Biobanking in health care: evolution and future directions

The history of cell line biobanking started with generation of the HeLa cell line in 1951 at Johns Hopkins Hospital, where the medical staff obtained the first cancer cell line from a patient named Henrietta Lacks (HeLa). Cancer cells were obtained without her consent after surgery and cultured in the lab using an experimental approach that made the cells immortal, thereby achieving, for the first time, an in vitro cancer model system for research [8]. The HeLa cell line is used worldwide in the research laboratories because it is easy to grow and offers an optimal and stable model system for in vitro research experiments. Thanks to the epochal HeLa cell line, scientific results have been gained with crucial advantages for global health, including the development of polio vaccines [9]. The success of the HeLa cell line encouraged the stabilization and use of immortalized cell lines for medical research, particularly during the 70 s and 80 s when several continuous cancer cell lines were established. The growing number of newly generated cell lines underscored the need for an impartial organization that could guarantee the origin and quality of each model system. In this way, the evolution of cell line biobanks highlights the importance of standardizing technical procedures and ensuring data reproducibility in medical research. The first step for procedure harmonization was the institution of official cell line biobanks. The first was the American Type Culture Collection (ATCC) that was established with the aim of providing certified model systems to the scientific community. Currently, major cell line repositories include: (i) the ATCC (USA), (ii) the Leibniz-Institute DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen); (iii) the European Collection of Cell Cultures (ECACC); (iv) the Japanese Cancer Research Resources Bank (JCRB); RIKEN BioResource Center (Japan); and (v) the Korean Cell Line Bank (KCLB). These repositories still represent international references that guarantee authentic and highly controlled in vitro model systems for medical research, with appropriate certification including disease-associated data, karyotyping, immunoprofiles, unique molecular or genetic alterations, growing conditions, and mycoplasma testing [10]. Expanding the scientific interest toward other sources of biological material, it is important to consider the evolutionary process of biospecimen biobanks.

The first specimen biobanks started as university-based repositories for specific research projects. They were established by researchers with access to patient populations who took advantage of the availability of “left over” aliquots to be stored for immediate or future use. Samples were stored in one or a few freezers, and associated data were recorded in a laboratory notebook or basic database [11]. However, albeit these small collections lack the big data (clinical, genomic etc.), standard operative procedures and automated sample processing, they could be seen as the initial phase of the evolutionary process of human biobanking.

Over the time, technological advances, notably automated sample processing, computerization, and the advent of the World Wide Web revolutionized the management of biobanks, which developed into more complex units. According to The Research Centre Institute for Prospective Technological Studies of the European Commission (EUR 24361 EN-2010), 40% of European biobanks started during the 90 s, while 37% began after 2000. This trend coincided with completion of the human genome sequencing project, which led to genome-wide association studies focused on identifying disease susceptibility genes and diagnostic biomarkers. After this event, biobanking institution grew exponentially, assuming a unique role in contemporary scientific research. Indeed, different research programs have benefited from biobank specimens. One famous case is the development of the trastuzumab antibody (Herceptin). Starting from the evaluation of tumor specimens stored at National Cancer Institute’s Cooperative Breast Cancer Tissue Resource, it became one of the drugs effectively used to treat specific subtypes of breast cancer [12]. More recently, biobanks played a critical role in the development of Cancer Genome Atlas (TCGA), a publicly funded project that aims to create a comprehensive “atlas” of cancer genomic profiles by cataloguing major cancer-causing genomic alterations [13]. Thanks to human specimens with associated clinical data, it is possible to analyze large cohorts of over 30 tumors with large-scale genome sequencing. This approach has led to the identification of several novel molecular alterations in cancer, and tumor subtypes can be classified according to distinct genomic alterations, allowing a precision medicine approach for patient care.

Although the literature offers different classifications, a more general distinction can be made between population-based and disease-oriented biobanks. The first focuses on studying the possible future development of common and complex diseases, while the second is based on specific diseases, primarily cancer [4, 14, 15].

