Guideline On High-throughput Sequencing Method Performance Evaluation

Appendix 1
Guideline on High-throughput Sequencing Method Performance Evaluation
1.Introduction
This guideline is intended to guide the applicants to prepare and write the registration
application materials for the performance evaluation of tumor-associated gene
detection reagents, and also provide a reference for the technical review department
to review the registration materials during the technical review.
This guideline is aimed at the general requirements for the performance evaluation of
tumor-associated gene detection reagents. Applicants should determine whether the
content is applicable according to the specific characteristics of the product. If not
applicable, the reasons and the corresponding scientific basis should be specified. The
contents of the registration application should be enriched and refined based on the
specific features of the product.
This guideline is a guiding document for applicants and reviewers, but does not include
administrative matters involved in registration and approval process, and is not
enforced as a regulation. If there are other methods that can meet the requirements
of relevant laws and regulations, it can also be adopted, but it is required to explain
the reasons in detail, verify the scientific rationality, provide detailed research
materials and verification materials, and relevant personnel should use the guidelines
under the premise of complying with relevant regulations.
This guideline is formulated under the current regulations and standards system as
well as the current level of cognition. With the continuous improvement of laws,
standards, and the continuous development of science and technology, the relevant
content of this guiding principle will be adjusted in due course.

2. The Scope of Application
   The performance evaluation of tumor-associated gene detection reagents described
   in this guideline mainly refers to high-throughput sequencing, i.e. next-generation
   sequencing (NGS), also known as massively parallel sequencing (MPS), to conduct in
   vitro detection of tumor-associated gene variants in tumor cells in human tissues. NGS
   for detecting somatic mutations is widely used for molecular detection related to
   tumor diagnosis, including sequencing DNA/RNA of specific genes to find changes in
   mutant genes associated with clinical diagnosis and treatment of tumors. Tumor gene
   mutation types include point mutations, insertions, deletions, gene rearrangements,
   copy number variants, and other generalized gene mutations.
   IVD detection based on the principle of NGS sequencing may include the following
   steps: sample collection, processing and preservation, DNA extraction, DNA processing,
   library preparation, sequencing and base recognition, sequence alignment/mapping,
   mutation recognition, and filtering, variant annotation and interpretation, as well as
   the generation of the test report. At the same time, some products may also include
   the software part, but the above-related steps are not necessarily all included and
   should be judged according to the specific design process of the product. For each step
   of detection, the applicants need to establish specific acceptable quality assessment
   indicators and conformity standards in combination with product design and clinical
   significance. In addition, to meet specific intended uses of the product, applicants
   must determine the appropriate reagents, consumables, instruments, and software
   through scientific and appropriate detection performance studies. Based on the above
   considerations, any differences in the design and workflow of NGS detection products
   may result in different results, so applicants need to describe the relevant detection
   performance indicators clearly.
   The original intention of analytical performance evaluation is to put forward
   assumptions about effectiveness and safety-related issues of the product performance,
   and then to confirm it through research. NGS has advantages in sequencing flux and
   discovery of unknown gene mutations; but in the application and use of NGS
   technology, there are challenges of the related clinical sample collection and
   processing, NGS detection content, sequencing process, data analysis, result report,
   technical quality certification and verification, etc. in all aspects.
   This guideline will focus on detecting clinically significant somatic variation in solid
   tumors and ensuring high-quality sequencing results. Applicants should focus on the
   interests of patients, fully integrate clinical oncologists’ views on precise diagnosis and
   treatment, and fully consider the operability of promotion and application in China.
   Applicants can use a diverse combination of targeted gene detection. The information
   generated by the targeted genome may be used to diagnose classifications, guide
   treatment decisions, and/or provide prognostic evaluations for specific tumors, and
   the number of genes involved in different products may vary widely.
   This guideline applies to products that undergo a first registration declaration and
   related licensing changes. This guideline does not apply to the following detections:
   germline-based source detection, whole exome detection, tumor mutation burden
   (TMB) detection, non-targeted sequencing, de novo sequencing, microbial infectionassisted diagnosis, free DNA detection, direct-to-consumer sequencing, and fetal
   detection, microbial genome identification and drug resistance testing,
   preimplantation testing, disease risk (including genetic risk) assessment and prediction,
   direct RNA sequencing, screening, independent diagnostic purposes, tumor genome
   sequencing, healthy individual sequencing detection.
3. NGS Performance Evaluation
   3.1 Summary
   Summary mainly includes the intended use of the product, product description,
   description of biosafety, summary evaluation of research results, and introduction of
   similar products during the market.
   As a general principle of IVD testing, applicants should first define the intended use
   and testing performance of the declared product. The intended use of the product will
   directly affect the design and detection performance of the testing and the types of
   genes that are sequenced and/or reported. Applicants need to proactively determine
   the research indicators (e.g., accuracy) that should be performed and the criteria that
   each research indicator should meet. After the design and development processes are
   complete, verify that if it meets the predefined performance. If the test does not meet
   any of the predefined performance criteria, it should be modified and revalidated.
   Through repeated design, development and verification, until the test meets the set
   requirements. Throughout the process, applicants need to document all research
   protocols, research processes and results, and the rationale for each study design.
   Applicants should list the relevant parameters of the test sample level. It is
   recommended to be clear in the form of a list. See Attached Table 1 for details.
   The applicant shall provide the SOP file of the entire inspection process, and describe
   in detail the process of sample collection processing, library preparation, sequencing,
   data analysis, and result reporting included in the whole process of NGS technology
   testing. Bioinformatics analysis describes and records data processing and analysis,
   including all processes of mutation identification, filtering and annotation. Identify all
   software to be used, including sources (e.g., internally developed and third-party) and
   any modifications. It is recommended that all required software and/or database
   names and their functions be described in a list.
