Copy of Example 2 - Modeling Multi-target Microbial Detection/Identification Test

Introduction

Persistent infection with Human Papillomavirus (HPV) is the main cause of cervical cancer and is a precursor to cervical intraepithelial neoplasia (CIN). HPV is a nonenveloped, double-stranded DNA virus, and more than 150 different types of HPV have been identified. A subset of these types is considered high-risk ( for the development of cervical cancer and its precursor lesions, and is routinely screened clinically in the context of cervical cancer prevention. The HPV test data modeled in this example represent a class of commercially available assays that are able to screen and detect multiple types of high-risk HPV from a subject specimen in a single analysis. https://www.uspharmacist.com/article/human-papillomavirus-cervical-cancer-and-prevention

In this example, the HPV test screens for a specific region in the viral DNA and carries out multiplexed amplification of target DNA by real-time Polymerase Chain Reaction (PCR). The targeted, pre-specified HPV types detected in this assay are: HPV types 31, 33, 35, 39, 51, 52, 56, 58, 59, 66 and 68.

Problem Statement

For most of CDISC specimen-based domains, the "analyte" under assessment is mapped to the --TEST/--TESTCD variables, and in the context of the MB domain, the analyte is typically the microorganism of interest. For multi-target microbial detection tests (where the assay is able to simultaneously detect and evaluate multiple organisms), this requires the pre-coordination of the targeted, pre-specified microorganisms into the MBTEST/MBTESTCD variables. In reference to the high-risk HPV screening test described in the Introduction section, this means a MBTEST value such as the follow would need to be created: MBTEST = Human Papillomavirus Types 31/33/35/39/51/52/56/58/59/66/68 DNA.

There are several issues associated with this approach:

1. CDISC Conformance Rules dictate that <=8-character TESTCD/PARMCD and <=40-character TEST/PARM values be available for data submission to regulators. Because of the 8- and 40-character limit, this MBTEST value will require heavy abbreviation/truncation; this also makes the development of a unique and meaningful MBTESTCD impossible, therefore in this instance, the MBTESTCD value will be replaced with the NCI C-code to ensure uniqueness. The abbreviated MBTEST and C-code dubbed MBTESTCD significantly undermine their meaningfulness and clarity, and make data/dictionary management of the TEST/TESTCD values difficult.

2. A quick LOINC search yields 27 unique HPV multi-target screening combinatorial tests. The pre-coordinated --TEST/--TESTCD approach would lead to the development of 27 unique MBTESTs where the MBTESTCDs will most likely be replaced with their NCI c-codes.
3. Overloading the --TEST/–TESTCD variables. Because there is a lack of clear definition on what is considered as the “analyte” under assessment as well as standard variables support in the MB domain, microbial antigens targeted in detection assays have traditionally been pre-coordinated into the MB topic paired variables. The types of microbial antigens targeted for detection are (but are not limited to) nucleic acids (e.g. DNA, RNA, mRNA etc.), antigens (e.g. specific toxins, proteins, etc.), and genes of interest (e.g. name of genes). Examples are Cytomegalovirus Phosphoprotein 65 Antigen, MERS-CoV RNA, Clostridium difficile tcdA (gene) DNA, Clostridium difficile B Toxin, etc. The pre-coordination of the antigenic types into --TEST/--TESTCD appears less of a problem when the assay targets a single organism; for multi-target tests, this approach exacerbates overloading of the MBTEST/MBTESTCD. For example, for this high-risk HPV screening test, Human papilloma virus 16+18+31+33+35+39+45+51+52+56+58+59+66+68 E6+E7 mRNA (loinc code: 94425-6), this test specifically detect the oncoprotein E6- and E7-producing mRNAs in the pre-specified subset of HPV types. The MBTEST for this assessment should look like the following: MBTEST = Human Papillomavirus Types 31/33/35/39/51/52/56/58/59/66/68 E6/E7 mRNA. Note the pre-coordination of the multiple HPV types as well as the antigenic types (E6/E7 mRNA) being detected in MBTEST.
   • See Kathleen's feedback in the Comment section of this page.

Resolution: post-coordination for multi-target tests

The MBTEST/CD for this class of tests should describe the overall type (or taxonomic classification if available) of the microorganism(s) being assessed. The specific multiple microorganism subtypes/strains/genotypes, etc., and the multiple microbial antigens will be post-coordinated into two NSVs: --MOTRG and --MAGTRG

Example 1: This example describes a class of pooled hrHPV screening tests that are able to detect multiple high-risk HPV types but are unable to differentiate between the specific types.

