Combined Transarterial Chemoembolization and External Beam Radiotherapy for Identifying Surgical Candidates for Hepatocellular Carcinoma with Macroscopic Vascular Invasion: A Propensity Score–Weighted Analysis
Article information
Abstract
Purpose
This study aimed to evaluate the role of hepatic resection in patients with objective responses after combined transarterial chemoembolization (TACE) and radiotherapy (RT) for hepatocellular carcinoma (HCC) with macroscopic vascular invasion (MVI).
Materials and Methods
We retrospectively reviewed the patients treated with combined TACE and RT for HCC with MVI between 2010 and 2015. Some of the patients with objective responses underwent hepatic resection or liver transplantation; to investigate the impact of surgery, patients with objective responses who did not undergo surgery were selected as the control group. Survival outcomes were compared using a propensity score–based stabilized inverse probability of treatment weighting method.
Results
Out of the 170 patients with objective responses after combined TACE and RT, 41 patients underwent surgery, including eight liver transplantations. The unweighted surgery group was younger and had a higher proportion of solitary tumors and unilateral vascular involvement. After adjustment, the 3-year overall survival (OS) rates were 61.0% and 28.6% in the surgery and non-surgery groups, respectively. The most important prognostic factor for OS was surgery (adjusted Cox hazard ratio [HR], 0.28; 95% confidence interval [CI], 0.17 to 0.46; p < 0.001). Complete response after TACE and RT (vs. partial response) was also a significant prognostic factor for OS (adjusted HR, 0.41; 95% CI, 0.27 to 0.61; p < 0.001). There was no surgical mortality. Four patients (9.8%) required additional surgery due to bleeding or graft failure.
Conclusion
Hepatic resection was significantly associated with improved OS in patients who showed objective responses after receiving combined TACE and RT for HCC with MVI.
Introduction
A substantial number of newly diagnosed hepatocellular carcinoma (HCC) patients are found in an advanced stage, accompanied by macroscopic vascular invasion (MVI). Systemic therapies are considered a standard treatment for advanced-stage HCC in the Barcelona Clinic Liver Cancer staging system [1]. However, there are still limitations regarding insufficient response rates and treatment outcomes after systemic therapies, especially in patients with HCC accompanied by MVI.
Vascular invasion is not only considered a significant factor for recurrence but is also associated with a decline in hepatic function due to the interruption of portal flow. As a result, surgical treatment is generally regarded as an unsuitable option. Nevertheless, hepatic resection was attempted to improve the outcome of HCC with MVI with some promising results. In large-scale retrospective studies, the median overall survival (OS) ranged from 6.0 to 29.4 months, with a 1-year OS rate of 18.0% to 58.7% as reported [2,3]. Accordingly, the Eastern guidelines for the management of HCC recommend surgery for vascular invasion in select cases [4-6]; however, a consistent indication for identifying surgical candidates has not yet been established.
External beam radiotherapy (RT) with or without locoregional modalities, has also been widely used for HCC with MVI [7,8]. A recent randomized clinical trial demonstrated superior progression-free survival (PFS) and OS rates with the combination of transarterial chemoembolization (TACE) and RT compared to sorafenib [9]. Favorable responses after TACE and RT were observed in more than half of these patients [10], and in some of these good responders, downstaging might have enabled curative resection, leading to better long-term survival. In the present study, we evaluated the efficacy and safety of hepatic resection in patients who showed a good response after combined TACE and RT for HCC with MVI.
Materials and Methods
1. Patients
We retrospectively reviewed patients who were treated with combined TACE and respiratory-gated three-dimensional conformal RT as a first-line treatment for HCC with MVI between January 2010 and December 2015 at Asan Medical Center (Seoul, Korea). Only patients with no extrahepatic metastasis, preserved liver function of Child-Pugh classification A, and an Eastern Cooperative Oncology Group (ECOG) performance status of 0-2 were included in the present study population as potential surgical candidates. After the combined TACE and RT followed by several sessions of TACE, some of these patients showed a good response and eventually underwent hepatic resection or liver transplantation (LT) (i.e., surgery group). On the other hand, patients who achieved at least a partial response (PR) as the best response to TACE and RT according to the modified Response Evaluation Criteria in Solid Tumors (mRECIST) [11] but did not undergo surgery were analyzed as the non-surgery group. Other exclusion criteria were as follows: (1) double primary cancer at diagnosis; (2) isolated MVI without a definite primary HCC; (3) incomplete RT administered at less than 80% of planned dose; (4) stereotactic body RT regimen; (5) missing values of pretreatment α-fetoprotein, and (6) lost to follow-up before 3 months from the start date of RT.
