A High Percentage of NSCLC With Germline CHEK2 Mutation Harbors Actionable Driver Alterations: Survey of a Cancer Genomic Database and Review of Literature

Introduction Germline CHEK2 mutations are rare and have not been associated with increased risk of NSCLC. Methods We identified two sequential primary NSCLCs harboring distinct actionable driver alterations (EGFR E746 _S752 delinsV and CD74-ROS1) in a patient with NSCLC with a novel germline CHEK2 mutation S5fs∗54 (c.14_20delCGGATGT). We queried a genomic database of NSCLC samples profiled by plasma next-generation sequencing (Foundation Medicine Inc.) and performed a literature search of germline CHEK2 mutations in NSCLC. Results Of 6101 patients with unique NSCLC profiled by plasma next-generation sequencing, 53 cases (0.87%) of germline CHEK2 mutation were identified (male-to-female ratio, 49%:51%; median age = 75 y). The median allele frequency of CHEK2 was 49% (interquartile range: 49%–51%). Ten unique CHEK2 germline mutations were identified. Literature review identified 15 additional cases of germline CHEK2 mutations in NSCLC. Overall, a total of 70 CHEK2 germline mutations (21 unique CHEK2 alterations) were identified. Among these 70 CHEK2 germline mutations, 54.3% were amino acid substitutions (point mutation), 40.0% were frameshift mutations, and 5.7% were splice site mutations. Of these 70 total cases assessed, 29 (41.4%) potentially actionable driver alterations were identified with KRAS G12C mutation (27.6%) being the most common and KRAS G12A/C/D/R/S/V mutations together constituting 51.7% of these driver mutations. Conclusions Germline CHEK2 mutations are rare in NSCLC. A large proportion of these cases harbor actionable driver alterations. The relationship between germline CHEK2 mutations and actionable driver alterations in NSCLC may be worth further investigation.


Introduction
The CHEK2 gene encodes checkpoint kinase CHK2, activated mainly in response to DNA double-stranded breaks. There are several germline hotspot mutations (IVS2þ1G>A; 1100delC; I157T) in CHEK2 that are associated with hereditary breast and prostate cancers. 1 Nevertheless, no germline CHEK2 (gCHEK2) mutations (gCHEK2m) have been associated with hereditary lung cancer and there is very limited literature on gCHEK2m and NSCLC. During routine clinical care, we identified a patient with NSCLC with a rare gCHEK2m who developed two sequential primary lung cancer with different actionable driver alterations, which prompted us to investigate the relationship between gCHEK2m and actionable driver alterations in NSCLC.

Materials and Methods
We queried the Foundation Medicine genomic database of NSCLC plasma samples that were profiled by next-generation sequencing (NGS) to assess frequency of CHEK2 mutations and any associated actionable driver alterations. CHEK2 germline status was determined using a previously described research algorithm. 2 We further performed a PubMed search for gCHEK2m in NSCLC.

