Transarterial chemoembolization as an alternative to radioembolization is associated with earlier tumor recurrence than in radioembolization-eligible patients

Introduction Although transarterial radioembolization (TARE) using yttrium-90 (90Y) is a treatment option for large hepatocellular carcinoma (HCC), a fraction of patients are ineligible for TARE due to high lung shunt fraction (LSF). Methods We evaluated if treatment with transarterial chemoembolization (TACE), owing to TARE ineligibility was associated with early HCC progression. Consecutive patients with HCC who were initially TARE candidates were included. Patients with vascular invasion or metastasis were excluded. Primary endpoints were time-to-progression (TTP) and overall survival (OS). The secondary endpoint was objective response rate. Results In total, 175 patients were included: 144 underwent TARE (TARE-eligible group) and 31 underwent TACE due to high LSF (TARE-ineligible group). This latter group had larger tumors (13.8 cm vs. 7.8 cm, P<0.001) and higher MoRAL scores (1,385.8 vs. 413.3, P=0.002) than the TARE-eligible group. After balancing baseline characteristics with an inverse probability of treatment weighting (IPTW), the TARE-ineligible group showed shorter TTP [adjusted hazard ratio (aHR)=2.16, 95% confidence interval (CI)=1.14–4.07, P=0.02] and OS (aHR=1.80, 95% CI=0.85–3.80, P=0.12), although the latter was not statistically significant. The TARE-ineligible group had a significantly lower objective response rate than the TARE-eligible group (9.7% vs. 56.9%, P<0.001). Conclusion TARE-ineligible patients had larger tumors and higher MoRAL scores than TARE-eligible patients. Treatment with TACE, owing to high LSF, was associated with a shorter TTP even after balancing tumor size and MoRAL scores.


Introduction
Hepatocellular carcinoma (HCC) is the most common primary liver cancer, accounting for approximately 90% of all liver cancers. Annually, 850,000 patients are newly diagnosed with HCC, which is the second leading cause of cancer-related deaths worldwide (1). Several curative and palliative treatments are available for HCC and are selected based on tumor stage, liver function and patient performance (2).
Nevertheless, some patients do not undergo TARE treatment for a variety of reasons. High lung-shunt fraction (LSF) is the leading cause of TARE-ineligibility (9). Other minor causes of TARE-ineligibility include HCC adjacent to bowel adhesion as a result of prior intraabdominal surgery or procedure, considerable arterioportal shunt that may result in extrahepatic deposition of radioactive microspheres, and poor liver function (10,11). High-energy b-rays may cause fatal radiation lung injury and severe radiation pneumonitis if substantial numbers of 90 Y particles pass through shunts and reach the lung (12,13). Thus, TARE is highly contraindicated in patients with high LSF, while a pretreatment simulation study before TARE treatment is required to evaluate TARE eligibility, unlike TACE. The lung shunt level is estimated using LSF of technetium-99m ( 99m Tc)-labeled macroaggregated albumin (MAA) scintigraphy (14). Patients who are ineligible for TARE using this pretreatment test are treated with other modalities, mostly TACE. Because the majority of TARE-eligible patients are Barcelona Clinic Liver Cancer (BCLC) B. In general, these BCLC B patients are unsuitable for hepatic resection and indication for TACE (15,16). High LSF is not a contraindication for TACE since the risk of pulmonary complication due to intratumoral shunting is extremely rare (17). In addition, TACE is performed prior to TARE for patients with high initial LSF to lower LSF in real-world clinical practice (18). This indicates that TACE may be performed safely on patients with high LSF and could successfully embolize abnormally dilated vessels within HCC (19).
In this study, we aimed to compare the outcomes between patients who were eligible to TARE and underwent TARE and those who were ineligible to TARE and underwent TACE as an initial treatment for HCC.

