–3 months until tumor recurrence was noted. The total number of transarterial chemoembolization sessions performed in each patient was recorded. To evaluate procedure-related mortality, death within 1 month after the last transarterial chemoembolization session was
Volume XX
’
Number X
’
Month
’
2016
3
analyzed. Regular follow-up of liver function was performed and included the international normalization ratio (INR). As a measure of medical cost, the number of hospital admissions and total duration of hospital stay during serial follow-up were recorded, including not only admissions for the transarterial chemoembolization sessions but also any admissions for supportive care of patients with cancer-related symptoms or deteriorated liver function.
by a larger size of the largest HCC nodule (7.3 cm vs 5.4 cm) and higher prevalence of portal vein invasion (31.7% vs 10.4%) and therefore more advanced Barcelona Clinic Liver Cancer stages (0, 1.7% vs 2.4%; A, 28.3% vs 40.0%; B, 38.3% vs 47.2%; C, 31.7% vs 10.4%). Serum αfetoprotein levels were not significantly different between the groups. After propensity score matching, 104 pairs of patients from the two groups were matched and showed comparable baseline variables (Table 1).
Statistical Analysis
Transarterial Chemoembolization Interval and Hospitalization
Student t test or Mann-Whitney U test was used to compare quantitative variables, and χ2 or Fisher exact test was used to compare qualitative variables. Student t test was used when the variables showed a normal distribution according to visual histogram inspection, and MannWhitney U test was used for variables showing a skewed distribution. Multivariate analysis was performed with Cox proportional hazards model analysis to assess factors associated with overall survival. Overall survival was compared using the log-rank test. To minimize any potential selection bias and confounding factors, propensity score matching was performed. Variables used to derive the propensity score included age; sex; number of tumors; diameter of the largest tumor; presence of portal vein invasion; INR; platelet count; Child-Pugh score; and serum levels of aspartate aminotransferase, alanine aminotransferase, albumin, total bilirubin, and α-fetoprotein. The propensity score was estimated nonparametrically using the matchit function in the R package MatchIt. Matching was performed using the nearest neighbor matching method, using a caliper width of 0.2 multiplied by the standard deviation for the linearly transformed propensity scores (logit transformation). All statistical analyses were performed using IBM SPSS Statistics for Windows, version 20 (IBM Corporation, Armonk, New York) or R software version 3.1.1 (R Foundation for Statistical Computing, Vienna, Austria). All reported P values were two-sided, and a P value o .05 was considered significant.
RESULTS Clinical Characteristics The present study comprised 305 patients with HCC; 180 belonged to the short-interval group, and 125 belonged to the conventional-interval group. The mean age and sex ratios were similar between the two groups. Baseline biochemical results were comparable between the two groups, including the serum alanine aminotransferase level (45.7 IU/mL vs 45.2 IU/mL), albumin level (3.5 g/ dL vs 3.4 g/dL), total bilirubin level (1.1 mg/dL vs 1.1 mg/dL), and INR (1.10 vs 1.12). However, the platelet count was significantly higher (166.1K/mm3 vs 129.8K/ mm3) and Child-Pugh class A were more frequent in the short-interval group. The short-interval group had tumors of more locally advanced stages, characterized
The median interval between the first and second transarterial chemoembolization session was 51 days (interquartile range, 38–70 d). As shown in Figure 1, the transarterial chemoembolization interval was significantly different between the two physicians because they had different transarterial chemoembolization management policies: physician A, a median of 70 days (interquartile range, 61– 82 d); physician B, a median of 42 days (interquartile range, 36–55 d; P o .001). The mean total number of transarterial chemoembolization sessions was significantly higher in the short-interval group (6.6 vs 5.5, P ¼ .011). However, the total number of hospital admissions (8.7 vs 7.8, P ¼ .066) and total hospital stay (55.6 d vs 54.5 d, P ¼ .767) were comparable between the two groups. The short-interval group showed less partial response (73.9% vs 89.6%) but more stable disease (20.0% vs 8.0%) and progressive disease (6.1% vs 2.4%) after the first transarterial chemoembolization session according to mRECIST (P ¼ .003). After propensity score matching, responses to transarterial chemoembolization were not statistically different between the two groups (P ¼ .157), and other transarterial chemoembolization and hospitalization parameters showed similar patterns in the full and the propensity score–matched cohorts (Table 2).