Results
Our index patient is a 64-year-old never-smoker Caucasian woman diagnosed with having a T3 (two separate nodules) N0M0 NSCLC (adenocarcinoma) at 59 years old and underwent a curative right upper lobe (RUL) lobectomy. No molecular profiling was performed at the time owing to early stage disease. She received four cycles of cisplatin/pemetrexed chemotherapy followed by regular surveillance computed tomography (CT) scans. After 4 years, she presented with chest pain. A 3.3 cm left breast mass and confluent left axillary lymphadenopathy were found on a CT scan, and the biopsy result revealed moderately metastatic adenocarcinoma with mucinous features with immunohistochemistry staining consistent with lung primary. Results of complete staging with magnetic resonance imaging of the brain and positron emission tomography scan revealed only the left breast, axillary, and mediastinal lymphadenopathy (Fig. 1A). Result of initial limited molecular profiling revealed ROS1 fluorescence in situ hybridization positivity. The patient was assumed to have developed metastases from the previous RUL adenocarcinoma and was started on entrectinib 600 mg daily. The patient transferred care to our institution, and results of comprehensive tissue DNA and RNA NGS (Caris Life Science, Phoenix, AZ) detected a CD74-ROS1 fusion variant (C6, R34). A CHEK2 S5fs* mutation was detected in the tumor (allele frequency [AF]: 42%). Plasma genotyping (Foundation Medicine, Boston, MA) revealed CHEK2 S5fs*54 (c.14_20delCGGATGT) at an AF of 50.2% consistent with a heterozygous germline mutation (Fig. 1B). Family history revealed one cousin with breast adenocarcinoma with the same CHEK2 mutation and her daughter was found to be a CHEK2 mutation carrier. Genetic testing was suggested to our patient and her siblings because there is a 50% chance of her siblings being carriers which would increase their children's chance of cancer.
Because of the atypical presentation of CD74-ROS1 NSCLC with the left breast and lymphadenopathy without a clear primary tumor, we requested retrospective DNA and RNA NGS of the previously resected RUL sample (Caris Life Science, Phoenix, AZ) which revealed EGFR E746 _S752 delinsV (AF: 16%) and no evidence of CD74-ROS1. CHEK2 S5f* mutation AF was 81%. Thus, our patient number (#)1 has two separate NSCLC primaries separated by approximately 4 years.
We then further identified a second 70-year-old female Caucasian never-smoker patient with a germline CHEK2 T367fs*15 (1100delC) mutation. The patient was diagnosed with having a 1.3 cm left upper lobe welldifferentiated NSCLC with KRAS G12C when she was 66 years old. She underwent left upper lobe lobectomy and continues having CT surveillance without recurrence. In addition, her CHEK2 mutation (AF: 69%) was confirmed to be germline by Ambry Genetics (Aliso Viejo, CA) hereditary cancer plasma genotyping assay. Details of both cases are summarized in Figure 1C.
These two cases prompted us to further investigate the incidence and correlation between CHEK2 mutations and actionable driver alterations in NSCLC by querying the Foundation Medicine genomic database. Of 6101 patients with unique NSCLC profiled by plasma genotyping, 53 cases (0.87%) of gCHEK2m were identified (male-to-female ratio, 49%:51%; median age ¼ 75 y). Ten (excluding our case #1) unique CHEK2 germline mutations were identified, including most often I157T and T367fs*15 (each 36% of cases). The median gCHEK2m AF was 49% (interquartile range: 49%-51%) ( Fig. 2A). There was one CHEK2 mutation (E457fs*33) predicted to be homozygous (AF ¼ 67%). Among these 53 gCHEK2m, 49 occurred in samples with detectable circulating tumor DNA (ctDNA). Actionable driver alterations (KRAS G12C  with gCHEK2m. In addition, six potentially actionable co-occurring KRAS mutations were identified (G12D Finally, survey of the literature identified two case reports and one family with homozygous CHEK2 mutation with multiple cancers, including NSCLC, [3][4][5] and two germline surveys of CHEK2 mutations in NSCLC. 6,7 Tian et al. 6 identified seven gCHEK2m among 1764 Chinese patients (0.40%) with NSCLC. Liu et al. 7 identified four among gCHEK2m 1026 Chinese patients (0.39%) with NSCLC. Combining the two Chinese survey studies and the Foundation Medicine database, gCHEK2m were found in 0.72% (64 of 8891) of NSCLC. Overall, literature survey identified 15 additional NSCLC gCHEK2m cases, 11 in the Chinese surveys and four additional reported cases (Table 1). Furthermore, six of the 15 cases had actionable driver alterations (either EGFR or KRAS G12C mutation) ( Table 1). Overall, a total of 70 gCHEK2m were identified (53 in the Foundation Medicine data set, 15 in the literature, and two reported herein), representing 21 unique mutations, the most common of which were T367fs*15 (29%) and I157T (27%) (Fig. 2B and C). Among the 70 gCHEK2m, 54.3% were amino acid substitutions (point mutations), 40.0% were frameshift mutations, and 5.7% were splice site mutations. Of these 70 total cases assessed, a total of 29 (41.4%) potentially actionable driver alterations were identified; KRAS G12C mutation (27.6%) was the most common and KRAS G12A/C/D/R/S/V mutations constituted an additional 51.7% of driver-positive cases (Fig. 2D). We did not consider KRAS V14L, D33E, and F156L actionable alterations currently.