Material and methods Patients
Consecutive patients with HCC who underwent pretreatment simulation studies as candidates for TARE, between September 2009 and March 2021 in a tertiary referral center (Seoul National University Hospital, Seoul, South Korea), were eligible for inclusion. Among eligible patients, those who had low LSF and underwent TARE or who had high LSF and underwent TACE were included. A HCC diagnosis was made according to radiographic or histological findings and followed American Association for the Study of Liver Disease guidelines (15). Exclusion criteria: (i) tumor thrombosis in the portal or hepatic vein, (ii) extrahepatic metastasis, (iii) poor liver function classified as Child-Pugh class C, (iv) poor performance status with a European Cooperative Oncology Group score ≥3, and (v) previous or current malignancies except HCC. Patients whose small HCC was treated with TARE without a pretreatment simulation test were also excluded. Patients who were ineligible for TARE due to anatomical variation (n=2, 5.9% of TARE-ineligible patients) and acute hepatic decompensation (n=1, 2.9% of TARE-ineligible patients) were also excluded ( Figure 1). HCC-related factors (tumor size, BCLC stage) and laboratory data were collected retrospectively (platelet count, aspartate aminotransferase levels, alanine aminotransferase levels, prothrombin time, and albumin, etc.). Model to predict tumor recurrence after living donor transplantation (MoRAL) scores were calculated to estimate the prognosis of HCC: MoRAL score = 11·√protein induced by vitamin K absence-II (PIVKA-II) + 2·√alpha-fetoprotein (AFP). The MoRAL score according to reports, indicates the MoRAL score reflects the aggressiveness and burden of HCC tumors, and is associated with clinical outcomes following various treatments, including TACE and radiofrequency ablation (20)(21)(22).
The study was performed in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Seoul National University Hospital (No: H-2101-093-1189). Written informed consent from patients was waived because clinical data were anonymously analyzed.

Procedures
For TARE candidates, pretreatment simulation tests were performed using hepatic angiography and 99m Tc-labeled MAA (Curium Pharma, London, United Kingdom) scintigraphy. Angiographic evaluations identified feeding arteries supplying the tumor and also non-target arteries that may cause aberrant 90 Y deposition. By injecting 99m Tc-MAA into the optimal position in hepatic arteries, LSF was assessed. The recommended limit was either a lung dose >30 Gy/treatment or a cumulative lung dose of 50 Gy in TheraSphere ® (Boston Scientific, Natick, MA) or 20% of lung shunting for SIR-Spheres ® (Sirtex Medical, Lane Cove, Australia) (23,24). TARE was performed by two experienced interventional radiologists (H.C.K. and M.L.) who had >10 years of experience. Radioactive microspheres (TheraSphere ® and SIR-Spheres ® ) were infused according to previous protocol (23,25). The mean target tissue dose by single compartment dosimetry ranged between 80-360 Gy for glass microspheres. When the patients had good liver function (Child-Pugh class A) and at least 30% of whole liver volume can be saved from irradiation, boosted radioembolization (>150 Gy of mean target tissue dose) was performed to enhance tumor response. Partition dosimetry was adopted for resin microspheres, normal liver dose was kept under 60Gy and absorbed tumor dose ranged between 120-360 Gy.
For TARE-ineligible patients, either conventional or drug-eluting bead (DEB)-TACE was performed according to the operator's decision based on multiple factors (e.g., the insurance policy of South Korea, the actual cost of the procedure, etc.). Both conventional and DEB-TACE were conducted superselectively utilizing a microcatheter with 1.7-2.0 F tip (Progreat Lambda or Alpha; Terumo, Tokyo, Japan) under cone beam CT guidance. In conventional TACE, an emulsion comprising 2-10 mL of iodized oil (Lipiodol; Guerbet, Roissy, France) and 10-50 mg doxorubicin hydrochloride powder (Adriamycin RDF; Ildong Pharmaceutical Co., Seoul, Korea) was slowly injected until tumors were completely stained. As an additional embolization, gelatin sponge particles (150-350 mm or 350-560 mm) was infused until near-stasis was achieved. In DEB-TACE, one or two vials of DEB agent (DC Bead; Boston Scientific, Natick, MA) were used per patient: 70-150 or 100-300 mm-sized DEB agent was administered to each patient based on the interventionist's discretion. Similar to the gelatin sponge used in conventional TACE, DEB agent was slowly infused through tumorfeeding arteries until near-stasis was achieved. Each vial of DEB agent used in the DEB-TACE, was loaded with 50 mg of doxorubicin for 1 hour. This doxorubicin-loaded DEB agent was suspended in a mixture containing 25 mL of normal saline and 25 mL of iodinated contrast agent. No patient utilized anti-reflux device. Further specific TACE procedure is explained elsewhere (26)(27)(28). When residual tumor or disease progression occurred after initial HCC treatment, patients received additional treatments after shared discussion.

