GI involvement in HCC is rare and is sometimes identified as one of the multiple lesions in end-stage HCC with peritoneal dissemination or extrahepatic metastases [2]. Moreover, because the duodenum and stomach are anatomically adjacent to the liver, they represent sites that are particularly susceptible to direct invasion by HCC [5]. Such cases typically involve large tumors with aggressive invasive potential, indicating an advanced stage [6]. In this case, the simultaneous detection of lung metastases and the absence of disseminated lesions outside the duodenum during surgery led to clinical suspicion of hematogenous duodenal metastasis.

The mechanisms of hematogenous metastasis to the duodenum remain unclear. However, several possibilities can be postulated. First, portal hypertension may play a role; irrespective of liver cirrhosis, major hepatic resections are known to result in postoperative splenomegaly [7], suggesting portal venous stasis or retrograde flow. Second, transarterial chemoembolization (TACE) and systemic therapies may promote the growth and dissemination of aggressive tumor clones by upregulating angiogenic factors or selecting for resistant subclones [8]. Third, although no studies have directly investigated this issue, hepatic resection and TACE may promote the development of collateral circulation and a relative increase in arterial inflow to the pancreaticoduodenal region, potentially facilitating hematogenous metastasis to this area [9].

Importantly, the duodenum possesses unique anatomical features, including the confluence of the bile and pancreatic ducts, necessitating a distinct clinical and therapeutic approach compared with that of other GI sites. A PubMed search using “duodenal metastases” and “duodenal involvement” of HCC as keywords yielded 30 publications. These studies, together with our case, involved 61 patients and are summarized in Table 1 [4, 6, 10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37]. The mean patient age was 62.6 ± 8.2 (range, 34 − 83) years, 86.8% were male. The most common route of involvement was direct invasion in 43 patients, followed by lymphatic or hematogenous metastases in 6, hematogenous metastases in 3, peritoneal dissemination in 2, and lymph node metastases in 1. Portal vein thrombosis was observed in 12 patients. At the time of diagnosis, 44 patients (72.1%) had received some form of prior treatment, regardless of whether duodenal involvement was due to direct invasion or metastasis (70.5% vs. 68.6%). A history of immune checkpoint inhibitor therapy was observed only in our case.

Excluding 8 cases with unavailable treatment data, management strategies for duodenal involvement included supportive care in 25 patients (47.2%), curative-intent surgery in 8 (15.1%), gastrojejunostomy in 2 (3.8%), TACE or transarterial embolization in 6 (11.3%), radiotherapy in 2 (3.8%), sorafenib in 1 (1.9%), and intra-arterial infusion chemotherapy in 1 (1.9%). The overall median survival was 3.0 months (range, 0.2–84). The median survival was 2.8 months in the direct invasion group and 4.0 months in the hematogenous or lymph node metastasis group (Mann–Whitney U test, p = 0.89). These outcomes were markedly shorter than the reported overall survival of patients with distant metastasis of HCC (OS, 17.6 months) [38]. In contrast, patients who underwent curative-intent surgery (n = 8) demonstrated substantially longer survival than those managed with non-surgical treatments (n = 9), with median survivals of 29.0 months (range, 8–84) and 4.0 months (range, 1–42.7), respectively. Curative surgical procedures included partial duodenal resection, distal gastrectomy, lymphadenectomy, and hepatopancreatoduodenectomy. Accordingly, curative resection is recommended for patients without other factors precluding resectability and with an acceptable operative risk.

In our case, the patient had multiple intrahepatic metastases and pulmonary metastases in addition to duodenal metastasis, and his overall condition was compromised by malnutrition and obstructive jaundice. Although a palliative gastrojejunostomy was performed to enable oral intake, he developed portal vein thrombosis due to progressive coagulopathy and died in the early postoperative period. Given the patient’s preoperative general condition and anticipated prognosis, a less invasive palliative approach should have been selected.

Endoscopic duodenal stent placement for malignant duodenal obstruction provides a clinical success rate comparable to that of gastrojejunostomy, despite a relatively high rate of re-obstruction [39]. Accordingly, it is considered a suitable option for patients with poor prognosis and compromised performance status. Lee et al. reported 6 cases of HCC with duodenal involvement treated with self-expandable metallic stents [29]. Obstructive jaundice occurred in 1 patient, whereas 5 patients were able to resume at least a soft diet, and stent patency was maintained for a mean of 51 days (range, 10–119).

Multimodal therapy for HCC has expanded with the advent of immune checkpoint inhibitors, resulting in improved survival outcomes [40, 41]. As an increasing number of patients proceed to subsequent lines of therapy, uncommon metastatic patterns, such as those observed in the present case, are likely to be encountered more frequently. Accordingly, further accumulation of evidence regarding the clinical features and treatment outcomes of these cases is warranted.

In conclusion, because the prognosis of unresectable duodenal involvement from HCC is markedly poor, palliative surgical interventions should be undertaken with careful consideration of the patient’s overall condition.