Convolutional neural network for classifying primary liver cancer based on triple-phase CT and tumor marker information: a pilot study

14106 Background: Due to extent of tumour and underlying liver disease primary liver cancer, hepatocellular(HCC) as well as cholangiocellular(CCC), has extremely poor prognosis. The tyrosinekinase and mTOR inhibitor sirolimus, is registered for preventing allograft rejection. It retard experimental liver tumours in rats, and has antiangiogenic effects as well. Methods: Patients with a nonoperable HCC or CCC, with a Karnofsky >70, and the diagnosis confirmed by biopsy or alfa-fetoprotein (>700ug/ml) were eligible. Ethic committee approved the study and informed consent was signed. The daily Sirolimus dose was adjusted to a 6–10 μg/L through-concentration.Twenty patients were to be enrolled in a nonrandomized study. The primary endpoint was to detect a radiological tumor response rate, defined by RECIST criteria, and secondary endpoint was to study the safety profile. CT or MRI were repeated every third month. The study was approved by the Etic Committee of Göteborg University. Results: Eleven patients with HCC and nine with CCC were included. Overall respone rate was 5%, which ocurred in a patient patient with HCC. The PR lasted for 15 months (RR 9% for HCC). Four patients with HCC had SD for median 7 months (range 5–18). Of the patients with CCC four had SD. Median duration of SD was 4 months (range 1 - 7). Median survival for patients with HCC was 7 months (range 2–20) and CCC 4 months (range 2–24) Discussion: There is low evidence supporting the use of chemotherapy for primary liver cancer. For HCC, one PR and 4 SD in this pilot study indicate that sirolimus might be of interest to study in a phase II trial, since the toxicity of sirolimus was low. In the european SILVER study,the impact of sirolimus for patients transplanted for HCC will be clarified. Experimentally, the antiangiogenetic effects of sirolimus has been boosted by combining with 5-FU and gemcitabine. This might translate to impoved response rate also for liver cancer. Conclusions: This pilot study suggests that sirolimus have a tumour effect in patients with HCC. No significant financial relationships to disclose.

Download Full-text

Automatic detection of arterial occlusion on digital subtraction angiography: pilot study with convolutional neural network

10.26226/morressier.5e8335ba7cb08a046ef7c6fc ◽

2020 ◽

Author(s):

Yannan Yu

Keyword(s):

Neural Network ◽

Pilot Study ◽

Convolutional Neural Network ◽

Digital Subtraction Angiography ◽

Arterial Occlusion ◽

Automatic Detection ◽

Digital Subtraction

Download Full-text

Automatic Segmentation Algorithm of Magnetic Resonance Image in Diagnosis of Liver Cancer Patients under Deep Convolutional Neural Network

Scientific Programming ◽

10.1155/2021/4614234 ◽

2021 ◽

Vol 2021 ◽

pp. 1-13

Author(s):

Jinling Zhang ◽

Jun Yang ◽

Min Zhao

Keyword(s):

Neural Network ◽

Image Segmentation ◽

Magnetic Resonance ◽

Liver Cancer ◽

Convolutional Neural Network ◽

Cancer Patients ◽

Segmentation Algorithm ◽

Image Segmentation Algorithm ◽

Single Sequence ◽

Mri Image

To study the influence of different sequences of magnetic resonance imaging (MRI) images on the segmentation of hepatocellular carcinoma (HCC) lesions, the U-Net was improved. Moreover, deep fusion network (DFN), data enhancement strategy, and random data (RD) strategy were introduced, and a multisequence MRI image segmentation algorithm based on DFN was proposed. The segmentation experiments of single-sequence MRI image and multisequence MRI image were designed, and the segmentation result of single-sequence MRI image was compared with those of convolutional neural network (FCN) algorithm. In addition, RD experiment and single-input experiment were also designed. It was found that the sensitivity (0.595 ± 0.145) and DSC (0.587 ± 0.113) obtained by improved U-Net were significantly higher than the sensitivity (0.405 ± 0.098) and DSC (0.468 ± 0.115, P < 0.05 ) obtained by U-Net. The sensitivity of multisequence MRI image segmentation algorithm based on DFN (0.779 ± 0.015) was significantly higher than that of FCN algorithm (0.604 ± 0.056, P < 0.05 ). The multisequence MRI image segmentation algorithm based on the DFN had higher indicators for liver cancer lesions than those of the improved U-Net. When RD was added, it not only increased the DSC of the single-sequence network enhanced by the hepatocyte-specific magnetic resonance contrast agent (Gd-EOB-DTPA) by 1% but also increased the DSC of the multisequence MRI image segmentation algorithm based on DFN by 7.6%. In short, the improved U-Net can significantly improve the recognition rate of small lesions in liver cancer patients. The addition of RD strategy improved the segmentation indicators of liver cancer lesions of the DFN and can fuse image features of multiple sequences, thereby improving the accuracy of lesion segmentation.

Download Full-text