1 Introduction

Primary liver cancer is a common and highly malignant tumor worldwide. According to global cancer statistics for 2023, primary liver cancer is the sixth most common malignancy and third leading cause of cancer-related death. Hepatocellular carcinoma (HCC) is the predominant subtype of primary liver cancer, accounting for 75–85% of cases [1]. At diagnosis, most HCC cases are already in advanced stages, limiting the opportunity for radical treatment [2].

Early-stage HCC can be treated with local therapies, such as radiofrequency ablation, partial hepatectomy, or liver transplantation; however, recurrence rates remain high, and the 5-year survival rate is typically below 30 [3]. For patients with recurrent or unresectable advanced HCC, systemic and local treatments are the main therapeutic options [4]. Consequently, combined treatments, including targeted-immunotherapy combinations such as atezolizumab with bevacizumab or pembrolizumab with lenvatinib, are emerging as key options [5, 6].

Non-invasive imaging modalities, such as computed tomography and magnetic resonance imaging (MRI), are essential for diagnosing and assessing treatment response in HCC, but their high cost limits frequent use. Alpha-fetoprotein (AFP), expressed in approximately 70% of patients with HCC, is a widely used serum biomarker in clinical practice [7] due to its specificity, safety, and low cost. AFP levels are closely associated with HCC development and progression. Elevated AFP is generally linked to a higher risk of disease progression and poor prognosis [8]. Furthermore, AFP response during treatment predicts the efficacy of both systemic and local therapies [9]. Studies have shown that early decreases in AFP following targeted or combined immunotherapy correlate with better outcomes and prognosis [10].

Clinical data from our center indicate that dynamic changes in AFP levels correlate with tumor size and treatment efficacy. Recently, however, we observed a subset of advanced patients with HCC receiving targeted immunotherapy who exhibited AFP decreases at certain stages despite disease progression, with imaging follow-ups ruling out pseudoprogression. This unusual AFP trajectory has not been reported in current research. Understanding this phenomenon may offer new insights into HCC immune escape mechanisms and prompt a re-evaluation of AFP’s role in monitoring HCC treatment. Thus, we collected data from patients with advanced HCC who underwent systemic and/or local treatments, and conducted a detailed analysis of their characteristics, treatment responses, prognosis, AFP changes, and potential causes of abnormal AFP trajectories.

2 Materials and methods

2.1 Clinical data

We retrospectively collected data from patients with advanced HCC treated at the First Affiliated Hospital of Zhengzhou University between March 2021 and May 2023. All patients received systemic and/or local treatment for a minimum of 1 month.

The inclusion criteria were as follows: (1) histologically, cytologically, or clinically diagnosed HCC; (2) HCC not suitable for surgical resection; (3) at least one measurable target lesion at time of initial treatment evaluation; (4) Child–Pugh liver function score ≤ 9; (5) Eastern Cooperative Oncology Group performance status score ≤ 1; and (6) baseline AFP level ≥ 20 ng/mL.

The exclusion criteria were as follows: (1) malignant tumors originating from other organs; (2) autoimmune disorders requiring immunosuppressive treatment; (3) multi-organ failure; (4) prior systemic anti-tumor therapy, including chemotherapy, targeted therapies, or immune checkpoint inhibitors; and (5) incomplete follow-up data.

This study was approved by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University (2023-KY-0146).

2.2 Procedures

2.2.1 Grouping

This study included 139 patients. Eleven patients, whose AFP levels decreased despite HCC progression, were classified into the AFP abnormal trajectory group. The remaining 128 patients, whose AFP levels changed in accordance with tumor progression, were categorized into the AFP normal trajectory group. The abnormal AFP trajectory is defined as a ≥ 20% absolute - value reduction from the previous level, or stability (absolute - value change < 20%), alongside HCC progression.

