Although, an ultrasound (US) examination has been widely applied for the diagnosis of a variety of diseases, the quality of examination is highly operator-dependent; missing a diagnosis and misdiagnosis of lesions can lead to a serious consequence. Artificial intelligence (AI)-aided US diagnosis is one valuable means of preventing human error. We have been constructing a nationwide US image database of liver and breast tumors, and cardiac diseases. Taking advantage of this large database, we have created the prototype of a real-time detection system for liver and breast tumors using AI. Significantly, we are focusing on detection of diseases that are the leading cause of death, such as malignant tumors and heart disease. Therefore, development of AI-assisted technologies in these fields could carry a potential enough to reduce cancer and heart disease mortality in the aging society.

Recently developed deep learning-based method lead a paradigm shift of diagnostic radiology. CT and MR images, which produced by image-reconstruction, go well with artificial intelligent so that many kinds of ideas for the application of AI in diagnostic radiology have been tested and published. Some of them have been already in practical use. Here, I picked up 3 areas of research in diagnostic radiology of the liver with AI, which are 1) image reconstruction/processing, 2) staging liver fibrosis, and 3) supporting image interpretation. Recent publication on these areas are reviewed in this manuscript.

The aim of this study is to clarify the significance of estimating the volume of new lesions emerging following the treatment of hepatocellular carcinoma (HCC) with TACE. Between April and September 2010, 50 patients underwent TACE for HCC. Response to TACE was assessed using post-treatment CT and/or MRI. Tumor markers were classified as effective (E) or non-effective (NE). In method A, patients with new lesions following TACE were classified as NE. In method B, patients in whom the volume of new lesions and remaining viable lesions was greater than 50% of the pretreatment volume were classified as NE. The cumulative survival rate of the E group was significantly higher than that of the NE group in both methods. However, the median survival time of ANE and BNE groups were 21.8 and 12.6 months, respectively. The median survival time of seven patients of ANE/BE was 29.0 months. Method B (Estimating the volume of new lesions after TACE) is useful in clinical settings, because it may more accurately reflect the prognosis than method A.

We performed a survey of 109 patients with Wilson disease (male 57, female 52; mean age, 34 years [range, 8-67]; mean age of diagnosis, 12 years [range, 2-50]). Among these, 73 and 59 patients were diagnosed and currently seen by pediatricians, respectively, and 10 and 29 were diagnosed and currently seen by gastroenterologists, respectively. These numbers suggest that the number of patients currently being treated by pediatricians has decreased, while those currently being treated by gastroenterologists has increased, as compared with a diagnostician. Thirty-four percent (28/82) of patients aged ≥15 years have transitioned from pediatric to adult care; 10 were dissatisfied with their current physicians for reasons such as not having sufficient knowledge about Wilson disease. Sixty-six percent (54/82) of patients were still treated by their pediatricians (mean age, 33 years). These results suggest that problems at transition include independence support to the patients, provision of disease information, and cooperation with Wilson disease-related organizations.

A 55-year-old woman, with a history of convulsions, was diagnosed with advanced hepatocellular carcinoma. She was treated with lenvatinib at a dose of 8 mg daily, as she was refractory to transcatheter arterial chemoembolization. Eight weeks after starting her lenvatinib therapy, she developed fatigue equivalent to CTCAE grade 2, with an elevation in her plasma acyl carnitine (AC) to free carnitine (FC) ratio. She was then prescribed oral levocarnitine supplementation (1500 mg per day). Her fatigue improved and the initial dose of the lenvatinib was administrated. Six months later, valproic acid was initiated due to epilepsy. One month later, her fatigue was exacerbated by a carnitine deficiency, with a severe elevation of the AC/FC ratio. Following further oral levocarnitine supplementation (3000 mg per day), her fatigue improved, with no reduction in lenvatinib dose. In conclusion, lenvatinib therapy may promote drug-induced carnitine insufficiency, while levocarnitine supplementation may be useful in improving fatigue symptoms.

A 75-year-old woman, with cirrhosis due to non-alcoholic steatohepatitis, was admitted to our hospital with a chief complaint of tarry stools. Emergent upper gastrointestinal endoscopy revealed widespread varices observed from the third to fourth portions of the duodenum. We performed balloon-occluded retrograde transvenous obliteration (B-RTO); however, the embolization was difficult due to the outflow of sclerosing agent to the portal vein owing to intestinal peristalsis. As an alternative treatment, we performed percutaneous transhepatic obliteration. The sclerosing agent accumulated in the varices by approaching from the portal side, and the varices were treated completely. No cases have been reported wherein B-RTO failed because of intestinal peristalsis. Duodenal varices, which are in wide contact with the intestines, may be affected by intestinal peristalsis during B-RTO. For the treatment of duodenal varices, it is necessary to plan a treatment strategy which considers the intestinal tract in close proximity to the varices, in addition to hemodynamics.

Radiofrequency ablation (RFA) therapy for liver cancer have been used in the worldwide. On the other hand, first-generation microwave ablation (MWA) therapy was mainly used for small liver cancer prior RFA goes on the market. In this study, we evaluated the early experience using next-generation microwave ablation therapy for liver cancer. Eighty patients with 105 tumors were enrolled. The median (interquartile range) tumor diameter was 8 (11-24) mm. The maximum ablation power (60/75/100W) were 4/29/24 tumors, ablation time was 4.0 (3.0-5.0) minutes, and insertion time (one/two or more) were 91/14, respectively. There was no severe adverse event associated with this therapy. In conclusion, the MWA therapy has safety and utility for the treatment of liver cancer.