Target areas of DNA extraction are the most important consideration for pathologists before NGS DNA preparation, and enrichment of tumor cells is vital when selecting the region of interest. Extracting DNA from locations below the standard for tumor cell fraction can lead to false negatives. If highly probable mutations are not found during the NGS, such as APC mutation in colorectal cancer, the tumor cell fraction at the DNA extraction point should be suspected of being too low. Furthermore, mixed normal cells with diploid genomes may dilute the amplification signal; therefore, thorough evaluation of tumor cell fraction and correction of the measured copy number are vital. Sample selection should therefore be made under close communication between clinicians and pathologists.

When comparing the primary and recurring/metastatic sites in each specimen, it is important to consider that primary tumors and metastatic tumors before treatment with systemic chemotherapy share a large proportion of key tumor-related genetic alterations [7–9]. Nevertheless, recurred or progressed tumors can be accompanied by genomic evolution. Although the majority of the additional genomic alterations are likely to be passenger [10], there can be therapeutically relevant alterations that mediate drug-resistance or predict patient prognosis. Therefore, re-biopsies at recently obtained relapse sites are more likely to reflect the solid tumor’s genetic features at the time of NGS. However, if clinical circumstances prevent the obtainment of recurring/metastatic tissue, primary tumors may be used.

Tumor purity, described as the proportion of tumor cells, is an important factor influencing the reliability of the examination because low tumor purity can potentially lead to false-negative outcomes. Suitable NGS samples should have the highest tumor purity and cellularity among biopsy and surgical samples, and surgical samples are generally preferred. However, the quantity of viable tumor cells in the surgical sample should be considered during selection, which may be too low if neoadjuvant therapy was conducted. The minimum requirement for tumor purity varies according to the NGS analysis method and technical sensitivity; however, a higher content of nucleic acids and tumor cells with higher yields reduce the chance of false negativity. Regardless, some samples such as sclerotic or cystic tumors tend to have low cellularity or may contain necrotic regions that can affect nucleic acid yield. All factors should therefore be reviewed if a false-negative result occurs. NGS requires a minimum quantity of 100 ng of DNA for hybrid capture and 10 ng for amplicon. Notably, hybrid capture requires at least 1 mm² of tumor tissue under exhaustive conditions, necessitating more than 2 mm (diameter) endoscopic biopsy tissues of three sections and at least two unstained slides of surgical tissue [11]. However, the minimum amount is dependent on the quality of the samples, so securing unstained slides at 1.5× the minimum requirement is recommended (i.e., 10 biopsy tissue slides, five surgical tissue slides). These criteria are reference only, establishing an optimized protocol tailored to the institution’s experience is advised.

Tissue storage and fixation is crucial to achieve high-quality samples suitable for NGS. Delayed or insufficient fixation in surgical samples, in particular, may lead to DNA degradation [12]; sufficient fixation can be achieved by placing the sample in buffered formalin suitable for the volume of the tissue within 1 hour of acquisition. Additionally, high-quality preoperative biopsy samples may be used if surgical tissue has been improperly fixated. The possibility of cytosine deamination (C:G > T:A mutations) causing background noise should be considered if tissues are stored for long periods under formalin-fixed paraffin-embedded conditions [13,14].

Liquid biopsies have been introduced into clinical practice for non-invasive genome analysis, treatment response monitoring, identification of drug-resistant mechanisms, early detection of recurrence, and overcoming intra-tumoral heterogenicity by supplementing the limitations of NGS examinations using existing tumor tissues [15].

Genetic aberrations detected by NGS are classified according to clinical usefulness and then applied to treatment. Although a scoring system for clinical usefulness has been established in several countries, the need for Korea-specific criteria has been highlighted [16,17]. Therefore, this current statement proposes the KPMNG scale of Clinical Actionability of molecular Targets (K-CAT), referring to the existing tier system and considering its applicability to Korean patients (Tables 2 and 3). Level 1 of this scale refers to recognition of a therapeutic agent as a standard treatment for a specific target gene alteration, approved by the Ministry of Food and Drug Safety (MFDS) in Korea or corresponding international regulatory agencies (U.S. Food and Drug Administration [FDA], European Medicines Agency [EMA], etc.) or supported by prospective, randomized, phase III trials showing the benefit of survival endpoints. Treatment drugs targeted for genes corresponding to Level 1 should be subject to reimbursement application. Level 2 denotes a target gene with clinical benefits in prospective phase 1 or 2 studies. A level 2 actionable gene needs to be considered subject to MFDS approval and potential off-label drug use after approval from the Health Insurance Review and Assessment Service (HIRA). Level 3 denotes a gene with potential for clinical benefit, confirmed by retrospective outcomes of specific therapeutic agents in patients with corresponding genes in the same or different indications. Oncologists may consider clinical trial participation or applying for the Expanded Access Program for level 3 results. Level 4 is defined as securing evidence for a therapeutic drug at the laboratory level for the corresponding genes with limited immediate application to clinical treatment decisions. Level G is classified as indication of suspected actionable germline variant, for which following germline tests are recommended. Level R indicates a drug-resistant gene for a specific targeted anticancer drug. This classification of resistant genes can also be used as conditions for insurance coverage and approval for anticancer drugs.

