Transarterial chemoembolization (TACE) is the standard treatment for intermediate-stage hepatocellular carcinoma (HCC). It is unknown whether conventional TACE (cTACE) should be continued or abandoned after initial nonresponse for intermediate-stage HCC. The optimal number of sessions before abandoning cTACE remains debated.

Purpose

To define the number of sessions that patients who do not respond to treatment (hereafter, nonresponders, with stable disease [SD] or progressive disease [PD]) should undergo before abandoning cTACE.

Materials and Methods

Patients with intermediate-stage HCC and Child-Pugh A liver function who underwent consecutive cTACE sessions between January 2005 and December 2012 were retrospectively included from three centers. Radiologic response rate to each session and its correlation with overall survival were evaluated. Response was assessed by modified Response Evaluation Criteria in Solid Tumors. A nomogram constructed by using tumor size, tumor capsule, and α-fetoprotein to predict patients who responded to treatment (hereafter, responders) was validated with sensitivity and specificity.

Results

This study evaluated 4154 patients (mean age, 58 years ± 6 [standard deviation]; 3777 men; primary cohort, 3442 patients [mean age, 58 years ± 6; 3129 men]; validation cohort, 712 patients [mean age, 58 years ± 7; 648 men]). Response rate to first cTACE was 35.6% (1227 of 3442, primary cohort) and 36.7% (261 of 712, validation cohort). For patients with SD who were nonresponders to first cTACE, the response rates after second cTACE were 46.1% (719 of 1560) and 48.4% (147 of 304); for patients with SD who were nonresponders to the second cTACE session, the response rates after the third cTACE session were 58.3% (591 of 1014) and 48.5% (98 of 202). For patients with SD who were nonresponders to third, fourth, and fifth cTACE sessions, response rates after fourth, fifth, and sixth cTACE sessions were less than 10%. All response rates in patients with PD who were nonresponders to the next cTACE were less than 5%. Responders to first, second, and third cTACE had higher 5-year overall survival than nonresponders (all P < .001) but responders to the fourth cTACE did not (P = .21). The sensitivity and specificity of a nomogram predicted responders to third cTACE: 75.0% and 79.4% (internal validation) and 78.6% and 87.0% (external validation), respectively.

Conclusion

Three sessions were recommended before abandoning conventional transarterial embolization (cTACE) for intermediate-stage hepatocellular carcinoma. The nomogram developed in this study identified responders to third cTACE.

Online supplemental material is available for this article.

See also the editorial by Georgiades in this issue.

Summary

For nonresponding intermediate-stage hepatocellular carcinoma, three sessions of transarterial chemoembolization were recommended before abandonment.

Key Results

■ In a retrospective analysis of 4154 patients treated at three centers, 50% of patients who did not respond (nonresponders) to the first two transarterial chemoembolization (TACE) sessions responded to the third TACE session, but less than 10% of patients who were nonresponders to the third TACE responded to subsequent sessions.
■ Compared with nonresponders, responders to first, second, and third TACE had higher 5-year overall survival (23.1% vs 12.4%, 24.7% vs 2.4%, and 9.1% vs 3.2%, respectively; all P < .001) but responders to the fourth TACE session did not (0% vs 0%; P = .21).

Introduction

Transarterial chemoembolization (TACE) is recommended for patients with intermediate hepatocellular carcinoma (HCC) (1). Repeated TACE may maximize tumor response and prolong survival (2,3), especially in patients with large and multifocal tumors (4,5). However, some patients do not respond to TACE after several sessions and even exhibit tumor progression during repeated TACE treatments; thus, the optimal time for other potential treatments may be lost. Repeated TACE sessions often lead to treatment-related complications (6), which may negatively influence the survival benefit conferred by TACE. Moreover, the subsequent alternative treatments for patients who are nonresponsive to TACE (hereafter, nonresponders) are limited (7–11). Although many new target drugs and immunotherapies including lenvatinib, regorafenib, cabozantinib, and nivolumab have been shown to be effective as first-line and second-line treatments for advanced HCC (12–15), their resultant survival benefit is modest with high cost. For patients who respond to TACE (hereafter, responders), it is widely acknowledged that TACE should continue. But, to our knowledge, there is no consensus on the number of TACE sessions that should be performed for nonresponding patients before switching to another treatment (16,17).

Georgiades et al (18) found that almost 50% of nonresponders to the first TACE showed a response and improved survival after a second TACE. Thus, they suggested that at least two TACE sessions should be performed before abandoning TACE therapy. The updated European Association for the Study of the Liver guideline also suggests that TACE should not be repeated when substantial necrosis is not obtained after two sessions of TACE, although this recommendation lacks sufficient robust clinical supporting evidence (1). Moreover, several studies with small sample sizes showed that patients who had no response (who had stable disease [SD] or progressive disease [PD]) to the second TACE achieved a radiologic response after the third session and ultimately achieved a good prognosis (5,16,17,19). A previous study (19) evaluated the response rates and clinical outcomes in patients with HCC after the first, second, and third TACE sessions. That study found only 26% of patients underwent a third session of TACE, but 55% of these patients obtained a complete response. However, this study had a small sample size (151 patients, 22 of whom underwent three TACE sessions) and lacked external validation, making a robust conclusion difficult. Therefore, to determine the optimal number of TACE sessions that should be performed for nonresponders to TACE, the radiologic responses and survival outcomes of patients with HCC to different numbers of TACE sessions must be evaluated in a large population.

