Evaluation of the effect of Operator Experience on Outcome of Hepatic Artery Embolization of Hepatocellular carcinoma in a Tertiary Cancer Center

Abstract

Rationale and Objectives:

There is lack of information on the learning curve and the effect of operator’s experience on the quality outcomes of hepatic artery embolization(TAE). The aim of this study was to evaluate the effect of operator experience on outcomes of TAE of hepatocellular carcinoma(HCC).

Materials and Methods:

Demographic characteristics and outcomes including overall survival(OS), time to local tumor progression(TLP) and complications of HCC patients treated with TAE were collected. Operators’ experience was measured in years based on the years after completion of fellowship and the date of first embolization and was divided into 5 groups: G1, Less than five year of experience at the time of first embolization(YOE), G2, 5–10YOE, G3, 10–15YOE, G4, 15–20YOE and G5, more than 20YOE. The effects of operator’s experience and outcomes were assessed using linear regression.

Results:

From January 2012 to January 2015, 93 patients(age range=30–86yrs) were treated. Number of patients treated by each group was: G1=12; G2=8; G3=23; G4=5; and G5=45. All groups were similar in regards to degree of cirrhosis, Barcelona Clinic Liver Cancer staging and Child–Pugh score (p > 0.05). Median TLP was 8.8 months. TLP was 7.0, 6.8, 19.2, 7.9 and 8.2 months in G1, G2, G3, G4 and G5, respectively;(p= 0.56). OS for 1, 2 and 3 years was 75%, 56% and 42% for G1; 87%, 54% and 54% for G2; 91%, 71% and 45% for G3, 100%, 50% and 0 for G4 and 84%, 65% and 40% for G5.

Conclusion:

Among IR fellowship-trained operators in a tertiary cancer center, OS, TLP and post procedure complications of TAE were not affected by the years of post-fellowship experience.

Introduction

Transarterial hepatic embolization (TAE) and chemoembolization (TACE) are two minimally invasive procedures that interventional radiologists (IR) have been using in the past 3–4 decades. Both procedures have been used extensively to treat non-surgical patients with hepatocellular carcinoma (HCC) (1). TACE has been shown to be safe and effective in improving the overall survival when compared to best supportive care through randomized clinical trials (2, 3). Most recently in a randomized clinical trail, TAE was compared with Doxorubicin-eluting microsphere chemoembolization (DEB-TACE) and the study demonstrated that TAE was as effective and safe as DEB-TACE (2).

The principles of how to perform TAE and TACE are part of the teaching curriculum of any IR fellowship training program. The technical success in most interventions is related to skills that are obtained during training and continue to improve after finishing training and during independent practice. IR fellows are trained on how to use wires, microwires, catheters and microcatheters. Additionally, IR fellows will get educated on principles of embolization and different embolizing options.

Experience from radiofrequency ablation (RFA), another minimally invasive procedure for treatment of HCC, demonstrated a significant learning curve (4, 5). Learning curve has been also established with other types of technology-based minimally invasive treatments like laparoscopic surgery (6). Therefore, it can be speculated that such a learning curve also applies to TAE. However, to our best knowledge there is no published information on the effect of operator experience on the outcomes of TAE in terms of outcomes.

The aim of this study was to evaluate the effect of operator experience on the outcome of TAE in terms of overall survival, tumor recurrence and post-procedure complications.

Materials and Methods

Patient selection

Data on patients with HCC who underwent TAE from January 2012 to January 2015 was collected. Only patients whose treatments were performed by a single operator over the course of their disease were included.

Demographics characteristics, tumor size and pattern of involvement, percentage of liver involvement, Child-Pugh score, and Barcelona Clinic Liver Cancer (BCLC) classification were collected. Performance status of the patients was evaluated using the Eastern Corporative Oncology Group (ECOG) performance status scoring system. Technical aspects of embolization including the type, size and volume of embolization material were also collected.

