Surveillance Strategy for Patients With BCLC Stage B Hepatocellular Carcinoma After Achieving Complete Remission: Data From the Real World

Purpose: There is a lack of consensus on the surveillance strategy for Barcelona Clinic liver cancer (BCLC) stage B hepatocellular carcinoma (HCC) patients with complete remission (CR). We performed a real-world, retrospective analysis of the surveillance strategy for BCLC stage B HCC patients after radical therapy with CR to support clinical decision-making. Materials and Methods: We analyzed 546 BCLC stage B HCC patients with CR after radical treatments (surgery/ablation) at Sun Yat-sen University Cancer Center, from January 2007 to December 2019. The intensity of surveillance interval was defined as the mean of surveillance interval within 2 years. The primary endpoint of the study was overall survival (OS) and extra-Milan criteria relapse. Results: During a median follow-up time of 23.9 months (range = 3.1–148.3 months), there were 11.9% of patients died, 56.6% of patients developed recurrence, the vast majority of patients experienced recurrence within 2 years, and 27.8% patients developed extra-Milan criteria recurrence. The median disease-free survival and OS were 33.6 and 60.0 months, respectively. Patients were divided into regular surveillance group (RS) (≤4.3 months) and irregular surveillance (IRS) group (>4.3 months) based on the optimal cutoff value of the intensity of surveillance interval. The RS group owned a lower incident of extra-Milan criteria relapse and smaller and fewer tumors at recurrence than IRS group, which contributed to the prolonged OS. Besides, the cutoff values of surveillance interval that could lead to significant differences in the incidence of extra-Milan criteria relapse during 0–6, 6–12, and 12–18 months after CR were 2.6, 2.9, and 3 months, respectively. Conclusions: The average surveillance interval for patients with BCLC stage B HCC achieved CR should not exceed 4.3 months during the first 2 years' follow-up. During three different phases of the initial 18 months after CR, individualized surveillance showed intervals no more than 3 months were required to reduce the incidence of extra-Milan criteria relapse.


INTRODUCTION
Hepatocellular carcinoma (HCC) is the sixth most commonly diagnosed cancer and the fourth leading cause of cancer death worldwide in 2018 (1). Rates of both incidence and mortality are two to three times higher among men in most regions (1). Hepatitis B virus (HBV) and hepatitis C virus (HCV) are considered to be the main pathogens for the development of HCC, especially in Asia (2). Recently, the number of patients with HCC originating from HCV has increased year by year, and the number of HCC patients owing to HBV has decreased (3).
The Barcelona Clinic liver cancer (BCLC) stage B (intermediate stage) (4) patients account for ∼19.4% of total HCC (5). Also, the BCLC stage B represents a heterogeneous group of patients (6), which were more complicated and experience relapses earlier than BCLC stage A. The main factors were the span of liver function score (Child-Pugh: [5][6][7][8][9], the difference of tumor size (diameter 3-10 cm or more), tumor number (2-20 or more), and the difference of tumor distribution (single lobe limited or double lobe diffused).
Untreated patients at BCLC stage B present a median survival of 16 months or a survival rate of 49% at 2 years (7,8). Chemoembolization extends the survival of these patients to a median of up to 19-20 months (8, 9). Surgery and ablation comprise potentially curative treatment modalities for BCLC stage B HCC patients (10). Besides, patients in this stage achieving downstaging from combined treatments or TACE (transarterial chemoembolization) can be suitable for radical treatments (6,11). Unfortunately, the median survival of BCLC stage B HCC patients after curative treatment was 45 months (12). Tumor recurrence after curative surgery occurs in 50-70% of patients, which constitutes either intrahepatic metastases (often within 2 years after surgery) or a new HCC in the remaining cirrhotic liver (10,13). Factors of early or late recurrence or/and metastases were complex, including tumor size and history of rupture, etc. (11,14,15).
Patients with recurrence after radical therapies may still be candidates for curative therapies (10,16,17). Early diagnosis of recurrence is more likely to receive curative treatment and achieve better disease control and prolonged survival (18). Although recent guidelines recommend surveillance strategies (10,19)  strategies. Therefore, the impact of irregular surveillance (IRS) in the real world on patient survival is also unclear.
Based on this background, we performed a real-world, retrospective analysis of the surveillance strategy for BCLC stage B HCC patients after radical therapy with CR to support clinical decision-making.