In our study, we evaluated the radiologic responses and survival outcomes of patients with intermediate-stage HCC to different numbers of conventional TACE (cTACE) sessions (up to six TACE sessions) in a large cohort (4154 patients) with a long follow-up duration and validated a model to identify responders to the last recommended TACE session.

Materials and Methods

Our study was approved by the ethics committee and performed in accordance with the standards of the Declaration of Helsinki. Written informed consent was obtained for the clinical procedure from the patients.

Patients and Design

Between January 2005 and December 2012, consecutive patients with intermediate HCC who underwent TACE as the initial treatment at one academic hospital (First Affiliated Hospital, Sun Yat-sen University) were retrospectively screened for eligibility according to our criteria. The diagnosis of HCC was on the basis of noninvasive or histologic criteria outlined by the most current American Association for the Study of Liver Disease guidelines at the time of treatment (20,21).

Patients who met the following criteria were included: age 18 years or older at the time of the first cTACE session, Barcelona Clinic Liver Cancer stage B (multinodular tumors beyond Barcelona Clinic Liver Cancer stage A without macrovascular invasion or extrahepatic metastases), underwent consecutive cTACE treatments, Child-Pugh class A liver function before first cTACE, performance status score of 0, and detailed follow-up imaging data available for evaluating the radiologic response after cTACE.

Exclusion criteria were as follows: died before the evaluation of the radiologic response of first cTACE (n = 2) because we sought to control for guarantee-time bias, which reflects that patients who die early do not have an opportunity to enter the group of responders, resulting in shorter survival among nonresponders than responders; underwent any anticancer treatments before cTACE; history of any other concurrent malignancies; severe coagulopathy (prothrombin activity < 40% or a platelet count of < 40 000/mm³); or complications with severe dysfunction of the heart, kidney, or other organs.

We investigated the following outcomes in this cohort: the radiologic response to first cTACE; the overall survival of responders versus nonresponders to first cTACE; among the nonresponders to the previous first, second, third, fourth, and fifth session of cTACE but who underwent first cTACE, their radiologic responses to first cTACE session (eg, the radiologic responses to second cTACE among nonresponders to first cTACE); and among the nonresponders to the previous first, second, and third cTACE session but who underwent the first cTACE session, the overall survival of responders versus nonresponders to the first cTACE session. Finally, we developed a prediction model for identifying responders to the last recommended cTACE session. To validate the recommended number of cTACE sessions and the developed model, we used data from patients at the other two academic centers (Cancer Center, Sun Yat-sen University; Dongguan People’s Hospital) who met the above inclusion and exclusion criteria between January 2011 and December 2011 as the validation cohort. The workflow of our study design is reported in Figure 1.

Flowchart of the (a) primary cohort and (b) validation cohort in our study. cTACE = conventional transarterial chemoembolization, HCC = hepatocellular carcinoma. — Figure 1a: Flowchart of the **(a)** primary cohort and **(b)** validation cohort in our study. cTACE = conventional transarterial chemoembolization, HCC = hepatocellular carcinoma.

TACE Procedure

Contrast-enhanced CT (Aquilion 64; Toshiba Medical System) or 3.0-T MRI (Siemens Healthcare) was performed within 2 weeks before each cTACE session to evaluate the tumor conditions. In each center, cTACE was performed on demand and was scheduled at an interval of 6–12 weeks with the demonstration of viable tumors at contrast-enhanced CT or MRI in patients with acceptable clinical and laboratory findings (eg, performance status and liver function). cTACE was stopped if one of the following end points was reached: complete response, technical impossibility to embolize the residual tumor (eg, a tumor only supplied by extrahepatic collateral arteries), patient developed contraindications to cTACE (ie, deterioration of liver function to Child-Pugh C or Child-Pugh B decompensated cirrhosis; performance status score ≥ 2; development of vascular invasion, extrahepatic metastases, or extensive liver involvement; and impaired portal vein blood flow), or total resection or ablation of the tumor by subsequent surgical procedure or local ablation.

TACE was performed by two radiologists with more than 10 years of experience in interventional therapy at each center (22). The detailed procedure is described in Appendix E1 (online).

Response Assessment

The radiologic tumor response to cTACE was evaluated 4–6 weeks after each treatment session by using modified Response Evaluation Criteria in Solid Tumors (23,24). Assessment of tumor response was performed in the target lesions at each center by two experienced radiologists (Jiaping Li and S.F., with 25 and 20 years of experience in liver imaging, respectively) who were blinded for other imaging test results independently within 2–4 days. Any assessment results inconsistency was resolved by discussion. Tumor response was quantitatively defined as complete response, partial response, SD, or PD (24). Patients with complete response or partial response were defined as responders, whereas patients with SD or PD were defined as nonresponders.

Patient Follow-up

The patients were followed up at contrast-enhanced CT or MRI 4–6 weeks after each cTACE treatment. When the cTACE treatment ended, patients were followed up once every 3–4 months thereafter. At each follow-up visit, blood tests including liver function tests and α-fetoprotein measurements were performed. Chest radiography was performed every 6 months. Chest CT and bone imaging were performed when the patient was clinically suspected of having extrahepatic metastasis. If cTACE could not be continued because of technical issues or contraindications, patients were recommended to undergo treatments such as best supportive treatment, sorafenib, and chemotherapy. Detailed postprogression therapies are shown in Figure 1.