Transarterial hepatic embolization Procedure

TAE procedure was performed using the previously described technique (7). Patients were hydrated prior to the procedure and received an antiemetic (palonosetron hydrochloride 0.25 mg IV) and antibiotic (cephazolin 1 g IV). Embolization was performed using conscious sedation or general anesthesia. In general, a combination of 40–120 µm, 100–300 µm Embosphere (Merit Medical, South Jordan, Utah) and 100 µm poly vinyl alcohol particles (Cook, Bloomington, Indiana) was used to embolized all the arterial supply to the tumor. Only one lobe was embolized at each treatment session and in patients with bilobar disease, the second lobe was treated in 4–6 weeks after the first treatment. Embolization was initiated using the smallest particle (40–120 µm) while the supplying artery was selectively or sub-selectively catheterized. As previously described, embolization was performed till complete stasis was achieved in the supplying vessel. If after using 10 ml of 40–120 µm stasis was not achieved, then the larger size particle (100–300 µm) was used and this was followed by poly vinyl alcohol particles to achieve complete stasis. Similar technique was used by all the operators.

Operator Experience

From 2012 to 2015 the IR team consisted of 16 attendings (IR MD). All operators have completed at least a one-year fellowship in IR. Operator experience was measured in years after completion of IR Fellowship and date of first embolization for each patient. Depending on the years of operator’s experience (YOE) at the time of the first embolization the IR MDs were divided into 5 groups: Group 1, Less than five YOE; Group 2, 5–10 YOE; Group 3, 10–15 YOE; Group 4, 15 to 20 YOE and Group 5 more than 20 YOE. Since the YOE was based on the date of first embolization there was no overlap between the groups.

Outcome

Outcomes analyzed were time to local tumor progression (LTP), local tumor progression free survival (LTPFS), overall survival (OS) and post-procedure complications rates. Post-procedure complications were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.

Statistical analysis

Primary endpoint was overall survival. Secondary endpoints were LTP and complications. OS, LTP and post-procedure complications were assessed using logistic regression. The effects of operator’s experience and outcomes were assessed using linear regression. Means were compared using Students T. X2 was used for categorical variables. Life tables and Kaplan Meier tables and Mantel-Cox statistics were used for LTP and OS.

Institutional research board approval was obtained for this retrospective study on 5/03/2016.

Results

Demographic characteristic

Demographic characteristics of the patients are shown in table 1. Ninety-three patients with HCC were included. Fifty-seven (61%) patients were male. Mean age was 67 ± 12 years (range 30–86). Seventy-one (76%) patient had cirrhosis. Hepatitis C (34/71; 47%) and Hepatitis B (16/71; 22.5%) were the most common etiologies. Eighty-nine (96%) patients were Child-Pugh score A and the rest (4%) were Child-Pugh B. Seventy-eight percent of the patients treated presented with BCLC stage A/B (73/93) (Table 1). There was no statistically significant difference between different experience groups in regards to degree of cirrhosis, BCLC staging and Child–Pugh score (p > 0.05; Table 1.)

Table 1.

Demographic characteristics of the 5 groups of HCC patients treated with TAE.

	G1	G2	G3	G4	G5	P
N	12	8	23	5	45
Age (y)	68.7	67.1	65.8	64.2	67.9	0.891
Gender
Female	2	0	7	2	15	0.427
Male	10	8	16	3	30
Cirrhosis						0.504
No	3	3	7	0	9
Yes	9	5	16	5	36
Child-Pugh score						0.872
A	12	8	23	5	41
B	0	0	0	0	4
C	0	0	0	0	0
BCLC						0.833
A	3	2	9	1	16
B	6	4	10	4	18
C	3	2	4	0	11
% liver involvement						0.789
<25%	7	6	15	4	28
>25, <50	5	1	5	1	15
>50, <75	0	1	2	0	2
>76 to 100	0	0	1	0	0
Extrahepatic disease						0.6
No	12	8	23	5	45
Yes	0	0	0	0	0
Previous surgery						0.060
Yes	5	4	12	1	9
No	7	4	11	4	36
ECOG						0.602
0	11	7	22	13	36
1	1	1	0	0	0
2	0	0	0	0	0
3	0	0	0	0	0
Systemic Chemotderapy						0.366
No	11	6	22	5	42
Yes	1	2	1	0	3

ECOG status was 0 in 89 patients and 1 in 2 patients. Two patient’s ECOG status was not documented. ECOG status was similar between groups (p > 0.05).