This example fills a gap in the current MB domain model. The MBTEST = Microbial Organism Identification, is used as the open-ended question for "what microbe does the patient have in the sample?" I believe this test was originally designed for bacterial detection/identification studies where a subject's sample is grown on a selective media/plate in order to identify the different bacteria. The results for this test are the specific bacteria identified on the culturing media. This approach however does not work well for viral studies where most viral detection/identification tests are performed by PCR, and are typically tests that screen for multiple "pre-specified" viruses. Therefore, for the viral detection/identification tests, it is inappropriate to use MBTEST = Microbial Organism Identification, because this is not a completely open-ended question - you are looking for the presence of a microprogram from a group of pre-specified microorganisms. Take the hrHPV screening test as an example, the purpose of this test is to detect which high-risk HPV subtype(s) the patient may have. This PCR assessment is designed to detect a "scope/range of pre-specified high-risk HPV subtype targets", and the result will indicate which HPV subtype(s) from the pre-specified group are present in the patient's sample. 

In other words, the scope of the question goes from completely open-ended, to semi open-ended. The example below illustrates how to model the pre-specified, multi-target viral detection/identification tests.

New Variable Proposal

Impact on Controlled Terminology

1. Existing multi-target MBTESTs: NO deprecation.
2. New multi-target MBTEST requests: the MB team will continue to create new, pre-coordinated, multi-target tests when the team can fit all the targets into the MBTEST and the TEST is **meaningful**.
   1. CT rules on a "meaningful" pre-coordinated MBTEST:
      1. The MBTEST should clearly describe the microorganism(s) being assessed, either using full taxonomic name or scientifically accepted acronym/abbreviation.
      2. The MBTEST should clearly describe the antigenic target(s) using scientifically accepted full name or acronym/abbreviation.
      3. If excessive abbreviation/truncation is required to fit all targets into the MBTEST which makes the test less human-readable (and in turn a less useful SDTM mapping asset), or users are required to rely on the spelled-out CDISC synonyms to understand the intent of the value in MBTEST, a pre-coordinated MBTEST should be avoided.
      4. It is the MB team's responsibility to determine meaningfulness for a pre-coordinated MBTEST and whether or not to develop controlled tests, on a case by case basis. If the MB team is unable to develop a meaningful MBTEST, the team will provide mapping recommendations in the denied request message.
3. The post-coordination, additional variable approach will be recommended when a pre-coordinated and meaningful MBTEST is not possible, as in the case of the HPV multi-target test.
   1. CT Rules on post-coordinated MBTEST:
      
      1. The MBTEST should describe, on a high-level, and to the team's best ability, the overall type or taxonomic grouping commonality (if available) of the microorganism(s) being detected. Examples: MBTEST = HPV Type DNA (when the asseemsnet detects HPV Type 22+45+18+16+58+66 DNA); MBTEST = Coronavirus RNA (when the assessment detects SARS-related Coronavirus RNA+MERS RNA); MBTEST = Campylobacter DNA (when the test detects Campylobacter aviculae+Campylobacter bilis+Campylobacter corcagiensis+Campylobacter gracilis).
      2. For MBTEST, the abbreviation "MLTTRG (short for Multi-Target)" should be added to the end of the MBTEST, e.g. Coronavirus RNA, MLTTRG. For MBTESTCD, the letter "M" (which is equivalent to MLTTRG) is used at the end of the TESTCD, e.g. HRHPDNM.
      3. When "MLTTRG" appears in the MBTEST, this indicates a test that targets multiple microorganism(s) and/or microbial antigen(s) which have been post-coordinated. Users should map the specific multiple targets to --MOTRG and/or --MAGTRG.
      4. For unrelated multiple targeted microorganisms or when taxonomic commonality is difficult to determine, a generic MBTEST such as Bacteria, MLTTRG; Virus, MLTTRG, may be created. Example: MBTEST = Bacteria, MLTTRG (when test detects Shigella spp+Campylobacter spp+Clostridium difficile+EIEC+ETEC).

When to pre-coordinate vs post-coordinate MBTEST

1. For microbial detection/identification tests that target a "single" microorganism and a "single" specific antigen, the MBTEST should be pre-coordinated. (e.g. Clostridium difficile A Toxin, Hepatitis B Virus Core Antigen, Clostridium difficile tcdA DNA)
2. For microbial detection/identification tests that target "multiple" microorganisms and/or "multiple" specific antigens, the MBTEST should be:
   1. Pre-coordinated if a meaningful MBTEST can be created.
   2. Post-coordinated if a meaningful MBTEST is not possible, where the multiple pre-specified organisms should be mapped to MBMOTRG and the multiple pre-specified antigens should be mapped to MBMAGTRG.

Final Decision

Others

The Evolution of MBTESTs_updated 2022-05-23.pptx