Demographics, etiologic causes of HCC, liver function parameters, tumor characteristics, and general performance status for each patient (Table 1) were recorded and updated before every treatment session. Tumor characteristics including the number of lesions, tumor size, tumor distribution, and extent of the tumor and tumor capsule were defined by a radiologist experienced in liver disease and a hepatologist experienced in liver imaging after reviewing radiologic images.

**Table 1: Baseline Characteristics of Patients at First Conventional Transarterial Chemoembolization Session**

The adverse events were evaluated according to National Cancer Institute Common Toxicity Criteria, version 4.0 (25). TACE-related death was defined as death within 30 days after the corresponding treatment session. Overall survival was defined as the interval from the date of each cTACE to the date of death or last follow-up. This study was censored on September 15, 2019.

Statistical Analysis

All statistical analyses were performed by using software (R version 3.2.2, Bell Laboratories, http://cran.r-project.Org/bin/windows/base/old/3.2.2; and SPSS 20.0 for Windows, SPSS). Statistical significance was indicated by a two-sided P value less than .05. Differences in baseline characteristics between training and validation cohorts were compared with the t test for continuous variables and with the χ² test for categorical variables. In the entire population, the radiologic response rates after each first cTACE session were estimated in patients with SD or PD at the previous first, second, third, fourth, and fifth cTACE session. Patients were divided into two groups according to their radiologic response to each cTACE session: responders (complete response and partial response) and nonresponders (SD and PD). The analysis of overall survival was performed between these two groups. Survival curves were generated by the Kaplan-Meier method and compared by the log-rank test. Univariable and multivariable Cox proportional hazard regression models were developed to identify significant prognostic factors of overall survival. On the basis of our analyses, we sought to establish the optimal number of cTACE sessions that should be performed before abandoning this therapy.

Development and Validation of the Nomogram in Predicting Radiologic Response to Previous cTACE

We performed the following analyses in the subpopulation who did not respond to the previous first cTACE but underwent previous cTACE. In this subpopulation, we developed the nomogram in the training cohort and validated it in the internal validation cohort and external validation cohort. The least absolute shrinkage and selection operator regression method was used in the training data set to select the most significant predictive factors of a radiologic response to previous cTACE (26). Univariable and multivariable logistic regression analyses were performed to analyze the associations of these factors with the radiologic response to previous cTACE. To provide the clinician with a quantitative tool to predict individual probability to respond to previous session of cTACE, a nomogram was established on the basis of the results of the multivariable logistic regression analysis. The performance of the nomogram was evaluated with the C-index, the area under the receiver operating characteristic curve, sensitivity, specificity, and a calibration curve that compared nomogram-predicted versus observed probability of the radiologic response to previous cTACE. Bootstraps with 1000 resamples were used for the validation of nomogram and calibration curve construction.

Results

Baseline Characteristics

A total of 5220 patients with HCC underwent TACE during the study period, and 885 and 181 patients were excluded from, respectively, the primary and validation cohorts (Fig 1). Finally, 3442 (mean age, 58 years ± 6; 3129 men) and 712 (mean age, 58 years ± 7; 648 men) patients formed the primary and validation cohorts, respectively. The baseline characteristics of the patients at the time of first cTACE in these cohorts are described in Table 1. Hepatitis B virus infection was the most common cause of HCC (3484 of 4154; 83.9%), and approximately half of the patients in each of the two cohorts underwent antiviral treatment with nucleoside analog.

Optimal Number of TACE Sessions before Abandoning cTACE

Radiologic response.—A median of four treatment cycles (range, three to six cycles) of cTACE was performed in patients in the two cohorts. The detailed patient treatment is shown in Figure 2 and the detailed radiologic responses are summarized in Table 2. The response rate to each cTACE between primary and validation cohorts was comparable (all P > .05; Table E1 [online]).

A summary of the treatments of patients with hepatocellular carcinoma (HCC) in (a) the primary cohort and (b) the validation cohort. BSC = best supportive care, cTACE-1 = first conventional transarterial chemoembolization (cTACE) session, cTACE-2 = second cTACE session, cTACE-3 = third cTACE session, cTACE-4 = fourth cTACE session, cTACE-5 = fifth cTACE session, cTACE-6 = sixth cTACE session, RT = radiation therapy. — Figure 2a: A summary of the treatments of patients with hepatocellular carcinoma (HCC) in **(a)** the primary cohort and **(b)** the validation cohort. BSC = best supportive care, cTACE-1 = first conventional transarterial chemoembolization (cTACE) session, cTACE-2 = second cTACE session, cTACE-3 = third cTACE session, cTACE-4 = fourth cTACE session, cTACE-5 = fifth cTACE session, cTACE-6 = sixth cTACE session, RT = radiation therapy.