Population-based biobanks, such as Danish National Biobank [16], Estonia Biobank [17, 18], and UK Biobank [19], collect biological samples primarily from volunteer—without specific inclusion or exclusion criteria. The aim of population-based biobanks is to examine the role of individual genetic susceptibility and exposure to external factors in the development of specific diseases by combining molecular data with other associated data (clinical data, laboratory test results, questionnaire data, imaging data).

Disease-oriented biobanks were created to promote the study of human illness pathogenesis to identify possible therapeutic strategies. Thanks to the integration of different data encompassing a large number of biological samples, the research groups are able to develop large-scale research projects and collect information on the well-being of the examined population. In 1982 at the University of California, San Francisco (UCSF), an intervention strategy against AIDS was promoted with the establishment of the AIDS Specimen Bank. Researchers in epidemiology, infectious diseases, and pathology worked together to help discover the causative agent for AIDS with the goal of introducing new therapy protocols. The resulting network is now a major resource for investigators at UCSF and represents a global reference in this research field [20].

The advent of the World Wide Web considerably changed biobank administration, which opened new frontiers. By cataloguing extensive information about human samples, including annotation and location tagging, these virtual biobanks enable easy information sharing without the need to physically use biological samples, permitting simple sharing of medical data and allowing the development of networks for better cooperation between national and international biobanks [21]. Furthermore, this process negates the need to transport samples between two locations for a specific study, minimizing the risk of contamination.

While biobank foundation was an important medical research milestone, it is important to consider that their evolution has been decentralized. The different national directives established by local governance (data protection rules) [22] and, from a technical point of view, the different methods for collection/storage/qualification of basic data, have generated a high level of heterogeneity between biobanks [23]. These aspects could be potential barriers to the overall goal of an international research framework that aims to facilitate access to human biological materials. The life cycle stages of biosamples (intended as the collection, accession, acquisition, identification, preservation, long-term storage, quality control (QC), transport, and disposal of biomaterials) are a major source of heterogeneity. Result comparisons become difficult when different procedures for biosample management are applied. For this reason, harmonization of at least minimum standards for access and compensation need to be reached to facilitate large-scale, efficient use of human biological samples [24].

Qualification is the process used to verify the quality of biospecimens collected into a biobank. It is crucial for basic data and avoiding the introduction of uncontrollable variables that could make biospecimen samples incomparable in terms of quality. Considering that downstream analysis depends on the biobank research area (population-based or disease-oriented) and that different QC assays are used for biospecimen qualification, high heterogeneity between biobanks has developed over time.

There are two main approaches for qualification: the first includes the collection of biosamples with careful pre-analytical annotation (SPREC) [25], the second refers to retrospective collections with appropriate QC and sample qualification or quality stratification. The term “sample qualification” refers to the examination and validation of a single biospecimen or a collection on the basis of objective analytical parameters. “Quality stratification” refers to a process of examining and classifying biospecimens in different categories corresponding to specific in vivo parameters (e.g., protein content) or ex vivo pre-analytical conditions (e.g., pre-centrifugation conditions) [26].

Qualification methods and parameter measurements are strictly dependent on the downstream analysis they are collected for. Testing parameters of biosamples are specific for disease areas or for a specific downstream analytical platform, so not all parameters are tested everywhere, and parameter testing depends on the locally adopted standards. For example, a coagulation disease-oriented biobanking could test different parameters (e.g., fibrinogen, prothrombin, plasminogen activator inhibitor type-1) with different methods (clot detection, enzyme immunoassay [EIA], fluoro-immunoassay) for QC of plasma biosamples. Furthermore, different test methods entail different sensibility/specificity and threshold parameters, making data reproducibility difficult. Together with the different parameters that can be measured to qualify the same biosample typology, there are some measurement methods that are difficult to objectify, for example immunohistochemistry (IHC), immunocytochemistry (ICC), or microscopy for the qualification of tissue samples, cell suspensions, cell lines, or stem cells [27].

This paragraph aims at discussing some aspects related to the heterogeneity for collecting tissues, cells, blood, and nucleic acids.

