Patients with hepatocellular carcinoma achieving a complete response to sorafenib: Three case reports and review of literature

Abstract

BACKGROUND

Sorafenib has been the conventional treatment for advanced hepatocellular carcinoma (HCC) since 2008. While radiological complete responses are extremely rare, improved supportive care and multidisciplinary approaches in clinical practice may explain the recent increase in case reports and retrospective series documenting such responses.

CASE SUMMARY

This case series describes 3 patients with advanced HCC who achieved durable complete responses using first-line sorafenib therapy, even in the presence of portal vein thrombosis or extrahepatic spread, and highlights the potential for sustained remission in selected patients. Dermatologic toxicity and non-viral etiology may correlate with favorable outcomes; however, reliable predictive biomarkers for sorafenib response are lacking.

CONCLUSION

Future research into the etiology and molecular differences in HCC is necessary to develop more personalized therapy options.

Key Words: Hepatocellular carcinoma; Sorafenib; Complete response; Tyrosine kinase inhibitor; Case report

Core Tip: Complete responses to sorafenib in advanced hepatocellular carcinoma are extremely rare. This case series reports 3 patients who, despite having advanced hepatocellular carcinoma with features like portal vein thrombosis and extrahepatic spread, achieved durable complete responses on sorafenib therapy. Predictive biomarkers for sorafenib response remain unknown, although dermatologic toxicity and non-viral etiology may be associated with favorable outcomes. These findings support the continued utility of tyrosine kinase inhibitors in selected patients, particularly when immunotherapy is contraindicated or unavailable.

INTRODUCTION

Primary liver cancer is the sixth most commonly diagnosed cancer worldwide and the third leading cause of cancer-related death. Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer, accounting for 75%-85% of all cases[1]. Although curative methods are available for the 40%-50% of patients diagnosed at an early stage, most individuals will eventually require systemic therapy. The prognosis of HCC remains poor, with a 5-year survival rate under 20% in its advanced stages, and a high mortality-to-incidence ratio[2]. Liver cirrhosis is a primary risk factor for HCC, and managing both conditions presents significant treatment challenges. The most important factors in the development of cirrhosis and subsequent HCC are hepatitis B and C, alcohol consumption, and metabolic dysfunction-associated fatty liver disease, with metabolic dysfunction-associated fatty liver disease significantly increasing in incidence[3]. The Barcelona Clinic Liver Cancer (BCLC) staging system determines treatment for liver cancer. Patients with advanced-stage disease (BCLC C) or whose disease is unsuitable for locoregional therapy are treated with systemic drugs as the standard of care. Current first-line options according to European Society for Medical Oncology guidelines include combinations of atezolizumab and bevacizumab, tremelimumab and durvalumab, ipilimumab and nivolumab, and camrelizumab and rivoceranib.

Immunotherapy combinations have demonstrated superior overall survival (OS), progression-free survival (PFS), overall response rates (ORRs), and complete response (CR) rates compared to tyrosine kinase inhibitors (TKIs). TKIs such as sorafenib and lenvatinib remain effective first-line alternatives when immunotherapy is contraindicated[4]. TKIs are multikinase inhibitors that mainly target the vascular endothelial growth factor receptor (VEGFR), platelet-derived growth factor receptor, rapidly accelerated fibrosarcoma protein receptor, fibroblast growth factor receptor, KIT receptor, and rearranged during transfection receptor[5]. Their mechanism of action involves the inhibition of tyrosine kinase receptors, which are responsible for cellular proliferation and angiogenesis through the signal transduction pathway. Some TKIs also exert immunomodulatory effects by influencing the tumor microenvironment[6]. Sorafenib, a VEGFR, platelet-derived growth factor receptor, and RAF kinase inhibitor, has been the standard of care for first-line treatment of advanced HCC since 2008.

