surrogate sample - mdic

For Public Comment Period: April 5, 2017 – June 30, 2017 1

2

Surrogate Sample DRAFT Framework for Public Comment

Clinical Diagnostics Surrogate Sample Working Group 4-5-2017

MDIC Clinical Dx Surrogate Sample Draft Framework

20170405 MDIC Clinical Dx Surrogate Sample Draft Framework for Public Comment Page 1

Introduction 3

The efficiency, predictability, and cost of in vitro diagnostic (IVD) development directly impacts patient access 4 to IVDs. The collection of human specimens for testing is a central aspect of IVD validation. When specimens 5 are difficult to obtain or the use of specimens introduces variables into the evaluation of analytical or clinical 6 performance, surrogate samples may be used to improve the timeliness and efficiency of IVD validation. 7 8 The use of surrogate samples as an element of IVD analytical and clinical studies is not new. A review of 9 historical data, knowledge, and experience of IVD innovators demonstrates surrogate samples play an 10 important role in the validation of IVDs1. However, the lack of a common language and framework regarding 11 the use of surrogate samples in IVD performance studies reduces the predictability, timeliness, and efficiency 12 of IVD development, thus increasing cost and reducing patient access. The Surrogate Sample Framework sets 13 forth a common language and methodologies for the use of surrogate samples in IVD analytical and clinical 14 performance studies. 15 16 17 Reasons to use Surrogate Samples 18

IVD innovators rely upon surrogate samples for a variety of reasons. Surrogate samples provide an option 19 when naturally occurring specimens of biological origin are rare or difficult to acquire. For example, for 20 quantitative tests, acquiring a sufficient number of specimens across the measurement interval of a test, and 21 especially at the high end, the low end, or medical decision points may prove challenging when there is a low 22 prevalence of such specimens in the general population. Surrogate samples can help address these challenges. 23 24 Ethical considerations, such as invasive sampling methods or drawing large volumes of clinical specimens from 25 sick patients, are another reason to use surrogate samples. Additionally, surrogate samples may simplify 26 sample acquisition and characterization activities, as well as meet scientific demands of studies requiring large 27 volumes of a sample. In summary, surrogate samples are used for a variety of reasons including, but not 28 limited to: 29 30

Inadequate stability of naturally occurring specimens 31

Minimizing biological variability to meet the scientific rigor demanded by the study 32

Analyte does not naturally occur at the level demanded by the study, for example, at the extremes of 33 the measurement interval of a test or at the medical decision level (MDL) or test cut-off 34

Insufficient volume of naturally occurring specimens 35

Covering the measurement interval of a test 36

Supplementing the number of positive samples to complete studies 37 38

39

1 See Survey Report at www.mdic.org/clinicaldx/survey/ and Submission Analysis Report at www.mdic.org/mdicx/#archive



Terminology 40

The authors of this paper elected to use the phrase “surrogate sample” to describe the samples that are the 41 subject of this Framework. A review of relevant IVD and medical device vocabularies (e.g., JCGN, ISO, CLSI) 42 revealed that a common definition of “surrogate sample” does not exist. Thus, the authors chose to define the 43 term, adopting the following definition: 44 45

Surrogate Sample: material used as a substitute for body fluid or tissue taken for examination from a 46 patient. 47

48

Examples include, but are not limited to: 49

Materials supplemented (spiked) with an analyte of interest 50

Pooled specimens of biological origin 51

Material created to have functional characteristics of a body fluid or a tissue 52

Material comprised of a combination of an analyte that emulates the analyte of interest and a material 53 created to have the functional characteristics of a body fluid or tissue 54

55 56 Within this broader definition, surrogate samples are further characterized as: 57 58 Supplemented: Individual specimens of biological origin spiked with an analyte of interest 59 60 Pooled: Combined individual specimens of biological origin that may or may not be spiked with 61

an analyte of interest 62 63 Simulated: Treated materials of biological origin or artificial materials created to have functional 64

characteristics of body fluid or tissue that may or may not be spiked with an analyte of 65 interest 66

67 68 69 Terms such as contrived, altered, supplemented, or simulated are frequently used within the industry 70 interchangeably with the term surrogate or as a subcategory of samples within the broader category of 71 surrogate samples. As observed by the authors, the use of different terminology can impede the development 72 of sound, scientific strategies for the use of surrogate samples. Therefore, we recommend adoption of the 73 above definition of surrogate sample by the industry, including recognition and use in standards, guidance, and 74 other relevant vocabularies. 75 76 For purposes of this Framework, we have opted to use the more commonly used term “analyte,” as defined in 77 ISO 17511, rather than term “measurand.” Additionally, we have decided not to use the term “commutability” 78 to describe a demonstration of comparable performance between a surrogate sample and specimen. 79



Commutability has a distinct definition in the industry2 and has been deemed by us to be more relevant to 80 studies of method comparisons rather than for establishing validity of results from patient specimens by a 81 specific test. Thus, we have not used the term “commutability” in the context of surrogate samples for 82 purposes of this Framework. 83 84 We refer readers to the glossary for definition of other terms used in this Framework. Relevant IVD and 85 medical device vocabularies (e.g., JCGN, ISO, CLSI) have been used to define terms. 86 87 Framework Organization 88

This Framework begins with general principles applicable to surrogate samples and discussion of a general 89 surrogate sample hierarchy. It is further organized by performance study type. Each section expands upon the 90 use of surrogate samples and related points to consider that are unique to the study. The reader is advised to 91 consider the objectives of each study type, and recognize that what may be a suitable surrogate sample type 92 for one performance study is not necessarily suitable for another study. 93 94 95 General Principles for Design and Selection of Surrogate Samples 96

1. Identify the objective for using surrogate samples. 97

a. Supplement specimens 98

b. Replace specimens 99

2. Identify critical factors related to the performance study type, such as: 100

a. Preserve biological variability 101

b. Remove or reduce biological variability 102

c. Meet scientific demands of study type requiring large volumes of sample 103

d. Introduce a variable that can only be studied in a surrogate setting to understand how the test 104

performs under the surrogate conditions. 105

3. Consider factors based on whether the test is qualitative or quantitative 106

4. Consider the effect of extraction method, if applicable, in increasing or decreasing biological 107

variability 108

5. When selecting the surrogate sample matrix consider which functional characteristics of the naturally 109

occurring specimen are most relevant for the study type. Consider whether there are aspects of the 110

surrogate sample that would interfere or otherwise influence test performance in a way that is 111