The use of samples collected with standardized and validated protocols is a prerequisite to enable robust biological interpretation for data analysis and interpretation. For this purpose, the US National Cancer Human Biobank took a huge step forward in 2012 by introducing the first standard operating procedures (SOPs) for biobanks (http://​biospecimens.​cancer.​gov/​resources/​SOPs). These were introduced according to the understanding that a lack of standardized, high-quality biosamples has slowed the progress of clinical research, and they still serve as an example for biobanks worldwide. The introduction of these specific procedures raised the need to harmonize national and international procedures concerning biobanking; this improvement is still one of the main goals of biobanks worldwide.

The publication of ISO 20387 (ISO 20387:2018 “Biobanking—General requirements for biobanking”) can be considered an important milestone for procedure harmonization at international level. ISO 20387 represents a reference standard for biobanks to adopt common strategies for the organization and processing of biological samples to achieve minimum standardization requirements. The general purpose of the ISO 20387:2018 guideline is to make available biological material capable of guaranteeing the reproducibility and comparability of scientific research results by standardizing the life cycle stages of the biological materials. Its point-by-point indications provide specific tools related to policies, processes, and procedures covering the life cycle of biological materials and their associated data from collection to storage, reception and distribution, transport and traceability, preparation and preservation of biological samples, QC of processes, and method validation and verification.

In conclusion, while critical advances in health care have been achieved using biospecimens from biobanks, more qualified biospecimens are still needed for discoveries to promote translational research. In this context, it is critical to standardize the processes of quality assessment, consenting, sample collection, storage, and access. The implication is clear: if more well-characterized, high-quality samples are available, research will advance and impact health care delivery.

Cases related to human rights violations such as those mentioned in the Nuremberg process (1945) on criminal medical scientists have raised the need for international regulations on human research. In 1964, the Declaration of Helsinki guidelines introduced the principles that “research protocols should be always reviewed by an independent committee prior to initiation” (an Internal Review Board or IRB) and that “research with humans should be based on results from laboratory animals and experimentation.” In addition, IC was declared a right for any clinical research participant. The Declaration of Helsinki represented a milestone in the preservation of human rights (last revision, 2013) and provides general guidelines for medical research on humans [53]. It sets ethical principles including the importance of protecting the dignity, autonomy, privacy, and confidentiality of research subjects, as well as obtaining IC for the use of identifiable human biological material and data. However, starting from the last revision of Declaration of Helsinki (2013), clinical research infrastructure networks have been evolving, and there are growing needs to manage large collections of human samples and related data and plan new research strategies and predictive analyses [54].

Although biobanks and health databases have enormous potential, they are considered dangerous to human rights due to the accessibility of sensitive data. To face this issue, the World Medical Association (WMA) published the Declaration of Taipei to provide guidelines on the collection, storage, and use of identifiable data and biological material beyond the individual care of patients [55]. This declaration defines a biobank as “a collection of biological materials and associated data” and a health database as “a system for collecting, organizing and storing health information.” It describes human biological materials as samples obtained from living or deceased individuals that can provide biological information about a specific subject analyzed. The collections in health databases and biobanks are both obtained from individuals and populations, giving rise to similar concerns regarding dignity, autonomy, privacy, confidentiality, and discrimination. This document suggests that there must be harmonization between advancing scientific progress and protecting individuals’ rights related to data obtained from their biological samples. It is the first international guideline to provide ethical directions about the complex issues that arise with activities associated with human databanks and biobanks [56].

Appropriate IC is one of the most intensely discussed topics within the context of biobank research [57]. The goal is to enable a competent individual to decide whether to participate in a research program. This means that it is crucial to understand all that the consent entails to ensuring that the individual’s decision is effectively “informed” [58]. Unfortunately, there is no international consensus due to differences among the legal system of each country. In Europe, the General Data Protection Regulation (GDPR-2018) provided important indications about IC. This legal framework sets guidelines for the collection and processing of personal information of individuals within the European Union (EU). It defines consent as “an unambiguous indication of a data subject’s wishes that signifies an agreement to the processing of personal data relating to him/her whereby that consent needs to be given in clearly defined.” Article 7 describes the essential conditions regarding consent (to be valid). Principal points are:

Despite the GDPR2018 guideline, optimization of consent for future studies is an argument of intense debate in the biobanking research field [59]. Indeed, classical IC that is focused on a specific research project is considered insufficient in biobanking. To date, most biobanks adopted a “broad consent” model [59], which is the agreement for utilization of his/her sample for current or future studies within a specified framework without the need for contacting the patient. Broad consent provides researchers sufficient flexibility to pursue a wide range of future, scientific agendas, but it implies that the patient acknowledges that he/she will not receive any feedback on incidental findings. With the development of information technology tools, a novel consent model has been proposed termed dynamic consent. This type of consent requires tools for easily accessible and constant contact with the patient to manage reconsent for each new project. Dynamic consent enhances autonomy and helps meet the desire for increased user participation in research programs. Indeed, dynamic consent uses modern communication strategies to inform, involve, offer choices, and obtain consent for every research project based on biobank resources. This approach focuses on new possibilities for constant communication and has the potential to increasingly involve participants in the use of their biological samples [60], but it also reveals several weaknesses such as therapeutic misconception [61]. Comparing dynamic and broad consent, the latter still represents a good ethical solution for biobank research. Nevertheless, improvements are needed in the broad model, and criticism can be met with adapting some of the modern communication strategies proposed for the dynamic consent approach [62]. More research in this field is required, especially to test and improve documents used in the IC process. According to Bossert et al. [58] it is important to develop bioethics research studies to: (i) assess IC readability; (ii) evaluate readers’ understanding and reactions; (iii) use test results to revise the IC document, and, consequently; (iv) re-evaluate the revised document. We believe that considering reader feedback and applying this proposed research scheme could be a valuable strategy for the development and improvement of IC and other kinds of written information.

Research biobanks collect samples and data of human origin for their own research or that of third parties. Ownership of biological specimens is an important unsolved topic discussed in the literature. This argument presents ethical and legal issues in biobanking and, hopefully, all those involved in biobanking (patients/subject, researcher/clinicians, and institutional infrastructures) should be considered. Starting from the worldwide-recognized concept that no person can own another person, as this would constitute slavery and violate article 4 of the Universal Declaration of Human Rights [63], several biobanks have agreed to be custodians instead of owners of biological samples [64]. This statement is in line with that of International Agency for Research on Cancer (IARC) declaring, as a general rule, “no ownership of biological samples exists, and the biobank should assign ownership or custodianship based on national and institutional guidelines.” However, it is important to consider that local scientists who contribute to the biobank by collecting the biological samples could view the samples as “theirs.” Indeed, principal investigators may ask for priority access to specific collections or impose co-authorship in publication or veto rights to maintain a degree of control over biosamples due to scientific competition. This argument remains a theme of intense debate in the scientific literature and will be re-discussed in “Issues for sample access”.

Biobank sustainability is another crucial aspect to be explored. A research project usually involves a study on a specific topic, which foresees predictable experimental costs, predetermined procedures, and a finite period of time. Conversely, biobanks have costs that are not fully definable and a financial management system that evolves over time. Biobanks are costly in regards to staffing, equipment, service contracts, consumables, and expertise [65]. To maintain sustainability, biobanks could run as business units. Indeed, some biobanks leverage the financial potential of their specimens and data, but this may lead to ethical and legal issues [66]. International biobanks have agreed that the human samples stored cannot be used for commercial purposes [67].

Most biobanks implement a cost recovery system by charging investigators access to specimens and related data. Several biobanks have developed self-financing strategies; for example, the UCSF AIDS Specimen Bank (ASB) developed a recharge methodology to recover costs associated with sample processing, storage, consumables, software and hardware maintenance, data management, and data sharing (https://​cfar.​ucsf.​edu/​cores/​asb/​services). Other examples of biobanks that use cost recovery strategies are: the Wales Cancer Bank (https://​www.​walescancerbank.​com/​cost-recovery-fees.​htm), Australian Prostate Cancer BioResources (https://​www.​apcbioresource.​org.​au/​for-researchers/​apply-for-tissue-access/​), and the Ontario Tumor Bank (https://​ontariotumourban​k.​ca/​researchers/​access-fees).