The phase 3, prospective, multicenter randomized SHARP trial demonstrated a benefit in OS, with a median survival of 10.7 months for patients receiving treatment compared to 7.9 months in the placebo group[7]. These findings were later confirmed by a separate Phase 3 Asia-Pacific study[8]. While no cases of CR were reported in these two studies, sporadic cases of CR have been observed in other trials where sorafenib was used as a comparator[9], as well as in retrospective analyses[10,11] and case reports[12-18]. Real-world data and recent studies have revealed a considerable improvement in OS and ORR with sorafenib, surpassing the outcomes initially reported in its registration trials. The exact reasons for this upward trend are not fully understood, but potential contributing factors may include advancements in supportive care, a rise in multidisciplinary management approaches, and effective treatments for hepatitis C[9,19]. This case series presents 3 patients undergoing treatment at the University Hospital Centre Zagreb (Zagreb, Croatia). All of the patients were diagnosed with advanced HCC, began treatment with sorafenib, and experienced a CR. The third case was previously described in a separate case report[12].

CASE PRESENTATION

Chief complaints

Case 1: Persistent right upper quadrant abdominal pain lasting for more than 1 year.

Case 2: The patient was asymptomatic and presented during routine follow-up for known liver cirrhosis.

Case 3: Right lumbar pain of several weeks’ duration.

History of present illness

Case 1: A 63-year-old man was evaluated for ongoing pain in the upper quadrant of his abdomen in April 2020. Abdominal ultrasound revealed a 3.5-cm lesion in segment IVb of the liver. A subsequent computed tomography (CT) scan confirmed imaging features consistent with HCC. Preoperative alpha-fetoprotein (AFP) was elevated at 111 ng/mL, exceeding the normal range of < 10 ng/mL.

Case 2: A 55-year-old man with a history of cryptogenic liver cirrhosis, first diagnosed following variceal bleeding in 2018.

Case 3: The patient was a 73-year-old man who reported to the Emergency Department in August 2020 with right lumbar pain. The patient had liver cirrhosis; however, no underlying etiology except metabolic dysfunction-associated steatotic liver disease was found.

History of past illness

Case 1: The patient had a history of chronic hepatitis C virus infection that had been previously treated with interferon in the 1990s.

Case 2: The patient had a medical history of arterial hypertension and cholelithiasis.

Case 3: There was no significant past medical history.

Personal and family history

The three patients reported no family history of malignancy or liver disease or other significant comorbidities.

Physical examination

All three patients’ physical examination findings were unremarkable.

Laboratory examinations

Case 1: Baseline laboratory evaluation demonstrated preserved hepatic function and elevated AFP (111 ng/mL) prior to surgery.

Case 2: At the initiation of systemic therapy, the AFP level was significantly elevated at 2500 ng/mL.

Case 3: Serum AFP was 17.3 ng/mL, with all other laboratory parameters falling within normal range.

Imaging examinations

Case 1: Initial abdominal ultrasound and subsequent multiphasic CT demonstrated a 3.5-cm hepatic lesion in segment IVb, which displayed imaging characteristics consistent with HCC.

Case 2: Two lesions, each 5 cm in size, were found on ultrasound in liver segments VIII and IV during routine follow-up in February 2021. Lesions had radiological characteristics consistent with HCC on follow-up CT.

Case 3: Further workup, which included a multislice CT scan of the thorax, abdomen, and pelvis, identified a 7-cm mass in the liver’s caudate lobe, accompanied by portal and splenic vein thrombosis (Figure 1A and B). Magnetic resonance imaging (MRI) of the abdomen showed no additional lesions. Due to the lesion’s proximity to major vascular structures, a biopsy was not performed.

Figure 1 Computed tomography and magnetic resonance imaging scan results. A: Computed tomography scan showing thrombosis of the portal, splenic, and superior mesenteric veins with a probable tumor thrombus[12]; B: Computed tomography scan showing an expansive caudate lobe (5.4 cm × 3.7 cm), portal vein thrombosis, and fatty infiltration of the rest of the liver parenchyma[12]; C: Computed tomography scan showing necrosis of the largest liver lesion; D: Post-procedural computed tomography with partial regression of the lesions; E: Magnetic resonance imaging of the necrotic liver lesion; F: Last follow-up computed tomography scan, which showed no signs of a viable tumor. Citation for Figure 1A and B: Adžić G, Prejac J, Pleština S. Long-term follow-up of complete remission in a patient with advanced hepatocellular carcinoma treated with sorafenib: A case report. Front Oncol 2023; 13: 1260989. Copyright © 2023 Adžić, Prejac and Pleština. Published by Frontiers Media SA (Supplementary material).