**Table 2: Radiologic Response of Nonresponders to Previous cTACE after Next cTACE Session**

For the primary cohort (n = 3442), the response rate to first cTACE was 1227 of 3442 (35.6%; 95% CI: 33.7, 37.7). Among patients with SD to first cTACE, the response rate after second cTACE was 719 of 1560 (46.1%; 95% CI: 42.8, 49.6), whereas for patients with PD, the corresponding response rate was 18 of 655 (2.7%; 95% CI: 1.6, 4.3). Among patients with SD to second cTACE, the response rate after third cTACE was 591 of 1014 (58.3%; 95% CI: 53.7, 63.2), whereas for patients with PD to second cTACE, the corresponding rate was 10 of 364 (2.7%; 95% CI: 1.3, 5.1). Subsequently, among patients with SD to third cTACE, fourth cTACE, and fifth cTACE sessions, the corresponding rates after fourth cTACE, fifth cTACE, and sixth cTACE sessions were all low, at 32 of 343 (9.3%; 95% CI: 6.4, 13.2), one of 32 (3%; 95% CI: 0.08, 17.4) and zero of eight (0%; 95% CI: 0, 46.1), respectively.

For the validation cohort (n = 712), the response rate to first cTACE was 261 of 712 (36.7%; 95% CI: 32.3, 41.4). Among patients with SD to first cTACE, the response rate after second cTACE was 147 of 304 (48.4%; 95% CI: 40.9, 56.8), whereas for patients with PD, the corresponding response rate was six of 147 (4.1%; 95% CI: 1.5, 8.9). Among patients with SD to second cTACE, the response rate after third cTACE was 98 of 202 (48.5%; 95% CI: 39.4, 59.1), whereas for patients with PD to second cTACE, the corresponding rate was two of 92 (2%; 95% CI: 2.6, 7.9). Nevertheless, among patients with SD to third, fourth, and fifth cTACE sessions, the corresponding rates after fourth, fifth, and sixth cTACE sessions were all low, at four of 53 (8%; 95% CI: 2, 19), one of 12 (8%; 95% CI: 2, 47), and zero of six (0%; 95% CI: 0, 62), respectively.

Overall, approximately 50% (866 of 1864) of patients who were nonresponsive to one or two cTACE sessions responded to the second cTACE and third cTACE sessions, respectively, whereas less than 10% of patients who were nonresponsive to three, four, or five cTACE sessions responded to fourth, fifth, and sixth cTACE sessions, respectively, indicating that three cTACE sessions may be optimal for observing a response for nonresponders.

Overall survival stratified by radiologic response.—The median follow-up duration was 34.6 months (range, 6.0–96.0 months) for the primary cohort and 34.2 months (range, 5.7–90.0 months) for the validation cohort. The median overall survival of primary and validation cohorts was not significantly different (all P > .05; Table E1 [online]). The median overall survival of responders to the first cTACE session was comparable to that of responders to the second cTACE session in both the primary (41.2 vs 43.2 months, respectively; P = .34) and validation cohorts (42.5 vs 42.5 months, respectively; P = .32; Table 3).

**Table 3: Survival Outcomes for Responders and Nonresponders in Primary and Validation Cohorts**

The median overall survival of responders to the third cTACE session was also comparable to that of responders to the first cTACE (43.4 vs 42.5 months, respectively; P = .46) and second cTACE (43.4 vs 42.5 months, respectively; P = .24) in the validation cohort despite that the overall survival was about 6 months shorter than that of responders to first cTACE (35.7 vs 41.2 months, respectively; P < .001) and second cTACE (35.7 vs 43.2 months, respectively; P < .001) in the primary cohort. Concordantly, the median overall survival of responders in the primary and validation cohorts, respectively, to the fourth cTACE session was significantly shorter than that of responders to the first cTACE session (16.6 vs 41.2 months, P < .001; and 18.7 vs 42.5 months, P < .001), second cTACE (16.6 vs 43.2 months, P < .001; and 18.7 vs 42.5 months, P = .004), and third cTACE session (16.6 vs 43.4 months, P < .001; and 18.7 vs 43.4 months, P < .001).

In the primary cohort, the 1-, 3-, and 5-year overall survival rates were higher among responders to first cTACE than nonresponders (P < .001; Table 3, Fig 3). For patients with SD or PD to first cTACE but who underwent second cTACE, the 1-, 3-, and 5-year overall survival rates after second cTACE were higher for responders than for nonresponders (P < .001; Table 3, Fig 3). For patients with SD or PD to second cTACE but who underwent third cTACE, the responders to third cTACE achieved higher 1-, 3-, and 5-year overall survival rates than did the nonresponders (all P < .001; Table 3, Fig 3). However, for patients with SD or PD to third cTACE but who underwent fourth cTACE, the responders to fourth cTACE yielded similar long-term survival outcomes to the nonresponders, without a difference (P = .21; Table 3, Fig 3).

Kaplan-Meier survival curves of responders and nonresponders to conventional transarterial chemoembolization (cTACE) for patients with intermediate-stage hepatocellular carcinoma. Kaplan-Meier survival curves of responders and nonresponders to first cTACE in the (a) primary and (b) validation cohorts; Kaplan-Meier survival curves of responders and nonresponders to second cTACE among nonresponding patients to first cTACE but who underwent second cTACE in the (c) primary and (d) validation cohorts. Kaplan-Meier survival curves of responders and nonresponders to third cTACE among nonresponding patients to second cTACE but who underwent third cTACE in the (e) primary and (f) validation cohorts. Kaplan-Meier survival curves of responders and nonresponders to fourth cTACE among nonresponding patients to third cTACE but who underwent fourth cTACE in the (g) primary and (h) validation cohorts. Cum = cumulative. — Figure 3a: Kaplan-Meier survival curves of responders and nonresponders to conventional transarterial chemoembolization (cTACE) for patients with intermediate-stage hepatocellular carcinoma. Kaplan-Meier survival curves of responders and nonresponders to first cTACE in the **(a)** primary and **(b)** validation cohorts; Kaplan-Meier survival curves of responders and nonresponders to second cTACE among nonresponding patients to first cTACE but who underwent second cTACE in the **(c)** primary and **(d)** validation cohorts. Kaplan-Meier survival curves of responders and nonresponders to third cTACE among nonresponding patients to second cTACE but who underwent third cTACE in the **(e)** primary and **(f)** validation cohorts. Kaplan-Meier survival curves of responders and nonresponders to fourth cTACE among nonresponding patients to third cTACE but who underwent fourth cTACE in the **(g)** primary and **(h)** validation cohorts. Cum = cumulative.