Despite adopting this financial management system, the majority of biobanks are unable to fully recover their costs and have relied on the support of government policies and donations. A possible solution could be supporting biobanks mostly by local public bodies and partially by a cost recovery mechanism in order to favor efficient sustainability [68].

Research infrastructures are crucial for scientific research and are an important element of science policy. Biobanks play a central role in an increasingly complex research environment and are required to develop and implement accountable and transparent management procedures. Moreover, biobanking can be considered as an important transition from local research tools to complex international research infrastructures. Coordinating this process across countries is a critical process.

The coordination and structural components of biobanking infrastructures are diverse, with different goals, governance, and structures. Generally, biobanks are organized to acquire baseline and follow-up data of participants in cohorts classified by clinical features (e.g., pathology, disease stage, therapy, etc.), as well as sexual habits, work conditions, exposure to pollutants, and lifestyle habits. Moreover, data associated with samples are not limited to the moment of biospecimen collection but may be updated continuously by allowing its use in future studies. These considerations should motivate biobanks to have a dynamic organization. One of the primary needs in this field is the national and international standardization of procedures and management. In this context, The International Society for Biological and Environmental Repositories (ISBER) founded in 2000 (http://​www.​isber.​org/​) was one of the first and most important international organizations devoted to addressing the harmonization of scientific, technical, legal, and ethical issues relevant to repositories of biological and environmental specimens. ISBER was created by researchers, biobank managers (including those working with human, animal, plant, and environmental repositories and directors), National Institutes of Health representatives, bioinformatics managers, patient advocates, lawyers, and others interested in biobanking to share knowledge in the field. Indeed, ISBER fosters collaboration; creates education and training opportunities; provides an international showcase for state-of-the-art policies, processes, and research findings; and facilitates innovative technologies, products, and services. Each year, ISBER promotes new working groups to address areas such as biospecimen science, IC, informatics, rare diseases, and automated repositories. Moreover, they discuss important topics such as cost recovery, training, facilities, equipment, safety, quality assurance, QC, shipping, ethical issues, and specimen collection, processing, retrieval and culling [70].

After a preparatory phase project in 2008 [71], the European Biobanking and BioMolecular resources Research Infrastructure (BBMRI)-European Research Infrastructure Consortium (ERIC) was activated in 2011. The BBMRI-ERIC plays a critical role in health research and is composed of 19 European member states and 1 international organization, that is the IARC [71]. The BBMRI infrastructure was designed to operate across European countries. The goal is to link biobanks across Europe to foster cooperation and research that benefits both patients and European citizens. Indeed, it aims at improving interoperability of the existing population-based or clinically oriented biological samples from different European populations or patients with rare diseases. These collections include data on the health status, nutrition, lifestyle, and environmental exposure of study subjects. To fully realize the enormous potential of European biobanks for the benefit of society, the Stakeholders’ Forum was formed as a platform for clinical, ethical, and legal experts; the biotech and pharmaceutical industries; and patient advocacy groups. According to their suggestions, The BBMRI-ERIC seeks to develop standards and guidelines that properly respect individual values, such as protection of privacy and IC, with shared values of facilitated access to progress in healthcare and disease prevention [72]. According to Mayrhofer et al. the ultimate goal is that the BBMRI-ERIC will impact partnerships with patients/donors/participants, who know that their own tissues, samples, and personal data can yield discoveries and advances in medicine, diagnostics, and therapies [72]. In return, the BBMRI-ERIC is responsible for ensuring that the samples and data entrusted for research are used in the best way possible to advance knowledge and improve health care.

