Metastasis in Patients With Hepatocellular Carcinoma

Prevalence, Determinants, Prognostic Impact and Ability to Improve the Barcelona Clinic Liver Cancer System

Chia-Yang Hsu; Po-Hong Liu; Shu-Yein Ho; Yi-Hsiang Huang; Yun-Hsuan Lee; Rheun-Chuan Lee; Teddy S. Nagaria; Ming-Chih Hou; Teh-Ia Huo


Liver International. 2018;38(10):1803-1811. 

In This Article


Three major parameters, PS, tumour burden and the severity of cirrhosis, are used in the BCLC system to classify patients into stage 0, A, B, C or D. Patients with vascular invasion or metastasis are controversially designated as BCLC stage C, and there is no clear criteria regarding tumour burden for stage D patients. Unlike other human malignancies, HCC patients with distant metastasis are not considered terminal and still recommended to have anticancer treatment per current guidelines.[1,11] A unique characteristic of HCC is that a variable proportion of BCLC stage D patients could have significantly smaller tumour burden compared to stage C patients. This design of BCLC would make the predicted prognosis between stage C and D patients overlap substantially.[16,17] Recently, the co-existence of vascular invasion and metastasis was found to associate with a significantly worse prognosis of BCLC stage C patients, and another independent study showed that tumour burden is a strong prognostic factor for stage D patients.[3,9] These results prompted us to perform an in-depth investigation regarding the associated elements and prognostic effect of metastasis and also the possibility of integrating tumour behaviour into the BCLC system to further improve its predictive accuracy. Importantly, the current study confirms the synergistic effects of vascular invasion and metastasis on survival prediction, and the prognostic ability of BCLC system can be further improved by merely re-allocating patients with the co-existence of vascular invasion and metastasis from stage C to D.

Consistent with previous studies, patients with metastasis had more extensive tumour burden as determined by multiple tumours, larger tumour diameter, higher serum AFP level and larger total tumour volume.[18,19] Patients with metastasis were generally younger in our cohort; this might be related to the fact that patients with metastasis had a higher prevalence rate of hepatitis B, which is known to link with earlier development of HCC compared to those with hepatitis C.[20] In this large cohort study, we also notice that patients with metastasis had more severe cirrhosis, which could result from the fact that metastasis-related tumour burden may exacerbate liver dysfunction by decreasing the proportion of normally functioning parenchyma. Notably, the most common sites of metastasis of our patients were lungs and lymph nodes, followed by bones, adrenal glands and omentum/peritoneum, which are coherent to the results from another study.[9] Given so, patient survival was not influenced by the site of metastatic organs in the current and previous study.[21]

In the present study, there were 122 patients with metastasis receiving active intrahepatic anticancer treatments including surgical resection, ablation and TACE with an attempt to improve their life quality and prolong survival; this aggressive approach is supported by some reports from Asian countries. Further well-designed clinical trials are warranted to investigate the effect and select the candidates for intrahepatic therapy in patients with metastasis.[5,21,22]

A substantially high proportion of HCC patients who are diagnosed with extrahepatic involvement may clinically present with a single tumour, small nodule, absence of vascular invasion or low serum AFP level.[5,21] To further survey the indicator of metastasis, we stratified patients by TTV and investigated the associations between AFP, vascular invasion, tumour size and number with metastasis.[13,23,24] Interestingly, vascular invasion was highly associated with metastasis, both factors reflecting the progression of HCC, independent of the status of TTV.[9,25] Our results suggest that patients with smaller tumour burden but high serum AFP level and multiple nodules should be actively screened for clinical metastasis.

α-foetoprotein is a bio-hallmark of HCC and has been linked with cell proliferation, angiogenesis and increased resistance of cells against tumour necrosis factor-related apoptosis.[26,27] A higher AFP level was also reported to associate with a more aggressive HCC phenotype.[28,29] Importantly, it remains significantly connected to metastasis among patients with relatively small tumour burden. This finding indicates a similar concept as demonstrated in our previous study which showed high AFP to total tumour volume ratio may predict early tumour recurrence after surgical resection.[30] Alternatively, the size of most massive tumour did not result in a higher risk of metastasis after stratified by TTV; again, this finding supports that TTV may be a more feasible parameter to evaluate tumour burden compared to the size of primary tumour in HCC patients.[23,24]

Vascular invasion and metastasis are well-known predictors of poor survival in HCC patients.[3] In our study, patients with a combination of both factors had a significantly worse outcome compared to patients with vascular invasion alone or metastasis alone. This finding is consistent with a recently published study from the USA.[9] Patients with coexistence of vascular invasion and metastasis often have larger tumour burden, more severe cirrhosis and a destitute general condition. However, the currently recommended anticancer treatment for this group of patients are still highly debatable. Based on their poor prognosis and the lack of ideal treatment, we propose a modified BCLC system which moves patients with a combination of vascular invasion and metastasis from BCLC stage C to stage D without making alterations of cirrhosis and PS. As a result, 181 (5%) patients were re-classified. Based on the AIC analysis, this simple modification helps to improve the prognostic power of BCLC system without additional laboratory work or imaging studies.

This study has a few limitations. Firstly, our results are from a cohort where hepatitis B is the leading cause of chronic liver disease. External validation is required before the results can be widely interpreted in countries with high prevalence of alcoholic liver disease, non-alcoholic fatty liver disease or hepatitis C. Secondly, specific anticancer treatments were not included in survival analysis. Of note, re-allocating patients from BCLC stage C to stage D does not exclude these patients' eligibility to receive other anticancer treatments. Further studies directed towards the association between survival and treatments are needed to clarify the clinical benefits. Thirdly, the presences of vascular invasion and metastasis were determined at the time of diagnosis to focus on their prognostic impact on survival and staging system; metastatic lesions detected during follow-up were not considered in this study. Last, only patients with symptoms or clinical suspicions of metastasis underwent detailed surveys such as MRI, bone scan and biopsy. It is possible that some patients with brain, bone or limb metastasis might not be timely revealed at baseline.

In conclusion, metastasis in HCC is not uncommon. We identify indicators of metastasis in HCC patients with small to large tumour burden. Our study also confirms the unique behaviour of vascular invasion and metastasis, and their combination may predict an even worse outcome compared to patients with vascular invasion alone or metastasis alone. Re-allocating BCLC stage C patients with the coexistence of vascular invasion and metastasis to stage D could further improve the predictive accuracy of BCLC system at the time of diagnosis.