We aimed to develop the deep learning (DL) predictive model for postoperative early recurrence (within 2 years) of hepatocellular carcinoma (HCC) based on contrast-enhanced computed tomography (CECT) imaging. This study included 543 patients who underwent initial hepatectomy for HCC and were randomly classified into training, validation, and test datasets at a ratio of 8:1:1. Several clinical variables and arterial CECT images were used to create predictive models for early recurrence. Artificial intelligence models were implemented using convolutional neural networks and multilayer perceptron as a classifier. Furthermore, the Youden index was used to discriminate between high- and low-risk groups. The importance values of each explanatory variable for early recurrence were calculated using permutation importance. The DL predictive model for postoperative early recurrence was developed with the area under the curve values of 0.71 (test datasets) and 0.73 (validation datasets). Postoperative early recurrence incidences in the high- and low-risk groups were 73% and 30%, respectively (p = 0.0057). Permutation importance demonstrated that among the explanatory variables, the variable with the highest importance value was CECT imaging analysis. We developed a DL model to predict postoperative early HCC recurrence. DL-based analysis is effective for determining the treatment strategies in patients with HCC.

Focal nodular hyperplasia (FNH) is the second most common benign hepatic tumor and is very rarely complicated by hemorrhage or rupture. Although thought to be extremely rare, there have been several reports of hemorrhage caused by ruptured FNH. Herein, we report the case of a patient with ruptured FNH, who subsequently developed hemorrhage during follow-up. A 32-year-old man was admitted to our department for an asymptomatic hepatic tumor in segments 4 and 5 (S4/5), which measured 8 cm in diameter and observed to project from the liver. Imaging and pathologic examination of a biopsy specimen confirmed the diagnosis of FNH. Three years after the diagnosis, the patient was readmitted to our hospital because of sudden onset of upper abdominal pain. Dynamic abdominal computed tomography revealed ascites around the tumor with high-density areas that were considered to represent hematoma caused by ruptured FNH. Transcatheter arterial embolization (TAE) was performed to stop the hemorrhage. One month after TAE, S4/5 of the liver was resected; macroscopic findings revealed that a large part of the tumor was composed of necrotic tissue and hematoma. Pathological examination using hematoxylin–eosin staining and immunohistochemical examination indicated a final diagnosis of FNH rupture and hemorrhage. Although a well-established diagnosis of FNH usually requires no treatment or surveillance, careful examination remains necessary when the FNH is large and projects from the liver because of the possibility of rupture and hemorrhage.

The efficacy of indocyanine green (ICG) fluorescence imaging for visualizing hepatic tumors in robot-assisted hepatectomy (RAH) should be validated. This study included 30 consecutive patients with 33 collective tumors who underwent RAH. ICG was administered at a dose of 0.5 mg/kg before surgery. ICG fluorescence imaging was performed intraoperatively. In total, 28 patients with a combined total of 31 tumors underwent ICG fluorescence imaging. Further, 26 (84%) tumors were identified on hepatic surfaces prior to hepatic transection. The fluorescence signals of eight tumors were detected on hepatic raw surfaces during parenchymal dissection, thereby enabling surgeons to adjust the transection planes to ensure appropriate surgical margins. One patient with intrahepatic cholangiocarcinoma tested positive for cancer cells at the dissected stump of the bile duct. However, in all patients in whom ICG fluorescence imaging was used, negative surgical margins were achieved at the site of the dissected hepatic parenchyma. On the other hand, one of two patients with ICG contraindications had a positive surgical margin surrounding the dissected hepatic parenchyma. The median operative time and volume of blood loss were 259 (range: 124–594) min and 150 (range: 1–1150) mL, respectively. ICG fluorescence imaging facilitates the easy identification of hepatic tumors, even in RAH. Hence, it can be useful for confirming appropriate surgical margins.