Management and Prognosis of HCC Patients With Elevated Tumor Markers but no Imaging Recurrence After Radiofrequency Ablation.

1
Introduction
Liver cancer is the seventh most common malignancy worldwide and the third leading cause of cancer‐related death [1]. HCC is the most prevalent type of primary liver cancer, accounting for 80% of all liver malignancies [2]. The guidelines for HCC management recommend treatments according to tumor stages and liver function reserve and surgical resection is recommended for cases of solitary HCCs with preserved liver function [3, 4, 5]. Radiofrequency ablation (RFA) has emerged as a nonsurgical option for the management of small HCCs with a high rate of success because it can preserve liver function, affording survival outcomes comparable to those achieved with surgical resection [6, 7, 8].
Even after curative treatment with resection or ablation, HCC recurs frequently owing to de novo carcinogenesis of the liver or intrahepatic metastases [9, 10, 11]. Therefore, post‐treatment surveillance to detect recurrence is essential for improving patient outcomes [11]. For that purpose, tumor markers, such as alpha‐fetoprotein (AFP), lens culinaris agglutinin‐reactive fraction of alpha‐fetoprotein (AFP‐L3), and des‐gamma‐carboxy prothrombin (DCP), have been widely used in conjunction with imaging modalities in Japan, because tumor marker elevation has been recognized as a strong indicator of HCC recurrence during post‐treatment surveillance [12, 13, 14, 15]. However, in clinical practice, some patients exhibit persistent tumor marker elevation without evidence of recurrence in imaging studies. According to the guidelines of the American Association for the Study of Liver Diseases, patients with significantly elevated AFP levels (≥ 200 ng/mL) but no visible liver mass on imaging are recommended to undergo repeat abdominal imaging using an alternative modality (such as multiphasic MRI if CT was initially performed) and additional imaging of the chest and pelvis, with positron emission tomography‐CT as a potential further diagnostic option when additional tests fail to identify the cause of AFP elevation [3]. However, the optimal frequency of liver imaging studies, how to explore extrahepatic lesions, and the characteristics of HCCs detected during intensive examinations for that situation remain to be fully elucidated. Therefore, in this study, we aimed to assess the outcomes of patients with HCC who had persistently elevated tumor markers in the absence of detectable tumors following curative RFA.

2
Patients and Methods
2.1
Study Protocol
This retrospective study was conducted in accordance with both the Declarations of Helsinki and Istanbul, and in accordance with the ethical guidelines for epidemiological research established by the Japanese Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labour and Welfare. The study design was included in a comprehensive protocol of retrospective studies at the Department of Gastroenterology, the University of Tokyo Hospital, and was approved by the ethics committee of the University of Tokyo Medical Research Center (approval no. 2058). The requirement of informed consent was waived because of the retrospective nature of the study. The following statements were posted on a website (https://gastro.m.u‐tokyo.ac.jp/disclosure/), and participants who did not agree to the use of their clinical data could request the deletion of their data.