FINAL DIAGNOSIS

All three patients, based on the radiological characteristics, a diagnosis of HCC was established.

TREATMENT

Case 1

The lesion was surgically resected in May 2020. The pathohistological report confirmed a diagnosis of HCC characterized by significant invasion into small blood vessels and the capsule surrounding the liver. The remaining liver tissue showed no sign of cirrhosis. The patient was closely monitored until October 2021, at which point new lesions were found in liver segments III, VIII, I, and IV, accompanied by portal vein thrombosis. The AFP value at the time was 7.5 ng/mL. The patient’s case was presented at a multidisciplinary team (MDT) meeting, leading to initiation of treatment with 400 mg sorafenib twice daily. At the initial radiological evaluation in February 2022, the largest lesion in liver segment III exhibited slight progression caused by necrosis. All other lesions remained stable in size, so treatment with sorafenib was continued (Figure 1C). Subsequent imaging revealed partial regression of all lesions. The patient continued sorafenib treatment for more than 2 years, with no signs of radiological progression and AFP values within the normal range. The only adverse event was grade 1 diarrhea, as per the Common Terminology Criteria for Adverse Events (CTCAE), which was successfully managed with symptomatic treatment without the need for dose reduction or interruption.

Case 2

Since the disease did not meet the Milan criteria, radiofrequency ablation was performed in March 2021. Follow-up positron emission tomography (PET)-CT imaging revealed progression in the size of both lesions, which remained metabolically active. The case was presented at the MDT meeting, and transarterial chemoembolization was recommended. Post-procedural CT confirmed partial regression of the lesions (Figure 1D). Another PET-CT scan was performed in April 2022. Although there were no metabolically active lesions in the liver, two new lesions were found: A metabolically active nodule near the right ventricle in the pericardium and a metabolically active lymph node adjacent to the right internal thoracic artery, suggesting extrahepatic disease spread. At that time, the AFP level was significantly elevated at 2500 ng/mL. The patient maintained normal liver function, as indicated by a Child-Pugh (CP) score of A5, and exhibited an Eastern Cooperative Oncology Group performance status of 0. Systemic treatment with 400 mg sorafenib, administered twice daily, started in May 2022. The dose was initially halved to 200 mg twice daily within the first month because of skin toxicity, specifically a grade 2 hand-foot skin reaction (HFSR). However, the patient was able to resume the full dose after 2 months.

Case 3

An atypical caudate lobe resection was performed in September 2020, following the decision that surgical removal was the optimal treatment approach. According to the pathohistological report, the tumor was a grade 3 HCC and staged at T4N0M0, with no perineural or lymphovascular invasion observed; however, the surgical margin was close, measuring less than 1 mm. On the October 2020 post-resection MRI, a new 1.3-cm lesion was identified in the VI segment, along with portal vein tumor thrombosis. Following a review by the MDT, a decision was made to initiate systemic treatment with sorafenib. The patient’s AFP level was 4.1 ng/mL; his CP class was A5, and his Eastern Cooperative Oncology Group performance status was 0. Sorafenib therapy was initiated in October 2020 at the standard dosage of 400 mg twice daily. After the first cycle, the patient developed a maculopapular rash on his hands and feet. After the sixth cycle, the patient developed grade 2 diarrhea, according to CTCAE guidelines. As a result, the sorafenib dosage had to be reduced to 200 mg twice daily.

OUTCOME AND FOLLOW-UP

Case 1

An MRI scan in March 2024 showed that the largest lesion in liver segment III was 55 mm. This area, along with all of the other lesions, appeared to have undergone complete necrosis with no viable tumor cells found (Figure 1E). The patient was re-evaluated by the MDT as a potential liver transplantation candidate; however, the procedure was deemed surgically unfeasible due to cavernous transformation of the portal vein. Sorafenib was discontinued in June 2024, and the patient remains under regular follow-up. As of the most recent assessment in June 2025, there are no signs of active disease.

Case 2

The patient’s treatment continued until October 2024, but was subsequently paused and ultimately discontinued due to the development of grade 2 thrombocytopenia per CTCAE. A follow-up PET-CT scan conducted in October 2024 revealed no evidence of metabolically active lesions. The last imaging follow-up was a CT scan in April 2025, which showed no signs of a viable tumor, and the AFP value was 2 ng/mL (Figure 1F).

