Ijrar.org



Role of Artificial Intelligence in 21st Century Cancer Research: A Review1Arnab Kar, 2Susanta Roy Karmakar, 3Subir Chandra Dasgupta, 4Sujit Kumar Bhowal*1Post Graduate student of Zoology, 2,4Assistant Professor of Zoology, 3Professor of Zoology1, 2, 3, 4Department of Zoology,Maulana Azad College, Kolkata- 700013, India*Corresponding authorEmail: sujit.sb@Telephone: 7596812940 Abstract: From the last century scientists from all over the globe have researched tirelessly to defeat the deadly disease, Cancer. In the past few years Artificial Intelligence left a satisfactory impression in some fields of medical science. Nowadays, Artificial Intelligence have fascinating role in cancer detection, molecular nature of cancer, which varies patient to patient, high resolution medical imaging, genomic profiling, dose detection during chemo- and radiotherapy, calculation of survival rate etc. Machine learning, especially deep learning, has the ability to discover malignancies and decision making capacity, which helps in drug design. Technical development of computer science, technologies, statistics continuously improve the accuracy of such analysis which is far better than traditional analysis. Recent development of algorithms and Deep Neural Networks help in cancer prognosis and diagnosis. The problem arising in research also discussed here. However, we are hopeful that in the near future we will be able to modify this system to diagnose cancer as early as possible and treat cancer in the most severe stage; to make it the most popular strategy in cancer treatment.Index Terms - Algorithm; Artificial Intelligence; Cancer; Deep learning; Deep Neural Network; Dose detection; Machine learning.INTRODUCTION:Cancer after its discovery diverts our maximum attention towards it due to its high fatality rate and complex treatment procedure. According to the World Health Organization (WHO), the global cancer burden is estimated to have risen to 18.1 million new cases and 9.6 million deaths in 2018. One in 5 men and one in 6 women worldwide develop cancer during their lifetime, and one in 8 men and one in 11 women die from the disease (WHO, 2018). Cancer has some unique property such as it can avoid therapies due to its drug resistant property; it interacts with its micro environment and maintains a cancer stem cell niche which is very much difficult to destroy. For this reason cancer relapses within some days after treatment stopped. Another problem associated with cancer is that it is very much difficult to detect a cancer at the early stage of growing. Following this complexity, researchers are now trying hard to develop strategies that are hard to be avoided by cancer.Artificial intelligence (AI) was first coined by John McCarty (Fig.1) in a conference at Dartmouth College. This new technology, which originated from extensive interlinking between engineering, computer science and other applied sciences, is developing main fields of applied research with technology transfer in robotics, natural language processing, machine learning, computer vision etc. (Coccia, 2019). Development of Metal Oxide Semiconductor and very large scale integration helped to build Artificial Neural Network in around 1980s and improve machine learning efficiency. Around the 1970s a system called MYCIN developed which has a significant role in medical applications. In 2010, IBM (International Business Machines) introduced a supercomputer called WATSON that helps in cancer detection and treatment recommendation like a virtual oncologist. Moreover, there are several AI systems developed for this structure.Actually, AI analyzes a huge amount of data that cannot be perceived by the human brain, and makes decisions based on its previous experiences. The bigger no. of data makes more accurate decisions. It is also used for non-invasive cancer detection with high accuracy in melanoma, cervical cancer, uterine cancer, breast cancer. The basic work of AI is done by Artificial Neural Network (ANN) which acts like a biological neural network; it performs as a nonlinear data analysis tool where complex relationships between input and output are modelled. ANN is the foundation of machine learning (ML) which is a type of AI, not explicitly programmed to perform a specific task used to construct predictive algorithm models that learn logical patterns from mass historical data so as to predict the survival rate of a patient. ML has been used extensively for improving prognosis. The more data an ML model is exposed to, the better it performs over time (Huang, et al., 2019).Deep learning (DL) which is a subset of ML (Fig.2) has attracted more attention in the last few years. DL uses ANN to design an algorithm from a large amount of input data like the human brain. DL often makes decisions based on Medical imagery like CT scan, Magnetic Resonance Imaging (MRI) etc. AI