INR Levels at Specific Time Points after First Transarterial Chemoembolization The mean INR level of the study subjects showed a trend toward aggravation and was similar between the shortinterval and conventional-interval groups at 3, 6, and 12 months after the first transarterial chemoembolization session. The same was true for the propensity score– matched cohort (Table 3).
Factors Associated with Survival Cox regression analyses revealed that tumor diameter 4 5 cm was a significant independent predictor of patient survival (adjusted hazard ratio [HR], 1.60; 95% confidence interval [CI], 1.21–2.12; P ¼ .001), along with partial response based on mRECIST (HR, 0.63; 95% CI, 0.46–0.87; P ¼ .005) and serum α-fetoprotein 4 400 ng/ mL (HR, 1.47; 95% CI, 1.11–1.93; P ¼ .007) (Table 4). Preserved liver function characterized by Child-Pugh class A showed a statistically marginal association with overall survival (HR, 0.72; 95% CI, 0.51–1.03; P ¼ .069).
4
’
Multiple Transarterial Chemoembolization Sessions and HCC
Kim et al
’
JVIR
Table 1 . Baseline Characteristics of Study Patients Full Cohort
Characteristics Age (y)* Male sex
Short-Interval Group (n ¼ 180) 56.1 ⫾ 10.4 156 (86.7%)
Conventional-Interval Group (n ¼ 125) P Value 56.7 ⫾ 9.5 106 (84.8%)
Etiology HBV (%) HCV (%) Others (%) ALT (IU/mL)* Albumin (g/dL)* Total bilirubin (mg/dL)*
Child-Pugh class A (%) B (%)
.654 .645
47 (26.1%) 45.7 ⫾ 28.8 3.5 ⫾ 0.5 1.1 ⫾ 1.0
183 (11–4,555)
94 (75.2%) 8 (6.4%) 23 (18.4%) 45.2 ⫾ 7.2 3.4 ⫾ 0.5
Conventional-Interval Group (n ¼ 104) P Value
57.4 ⫾ 10.9 90 (86.5%)
56.4 ⫾ 9.9 89 (85.6%)
68 (65.4%) 11 (10.6%)
80 (76.9%) 4 (3.8%)
1.1 ⫾ 0.6 1.12 ⫾ 0.11 129.8 ⫾ 66.4 74 (14–684)
.858 .148 o .001 .083
45.1 ⫾ 26.6 3.5 ⫾ 0.5
20 (19.3%) 43.9 ⫾ 24.7 3.5 ⫾ 0.5
.724 .870
1.1 ⫾ 0.6
1.1 ⫾ 0.6
.516
1.10 ⫾ 0.10 138.2 ⫾ 59.3
1.10 ⫾ 0.11 134.2 ⫾ 66.8
.504 .650
136 (11–943)
82 (17–581)
93 (89.4%)
87 (83.7%)
.048 160 (88.9%)
101 (80.8%)
20 (11.1%)
24 (19.2%)
2.4 ⫾ 1.7 7.3 ⫾ 4.3
2.7 ⫾ 2.6 5.4 ⫾ 3.7
.270 o .001
Portal vein invasion (%)
57 (31.7%)
13 (10.4%)
o .001
11 (10.6%)
17 (16.3%)
2.6 ⫾ 2.6 5.4 ⫾ 3.8
2.6 ⫾ 1.8 5.6 ⫾ 3.4
9 (8.7%)
13 (12.5%)
o .001 3 (1.7%)
3 (2.4)%
3 (2.9%)
3 (2.9)%
50 (40.0%)
41 (39.4%)
40 (38.5%)
B (%) C (%)
69 (38.3%) 57 (31.7%)
59 (47.2%) 13 (10.4%)
51 (49.0%) 9 (8.7%)
48 (46.2%) 13 (12.5%)
24.4 (12.7–38.4)
.412
4 .999 .660 .367 .842
51 (28.3%)
23.6 (14.8–34.6)
.958 .223
A (%)
Follow-up (mo)†
.498 .841 .091
25 (24.0%) .883 .196
No. HCC nodules* Size of HCC nodule (cm)* BCLC stage 0 (%)
Short-Interval Group (n ¼ 104)
.165 117 (65%) 16 (8.9%)
1.10 ⫾ 0.11 INR* Platelet count ( 1,000/mm3)* 166.1 ⫾ 98.7 α-Fetoprotein (ng/mL)†
PS-Matched Cohort
25.3 (14.8–34.6)
25.0 (12.5–37.5)
.821
