Automated diagnosis and classification of liver cancers using deep learning techniques: a systematic review

Clinical decision making and research in hepatocellular carcinoma: Pivotal role of imaging techniques

This review examines the recent developments in deep learning (DL) techniques applied to multimodal fusion image segmentation for liver cancer. Hepatocellular carcinoma is a highly dangerous malignant tumor that requires accurate image segmentation for effective treatment and disease monitoring. Multimodal image fusion has the potential to offer more comprehensive information and more precise segmentation, and DL techniques have achieved remarkable progress in this domain. This paper starts with an introduction to liver cancer, then explains the preprocessing and fusion methods for multimodal images, then explores the application of DL methods in this area. Various DL architectures such as convolutional neural networks (CNN) and U-Net are discussed and their benefits in multimodal image fusion segmentation. Furthermore, various evaluation metrics and datasets currently used to measure the performance of segmentation models are reviewed. While reviewing the progress, the challenges of current research, such as data imbalance, model generalization, and model interpretability, are emphasized and future research directions are suggested. The application of DL in multimodal image segmentation for liver cancer is transforming the field of medical imaging and is expected to further enhance the accuracy and efficiency of clinical decision making. This review provides useful insights and guidance for medical practitioners.

Surveillance for Hepatocellular Carcinoma in Patients With Cirrhosis

Liver segmentation from computed tomography (CT) images has grown significantly in importance in the field of medical image processing in the last few years. It is the first and most crucial step in any computerized technique for the automatic detection of liver disease, liver volume measurement, and 3D liver volume rendering. The diagnosis and treatment of liver cancer depend heavily on the segmentation of the liver from CT images to get liver volumetric data, but manual segmentation is a strenuous and time-consuming process. The procedure can be accelerated, simplified, and made less error-prone by using deep learning methods. Image segmentation based on deep learning techniques has gained widespread acceptance due to its robustness, efficiency, and it's reproducible nature. Therefore, in this paper, using UNet, MONAI (Medical Open Network for Artificial Intelligence) and PyTorch framework, a deep-learning model to segment the liver from publicly available CT scan dataset was developed. The same ideas that underlie this model for segmenting the liver will allow to create models for segmenting other organs or malignancies using CT data. The goal is to develop a liver segmentation model that can quickly and accurately extract the liver from any given CT image with an accuracy that is on par of manual segmentation performed by a skilled radiologist.

Designing of VehiNet Using Convolutional Neural Networks and Deep Learning Techniques

The relationship between diabetes mellitus and primary liver cancer was investigated in a case-control study conducted in Italy between 1984 and 1996 on 428 cases with incident, histologically confirmed hepatocellular carcinoma, 59 with gallbladder and bile duct cancer, and 1,502 control subjects in the hospital for acute non-neoplastic diseases. Sixty-four cases of hepatocellular carcinoma vs. 87 controls reported a history of diabetes, corresponding to an odds ratio (OR) of 2.3 after allowance for age, sex and area of residence, and of 2.1 [95% confidence interval (CI) = 1.4-3.2] after further allowance for alcohol and tobacco consumption, history of hepatitis and liver cirrhosis, body mass index and history of liver cancer in first-degree relatives. The ORs were similar both for subjects diagnosed with diabetes below age 45, who most likely had insulin-dependent diabetes, and for those diagnosed later, who were likelier to have non-insulin-dependent diabetes. The OR was 2.3 for subjects whose diabetes was diagnosed <5 years before diagnosis of liver cancer, 1.9 for those diagnosed 5-9 years in advance and 2.2 for those diagnosed since 10 years or more. Five cases of gallbladder and bile duct cancer reported a history of diabetes: the corresponding OR was 1.2 (95% CI 0.5-2.9). The OR of hepatocellular carcinoma was 2.4 for males and 2.0 for females, 3.0 for subjects diagnosed with liver cancer under age 60 and 1.8 for those diagnosed at age 60 or over. None of the other covariates considered, including education, history of hepatitis, liver cirrhosis and alcohol drinking showed any meaningful modifying effect or interaction. The potential pathogenic mechanisms include liver alteration-and consequent cell proliferation-in subjects with diabetes. Thus a history of diabetes mellitus could explain about 8% (95% CI 5-11) of cases of liver cancer in this population.