2.2.2 Evaluation

AFP levels were recorded at each stage of the study, from diagnosis until the end of the follow-up period. Upon enrollment, enhanced computed tomography and MRI were used to assess tumor size. Treatment efficacy was evaluated 3 months post-treatment using the RECIST 1.1 criteria. The overall response rate (ORR) was defined as the proportion of patients achieving complete response or partial response, while the disease control rate (DCR) included complete response, partial response, and stable disease. Progression-free survival (PFS) was defined as the time from initial treatment to tumor progression or death, while overall survival (OS) was defined as the time from treatment initiation to death from any cause or the end of the follow-up period.

2.3 Statistical analyses

Statistical analysis was performed using SPSS (version 26.0; IBM Corp., Armonk, NY, USA) and GraphPad Prism (version 9.0; La Jolla, CA, USA). Data that were not normally distributed are presented as medians (interquartile ranges), while normally distributed data are presented as means ± standard deviations. Categorical data are presented as percentages and analyzed using the χ2or Fisher’s exact test. The Kaplan–Meier method was used to calculate the median PFS (mPFS) and median OS (mOS), and survival curves were plotted. Differences between groups were assessed using the log-rank test. Logistic regression analysis was used to evaluate risk factors associated with the abnormal AFP trajectory phenomenon, and results are expressed as odds ratios (ORs) with 95% confidence intervals (CIs). A P-value of < 0.05 was considered statistically significant.

3 Results

3.1 General clinical data

Overall, 139 patients were included in the study, all of whom received systemic and/or local treatment at our hospital between March 2021 and May 2023. Of these patients, 111 (79.9%) were male and 28 (20.1%) were female, with a mean age of 55.3 ± 9.2 years. The median baseline AFP level was 515.0 (range: 50.0–5176.0) ng/mL. Among the patients, 126 (90.6%) had hepatitis B, and two (1.4%) had hepatitis C. Fifty patients (35.9%) had portal vein thrombosis, and 54 (38.8%) had extrahepatic metastases. In terms of staging, 64 patients (46.1%) were classified as Barcelona Clinic Liver Cancer Stage C, 40 patients (28.8%) had Child–Pugh B grade, and 98 patients (70.5%) had an Eastern Cooperative Oncology Group performance status score of 0.

In the AFP normal trajectory group (n = 128), 14 patients (10.9%) received interventional therapy, 25 patients (19.5%) received targeted therapy, 20 patients (15.7%) received a combination of targeted and immunotherapy, and 69 patients (53.9%) received a combination of interventional therapy, targeted therapy, and immunotherapy. In the AFP abnormal trajectory group (n = 11), one patient (9.1%) received interventional therapy, one patient (9.1%) received targeted therapy, seven patients (63.6%) received combined targeted therapy and immunotherapy, and two patients (18.2%) received a combination of interventional therapy, targeted therapy, and immunotherapy.

Baseline characteristic analysis of the 139 patients revealed no significant differences between the two groups, except for treatment modality (P = 0.008) and the neutrophil-to-lymphocyte ratio (NLR; P = 0.023; Table 1).

Table 1 Baseline patient characteristics
Full size table

3.2 Evaluation of ORR and DCR

Tumor response was evaluated 3 months after initial treatment. Of the 139 patients, none achieved a complete response; 24 (17.3%) showed partial response, 90 (64.8%) exhibited stable disease, and 25 (17.9%) showed progressive disease. The ORR in the AFP abnormal trajectory group was significantly lower than in the AFP normal trajectory group (0% vs. 18.8%, P = 0.037), and the DCR was also significantly lower (54.5% vs. 84.3%, P = 0.027). The differences between groups were statistically significant (Table 2).

Table 2 Comparison of treatment effects between the alpha-fetoprotein abnormaltrajectory and normal trajectory groups
Full size table

3.3 Evaluation of PFS and OS

The mPFS for the AFP abnormal trajectory group was 2.8 (95% CI: 1.34–4.25) months, significantly shorter than the mPFS of 14.2 (95% CI: 10.57–17.82) months observed in the AFP normal trajectory group (P < 0.001). Similarly, the mOS for the AFP abnormal trajectory group was 14.0 (95% CI: 4.95–23.04) months, markedly lower than the 32.0 (95% CI: 16.4–47.5) months reported for the AFP normal trajectory group (P = 0.039). These findings underscore significant survival disparities between the two groups (Fig. 1).