Note–Data are presented as n (%) or mean ⫾ SD or median (interquartile range). ALT ¼ alanine aminotransferase; BCLC ¼ Barcelona Clinic Liver Cancer; HBV ¼ hepatitis B virus; HCC ¼ hepatocellular carcinoma; HCV ¼ hepatitis C virus; INR ¼ international normalized ratio; PS ¼ propensity score. n Student t test was used. † Mann-Whitney U test was used.
0.97–1.61; P ¼ .121) and propensity score–matched (HR, 1.00; 95% CI, 0.74–1.34; P ¼ .992) cohorts.
Survival Analysis
Figure 1. Graphic depiction of the interval between the first and second transarterial chemoembolization sessions according to physician.
A first transarterial chemoembolization interval of o 60 days was not a significant predictor of overall survival in univariate analysis of the full (HR, 1.23; 95% CI,
The median survival times for patients with shortened and conventional intervals were 28.5 months and 28.6 months, respectively. According to Kaplan-Meier analysis, patients with different transarterial chemoembolization intervals showed similar survival results in both the full (P ¼ .121) and the propensity score–matched (P ¼ .993) cohorts (Fig 2a, b). Patients were subdivided into Child-Pugh class A (n ¼ 261) and Child-Pugh class B (n ¼ 44) groups. Although the overall survival of patients with Child-Pugh class A was similar between the two groups (P ¼ .403), patients with Child-Pugh class B in the short-interval group showed inferior survival results (P ¼ .005) (Fig 3a, b). There were no significant survival differences between the two groups regardless of whether there was portal vein invasion or no portal vein invasion (P ¼ .290 or P ¼ .851, respectively) (Fig 3c, d). Patients with Child-Pugh class A showed similar survival outcomes in the propensity
Volume XX
’
Number X
’
Month
’
2016
5
Table 2 . Profiles of Transarterial Chemoembolization Sessions and Hospitalizations Full Cohort Short-Interval Group (n ¼ 180)
Characteristics
PS-Matched Cohort
Conventional-Interval Group (n ¼ 125)
P Value
Response to first transarterial chemoembolization
Short-Interval Group (n ¼ 104)
Conventional-Interval Group (n ¼ 104)
.003
P Value .157
based on mRECIST PR SD
133 (73.9%) 36 (20.0%)
PD
11 (6.1%)
Interval between first and second transarterial
112 (89.6%) 10 (8.0%)
83 (79.8%) 16 (15.4%)
3 (2.4%)
41.1 ⫾ 8.0
78.7 ⫾ 20.5
6.6 ⫾ 3.6
8.7 ⫾ 4.3 55.6 ⫾ 30.0
5 (4.8%)
93 (89.4%) 8 (7.7%) 3 (2.9%)
o .001
41.6 ⫾ 8.3
78.7 ⫾ 20.3
o .001
5.5 ⫾ 3.4
.011
6.4 ⫾ 3.3
5.4 ⫾ 3.3
.043
7.8 ⫾ 4.0 54.5 ⫾ 33.1
.066 .767
8.8 ⫾ 4.3 54.2 ⫾ 27.9
7.9 ⫾ 3.9 56.4 ⫾ 35.1
.092 .614
chemoembolization (d)* Total transarterial chemoembolization sessions* Total admissions* Total hospital duration (d)*
mRECIST ¼ modified Response Evaluation Criteria in Solid Tumor; PD ¼ progressive disease; PR ¼ partial response; PS ¼ propensity score; SD ¼ stable disease. n Student t test was used.