The presence of MVI was determined by liver dynamic computed tomography (CT) or magnetic resonance imaging (MRI) using hepatocyte-specific contrast agent by the following criteria outlined in a previous study [9]: (1) intraluminal filling defect in a portal vein (PV), hepatic vein, or inferior vena cava continuous with the intrahepatic tumor, and (2) enhancement of the filling defect on the arterial phase, and washout on the portal or delayed phases.
2. Treatment
Initially, cisplatin-based TACE was performed. After infusing 2 mg/kg of cisplatin (Dong-A Pharmaceutical) by selective catheterization of the feeding artery, embolization was performed using an emulsion of 3-20 mL of iodized oil (Lipiodol, Guerbet) and gelfoam slurry (Upjohn). Transarterial chemotherapeutic infusion without gelfoam slurry was done according to the PV blood flow.
Respiratory-gated 3-dimensional conformal RT was initiated within 3 weeks after the first TACE. The gross tumor volume included intravascular and adjacent tumors within a 2 cm margin. A total dose of 40-45 Gy with 2.5-3 Gy per fraction was prescribed for the planning target volume, which extended 7 mm from 30%-70% of the respiratory phase internal motion of the gross tumor volume. The detailed procedures of TACE and RT were described in our previous reports [9,10,12,13]. Subsequent TACE for residual viable HCC was repeated every 6-8 weeks.
Based on the evaluation of tumor extent before each session of TACE, if it was determined to be resectable by the hepatobiliary or transplant surgeons in the multidisciplinary liver cancer team, a surgical treatment option was offered to the patient. Surgery with curative intent was considered if all viable tumors could be managed by anatomical liver resection, either alone or in conjunction with other local therapies such as radiofrequency ablation, provided that a sufficient remnant liver volume was expected. LT was also performed in selected patients when a suitable donor was available. The extent of hepatic resection was determined by the proportion of future liver remnant volume and hepatic functional reserve. Depending on the extent of PV invasion, either en bloc resection of involved PV branches without tumor thrombus exposure or thrombectomy, which includes extracting the tumor thrombus and forceful flushing to confirm complete removal, was performed.
3. Evaluation
The radiologic response of the target lesions was evaluated according to the mRECIST using dynamic liver CT or MRI after the completion of RT and before each subsequent TACE session, and the best overall response observed during follow-up was recorded. Newly developed intrahepatic and extrahepatic recurrences were defined as progression. The primary outcome measure was the comparison of OS rates between the surgery and non-surgery groups, and the secondary outcome variables included PFS rates and treatment-related toxicities. Adverse events related to RT were evaluated using the Common Terminology Criteria for Adverse Events (CTCAE) ver. 4.03. Non-classic radiation-induced liver disease was graded according to the CTCAE or by an increase of the Child-Pugh score by ≥ 2 without evidence of disease progression within 3 months after RT. The major complication after surgery was defined as events of grade 3 or higher according to the Clavien-Dindo classification [14].
4. Statistical analysis
A propensity score-based stabilized inverse probability of treatment weighting (IPTW) method was used to minimize selection bias arising from confounding factors between the surgery and non-surgery groups. Binary logistic regression including all covariates was used to generate propensity scores that were assigned to the treatment groups. Weights for patients who underwent surgery (1/propensity score) and those who did not (1/[1−propensity score]) were calculated, and stabilized weights were then applied to reduce the variability of weights in the IPTW model [15]. Balance among covariates was evaluated using absolute standardized mean differences (SMD); an SMD of ≤ 0.2 was considered acceptable. We compared survival outcomes using a weighted log-rank test [16] and estimated the hazard ratio (HR) with a weighted Cox proportional hazard (PH) model. To mitigate the guarantee time bias caused by selecting only patients with a good response after TACE and RT, survival times were calculated 3 months after the start date of RT, corresponding to the interval between the first response assessment after RT. Also, the surgery was regarded as a time-dependent variable in the survival analyses and weighted Cox PH models because it was not performed at the same time point, and there were disease progression events before the surgery. Variables associated with survival (p ≤ 0.1) in univariable analyses were included in multivariable analyses using backward elimination for further covariate adjustment. Two-sided p-values less than 0.05 were considered statistically significant. All statistical analyses were performed using R ver. 4.0.3 (R Foundation for Statistical Computing, http://cran.r-project.org).