The relationship between diabetes mellitus and primary liver cancer was investigated in a case-control study conducted in Italy between 1984 and 1996 on 428 cases with incident, histologically confirmed hepatocellular carcinoma, 59 with gallbladder and bile duct cancer, and 1,502 control subjects in the hospital for acute non-neoplastic diseases. Sixty-four cases of hepatocellular carcinoma vs. 87 controls reported a history of diabetes, corresponding to an odds ratio (OR) of 2.3 after allowance for age, sex and area of residence, and of 2.1 [95% confidence interval (CI) = 1.4–3.2] after further allowance for alcohol and tobacco consumption, history of hepatitis and liver cirrhosis, body mass index and history of liver cancer in first-degree relatives. The ORs were similar both for subjects diagnosed with diabetes below age 45, who most likely had insulin-dependent diabetes, and for those diagnosed later, who were likelier to have non-insulin-dependent diabetes. The OR was 2.3 for subjects whose diabetes was diagnosed <5 years before diagnosis of liver cancer, 1.9 for those diagnosed 5–9 years in advance and 2.2 for those diagnosed since 10 years or more. Five cases of gallbladder and bile duct cancer reported a history of diabetes: the corresponding OR was 1.2 (95% CI 0.5–2.9). The OR of hepatocellular carcinoma was 2.4 for males and 2.0 for females, 3.0 for subjects diagnosed with liver cancer under age 60 and 1.8 for those diagnosed at age 60 or over. None of the other covariates considered, including education, history of hepatitis, liver cirrhosis and alcohol drinking showed any meaningful modifying effect or interaction. The potential pathogenic mechanisms include liver alteration—and consequent cell proliferation—in subjects with diabetes. Thus a history of diabetes mellitus could explain about 8% (95% CI 5–11) of cases of liver cancer in this population. Int. J. Cancer 73:204–207, 1997. © 1997 Wiley-Liss, Inc.

Hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) are the two most common pathology subtypes of primary liver cancer (PLC). Identifying DNA methylation biomarkers for diagnosis of PLC and further distinguishing HCC from ICC plays a vital role in subsequent treatment options selection. To obtain potential diagnostic DNA methylation sites for PLC, differentially methylated CpG (DMC) sites were first screened by comparing the methylation data between normal liver samples and PLC samples (ICC samples and HCC samples). A random forest algorithm was then used to select specific DMC sites with top Gini value. To avoid overfitting, another cohort was taken as an external validation for evaluating the area under curves (AUCs) of different DMC sites combination. A similar model construction strategy was applied to distinguish HCC from ICC. In addition, we identified DNA Methylation-Driven Genes in HCC and ICC via MethylMix method and performed pathway analysis by utilizing MetaCore. Finally, we not only performed methylator phenotype based on independent prognostic sites but also analyzed the correlations between methylator phenotype and clinical factors in HCC and ICC, respectively. To diagnose PLC, we developed a model based on three PLC-specific methylation sites (cg24035245, cg21072795, and cg00261162), whose sensitivity and specificity achieved 98.8%,94.8% in training set and 97.3%,81% in validation set. Then, to further divide the PLC samples into HCC and ICC, we established another mode through three methylation sites (cg17769836, cg17591574, and cg07823562), HCC accuracy and ICC accuracy achieved 95.8%, 89.8% in the training set and 96.8%,85.4% in the validation set. In HCC, the enrichment pathways were mainly related to protein folding, oxidative stress, and glutathione metabolism. While in ICC, immune response, embryonic hepatocyte maturation were the top pathways. Both in HCC and ICC, methylator phenotype correlated well with overall survival time and clinical factors involved in tumor progression. In summary, our study provides the biomarkers based on methylation sites not only for the diagnosis of PLC but also for distinguishing HCC from ICC.

Aim. In order to determine the expression level of immunohistochemical markers of matrix metalloproteinase-9 (ММР-9) and its tissue inhibitor TIMP-1, and area of immunopositive cells а comprehensive histopathological, histological and immunohistochemical study of liver biopsy specimens was carried out in 94 patients, of which 55 (58.51%) patients had hepatocellular carcinoma and 39 (41.49%) – cholangiocellular carcinoma of liver.Methods and results. In 92.73% of patients with hepatocellular carcinoma and in 89.74% of patients with cholangiocellular carcinoma in malignant cells cytoplasmic expression of MMP-9 was determined. MMP-9 immunopositive cells took 59,33±22,57% of area of the tissue section of hepatocellular carcinoma and 52,71±20,86% of area of the tissue section of cholangiocellular liver cancer. In 80% of hepatocellular carcinoma patients and 74.36% of patients with cholangiocellular carcinoma in tumor cells cytoplasmic expression of TIMP-1 was revealed. TIMP-1 immunopositive cells constituted 21,94±6,27% of area of the tissue section of hepatocellular liver cancer and 33,05±13,85% of area of the tissue section of holangiocellular liver cancer.Conclusion. In both types of liver cancer direct strong relationship was noted between high levels of expression of MMP-9 by the tumor cells and low expression of TIMP-1, which indicates high invasive potential and aggressiveness of hepatocellular carcinoma and cholangiocellular carcinoma.