Patients
This study met the requirements of the Declaration of Helsinki and was approved by the Institutional Review Board of Sun Yat-sen University Cancer Center. We retrospectively analyzed BCLC stage B HCC patients who underwent radical therapy (surgery/ablation) from an institutional database at Sun Yat-sen University Cancer Center, from January 2007 to December 2019. A total of 2,193 consecutive patients were initially considered eligible. All cases were diagnosed as HCC according to pathology or clinical criteria (10,19). This study included BCLC stage B HCC patients who received radical treatment (surgery/ablation) and achieved CR. Multidetector computed tomography (CT) and/or magnetic resonance imaging (MRI) were performed routinely to evaluate the local or distant extension of the primary tumors. Patients who visited our hospital at least 3 months after radical treatment were candidates for this study. CR is defined as no recurrence within 3 months after radical treatments. Patients were also excluded if they met any of the following criteria: age <18 or >75 years, non-HCC, mixed liver cancer, non-BCLC stage B, non-radical treatment, non-CR, died of postoperative complications. After excluding 1,637 patients according to the exclusion criteria, 546 patients were finally included in the study. All patients received radical treatment, including surgery and ablation. Some patients were treated with TACE before having undergone radical treatment, whereas others received a one-stage radical treatment.

Surveillance Strategy
After radical operation, patients were informed to perform multiphasic, high-quality, cross-sectional imaging of the chest, abdomen, and pelvis every 3-6 months for 2 years and then followed up every 6-12 months as recommended by the guideline (19). Recurrence was defined as radiological evidence of intraabdominal or abdominal soft tissue around the surgical site, or else distant metastasis. Besides, the date of each surveillance was recorded, and the end point of the surveillance was the time of tumor extra-Milan criteria recurrence and death. Intensity of surveillance interval was defined as the mean of surveillance interval within 2 years. For patients who died, survival time after curative treatment and the result of death were recorded.

Statistical Analysis
Overall survival (OS) and extra-Milan criteria recurrence were measured from the date of CR to death or extra-Milan criteria recurrence or last follow-up evaluation. Continuous variables were presented as mean ± standard deviation and analyzed using the Student t-test. Categorical variables were analyzed using the χ 2 or Fisher exact test, as appropriate. Survival rates were estimated by the Kaplan-Meier (K-M) method. Differences in OS were assessed for significance using the log-rank test. The Cox proportional hazards regression model was used to determine the factors associated with survival. As per initial design, all variables with a P < 0.05 by univariable analysis were entered in the multivariable analysis. Finally, only one variable was found to be associated with survival, and multivariable analysis could not be performed. Optimal cutoff for analysis was selected using X-Tile. Statistical analyses were performed using SPSS 18 for Windows (SPSS Inc., Chicago, IL, USA). Pictures were drawn using GraphPad Prism version 6.0 and R-3.6.3 software. All P-values were two-sided, and P < 0.05 was considered significant.

Clinical and Pathological Characteristics of All HCC Patients
We enrolled 546 BCLC stage B HCC patients who achieved CR after radical treatment. The clinical and pathological characteristics of all the 546 patients are listed in Table 1. The 546 patients were followed 2,115 times, with an average of 4 times per person over 2 years. According to the best cutoff, regular surveillance group (RS) (n = 441) was defined as receipt of repeated CT/MRI with mean interval ≤4.3 months within 2 years. The IRS group (n = 105) was defined as receipt of repeated CT/MRI with mean interval >4.3 months within 2 years. Overall demographics were similar, but RS patients with a higher proportion of poor differentiation (P < 0.001).

Follow-Up and Assessment of Prognosis of All HCC Patients
Median follow-up time was 23.9 months (range = 3.1-148.3 months), and the median disease-free survival (DFS) and OS were 33.6 and 60.0 months, respectively; 11.9% of patients (65/546) died, with a 2-years OS rate of 88.0%, and the 5-years OS rate was 87.0%; 56.6% of patients (309/546) developed a recurrence, with 2-years DFS rate of 53.0% and 5-years DFS rate of 46.0% (Figures 1A,B). The 1-, 3-, and 5-years survival rates     (Figures 1A,C). Moreover, the hazard of relapse reached its peak in the first 2 years (Figure 1D). Thus, it makes sense to focus on surveillance during the first 2-years after curative treatments to detect early recurrence at a potentially more treatable stage.

Univariate and Multivariate Analyses of Prognostic Factors for Survival of HCC Patients With Relapse
To further assess the association between surveillance interval and survival, further analysis was performed on relapsed patients.
The correlation analysis demonstrated that patients in the IRS group owned a higher incidence of extra-Milan criteria recurrence (P = 0.004), a larger size of the recurrent tumor (P = 0.011), and a higher proportion of multinodular tumors (P = 0.003) ( Table 4) and less likely to receive secondary treatments after recurrence (P = 0.001). Moreover, the violin plot also   The meaning of the bold values provided was p < 0.05. indicated that the IRS group owned a larger size of the recurrent tumor (Figure 3).