Biobanks are heterogeneous in their design (population or disease oriented) and use (epidemiology, translational, pharmaceutical research). They may contain data and samples from family studies, patients with a specific disease, clinical trials of new interventions, or they may be part of large-scale epidemiologic collections. Inevitably, data and samples are collected under various conditions and standards and for different purposes. Given the challenges of data collection and sample storage within studies, there has been little standardization across biobanks. A number of international initiatives to provide guidance and protocols were recently developed to address this issue [73]. The goal of procedure standardization and harmonization is to facilitate data sharing among different resources, thereby increasing effective sample size and statistical power, especially for rare diseases [73]. Quality programs need to be implemented to minimize the impact of variability on the integrity of the samples and, where possible, consideration should be given to future proofing the collection. In this context, in 2014, the ISO/Technical Committee (TC) 2761/Working Group (WG) 2 Biobanks and Bioresources was formed to ‘‘elaborate a package of International Standards in the Biobanking field, including human, animal, plant, and microorganism resources for research and development, but excluding therapeutic products.’’ The ISO/DIS 20387 was intended to be applicable to “all organizations performing biobanking activities, including biobanking of human, animal, plant and microorganism resources for research and development.” It provides a set of requirements to “enable biobanks to demonstrate that biobanking entities operate competently and are able to provide biological resources (biological material and associated data) of appropriate quality.” These requirements include personnel competence, method validation, and QC. It also includes general requirements for the competence, impartiality, and consistent operation of biobanks including QC to ensure that biological material and data collections are of appropriate quality. The ultimate aim of biobanking is to build and run quality-controlled storage facilities and infrastructures to enable future biomedical research. A proposed workflow model for the collection, storage, and distribution of biological specimens in biobanking is displayed in Fig. 2. Briefly, biological specimens are accepted and coded, and recorded on appropriate management software along with corresponding clinical information from patients. Successively, robotic process automation (Fig. 2a) can be used for sample aliquoting (preferably micro-aliquoting) before storage at low (− 80 °C) or ultralow (liquid nitrogen vapor phase, − 150 °C) temperature (Fig. 2b). The biobank software program should be able to manage biological samples and related information. To fulfill this purpose, it should have three basics features: (i) biological specimen management (consent management; non-conformity management; and biological resource history related to patients, samples, aliquots, derivatives, storage, and request management); (ii) traceability (follow-up for sample requests, annotation of collections with links to patient files [clinical, genetic, imaging data]); and (iii) interoperability (interface with temperature monitoring systems, interface with the hospital/laboratory to obtain further clinical data). Finally, the application of a material transfer agreement regulates biomaterials transfer between biobanks and recipients (research groups internal or external to the biobank infrastructure) to maintain specific quality and traceability standards for the samples [74] (Fig. 2c).

Human biological materials and related data stored in biobanks are valuable resources for biomedical research. Transparent, effective, and efficient governance structures and procedures for access, compensation, and priority setting are needed, but recent debates highlight challenges in their practical application [75]. Access to biological samples and related data have been much debated in recent years; indeed, access policies are neither standardized nor harmonized. Langhof et al. [76] recently highlighted the need for governance structures accepted by all stakeholders (patients/donors, researchers, research funders, public, and others) to ensure appropriate sample access for research. The authors proposed specific topics based on interviews with biobank experts on the BBMRI-ERIC infrastructure. Regarding biological sample sharing, they have raised different aspects that are debated but not yet harmonized, among the main ones:

The growth of imaging biobanks is linked to the capability of high-throughput computing to extract numerous quantitative features from bioimages obtained with advanced CT, MR, and PET [81, 82] acquisition technologies. This promising area of research is defined as radiomics and focuses on evaluating extracted features as novel IBs for assessing physiological or pathological processes as well as pharmaceutical responses to a therapeutic intervention [83]. Some imaging features are already routinely used for patient monitoring in oncology, including tumor volume measurement, perfusion grade, Brownian motion of water molecules within a voxel of tissue, texture analysis, MR spectroscopy, stiffness, glucose metabolism, CT density, MR signal intensity, and MR fingerprinting. IBs could be very useful for patient management since they are non- or minimally invasive and could decrease the need for more invasive procedures like biopsy. Generally, IBs are considered as the expression of bio-signals because they are extracted from the analysis of an electromagnetic, photonic, or acoustic signal emitted by the patient. In this way, it is possible to consider IBs as a unique expression of a disease phenotype. According to Neri et al. imaging biobanks need to include data, metadata, raw data, measurements, and biomarkers derived from image analysis to allow feature extraction [80]. In addition, other disease-related factors (e.g., patient prognosis, pathological findings, genomic profiling, etc.) should be included for full validation and qualification in routine research or clinical use. This is important considering what was proposed in 2015 by the European Society of Radiology [84], which defined imaging biobanks as “organized databases of medical images, and associated IBs, shared among multiple researchers, and linked to other bio-repositories,” and suggested that “biobanks (which focus only on the collection of genotype data) should simultaneously come with a system to collect related clinical or phenotype data”. This definition highlights the roles that imaging biobanks play in ensuring standardization of data acquisition and analysis, accredited technical validation, and the transparent sharing of biological and clinical validation data. These steps are critical for effective translation of an IB into clinical practice according to Bonmatí et al. [85]. They highlight how clinical validation represents the bottleneck that selects useful and useless IBs. The awareness of imaging biobanks importance is gradually increasing, especially in the oncologic community, because they could be critical for crossing the translational gaps necessary to apply novel IBs in clinical practice [86].