K-CAT aimed to present methods and grounds for NGS results for application to treatment decisions reflecting the domestic situation in Korea. Gene classification depends on whether studies were prospective, were approved by the MFDS, the magnitude of survival benefits, data on other indications and the preclinical data. Table 3 presents the criteria reflected in each scale. K-CAT list of genes for each cancer type will be addressed separately in the next series of KPMNG guidelines.

The testing environment for the NGS differs by each institution in terms of sequencers (Illumina NextSeqDX or Thermofisher ion S5), chemistry used in panels (amplicon or hybrid capture method), test range (number of genes included in the panel and kind of variants to be detected), and detection criteria of variants. Regarding that detection results may vary in the boundary conditions, it requires further caution when comparing and interpreting the results provided by different institutions.

From sample collection to the analysis step, the final results can be affected by diverse variables [18]. To be specific, the quantity and quality of tumor samples as well as the NGS library preparation, sequence alignment, and analysis software can cause the variables, which can lead to various false positives in the results. Although such false positives are filtered out as much as possible throughout both analysis and manual curation steps, some indecisive variants may be included in the final reports as “variants of uncertain/unknown significance (VUS).” Therefore, VUS should be interpreted with further caution.

Above all, since most NGS tests do not use matched normal samples, germline and somatic variants cannot be easily distinguishable, and therefore potential germline variants are possibly included in the results. Hence, if a pathogenic germline variant is suspected, it is recommended to confirm the variant in normal tissues or blood samples via appropriate genetic counseling.

In general, the limit of detection (LOD) to report variant detection is set to a variant allele frequency (VAF) threshold of 3%–5% [18]. However, in cases of tumor purity is low, the final results may include significant variants (such as hotspot or oncogenic) below the LOD (1%–2%). The VAF of the detected somatic variants should be considered based on the tumor purity level. In particular, the purity level and VAF have similar distributions. If VAF is approximately half of the tumor purity level, the variant is considered a heterozygous loss variant. Variants with different levels of the VAF other than the aforementioned cases can be interpreted as variants that emerged from subclones.

Fusions are fundamental markers for cancer diagnosis, patient prognosis, and targeted therapy. Most fusion breakpoints occur between intronic regions. For accurate detection of fusions, gene panels should be designed with optimization to cover either whole regions of a gene or specific exonic/intronic regions where fusions frequently occur. However, intronic regions are hardly covered by the sequencing due to repeated sequences. In this respect, no fusion is detected if breakpoints are located in the uncovered regions. To overcome these limitations, the use of RNA panels has been gradually expanded recently [19].

In most cases of copy number variations (CNVs), only variants with amplifications or homozygous deletions (0 copy) are reported according to the reliability of the results, and variants with heterozygous deletions (1 copy) are usually not reported due to possible experimental errors. Additionally, thresholds for genes that are frequently amplified (ERBB2 and MET) and those that are not frequently amplified are flexibly applied, respectively. In the “frequently amplified” genes, it can be suspected to have an amplification even though the copy number is low. Thus, in these cases, confirmation by comparing results obtained from alternative methods (such as fluorescent in situ hybridization) is highly recommended. Besides, CNVs are affected by the number of target probes and amplicons, theoretically. Thus, the copy numbers estimated by using single probes (usually targeting the hotspot regions of a gene) are not accurate as of the average estimation from probes that covers the whole exonic regions.

Although there is no standardized numerical definition for the number of collective tumor mutations per 1 Mb, tumor mutational burden (TMB) has arisen as new biomarker for immune checkpoint inhibitors [20–22]. Regarding that the types of variants included in the TMB definition and calculation are varied as “missense variants only,” “including nonsense and frameshift variants,” or else, it is necessary to check the definition of TMB applied in the panel test. Furthermore, since the thresholds to classify as “TMB-high” are varied from 7.4 mutations per Mb (mut/Mb) to 20 mut/Mb depending on cancer types, it is important to confirm the appropriate LOD for each type of cancer.

Tumors with high microsatellite instability are another key factor for immunotherapy along with TMB. To measure and determine the stability of microsatellite instability, various markers (such as BAT-25, BAT-26, D2S123, D5S346, D17S250, NR21, NR24, and NR27) are used. Since detection criteria are varied by each panel type, the LOD should be thoroughly reviewed [23–26].