also gained huge popularity in non-invasive cancer detection and advanced multi drug systems in recent days. There is a huge prospect of artificial intelligence combined with nanotechnology which makes a major breakthrough in cancer research.TYPE OF MACHINES:Machine Learning (ML) algorithm is the important tool which is used in nearly all AI applications in cancer research at first the machine is learnt by huge amounts of input data (Fig.3). Machine learning has 2 types:-Supervised LearningSupervised learning is used to train ML algorithms with a labelled dataset of inputs and outputs. An algorithm makes itself familiar with general rules that map input to output. Supervised ML algorithms can learn patterns hidden in inputted data for categorical outputs (classification) and continuous data (regression).Unsupervised LearningUnsupervised ML algorithms use unlabeled data, with an attempt to discover any structure in the input data. Usually Unsupervised ML algorithms help to simplify (dimensionality reduction) or organize (clustering) data.Usually, in cancer prediction and prognosis ML algorithm uses supervised learning. However, this supervised learning acts as classifiers that classify on the basis of conditional probabilities (Wishart et al., 2006). There are some major types of conditional algorithms, such as:-Artificial neural networks (ANN – Rummelhart et al.,1986).Genetic algorithms (GA – Holland, 1975).Decision trees (DT – Quinlan, 1986).Linear discriminant analysis (LDA) methods.k-nearest neighbor algorithmsArtificial Intelligence in Cancer Prognosis Prediction:One of the major obstacles that clinicians face in cancer prevention is that cancer gone unnoticed in most cases because they have nothing but their work experience for diagnosis of cancer and in numerous cases patients die due to treatment stress. So, in recent years with the development of modern AI technologies and machine algorithm prognosis, prediction of a specific cancer has become more accurate (Fig.4).In Brain Cancer:Brain malignancies predicted by an AI method MIRSPSO (mutual information and rough set of particle swarm optimization) which uses Support Vector Machine (SVM) algorithm and Random Forest (RF) algorithm separately. Both of these algorithms make predictions based on learned patterns (inputs). When examined in 700 brain cancer patients this method predicted with 88.5% accuracy i.e. far better than conventional statistical methods (Zhao et al.,2019).3.2 In Oral Cancer:Oral squamous cell carcinoma is a challenging problem in third world countries. Prognosis of this disease can successfully be done using AI technologies. In a study (Zain et al., 2013) uses a combination of ANFIS (Adaptive neuro- fuzzy inference system), artificial neural network, support vector machine, logistic regression to predict death rate after a particular treatment. Clinicopathologic variables from the OCRCC database and genomic variables from Immunohistochemistry (IHC) staining are inputted into the algorithmic model. Three features viz. drink, invasion and p63 achieved the best accuracy (accuracy= 93.81%; AUC= 0.90) for prognosis of the oral cancer.In another study (Rajpoot et. al., 2019) developed a digital marker using an abundance of tumor infiltrating lymphocyte (Fig.5) biomarkers as prognostic indicator. five state-of-the-art convolutional neural network models ResNet50, DenseNet, Inception, Xception and MobileNet acts as classifier and the total no. of disease free survival predicted using Kaplan-Meier (KM) curves and Cox hazard analyses by conducting univariate and multivariate analysis of digital, clinical, and pathological parameters (Fig.6).3.3 In Breast Cancer:There are several studies conducted by several researchers to utilize AI in predictive prognosis of breast cancer. Most of the research concentrated on the use of Deep Neural Network (DNN) to analyze multidimensional data. (Ching et al., 2018) developed a new ANN framework known as Cox-net (a neural network extension of the Cox regression model) to predict patient survivability from high throughput transcriptomics data. Cox-net digs out extensive biological information, at both the gene levels and cellular pathway, by analyzing features represented in the hidden layer nodes in Cox-net.Chi et al., 2007 designed MDNNMD (multimodal DNN by integrating multi-dimensional data) for the prognosis of breast cancer (Fig.7). The design of the method’s architecture and the fusion of multi- dimensional data are the novelties of this system and make this system fruitful. The results of the comprehensive performance evaluation reflect that this proposed method outperformed all the other prediction methods using single dimensional data.3.4 In Prostrate Cancer:Urologists are able to develop AI technology based on DNN algorithms to monitor prostate cancer. For an independent test dataset three criteria of a