Multiphase convolutional dense network for the classification of focal liver lesions on dynamic contrast-enhanced computed tomography.

BACKGROUND: The accurate classification of focal liver lesions (FLLs) is essential to properly guide treatment options and predict prognosis. Dynamic contrast-enhanced computed tomography (DCE-CT) is still the cornerstone in the exact classification of FLLs due to its noninvasive nature, high scanning speed, and high-density resolution. Since their recent development, convolutional neural network-based deep learning techniques has been recognized to have high potential for image recognition tasks. AIM: To develop and evaluate an automated multiphase convolutional dense network (MP-CDN) to classify FLLs on multiphase CT. METHODS: A total of 517 FLLs scanned on a 320-detector CT scanner using a four-phase DCE-CT imaging protocol (including precontrast phase, arterial phase, portal venous phase, and delayed phase) from 2012 to 2017 were retrospectively enrolled. FLLs were classified into four categories: Category A, hepatocellular carcinoma (HCC); category B, liver metastases; category C, benign non-inflammatory FLLs including hemangiomas, focal nodular hyperplasias and adenomas; and category D, hepatic abscesses. Each category was split into a training set and test set in an approximate 8:2 ratio. An MP-CDN classifier with a sequential input of the four-phase CT images was developed to automatically classify FLLs. The classification performance of the model was evaluated on the test set; the accuracy and specificity were calculated from the confusion matrix, and the area under the receiver operating characteristic curve (AUC) was calculated from the SoftMax probability outputted from the last layer of the MP-CDN. RESULTS: A total of 410 FLLs were used for training and 107 FLLs were used for testing. The mean classification accuracy of the test set was 81.3% (87/107). The accuracy/specificity of distinguishing each category from the others were 0.916/0.964, 0.925/0.905, 0.860/0.918, and 0.925/0.963 for HCC, metastases, benign non-inflammatory FLLs, and abscesses on the test set, respectively. The AUC (95% confidence interval) for differentiating each category from the others was 0.92 (0.837-0.992), 0.99 (0.967-1.00), 0.88 (0.795-0.955) and 0.96 (0.914-0.996) for HCC, metastases, benign non-inflammatory FLLs, and abscesses on the test set, respectively. CONCLUSION: MP-CDN accurately classified FLLs detected on four-phase CT as HCC, metastases, benign non-inflammatory FLLs and hepatic abscesses and may assist radiologists in identifying the different types of FLLs.

The Predictive Value of Albumin-to-Alkaline Phosphatase Ratio for Overall Survival of Hepatocellular Carcinoma Patients Treated with Trans-Catheter Arterial Chemoembolization Therapy.

Background: We have previously reported the prognostic value of the albumin-to-alkaline phosphatase ratio (AAPR) for advanced hepatocellular carcinoma (HCC) patients who are not receiving any standard anticancer therapy. However, the prognostic value of the AAPR for HCC patients treated with trans-catheter arterial chemoembolization therapy (TACE) was not investigated. Methods: We retrospectively analysed 372 HCC patients treated with TACE (the training cohort) and applied receiver operating characteristic curves (ROC curves) to identify the best cut-off value for the AAPR in this cohort. Then, univariate analyses by the Kaplan-Meier method and multivariate analysis by a Cox proportional hazards regression model were conducted. Both comparisons of the ROC curves and the likelihood ratio test (LRT) were employed to evaluate the abilities of different factors in predicting the survival of patients in this cohort. Finally, the prognostic value of the AAPR was validated in two cohorts: one included 202 HCC patients treated with supportive care (validation cohort I), and the other included 82 HCC patients treated with TACE (validation cohort II). Results: We identified 0.439 as the best cut-off value of the AAPR by ROC curve analysis. An AAPR > 0.439 was significantly correlated with a lower frequency of Child-Pugh grade B, portal vein tumour thrombus (PVTT), T3-4 and lymph node metastasis (P < 0.05). The median overall survival (OS) of the patients with an AAPR > 0.439 was significantly longer than that of those with an AAPR ≤ 0.439 (58.4 m vs 17.8 m, respectively, P < 0.001). The AAPR was identified as an independent prognostic factor after univariate and multivariate analyses (HR = 0.636, P = 0.003). The independent prognostic value of the AAPR was also confirmed in validation cohorts I and II. Additionally, we substituted the AAPR for the Child-Pugh grade in the CLIP system and integrated the AAPR into the TNM system. We found that the area under the curve (AUC) of the AAPR-CLIP system was significantly larger than that of the CLIP and the TNM when predicting 3-month, 6-month, 1-year and 2-year survival (P < 0.05). There was no significant difference between the AUCs for the AAPR-CLIP and the AAPR-TNM. The LRT suggested that both AAPR-CLIP and AAPR-TNM had significantly larger χ2 values and smaller AIC values than that of their corresponding primary system (P < 0.05). Conclusions: The AAPR was an independent prognostic index for the HCC patients treated with TACE. Both AAPR-CLIP and AAPR-TNM outperformed their corresponding primary system in predicting OS in the current study.