Recent technological advancements allow deep learning to be employed in practically every aspect of life. Because deep learning techniques are so precise, they can be used in medicine to classify and detect various diseases. The coronavirus (SARSCoV2) epidemic has recently affected global health systems. One of the ways to diagnose SARSCoV2 is with a chest X-ray. This paper proposes a deep learning technique to distinguish SARSCoV2 positive and normal cases. In this study, we fine-tuned the deep learning models and hyperparameters, and the fine-tuned deep learning models performed significantly better. To classify X-ray images, we developed a system based on deep learning algorithms that includes five models: Xception, VGG19, ResNet50, DenseNet121, and Inception. We offer deep learning models and algorithms that have been trained and evaluated to support medical efforts and reduce medical staff workload when dealing with SARSCoV2. In addition, the classification model that was proposed yields positive results because it makes use of accurate classification of the SARSCoV2 disease based on medical images. Additionally, the performance of our proposed CNN classification method for medical imaging was evaluated using various edge-based neural networks. The accuracy of tertiary classification with CNN will decrease as the number of classes in the training network grows. In tertiary classification, which includes normal and SARSCoV2 positive images, the proposed model achieved a 0.9897 accuracy. The proposed algorithm achieves a high level of classification accuracy when using the DenseNet121 model for binary classification.

Deep and machine learning techniques for medical imaging-based breast cancer: A comprehensive review

In the twentieth session of the National Clinical Oncology Conference and 2017 annual meeting of Chinese Society of Clinical Oncology (CSCO), many progre-ssions have been made in the diagnosis and treatment of primary liver cancer. (1) A preliminary report of the real world study of primary liver cancer in China was released. (2) Emerging diagnostic techniques and predictive models can detect preoperative liver cirrhosis and risk of postoperative liver failure in patients with hepatocellular carcinoma (HCC). (3) Photodynamic diagnostic technology can ensure histological clearance with nega-tive surgical margin and preserve as much liver parenchyma as possible. (4) Lenvatinib is expected to be the first-line targeted drug for the treatment of patients with unresectable HCC following sorafenib. (5) Immunotherapy with Nivolumab and Pembro-lizumab is still a hot topic in the field of HCC management. (6) Though a vast heterogeneity of patterns of liver resection for HCC exists between two large centers from the East and the West, their surgical safety and long-term efficacy are actually comparable. (7) Routine lymphadenectomy, active surgical intervention for patients with late-stage tumors, ensuring the negative margins of resection and comprehensive treatment of postoperative recurrence are expected to improve the long-term outcomes of intrahepatic cholangiocarcinoma (ICC). Key words: Liver neoplasms, primary; Sorafenib; Lenvatinib; Targeted therapy; Immunotherapy

Glucose-regulated protein 78 (GRP78), a molecular chaperone widely elevated in human cancers, is critical for endoplasmic reticulum (ER) protein folding, stress signaling, and PI3K/AKT activation. Genetic knockout models of GRP78 revealed that GRP78 maintains homeostasis of metabolic organs, including liver, pancreas, and adipose tissues. Hepatocellular carcinoma (HCC) and cholangiocarcinoma (CC) are the most common liver cancers. There is a lack of effective therapeutics for HCC and CC, highlighting the need to further understand liver tumorigenic mechanisms. Phosphatase and tensin homolog (PTEN), a tumor suppressor that antagonizes the PI3K/AKT pathway, is inactivated in a wide range of tumors, including 40-50% of human liver cancers. To elucidate the role of GRP78 in liver cancer, we created a mouse model with biallelic liver-specific deletion of Pten and Grp78 mediated by Albumin-Cre-recombinase (cPf/f78f/f). Interestingly, in contrast to PTEN, deletion of GRP78 was progressive but incomplete. At 3 months cPf/f78f/f livers showed hepatomegaly, activation of lipogenic genes, exacerbated steatosis, and liver injury, implying that GRP78 protects the liver against PTEN-null mediated pathogenesis. Furthermore, in response to liver injury, we observed increased proliferation and expansion of bile duct and liver progenitor cells in cPf/f78f/f livers. Strikingly, bile duct cells in cPf/f78f/f livers maintained wild-type GRP78 level while adjacent areas showed GRP78 reduction. Analysis of signaling pathways revealed selective c-Jun N-terminal protein kinase activation, β-catenin downregulation, along with platelet-derived growth factor receptor α upregulation, which was unique to cPf/f78f/f livers at 6 months. Development of both HCC and CC were accelerated and evident in cPf/f78f/f livers at 8-9 months, coinciding with intense GRP78 expression in the cancer lesions, and GRP78 expression in adjacent normal areas reverted back to the WT level. In contrast, livers of Albumin-Cre; Grp78f/f (c78f/f) mice showed no malignancy even at 14 months. These studies reveal GRP78 is a novel regulator for PTEN-loss mediated liver injury and cancer progression. Citation Format: Wan-Ting Chen, Genyuan Zhu, Kyle Pfaffenbach, Gary Kanel, Bangyan Stiles, Amy S. Lee. GRP78 as a regulator of liver steatosis and cancer progression mediated by loss of the tumor suppressor PTEN. [abstract]. In: Proceedings of the Third AACR International Conference on Frontiers in Basic Cancer Research; Sep 18-22, 2013; National Harbor, MD. Philadelphia (PA): AACR; Cancer Res 2013;73(19 Suppl):Abstract nr C61.