malignant tissue i.e. its presence, extent, and Gleason grade (used to determine aggressiveness of prostate cancer) (Fig.8) were predicted by the DNN. Strom et al., 2020 developed this algorithm when digitized 6682 slides from needle core biopsies from 976 randomly selected participants aged 50– 69 in the Swedish prospective and population-based STHLM3 diagnostic study and another 271 from 93 men from outside the study. The captured images were used to train deep neural networks to analyze prostate biopsies.Whole slide images were classified through a segmentation algorithm based on Laplacian filtering to identify the region corresponding to tissue sections. Two convolutional DNN ensembles (each consisting of 30 inception V3 models) were used which were pre-trained on ImageNet with classification layers adapted to expectable outcome.3.5 In Cervical Cancer:The National Cancer Institute developed a ML algorithm to predict cervical cancer. In normal, Visual Inspection after Application (VIA) of acetic acid has been done to visualize precancerous and cancerous cells in the cervix. Sometimes in very complicated stages cervical cytology (Pap tests) and colposcopy have been done. The deep learning algorithm Faster R- CNN (Faster Region based convolutional neural network) was used and the machine was trained by images collected by cervicography from a population-based longitudinal cohort of 9406 women, all of whom ages 18–94 years in Guanacaste, Costa Rica (Schiffman et. al., 2019). The patients were subdivided into two groups (Fig.9). In one group (cases) more severe patients CIN2+ and CIN2 were kept, in another group patients with <CIN2 were kept. For training set images from 189 images from cases and 555 images from control fed to the machine to design algorithm. Therefore validation test last images taken from women during follow up used; 82 from cases and 242 from control. Lastly for screening test at first images used from 85 women in the case group and 8,174 from the control group. The ML algorithm complete the screening with higher accuracy (area under the curve [AUC] = 0.91, 95% confidence interval [CI] = 0.89 to 0.93) than original cervigram interpretation (AUC = 0.69, 95% CI = 0.63 to 0.74; P <.001) or conventional cytology(AUC = 0.71, 95% CI = 0.65 to 0.77; P<.001).3.5 In Gastric Cancer:ANN has been shown to be a more powerful statistical tool for predicting the survival rate of gastric cancer like all other cancer prediction system patients compared to the Cox proportional hazard regression model. Oh et al., 2018 used a survival recurrent network (SRN) to predict survival rate, and the results corresponded closely with actual survival. The American Joint Committee on Cancer (AJCC) also certified this system as a powerful tool to make predictions for gastric cancer. SRN model provides an individualized prediction based on numerous factors rather than only tumor factors followed by TNM staging in traditional models; basically, patient grouping is not necessary in SRN technique. Biglarian et al., 2011, analyzed 436 registered gastric cancer patients who had had surgery between 2002 and 2007 at the Taleghani Hospital, Tehran (Iran), to predict the survival time using Cox proportional hazard and ANN techniques. Moreover, the Cox regression analysis also revealed that the survival rates significantly associated with the patient’s age at the time of diagnosis, high-risk behaviors like extent of wall penetration, distant metastasis, and tumor stage. Actually, using these features the system makes almost true predictions which are 83.1%accurate.3.6 In Lung Cancer:In most of the cases it is very difficult to understand the survivability of lung cancer patient after treatment. Lynch et al., 2017 used a supervised learning procedure to SEER (Surveillance, Epidemiology, and End Results) program database to classify lung cancer patients in terms of survival. The database uses different algorithms including linear regression, Decision Trees, Gradient Boosting Machines (GBM), Support Vector Machines (SVM), and a custom ensemble. For this study, researchers chose linear regression, Decision Trees, ensemble learning algorithms, Random Forest and Generalized Boosting Machines as logic-based methods, Support Vector Machine (SVM) using a polynomial kernel function as a non-probabilistic method; and at the final stage of prediction, custom ensemble method used a weighting function to sum the prediction of each of the five individual models. The final result showed that the GBM algorithm is the best ML procedure to perform the task.ARTIFICIAL INTELLIGENCE IN CANCER DIAGNOSIS AND IMAGING:Due to the limited ability of the human brain to analyse large amounts of data AI models are used to make cancer diagnosis more accurate. Clinicians are very much dependent on their work experience and personal knowledge to diagnose cancer but in