2 CLSI. Characterization and Qualification of Commutable Reference Materials for Laboratory; Approved Guideline. CLSI

document EP30-A. Wayne, PA: Clinical and Laboratory Standards Institute; 2010.



inconsistent with naturally occurring biological specimens. When the matrix is artificially prepared, 112

describe the characteristics of the naturally occurring specimen that are represented or mimicked by 113

the surrogate sample. Consider factors to minimize differences between specimens and surrogates, 114

such as viscosity or endogenous cellular components. 115

6. Consider existing CLSI guidelines regarding the use of surrogate samples specific to the study type and 116

FDA guidance regarding the use of surrogate samples specific to the analyte type. 117

7. Keeping the above principles in mind, consider a hierarchal approach to the selection of a surrogate 118

to identify a surrogate similar to the specimens that will be used in clinical testing. 119

When constructing surrogate samples, it may be necessary to employ special methodologies. For example, 120 when formalin-fixed paraffin embedded (FFPE) tissue is the specimen type, as might be used for cancer 121 diagnostic tests, DNA or RNA containing the mutations, insertions, deletions, or sequences of interest may be 122 added/spiked into the samples. This could include cell line DNA, plasmids, or transcripts cell lines that are 123 formalin fixed and paraffin embedded. For swab specimens, addition of the analyte directly to the swab may 124 not be feasible. In these cases, the analyte may need to be added to an appropriate liquid matrix in which 125 various dilutions are then coated onto the swab, or surrogate swab samples may be created by dipping a swab 126 in a solution or collection media spiked with the analyte of interest. 127 128 While the use of surrogate samples is well-established for certain technologies, analytes, or studies, there are 129 other areas where the use has been less frequent. In these areas, a demonstration of functional 130 characterization may prove useful in providing scientific evidence to support the use of a particular surrogate 131 sample type. 132 133 134 Surrogate Sample Hierarchy 135

CLSI Guideline EP06-A “Evaluation of the Linearity of Quantitative Measurement Procedures: A Statistical 136 Approach; Approved Guideline” establishes a hierarchy to consider when selecting surrogate samples for 137 linearity performance evaluations. We believe this hierarchy provides a reasonable path for consideration for 138 all other study types as well. In concert with three categories of surrogate samples, identified above—139 supplemented, pooled, and simulated—and consideration of CLSI EP06-A, we have constructed a general 140 surrogate sample hierarchy to serve as a starting point for all study types. Therefore, after considering the 141 general principles and the performance study specific objectives, evaluate the hierarchy and select the best 142 option for surrogate sample preparation and use. Refer to CLSI guideline EP06-A for specific elements to 143 consider when preparing surrogate samples within the hierarchy. 144



145

Figure 1 146

147



Performance Study Types 148

In addition to the general principles, there are factors to consider in conjunction with specific performance 149 study objectives. Additionally, application of the hierarchy varies with specific performance study objectives. 150 This Framework outlines these factors and the application of the surrogate sample hierarchy by performance 151 study type. 152 153 Based on our research3, the use of surrogate samples has been more prevalent in certain study types. 154 Whereas in other study types, the use has been more situation dependent4. This Framework reinforces 155 common usage embodied in standards and practice. Further, it is designed to formalize collective scientific 156 knowledge through the use of uniform principles and terminology for those study types where surrogate 157 sample use has been more prevalent. For those study types, in which surrogate sample use has been more 158 situation dependent, this Framework sets forth practical applications aligned within the general principles and 159 surrogate sample hierarchy. 160 161 This Framework does not provide comprehensive guidelines on study design, unless relevant to the use of 162 surrogate samples. Readers are referred to scientific standards, guidance documents, and guidelines for 163 factors to consider when designing performance studies. 164 165 The following performance studies are addressed in this Framework: 166

1. Linearity/Assay Analytical Measuring Interval 167

2. Analytical Specificity/Interference/Cross-Reactivity/Competitive Inhibition 168

3. Precision/Reproducibility 169

4. Detection Limit 170

5. Matrix Comparison 171

6. Method Comparison 172

7. Specimen Stability 173

8. Reagent Stability 174

9. Instrument Carry-Over Studies 175

10. Clinical Performance 176

177

3 See Survey Report at www.mdic.org/clinicaldx/survey/ and Submission Analysis Report at www.mdic.org/mdicx/#archive 4 See Survey Report at www.mdic.org/clinicaldx/survey/ and Submission Analysis Report at www.mdic.org/mdicx/#archive



Linearity / Assay Analytical Measuring Interval 178

Study Type 179 Linearity/assay analytical measuring interval is the range within a test providing results that are directly 180 proportional to the concentration of analyte in the test sample. 181 182

Objective 183

The objective of this study is to determine the range across which the average value of test replicates is 184 proportional to the true value of the concentration of the analyte. Increasing the number of replicates at 185 each analyte concentration or increasing the number of analyte concentrations across the target range can 186 be used for more definitive examinations of linearity. 187 188

Principles of the Use of Surrogate Samples 189 Surrogate samples are widely used in linearity/assay analytical measuring interval studies to produce 190 samples in which there is a known proportional relationship in a representative matrix and that span the 191 desired range of the test. The preferred type of surrogate sample is a high concentration specimen diluted 192 with a negative patient sample to achieve the known proportional relationship of the analyte 193 concentration. Sometimes this will not be possible and the surrogate specimen hierarchy should be 194 followed, selecting the most preferred surrogate sample to help ensure the surrogate sample is 195 representative. 196 197 How to Use 198 To provide sufficient volume of sample with an analyte concentration at different levels across the test’s 199 measuring interval, surrogate samples can be prepared in a number of ways, including: dilution of a 200 specimen with a high concentration of analyte with a matrix that has a very low amount of analyte; 201 dilution of pure analyte with a matrix that has a very low amount of analyte; dilution of a specimen or pure 202 analyte in a surrogate matrix. Preservation of patient-to-patient variability is a key consideration. 203 204 When to Use 205 Because the linear range experiment requires enough samples to prepare dilutions across the linear range, 206 the use of surrogate samples is generally accepted. In all cases the surrogate samples should be 207 representative of a specimen. 208 209 Additional Considerations 210 When using surrogate samples for linearity/assay analytical measuring interval preservation of biological 211 variability should be considered. For this reason, first consider use of a large volume from one patient 212 diluted prior to the use of pooled samples. 213 214 Reference 215

CLSI. Evaluation of the Linearity of Quantitative Measurement Procedures: A Statistical Approach; 216 Approved Guideline. CLSI document EP06-A. Wayne, PA: Clinical and Laboratory Standards Institute; 217 2003. 218