Case 3

On the first follow-up CT scan, no lesion was visualized in the VI segment, a finding that remained consistent across subsequent CT scans repeated every 2 months. Additionally, the AFP levels remained within the normal range. The patient was treated with sorafenib until February 2024, at which point treatment was stopped. Follow-up imaging in October 2024 revealed no evidence of active disease, and the patient remains under close surveillance.

DISCUSSION

All patients were treated with sorafenib, which was the standard of care in the Republic of Croatia at the time due to the unavailability of immunotherapy. Notably, all patients in this case series achieved a CR. Although the treatment of HCC has significantly advanced with the introduction of immunotherapy, a subset of patients continues to receive TKIs as first-line therapy. This is typically due to contraindications to immunotherapy, concerns about potential side effects, or specific exclusion criteria from clinical trials, such as a history of variceal bleeding or the presence of portal vein thrombosis[9]. It is also important to note that contemporary immunotherapy trials have shown considerable improvement in OS and ORR using sorafenib as a comparator compared to earlier studies, such as SHARP[7] and Asia-Pacific trial[8], especially in patients with non-viral HCC etiology[9,19]. Recently, propensity score matching analysis in patients with non-viral HCC suggested that lenvatinib may offer better PFS and OS rates compared to the combination of atezolizumab and bevacizumab. The analysis also indicated that sorafenib performs similarly to the atezolizumab and bevacizumab combination in terms of survival outcomes[20]. These findings support the role of sorafenib as an effective first-line treatment option for patients.

However, the question of when to choose a TKI over immunotherapy in the first-line setting remains unresolved, primarily due to the lack of reliable biomarkers to predict treatment response. To date, AFP level ≥ 400 ng/mL remains the only validated positive predictive biomarker, specifically for the use of ramucirumab in the second-line setting[21]. Two studies have shown that tracking the decline of AFP levels could serve as a surrogate marker for PFS in patients with baseline levels > 20 μg/L; specifically, a reduction of > 20% after 6 weeks of sorafenib is notable[22,23]. However, not all patients with metastatic disease present with elevated AFP levels at diagnosis; importantly, in all 3 of our cases, AFP was below the limit of detection when a CR was achieved (Table 1). Both CP and BCLC staging systems have consistently shown across prospective studies, registries, and pooled trial analyses to predict survival in patients with HCC receiving sorafenib, with better outcomes observed in patients with preserved liver function (CP-A) and less advanced disease (BCLC-B). Patients with more severe liver impairment (CP-B) and advanced cancer (BCLC-C) experience significantly less benefit. This prognostic pattern holds true for newer immune checkpoint inhibitor combination therapies as well[24,25].

Table 1 Serological alpha-fetoprotein values at different timepoints.

Timepoint1	Case 1	Case 2	Case 3
At diagnosis	111	2508	17.3
At initiation of sorafenib treatment	7.5	2837	4.1
At first radiological appearance of a complete response	< 2.0	2.1	< 2.0
Last available measurement	3.0	< 2.0	< 2.0

¹Reference range: < 10 μg/mL.

Some evidence suggests that patients who experience adverse events, particularly dermatologic and gastrointestinal toxicities, may have a favorable response to sorafenib treatment. Among these, HFSR is most consistently associated with better survival, as multiple retrospective and prospective analyses have demonstrated significantly longer PFS and OS in patients who develop early or higher grade HFSR compared to those without this dermatologic toxicity[26,27]. This association is believed to reflect adequate systemic drug exposure and effective inhibition of the rapidly accelerated fibrosarcoma/VEGFR pathway within both tumor and endothelial cells. Similarly, gastrointestinal adverse events, most notably diarrhea, have also been correlated with favorable treatment outcomes, possibly due to shared mechanisms of kinase inhibition affecting epithelial and vascular signaling pathways[28].