In the validation cohort, responders to third cTACE achieved better long-term survival outcomes than nonresponders (P < .001), whereas responders to fourth cTACE yielded similar outcomes to nonresponders (P = .91; Table 3, Fig 3). In the survival comparisons of fourth cTACE, we estimated the power of the sample size in detecting statistical significance and found that the power was 6.6% in the primary cohort and 5.8% in the validation cohort.

Radiologic response to third cTACE session was a negative prognostic factor of death.—To analyze the association of radiologic response with death, we performed univariable and multivariable analyses. Univariable and multivariable analyses demonstrated that the following were independent prognostic factors of death (Table 4): ascites (hazard ratio, 2.1; 95% CI: 1.6, 2.9; P < .001), number of tumors (hazard ratio, 1.9; 95% CI: 1.6, 2.3; P < .001), tumor capsule (hazard ratio, 0.77; 95% CI: 0.65, 0.91; P = .003), up to seven criteria (hazard ratio, 0.79; 95% CI: 0.66, 0.94; P = .01), α-fetoprotein level (hazard ratio, 1.9; 95% CI: 1.6, 2.3; P < .001), and radiologic response to third cTACE (hazard ratio, 0.15; 95% CI: 0.12, 0.20; P < .001).

**Table 4: Predictors of Death for Intermediate-Stage HCC in Patients Undergoing Consecutive cTACE Sessions**

Identifying Potential Responders to Third cTACE

In our analyses, 1672 patients who did not respond to the first two cTACE sessions but who underwent a third cTACE session were included, with 918 patients in the training cohort, 460 in the internal validation cohort, and 294 in the external validation cohort (Fig E1, Table E2 [online]).

Subgroup analysis of the radiologic response.—Subgroup analysis showed that the response rates after third cTACE were higher for patients in the training and internal validation cohort (Table E3 [online]), respectively, with tumors 5 cm or smaller (330 of 776 [42.5%; 95% CI: 38.1, 47.4] vs 130 of 602 [21.6%; 95% CI: 18.0, 25.6], P < .001), with α-fetoprotein levels of 20–200 ng/mL versus less than 20 ng/mL and more than 200 ng/mL (263 of 611 [43.0%; 95% CI: 38.0, 48.6] vs 63 of 168 [37.5%; 95% CI: 28.8, 48.0] and 134 of 597 [22.4%; 95% CI: 18.8, 26.6%], respectively; P < .001), or with tumors in which the capsule was present (304 of 674 [45.1%; 95% CI: 40.2, 50.5] vs 155 of 544 [28.5%; 95% CI: 24.2, 33.4], P < .001). Subgroup analysis yielded similar results for the external validation cohort (Table E3 [online]).

Nomogram for predicting the radiologic response to third cTACE.—By considering the prognostic effect of the radiologic response in overall survival, we sought to develop a nomogram to predict the radiologic response to third cTACE. Least absolute shrinkage and selection operator was used, and tumor size, tumor capsule, and α-fetoprotein level were found to be associated with the radiologic response to third cTACE in the training cohort. A nomogram integrating these three factors was constructed (Fig 4a). The C-index for predicting the radiologic response to third cTACE was 0.84 (95% CI: 0.81, 0.86) in the training cohort, 0.85 (95% CI: 0.82, 0.89) in the internal validation cohort, and 0.85 (95% CI: 0.83, 0.89) in the external validation cohort. The area under the receiver operating characteristic curve of the nomogram in predicting response to third cTACE in the internal validation cohort was 0.77 (95% CI: 0.72, 0.82) with a sensitivity of 75.0% (95% CI: 65.3, 83.1) and a specificity of 79.4% (95% CI: 73.0, 84.8). The area under the receiver operating characteristic curve of the external validation cohort was 0.83 (95% CI: 0.79, 0.86) with a sensitivity of 78.6% (95% CI: 72.6, 83.9) and a specificity of 87.0% (95% CI: 82.1, 91.0). The calibration curves fit well between the nomogram prediction and the actual observations in the training and internal and external validation cohorts (Fig 4b–4d).