AI technology Integrative processing and extraction can allow more accurate disease diagnosis due to the efficiency and effectiveness of learning and training large samples. Different AI setups are able to state the molecular nature and the treatment sensitivity of a tumour. Deep Convolutional Neural Network (DNCC) used to diagnose thyroid cancer; and shows high sensitivity and increased specificity compared to radiologists. DNCC has diagnosed the tumour from sonographic images taken during clinical trials (Li ei al., 2019).Orringer et al., 2020 studied about the efficiency of AI technique in intraoperative brain tumour diagnosis. Actually, in the traditional way H-E staining of processed tissue was done which is time and labour consuming. DNCC aided technique can complete diagnosis within 3 minutes and replace traditional methods which took 40 minutes. To develop ML algorithms the machine was trained with 2.5 million Stimulated Raman Histology (SRH) images. They repeated the same method multiple times to every common type of brain tumour and got an overall success (Fig.10).Prior to AI, dermatologists depend only on biopsy results to confirm melanoma. But use of CNN helps a lot to diagnose melanoma through noninvasive ways. Uhlmann et al., 2018 have studied to train CNN, validate and test the DL algorithm for the diagnostic classification of dermoscopic images of lesions of melanocytic origin. The researchers use Google’s Inception v4 CNN architecture and train the machine with 300 images test set (comprises 20% melanoma image and 80% benign tumor image). Main outcome measures were sensitivity, specificity and diagnostic classification. The efficiency of the machine was compared with the knowledge of 58 international dermatologists.In another study Phillip et al., 2019 studied to examine the accuracy of artificial intelligence neural network DERM (Deep Ensemble for Recognition of Melanoma) to detect malignant melanoma from dermoscopic images of pigmented skin lesions (Fig.11) and to show how this outperforms doctors’ ability of detection assessed by meta-analysis. Particularly those deep learning techniques which identify and assess features of pigmented lesions that are associated with malignant melanoma used to develop DERM. The novelty of deep learning is that it can learn and auto-determines features which are associated with malignant melanoma directly from the data; unlike earlier machine learning methods which learn features pre-determined by researchers. The algorithm was experienced and accepts a dataset of archived dermoscopic images of skin lesions, using 10- fold cross-validation. This preliminary analysis demonstrates the ability of an AI-based system i.e. DERM algorithm to learn features of a skin lesion that are associated with malignant melanoma, which can then be applied to the identification of malignant melanoma (Fig.12)In Qiu et al., 2017 employed an AI system known as to develop and test computer-aided diagnosis (CAD) scheme of mammograms (Fig.13) for classifying between malignant and benign masses. An image dataset involving 560 regions of interest (ROIs) extracted from digital mammograms was used. In the algorithm setup an 8 layer deep learning network consists of 3 pairs of convolution- max-pooling layers for automatic feature extraction and a multiple layer perceptron (MLP) classifier which help in feature categorization were implied to each ROI to process the DL.To improve the feature robustness each convolutional layer is connected with a max pooling layer. The output comes from the sixth layer mix totally connected with a MLP classifier which then generates a classification score to prognosticate the likelihood of malignancy of a particular ROI. Again, a four-fold cross validation method was applied to train and test this deep learning network for better accuracy.This study demonstrates the feasibility of applying a CAD scheme based on deep learning to categorize malignant and benign breast masses without a lesion segmentation, image feature computation and selection process.AI AND GENOME ANALYSIS IN CANCER:Implication of AI to classify cancer is a modern approach, where expression of relevant genes involved in various types of cancer is detected using ML algorithms. This system is also helpful to detect the origin of cancer in a patient’s body when cancer is detected in a critical stage.The ML algorithm trained with features subsets (features correlated with the cancer class) to build a mathematical model. ML evaluates features subsets in two ways called filter method, and wrapper method. General characteristics of training data set considered by filter while, wrapper approaches measure biases of ML algorithm (Mewes et al., 2005).AlphaGo (Hassabis et al., 2016) object recognition is a deep learning algorithm that outperforms human analytical power. DL used to find out complex structures