Assays for Qualification and Quality Stratification of Clinical Biospecimens Used in Research: A Technical Report from the ISBER Biospecimen Science Working Group

This technical report presents quality control (QC) assays that can be performed in order to qualify clinical biospecimens that have been biobanked for use in research. Some QC assays are specific to a disease area. Some QC assays are specific to a particular downstream analytical platform. When such a qualification is not possible, QC assays are presented that can be performed to stratify clinical biospecimens according to their biomolecular quality.


Introduction
C linical biospecimens used in research are subject to two types of laboratory analyses. The first of these is the analysis of established clinical biology/pathology parameters where reference ranges are usually known and methods are validated (e.g., CLIA or ISO15189 accreditation). Results of these analyses are necessary to support any research on novel clinically relevant biomarkers (definition of true positive and negative cases, use as a reference method). The second type is analysis of research parameters where there are usually no established reference ranges, and often methods are not validated by the laboratory as extensively as clinical biology/pathology methods. 1 Results of these analyses are used to discover novel clinical endpoint correlates (biomarkers).
In vivo and in vitro pre-analytical variations have a more or less significant impact on the output of analyses, depending on the biospecimen type, the pre-analytical variable, and the analyte of interest. According to the type of analysis above, the word ''significant'' has a different meaning. In the first typethe analysis of clinical biology/pathology parameters-''significant'' means clinically consequential at the diagnostic level. In the second type-analysis of research parameters-''significant'' means statistically significant. Examples illustrating this concept are shown in Table 1.
In some cases, the impact may be molecule-and even epitope-specific, for example tissue ischemia time may influence specific phospho-epitopes differently. A standard biospecimen research experimental protocol has been proposed for this type of research. 2 Therefore, in all research comparing different groups of samples for biomarker discovery, it is critical that all samples are of comparable quality to avoid the introduction of uncontrolled variables and increase the power of analysis of biomarkers. There are two approaches to this end: either sample collections with careful pre-analytical annotations (SPREC), 3 or retrospective collections with appropriate quality control (QC) and sample qualification or quality stratification. A combination of the two approaches to control compliance of procedures with specified SPRECs is also possible.
Biobanks underpin all three layers of biomarker discovery, validation, and use in clinical practice. In the biomarker discovery phase, biospecimens collected and processed with one Standard Operating Procedure (SOP), and corresponding to one quality category, should be used in order to avoid preanalytical bias and increase the power of research. However, in the biomarker validation phase, biospecimens collected and processed with more than one known and documented SOPs and corresponding to more than one quality category should be used in order to validate the robustness of a bio-marker to relevant pre-analytical variations. Finally, in the biomarker clinical implementation phase, biospecimens collected and processed via validated SOPs should be used in order to ensure successful and accurate clinical diagnostic results. For these reasons, during recent years, biobank managers, auditors, and funding bodies have been asking what assays can be performed in order to assess the quality of biospecimens objectively. This technical review provides answers to this question. Although gaps exist, this review shows that many tools are already available and can be used for specimen qualification.

Methods
For the purposes of this technical report, the members of the International Society for Biological and Environmental Repositories (ISBER) Biospecimen Science Working Group held face-to-face meetings and teleconferences between 2013 and 2015. The chair of the Working Group performed a thorough literature review and compiled a list of relevant and effective QC attributes for different categories of biospecimens. This list was reviewed and complemented by members of the Working Group. When the information is based on published evidence, the corresponding reference is given. When no reference is given, the information corresponds to current practice or to the corresponding author's opinion.
The following definitions were used: Biospecimen: any biological specimen, which may be a: B Primary sample: specimen directly collected from the donor (e.g., whole blood, urine, solid tissue); B Simple derivative: sample prepared through a simple laboratory manipulation (e.g., after centrifugation of collection tubes or mechanical disruption of tissues) without the addition of chemical substances, and without cell disruption or cell selection as part of a multi-step process; or B Complex derivative: derivative whose isolation requires usage of multiple steps and/or addition of chemical substances (e.g., nucleic acids, proteins, lipids, sorted cells, cultured cells, immortalized cells). Qualification: process of examination of a biospecimen or a collection of biospecimens, and verification, based on objective analytical evidence, of their suitability for research use, either in a specific disease area or on a specific downstream analytical platform. Quality stratification: process of examination of a biospecimen or a collection of biospecimens, and their classification, based on objective analytical evidence, into distinct categories, each category corresponding to a specific in vivo biological characteristic (e.g., level of inflammation, % tumor, protein content) or to a specific ex vivo pre-analytical condition (e.g., pre-centrifugation conditions). Biomolecular integrity: quality status of a biospecimen, reflecting whether biomolecules of interest have not un-dergone either statistically or clinically significant changes relative to their in vivo state/levels. Commutability: equivalence of analytical methods, based on objective evidence.
The term ''qualification'' is used qualitatively. Therefore, a biospecimen is or is not qualified for use in research in a specific disease area or on a specific analytical platform.
The term ''quality stratification'' is used quantitatively. Therefore, one or more thresholds apply in order to stratify biospecimens in two or more quality categories. These quality categories correspond to defined in vivo or in vitro conditions. When qualification is not possible because of lack of relevant assays, then quality stratification can be made. In some cases, qualification can be achieved for biomarker research in a specific disease area (Table 2) or on a specific downstream analytical platform. For primary samples, qualification depends on their biomolecular integrity. For simple or complex derivatives, qualification depends both on the biomolecular integrity of the primary sample from which the derivative has been extracted and on the efficiency/performance of the extraction, culture,    cryopreservation, or other laboratory manipulation (e.g., cfDNA from plasma; Fig. 1).