(a) A nomogram for predicting the radiologic response to third conventional transarterial chemoembolization (cTACE) in patients with intermediate-stage HCC who did not respond to either of the first two cTACE sessions. (b–d) Calibration curves for predicting the radiologic response to third cTACE in the (b) training, (c) internal validation, and (d) external validation cohorts. The nomogram-predicted probability of the radiologic response to third cTACE is plotted on the x-axis; the actual radiologic response to third cTACE is plotted on the y-axis. AFP = α-fetoprotein. — Figure 4a: **(a)** A nomogram for predicting the radiologic response to third conventional transarterial chemoembolization (cTACE) in patients with intermediate-stage HCC who did not respond to either of the first two cTACE sessions. **(b–d)** Calibration curves for predicting the radiologic response to third cTACE in the **(b)** training, **(c)** internal validation, and **(d)** external validation cohorts. The nomogram-predicted probability of the radiologic response to third cTACE is plotted on the x-axis; the actual radiologic response to third cTACE is plotted on the y-axis. AFP = α-fetoprotein.

Complications after Third cTACE

Hypoalbuminemia, anemia, fever, thrombocytopenia, and elevated aspartate aminotransferase or alanine aminotransferase were the five most common complications after third cTACE (Table E4 [online]). The most common grade 3–4 adverse events in the training, internal validation, and external validation cohorts were, respectively, elevated aspartate aminotransferase or alanine aminotransferase (207 of 918 [22.5%], 98 of 460 [21.3%], and 65 of 294 [22.1%]), thrombocytopenia (51 of 918 [5.6%], 32 of 460 [7.0%], and 19 of 294 [6.5%]), and pain (41 of 918 [4.5%], 26 of 460 [5.6%], and 18 of 294 [6.1%]). Two other complications associated with liver function (hyperbilirubinemia and hypoalbuminemia) were mainly grade 1–2 severity. The complications after the other cTACE sessions are summarized in Tables E5–E9 (online). Two patients died of acute myocardial infarction and acute pulmonary embolism before the evaluation of first cTACE.

Discussion

Transarterial chemoembolization (TACE) is the standard treatment for intermediate-stage hepatocellular carcinoma (HCC). But, to our knowledge, it remains unknown whether conventional TACE (cTACE) should be continued or abandoned after initial nonresponse for intermediate-stage HCC. The optimal number of sessions before abandoning cTACE is also debatable. Our study demonstrated that approximately 50% of nonresponders to the first two cTACE sessions and with good liver function could respond to third cTACE with better long-term survivals than persistent nonresponders to the first three cTACE sessions. However, fewer than 10% of nonresponders to third cTACE responded to the fourth or subsequent cTACE sessions, indicating that three cTACE sessions should be performed before switching to another treatment for nonresponding patients. The radiologic response to third cTACE determined overall survival. Our developed nomogram constructed by tumor size, α-fetoprotein level, and tumor capsule was effective in identifying patients who would potentially respond to third cTACE (sensitivity, 75.0%, and specificity, 79.4%, in the internal validation cohort; sensitivity, 78.6%, and specificity, 87.0%, in the external validation cohort).

Half of the nonresponders to the first two cTACE sessions responded to third cTACE, but a low proportion of nonresponders to third cTACE responded to the fourth cTACE session. These patients had high tumor burdens with large (mean size, 6.0 cm; range, 2.0–15.0 cm) and multiple lesions (>3), in which two cTACE sessions might not induce substantial necrosis. The third cTACE session appeared to be the testing session in which patients sensitive to cTACE treatment were identified, and we observed a dramatic decline in the proportion of patients who responded to the fourth cTACE if they did not respond to third cTACE. Unlike the recommendation from the European Association for the Study of the Liver that cTACE should not be repeated when substantial necrosis is not achieved after two rounds of treatment (1), we identified a subgroup of patients who were nonresponsive to the first two cTACE sessions but in whom an additional session was effective for inducing a radiologic response. These responding patients were mainly those with SD (50% and <10% of responding rate for SD and PD, respectively). For patients with SD whose tumors were under good control with a disease pathologic profile that may not be aggressive, it is possible that one additional session of cTACE (third cTACE) could induce vulnerability of the tumor to embolization and chemotherapy, resulting in a tumor response to third cTACE. For patients with PD whose tumors were refractory to cTACE and with an underlying biology that may be tolerant to ischemic stress and nonsensitive to chemotherapeutic agents, one additional cTACE session is trivial and likely to be ineffective. Therefore, the appropriate candidates to undergo at least three cTACE sessions among nonresponders to the first two cTACE sessions are those with SD. Moreover, among the patients with SD, patients with tumors 5 cm or smaller, with an α-fetoprotein level of 200 ng/mL or less, or with a tumor capsule might be more suitable candidates for undergoing a third cTACE.

Even patients with SD after the second cTACE obtained a good prognosis if they responded to third cTACE with a median overall survival of 43.4 months. Tumor response to third cTACE is a major determinant of long-term survival and not all patients could achieve a response from third cTACE, so we developed a nomogram to identify potential responders to third cTACE. Our nomogram has a high predictive accuracy and is easy to use, with only three variables integrated: tumor size, serum α-fetoprotein level, and tumor capsule. Our nomogram had several differences compared with the previously reported Assessment for Retreatment with TACE score and a score consisting of α-fetoprotein, Barcelona Clinic Liver Cancer stage, Child-Pugh class, and radiologic response (referred to as ABCR) (27–29). First, our nomogram does not consider the radiologic response to the previous TACE session because the tumor response to the second cTACE might not necessarily be correlated with the response to the third session in our study. Second, Assessment for Retreatment with TACE and ABCR scores incorporate not only the radiologic response and tumor characteristics but also liver function parameters. Assessment for Retreatment with TACE and ABCR are controversial regarding the weighting of the radiologic response and the Child-Pugh score increase in patient stratification (27,28). However, different weightings of these factors lead to different results, which may make these scores difficult to use and limit their generalization to different populations. Third, to our knowledge, our nomogram was established on the basis of the largest population with relatively homogeneous patients at Barcelona Clinic Liver Cancer stage B (n = 918), whereas Assessment for Retreatment with TACE and ABCR scores were on the basis of 107 and 139 patients, respectively, with different Barcelona Clinic Liver Cancer stages.