Technical Aspects

The number of patients treated by each experience group was as follows: G1, 12 patients; G2, 8 patients; G3, 23 patients; G4, 5 patients; G5, 45 patients. In the technical variables recorded, the total number of embolizations per patient was 2.2 for G1, 3.7 for G2, 2.2 for G3, 2.6 for G4 and 2.8 for G5 (p=0.236). The mean number of embolization vials used was 3.3 for G1, 2.5 for G2, 2.2 for G3, 2.0 for G4 and 2.4 for G5 (p = 0.425)

Outcomes

Local tumor progression free survival:

In G1 LTP free survival was 46%, 18% and 18% for one, two and three years respectively. In G2 LTP free survival was 28%, 0% and 0% for one, two and three years respectively. In G3 LTP free survival was 60%, 32% and 26% for one, two and three years respectively. In G4 LTP free survival was 25%, 0% and 0% for one, two and three years respectively and for G5 LTP free survival was 40%, 24% and 12% for one, two and three years respectively (Figure 1). Median time to local progression for the entire HCC group was 8.8 months. For G1 was 7.0 months; 6.8 months for G2, 19.2 months for G3, 7.9 for G4 and 8.2 months for G5; (p= 0.56).

Figure 1.

Comparison of Time to local progression (TLP) between the 5 groups.

Overall survival:

OS for 1, 2 and 3 years was 75%, 56% and 42% for group 1; 87%, 54% and 54% for group 2; 91%, 71% and 45% for G3, 100%, 50% and 0 for G4 and 84%, 65% and 40% for G5 (Figure 2). Median OS for the entire HCC group was 30 months. Median OS was 29.5 months for G1, 52.4 months for G2, 32 months for G3, 18.5 months for G4 and 30.3 months for G5; (p = 0.75).

Figure 2.

Comparison of overall survival between the five groups.

Complications:

There were 10 (10.7%) grade III complications requiring additional treatment. Distribution of complications between the five experience groups was similar with no statistically significant different (p > 0.05).

Discussion

A learning curve effect has been reported in a variety of new techniques in different specialties including interventional radiology, laparoscopic surgery and interventional cardiology (4–6, 8). In these studies, outcomes improved as the operators gained experience and ultimately a steady state was reached over time. Poon et al evaluated clinical data and treatment outcomes of 100 patients with liver tumor that were treated with RFA (5). They concluded that after 50 cases, the accumulated experience resulted in lower complication rate, higher complete ablation and shorter hospital stay. Lee et al used Taiwan national database and stratified HCC patients receiving RFA according to the yearly volume of ablations at the treating institution (4). They observed that local recurrence at 5 years in the high-volume group (more than 79 cases per year) was significantly lower compared to the low-volume group (less than 10 cases per year); 65.8% vs 72.1%; respectively. However, they did not describe a steady-state in RFA procedure learning curve.

In the first reports on TAE from 1988, Lin et al performed a three-arm randomized clinical trial comparing repeated TAE with TAE plus 5-Fluorouracil (5-FU) and monthly 5-FU (9). This study is among the first studies to evaluate the efficiency of TAE in treating HCC. The one-year survival of the TAE-only group was 42% which was significantly higher than the TAE + 5-FU or 5-FU arm (21 and 13%, respectively). Four years later in 1992, Kawai et al used lipiodol plus Gel sponge for TAE and reported a one-year survival of 74.4% which was almost twice (32.4% more) what was obtained by the first reported results (10). In 2013, Meyer et al demonstrated a 1-year survival rate of 68% for HCC patients that were treated with TAE using 50–150 µm polyvinyl alcohol (11). Finally, in the most recent years, Brown et al demonstrated a mean survival of 20.8 months in HCC patients being treated by TAE using 100–300 µm Bead Blocks (2). Therefore, there has been significant improvement in survival from the initial report approximately 29 years ago and could be speculated that a learning curve exists. Additionally, improvement in technology and better understanding of the nature of the disease has improved the efficiency of TAE in treating HCC.