Table 3 . Comparison of INR at Specific Time Points after First Transarterial Chemoembolization Full Cohort
Baseline INR INR at 3 mo from first
PS-Matched Cohort
Short-Interval
Conventional-Interval
Short-Interval
Conventional-Interval
Group (n ¼ 180)
Group (n ¼ 125)
P Value
Group (n ¼ 104)
Group (n ¼ 104)
P Value
1.10 ⫾ 0.11 1.14 ⫾ 0.14
1.12 ⫾ 0.11 1.15 ⫾ 0.25
.148 .364
1.10 ⫾ 0.10 1.13 ⫾ 0.14
1.10 ⫾ 0.11 1.16 ⫾ 0.27
.504 .390
1.17 ⫾ 0.31
1.16 ⫾ 0.19
.818
1.15 ⫾ 0.19
1.16 ⫾ 0.20
.623
1.21 ⫾ 0.35
1.19 ⫾ 0.22
.611
1.17 ⫾ 0.24
1.17 ⫾ 0.20
.958
transarterial chemoembolization INR at 6 mo from first transarterial chemoembolization INR at 12 mo from first transarterial chemoembolization INR ¼ international normalized ratio; PS = propensity score. Note–Student t test was used.
score–matched cohort (P ¼ .628). For the patients with Child-Pugh class B in the matched cohort, there was a nonsignificant trend for inferior survival in the short-interval group (P ¼ .117) (Fig 4a–d). There was no difference in the proportions of patients who died within 30 days after the last transarterial chemoembolization session (1.1% [2/180] in the short-interval group and 1.6% [2/125] in the conventional-interval group, P 4 .99).
DISCUSSION Although there has been no appropriate study of the interval between consecutive sessions of transarterial chemoembolization, these sessions have conventionally been performed
with a median interval of 2 months (7); guidelines recommend transarterial chemoembolization intervals of 3–4 months (2) or 2–3 months (3). From the oncologic point of view, transarterial chemoembolization intervals of 4 2 months may be insufficient, particularly given that systemic chemotherapy is delivered every 3 weeks, because local progression or metastasis of viable tumors will almost always occur in that time (7). However, there has been concern that transarterial chemoembolization applied with an interval o 2 months might result in further deterioration of liver function and lead to poor survival outcomes as a result of liver failure. In the present study, the survival outcomes and safety profiles of patients with HCC undergoing repeated
6
’
Multiple Transarterial Chemoembolization Sessions and HCC
Kim et al
’
JVIR
Table 4 . Multivariate Regression Analysis of Overall Survival in Full Cohort Overall Mortality Univariate Analysis
Multivariable Analysis
Variables
HR
95% CI
P Value
HR
95% CI
P Value
Age (y) Male sex
0.99 0.95
0.98–1.00 0.65–1.40
.122 .807
— —
— —
— —
Etiology (HBV)
1.09
0.81–1.44
.603
—
—
—
Child-Pugh class A Multiple tumors (vs single)
0.74 1.08
0.52–1.04 0.83–1.42
.083 .565
0.72 —
0.51–1.03 —
.069 —
Tumor diameter Z 5 cm
1.85
1.41–2.42
o .001
1.60
1.21–2.12
.001
Presence of portal vein invasion α-fetoprotein 4 400 ng/mL
1.64 1.70
1.21–2.22 1.31–2.22
.002 o .001
— 1.47
— 1.11–1.93
— .007
PR based on mRECIST
0.60
0.44–0.82
.001
0.63
0.46–0.87
.005
Transarterial chemoembolization interval o 60 d
1.23
0.97–1.61
0.121
—
—
—
Note–Cox proportional hazards model was used for all analyses. CI ¼ confidence interval; HBV ¼ hepatitis B virus; HR ¼ hazard ratio; mRECIST ¼ modified response evaluation criteria in solid tumor; PR ¼ partial response.