in massive data set like splice site promoters and enhancers. More accuracy may come if convolutional neural networks are used in DL methods. (Sun et al., 2019) develop genomic deep learning (GDL) algorithm to classify cancer based on relationship of mutation (at oncogene or tumor suppressor gene) and cancer. To develop GDL algorithm researchers used 6083 WES (whole exon sequencing) files from 12 different types of cancer from TCGA (The Cancer Genome Atlas) and 1991 WES of healthy samples from IGSR (International Genome Sample resource). After that, 12 specific models were prepared which can distinguish a specific cancer from others 97.47% accurately; a mixture model to distinguish between all 12 types of cancer based on GDL, 70.08% accurately and a total specific model which can identify healthy and cancer tissue 94.7% accurately.AI IN PERSONALISED MEDICINE DESIGNING:In the age of high-throughput, data-intensive biomedical research assays, it is necessary for researchers to develop strategies for analyzing, integrating, and explaining the massive amounts of data generated. Given how important data-intensive assays are to reveal appropriate intervention targets and strategies for treating an individual with a disease, AI can play a significant role in the development of personalized medicines (Fig.14).Different approaches like SVM, DL, NLP (Neural Language Processing) may be integrated to translate high dimensional data into clinically actionable data which acts as a foundation tool in precision medicine. Next generation sequencing has revolutionized medical research and enable multilayer studies that integrate genomic data of high dimensionality such as DNA- seq, RNA- seq and multi- omics data such as proteome, epigenome and micro-biome (Chari et al., 2010). The integrative analysis done by several AI technologies such as BIOXCEL Therapeutics, XTALPI AI etc. help in designing and development of precise medicine.In the recent past, Microsoft has built up collaboration with research groups of Jackson Laboratory to develop a database called the Clinical Knowledgebase (CKB). A machine learning called JAX researchers can curate CKB, which stores information about genetic mutations that impulse cancer and drug responses in cancer patients. The database helps oncologists find out what matches exist between a patient’s known cancer-related genomic mutations and drugs that target them, allowing providers to choose the most appropriate treatments (Kent, 2019).AI TO DESIGN DRUG DELIVERY SYSTEM:Over the last decades, targeted delivery of therapeutics with maximal efficiency and minimal side effects through different systems has attracted a growing interest. Target fishing (TF) is a technique which helps to link biological targets with novel drug compounds as well as rapid prediction and identification of biological targets. To design the control system neural networks, fuzzy logic, integrators, and differentiators may be applied. Drugs may be delivered using focused ultrasound, micropump mechanism, and targeted delivery by micro-robots (Fig.15).Advance wireless communication systems also provide a high degree of flexibility to the entire drug delivery system and receive and send data from external sources to monitor and control drug delivery. Microfluidic technology permits the development of smart drug delivery systems, like Janus micro- or nanoparticles capable of delivery of multiple drugs (Staples et al., 2006). For programmed drug delivery, electronic components, wireless communication hardware, and power supply have been embedded in a microchip implant (MicroCHIPS, Inc.) followed by a pulsatile release for about six rmation technology, wireless communication, and artificial neural networks (ANNs) contribute to modelling smart drug delivery systems that are very helpful for overcoming the limitations of conventional treatment strategies. ANN helps to monitor the factor responsible for cytotoxicity when a nanoemulsion system is applied. Model neurons are stimulated to create interconnected processing element of ANNs which are helpful to generate software for mimicking the biological processes, generating the control algorithms, Pharmacodynamic/ pharmacokinetic modeling, controlled drug delivery, and evaluating the effectiveness of treatment strategies (Murtoniemi et al., 1994). Ligand based targets are predicted by ML approaches which give rise of Target Fishing (TF) in which ligand dataset , target protein, and relationship between the ligands may be used to predict protein targets of novel compounds with biological activities (Sherrif et al., 2004).CONCLUSION:World famous scientist Stephen Hawking once said, “The development of full artificial intelligence could spell the end of the human race. Once humans develop artificial intelligence, it will take off on its own and redesign itself at an ever-increasing rate. Humans, who