Discussion
We believe that this is the most comprehensive study of gCHEK2m in NSCLC to date. Combined survey of plasma samples from a commercial genomic database and survey of the literature revealed that gCHEK2m are rare (<1%) in NSCLC. Furthermore, the germline alterations occurred throughout the entire CHEK2 gene and consisted of amino acid changes, frameshift mutations, and splice site mutations. Importantly, more than 40% NSCLC with gCHEK2m harbored actionable drivers with proven therapeutics treatments that could potentially extend to 51% if considering other KRAS G12 isotype mutations.
The CHEK2 mutation 1100delC and CHEK2 I157T mutations associated with hereditary breast cancer are the two most common CHEK2 mutations identified in NSCLC. 1,8 CHEK2 S5Lfs*54 from our patient #1 is extremely rare, limited information has been published in the literature listed by the National Library of Medicine, and it is only reported four times in the National Library of Medicine database by commercial sequencing companies, 9 including in a survey of hereditary genetic mutations potentially predisposing to pancreatic adenocarcinoma. 10 This frameshift mutation resulted in a deletion of seven bases with a frameshift starting at the fifth amino acid and resulted in a stop codon downstream by 54 bases (S5fs*54), and it is expected to result in loss of function by premature protein truncation or nonsense-mediated mRNA decay and is considered pathogenic.
The initial NSCLC of our patient #1 was found to harbor EGFR exon 19 deletion mutation in retrospect. At the time of diagnosis, the standard of care was four cycles of adjuvant chemotherapy. The recently published ADAURA trial indicated that 3 years of adjuvant osimertinib significantly improved 2-year disease-free survival rate to a similar extent with or without adjuvant cisplatin-based chemotherapy. 11 Thus, our patient #1, if diagnosed in 2021, could potentially have avoided adjuvant cisplatin-based chemotherapy. 12 We wonder if the cisplatin-based chemotherapy may have promoted the genesis of the second ROS1þ NSCLC given that cisplatin induces DNA strand breaks and the gCHEK2m reduces the corrective response to DNA breaks.
Plasma genotyping will identify gCHEK2m at AF 50% as shown in Figure 2A while tumor genotpying the AF of gCHEK2m could vary widely though a high AF should alert clinicians to the potential presence of gCHEK2m. One of the limitations of this study is that not many patients with NSCLC undergo plasma NGS genomic profiling, not all plasma NGS genomic profiling assays include CHEK2, and not all NGS assays sequence the whole CHEK2 gene. Thus, the frequency of gCHEK2m among NSCLC may be underestimated in the literature. Given the predominance of EGFR mutations among Asians, a comprehensive NGS may not be routinely performed. Indeed, no KRAS mutations were reported among Chinese patients with NSCLC with gCHEK2m. Furthermore, actionable driver alterations could only be detected in patients who are tumor ctDNA "shedders," which is often a surrogate of tumor burden. For example, in our patient case #1, the ROS1 fusion was not detected in the plasma as she was already receiving effective treatment at the time of plasma genotyping.
In summary, gCHEK2m are rare in NSCLC but they span the entire CHEK2 gene with 21 unique mutations identified comprising point mutations, frameshift mutations, and splice site mutations. At least 40% of these NSCLC with gCHEK2m harbor diverse actionable driver alterations. Whether there should be increased surveillance for lung cancer in gCHEK2 carriers remained to be determined given that NSCLC has rarely been reported in