Figure 2. Overall survival rate of the full cohort (a) and the propensity score–matched cohort (b).
on-demand transarterial chemoembolization on the basis of different intervals between transarterial chemoembolization sessions were evaluated. The study showed that there was no significant difference in overall survival between patients with transarterial chemoembolization intervals o 60 days and patients with transarterial chemoembolization intervals of 60 days or longer. The short-interval resulted in inferior survival outcomes only in the subgroup of patients with Child-Pugh class B in the full cohort, whereas patients with Child-Pugh class B in the propensity score–matched cohort showed only a trend for inferior survival. Also, INR levels at specific time points after transarterial chemoembolization were not different between the two interval groups. Despite higher numbers of transarterial chemoembolization sessions in the short-interval group than in the conventional-interval group, no increases in the total number of admissions and total hospital duration
were observed in the full and propensity score–matched cohorts. In addition, in the present study, aggressive transarterial chemoembolization with a shortened interval tended to be applied in patients with advanced HCC, which is characterized by larger nodules, more frequent invasion of major vessels, and a more unfavorable response to transarterial chemoembolization, although the liver function of these patients was relatively preserved. It is conceivable that short-interval transarterial chemoembolization might be better at controlling and curbing tumor progression, owing to the finding that overall survival was not inferior in the short-interval group with unfavorable tumor parameters. However, this effect did not translate into better survival in the multivariate and propensity score–matched analyses. In conjunction with the absence of concrete evidence that short-interval transarterial chemoembolization better
Volume XX
’
Number X
’
Month
’
2016
7
Figure 3. Overall survival rate in patients with Child-Pugh class A (a), with Child-Pugh class B (b), with portal vein invasion (c), and with no portal vein invasion (d) in the full cohort.
prevented tumor progression, there would be no strong rationale for recommending short-interval transarterial chemoembolization. In particular, the shortened transarterial chemoembolization interval led to poor survival outcomes in patients with Child-Pugh class B, suggesting that the risk of deterioration of liver function caused by a shortened transarterial chemoembolization interval outweighs the potential benefit of aggressive tumor control with frequent transarterial chemoembolization sessions in patients with Child-Pugh class B. Although a statistically significant difference was not reached in patients with Child-Pugh B in the propensity score– matched cohort, possibly because of the small number of patients included in the analysis, application of shortinterval transarterial chemoembolization in patients with Child-Pugh B still seems to be hazardous considering the trend for inferior survival outcomes. In terms of safety, serial INR levels after the first transarterial chemoembolization session and treatmentrelated mortality—described as death within 30 days of
undergoing transarterial chemoembolization—did not differ between the two study groups. In addition, the finding that short-interval transarterial chemoembolization involved more sessions of this treatment without significantly increasing the number of hospitalizations or duration of hospital stay may indicate the overall safety and potential clinical efficacy of a short-interval approach. We assume that better tumor control in the short-interval group may have decreased the number of admissions and duration of hospital stay for supportive care related to cancer progression. In contrast to previous concerns, these results imply that the shortened transarterial chemoembolization interval does not cause a further decline in liver function and increased patient mortality, especially in cases with preserved liver function. To the best of our knowledge, the current findings are novel because no relevant study has previously directly compared clinical outcomes according to the interval between transarterial chemoembolization procedures. A
8
’
Multiple Transarterial Chemoembolization Sessions and HCC
Kim et al
’
JVIR
Figure 4. Overall survival rate in patients with Child-Pugh class A (a), Child-Pugh class B (b), portal vein invasion (c), and no portal vein invasion (d) in the propensity score–matched cohort.