Results
1. Patient characteristics
Among patients treated with combined TACE and RT as a first-line treatment for MVI, 292 patients were identified with the ECOG performance status 0-2, the Child-Pugh class A liver function, and no extrahepatic metastasis. After excluding patients without radiologic response after TACE and RT, 170 were included in the final analysis. A patient selection flowchart, including minor exclusions, is depicted in Fig. 1. The comparison of baseline characteristics between non-responders, who were not included in the final cohort, and responders is presented in S1 Table.
Patient selection flow chart. AFP, α-fetoprotein; ECOG PS, Eastern Cooperative Oncology Group performance status; HCC, hepatocellular carcinoma; MVI, macrovascular invasion; PD, progressive disease; RT, radiotherapy; SBRT, stereotactic body radiotherapy; SD, stable disease; TACE, transarterial chemoembolization.
The most common etiology of our study population was hepatitis B virus. A total of 41 patients underwent surgery, including 8 LTs. Surgery was performed at a median of 5.2 months (interquartile range [IQR], 2.7 to 8.9) from the start date of RT, and additional treatments were also administered between RT and surgery (median, 2; IQR, 1 to 4). Before IPTW, the surgery group was significantly younger (median, 53 vs. 55 years; p=0.047), had a higher rate of single tumor (51.2% vs. 31.8%, p=0.039), and a higher rate of unilateral PV involvement or less (70.7% vs. 51.2%, p=0.044) than the non-surgery group. After applying the stabilized weights, all covariates were balanced, with SMDs less than 0.2 (Table 1). The distribution of propensity scores and stabilized weights of both groups is shown in S2 Fig.
Baseline patient characteristics
2. Survival outcomes and prognostic factors
Fig. 2 shows the survival outcomes of the two groups. The median follow-up duration was 27 months (range, 5 to 143 months). The adjusted 1-, 3-, and 5-year OS rates were higher in the surgery group than in the non-surgery group (93.8% vs. 79.1%, 61.0% vs. 28.6%, 56.3% vs. 17.7%, respectively; p < 0.001 by weighted log-rank test). Adjusted median OS times of the surgery and non-surgery groups were 81.2 months (95% confidence interval [CI], 33.8 to not reached) and 21.5 months (95% CI, 18.9 to 27.5), respectively. The adjusted PFS rates of the surgery group were also significantly higher than those of the non-surgery group (1- and 3-year PFS rates of 49.2% vs. 35.3%, 29.7% vs. 11.8%, respectively; p=0.036 by weighted log-rank test).
Overall survival (A) and progression-free survival (B) of the surgery and non-surgery groups. Crude (solid line) and weighted survival curves (dashed line) of each group are presented. The number at risk was different from the number of original sizes of each group because the survival analysis was conducted in a time-dependent manner.
In multivariable analysis using the weighted Cox PH method, the model that included maximum tumor size, radiologic response, and surgery was the most efficient in predicting OS, for which both radiologic response and surgery were identified as independent prognostic factors (the adjusted HR for surgery, 0.28; 95% CI, 0.17 to 0.46) (Table 2). The results of multivariable analysis before IPTW for OS are presented in S3 Table.
Univariable and multivariable analyses of prognostic factors for overall survival after IPTW
In a subgroup analysis according to radiologic response in which new weights were applied for each subgroup), the significant benefit in OS of surgery was maintained in the PR group (p < 0.001 by weighted log-rank test) (Fig. 3). In the complete response (CR) group, surgery initially appeared to result in higher survival rates; however, after a long-term follow-up, the intersection of the two survival curves was identified with a p-value of 0.190 using the weighted log-rank test (Fig. 3). Multivariable analysis for OS in the surgery group found no significant factors except for tumor size, which showed marginal significance (S4 Table).