To date, most existing imaging biobanks have been oncologic oriented because IBs developed in diagnostic imaging are clinically used for oncologic purposes (e.g., glucose metabolism, tumor volume) [86]. An example of an oncological imaging biobank is the US-based Cancer Imaging Archive (TCIA), developed by the Frederick National Lab for Cancer Research [87]. TCIA is a service in which de-identified medical images from cancer patients are hosted in a large archive. The data are organized as “Collections” in which patients share a disease (e.g., lung cancer), an image modality (MRI, CT, among others), or a research focus. Digital Imaging and Communication in Medicine (DICOM) is the primary file format used by TCIA for image storage. Supporting data related to the images, such as patient outcomes, treatment details, genomics, pathology results, and expert analyses are also provided when available [87]. One major feature of TCIA is the open access; DICOM datasets and other “-omics” information are available for public download. Researchers can download datasets for many purposes, even to develop and test new IBs. There is currently no European counterpart to TCIA, but it would be desirable for imaging biobanks in Europe to become public to allow data sharing and stimulate the exploitation of imaging data from researchers. To this aim, it will be important to develop procedures to harmonize instruments for data collection and mining and perform comparative analyses.

The latest goal of multi-omics biobanks is integrating imaging and biological data to provide a deep association between phenotype and genotype with possible IBs. To obtain such integration, IB data from different “systems” should be in manageable formats (e.g., standardized quantitative MRI and PET protocols and measures, obtained and combined by separate or concurrent acquisitions) to be integrated with other “-omics” data, including the genomic profile of the same patient [88]. In this context, the bioinformatics approach plays a pivotal role in analyzing, correlating, and interpreting a large amount of data that cannot be managed with only on human capabilities. Based on these considerations, the immediate purposes of imaging biobanks are to enable generation of IBs for use in research studies and support biological validation of existing and newly identified IBs. Thanks to the development of imaging biobanks and the technological tools required for radiomic analyses, it is possible to integrate radiomic features with genetic data in a novel research approach that is defined as radiogenomics. This term applies in two different fields of study: “radiation genomics” and “imaging genomics.” Radiation genomics refers to the study of genetic variation (e.g., single-nucleotide polymorphisms) in relation to a cancer patient’s risk of developing toxicity following radiation therapy. It is also used in the context of studying the genomics of tumor response to radiation therapy. Imaging genomics refers to the extraction of IBs that can identify disease genomics, especially cancer without a biopsy sample [89]. Radiogenomics represents the evolution of radiology–pathology correlation from an anatomic–histologic level to a genetic level [90]. Imaging features are correlated with genomic data often obtained through high-performance molecular techniques such as NGS technologies, DNA sequencing, and microarray. Radiogenomics can better characterize tumor biology and capture intrinsic tumor heterogeneity with relevant implications for patient care. Existing radiogenomics studies were mainly concerned with oncologic diseases such as glioblastoma multiforme [91], lung cancer [92], prostate cancer [93], and breast cancer [94]. In this context, the increasing use of bioresources from certified biobanks associated with imaging will allow the merging of these two data types to improve patient management and provide personalized medicine as hypothetically conceived in Fig. 3. In this scenario, the use of hybrid biobanks that integrate imaging and molecular data is an interesting potential approach to improve clinical management on a personalized background.