2.2
Patient Enrollment and Data Collection
Data were collected for patients with primary HCC who underwent RFA in our department from January 1, 1999, to December 31, 2015, from a designated computerized database [16]. HCC was diagnosed using dynamic CT or MRI findings according to Japanese Clinical Practice Guidelines for Hepatocellular Carcinoma [17]. Early enhancement in the arterial phase and washout in the delayed phase were considered indicative of HCC. If the diagnostic imaging tests could not help reach a definitive diagnosis of HCC, an ultrasound‐guided tumor biopsy was performed. Pathological diagnosis was made based on General Rules for the Clinical and Pathological Study of Primary Liver Cancer [18].
Patients were treated with curative intent by RFA, and treatment efficacy was evaluated using dynamic CT or MRI. The detailed RFA procedure has been described elsewhere [19]. After curative treatment with RFA for primary HCC, patients underwent post‐treatment surveillance. The follow‐up protocol consisted of the determination of tumor markers, serum AFP, AFP‐L3, and DCP, and contrast‐enhanced CT or MRI every 4 months [20]. When HCC remained undetected despite elevated tumor markers, the follow‐up interval was reduced to 2–3 month or chest CT scans were added at the discretion of the outpatient physician. Patients were followed up until death or lost to follow‐up before December 31, 2021. Tumor recurrence was defined using the same criteria as those for the initial diagnosis of HCC. Patient status was categorized as shown in Figure 1. Patients' status was considered tumor‐free until a viable tumor was found by imaging modalities. When tumor recurrence was identified, patients were regarded as tumor‐free if there was no residual HCC on imaging after recurrence treatment. Otherwise, patients were categorized as having a residual tumor, which was attributable to unsuccessful local ablation owing to technical difficulties or tumor multiplicity, vascular invasion, and extrahepatic metastasis.
We identified cases in which tumor marker levels exceeded the cut‐off value during follow‐up in patients who were considered tumor‐free after treatment of initial or recurrent HCC. Cases in which tumor markers did not return to normal levels after curative treatment were not excluded; when tumor markers remained above the cutoff at the first visit after curative treatment, elevation detected at subsequent visits was regarded as tumor marker elevation, given that persistent tumor marker elevation after treatment has been reported to be associated with a high risk of recurrence [11]. The cut‐off values set were as follows: AFP of ≥ 20 ng/mL for patients with hepatitis C who achieved a sustained virological response and ≥ 200 ng/mL for others, AFP‐L3 of ≥ 15%, and DCP of ≥ 200 mAU/mL, excluding for those treated with warfarin [21]. These cases were then classified into three groups [1] recurrence diagnosed within 180 days after tumor marker elevation (early recurrence group) [2], recurrence diagnosed more than 180 days after tumor marker elevation (late recurrence group) (Figure SS1), and [3] no recurrence identified during follow‐up (no recurrence group). The unit of analysis in this study is a case of tumor marker elevation, not an individual patient; the same patient may contribute multiple cases if tumor‐free status was achieved after curative treatment.
For these cases, we collected data on tumor marker levels at initial elevation and at recurrence, tumor size and number, presence of vascular invasion or extrahepatic metastasis, and treatment modalities for recurrence.
The primary outcome of this study was tumor characteristics at recurrence (size, number, presence of vascular invasion or extrahepatic metastasis) and curability of treatment. The secondary outcome was patient survival.

2.3
Statistical Analysis
Data are expressed as medians with 25th to 75th percentiles unless otherwise indicated. Numbers and percentages were used for qualitative variables. We estimated the recurrence rate after tumor marker elevation using the Kaplan–Meier method. Patient characteristics at tumor marker elevation and recurrent HCC characteristics were compared between the early recurrence group and the late recurrence group. Given that AFP cutoff values were defined according to etiology and SVR status, AFP positivity rates were also compared between the two groups after stratification by etiology and SVR status. As a subgroup analysis, we compared the characteristics between the non‐recurrence group and the late recurrence group. Continuous variables were compared using the Wilcoxon test, and categorical variables were compared using the chi‐square test or Fisher's exact test, as appropriate. A two‐sided P value of < 0.05 was considered statistically significant. We also estimated the survival after HCC recurrence with the Kaplan–Meier method. Statistical analyses were performed using the R software (Ver. 4.1.3; R Development Core Team, Vienna, Austria).