previous study by Ernst et al (8) compared survival outcomes between on-demand transarterial chemoembolization and transarterial chemoembolization performed every 2 months with a fixed schedule. In that study, transarterial chemoembolization was more frequently performed in the fixed-schedule group and showed inferior survival. However, the relatively narrower interval of the fixed-schedule group was predominantly induced by a longer interval between sessions for the on-demand transarterial chemoembolization caused by delaying the next transarterial chemoembolization session in patients with a complete response. Such patients were excluded from the current study. Furthermore, this relatively narrow interval was 2 months, which was used as a cutoff to discriminate between the short-interval and conventional-interval groups. Although no clinical benefit of short-interval transarterial chemoembolization in prolonging patient survival was demonstrated, the current study is clinically important because it provides evidence that frequent transarterial chemoembolization could be a contraindication in
patients with Child-Pugh class B and that more aggressive forms of transarterial chemoembolization could be considered in patients with preserved liver function, together with the finding that a short transarterial chemoembolization interval did not increase the number of hospitalizations and duration of hospital stay. Furthermore, the study results provide clinical evidence that the shortest time required for the recovery of liver function to perform the next transarterial chemoembolization session safely is o 60 days. Based on the current results, a future prospective study should try to confirm the efficacy and safety of short-interval transarterial chemoembolization regarding tumor progression and survival outcomes in selected patients with preserved liver function who need aggressive tumor control because of large tumor burden. In addition to the retrospective nature of the present study, there are several other limitations. First, the possibility of a selection bias remained. Short-interval transarterial chemoembolization tended to be applied to patients with local progression of tumors and preserved liver function, which might have directly affected study
Volume XX
’
Number X
’
Month
’
2016
outcomes. To overcome the effect of a selection bias and evaluate the effect of the transarterial chemoembolization interval on survival, statistical approaches, including propensity score–matched analysis, were implemented. After propensity score matching, all the baseline characteristics, including tumor parameters, were comparable, and the responses to transarterial chemoembolization were not statistically different between the two groups, providing a sound background to the comparison of the two groups in a nonrandomized setting. Second, the first transarterial chemoembolization interval may not fully represent the entire HCC management policy. Although the two physicians who treated study patients had clearly different management policies, multiple factors, including patient tolerability, liver function after transarterial chemoembolization sessions, and extent of remnant viable tumor, might have also played a role in determining when the subsequent transarterial chemoembolization session was performed. Furthermore, a mean value of the total intervals between transarterial chemoembolization could not be incorporated because transarterial chemoembolization was performed on-demand. However, considering the complexities of the application of transarterial chemoembolization for the treatment of HCC (4–7) and the current finding that the total number of transarterial chemoembolization sessions significantly differed according to the first interval between the transarterial chemoembolization sessions, the first transarterial chemoembolization interval may arguably be the only realistic parameter that reflects the management strategy regarding the interval between transarterial chemoembolization sessions. Third, patients with portal vein invasion were included in the analyses, even though portal vein invasion is a contraindication for transarterial chemoembolization under Western guidelines (1,2). However, evidence is accumulating regarding the efficacy of transarterial chemoembolization in patients with portal vein invasion (10,13–15). Furthermore, it was shown that there was no survival difference regarding the transarterial chemoembolization interval according to whether the patient had portal vein invasion. Fourth, as for the transarterial chemoembolization protocol, infusion of chemotherapeutic material through the lobar artery, which may have an additional antitumor effect because of possibly wider delivery of chemotherapeutic material, is not standard and requires scientific verification in future studies. However, it is unlikely that chemotherapy infusion at the lobar artery level affected the study results because this procedure was identically performed in both study groups. Although the follow-up principle of performing dynamic CT scanning 1 month after transarterial chemoembolization sessions was identical for both physicians, the follow-up intervals might not have been the same because of the differences in the frequency of transarterial chemoembolization. However, because unequal follow-up imaging was the consequence
9
rather than the cause of the different management policy, it is unlikely that follow-up imaging variations would affect the clinical outcome. Lastly, the number of study subjects with Child-Pugh class B was possibly too small to draw a conclusive inference. However, we believe that our data are still valuable in that we at least showed that short-interval transarterial chemoembolization in patients with Child-Pugh class B requires extreme caution and could be hazardous. In addition, no other data are currently available regarding this issue. In conclusion, patients with HCC undergoing repeated on-demand transarterial chemoembolization with an interval of o 60 days show similar overall survival rates without a significant increase in the number of hospitalizations and duration of hospital stay compared with patients with transarterial chemoembolization intervals of Z 60 days. Subgroup analysis showed that a transarterial chemoembolization interval of o 60 days led to a nonsignificant trend toward inferior survival outcomes for patients with Child-Pugh class B.
