Introductory Chapter: Application of Bioinformatics Tools in Cancer Prevention, Screening, and Diagnosis

1. Introduction

Cancer is a leading cause of death worldwide, with nearly 10 million deaths in 2020, accounting for one in six deaths. Breast, lung, colon rectum, and prostate are considered the most common cancer types [1]. Around one-third of deaths from cancer are due to environmental factors and lifestyle habits, such as tobacco use, high body mass index, alcohol consumption, low fruit and vegetable intake, and lack of physical activity [2]. In addition, 10% of cancer cases are due to genetic factors and around 10% of cancer-causing infections, such as human papillomavirus (HPV) and hepatitis, are responsible for approximately 30% of cancer cases in low- and lower-middle-income countries [3]. Indeed, HPV infection is the main cause of cervical cancer, cancer that can be cured if detected early and treated effectively [4]. The multifactorial character of the disease with the huge amount of data that has been generated during the last decades covering all risk factors behind cancer disease allowed bioinformatics to play an essential role in Cancer research and made oncology a success story in translating and using OMICs data, including genomics, transcriptomics and proteomics data, in clinical settings [5].

2. Use of bioinformatics integrative approaches in oncology

Numerous research groups worldwide have attempted to develop strategies to identify novel diagnostic and prognostic markers for different cancer types based on computational integrative analyzes and tools. One of the most powerful computational approaches is meta-analysis, where multiple studies interrogating a common hypothesis are analyzed together [6]. Several studies have applied meta-analysis methods to cancer microarray data in order to identify differentially expressed genes (DEGs) between cancer patients and controls. These methods can be applied to identify robust gene-expression signatures in a single cancer type and/or to look for common expression patterns across different types of cancer. In 2004, Rhodes and co-workers investigated and analyzed 40 published cancer microarray data sets, comprising 38 million gene expression measurements from >3700 cancer samples [7]. With the advent of high throughput sequencing technology, known as NGS, RNA sequencing (RNASeq) has been used in several aspects of cancer research and therapy including the discovery of biomarkers, the characterization of cancer heterogeneity and evolution, cancer immunotherapy, and the investigation of drug resistance [8]. High throughput sequencing technology has the advantage of fast-speed sequencing at low cost and with high accuracy compared to the former Sanger technology. Compared to microarray, RNASeq can also detect unknown gene expression sequences [9]. Gene expression profiling often generates large gene-expression signatures that need to be functionally analyzed to identify a handful of genes of interest that are selected for experimental validation. Several methods have been developed allowing systematic functional analysis of gene expression signatures including Gene Ontology (GO) [10, 11], KEGG [12], TransPath [13], and GenMAPP [14]. Finally, to better understand complex biological processes, such as cancer initiation and progression, it is important to consider the integration of transcriptomic data in the context of complex molecular networks. This implies the mapping of interactomes involving protein-protein interaction with the gene expression signature to identify induced or repressed interactome subnetworks on the basis of known and predicted protein-protein interactions [15].

3. Data science in oncology

In the past decade, Artificial intelligence (AI), particularly, machine learning (ML) has grown rapidly in the context of data analysis and computing allowing applications and platforms to function in an intelligent manner (https://pubmed.ncbi.nlm.nih.gov/34278328/). ML is a field that refers to a broad range of learning algorithms that perform intelligent predictions based on learning from a subset of data [16]. AI has recently altered the landscape of cancer research and medical oncology using traditional ML algorithms and cutting-edge Deep Learning (DL) approaches [17]. Indeed, ML algorithms including Random Forest (RF), Gradient Boosting Machine (GBM), and Neural Network (NN) have been used to optimize cancer classification [18]. Furthermore, DL-based algorithms have been widely applied in medical imaging to accurately diagnose breast cancer [19], colorectal cancer [20], lung cancer [21], and others [22]. Moreover, AI systems have been developed and used to diagnose early gastric cancer (EGC) from 4667 magnifying image-enhanced endoscopy images, including 1950 EGC images from 1042 cases and 2717 noncancerous images from 769 cases [23].

4. Tools and databases

Several publicly accessible databases containing cancer related data, and integrating tools for delivering and analyzing information and data, as well as specialized databases dedicated to specific types of cancer, have been developed during the last decades. Most commonly used and prominent ones include the International Cancer Genome Consortium (ICGC) [24] and The Cancer Genome Atlas (TCGA) [25]. A detailed list of publicly available databases and their descriptions has been reported by Pavlopoulou and co-workers [26]. Recently, a novel database integrating RNA-seq, DNA methylation, and related clinical data from over 10,000 cancer patients in the TCGA study as well as from normal tissues in the GTEx study has been developed and made freely available through [27, 28]. Concerning bioinformatics and computational tools for cancer risk prediction, numerous resources have been developed including the International Breast Cancer Intervention Study (IBIS) [29], the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA) [30], the BRCAPRO [31] and the Breast Cancer Surveillance Consortium (BCSC) risk model [32]. A comprehensive list of web tools and web servers for cancer genomic study and cancer prognosis analysis has been provided by Yang and coworkers [33] and Zheng and colleagues [34].

5. Precision oncology application

Molecular and genetic profiling of tumors play an increasingly important role not only in cancer research but also in the clinical management of cancer patients [35]. Multi-omics approaches hold the promise of improving diagnostics, prognostics, and personalized treatment using highly reproducible and robust bioinformatics methods of complex data management and integration to go from the primary analysis of raw molecular profiling data to the automatic generation of a clinical report and its delivery to decision-making clinical oncologists [36]. The initial results coming out from these efforts are promising, but it has also become explicit that the exploitation of the full potential of precision oncology faces many challenges. One major bottleneck resides in the efficient and precise annotation of variants [37]. This challenge requires the use of databases containing well-curated variants as well as their interactions with potential drugs. The second challenge is the rapid development of molecular profiling techniques coming with novel challenges in terms of the development of new bioinformatics tools, pipelines, and workflows adapted to each of these new techniques [38]. Moreover, multi-omics approaches are providing more insights into dysregulated pathways, increasing the level of confidence in reporting actionable variants when they can be confirmed by RNA, protein, or epigenetic profiling. However, the availability of diverse multi-omics data is currently posing new bioinformatics challenges to integrate multiple data sets and identifying potentially efficient treatments [39]. Finally, interpreting the clinical significance of genomic variants and transcriptional changes is a laborious task that cannot be fully automated in a reliable way and therefore needs a multidisciplinary team to apply clinical interpretation to select relevant variants and to recommend targeted, personalized therapies [40]. That being said, bioinformatics still holds the hope to make the intersection of cancer research and medical applications for better clinical management of patients.