are limited by slow biological evolution, couldn't compete and would be superseded.” Though AI is the ultimate solution in some complicated cases, more clinical trials should be done to make it successful; in the sense compatible otherwise it will get an opportunity to destroy the entire human race. Now human employment is standing before a question mark due to rapid implementation of AI. Both the Government and Private sectors should make a brief policy to restrict its use. Another existing problem is the “black-box problem”, i.e. our knowledge cannot be able to detect how AI makes a particular inference. Still AI depends on other techniques for learning; it can learn about cancer patterns by itself and this drawback can raise questions about its authenticity. However, scientists should strictly follow the ethical guideline. Designing algorithms obviously is a tough work and should be done under proper surveillance. As prevention is better than cure so we should give more attention towards cancer diagnosis than treatment. Hope in the near future we can make a cancer free world with the help of both human and artificial intelligence.FIGURS AND GRAPHS:4648201270Figure 2- Relationship between artificial Intelligence, Machine Learning, Deep LearningFigure 1- John McCarty (1927- 2011)97028078105Figure3-An example of how a machine learner is trained to recognize images using a training set542290-401320Figure 4-General work flow of AI for making predictionFigure 5- Tumour Infiltrating LymphocyteFigure-6 - A Novel Digital Score for abundance of tumour infiltrating lymphocytes predicts disease free survival in oral squamous cell carcinoma 9563100Figure-7- Multimodal DNNs to predict human breast cancer prognosis by integrating multi-dimensional data including gene expression profile, copy no. alteration (CAN) profile, and clinical data. The model consist of a triple modal DNN and it is combined predictive scores from each independent model. The graphical representations (A. Li et al, 2018) show a comparison of the ROC curve and AUC value. The results indicated that combining different data types and ensemble DNN methods is an efficient way to improve human breast cancer prognosis prediction performance.Figure 8- Prostrate CancerFigure 9 - The basic design of automated visual evaluation algorithm. Two models are trained: a cervix locator (top), and the automated visual evaluation detection algorithm (bottom). The final validation algorithm incorporated both cervix locator and automated visual evaluation.Figure 10: (a) Fresh surgical specimen; (b1) Image acquisition; (b2) Image processing; (c) Intra-operative diagnostic prediction.Figure 11-Melanoma diagnosis by machineFigure 12-ML algorithm can diagnose skin cancer from scanned imageFigure 14- Basic pathway to develop personalized medicineFigure 13- Diagnosis of Breast cancer by DL algorithmFigure 15- Basic design of nano-drug delivery system.ACKNOWLEDGMENT:The authors would like to thank The Principal, Maulana Azad College to his kind support. The authors are also thankful to all teaching staffs and non-teaching staffs of Maulana Azad College for their kind help.REFERENCES:Alessandra, M., Maria, D.D., Gulia, L., 2010. Oral pulsatile delivery: rationale and chrono-pharmaceuticle formulation, Int. J. Pharm, 398, p:1-8Bhattacharya, T., Brettin, T., Doroshow, J.H., Evrard, Y.A., Greenspan, E.J., Gryshuk, A.L., Hoang, T.T., Lauzon, C.B.V., 2019. AI Meets Exascale Computing: Advancing Cancer Research With Large-Scale High Performance Computing. Front Oncol. 9, p:984 doi:10.3389/fonc.2019.00984Biglarian, A., Hajizadeh, E., Kazemnejad, A., Zali, M., 2011. Application of artificial neural network in predicting the survival rate of gastric cancer patients. Iran J Public Health. 40(2)p:80–86.Chang, S.W., Abdul-Kareem, S., Merican, A.F., Zain, R.B., 2013. Oral cancer prognosis based on clinicopathologic and genomic markers using a hybrid of feature selection and machine learning methods. BMC Bioinformatics.14,p: 170. doi: 10.1186/1471-2105-14-170Chari, R., Coe, B.P., Vucic, E.A., Lockwood, W.W., Lam, W.L., 2010. An integrative multi-dimensionalgenetic and epigenetic strategy to identify aberrant genes and pathways in cancer. BMC Syst. Biol., 4,p:67. doi:10.1186/1752-0509-4-67Chi, C. L., Street, W. N., & Wolberg, W. H., 2007. Application of artificial neural network-based survival analysis on two breast cancer datasets. AMIA Annual Symposium proceedings. AMIA Symposium, 2007, p: 130–134.Ching T, Zhu X, Garmire LX (2018) Cox-nnet: An artificial neural network method for prognosis prediction of high-throughput omics data. PLoS Comput Biol 14(4): e1006076. doi:- 10.1371/journal.pcbi.1006076Coccia, M., 2019. Artificial intelligence technology in oncology: a new technological paradigm, Yale University School of Medicine, Global Oncology.Cruz, J. A., Wishart, D. S., 2006. Applications of Machine Learning in Cancer Prediction and Prognosis, Cancer Informatics: 2, p: 59–78Haenssle, H. A., Fink, C., Schneiderbauer, R., Toberer, F., Buhl, T., Blum, A., Kalloo, A., Hassen, A., Thomas, L., Enk, A., Uhlmann, L., Reader study level-I and level-II Groups, Alt, C., Arenbergerova, M., Bakos, R., Baltzer, A., Bertlich, I., Blum, A., Bokor-Billmann, T., Bowling, J., Zalaudek, I., 2018. Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists. Annals of oncology: official journal of the European Society for Medical Oncology. 29(8), p:1836–1842Hassanzadeh, P., Atyabi, F., & Dinarvand, R., 2019. The significance of artificial intelligence in drug delivery system design. Advanced drug delivery reviews, 151-152, p:169–190. doi:10.1016/j.addr.2019.05.001.news/microsoft-using-ai-to-accelerate-cancer-precision-medicineHeneghan, H. M., Miller, N., Kelly, R., Newell, J., & Kerin, M. J., 2010. Systemic miRNA-195 differentiates breast cancer from other malignancies and is a potential biomarker for detecting noninvasive and early stage disease. The oncologist, 15(7), p:673–682. doi:10.1634/theoncologist.2010-0103Herrero, R., Schiffman, M. H., Bratti, C., Hildesheim, A., Balmaceda, I., Sherman, M. E., Greenberg, M., Cárdenas, F., Gómez, V., Helgesen, K., Morales, J., Hutchinson, M., Mango, L., Alfaro, M., Potischman, N. W., Wacholder, S., Swanson, C., Brinton, L. A., 1997. Design and methods of a population-based natural history study of cervical neoplasia in a rural province of Costa Rica: the Guanacaste Project. Revista panamericana de salud publica, Pan American journal of public health, 1(5),p:362–375.Hollon, T. C., Pandian, B., Adapa, A. R., Urias, E., Save, A. V., Khalsa, S., Eichberg, D. G., D'Amico, R. S., Farooq, Z. U., Lewis, S., Petridis, P. D., Marie, T., Shah, A. H., Garton, H., Maher, C. O., Heth, J. A., McKean, E. L., Sullivan, S. E., Hervey-Jumper, S. L., Patil, P. G., Orringer, D. A., 2020. Near real-time intraoperative brain tumor diagnosis using stimulated Raman histology and deep neural networks. Nature medicine, 26(1), p:52–58.doi:10.1038/s41591-019-0715-9 Hsu, E. R., Klemm, J. D., Kerlavage, A. R., Kusnezov, D., Kibbe, W. A. (2017). Cancer Monoshot Data and Technology Team: Enabling a National Learning Healthcare System for Cancer to Unleash the Power of Data. Clinical pharmacology and therapeutics, 101(5), p:613–615. doi:10.1002/cpt.636, L., Bell D.,Antani, S., Xue Z., Yu, K., Horning P.M., Gachuhi, N., Wilson, B., Jaiswal, M.S., Befano, B., Long, L.R., Herrero, R., Einstein, M.H., Burk, R. D., Demarco, M., Gage, J. C., Cecilia A., Rodriguez, Wentzensen, N., Schiffman, M., 2019. An Observational Study of Deep Learning and Automated Evaluation of Cervical Images for Cancer Screening, JNCI: Journal of the National Cancer Institute, 111(9), p:923–932,doi:10.1093/jnci/djy225Hua, K. L., Hsu, C. H., Hidayati, S. C., Cheng, W. H., & Chen, Y. J. 2015. Computer-aided classification of lung nodules on computed tomography images via deep learning technique. OncoTargets and therapy, 8, 2015–2022. doi:10.2147/OTT.S80733Huang S., Yang J., Fong S., Zhao Q., 2019.Mining Prognosis Index of Brain Metastases Using Artificial Intelligence, Cancers, 11, 1140; doi:10.3390/cancers11081140Jhajharia, S., Varshney, H.K., Verma, S., Kumar, R., 2016. A neural network based breast cancer prognosis model with pca processed features. In: 2016 Int. Con. on Adv. in Computing, Communications and Informatics (ICACCI), p: 1896–1901Kann, B.H., Thomas, R. C., Thompson, R., Artificial Intelligence in Oncology: Current Applications and Future Directions, Oncology (Williston Park). 33(2), p:46-53.Levine, A.B., Schlosser, C., Grewal, J., Coope, R., Jones, S.J.M., Yip, S., 2019. Rise of the machines: advances in deep learning for cancer diagnosis. Trends Cancer, 5, p:157–169. doi:10.1016/j.trecan.2019.02.002Li, X. Zhang, S., Zhang, Q., Wei, X., Pan, Y., Zhao, J., Xin, X., Qin, C., Wang, X., Li, J., Yang, F., Zhao, Y., Yang, M., Wang, Q., Zheng, Z., Zheng, X., Yang, X., Whitlow, C.T., Gurcan, M.N., Zhang, L., Wang, X., Pasche, B.C., Gao, M., Zhang, W., Chen K., 2019, Diagnosis of thyroid cancer using deep convolutional neural network models applied to sonographic images: a retrospective, multicohort, diagnostic study, Lancet Oncol. 20, p:193–201.Lusci A, Pollastri G, Baldi P. Deep architectures and deep learning in chemoinformatics: The prediction of aqueous solubility for drug-like molecules. J. Chem. Inf. Model, 53, p:1563-75.Ly, D. Q., Paramonov, L., Davidson, C., Ramsden, J., Wright, H., Holliman, N., Hagon, J., Heggie, M., Makatsoris, C., 2011 Nanotechnol. Percept. 