Fig. 1
figure 1

Comparison of survival time between the AFP normal trajectory and abnormal trajectory groups. (A) Comparison of the median PFS between the two groups (B). Comparison of median OS between the two groups

Full size image

3.4 Characteristics of patients in the AFP abnormal trajectory group

Of the 11 patients in the AFP abnormal trajectory group, one receiving targeted therapy exhibited abnormal AFP trends during treatment, and another undergoing transarterial chemoembolization (TACE) showed abnormalities at treatment initiation. Among the two patients receiving TACE combined with targeted immunotherapy, one exhibited abnormalities initially and the other during treatment. Of the seven patients on targeted immunotherapy, four presented abnormalities initially, while three developed them during treatment. The median duration of the abnormal AFP phenomenon was 63 (range: 36–97) days, with two patients still exhibiting abnormalities at the end of follow-up. The median treatment time when the abnormal phenomenon occurred was 83 (range: 32–249) days. The albumin-bilirubin grade (ALBI) levels of these 11 patients mostly increased with the progression of liver cancer or fluctuated when the tumor was stable. No obvious regularity or characteristics were observed.

After the onset of abnormal AFP trajectories, treatment plans were not altered in two patients, and their HCC continued to progress. One patient discontinued targeted immunotherapy, but the disease advanced. Eight patients modified their treatment strategies: seven transitioned to triple therapy (TACE combined with targeted immunotherapy, including multiple changes in targeted drugs), but HCC progression persisted. One patient received radiofrequency ablation for new small lesions, leading to temporary disease stabilization. We statistically analyzed the specific medications used in the treatment of the Abnormal AFP Trajectory Group (Table 3).

Table 3 Specific treatment regimens for the abnormal AFP trajectory group
Full size table

To better comprehend this abnormal phenomenon, a detailed account of one case from the AFP abnormal trajectory group will be provided. A 54-year-old male with an initial AFP level of 176.85 ng/mL was diagnosed with HCC via MRI, which revealed a 41.7-mm tumor in the left liver lobe, multiple intrahepatic metastases, and lymph node involvement. The patient initiated TACE combined with targeted immunotherapy (camrelizumab + anlotinib). After 30 days of treatment, his AFP level dropped to 76.08 ng/mL, but MRI showed an increase in the left outer lobe lesion size to 69.7 mm andexpansion of other lesions. At 59 days, the AFP level declined further to 42.5 ng/mL, but MRI indicated continued lesion growth in the left and right lobes, along with portal vein invasion. After 84 days, AFP reached 34.20 ng/mL, while MRI revealed increased lesion sizes and counts in the right lobe, with some lesions expanding further. Despite persistent tumor progression, the treatment regimen remained unchanged. The patient exhibited the abnormal AFP trajectory for 84 days, with a survival duration of 6.7 months. The variations in serum alpha-fetoprotein levels during the diagnosis of HCC and disease progression in this patient were presented more intuitively through images (Fig. 2).

Fig. 2
figure 2

Changes in serum alpha fetoprotein levels during HCC diagnosis and disease progression

Full size image

3.5 Regression analysis of AFP abnormal trajectory phenomenon

Clinical data from 139 patients were analyzed to identify factors associated with abnormal AFP trajectories. Variables including gender, baseline NLR, presence of extrahepatic metastasis, major vascular invasion, Child–Pugh classification, and treatment modality were evaluated using univariate logistic regression.

Results indicated that combined targeted immunotherapy and chemotherapy (OR: 9.21, 95% CI: 1.04–81.30, P = 0.046) and elevated NLR (OR: 1.22, 95% CI: 1.05–1.43, P = 0.008) were risk factors for an abnormal AFP trajectory. Multivariate logistic regression analysis showed that combined targeted immunotherapy and chemotherapy (OR: 15.35, 95% CI: 1.22–192.90, P = 0.034) and elevated NLR (OR: 1.29, 95% CI: 1.03–1.62, P = 0.025) were independent risk factors for an abnormal AFP trajectory (Table 4).