ACKNOWLEDGMENT We thank Dr. Seungbong Han (Department of Applied Statistics, Gachon University) for full support of the statistical analysis.
REFERENCES 1. Bruix J, Sherman M. Management of hepatocellular carcinoma: an update. Hepatology 2011; 53:1020–1022. 2. European Association for the Study of the Liver; European Organisation for Research and Treatment of Cancer. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012; 56:908–943. 3. Omata M, Lesmana LA, Tateishi R, et al. Asian Pacific Association for the Study of the Liver consensus recommendations on hepatocellular carcinoma. Hepatol Int 2010; 4:439–474. 4. Vogl TJ, Naguib NN, Nour-Eldin NE, et al. Review on transarterial chemoembolization in hepatocellular carcinoma: palliative, combined, neoadjuvant, bridging, and symptomatic indications. Eur J Radiol 2009; 72:505–516. 5. Cheng AL, Amarapurkar D, Chao Y, et al. Re-evaluating transarterial chemoembolization for the treatment of hepatocellular carcinoma: consensus recommendations and review by an international expert panel. Liver Int 2014; 34:174–183. 6. Terzi E, Golfieri R, Piscaglia F, et al. Response rate and clinical outcome of HCC after first and repeated cTACE performed “on demand”. J Hepatol 2012; 57:1258–1267. 7. Marelli L, Stigliano R, Triantos C, et al. Transarterial therapy for hepatocellular carcinoma: which technique is more effective? A systematic review of cohort and randomized studies. Cardiovasc Intervent Radiol 2007; 30:6–25. 8. Ernst O, Sergent G, Mizrahi D, Delemazure O, Paris JC, L’Hermine C. Treatment of hepatocellular carcinoma by transcatheter arterial chemoembolization: comparison of planned periodic chemoembolization and chemoembolization based on tumor response. AJR Am J Roentgenol 1999; 172:59–64. 9. Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment for hepatocellular carcinoma. Semin Liver Dis 2010; 30:52–60. 10. Kim GA, Shim JH, Yoon SM, et al. Comparison of chemoembolization with and without radiation therapy and sorafenib for advanced hepatocellular carcinoma with portal vein tumor thrombosis: a propensity score analysis. J Vasc Interv Radiol 2015; 26:320–329. e326. 11. Shim JH, Lee HC, Kim SO, et al. Which response criteria best help predict survival of patients with hepatocellular carcinoma following
10
’
Multiple Transarterial Chemoembolization Sessions and HCC
chemoembolization? A validation study of old and new models. Radiology 2012; 262:708–718. 12. Jung ES, Kim JH, Yoon EL, et al. Comparison of the methods for tumor response assessment in patients with hepatocellular carcinoma undergoing transarterial chemoembolization. J Hepatol 2013; 58:1181–1187. 13. Tawada A, Chiba T, Ooka Y, et al. Efficacy of transarterial chemoembolization targeting portal vein tumor thrombus in patients with hepatocellular carcinoma. Anticancer Res 2014; 34:4231–4237.
Kim et al
’
JVIR
14. Yoon SM, Lim YS, Won HJ, et al. Radiotherapy plus transarterial chemoembolization for hepatocellular carcinoma invading the portal vein: long-term patient outcomes. Int J Radiat Oncol Biol Phys 2012; 82: 2004–2011. 15. Kim KM, Kim JH, Park IS, et al. Reappraisal of repeated transarterial chemoembolization in the treatment of hepatocellular carcinoma with portal vein invasion. J Gastroenterol Hepatol 2009; 24:806–814.