Our study had limitations. First, it was a retrospective study with inherent defects. Second, most of our patients had hepatitis B virus, and our findings may need to be validated in patients with other diseases such as hepatitis C and alcoholic liver disease. Third, there were surviving nonresponders to TACE sessions who did not get treated with a subsequent TACE, which would introduce potential selection bias in our analyses. Fourth, the sample size in some subgroups (ie, the survival analysis for fourth TACE) was limited and therefore lacked enough power to detect significance.

In conclusion, we recommend three conventional transarterial chemoembolization (cTACE) sessions before switching to another treatment for nonresponding patients with intermediate-stage hepatocellular carcinoma. Our developed nomogram shows good performance for predicting the radiologic response to the third cTACE session, which can help identify potential responding patients to third cTACE with a good prognosis, and indicate that nonresponding patients avoid third cTACE and receive other alternative treatments in time.

Disclosures of Conflicts of Interest: S.C. disclosed no relevant relationships. Z.P. disclosed no relevant relationships. Y.Z. disclosed no relevant relationships. M.C. disclosed no relevant relationships. Jiaping Li disclosed no relevant relationships. R.G. disclosed no relevant relationships. Jiali Li disclosed no relevant relationships. B.L. disclosed no relevant relationships. J.M. disclosed no relevant relationships. S.F. disclosed no relevant relationships. M.K. disclosed no relevant relationships.

Author Contributions

Author contributions: Guarantors of integrity of entire study, Z.P., Y.Z., Jiaping Li, Jiali Li, M.K.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, S.C., Z.P., Y.Z., Jiaping Li, Jiali Li, M.K.; clinical studies, S.C., Z.P., Y.Z., M.C., Jiaping Li, R.G., Jiali Li, S.F., M.K.; experimental studies, S.C., Y.Z., Jiali Li; statistical analysis, S.C., Z.P., Y.Z., Jiali Li, B.L.; and manuscript editing, S.C., Z.P., Y.Z., M.C., Jiaping Li, R.G., Jiali Li, J.M., M.K.

* S.C. and Z.P. contributed equally to this work.

Study supported by the National Natural Science Foundation of China (grants 81801703 and 81770608), National Science Fund for Distinguished Young Scholars (grant 81825013), National high level talents special support plan (“Ten thousand plan”), Young top-notch talent support program, Natural Science Foundation of Guangdong Province (grant 2018A030310282), and Kelin Outstanding Young Scientist of the First Affiliated Hospital, Sun Yat-sen University (2017).