MTBs are consultative bodies of cancer experts, including medical oncologists, pathologists, and bioinformaticians who gather within the institution/hospital network to share clinical insights and establish optimal treatment strategies based on complex genetic results [27]. Essential personnel for operating MTBs includes medical oncologists, pathologists, and bioinformaticians, and ideally include clinical geneticists, tumor biologists, and clinical research coordinators [28]. However, realistically, many institutions would not be able to operate MTB with such a configuration [29]. Therefore, each hospital should first organize an MTB with available experts, constituting at least a pathologist and an oncologist clinician. If institutional circumstances prevent the MTB from operating at complete configuration, a virtual MTB using an online platform could be an alternative [30,31].

The MTB aims to discuss potential therapeutic options tailored to the results of genetic analyses in patients with advanced solid cancer through a multidisciplinary approach. Essential discussions around the case presented in MTB concern: (1) quality assessment of the NGS, (2) review of clinically significant genetic abnormalities found in the NGS test, and (3) the recommended treatment based on the NGS results [27,28]. Reviewing the use of an appropriate sample and correct test methods comprise the discussions concerning quality within the MTB. Moreover, reviewing appropriate sample selection, cellularity, and mean target depth is important to assess the quality of the NGS. When clinically significant genetic abnormalities arise from the NGS, single nucleotide variants, CNV, and structural variation should be examined. Additionally, when interpreting each genetic aberrations, knowledge databases such as OncoKB (https://www.oncokb.org), CIVic (https://civicdb.org/home), cBioportal (https://www.cbioportal.org/), My Cancer Genome (https://www.mycancergenome.org/), Clinvar (https://www.ncbi.nlm.nih.gov/clinvar/), COSMIC (https://cancer.sanger.ac.uk/cosmic) may be consulted. Other important considerations include addressing findings that may be suspect of microsatellite instability, drug-resistant mutations, and germline mutations. The TMB should also be reviewed to be reflected in the treatment options. The full treatment spectrum should be considered at final treatment selection, including regulatory approved standard treatment, off-label drug use, clinical trials, and the Expanded Access Program. Decision-makers should also refer to crucial factors such as the patient’s previous treatment history, systemic medical condition, and disease course.

The results of the MTB discussion should be recorded in a unified and standardized report, including the patient’s basic information, minutes on the discussion and interpretation of NGS results, and treatment recommendations [27,28]. A virtual MTB report from the Korea Precision Medicine Networking Group Study (KOSMOS) on genomic mutation-based customized drug therapy for patients with advanced solid cancer is presented in Fig. 1.

An approved standard treatment with a high level of evidence matching NGS-determined genetic variation should be considered preferentially in the MTB. The K-CAT could be used to prioritize recommendations in the event of many identified genetic variations. A medical oncologist should make shared decision making among the proposed treatment considering the patient’s systemic condition, cancer type, disease course, comorbidities, and socioeconomic factors.

If a drug recommended by the MTB based on NGS results is not approved for the particular cancer type, drugs included in the insurance coverage list may be used with the approval of off-label drug use after the review by a multidisciplinary committee of six or more experts, including medical oncologists at each institution, followed by application for HIRA’s approval. Upon prior application, approval for use will be notified after deliberation by the multidisciplinary committee within 60 days of the date of receipt. Patients requiring prompt treatment outside of the pre-approval review period for chemotherapy may apply for post-approval (use before approval) [32].

In addition, if the drug recommended by the MTB is not approved in Korea or does not comply with off-label drug use application conditions, the patient may participate in clinical trials or try the Expanded Access Program. Information regarding clinical trials conducted in Korea may be searched online at the MFDS website (https://nedrug.mfds.go.kr/index), Korea Disease Control and Prevention Agency CRIS (https://cris.nih.go.kr/cris/index/index.do), and KCSG website (https://www.kcsg.org). The Expanded Access Program is a system that allows patients with cancer who can no longer receive standard treatment or participate in clinical trial, to use drugs that are currently under clinical development according to the discretion of an experienced physician after obtaining approval from the MFDS. Since the indications do not have to match those of clinical trials, this process may be considered if medical judgment determines that the benefit of using clinical trial drugs outweighs the risk, with willingness from the pharmaceutical company [33].

To further improve molecular profiling guided therapy, development of better decision-making aids based on the Korean health care situation is needed, as well as a platform for shared decision-making with patients and families [34–36]. Furthermore, it will be essential to build a nationwide clinico-genomic database, integrating genomic data and clinical data including tumor and patient characteristics, prescription data with corresponding treatment outcomes including survival [37–39]. Data privacy protection and appropriate data-sharing strategies should be accompanied to make clinico-genomic database usable to all stakeholders for future research, drug development, practice, drug approval, and reimbursement. Importance of integration and utilization of already established nationwide databases cannot be overemphasized [40].