Table 4 Regression analysis of influencing factors for alpha-fetoprotein abnormal trajectory phenomenon
Full size table

4 Discussion

Recently, combination therapies have become the standard treatment for advanced HCC [11]. AFP is a widely recognized biomarker for HCC diagnosis, monitoring, and prognosis, with changes in AFP levels closely correlated with treatment outcomes and prognosis [12]. However, an unusual phenomenon was observed during advanced HCC treatment, particularly in patients receiving combination regimens that include targeted immunotherapy. In these cases, AFP levels decreased despite continued HCC progression. Previous studies have noted that in the progressive disease (PD) subgroup, patients treated with ramucirumab had a significantly lower rise in AFP from baseline compared with those treated with placebo [13]. This suggests that despite disease progression, elevated AFP levels can be suppressed in some patients, and AFP may even decline. Here, for the first time, we analyse this phenomenon in detail.

In clinical practice, immune checkpoint inhibitors occasionally lead to tumor enlargement or new lesion formation, followed by subsequent tumor reduction or resolution, a phenomenon known as pseudoprogression [14]. This is believed to result from immune cell infiltration, tumor necrosis, hemorrhage, and lesion edema [15]. Pseudoprogression is primarily identified through imaging follow-up, with clinical trials reporting its incidence in melanoma (4.6–8.3%), non-small cell lung cancer (1.8–6.9%), and renal cell carcinoma (4.9–7.1%). In HCC, however, pseudoprogression is rare [16,17,18]. Herein, the observed abnormal AFP trajectory phenomenon in patients with HCC—characterized by AFP reduction concurrent with HCC progression—represents true progression rather than pseudoprogression. This is confirmed by continued imaging-documented tumor growth and later re-elevation of AFP levels alongside disease progression.

Among the 11 patients with abnormal AFP trajectories, nine were treated with targeted immunotherapy as part of their regimen. These abnormalities occurred either at therapy initiation or during treatment, with six patients exhibiting abnormalities early and five developing them later. The median duration of this phenomenon was 63 (range: 36–97) days, with two patients still demonstrating abnormal trajectories at the end of follow-up. For the remaining nine patients, AFP changes coincided with synchronous imaging evidence of tumor progression. This suggests that abnormal AFP trajectories occur at specific treatment stages rather than consistently throughout the disease course. These phenomena may be transient, making them difficult to detect without regular monitoring.

In this cohort of 139 patients with advanced HCC receiving systemic and/or local treatments, 11 patients (7.9%) exhibited abnormal AFP trajectories, indicating a relatively low incidence. This phenomenon may have been underreported in the past due to its transient nature and potential for being overlooked. However, in this study, the inclusion of regular follow-ups at 1–2-month intervals facilitated the identification of such abnormalities. These abnormal AFP trajectory phenomena reduce the reliability of AFP as a standalone marker for evaluating treatment efficacy and prognosis.

To improve detection and accurate assessment of tumor progression in patients with advanced HCC, dynamic analysis of AFP levels must be integrated with imaging studies. This combined approach is critical for identifying tumor progression and optimizing patient management strategies.

Of the 11 patients exhibiting abnormal AFP trajectories, eight altered their treatment plans following the onset of this phenomenon. Among these, seven patients transitioned to triple therapy involving TACE and targeted immunotherapy, accompanied by multiple adjustments in targeted drugs; however, disease progression persisted. In contrast, one patient achieved stable disease following radiofrequency ablation for newly developed small lesions. The ORR and DCR of the AFP abnormal trajectory group were significantly lower than those of the AFP normal trajectory group at 3 months post-treatment. Additionally, the mPFS and mOS were significantly shorter, underscoring the association of abnormal AFP trajectories with poor treatment efficacy, rapid disease progression, potential drug resistance, and an overall unfavorable prognosis. This phenomenon highlights the necessity of close clinical monitoring.