Results
The results are presented in the form of Tables for fluid  (Tables 3 and 4), tissue (Tables 5 and 6), and cytological biospecimens and their derivatives. Table 2 includes information on QC measurands for qualification for use of samples in specific disease areas. 4,5 The measurands in this table are molecules that are recognized biomarkers in the respective disease areas and are also known to be labile. Detection of the measurand above the method's level of detection is necessary (though not always sufficient) for qualification of a sample. As an example for reading Table 2, if Ab42 is undetectable in CSF samples, then these samples cannot be qualified for research in the area of neurodegenerative diseases.
Tables 3, 5, and 7 include information that can be used for the qualification of fluid, tissue, or cytological specimens, respectively, in the scope of different types of downstream analyses. In these tables, ''qualification parameter'' is the quality aspect of the biospecimen that is being evaluated; ''measurand'' is the molecule, or the morphological or functional characteristic that is being measured and whose positive or negative result is necessary for the qualification; ''scope of qualification'' is the type of downstream analysis for which the biospecimen is being qualified as fit-forpurpose; and ''measurement method'' is the type of method that is used to measure the measurand.
Tables 4, 6, and 8 include information that can be used for the quality stratification of a fluid, tissue, or cytological biospecimen, respectively. In these tables, ''qualification parameter'' is the quality aspect of the biospecimen for which the biospecimen is being stratified; ''measurand'' is the molecule, or the morphological or functional characteristic that is being measured and whose level is used to stratify the biospecimens in categories; ''quality stratification thresholds'' are the levels of the measurand, which are critical for the quality stratification; and ''measurement method'' is the type of method that is used to measure the measurand. The quality stratification thresholds listed in Tables 4, 6, and 8 classify the biospecimens into the categories of the qualification parameter given. The ''time xxx/ temperature yyy'' categories correspond to available experimental data, but they should be understood as ''time xxx/temperature yyy or equivalent conditions.'' The quality stratification thresholds listed in Tables 4, 6, and 8 are those corresponding to the measurement methods described in the references. Application of a threshold with a measurement method that is different from the method that has been used for the establishment of the threshold requires previous demonstration of the commutability of the methods.

Tissue type specificities
Assays for tissue qualification or quality stratification may be tissue type-specific. Some examples are given below. Fixation conditions have a significant impact on P-Akt and P-Erk1/2 in breast cancer tissue. 35 Ischemia has a significant impact on estrogen and progesterone receptors in breast tissue. 36,37 A Tissue Quality Index has been proposed for formalin-fixed, paraffin-embedded breast tissue in order to assess its cold ischemia time by immunohistochemistry. 38 Stathmin 2-20 has been proposed as indicator of degradation in brain tissue by matrix-assisted laser desorption/ionization time of flight mass spectrometry. 39 AKT-P has been proposed as indicator of postmortem conditions in brain tissue by western blot. 40 Superoxide dismutase in the liver and peptidyl-prolyl-cis-trans isomerase and insulin C-peptides in the pancreas have been associated with postmortem delay and assessed by two-dimensional difference in gel electrophoresis. 41

Discussion
This article proposes a biospecimen QC strategy, based on current state of knowledge, in the form of summary tables (Fig. 2).
The qualification and quality stratification assays presented in this technical report do not aim for an absolute assessment of the quality of samples, since a sample can be of high enough quality (fit-for-purpose) for one type of analysis (e.g., antibody analysis), but not for other types of analyses (e.g., metabolite analysis). Therefore, scientists should devote time and effort to understand and define what sample quality is needed to obtain consistent results with a given downstream analytical platform. As can be seen from Tables 3, 5, and 7, there are several gaps in the area of biospecimen qualification for use on specific analytical platforms. These include, for example, urine, saliva, or frozen tissue qualification for use in proteomic analyses, serum, plasma, or other body fluid qualification for use in miRNome analyses, or DNA qualification for use in methylation analyses. In the absence of such knowledge, this technical report offers a strategy for sample quality stratification so that bias due to samples of inconsistent quality levels can be minimized.    The information provided in this report is important because its application will enable and support bioprocessing method validation by providing relevant readouts (measurands); assessment of the quality of biospecimens of unknown history; biomarker discovery by ensuring use of qualified biospecimens or biospecimens belonging to a specific quality category; validation of biomarker robustness by using quality-stratified biospecimens belonging to different, defined quality categories; implementation of novel biomarkers in clinical practice; and characterization and production of clinical reference materials.
For the above purposes, QC measurands of clinical biospecimens can be assessed either by the biobanks themselves, or by subcontractors/collaborators who are accredited or successfully participate in relevant Proficiency Testing schemes. The results of the QC can be used by biobanks for qualification of legacy collections (the definition of cutoff values for acceptance of legacy collections or specific samples can be made and disclosed by the biobank), by end users for stratification of samples of different origins, or by funding agencies for assessment of the fitness for purpose of collections to be used in the context of grant allocation.
Author Disclosure Statement F.B. is listed as co-inventor in patent no. 0704237 and in the filed patent 15195301.5-1408 (on sCD40L and LacaScore, respectively).