7, p:199–217Lynch, C. M., Abdollahi, B., Fuqua, J. D., de Carlo, A. R., Bartholomai, J. A., Balgemann, R. N., van Berkel, V. H., & Frieboes, H. B., 2017. Prediction of lung cancer patient survival via supervised machine learning classification techniques. International journal of medical informatics, 108, p:1–8. doi:10.1016/j.ijmedinf.2017.09.013media.Murtoniemi, E., Merkku, P., Kinnunen, P., Leivisk, K., Yliruusi, J., 1994. Effect of neural network topology and training end point in modelling the fluidized bed granulation process, Int. J. Pharm. 110, p:101– 108Oh, S. E., Choi, M. G., & Seo, S. W., 2019. ASO Author Reflections: Use of the Survival Recurrent Network for Prediction of Overall Survival in Patients with Gastric Cancer. Annals of surgical oncology, 26(Suppl 3), p:539–540.doi.10.1245/s10434-018-7044-yPadovani de Souza, K., Setubal, J. C., Ponce de Leon F de Carvalho, A. C., Oliveira, G., Chateau, A., Alves,R. 2019. Machine learning meets genome assembly. Briefings in bioinformatics, 20(6), p: 2116–2129. doi: 10.1093/bib/bby072Phillips, M., Greenhalgh, J., Marsden, H., Palamaras, I. 2019. Detection of Malignant Melanoma Using Artificial Intelligence: An Observational Study of Diagnostic Accuracy. Dermatology practical & conceptual, 10(1), e2020011. doi:10.5826/dpc.1001a11Press release (N- 263), The International Agency for Research on Cancer (IARC), World Health Organization, 2018.Qiu, Y., Yan, S., Gundreddy, R. R., Wang, Y., Cheng, S., Liu, H., & Zheng, B. 2017. A new approach to develop computer-aided diagnosis scheme of breast mass classification using deep learning technology. Journal of X-ray science and technology, 25(5), p: 751–763.doi:10.3233/XST-16226Quinlan, J.R., 1986, Induction of decision trees. Mach Learn, 1, p:81–106. doi: 10.1007/BF00116251Ramesh, K., Gupta, S. Ahmed, S., Kakkar, V., 2016. A Comprehensive Study on Design Trends and Future Scope of Implantable Drug Delivery Systems. Int. J. of BioScience and Biotech. 8, p:11-20. doi: 10.14257/ijbsbt.2016.8.6.02.Sacha, G.M., Varona, P., 2013. Artificial intelligence in nanotechnology. Nanotechnology. 24, p:1–13.Shaban, M., Khurram, S. A., Fraz, M. M., Alsubaie, N., Masood, I., Mushtaq, S., Hassan, M., Loya, A., & Rajpoot, N. M. (2019). A Novel Digital Score for Abundance of Tumour Infiltrating Lymphocytes Predicts Disease Free Survival in Oral Squamous Cell Carcinoma.?Scientific reports,?9(1), 13341.doi:-10.1038/s41598-019-49710Sherriff, A., Ott, J., & ALSPAC Study Team, 2004. Artificial neural networks as statistical tools in epidemiological studies: analysis of risk factors for early infant wheeze. Paediatric and perinatal epidemiology, 18(6), p:456–463. doi:-10.1111/j. p:1365-3016.Silver, D., Huang, A., Maddison, C., Guez, A.,Sifre, L., Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., Dieleman, S., Grewe, D., Nham, J., Kalchbrenner, N., Sutskever, I., Lillicrap, T., Leach, M., Kavukcuoglu, K., Graepel, T., Hassabis, D. (2016). Mastering the game of Go with deep neural networks and tree search. Nature. 529. p:484489. doi: 10.1038/nature16961.Staples, M., & Daniel, K., Cima, M.J., Langer, R. 2006. Application of Micro- and Nano Electromechanical Devices to Drug Delivery. Pharmaceutical research. 23. p:847-63. doi: 10.1007/s11095-006-9906-4.Str?m, P., Kartasalo, K., Olsson, H., Solorzano, L., Delahunt, B., Berney, D. M., Bostwick, D. G., Evans, A. J., Grignon, D. J., Humphrey, P. A., Iczkowski, K. A., Kench, J. G., Kristiansen, G., van der Kwast, T. H., Leite, K., McKenney, J. K., Oxley, J., Pan, C. C., Samaratunga, H., Srigley, J. R., Eklund, M., 2020. Artificial intelligence for diagnosis and grading of prostate cancer in biopsies: a population-based, diagnostic study. The Lancet. Oncology, 21(2), p:222–232. doi:10.1016/S1470-2045(19)30738-7Sun D., Wang M., Li A., 2018. A multimodal deep neural network for human breast cancer prognosis prediction by integrating multi-dimensional data, IEEE ACM Trans. Comput. Biol. Bioinform. 16 (3), p:841–850. doi:10.1109/TCBB.2018.2806438Sun, Y., Zhu, S., Ma, K., Liu, W., Yue, Y., Hu, G., Lu, H., Chen, W., 2019. Identification of 12 cancer types through genome deep learning. Sci Rep, 9, 17256. doi:10.1101/528216.Sutradhar, K. B., Sumi, C. D., 2016. Implantable microchip: the futuristic controlleddrug delivery system. Drug delivery, 23(1),1–11.doi:10.3109/10717544.2014.903579Wang, Y., Tetko, I.V., Hall, M.A., Frank, E., Facius, A., Mayer, K.F.X., Mewes, H.W., 2005, Gene selection from microarray data for cancer classification- a machine learning approach, Computational Biology and chemistry, 29(1), p: 37-46Xu, J., Yang, P., Xue, S., Sharma, B., Sanchez-Martin, M., Wang, F., Beaty, K. A., Dehan, E., & Parikh, B. 2019. Translating cancer genomics into precision medicine with artificial intelligence: applications, challenges and future perspectives. Human genetics, 138(2), p:109–124. doi: 10.1007/s00439-019- 01970-5Zen, K., Zhang, C.Y., 2012. Circulating microRNAs: a novel class of biomarkers to diagnose and monitor human cancers. Med Res Rev, 32, p:326–48.b2b1aFig 12: Diagnosis of Breast cancer by DL algorithm ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download