References

1. European Association for the Study of the Liver. EASL clinical practice guidelines: Management of hepatocellular carcinoma. J Hepatol 2018;69(1):182–236. Crossref, Medline, Google Scholar
2. Park JW, Chen M, Colombo M, et al. Global patterns of hepatocellular carcinoma management from diagnosis to death: the BRIDGE Study. Liver Int 2015;35(9):2155–2166. Crossref, Medline, Google Scholar
3. Vogl TJ, Trapp M, Schroeder H, et al. Transarterial chemoembolization for hepatocellular carcinoma: volumetric and morphologic CT criteria for assessment of prognosis and therapeutic success-results from a liver transplantation center. Radiology 2000;214(2):349–357. Link, Google Scholar
4. Kim DY, Ryu HJ, Choi JY, et al. Radiological response predicts survival following transarterial chemoembolisation in patients with unresectable hepatocellular carcinoma. Aliment Pharmacol Ther 2012;35(11):1343–1350. Crossref, Medline, Google Scholar
5. Choi J, Shim JH, Shin YM, Kim KM, Lim YS, Lee HC. Clinical significance of the best response during repeated transarterial chemoembolization in the treatment of hepatocellular carcinoma. J Hepatol 2014;60(6):1212–1218 [Published retraction appears in J Hepatol 2015;62(1):252.]. Crossref, Medline, Google Scholar
6. Kwok PC, Lam TW, Chan SC, et al. A randomized clinical trial comparing autologous blood clot and gelfoam in transarterial chemoembolization for inoperable hepatocellular carcinoma. J Hepatol 2000;32(6):955–964. Crossref, Medline, Google Scholar
7. Ikeda M, Mitsunaga S, Shimizu S, et al. Efficacy of sorafenib in patients with hepatocellular carcinoma refractory to transcatheter arterial chemoembolization. J Gastroenterol 2014;49(5):932–940. Crossref, Medline, Google Scholar
8. Kim YI, Park JW, Kwak HW, et al. Long-term outcomes of second treatment after initial transarterial chemoembolization in patients with hepatocellular carcinoma. Liver Int 2014;34(8):1278–1286. Crossref, Medline, Google Scholar
9. Kang JK, Kim MS, Cho CK, et al. Stereotactic body radiation therapy for inoperable hepatocellular carcinoma as a local salvage treatment after incomplete transarterial chemoembolization. Cancer 2012;118(21):5424–5431. Crossref, Medline, Google Scholar
10. Hatooka M, Kawaoka T, Aikata H, et al. Comparison of outcome of hepatic arterial infusion chemotherapy and sorafenib in patients with hepatocellular carcinoma refractory to transcatheter arterial chemoembolization. Anticancer Res 2016;36(7):3523–3529. Medline, Google Scholar
11. Klompenhouwer EG, Dresen RC, Verslype C, et al. Safety and Efficacy of Transarterial Radioembolisation in Patients with Intermediate or Advanced Stage Hepatocellular Carcinoma Refractory to Chemoembolisation. Cardiovasc Intervent Radiol 2017;40(12):1882–1890. Crossref, Medline, Google Scholar
12. Kudo M, Finn RS, Qin S, et al. Lenvatinib versus sorafenib in first-line treatment of patients with unresectable hepatocellular carcinoma: a randomised phase 3 non-inferiority trial. Lancet 2018;391(10126):1163–1173. Crossref, Medline, Google Scholar
13. El-Khoueiry AB, Sangro B, Yau T, et al. Nivolumab in patients with advanced hepatocellular carcinoma (CheckMate 040): an open-label, non-comparative, phase 1/2 dose escalation and expansion trial. Lancet 2017;389(10088):2492–2502. Crossref, Medline, Google Scholar
14. Bruix J, Qin S, Merle P, et al. Regorafenib for patients with hepatocellular carcinoma who progressed on sorafenib treatment (RESORCE): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet 2017;389(10064):56–66. Crossref, Medline, Google Scholar
15. Kelley RK, Verslype C, Cohn AL, et al. Cabozantinib in hepatocellular carcinoma: results of a phase 2 placebo-controlled randomized discontinuation study. Ann Oncol 2017;28(3):528–534. Crossref, Medline, Google Scholar
16. Wang W, Zhao Y, Bai W, Han G. Response assessment for HCC patients treated with repeated TACE: The optimal time-point is still an open issue. J Hepatol 2015;63(6):1530–1531. Crossref, Medline, Google Scholar
17. Kim BK, Kim SU, Kim KA, et al. Complete response at first chemoembolization is still the most robust predictor for favorable outcome in hepatocellular carcinoma. J Hepatol 2015;62(6):1304–1310. Crossref, Medline, Google Scholar
18. Georgiades C, Geschwind JF, Harrison N, et al. Lack of response after initial chemoembolization for hepatocellular carcinoma: does it predict failure of subsequent treatment? Radiology 2012;265(1):115–123. Link, Google Scholar
19. 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(6):1258–1267. Crossref, Medline, Google Scholar
20. Bruix J, Sherman M, Llovet JM, et al. Clinical management of hepatocellular carcinoma. Conclusions of the Barcelona-2000 EASL conference. European Association for the Study of the Liver. J Hepatol 2001;35(3):421–430. Crossref, Medline, Google Scholar
21. Bruix J, Sherman M; Practice Guidelines Committee, American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma. Hepatology 2005;42(5):1208–1236. Crossref, Medline, Google Scholar
22. Peng ZW, Zhang YJ, Chen MS, et al. Radiofrequency ablation with or without transcatheter arterial chemoembolization in the treatment of hepatocellular carcinoma: a prospective randomized trial. J Clin Oncol 2013;31(4):426–432. Crossref, Medline, Google Scholar
23. Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment for hepatocellular carcinoma. Semin Liver Dis 2010;30(1):52–60. Crossref, Medline, Google Scholar
24. 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(6):1181–1187. Crossref, Medline, Google Scholar
25. National Cancer Institute Common Toxicity Criteria. Version 4.0. Cancer Therapy Evaluation Program. http://ctep.cancer.gov/protocolDevelopment/electronic_applications/ctc.htm. Published 2009. Accessed January 5, 2021. Google Scholar
26. Sauerbrei W, Royston P, Binder H. Selection of important variables and determination of functional form for continuous predictors in multivariable model building. Stat Med 2007;26(30):5512–5528. Crossref, Medline, Google Scholar
27. Sieghart W, Hucke F, Pinter M, et al. The ART of decision making: retreatment with transarterial chemoembolization in patients with hepatocellular carcinoma. Hepatology 2013;57(6):2261–2273. Crossref, Medline, Google Scholar
28. Adhoute X, Penaranda G, Naude S, et al. Retreatment with TACE: the ABCR SCORE, an aid to the decision-making process. J Hepatol 2015;62(4):855–862. Crossref, Medline, Google Scholar
29. Kudo M, Arizumi T, Ueshima K. Assessment for retreatment (ART) score for repeated transarterial chemoembolization in patients with hepatocellular carcinoma. Hepatology 2014;59(6):2424–2425. Crossref, Medline, Google Scholar

Article History

Received: May 19 2020
Revision requested: June 24 2020
Revision received: Sept 11 2020
Accepted: Oct 29 2020
Published online: Jan 19 2021
Published in print: Mar 2021

Video Summary »

Vol. 298, No. 3

Supplemental Material

Abbreviations

Abbreviations:
cTACE	conventional TACE
HCC	hepatocellular carcinoma
PD	progressive disease
SD	stable disease
TACE	transarterial chemoembolization

Metrics

Downloaded 1,127 times

Altmetric Score