Beyond clinical observations, metabolomic profiling studies have revealed that sorafenib induces profound reprogramming of tumor cell metabolism. One study demonstrated that plasma D-lactate accumulation correlated with favorable therapeutic outcomes, suggesting its role as a potential biomarker for a positive treatment response, whereas elevated glycerol levels were linked to the development of treatment resistance[29]. Tumor cells that efficiently eliminate glycolytic byproducts through D-lactate production tend to retain sensitivity to sorafenib. By contrast, the accumulation of glycerol indicates metabolic adaptation by the cancer cells, which reduces the drug’s efficacy and leads to resistance. However, a key limitation shared across all these indicators is their post hoc availability, as they become measurable only after therapy has been initiated and cannot be used to predict treatment response before therapy initiation. Consequently, efforts have shifted towards developing predictive models to guide treatment selection before sorafenib initiation.

Several established risk models exist specifically for patients receiving sorafenib, with the PROSASH and PROSASH-II models being the most extensively validated. These models incorporate a range of objective clinical variables, including serum albumin and bilirubin, baseline AFP levels, the presence of macrovascular invasion or extrahepatic spread, and the size of the largest intrahepatic tumor lesion. Of note, PROSASH-II demonstrated superior discriminatory performance compared to older models such as the Hepatoma Arterial-embolization Prognostic score and its variant Sorafenib Advanced HCC Prognosis in both internal and external validation studies[30]. However, the performance of these models remains suboptimal, particularly in identifying the minority of patients who may achieve substantial or complete radiological responses.

One of the prognostic markers being researched is the neutrophil-lymphocyte ratio, which serves as a marker of systemic inflammation, and if elevated, indicates worse outcomes[31]. C-reactive protein, another marker of systemic inflammation, not only promotes tumorigenesis and cancer progression but also influences the tumor microenvironment by promoting tumor growth and inhibiting adaptive immunity. C-reactive protein has been included with AFP in a new, although unvalidated, scoring system aimed at predicting outcomes for patients with HCC receiving programmed death-ligand 1 inhibitor treatments. However, this score’s applicability is limited because it produces similar predictive results for patients treated with sorafenib, failing to distinguish which treatment might be more effective for a given individual[32]. Inflammation represents a crucial component of both tumor biology and treatment response, influencing not only disease progression but also the efficacy of systemic therapies such as sorafenib and immunotherapy. Genetic determinants can be additional modulators of sorafenib efficacy. A recent study identified inherited variants in interleukin 23 receptor (IL-23R) and autophagy-related protein 10 (ATG-10) as significant predictors of therapeutic outcomes, with the IL-23R TT and ATG-10 TT genotypes associated with markedly higher response rates and longer survival[33]. These polymorphisms likely reflect interindividual differences in immune regulation and autophagic activity. Specifically, IL-23R variants likely change how T helper 17-driven inflammatory signals are modulated, whereas ATG-10 variants influence autophagy efficiency under sorafenib-induced metabolic stress.

These prognostic models show promise, but are not yet reliable tools for treatment selection because they have not been sufficiently validated across heterogeneous populations or confirmed through prospective studies. In this context, our findings provide additional observational insights. Two of the three patients (patients 2 and 3) experienced dermatologic toxicity and had non-viral HCC etiology. These specific characteristics are generally associated with improved outcomes under TKI therapy. Patients 1 and 3 experienced diarrheas, which is considered a potential on-treatment indicator of a positive response, whereas patient 2 did not. Conversely, patient 1 exhibited viral etiology and no dermatologic toxicity, demonstrating that these correlations are not definitive. Given the very small cohort, these observations should be interpreted cautiously and regarded as hypothesis-generating rather than confirmatory. Importantly, all 3 patients, despite presenting with unfavorable prognostic factors like portal vein thrombosis and extrahepatic metastases, achieved CRs. This finding further highlights the limitations of current prognostic models in accurately identifying exceptional responders to sorafenib.

Another important question is the optimal duration of sorafenib treatment following a CR. In this series, patients received sorafenib for 32, 29, and 40 months, respectively. In similar published case series, the median treatment duration was about 40 months[10]. Some reports suggest that discontinuing sorafenib after achieving a CR does not negatively impact OS or RR, while potentially improving quality of life[34]. However, due to the limited number of patients achieving a CR, current evidence is insufficient to support a clear recommendation regarding treatment discontinuation.

CONCLUSION

Our patients responded remarkably well to sorafenib treatment. However, if they developed HCC today, they would probably receive a different first-line therapy, and we can only speculate about the outcomes of that alternative treatment. Therefore, future research is necessary to better understand personalized therapy options for HCC, considering variations in etiology and molecular characteristics.