Analytical Specificity (Interference, cross-reactivity, 219

competitive inhibition) 220

221

222

Study Type 223 Analytical Specificity (Interference, cross-reactivity, competitive inhibition) studies are conducted to evaluate 224

the ability of a measurement procedure to determine solely the quantity it purports to measure. 225

226

Objective 227

The objective of this study is to demonstrate that test performance is not affected by potentially interfering 228

materials present in samples. 229 230 Principles of the Use of Surrogate Samples 231 Surrogate samples should be used when clinical specimens are not available or if it is not possible to obtain 232 specimens with known amounts of potentially interfering material. 233 234 How to Use 235 For analytical specificity, there are two types of samples to which potentially interfering materials could be 236 added: naturally occurring human specimens and surrogate samples. Once the potentially interfering materials 237 have been added to either sample type the result is a surrogate sample that can be used to assess analytical 238 specificity with the test. In some cases, the surrogate sample could be composed of a surrogate analyte, a 239 surrogate sample matrix, and a potentially interfering substance. 240 241 It should be noted that not all potentially interfering substances are added to a sample. Potentially interfering 242 substances may be introduced through the preparation of a sample. For example, the fixation process for 243 formalin-fixed paraffin embedded (FFPE) tissue may introduce materials that interfere with test performance 244 or otherwise diminish the integrity of the sample. In these cases, the potential interference can be evaluated 245 through the comparison of specimens or surrogate samples processed using different procedural conditions. 246 247 When to Use 248 In order of recommended preference, surrogate samples used for analytical specificity should be as similar to a 249 clinical specimen as possible. As a consequence, surrogate samples that are categorized as supplemented with 250 interferent should be given the highest priority, followed by the pooled category, and finally the simulated 251 category. 252 253 In some cases, the potentially interfering material may not have an established method with which the 254 material concentration can be measured. For these surrogate samples it is acceptable to determine the 255 concentration of the potentially interfering materials based on the amount added. 256 257 In other cases, the addition of the potentially interfering material may not properly reflect physiological 258 conditions. For example, adding a drug or diluted drug to a sample may not yield the biologically active 259 metabolite, adding triglycerides may not reflect chylomicron or micelle formation that would happen in a 260 naturally occurring lipemic specimen, or the binding constants for the potentially interfering material may 261 influence test results as with the use of caffeine or theophylline when measuring free versus protein bound 262 bilirubin. For these surrogate samples the sponsor may need to refer to literature for ‘best practices’ based on 263



historic or otherwise reasonable expectations of how the potentially interfering material has been used in 264 surrogate samples. 265 266 Finally, some potentially interfering conditions do not have a corresponding surrogate counterpart. For 267 example, necrotic tissue in FFPE tissue cannot be emulated using a surrogate sample. In these cases, there is 268 no suitable surrogate sample and clinical specimens containing the potentially interfering conditions should be 269 used. 270 271 Reference 272

CLSI. Interference Testing in Clinical Chemistry; Approved guideline—Second edition. CLSI document EP07-273 A2. Wayne, PA: Clinical and Laboratory Standards Institute; 2005. 274

275



Precision/Reproducibility 276

Study Type 277 Precision/reproducibility studies are conducted to establish the within-laboratory precision and site-to-site 278 variability (reproducibility) of a test. In addition, evaluation of different components of variance as within-run 279 (repeatability), between-run, between-day, and between-lot are also conducted. 280 281 Objective 282 Precision/reproducibility studies are used to establish the level of agreement between repeated 283 measurements on the same or similar sample under different conditions. These different conditions can 284 include factors such as analyst, instrument, time, location, or conditions of use. 285

286 Principles of the Use of Surrogate Samples 287 Surrogate samples are widely used in precision/reproducibility studies to obtain sufficient sample with 288 consistent levels of analyte to perform the required replicate measurements. The preferred type of surrogate 289 sample for precision/reproducibility studies are pooled patient samples or spiked individual patient samples. If 290 the desired amount of analyte is not available in the pooled patient samples, the concentration of analyte 291 could be increased with spiking or decreased by dilution with a negative sample pool. 292 293 How to Use 294 To provide sufficient volume of sample with a known analyte concentration at different levels across the test’s 295 measuring interval, surrogate samples can be prepared in a number of ways, including: spiking individual 296 patient specimens, pooling of samples of similar concentrations across the measuring range, spiking a pooled 297 sample with a high concentration analyte, or dilution of a specimen with a high concentration of analyte with a 298 matrix that has a very low amount of analyte; dilution of pure analyte with a matrix that has a very low amount 299 of analyte; dilution of a specimen or pure analyte in a surrogate matrix. 300 301 When to Use 302 Due to the volume of sample required to perform precision/reproducibility testing the use of surrogate 303 samples is generally accepted. In all cases the surrogate samples should be representative of a specimen. 304

305 Additional Considerations 306 The benefit of individual spiked patient samples to better represent the variability between samples should be 307 considered when selecting the type of surrogate samples to use. Spiking some amount of analyte into patient 308 individual specimens would retain the matrix variability. 309

310 Reference 311

CLSI EP05: 2014 (Evaluation of Precision of Quantitative Measurement Procedures) recognizes that 312 typically for precision studies pooled patient specimens are used and that the spiking or dilution of 313 samples is necessary to achieve analyte levels at the appropriate levels. 314