No difference between the outcomes of TAE was observed in the 5 different experience groups in this study. However, this data does not document that there is no learning curve in TAE. In the contrary, it might suggest that the learning curve has reached to steady-states in operators after completely the IR fellowship. The 4 IR MDs in group 1 each had performed approximately 60–120 supervised embolizations during their fellowship and therefore they might have reached the steady-state of embolization learning curve.

Another possible explanation for not findings statistically significant difference in OS or TLP between the groups might be the method that the patients are being managed and treated at this tertiary cancer center. Patients are treated in a multidisciplinary environment and all treatment decisions are made in a multidisciplinary fashion. Additionally, junior faculty have the advice and support of other faculties and the senior faculty are always available to assist with challenging cases, which is not commonly recorded in charts. Therefore, high volume procedure and multidisciplinary management are the two main advantages that junior faculties in tertiary cancer centers have. The concept of centralization and tertiary referral center has been previously studied and investigators have been able to demonstrate that high volume hospitals and multidisciplinary approach to cancer management is associated with lower mortality and longer survival (12, 13). A study looking at experience of surgeons supervising fellows performing open radical prostatectomy (ORP), showed that urology trainee supervised by a low volume attending had worse oncologic outcomes than attending lead surgeries, but when a high experience surgeon supervised the ORP there were no differences in oncologic outcomes (biochemical recurrence and prostate cancer related mortality) (14). Therefore, supervision with senior faculty could result in similar outcomes in TAE.

The final explanation for not detecting any difference between the outcomes among different level of experience might be the introduction of cone-beam CT (CBCT) scan and automatic vessel detection software. All of the operators had access to Flight Plan for Liver software (GE Healthcare, Chalfont St Gilles, England) which is capable of intraprocedural semi-automatic detection of arteries feeding the tumors (15, 16). Such technologies have demonstrated better sensitivity to detect tumor feeding vessels compared to manual analysis of digital substraction angiography and reformatted CBCT images and are able to assist the operators in better tumor feeding vessel detection (17, 18).

The limitations of the current study are retrospective nature of study with small number of patients being analyzed and no information about input of senior IR physicians. Another limitation is the retrospective design. Finally, we were not able to define when a steady state, if any is reached, for TAE procedure. A prospective study may be able to answer the above mentioned question.

In conclusion, effective training is crucial to allow new interventional radiologists to perform TAE safely and effectively and therefore should remain a high priority. Multidisciplinary approach and treatment should be advised in institutions treating cancer patients and supervision of senior attending helps improve the outcome of procedures performed by more junior faculty. Finally, this single-tertiary cancer center study, demonstrates that hepatic artery embolization for HCC can be effectively performed by Interventional radiologist with different degrees of experience and with similar outcome.

Abbreviations:

TAE

Transarterial hepatic embolization

TACE

Transarterial chemoembolization

DEB-TACE

Doxorubicin-eluting microsphere chemoembolization

HCC

Hepatocellular carcinoma

OS

Overall survival

TLP

Time to local tumor progression

LTPFS

Local tumor progression free survival

YOE

Year of experience

IR

Interventional Radiology

RFA

Radiofrequency ablation

BCLC

Barcelona Clinic Liver Cancer

ECOG

Eastern Corporative Oncology Group

IR MD

Interventional radiology attending

5-FU

5-Fluorouracil

ORP

Open radical prostatectomy

CBCT

cone-beam Computed tomography

IV

intravenous