Outcomes
Primary outcomes were time-to-progression (TTP) and OS, and the secondary outcome was objective response rate within 6 months after initial treatment with either TARE or TACE. Patient survival data of the patients were obtained from the Ministry of the Interior and Safety of Korea. Responses were assessed using modified Response Evaluation Criteria in Solid Tumors criteria (29). Treatment responses were assessed at 1, 3, and 6 months after initial HCC treatment using multiphase dynamic computed tomography or magnetic resonance imaging. All images were evaluated by one radiologist with >10 years of experience.

Statistical analysis
Chi-squared or Fisher's exact tests were used to compare categorical variables and the Mann-Whitney U test for continuous variables. OS and TTP were compared using the Kaplan-Meier method and log-rank test. The hazard ratio (HR) and its 95% confidence interval (CI) were estimated using the Cox proportional hazards model. Inverse probability of treatment weighting (IPTW) was applied to balance baseline characteristics between TARE-eligible and TARE-ineligible groups (30). The Cox proportional hazards model was used to evaluate independent survival risk factors. OS comparisons between TAREeligible and TARE-ineligible groups and univariable and multivariable Cox analyses were performed in an IPTW-weighted cohort. Correlation between TTP and OS was analyzed using Kendall test.
Analyses were performed using R 4.2.0 (R Foundation for Statistical Computing, Vienna, Austria). All statistical tests were two-sided, and P values <0.05 were considered statistically significant.

Study population
In total, 175 patients were included: 144 patients in the TAREeligible group received TARE [glass microsphere (TheraSphere ® ), Flow Chart. A total of 398 HCC patients undergone pretreatment 99mTc MAA scintigraphy and 223 patients were excluded. A total of 175 patients were analyzed. a The patient underwent a right posterior sectionectomy and caudate lobectomy due to the difficulty in targeting HCC (HCC feeding artery branched from pancreatic artery). b TACE was performed on patient due to anatomical variance (left hepatic artery branching from left gastric artery causing difficult to target). c Due to poor liver function, the patient received salvage living donor liver transplantation (Child-Pugh score 10). 99m Tc, technetium-99m; HCC, hepatocellular carcinoma; MAA, macroaggregated albumin; TACE, transarterial chemoembolization; TARE, transarterial radioembolization.
n=136; 90 Y resin microsphere (SIR-Spheres ® ), n=8], whereas 34 patients were ineligible for TARE. Among the 34 TARE-ineligible patients, 31 patients were ineligible to TARE due to high LSF (TAREineligible group), two patients underwent other procedures (hepatic resection and TACE) due to technical difficulty and one patient was not suitable for TARE due to abrupt deterioration of liver failure and underwent salvage living donor liver transplantation. All patients with high LSF underwent TACE: 28 patients underwent TACE and 3 underwent DEB-TACE. As shown (Table 1), significant differences in several baseline variables were observed, including tumor size,  MoRAL scores, platelet counts, aspartate aminotransferase levels, alanine aminotransferase levels, prothrombin time, and albumin before IPTW. However, after applying IPTW, variables were generally well-balanced, including tumor size ( Table 1).
We performed subgroup analysis based on different factors (i.e., age, BCLC stage, Child-Pugh class, tumor size, and MoRAL score) and similar results regarding TTP and OS were reproduced (Table 4).
Patients with residual tumors or disease progression after initial HCC treatment underwent different treatments ( Table 5)