3
Results
3.1
Patient Profiles
Our database search identified 1620 patients with treatment‐naive hepatocellular carcinoma (HCC) who achieved complete tumor ablation during the study period. During follow‐up, among 1620 patients, 907 cases of elevated tumor markers were observed in 509 patients. Among these, recurrence was not observed in 21 cases, but recurrence was observed in 886 cases. Recurrence within 180 days after tumor marker elevation occurred in 707 cases, and recurrence after 180 days following tumor marker elevation occurred in 179 cases (Figure 2). The median (interquartile range) interval between tumor marker elevation and HCC recurrence was 89.0 (81.0–97.0) days (Figure SS2).
The characteristics of the patients are shown in Table 1. The mean age was 71.7 ± 8.3 years in the early recurrence group, and 70.8 ± 9.2 years in the late recurrence group, with 404 (57.1%) and 100 (55.9%) being male, respectively. Regarding etiology, hepatitis C was more prevalent in the early recurrence group (77.4% vs. 68.2%), while non‐B non‐C hepatitis was more common in the late recurrence group (12.0% vs. 20.1%). At the time of initial tumor marker elevation, AFP ≥ 20 ng/mL was more frequently observed in the early recurrence group (71.0% vs. 58.1%, p < 0.01), as was AFP ≥ 200 ng/mL (29.4% vs. 20.1%, p = 0.02). AFP‐L3 ≥ 15% (60.0% vs. 62.6%, p = 0.58) and DCP ≥ 200 mAU/mL (19.7% vs. 16.2%, p = 0.34) showed no significant differences between the two groups.
When stratified by etiology and SVR status, among HCV non‐SVR patients, the AFP ≥ 200 ng/mL positivity rate was similar between the early and the late recurrence groups (37.2% vs. 37.3%, p = 1.0), as was AFP ≥ 20 ng/mL positivity (80.1% vs. 76.0%, p = 0.44). In HCV‐SVR patients, AFP ≥ 200 ng/mL positivity was similarly low in both groups (9.7% vs. 12.8%, p = 0.59), whereas AFP ≥ 20 ng/mL positivity was similarly high in both groups (62.5% vs. 61.7%, p = 1.00). Among HBV and NBNC patients, AFP ≥ 200 ng/mL positivity was significantly lower in the late recurrence group than in the early recurrence group (40.0% vs. 9.5% and 16.5% vs. 0%, respectively; both p < 0.01).
In the no‐recurrence group, 4 of 21 patients were lost to follow‐up, 14 of 21 patients died. Median (interquartile range) follow‐up time was 461 (326–696) days. As a subgroup analysis, we compared the characteristics between non‐recurrence group and late recurrence group. Except for Albumin levels, there were no significant differences between the groups (Table SS1).

3.2
HCC Recurrence Characteristics
Recurrence patterns are presented in Table 2. In the early and late recurrence groups, single intrahepatic recurrence occurred in 292 (41.3%) and 84 cases (46.7%); two or three recurrences in 255 (36.1%) and 60 cases (33.5%), and four or more recurrences in 133 (18.8%) and 24 cases (13.4%), respectively, showing no significant difference. Vascular invasion was observed in 13 (1.9%) and 6 cases (3.3%, p = 0.34), and distant metastasis in 14 (1.9%) and 5 cases (2.7%, p = 0.70), respectively. The mean tumor diameter at recurrence was 17.8 ± 8.4 mm and 17.2 ± 8.7 mm (p = 0.40), with tumors ≤ 20 mm in 499 (70.6%) and 128 cases (71.5%, p = 0.52), and ≤ 30 mm in 628 (88.8%) and 159 cases (88.8%, p = 0.39), respectively.

3.3
Treatment Modalities and Curability of Treatment
In the early and late recurrence groups, the treatment modalities used for recurrent HCC were RFA (n = 599, 84.7% vs. n = 148, 82.7%), transcatheter arterial chemoembolization (n = 77, 10.9% vs. n = 18, 10.1%), resection (n = 15, 2.1% vs. n = 7, 3.9%), systemic therapy (n = 5, 0.7% vs. n = 3, 1.7%), radiation (n = 4, 0.6% vs. n = 0, 0.0%), hepatic arterial infusion chemotherapy (HAIC) (n = 4, 0.6% vs. n = 0, 0.0%), best supportive care (n = 3, 0.4% vs. n = 3, 1.7%), with no significant difference (p = 0.20) (Table 3). Local cure was achieved in 83.5% (n = 590) and 78.8% (n = 141) of cases in the early and late recurrence groups, respectively, with no significant difference (p = 0.17).