315

316



Detection Limit (LoB, LoD, and LoQ) 317

Study Type 318 Test developers must determine the ability of a test to detect an analyte at the limits of test performance and 319 accurately determine when an analyte is not present. 320 321 Objective 322 Detection capability studies are used, depending on the type of test, to establish the Limit of Blank (LoB), Limit 323 of Detection (LoD), and/or Limit of Quantitation (LoQ). The studies are usually the first analytical studies 324 conducted in order to determine the appropriate spiking level of an analyte for other analytical studies (e.g., 325 interfering substances, reproducibility) for qualitative tests. 326 327 Principles of the Use of Surrogate Samples 328 The matrix selected should closely represent a true, individual clinical specimen. It is important that the matrix 329 selected for conducting detection capability studies does not contain extraneous properties or elements that 330 would contribute to the variability of results that would not typically be encountered when testing clinical 331 specimen or would otherwise confound test results obtained. Such exogenous substances may introduce bias 332 that could impact the ability of the study to establish the “true” detection limit of a test or test system. If the 333 matrix must be artificially prepared, the characteristics of the surrogate specimen should be clearly defined so 334 that the surrogate matrix resembles an appropriate patient specimen. 335 336 How to Use 337 For evaluation of LoB, for most material, it is preferred to collect individual patient specimen(s) known to be 338 negative for the intended analyte(s). The analyte intended for measurement is then added to the matrix either 339 directly from a high concentration stock or using a clinical specimen known to contain the intended analyte(s), 340 if quantitatively measurable in the latter case. For qualitative tests, from this spiked sample, serial dilutions can 341 be made to create the varying levels required to determine the detection limit; dilutions should preferably be 342 made with the unspiked, negative sample matrix. Should the required volume requirements not be met with a 343 single-source sample, pooling a minimal number of clinical specimens is preferred to preserve biological 344 variability. If a large amount of a known negative specimen is hard to obtain (e.g., cerebral spinal fluid), then a 345 simulated artificial matrix that closely resembles the naturally occurring specimen can be used. 346 347 When using negative patient specimens, it is important to reduce bias associated with the variability 348 introduced by such specimens. Specimens obtained from different patients in a clinical setting, collected to 349 create a representative clinical sample matrix pools, may contain different levels of endogenous and/or 350 exogenous components based on their genetics and/or disease state as well as different substances used to 351 manage their disease. Although presence of the endogenous elements may be beneficial for establishing 352 analytical sensitivity in the intended use population, in excess these substances may create a matrix effect, 353 which is a phenomenon that may generate bias in the results obtained from the study. Patient specimens used 354 to create negative pools should be free of obvious, potential interferents that generate bias. It is acceptable to 355 dilute positive specimens or spike the analyte into negative specimens in order to provide low-level samples at 356 desired levels of analyte. When spiking analyte(s) into the sample matrix, high concentrations of naturally 357 occurring or artificial material is preferred, to limit any initial dilution effects. 358 359 In several situations, (e.g., qualitative ID molecular/NAAT assays), the ability to create reproducible negative 360 sample matrix pools is critical. Therefore, it is preferred to use the same pooled/supplemental sample matrix 361 that is spiked at concentrations representing multiples of the LOD determined for the test (e.g., at 362



concentrations 2 to 3 times the determined LOD). This approach will help reproduce variability in results 363 associated with different specimen pools. 364 365 When to Use 366 When establishing analytical detection limits, surrogate specimens are needed to reliably determine these 367 limits since in many cases sufficient volumes of clinical specimens containing known concentration(s) of the 368 analyte(s) of interest cannot practically be found in a clinical population. 369 370 Additional Considerations 371 Preparation of surrogate specimens for use in detection limit studies may be challenging as the method of 372 preparation may introduce bias. For example, the spiking of very low levels of analytes may be challenging as 373 the ability to reliably add small amounts of the analyte may exceed the capability of the system used to add 374 these low levels. Therefore, it may be necessary to create serial dilutions of the matrix, whereby the analyte is 375 added at a high level to a matrix that is then diluted multiple times to achieve the appropriate level. 376 377 Some specimen types may create additional challenges where special methodologies may be required to 378 assess the detection limit. For example, when formalin-fixed paraffin embedded (FFPE) tissue is the specimen 379 type used for cancer diagnostic tests, it may be necessary to provide insertions or deletions of one or more 380 nucleotides. For swab specimens, the analyte may not be able to be added directly to the swab. In these cases, 381 the analyte may need to be added to an appropriate liquid matrix. Various dilutions are then coated onto the 382 swab. 383 384 Additionally, as large specimen pools are used for such detection limit studies, the stability of the specimen 385 matrix should be established so as not to introduce result variability through degradation of the matrix. 386 387 References 388

CLSI. Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures; Approved 389 Guideline—Second Edition. CLSI document EP17-A2. Wayne, PA: Clinical and Laboratory Standards 390 Institute; 2012. 391

Armbruster D, Pry T. Limit of blank, limit of detection, and limit of quantitation. Clin Biochem Rev; 2008 392 Aug; 29(Suppl 1): S49-S52. 393

Shiobara et al. Development of artificial cerebrospinal fluid: Basic experiments, and Phase II and III clinical 394 trials. J Neurol Neurophysiol. 2013; 4: 173. 395 396

397



Matrix Comparison Study 398

Study Type 399 Matrix comparison is a study for demonstration of similarity of specimen types. 400 401 Objective 402 The purpose of the matrix comparison study for a Candidate type of samples (e.g., plasma) and a Primary type 403 of samples (e.g., serum) for a quantitative test is an evaluation of a systematic difference between these two 404 types of samples at medical decision levels (MDLs). For a qualitative test, the purpose of the matrix 405 comparison study is an estimation of positive and negative percent agreements of the test results with the 406 Candidate samples vs test results with the Primary samples for a population. In the matrix comparison study, 407 usually every patient in the study provides Primary and Candidate specimens which are measured by the test 408 and results are compared. 409 410 When Surrogate Samples can be Considered for Use 411 Covering the measuring interval of the test 412 For a quantitative test, test values with Primary samples should cover the measuring interval of the test. 413 Sometimes, it is difficult to find patients who have Primary samples with high values or low values. Pairs of 414 surrogate samples can be used in the matrix comparison study. 415 416 Patients with Primary samples values close to MDL 417 In the study, it should be few Primary samples (3-5 samples) with test values close to the MDL. If it is difficult 418 to find patients with Primary samples values close to the MDL, pairs of surrogate samples can be used. 419 420 Not enough patients with positive test results for Primary samples 421 For a qualitative test, an estimation of the positive percent agreement (PPA) requires a certain number of 422 patients with positive test results for Primary samples. If it difficult to find patients with positive test results 423 with Primary samples, surrogate pairs of samples (Primary and Candidate samples) can be added for 424 estimation of PPA. 425 426 Samples close to the cutoff 427 In matrix comparison studies, it is required to have some percent of sample pairs with values close to the test’s 428 cutoff with Primary samples among all positive samples. If it is difficult to find patient samples close to the 429 cutoff, pairs of surrogate samples can be included in the matrix comparison study. 430 431 Principles of the Use of Surrogate Samples 432 For pairs of surrogate samples with high test values with Primary samples in the matrix comparison study of 433 quantitative test, add the minimal amount of analyte as possible to the Primary and Candidate samples. 434 435 For pairs of surrogate samples with low test value with Primary samples in the matrix comparison study of 436 quantitative test, add the minimal amount of the sample with zero concentration (as diluent) as possible. 437 438 The use of the surrogate samples in the matrix comparison study for a qualitative test should preserve the 439 percent of samples close to the cutoff, which is required in the matrix comparison study. When pairs of 440 surrogate samples are used in the matrix comparison study, first, it should be determined C5 –C95 interval (or 441 LoD) of the qualitative test for defining of “close to the cutoff.” If a spiking or dilution technique will be used to 442 prepare surrogate samples, the relationship between amount of spiking material (or amount of true negative 443



patient sample) and how close to the cutoff these Primary samples can be should be understood. If the 444 qualitative test does not have available signals, then the use of surrogate samples can be problematic. 445 446 Reference: 447