Subgroup analysis of overall survival according to radiologic response by the modified Response Evaluation Criteria in Solid Tumors; complete response (A) and partial response groups (B). Crude (solid line) and weighted survival curves (dashed line) of each group are presented. The number at risk was different from the number of original sizes of each group because the survival analysis was conducted in a time-dependent manner.
3. Toxicity
There were five cases (2.9%) of grade 3 elevations of aminotransferases and 23 cases (13.5%) of a worsening Child-Pugh score by ≥ 2 within 3 months of completing RT; none of these cases resulted in hepatic failure. Grade 3 gastrointestinal bleeding was observed in three patients (1.8%). There were seven patients (17.1%) who experienced major complications within 3 months after surgery (Table 3). Two patients required laparotomy to control postoperative bleeding, and two others, including one who needed re-LT due to graft failure, required admission to the intensive care unit (Table 3). However, surgery-related mortality was not reported.
Major surgical complications according to the Clavien-Dindo classification
4. Sensitivity analysis
To evaluate the robustness of the observed difference in OS between the surgery and non-surgery groups, several alternative analyses were conducted: (1) to isolate the effect of hepatic resection, patients who underwent LT were excluded from the surgery group; (2) patients with missing pretreatment α-fetoprotein values, who were excluded from the current IPTW analysis, were reintegrated after filling in the missing values using the multiple imputation method; (3) alternative statistical approaches including, the use of only significantly different covariates for generation of propensity scores, and 1:1 propensity score matching instead of IPTW, were also conducted. In every sensitivity analysis, the significance of the superior OS of the surgery group remained (S5 Fig.).
Discussion
As the presence of MVI is associated with poorer prognosis among patients diagnosed with advanced-stage HCC, further efforts are needed to identify more effective treatment modalities that can enhance the survival outcomes of such cases. In the previous randomized trial evaluating TACE and RT versus sorafenib for MVI, patients treated with sorafenib did not undergo surgery due to an unsatisfactory response, while curative resection was performed in five patients (11.1%) in the TACE and RT group [9]. Despite atezolizumab/bevacizumab or lenvatinib being recommended as the standard of care for advanced HCC, they still exhibit suboptimal response rates in cases with vascular invasion [17], and cases of curative surgery following systemic therapy are rarely observed. Therefore, combined TACE and RT, which offers potential for curative surgery due to effective local control and downstaging, may still play a distinct role for these patients. In the present study, approximately 24% of good responders and 14% of all patients with MVI could benefit from surgery, similar to our previous randomized trial. The favorable response rate of RT to MVI also prevents intrahepatic tumor spread and deterioration of hepatic function by maintaining portal flow. Therefore, it might also aid in preventing disease progression and preserving hepatic function after surgery. Promising survival outcomes (median OS, 18.9 months; 1-year OS rate, 69%-75%) were also reported from studies of surgery with perioperative RT [18-20]. Moreover, a significantly increased future remnant liver volume was observed after RT with intra-arterial chemotherapy for locally advanced HCC, which might enable more patients to receive hepatic resection after downstaging [21].
The patients who underwent surgery showed significantly higher OS rates than those in the non-surgery group, even after showing a favorable response to combined TACE and RT for HCC with MVI. Though some previous studies evaluating the efficacy of surgery after downstaging with locoregional modalities have shown excellent survival outcomes [21,22], identifying suitable candidates for curative surgery after downstaging remains challenging due to the variable prognosis of HCC with MVI [12]. Considering that radiologic response based on the mRECIST after locoregional or systemic therapies was reported as a significant prognostic factor [10,23], which is consistent with the present study, including patients with a poor response to previous locoregional treatments in the control group could potentially strengthen selection bias and complicate the interpretation of the role of surgery. The higher survival rate of the surgery group in the present study is particularly noteworthy because it resulted from a comparison to a control group consisting exclusively of patients with favorable responses, which not only confirms that surgery provides an additional survival benefit in this clinical setting but also suggests that this response-based approach may help identify patients who can expect better long-term survival outcomes with surgery.