Modified CLIP score with the albumin-bilirubin grade retains prognostic value in HBV-related hepatocellular carcinoma patients treated with trans-catheter arterial chemoembolization therapy.

Background: The Cancer of the Liver Italian Program (CLIP) score is commonly used for prognosis prediction of hepatocellular carcinoma (HCC). The CLIP includes the Child-Pugh grade, which is relatively subjective, for hepatic encephalopathy assessment. A newly developed scoring system called albumin-bilirubin grade (ALBI grade), consists of albumin and bilirubin to assess liver function reserve objectively. Here, we substituted the ALBI grade for the Child-Pugh grade to establish the ALBI-CLIP scoring system and validated its prognostic value in hepatitis B virus (HBV)-related HCC patients treated with trans-catheter arterial chemoembolization (TACE) therapy. Methods: We retrospectively analyzed HBV-related HCC patients who received TACE therapy. Baseline characteristics were collected and evaluated to classify patients according to ALBI-CLIP, CLIP and TNM systems. Univariate analyses using the Kaplan-Meier method and the log-rank test, as well as multivariate analysis using the Cox proportional hazards regression model, were conducted to detect independent prognostic factors for overall survival. Receiver operating characteristic (ROC) curves and a likelihood ratio test (LRT) were both utilized to compare the values of ALBI-CLIP, CLIP and TNM staging systems in predicting survival. Results: With a total of 389 patients included in the current study, 301 (77.4%) and 88 (22.6%) were classified as Child-Pugh grade A and B, respectively. However, 152 (39.1%), 227 (58.4%) and 10 (2.5%) patients were correspondingly classified into ALBI grade 1, 2 and 3. The areas under the curves of ALBI-CLIP, CLIP and TNM systems were 0.804, 0.778 and 0.734, respectively, for predicting 3-month survival; 0.796, 0.778 and 0.733, respectively, for 6-month survival; 0.697, 0.687 and 0.644, respectively, for 1-year survival; and 0.618, 0.612 and 0.569, respectively, for 2-year survival. The LRT indicated that the ALBI-CLIP and the CLIP had similar values of χ2 and Akaike information criterion (AIC) while the TNM system had the smallest χ2 value (χ2 = 12.1, 11.9, 10.5; AIC = 2620.2, 2620.5, 2621.1 for ALBI-CLIP, CLIP and TNM, respectively). Conclusions: In conclusion, our present study suggested that the ALBI-CLIP scoring system retained the prognostic value of the CLIP in HBV-related HCC treated with TACE therapy.

[Low dose contrast-enhanced 4D DSA with the 320-detector row computed tomography scanner for post-operative evaluation of the hepatic artery complications after liver transplantation].

OBJECTIVE: To evaluate the value of low-dose contrast-enhanced 4D DSA acquired from 320-detector row CT on the diagnosis of the hepatic artery complications after liver transplantation. METHODS: 320-detector row CT were performed in fifth patients with liver dynamic enhanced using volume CT body-perfusion protocol, hepatic artery phase images obtained after intravenous injection of 50 ml of contrast at a rate of 6 ml/s were acquired based on 4D DSA. The time-density curve (TDC) of the hepatic artery was delineated. The time to peak, peak contrast enhancement were recorded. The data of artery phase were sent to workstation to reconstruct the hepatic artery with VR, MIP and MPR. RESULTS: The delay mean time and the peak CT value of hepatic artery was 20 (10.0 - 24.2) s and 373 (310 - 440) HU; 4D DSA revealed hepatic artery pseudo-aneurysm (n = 2), and hepatic artery mild stenosis (n = 3), moderate stenosis (n = 4), severe stenosis (n = 1) and occlusion (n = 1). Compensatory circulation was seen in 2 cases. 4 patients were appeared segmental stenosis and hepatoportal arteriovenous fistulas (HPAVF). Hepatic arterial branch are decreased and opened in 9 cases and 4 cases. 4D DSA also displayed other signs including hepatic hemangioma and the splenic artery aneurysms (SAA) and ectasi (n = 1), donor-recipient hepatic artery mismatch (n = 3). CONCLUSION: 4D DSA provided a safe, noninvasive and accurate method for guiding the diagnosis of hepatic artery complication and play an effective role after liver transplantation.