No current CLSI document (CLSI EP35, regarding matrix comparison, is in development, projected 448 publication: end of 2017) 449

450



Method Comparison 451

Study Type 452 Evaluation of bias or systematic difference between two quantitative or qualitative tests. 453

454 Objective 455 The purpose of the method comparison study for a candidate quantitative test and the comparator test is an 456 evaluation of bias or systematic difference. For a qualitative candidate test, the purpose of the method 457 comparison study is an estimation of positive and negative percent agreements of the candidate test results 458 with a comparator test results for a population. The comparator can be a qualitative test or a quantitative test 459 with an appropriate cutoff. 460 461 When Surrogate Samples can be Considered for Use 462 Not enough volume 463 In the method comparison studies, each sample needs to be measured by the candidate test and by the 464 comparator test. Sometimes, volumes of neat patient samples do not allow this; therefore, pooling of patients 465 samples are used. 466 467 Covering the measuring interval 468 For a quantitative candidate test, samples in the method comparison study should cover the measuring 469 interval of the test. Sometimes, it is difficult to have samples with high values or low values; surrogate samples 470 can be used in the method comparison study to meet these needs. 471 472 Samples close to medical decision levels (MDLs) 473 For a quantitative candidate test, it should be few samples (3-5 samples) close to the MDL in the method 474 comparison study. If it is difficult to find patient samples close to the MDL, surrogate samples can be used. 475 476 Not enough samples positive by the comparator 477 For a qualitative candidate test, an estimation of the positive percent agreement (PPA) with some level of 478 uncertainty (confidence intervals) requires some number of samples positive by the comparator test. If it is 479 difficult to find patient samples positive by the comparator test, surrogate positive samples can be added for 480 estimation of PPA. An inclusion of surrogate samples can be limited to only some special type of studies and 481 most of the time, archived samples (not surrogate samples) should be considered. 482 483 Samples close to the cutoff 484 In some method comparison studies (e.g., migration studies), it is required to have a certain percentage of 485 samples close to the cutoff of the candidate test among all positive samples. If it is difficult to find patient 486 samples close to the cutoff, surrogate samples can be included in the method comparison study. 487 488 Principles in Preparation of Surrogate Samples 489 Preserve individual biological variability as much as possible (if possible, do not pool samples). For pooling, use 490 each individual sample only one time to preserve variability. 491 492 For very high concentration surrogate samples in the method comparison study of a quantitative Candidate 493 test, use patient specimens with the highest concentration of naturally occurring analyte to minimize the use 494 of spiking material available (e.g. high concentration stock solutions). For low value surrogate samples in the 495



method comparison study of the quantitative Candidate test, use patient specimens with low concentrations 496 of naturally occurring analyte to minimize the use of diluent. 497 498 For obtaining unbiased estimates of positive and negative percent agreements in the method comparison 499 study of the qualitative Candidate test, the percent of samples close to the cutoff of the Candidate test should 500 reflect the intended use population or requirements applied for the patient samples (e.g., migration study). 501 Therefore, the use of the surrogate samples in the method comparison study should preserve the percent of 502 samples close to the cutoff. First, when surrogate samples are used in the matrix comparison study, it should 503 be determined C5 –C95 interval (or LoD) of the Candidate test for defining of “close to the cutoff”. Second, the 504 percent of samples close to the cutoff of the Candidate test should be evaluated in the intended use 505 population. For this, consider all patient samples in the comparison study, which were randomly selected from 506 the intended use population, and estimate the percent of the patient samples close to the cutoff of the 507 Candidate test. The surrogate samples should have the same percent of the samples close to the cutoff. To 508 determine whether a patient sample result was close to the cutoff, one can use values of the signal from the 509 Candidate test (if available). If the signal of the Candidate test is not available (e.g., influenza test with visual 510 reading of the line), consider a signal of the Comparator test (if available) or quantitative values of the 511 quantitative Comparator. Third, if a spiking or dilution technique will be used to prepare the surrogate 512 samples, one should understand the relationship between the amount of spiking material (or amount of true 513 negative patient sample in dilution) and how close to the cutoff these samples can be. If the Candidate test 514 and Comparator test do not have available signals then the use of surrogate samples can be problematic, for 515 instance, a method comparison study for influenza test with visual reading of the line (Candidate) vs culture 516 (Comparator). 517 518 Additional Considerations 519 First, always perform separate statistical analyses for patient specimens and for surrogate samples. If 520 performances are similar, as described by estimates of biases (systematic differences) at MDLs for quantitative 521 test or PPA and NPA for the qualitative candidate test, then perform a statistical analysis for all samples 522 combined. 523 524 Note that for a quantitative candidate test, estimates of the slopes, intercepts, and biases (or systematic 525 differences) at MDLs using patient specimens and using the surrogate samples should be compared. If the 526 surrogate samples show less dispersion around the regression line compared to the dispersion around the 527 regression line for the patient samples, then confidence intervals about slope, intercept, and biases at MDL 528 with combined data can be misstated. 529 530 References 531

CLSI. Measurement Procedure Comparison and Bias Estimation Using Patient Samples; Approved 532 Guideline—Third Edition. CLSI document EP09-A3. Wayne, PA: Clinical and Laboratory Standards Institute; 533 2013. 534

CLSI. User Protocol for Evaluation of Qualitative Test Performance; Approved Guideline—Second Edition. 535 CLSI document EP12-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 2008. 536