Notably, nine out of 11 patients with abnormal AFP trajectories were receiving targeted immunotherapy when the abnormal trajectory emerged. Multivariate logistic regression analysis identified combined targeted immunotherapy as an independent risk factor for this phenomenon, with an OR of 15.35, indicating a strong association. Beyond its role as a molecular marker, AFP plays a critical role in HCC pathogenesis and progression. Specifically, AFP contributes to tumorigenesis by inhibiting the p53/Bax/caspase-3 apoptosis pathway, inducing the malignant transformation of hepatocytes, promoting tumor invasion and angiogenesis, and suppressing immune cell function. These processes collectively enable HCC cells to evade immune surveillance, survive, and metastasize [19, 20].

Therefore, it is speculated that during advanced HCC treatment—particularly with targeted immunotherapy—tumor cells may actively downregulate AFP expression as a mechanism to evade AFP-targeted immune responses. This abnormal trajectory phenomenon could represent a form of tumor immune evasion and self-protection, accompanied by the development of drug resistance. Further research is required to elucidate the precise molecular mechanisms underlying this phenomenon, which may provide new insights into overcoming treatment resistance in advanced HCC.

Neutrophils play a significant role in cancer progression by enhancing angiogenesis, promoting tumor cell proliferation and metastasis, and facilitating immune evasion [21]. A reduction in the proportion of lymphocytes suggests a weakened cell-mediated immune response to cancer [22]. Research indicates that neutrophils may undergo immunosuppressive changes in cancer, particularly in their interactions with T cells. Strategies to inhibit neutrophil mobilization and recruitment into tumors may enhance the efficacy of cancer immunotherapy. A low NLR has been identified as an important predictor of pembrolizumab efficacy [23], whereas a high NLR is recognized in clinical practice as an independent prognostic indicator of poor cancer outcomes and a marker of systemic inflammation [24, 25].

In the context of HCC, studies reveal that an elevated baseline NLR is associated with worse survival outcomes. For instance, among patients receiving TACE as first-line therapy for stage B HCC, a baseline NLR of ≥ 3 was linked to disease progression within 6 months. Similarly, in a cohort of 380 patients with newly diagnosed HCC undergoing TACE, those with a normal baseline NLR (≤ 2.4) had a mOS of 29.1 months, compared to 19.1 months [26] for those with an elevated NLR (> 2.4).

Herein, the baseline NLR in the AFP normal and abnormal trajectory groups was 2.54 and 3.98, respectively. The ORR, DCR, PFS, and OS were significantly lower in the AFP abnormal trajectory group. These findings indicate that patients in this group experienced poorer treatment efficacy and faster disease progression. Consequently, the baseline NLR should be considered a prognostic factor in HCC. High baseline NLR was also identified as an independent risk factor for the AFP abnormal trajectory phenomenon, which is consistent with previous studies linking NLR to immune escape mechanisms.

This study has several limitations. As a retrospective analysis, potential confounding factors may have influenced the results. Between March 2021 and May 2023, 139 patients were recruited. The abnormal trajectory phenomenon was detected in 11 hepatitis B - positive cases only. This likely stems from hepatitis B being the leading cause of HCC in China. Given the small sample size, the phenomenon might occur in non - viral HCC cohorts but was not observed in our study. Moreover, the small sample size of the abnormal - trajectory group and its internal heterogeneity may have undermined the study’s conclusions, underscoring the necessity for large - scale prospective clinical validation.

In conclusion, a subset of patients with advanced HCC exhibit abnormal AFP trajectories during treatment. This phenomenon is associated with HCC progression, resistance to current targeted immunotherapy, and poor prognosis. Modifying targeted immunotherapy regimens in these patients may not improve outcomes. Combined targeted immunotherapy and elevated baseline NLR are independent risk factors for the development of this phenomenon. Clinically, patients exhibiting these risk factors require close monitoring to optimize management strategies. In the future, it is necessary to conduct relevant experiments to clarify the mechanism by which AFP participates in immune escape during HCC treatment, so as to explore related treatment options.