   When an applicant plans to develop a targeted genetic test reagent, it is necessary to
   determine its intended use and the number of genes to be tested, including the type
   of sample to be tested, the applicable population, and which types of test information
   will be evaluated and reported. Other factors affecting the design, verification and
   quality control of the test should also be considered. The reagent components
   contained in the product and the reagents, equipment, and consumables that are
   required but not provided should be specified in the reagent composition specification.
   3.2 Research Data for Major Raw Materials
   The NGS detection process mainly includes sample collection processing, library
   preparation, sequencing, data analysis, and result reporting.
   3.2.1 Sample Collection Processing (If Applicable)
   The sample collection process is to ensure that the sample has a key link that can
   accurately reflect the patient’s signs. Applicants are required to submit information on
   the main raw materials related to sample collection and processing related reagents,
   such as sample transport preservation reagents, sample preparation reagents, and
   main components of nucleic acid extraction reagents.
   3.2.2 Library Preparation and Sequencing
   The sequencing library and sequencing process mainly include deoxynucleotide
   nucleoside, linker sequence, ligase, polymerase, reverse transcriptase, restriction
   endonuclease, primer, probe, linker, etc.; if the applicant researches the main raw
   materials, the applicant should detail the experimental process of sequencing library
   construction and provide functional studies on each of the major raw materials. The
   finished product should also conduct quality testing to confirm that it meets the
   standard requirements, and the entire production process should be stable and
   controllable by the applicant. If the applicant purchases the main raw materials, it shall
   detail the source of each raw material purchaser, submit the performance indicators
   and quality control data of each raw material issued by the purchased party, and detail
   the quality requirements of the applicants for the main raw materials purchased, and
   the supporting details of the raw material purchased as the main raw material of this
   product. Packaging materials and consumables for nucleic acid detection reagents
   should be free of deoxyribonuclease (DNase) and ribonuclease (RNase) contamination.
   3.2.3 Bioinformatics Analysis and Database Requirements
   NGS’s data analysis process, or bioinformatics process can generally be divided into
   four main operations: base recognition, sequence alignment, mutation recognition,
   and mutation annotation. Clearly define the reference sequence types, the assisted
   sequencing comparative splicing, and the confirmation of sequencing results. Different
   mutation types may require the development of different mutation/mutation
   detection algorithms. Second, the scope of the software tools and the type of
   verification required are based on the type of product design. Applicants are advised
   to submit traceability information, database type, integrity, real-timing, maintenance
   and upgrade schemes for the database (reference sequence), as well as to analyze
   software versions, algorithms, performance verification, and upgrade plans.
   3.2.4 Internal reference products are one of the important components to ensure
   product performance stability and traceability of detection values. Reference product
   research should include raw material selection, preparation process, valuation studies,
   evaluation indicators, and statistical analysis. Applicants should conduct accurate
   experimental verification of the source of the internal positive/negative reference
   materials and gene sequence settings, and provide relevant information and detailed
   verification data of the measurement procedure or reference method of the reference
   product traceability process.
   Specific requirements are as follows:
   3.2.4.1 Positive Reference Products and Positive Control Products
   3.2.4.1.1 Positive Reference Product
   Ideally, a positive reference product should include a quality control sample for each
   genotype of the product being declared. However, considering the large number of
   detectable genes based on NGS technology, applicants should design targeted and
   representative products based on factors such as the intended use, clinical significance,
   and gene type.
   A positive reference product should include all types of genes that require a definitive
   medical treatment, such as a reference set for a site that has a clear guiding role in the
   use of the drug, and at least include representative gene types from different gene
   types. Clinical samples or the DNA stock extracted from the cell line should be used as
   a raw material.
   For genes with significant clinical significance or potential clinical significance, due to
   the representative problem, the types of genes with different clinical diagnostic
   significance and different types of mutations in the genotype should be identified. The
   selection of mutation frequency and variation type should be representative, including
   different Exon, different gene mutations, etc. If the detection target area is large
   (greater than 1M), the overall quality and sensitivity of the detection area should be
   quality-controlled (especially for SNV), and mixed immortalized normal human
   leukocyte DNA storage solution (HapMap cell line) can be used.
   The variation abundance and mutation frequency of different mutations/variations,
   the concentration range setting should be representative and provide the basis for
   such a setting. The mutated form and copy number of the positive reference product
   should be confirmed by an effective method, and the acceptance criteria should be
   clearly accepted. When designing a positive reference product, the applicant can
   consider the setting and confirmation method of the limit of the detection reference
   item and provide relevant basis.
   3.2.4.1.2 Positive Control Product
   Imitating the target nucleic acid sequence of a patient sample and use it to control the
   entire detection process, including nucleic acid extraction (if applicable), library
   construction, sequencing, and data analysis. At present, the process of positive control
   detection and analysis should be consistent with the clinical sample.
   3.2.4.2 Negative Reference Products and Negative Control Products
   The negative reference product should be a mixture of normal FFPE samples, and the
   mixture or cell line of the normal sample of FFPE, which should be clearly free of the
   target region tumor mutated gene and the germline mutated gene in the target gene.
   3.2.4.3 Precision Reference Products
   The setting requirements of precision reference product can refer to the
   positive/negative reference product.
   3.2.4.4 No Template Control PCR Reagent Product (NTC)
   Applicants should establish acceptable standards for NTC against Qubit testing.
   Monitoring the presence of contamination during the test.