Specimen Stability 537

Study Type 538 Specimen stability is the ability of a specimen to maintain its measured characteristic(s) within acceptable 539 limits over an established period of time. 540 541 Objective 542 Assess the magnitude of analyte degradation through its measured characteristic(s) over time when stored 543 under defined conditions. 544 545 Principles of the Use of Surrogate Samples 546 Acceptable changes of bias or variability that may arise from sample stability are identified by assessing 547 analyte stability in a known surrogate matrix. Surrogate samples may be used to facilitate specimen stability 548 studies by providing sufficient sample volume to conduct tests or to obtain targeted analyte levels, such as 549 levels near the medical decision level (MDL). In these situations, selection from the surrogate hierarchy begins 550 at the level designated “pooled unspiked” followed next by the level designated “pooled spiked.” When the 551 use of surrogate samples is solely to address sample volume, thoroughly consider the scientific rationale as to 552 why the higher ranking sample type is not feasible before moving to the next level of the hierarchy. When the 553 use of surrogate samples is to address scientific challenges with the functional characteristics of naturally 554 occurring specimens (e.g., whole blood, cerebral spinal fluid), begin at the surrogate sample hierarchy level 555 designated “supplemented” and move down through each level of the hierarchy after thoroughly considering 556 the scientific rationale as to why the higher ranking sample type is not feasible. 557 558 How to Use 559 Although the use of surrogate samples to establish specimen stability is not well-documented in standards or 560 literature, assessment of analyte stability may occur with the use of a well-characterized matrix of known 561 stability and with reagents of known stability. Prepare surrogate samples with very high and very low 562 concentrations of the analyte, as well as at levels at the MDL. When preparing surrogate samples follow well-563 designed pre-analytical steps to mimic processing steps of native specimens. 564 565 When to Use 566 Consult the surrogate sample hierarchy at Figure 1, so that your selection begins with a surrogate that is most 567 representative of a specimen. Stability of the selected matrix should be known and understood to avoid the 568 introduction of bias. 569 570 Additional Considerations 571 Consider how well the selected matrix resembles a specimen in terms of viscosity or other key elements of the 572 naturally occurring specimen. Well-designed pre-analytical steps to mimic processing steps of native 573 specimens is an important consideration, so that spiked analyte is processed as the naturally occurring analyte 574 in native biological specimens. 575 576 Reference 577 CLSI. Evaluation of Stability of In Vitro Diagnostic Reagents; Approved Guideline. CLSI document EP25-A. 578

Wayne, PA: Clinical and Laboratory Standards Institute; 2009. 579



Reagent Stability 580

Study Type 581 Stability is the ability of an in vitro diagnostic (IVD) reagent to maintain its performance characteristics within 582 pre-established acceptance limits over time. Manufacturers establish shelf-life and in-use stability for 583 quantitative and qualitative IVD reagents. 584 585 Objective 586 IVD reagent stability is evaluated on the basis of ensuring that key performance metrics continue to meet 587 predefined acceptance criteria within the expiry date. Eliminating or reducing variation of other study 588 parameters permits the evaluation of the IVD reagents. The use of surrogate samples in stability studies plays a 589 key role in minimizing sample variability because it enables the creation of sufficient volume for the total 590 number of tests. 591 592 Principles of the Use of Surrogate Samples 593 Surrogate samples facilitate reagent stability studies because they ensure sufficient volume of sample for 594 testing over the study’s duration. Stability studies assess performance at the extremes of the reportable range, 595 which are often the regions where significant reagent changes are likely to occur, thus affecting stability, as 596 well as at the medical decision level (MDL). Surrogate samples allow for sufficient sample volume targeted at 597 the appropriate levels for testing. When an appropriate surrogate sample type is selected, stability studies may 598 be conducted with one hundred percent surrogate samples. As used in this context, an appropriate surrogate 599 sample is one of known stability duration within the unspiked or spiked with naturally occurring analyte pooled 600 section of the surrogate sample hierarchy at Figure 1. Because IVD reagents may show little or no significant 601 analyte drift when tested with non-native analyte material, design stability studies to use unspiked pooled or 602 spiked with naturally occurring analyte pooled surrogate samples when available. 603 604 How to Use 605 Patient sample pools, controls, and other suitable materials of known analyte stability are recognized as 606 appropriate surrogate sample types and are tested alongside associated calibrators.5 Surrogate sample types 607 may be neat, diluted, or supplemented to represent the MDL, low level of measuring interval, and high level of 608 the measuring interval for quantitative tests. For qualitative tests, surrogate sample types may be neat, 609 diluted, or supplemented to represent the MDL. Determine surrogate sample stability duration as a pre-610 requisite to use in reagent stability studies. 611 612 When to Use 613 Although the use of surrogate samples is generally acceptable in establishing reagent stability,5 consult the 614 surrogate sample hierarchy at Figure1, so that your selection begins with a surrogate that is most 615 representative of a specimen (e.g., patient sample pool). Calibrators and controls may be used as part of the 616 study. However, they should not be used at the exclusion of surrogate sample types ranking higher on the 617 hierarchy without clear scientific rationale as to why the higher ranking sample type is not feasible. For 618 surrogate samples selected from the spiked artificial pooled section or simulated section of the surrogate 619 sample hierarchy, it is advisable to include within the study native human specimen, unspiked pooled 620 surrogate samples, or spiked with naturally occurring analyte pooled surrogate samples comparable to one 621 surrogate sample level or the MDL. 622

5 CLSI. Evaluation of Stability of In Vitro Diagnostic Reagents; Approved Guideline. CLSI document EP25-A. Wayne, PA: Clinical and Laboratory Standards Institute; 2009.



623 Additional Considerations 624 Describe the surrogate sample in the stability testing plan. Consider differential recognition of the analyte 625 changing over time when using surrogate samples. When able, include native human specimen comparable to 626 one surrogate sample level or the MDL. Consider performance of surrogate samples in other testing to support 627 the use in stability studies when native human specimens are unavailable. 628 629 References 630

CLSI. Evaluation of Stability of In Vitro Diagnostic Reagents; Approved Guideline. CLSI document EP25-A. 631 Wayne, PA: Clinical and Laboratory Standards Institute; 2009. 632

ISO. In vitro diagnostic medical devices – Evaluation of stability of in vitro diagnostic reagents. ISO 23640. 633 Geneva, Switzerland: International Organization for Standardization; 2012. 634