Part of the additional survival benefit from surgery might be due to a reduction in the tumor burden of primary HCC, enabling successful salvage treatment in case of disease progression. In contrast to the substantial difference observed in OS favoring surgery, the PFS difference between the surgery and non-surgery groups was relatively smaller than the OS difference. Moreover, the PFS of hepatic resection alone, excluding LT, was not significantly different from that of the non-surgery group (S6 Fig.). Vigorous locoregional salvage treatments for intrahepatic recurrent HCCs may improve survival outcomes [24,25]. This assumption is also consistent with the finding that the OS benefit of surgery was not evident in the CR subgroup. Patients who achieve a CR have no residual tumor burden, similar to those who undergo surgery, which may potentially allow for more effective salvage therapy. Alternatively, in patients with residual tumors, surgical reduction of tumor burden might have played a complementary role in offsetting prognostic differences associated with radiologic response. This concept aligns with recent evidence suggesting that, in certain patients who demonstrate an excellent response or achieve pathological complete necrosis following downstaging therapy, surgical resection may not confer additional survival benefits [26]. These insights support the further refinement of surgical decision-making after achieving locoregional tumor control. In this context, administering TACE and RT as an initial treatment for MVI, followed by radiologic response assessment, may aid in estimating the tumor’s inherent biology and identifying patients who are more likely to benefit from surgery. Accordingly, a salvage-oriented approach, rather than routine consolidative surgery, may represent a promising strategy for selected patients following locoregional therapy.
The extent of surgery and surgical approaches for intravascular tumors should also be considered. Chong et al. [27] reported that major resection, defined as the resection of three or more adjacent anatomical segments, showed better disease-free and disease-specific survival. [27]. However, major surgery is typically deemed feasible only if at least 40% of the total liver volume can be preserved post-surgery, making it applicable to a select group of patients. LT may also be considered as an alternative surgical option for patients with HCC who have MVI after successful downstaging with prior locoregional treatments [28,29]. In the present study, eight patients who received LT showed a long-term OS rate of 75% at 3 years. For intravascular tumors, either peeling-off resection or en bloc resection with reconstruction would be considered. En bloc resection may increase the chance of complete resection but is associated with high morbidity and mortality rates. Peeling-off resection is usually less demanding but has the potential to leave remnant viable cancer cells after surgery [30].
The study has several limitations. Because of its retrospective nature, there is a risk of bias from unmeasured confounders even after IPTW adjustment, including the decision on resectability, which may vary among surgeons. Failure to assess the future remnant liver volume or surgical approach could obscure the surgical outcomes. Nevertheless, it is expected that a substantial portion of bias was alleviated due to the selection of only the good response group, which is a strong prognostic factor, as well as the control group and the vigorous effort to reduce bias through analytical methods including IPTW, time-dependent analysis, and further covariate adjustment. Another limitation of this study stems from the heterogeneity of salvage treatment strategies used after combined TACE and RT or in cases of postoperative disease progression. The individualized approach complicates the incorporation of these variables into survival analysis. Lastly, we could not find any clinical factors affecting the prognosis among the surgery group. Further exploration of factors affecting the outcome after surgery will help identify explicit indications for the surgery.
In conclusion, hepatic resection was found to have a significant association with improved OS rates and acceptable toxicities in patients showing a good response after combined TACE and RT for HCC with MVI. Further studies are needed to investigate the role of surgical approaches following successful locoregional treatments for patients with advanced-stage HCC who have MVI.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
This study was approved by the Institutional Review Board of Asan Medical Center (#2017-0491), and the requirement for informed consent was waived considering the retrospective nature of the study.
Author Contributions
Conceived and designed the analysis: Lee S, Yoon SM.
Collected the data: Lee S, Yoon SM.
Contributed data or analysis tools: Lee S, Jung J, Choi J, Kim SY, Kim JH, Lee D, Shim JH, Kim KM, Lim YS, Lee HC, Song GW, Park JH, Yoon SM.
Performed the analysis: Lee S, Jung J, Choi J, Yoon SM.
Wrote the paper: Lee S, Jung J, Choi J, Kim SY, Kim JH, Lee D, Shim JH, Kim KM, Lim YS, Lee HC, Song GW, Park JH, Yoon SM.
Conflicts of Interest
Conflict of interest relevant to this article was not reported.
Acknowledgments
We thank Dr. Joon Seo Lim from the Scientific Publications Team at Asan Medical Center for his editorial assistance in preparing this manuscript.