Hepatic artery infusion pump for nasopharyngeal carcinoma with liver metastasis

Nasopharyngeal carcinoma (NPC) is a common head and neck malignancy, with a high prevalence in South-East Asia, particularly in southern China [1]. It is estimated that in China, NPC accounts for about 1.34% of all new cancer cases per year and 1.03% of all cancer deaths per year, with a prevalence of 3.16/100,000 and 1.53/100,00 respectively [2]. Unlike other head and neck cancers, NPC presents with high incidence of locoregional recurrence or distant metastases [3], which is considered as the predominant cause of mortality [4]. The most common metastatic sites are bone, liver and lung [5]. Liver metastases of NPC often present as multifocal nodules, and unfortunately has worse prognosis compared with metastasis to bone or to lung, with median overall survival of 3–5 months [6, 7]. According to NCCN guidelines, treatment options for metastatic NPC include Clinical trials (preferred), Platinum-based combination chemotherapy or concurrent chemo/radiotherapy [8], but the clinical outcome is still limited. Thus how to improve the prognosis of NPC patients with liver metastasis remains a big challenge.

It is reported that perioperative hepatic arterial infusion pump chemotherapy is associated with longer survival after resection of colorectal liver metastases [9]. Previous work by our team showed that initial hepatic artery infusion (HAI) and systemic chemotherapy is helpful for colorectal cancer patients with liver metastases to obtain a high resection rate [10]. Hepatic-directed therapy is recommended for some liver-predominant disease in patients with neuroendocrine tumors of the pancreas [11]. Here, we use this HAI regime in NPC patients with liver-predominant metastases, aiming to verify its effectiveness and safety.

Several studies have been shown that HAI is effective in improving hepatic response rates in colorectal cancer liver metastasis. In the present study, we treated 16 consecutive NPC patients with predominant liver metastasis using HAI of Gemcitabine and FUDR. Our data show that the ORR was 87.5% and the mOS was 30 months. The majority (223/266, 83.8%) of toxicities and adverse events was judged on Grades1/2.

About one-third of patients with NPC suffer of distant metastases and resultant treatment failure [15]. Liver metastasis is the second common metastasis site, which accounts for 30% of such cases, following bone metastasis (70%) and followed by lung metastasis (18%) [16], but has the worst prognosis (with reported mOS of 3–5 months) [6, 7]. Thus it is of significance to explore new therapeutic strategy.

HAI has been proven both by our previous work and other authors to be helpful in some patients [9, 10]. It is also accepted as first-line treatment for unresectable colorectal cancer liver metastases [17]. Here we administered HAI for patients with NPC liver metastasis. Given that in our case series, the majority of patients suffered of unresectable dominant liver involvement (defined as liver involvement as the dominant site of metastasis, hepatic metastases constitute ≥ 50% of all tumor burden [18]): 13 patients (81.2%) had hepatic involvement > 25%, 13 patients (81.2%) had bilobar involvement, 14 patients (87.5%) had number of intrahepatic lesions > 5 (among whom 9 patients with number of intrahepatic lesions > 9), the ORR and DCR of intrahepatic lesions as well as survival analysis is encouraging.

Nowadays treatment options for metastatic NPC include Clinical trials, Platinum-based combination chemotherapy or concurrent chemo/radiotherapy [8], but prognosis is still poor. Some investigators have reported systemic monochemotherapy for metastatic NPC with ORR range from 28 to 48% [19-21], and some polychemotherapy protocols with ORR range from 42.7 to 73% [20, 22, 23]. It seems that patients in current study probably have survival benefit. Theoretically, controlling intrahepatic lesions decreases the total tumor burden, then yields survival benefit. It is reported that elimination liver metastasis of nasopharyngeal carcinoma might improve overall survival [24], which might have something in common with current study as respect to the underlying mechanism of survival benefit.

In current study, chemotherapy agents are administered through intra-arterial approach, directly targeting intrahepatic lesions. Especially for FUDR, the low-dose and long-period infusion may provide higher intact drug concentration in liver and minimal systemic toxicity [25].

Besides HAI, patients in current study also underwent comprehensive therapy to primary lesion. This combined treatment modality supplies control power both of the primary tumor and of the liver metastases, which may contribute to survival benefit. But there was still 3 patients (18.8%) who had extra-hepatic progression during the trial treatment, indicating the necessity to explore more effective treatment strategies.

Several side effects and adverse events could be observed or detected in current study, including (i) hematological toxicities such as leukopenia, neutropenia, anemia, thrombocytopenia; (ii) AST/ALT elevation; (iii) hyperbilirubinemia; (iv) gastrointestinal toxicities such as anorexia, nausea, vomiting, diarrhoea; (v) fatigue; (vi) abdominal pain. For all the toxicities mentioned above, the proportion for grades 1/2 toxicities is obviously higher than that for grades 3/4 toxicities. Some toxicities resulted from the same underlying mechanism as systemic chemotherapy. AST/ALT elevation appeared in 48 cycles (66.7%), but the majority of them (n = 36, 50%) was on grades 1/2, while the minority (n = 12, 16.7%) was on grades 3/4. This might reveal the safety of this strategy for liver function. Some authors reported biliary toxicity of HAI with FUDR [25]. In current study, reversible Grade1/2 Hyperbilirubinemia is observes in 9(12.5%) cycles, with no evidence of long-term biliary damage, probably because of adding DXM into FUDR solution. Technically, the gastrointestinal toxicities might derive from extra-hepatic infusion of FUDR to stomach or duodenum, so the main branch vessels from the associated arteries (e.g. Hepatic artery or GDA) were routinely embolized in operation, but this still could not be completely avoided. Fortunately, most patients with grades 3/4 toxicities continued HAI therapy, even though some needed dose reduction or gap prolongation. If indwelled directly in hepatic artery, the dislodgement of the infusion catheter tip may result in vessel damage, occlusion or aneurysm, finally lead to failure of the whole infusion system. Here we fixed the side-hole infusion catheter in GDA or in the relatively smaller branch of hepatic artery, positioned the side-hole in common hepatic artery, thus effectively avoided the potential risk and guaranteed the unidirectional infusion to liver. Other HAI related complications, such as bleeding, thrombosis or infection are not observed in these patients. In our opinion, these data, to some extent, could prove the safety of our HAI strategy.

This study has several limitations. We performed HAI for NPC patients with liver-predominant metastasis, whose intrahepatic disease status is more advanced compared to those who do not undergo HAI. As a retrospective study with small sample size, it lacks a control group with matched status of NPC liver metastases. The improvement of DCR of intra-hepatic lesions and survival benefit needs to be further investigated and confirmed by multi-center randomized study.

In conclusion, we introduced a new treatment strategy for NPC liver metastases, which may improve DCR of intra-hepatic lesions and prolong mOS. It can be considered as an option for suitable patients.