635



Instrument Carry-Over Studies 636

637 Study Type 638 For automated liquid handling systems used in clinical diagnostics, developers must demonstrate that carry-639 over and cross-contamination will not occur during use of the system with patient specimens. These types of 640 phenomena may contribute to a false negative, false positive, incorrect concentration level, indeterminate, or 641 no result. 642 643 It should be noted that these types of studies might not be necessary if system design prevents the possibility 644 of carry-over and cross-contamination. 645 646 647 Objective 648 The results of carry-over and cross-contamination studies should demonstrate that there is no detrimental 649 impact of the liquid handling system when used with patient specimens on the results of the tests run on the 650 automated analyzer. 651 652 Principles of the Use of Surrogate Samples 653 The matrix selected should closely represent a true clinical specimen. If matrices other than patient specimens 654 are used, the sample may not properly mimic the intended specimen type. For example, the viscosity of the 655 surrogate sample may not accurately reflect the opportunity for specimen/analyte carry-over effects. 656 Surrogate specimens spiked with a high amount of analyte may be used to simulate high positive (or high 657 concentration) samples. Individual patient samples known to be negative for the target analyte are also used. 658 The samples are to be used in series alternating with negative (or low concentration) samples in patterns 659 dependent upon the operational function of the device. 660 661 How to Use 662 The surrogate samples should be derived from the specimen types intended to be used with the automated 663 system. Therefore, the preferred specimen types are pooled patient specimens that are unspiked and spiked 664 with the target analyte. The surrogate specimens should be spiked with the analyte to an appropriate level in 665 order to measure the impact of sample/analyte carry-over. Based on manufacturer instructions, it may be 666 necessary to process the specimen (e.g., in a sample buffer) prior to analysis in the carry-over study. 667 668 Simulated sample matrices should be avoided when evaluating the impact of patient sample carry-over and 669 cross-contamination. 670 671 When to Use 672 When determining the impact of potential carry-over and cross-contamination effects using patient samples, 673 surrogate specimens are needed due to the lack of specimens with known, high concentrations of the target 674 analyte, in the clinical population. 675 676 Additional Considerations 677 Note that some specimen types may have inherent variability that may be difficult to mimic with a surrogate 678 specimen (e.g., excessively mucoidal specimens, sputum, hemolyzed blood samples). It may be necessary to 679 obtain negative specimens with these characteristics and spike in the appropriate levels of the target 680 analyte(s). 681



682 References 683

Broughton P. M. Carry-over in automatic analysers. J Automatic Chem, 1984; 6(2): 94-95. 684

Haeckel R. Proposals for the description and measurement of carry-over effects in clinical chemistry. Pure 685 Appl Chem. 1991; 63: 301-306. 686

687



Clinical Performance 688

Study Type 689 Clinical performance relates to the clinical sensitivity and sensitivity of the test. 690 691 Objective 692 Clinical sensitivity is the ability of an IVD test to identify the presence of a target marker associated with a 693 particular disease or condition, which is based on criteria that are independent of the test under consideration. 694 Clinical specificity is the ability of an IVD test to recognize the absence of a target marker associated with a 695 particular disease or condition, which is based on criteria that are independent of the test under consideration. 696 Clinical cut-off is the value that is understood to differentiate two conditions, usually health from disease. 697 698 Principles of the Use of Surrogate Samples 699 Since clinical performance is contingent upon the intended use of the test in defined populations, the use of 700 surrogate samples is typically reserved for use as a means to supplement clinical specimens, rather than 701 replacement. In cases of rare disease, surrogate samples may be considered for replacement purposes. 702 However, in such cases, carefully consider functional comparability. Preservation of biological variability is a 703 key consideration. Thus, it is advisable to begin at the top of the Surrogate Sample Hierarchy when considering 704 the use of surrogate samples for clinical performance studies. 705 706 How to Use 707 Use well-characterized surrogate samples. When preparing surrogate samples follow well-designed pre-708 analytical steps to mimic processing steps of native specimens. 709 710 When to Use 711 Quantitative Tests 712 Use when the analyte of interest is absent from specimens of healthy subjects. In those cases, the cutoff is 713 often set as the limit of blank (the lowest concentration of analyte that can be consistently detected) or the 714 limit of quantification (the lowest concentration of analyte that can be quantitatively determined with stated 715 accuracy).6 Supplementing the analyte of interest into individual specimens (preferred) or pools is commonly 716 used to establish these parameters. The indication for use would prescribe which is appropriate. 717 718 Similarly, surrogates can be used to better define the Confidence Interval (CI) for the cut-off values of the 719 reference interval. 720 721 Further, with increased used of nucleic acid diagnostic techniques, when “copy” numbers are used to establish 722 a cutoff, supplementing samples with known oligonucleotides are often employed. 723 724 Qualitative Tests 725 Qualitative tests are frequently used for the diagnosis of infectious diseases by identifying an analyte that is 726 not present in healthy subjects. For these tests, the cutoff is defined as the limit of blank. Supplementing the 727 analyte of interest into individual specimens (preferred) or pools is commonly used to establish these 728 parameters. 729 730

6 CLSI. Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures; Approved Guideline—Second Edition. CLSI document EP17-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 2012.



Additional Considerations 731 Surrogates prepared by supplementing analyte or pooling specimens, defined here, are best prepared with 732 matrices most similar to those of the clinical specimens used by the test. If the analyte is added to samples 733 that are devoid of material of interest, several patient sources are preferable for determining limits of 734 detection. In this way, any bias caused by patient-specific causes can be evaluated and compared to pre-735 determined acceptance criteria. If studies of aggregate properties are being conducted, such as in an 736 agreement study or a correlation, pools are often more convenient for preparation, especially if a large sample 737 volume is needed. In addition, when analytes are extracted before analysis, such as is often needed for 738 infectious disease or nucleic acid tests, surrogates can be prepared using the extraction medium as the solvent. 739 740 741



742

GLOSSARY



Glossary 743

Analyte: Component represented in the name of a measurable quantity. In this framework, we have used 744 analyte preferentially over the term measurand (ISO 17511) 745

Analytical Specificity: Ability of a measurement procedure to measure solely the measurand (ISO 17511) 746

Biospecimen: A biological sample fluid or tissue obtained from an organism (CLSI I/LA28) 747

Cut-off value (laboratory medicine): Quantity value used as a limit to identify samples that indicate the 748 presence or the absence of a specific disease, condition, or measurand (ISO/TS 17822-1:2014 (en)) 749

Detection Limit (Limit of Detection, LoD): The lowest concentration of analyte that can be consistently 750 detected. (Adopted from CLSI EP17-A2:2012) 751

i. Limit of Blank (LoB): The highest measurement result that is likely to be observed, with a 752 stated probability (typically, in ≥95% of samples tested under routine clinical laboratory 753 conditions and in a defined type of sample), for a blank sample (CLSI EP17-A2:2012) 754

ii. Limit of Quantitation (LoQ): The lowest amount of measurand in a material that can be 755 quantitatively determined with a stated accuracy under stated experimental conditions (CLSI 756 EP17-A2:2012) 757

Interference (analytical): Systematic error of measurement caused by an influence quantity, which does not by 758 itself produce a signal in the measuring system, but which causes an enhancement or depression of the value 759 indicated (CLSI MM17-A:2008) 760

In-use stability: Duration of time over which the performance of an IVD reagent within its expiration date 761 remains within specified limits after opening the container system supplied by the manufacturer, and put into 762 use under standard operation conditions (e.g., storage on the instrument). (CLSI EP25-A:2009) 763