Discussion
In this study, we compared clinical outcomes between patients eligible and ineligible for TARE as an initial treatment for large HCC. The TARE-ineligible group tended to have larger tumor sizes and higher MoRAL scores than the TARE-eligible group. The TAREineligible group had shorter TTP than the TARE-eligible group in crude analysis, even after balancing baseline variables including tumor size and MoRAL score with IPTW. The risk of death was 35%-92% higher in the TARE-ineligible group than the TAREeligible group, although it failed to reach statistical significance in multivariable and/or IPTW analyses. The objective response rate was significantly higher in the TARE-eligible group than the TAREineligible group. Our study was conducted in a single tertiary center where all treatments and decisions were made by highly experienced physicians and interventionists, something that would not have been possible in small-volume institutions. In terms of study design, these highly qualified specialists enable us to compare TACE and TARE treatments by minimizing the impact of human variables (low proficiency of TARE or TACE techniques).
As stated, patients in the TARE-ineligible group had larger tumors and higher MoRAL scores than the TARE-eligible group. As prognostic factors, large tumor size and high MoRAL scores in the TARE-ineligible group may account for shorter TTP or OS than the TARE-eligible group. Large HCC is associated with a poor prognosis (31,32). Two components of MoRAL scores, AFP and PIVKA, are associated with aggressive tumor behaviors and poor clinical outcomes (33)(34)(35).  Kaplan-Meier methodology was used to estimate time-to-progression between TARE-eligible and TARE-ineligible groups (A) before applying IPTW and (B) after applying IPTW. Propensity score for IPTW were computed using the following variables: tumor size, MoRAL score, Child-Pugh class, and BCLC stage. BCLC, Barcelona Clinic Liver Cancer; IPTW, inverse probability of treatment weighting; MoRAL, model to predict tumor recurrence after living donor transplantation. Kaplan-Meier methodology was used to estimate overall survival between the TARE-eligible and TARE-ineligible groups (A) before applying IPTW and (B) after applying IPTW. Propensity score for IPTW were computed using the following variables: tumor size, MoRAL score, Child-Pugh class, and BCLC stage. BCLC, Barcelona Clinic Liver Cancer; CI, confidence interval; IPTW, inverse probability of treatment weighting; MoRAL, model to predict tumor recurrence after living donor transplantation. LSF may have functioned as a confounder since a higher LSF was associated with more aggressive tumor behavior as well as worse clinical outcomes (36) in previous studies and TARE-eligibility was directly related to LSF. In our study, however, after baseline characteristics were balanced using IPTW, there was no association between LSF and clinical outcomes (i.e., time to progression and overall survival). Based on the traditional definition of confounder (37), which is defined as a pre-exposure variable associated with both exposure and outcome, LSF was not a confounder in our study. However, due to the inadequacy of our study design to investigate the confounding effect of LSF, validation requires additional study.
In previous studies, the OS of patients treated with TARE was longer than in patients treated with conventional TACE (6,7). Moreover, patients treated with conventional TACE, because they were ineligible Our study had several limitations. First, due to its retrospective nature, unintended biases may have been introduced. We applied IPTW and multivariable analysis to offset differences by balancing the characteristics of both groups to minimize the selection bias, and multiple subgroup analyses to validate our points. Also, since only the outcome of TACE for TARE-ineligible patients with high LSF was evaluated, it is critical not to overinterpret this as a sign that TACE is inferior.
In conclusion, TARE-ineligible HCC patients treated with TACE had shorter TTP than TARE eligible patients. Further study might be required to determine whether other treatment strategies, such as surgical resection and systemic therapy could improve outcome in TARE-ineligible patients with a high LSF.

Data availability statement
The data that support the findings of this study are available from the corresponding authors, upon reasonable request.

Ethics statement
The studies involving human participants were reviewed and approved by the Institutional Review Board of Seoul National University Hospital. The ethics committee waived the requirement of written informed consent for participation.

Author contributions
The corresponding authors (J-HL and H-CK) had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.  The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.