3.4
Cases of Vascular Invasion and Distant Metastasis
Six cases of vascular invasion were observed in the late recurrence group (Figure 3) two cases each with invasion of first‐order portal vein branches and of second‐order portal vein branches and one case each with invasion of third‐order portal vein branches and of bile duct invasion. Among them, three patients could not have undergone dynamic CT or MRI because of impaired renal function or allergy to contrast agents, which led to a delayed diagnosis. Treatment modalities included resection and best supportive care in two patients each and transcatheter arterial chemoembolization and systemic therapy in one patient each. We identified three cases of pleural seeding in the late recurrence group, all of which were treated surgically (Figure 4), and one case each of peritoneal dissemination and hilar lymph node with lung metastases, both of which were treated with systemic therapy (Table 4).

3.5
Prognosis After HCC Recurrence
The median (interquartile range) survival time after HCC recurrence was 3.55 years (3.27–4.10) in the early recurrence group, 3.31 years (2.77–4.01) in the late recurrence group, showing no significant difference (p = 0.2) (Figure 5).

4
Discussion
This study examined patients with HCC who had persistently elevated tumor markers but no signs of visible recurrence on imaging after complete RFA. No difference was observed between the early and late recurrence groups in the possibility of curative treatment at HCC recurrence, with curative treatment being feasible in most cases and no significant difference in post‐recurrence survival prognosis. To our knowledge, this is the first report describing long‐term tumor marker elevation before imaging‐based diagnosis of HCC recurrence.
With regard to tumor markers, the AFP positivity rate was significantly lower in the late recurrence group compared to the early recurrence group. In HCV non‐SVR patients, AFP ≥ 200 ng/mL positivity was consistently high in both the early and late recurrence groups (37.2% and 37.3%, respectively). In HCV‐SVR patients, however, AFP ≥ 200 ng/mL identified only 9.7% and 12.8% of recurrences, whereas AFP ≥ 20 ng/mL detected 62.5% and 61.7%, supporting the use of a lower cutoff following viral eradication [21]. In contrast, among HBV and NBNC patients, AFP ≥ 200 ng/mL positivity was significantly lower in the late recurrence group than in the early recurrence group (40.0% vs. 9.5% and 16.5% vs. 0%, respectively), suggesting that AFP ≥ 200 ng/mL may have limited utility for detecting late recurrences in these populations. However, the late recurrence group exhibited a high positivity rate for the AFP‐L3 fraction. Previous studies suggest that elevated AFP‐L3 serve as an early predictor of HCC development, even when AFP is low and imaging studies do not reveal HCC, and this finding appears consistent with our result [22]. AFP and PIVKA2 have been reported to be positive even in chronic hepatitis or cirrhosis without HCC, and such cases were also observed in no recurrence group of our study [23, 24]. Among the 21 no‐recurrence cases, the majority of patients died or were lost to follow‐up during the observation period, and some may represent HCC recurrences that had not yet become detectable by imaging at censoring; in others, tumor marker elevation may have reflected underlying hepatic inflammation or cirrhosis rather than HCC recurrence. However, recurrence was observed in most of the cases within two years after the tumor marker elevation, suggesting the usefulness of surveillance after curative treatment for HCC based on a combination of three tumor markers (AFP, AFP‐L3, DCP) and imaging studies, which is consistent with the rationale underlying the GALAD score—a validated composite index incorporating these three markers alongside age and sex that has demonstrated superior sensitivity over AFP alone for HCC detection in a recent phase 3 prospective study [25]. While this study defined tumor marker positivity using fixed cutoff values, future studies incorporating patient‐specific dynamic changes may yield further insights into optimal surveillance strategies.
In this study, the tumor characteristics at HCC recurrence did not significantly differ between the early and the late recurrence group. In our institution, we regularly repeat imaging examinations every 4 months after complete RFA [26]. When HCC remained undetected despite elevated tumor markers, the follow‐up interval was reduced to 2–3 months at the discretion of the outpatient physician, after which many cases of recurrences were detected on subsequent surveillance. As a result, the tumor size was smaller than 30 mm in diameter in nearly 90% of cases and could be treated curatively. The tumor volume doubling time in HCC has been reported to be heterogeneous, with approximately 30% of the tumors classified as rapidly growing (< 3 months) [27, 28]. Intervals shorter than 2 months may increase detection of tumors suitable for curative local treatment.
We observed five cases of distant metastasis. Among these, pleural seeding was observed in three cases, and resection was performed in all three. Reports suggest that resection of disseminated lesions may be beneficial in cases without intrahepatic lesions [29, 30, 31] and chest CT may contribute to improved prognosis. It is unclear whether the early detection of other extrahepatic metastases leads to an improvement in prognosis, because these conditions are indications for systemic therapy.
We observed six cases of vascular invasion. Long‐term survival with surgical resection has been reported in cases of vascular invasion limited to the second‐order branch [32, 33]. Therefore, detecting vascular invasion within that extension may contribute to improving the prognosis. Of the four cases with inoperable portal vein invasions in this study, three had imaging limitations owing to renal dysfunction or allergy to contrast media. The usefulness of non‐contrast MRI including diffusion‐weighted imaging (DWI), and contrast‐enhanced ultrasound (CEUS) has been demonstrated in patients with renal impairment [34, 35]. Careful surveillance using these alternative modalities is recommended for patients with imaging limitations.
Our study has several limitations that should be considered when interpreting the results. First, being a retrospective single‐institution study conducted at our large facility with advanced diagnostic equipment, we might have detected recurrences earlier than those feasible in institutions with different resources. Second, when HCC was not detected despite elevated tumor markers, the follow‐up interval was shortened at the discretion of the outpatient physician. Therefore, it is impossible to discuss the necessity of shortening the follow‐up interval. Third, our study focused on patients treated with RFA, and the results may not be directly applicable to patients treated with other modalities such as surgical resection or liver transplantation. Fourth, because the same patient may contribute multiple cases if tumor‐free status was achieved after curative treatment, the assumption of independence between observations may not strictly hold, which may affect the precision of statistical comparisons.
In conclusion, in post‐RFA surveillance for HCC, appropriate surveillance strategies can enable curative treatment for subsequent recurrence in many cases, regardless of elevated tumor markers without detectable intrahepatic recurrence. However, if intrahepatic lesions are not detected, chest CT should be considered in cases with possible pleural dissemination, and in patients in whom limited imaging modalities can be used, caution is needed regarding the possibility of vascular invasion.