The aim of this study is to develop a predictive model utilizing deep learning and machine learning techniques that will inform clinical decision-making by predicting the 1-year postoperative recovery of patients with lumbar disk herniation. The clinical data of 470 inpatients who underwent tubular microdiscectomy (TMD) between January 2018 and January 2021 were retrospectively analyzed as variables. The dataset was randomly divided into a training set (n = 329) and a test set (n = 141) using a 10-fold cross-validation technique. Various deep learning and machine learning algorithms including Random Forests, Extreme Gradient Boosting, Support Vector Machines, Extra Trees, K-Nearest Neighbors, Logistic Regression, Light Gradient Boosting Machine, and MLP (Artificial Neural Networks) were employed to develop predictive models for the recovery of patients with lumbar disk herniation 1 year after surgery. The cure rate score of lumbar JOA score 1 year after TMD was used as an outcome indicator. The primary evaluation metric was the area under the receiver operating characteristic curve (AUC), with additional measures including decision curve analysis (DCA), accuracy, sensitivity, specificity, and others. The heat map of the correlation matrix revealed low inter-feature correlation. The predictive model employing both machine learning and deep learning algorithms was constructed using 15 variables after feature engineering. Among the eight algorithms utilized, the MLP algorithm demonstrated the best performance. Our study findings demonstrate that the MLP algorithm provides superior predictive performance for the recovery of patients with lumbar disk herniation 1 year after surgery.

It has been suggested that use of the antihypertensive drugs angiotensin-converting enzyme (ACE) inhibitors and β-blockers may decrease the risk of primary liver cancer; thus, the objective of this study was to evaluate whether use of ACE inhibitors and/or β-blockers is associated with a lower risk of liver cancer. Nested case-control study. United Kingdom Clinical Practice Research Datalink. We identified 490 cases with hypertension and a first-time (incident) diagnosis of primary liver cancer between 1988 and 2011. To account for an induction period, the index date was defined as the date of the first recorded liver cancer diagnosis minus 1year. Controls were selected from patients with hypertension in the CPRD during the study period with a recorded diagnosis of hypertension who had no diagnosis of liver cancer and were free of any other cancer (except nonmelanoma skin cancer) before the index date; they were matched up to a 4:1 ratio to cases based on index date (same index date as that of their matched case), age (same year of birth), sex, general practice, and number of years of recorded history in the CPRD before the index date (1909 controls). Both cases and controls were required to have at least 2years of recorded activity in the database before the index date. Exposure was defined as receipt of two or more prescriptions for ACE inhibitors and/or β-blockers before the index date; the reference group was nonuse (0-1 prescription) of ACE or β-blocker prescriptions before the index date. We also examined the effect of duration of use and, separately, the effect of individual drugs within each medication class on risk of liver cancer, and conducted analyses restricted to patients without liver disease or diabetes mellitus. Conditional logistic regression analysis was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs). No association was found between use of ACE inhibitors and/or β-blockers and the risk of liver cancer compared with nonuse (adjusted OR 1.14, 95% CI 0.85-1.55). No significant differences were noted in risk by duration of use or by individual drugs, or after restricting the analyses to patients without diabetes or liver disease. Use of ACE inhibitors and/or β-blockers was not associated with reduced risk of primary liver cancer compared with nonuse of these drugs in persons with hypertension.