Linearity: The ability (within a given range) to provide results that are directly proportional to the 764 concentration (amount) of the analyte in the test sample (CLSI EP06-A:2003) 765

Matrix (of a material system): Components of a material system, except the analyte (ISO 15193) 766

Matrix comparison: Equivalence of specimen types 767

Matrix effect: Influence of a property of the sample, other than the analyte, on the measurement, and thereby 768 on the value of the measurable quantity (ISO 15194) 769

Measurand: Quantity intended to be measured (ISO 15193) 770

Method comparison: Evaluation of bias or systematic difference between two quantitative or qualitative tests 771

Precision (of measurement): Closeness of agreement between independent results of measurements obtained 772 under stipulated conditions 773

i. Repeatability: Precision under essentially unchanged conditions, often termed “within-run 774 precision” 775



ii. Reproducibility: Precision under changes in conditions (e.g., time, different laboratories, 776 operators, and measuring systems [including different calibrations and reagent batches]), (ISO 777 17511) 778

Reference Range or Reference Interval: The range of test values expected for a designated population of 779 individuals (42 CFR 293.2) 780

Shelf life: Period of time until the expiration (expiry) date, during which an IVD reagent in the packaging 781 configuration provided to the user maintains its stability under storage conditions specified by the 782 manufacturer (CLSI EP25-A:2009) 783

Sensitivity (of a measuring system): Quotient of the change in an indication of a measuring system and the 784 corresponding change in a value of a quantity being measured (JCGM 200:2012) 785

Specimen: Discrete portion of a body fluid or tissue taken for examination, study, or analysis of one or more 786 quantities or properties assumed to apply for the whole (ISO 15189) 787

Surrogate Material: Material used as a substitute for biologic material without having all of the functional 788 characteristics of such materials (CLSI H26) 789

Surrogate Sample: Material used as a substitute for body fluid or tissue taken for examination from a patient. 790 Examples include, but are not limited to: 791

• Materials supplemented (spiked) with an analyte of interest 792 • Pooled specimens of biological origin 793 • Material created to have functional characteristics of a body fluid or a tissue 794 • Material comprised of a combination of an analyte that emulates the analyte of interest and a 795 material created to have the functional characteristics of a body fluid or tissue. 796

797

Sources: 798

42 CFR 293.2 Laboratory requirements. General provisions. Definitions. 799

CLSI. Evaluation of Precision of Quantitative Measurement Procedures; Approved Guideline—Third Edition. CLSI 800

document EP05-A3. Wayne, PA: Clinical and Laboratory Standards Institute; 2014.CLSI. Evaluation of the 801

linearity of quantitative measurement procedures: A statistical approach; Approved Guideline. CLSI document 802

EP06-A. Wayne, PA: Clinical and Laboratory Standards Institute; 2003.CLSI. Interference Testing in Clinical 803

Chemistry; Approved Guideline—Second Edition. CLSI document EP07-A2. Wayne, PA: Clinical and Laboratory 804

Standards Institute; 2005.CLSI. Measurement Procedure Comparison and Bias Estimation Using Patient 805

Samples; Approved Guideline—Third Edition. CLSI document EP09-A3. Wayne, PA: Clinical and Laboratory 806

Standards Institute; 2005.CLSI. User Protocol for Evaluation of Qualitative Test Performance; Approved 807

Guideline—Second Edition. CLSI document EP12-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 808

2008. 809

CLSI. Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures; Approved Guideline—810

Second Edition. CLSI document EP17-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 2012. 811



CLSI. Evaluation of Stability of In Vitro Diagnostic Reagents; Approved Guideline. CLSI document EP25-A. 812


CLSI. Characterization and Qualification of Commutable Reference Materials for Laboratory Medicine; 814

Approved Guideline. CLSI document EP30-A. Wayne, PA: Clinical and Laboratory Standards Institute; 2010. 815

CLSI. Validation, Verification, and Quality Assurance of Automated Hematology Analyzers; Approved 816

Standard—Second Edition. CLSI document H26-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 817

2010. 818

CLSI. Quality Assurance for Design Control and Implementation of Immunohistochemistry Assays; Approved 819

Guideline—Second Edition. CLSI document I/LA28-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 820

2011. 821

CLSI. Verification and Validation of Multiplex Nucleic Acid Assays; Approved Guideline. CLSI document MM17-A. 822


ISO/TS 17822-1:2014 (en): In vitro diagnostic test systems—Qualitative nucleic acid-based in vitro examination 824

procedures for detection and identification of microbial pathogens—Part 1: General requirements, terms and 825

definitions 826

ISO 17511: In vitro diagnostic medical devices—Measurement of quantities in biological samples— 827

Metrological traceability of values assigned to calibrators and control materials 828

ISO 15194: In vitro diagnostic medical devices—Measurement of quantities in samples of biological origin—829

Requirements for certified reference materials and the content of supporting documentation 830

ISO 15193: In vitro diagnostic medical devices—Measurement of quantities in samples of biological origin—831

Requirements for content and presentation of reference measurement procedures 832

ISO 15189: Medical laboratories—Requirements for quality and competence 833

ISO 11843: Capability of detection 834

ISO 23640:2011 – Evaluation of stability in in vitro diagnostic reagents 835

ISO/Guide 30:2015 (en) Reference materials—Selected items and definitions 836

JCGM 200:2012 International vocabulary of metrology—Basic and general concepts and associated terms. 837

International Bureau of Weights and Measures. 838

839



MDIC has assembled a work group comprised of member organizations and other subject 840

matter experts to guide work on this project. 841

842

Industry: 843

Khatereh Calleja, JD, AdvaMed 844

Tremel Faison, BARDA 845

Ronald Freeze, PhD, Abbott 846

Ian Giles, MD, Sysmex 847

Patrick O’Donnell, Roche 848

Brad Spring, BD 849

Mark Del Vecchio, BD 850

April Veoukas, JD, Abbott (Working Group Chair) 851

852

FDA: 853

Marina Kondratovich, PhD, CDRH | OIR 854

Zivana Tezak, PhD, CDRH | OIR 855

Yun-Fu Hu, PhD, CDRH | OIR 856

857

Expert Advisors: 858

Susan Alpert, MD, PhD 859

Constantine Gatsonis, PhD 860

Fred Lasky, PhD 861

862

Program Director: 863

Carolyn Hiller, MDIC 864

865

© Medical Device Innovation Consortium 2017 866

surrogate sample - mdic

Documents