Author Contributions

Tomoharu Yamada: data curation, resources. Masaya Sato: data curation, resources. Makoto Moriyama: conceptualization, methodology, data curation, investigation, validation, formal analysis, visualization, project administration, writing – original draft, writing – review and editing. Mitsuhiro Fujishiro: supervision. Takuma Nakatsuka: data curation, resources. Ryosuke Tateishi: conceptualization, methodology, data curation, supervision, investigation, funding acquisition, project administration, writing – original draft, writing – review and editing.

Funding
This work was supported by Japan Agency for Medical Research and Development, JP22fk0210066, JP22fk0210090, JP25fk0210149. Welfare Policy Research Grants from the Ministry of Health, Labour and Welfare of Japan, 23HC2001.

Conflicts of Interest
Ryosuke Tateishi has received a research grant from Fujifilm Wako Pure Chemical Corporation.

Supporting information

Figure S1: Temporal relationship between tumor marker levels and HCC recurrence.
The schema illustrates cases where HCC recurred more than 180 days after tumor marker levels exceeded the cut‐off value during post‐curative treatment surveillance.
HCC: hepatocellular carcinoma

Figure S2: Cumulative incidence of HCC recurrence after tumor marker elevation.
Cumulative incidence curve showing the probability of HCC recurrence following tumor marker elevation. Numbers at the bottom indicate patients at risk at each time point.
HCC: hepatocellular carcinoma

Table S1: Characteristics of the non‐recurrence group and late recurrence group