Comparison of Surveillance Interval and 0-18 Months Extra-Milan Criteria Relapse of All HCC Patients
According to the hazard rate curve that the recurrence risk of BCLC stage B patients with CR was still high at 0-24 months ( Figure 1D). Moreover, although the K-M curve showed no difference between the RS and IRS groups for extra-Milan criteria  relapse in HCC patients ( Figure 2B) and HCC patients relapse (Figure 5B), the interval between 0 and 18 months of surveillance also appeared to be associated with extra-Milan criteria relapse ( Figure 5B). In the further analysis of patients with extra-Milan criteria relapse in 0-18 months, we found that the RS group could earlier detect extra-Milan criteria relapse (P = 0.046, HR = 0.602, 95% CI = 0.366-0.991) (Figure 6B), which significantly prolonged OS (P < 0.001, HR = 2.893, 95% CI = 1.647-5.082) (Figure 6A). Based on this, we further analyzed the surveillance interval of 0-6, 6-12, 12-18, and 18-24 months.

DISCUSSION
Currently, the European Society for Medical Oncology proposes that follow-up of patients who underwent radical treatments (resection or ablation) should consist of the clinical evaluation of liver decompensation and the early detection of recurrence by dynamic CT or MRI studies every 3 months during the 1st year and surveillance every 6 months thereafter (10). But the National Comprehensive Cancer Network offers a different view, recommending continuous surveillance every 3-6 months, for 2 years, and then every 6-12 months (19). However, these two guidelines are not sufficient to guide clinical practice, in which the follow-up strategies of the two clinical guidelines are quite different in terms of the surveillance interval, not for specific patients. Also, the guidelines do not specifically recommend surveillance intervals for BCLC stage B HCC patients with CR, which were more complicated and had relapses earlier than did those of BCLC stage A. Although patients are recommended for surveillance according to clinical guidelines, in the real world, for a variety of reasons, patients cannot fully follow the guidelines for surveillance strategies. Therefore, the impact of IRS in the real world on patient survival is unclear. Previous studies have indicated that earlier identification of disease may facilitate patient eligibility for investigational studies or other forms of treatment (19,20). HY K et al. demonstrated that the detection of small HCC eligible for curative treatment is increased by frequent surveillance (16,18). Besides, patients in the RS group were diagnosed at earlier stages than the IRS or non-surveillance groups, which had more chance for curative treatments (18). Moreover, AA M et al. also reported that a long surveillance interval compromises OS in high-risk patients who underwent curative thermal ablation for HCC within the Milan criteria (21). Besides, other tumors have the same results for which more intensive surveillance after surgery for esophagogastric adenocarcinoma, colorectal cancer, and non-small cell lung cancer translates into improved survival (22)(23)(24)(25). Although there is no high-level evidence, the cutoff of 2 years has been adopted to grossly classify early and late recurrences (14,26). In our result, we also found that the vast majority of patients experienced recurrence within 2 years. Other than that, we also proved that RS owned a lower incidence of extra-Milan criteria relapse and smaller and fewer tumors at recurrence than those of IRS group, which contributed to the prolonged OS. Thereby, the average surveillance interval for patients with BCLC stage B HCC who achieved CR should not exceed 4.3 months during the first 2 years' follow-up.
Over the past 20 years, the Milan criteria have been highly successful in selecting patients for good long-term survival and remain the criteria for potential transplant candidates for HCC (27). It is important to identify the possible predictive factors of within and extra-Milan criteria recurrences after radical treatments (28). Early diagnosis of extra-Milan criteria recurrence can enable patients to receive a more timely intervention after recurrence and control the development of tumors. In our study, we also found that the RS group could earlier detect extra-Milan criteria relapse and significantly prolonged OS in 0-18 months relapsed patients. Moreover, during 0-6, 6-12, and 12-18 months of the initial 18 months after CR, individualized surveillance intervals that no more than 3 months were required to reduce the incidence of extra-Milan criteria relapse. The interval of surveillance according to current guidelines is therefore insufficient, especially 12-18 months after CR.
As mentioned above, despite this study having many clinical implications, we should be clear that it is a retrospective study with its limitations. First, our study was conducted in a single center. The collection of multicenter data to expand the sample size is the next step that needs to be done. Moreover, the follow-up strategy of patients in different stages after radical operation needs to be further explored. Finally, RS could detect tumor recurrence at an early stage and prolong the survival of patients, which requires further clinical trials to verify it.
In conclusion, our results demonstrated that the surveillance interval for BCLC stage B HCC patients with CR after curative treatment should not exceed 4.3 months during the first 2 years' follow-up. Besides, during 0-6, 6-12, and 12-18 months of the initial 18 months after CR, individualized surveillance intervals of no more than 3 months were required to reduce the incidence of extra-Milan criteria relapse.