Driver mutation detected in cerebrospinal fluid despite negative liquid biopsy results

Abstract

We report a case where a KRAS G12V driver mutation was identified in cerebrospinal fluid (CSF) but not in peripheral blood cell-free DNA (cfDNA) in a patient with advanced lung adenocarcinoma and significant central nervous system involvement. A 67-year-old man presented with hemoptysis and was diagnosed with stage IVB TTF-1-positive lung adenocarcinoma with brain and bone metastases. Standard chemotherapy was ineffective. While cfDNA analysis detected only a RAD21 mutation, CSF analysis revealed the KRAS G12V mutation. Despite the identification, no effective targeted therapy was available. This case highlights that CSF may be more suitable than peripheral blood cfDNA for detecting driver mutations in patients with predominant CNS lesions.

Introduction

Brain metastases are a common and severe complication in patients with advanced cancers, significantly affecting prognosis and quality of life. Primary tumors that frequently metastasize to the brain include lung cancer, breast cancer, melanoma, renal cell carcinoma, and colorectal cancer [1]. Among these, lung cancer is the most common cause of brain metastases, accounting for approximately 40–50% of cases [2].

The nervous system is divided into the central nervous system (CNS), comprising the brain and spinal cord, and the peripheral nervous system (PNS), which includes all neural elements outside the CNS. This distinction is crucial because the CNS is protected by the blood–brain barrier (BBB), a selective barrier that restricts the passage of substances from the bloodstream into the brain [3]. The BBB limits the effectiveness of systemic therapies and the detection of tumor biomarkers in the blood, making CNS metastases particularly difficult to diagnose and treat compared to tumors in the PNS [4].

In the search for targetable genetic alterations in advanced-stage lung cancer, gene panel analysis using specimens from primary tumor tissue is desirable due to higher detection sensitivity [5, 6]. Although gene panel analysis using cell-free DNA (cfDNA) from peripheral blood is less invasive and preferable for patient comfort, its detection sensitivity for mutations—particularly fusion genes—is lower compared to tissue analysis [5]. Furthermore, when lesions are confined to the CNS, the sensitivity of cfDNA obtained from peripheral blood may decrease even further due to the BBB restricting the release of tumor DNA into the peripheral circulation [6].

Cerebrospinal fluid (CSF) circulating tumor DNA (ctDNA) has emerged as a promising tool for the detection and characterization of CNS metastases. Since the CSF is in direct contact with the CNS, ctDNA derived from tumor cells within the CNS can be detected more readily in the CSF than in plasma [7]. This suggests that in cases with significant central nervous system involvement, CSF may be a more suitable specimen for genetic analysis than peripheral blood.

We report a case where no driver alterations were detected in plasma cfDNA, yet a KRAS G12V driver mutation was identified through gene panel analysis of CSF. This case highlights the potential utility of CSF ctDNA analysis in detecting actionable mutations in patients with CNS metastases when plasma cfDNA analysis is negative.

Case presentation

A 67-year-old man presented with hemoptysis but had no significant medical, family, or psychosocial history. Imaging studies revealed a 2 cm nodule in the right upper lobe (Fig. 1A). Bronchoscopic biopsy confirmed TTF-1–positive adenocarcinoma. Staging evaluations, including FDG-PET/CT and contrast-enhanced brain MRI, showed left-sided pleural effusion (Fig. 1B), pelvic bone metastases, and a small brain metastasis (Fig. 1C). Based on the 8th edition of the TNM classification, he was diagnosed with stage IVB lung adenocarcinoma. EGFR and ROS1 fusion genes were analyzed using RT-PCR, and ALK immunohistochemistry was performed; no targetable driver mutations were detected. The paraffin-embedded tissue was exhausted by these tests, and the PD-L1 tumor proportion score was 20%.

Fig. 1.

FDG-PET/CT scan showing a nodule with intense FDG uptake in the right B3 segment (A). Mild FDG uptake observed in left-sided pleural effusion (B). Contrast-enhanced brain MRI revealing a small enhancing nodule in the right frontal lobe (white arrow head) (C)

First-line chemotherapy with carboplatin, pemetrexed, and pembrolizumab was initiated. After 2 months, progression of the primary lesion was observed, prompting second-line treatment with docetaxel. Subsequently, the patient developed dizziness, and worsening of multiple brain metastases was noted (Fig. 2A, B). Whole-brain radiotherapy (30 Gy) was administered. Third-line therapy with pemetrexed and bevacizumab was introduced, but dizziness recurred. As standard chemotherapy was ineffective, comprehensive genomic profiling (CGP) was considered. The patient's performance status (PS) was 2, and re-biopsy was difficult. Because the patient had brain metastases since the initial presentation and exhibited a recurrence of dizziness, carcinomatous meningitis was suspected. Additionally, due to the decline in PS, the findings suggested not only worsening of brain parenchymal metastases but also possible meningeal dissemination and progression to meningitis. Unfortunately, the brain metastases had not been evaluated by contrast-enhanced MRI of the head at this time, and there was no imaging evaluation of meningeal dissemination.

Fig. 2.

Post-treatment contrast-enhanced brain MRI demonstrating progression of brain metastases (white arrow heads) (A, B). Cerebrospinal fluid cytology showing clusters of atypical cells with nuclear pleomorphism (C)

Peripheral blood and cerebrospinal fluid (CSF) were collected simultaneously. The CSF was divided into two samples: one preserved for cytological examination to confirm malignant cells, and the other for genetic analysis. Peripheral blood was collected using specialized collection tubes and submitted to FoundationOne Liquid (Foundation Medicine, Massachusetts, USA). After malignant cells were confirmed in the CSF, the CSF sample was submitted to the Lung Cancer Compact Panel (LCCP) (DNA Chip Inc, Minato-ku, Tokyo) for genetic testing. FoundationOne Liquid was performed using peripheral blood cfDNA, detecting only a RAD21 W23* mutation with a variant allele frequency (VAF) of 1.5%; no other mutations were found. Simultaneously, the Lung Cancer Compact Panel (LCCP) was conducted using 2 mL of cerebrospinal fluid (CSF). The CSF was clear and colorless with a cell count of 25 cells/µL; adenocarcinoma cells were detected (Fig. 2C). Using CSF in the LCCP, a KRAS G12V mutation was detected with a VAF of 23.3%. Genetic analysis of the primary tumor was not possible, and KRAS G12V could not be confirmed in the primary lesion. However, KRAS G12V is an established driver mutation in lung cancer, and its identification through CSF analysis was considered successful.

Regrettably, molecular-targeted agents against KRAS G12V were still under development, and no promising treatment options were available. The patient passed away approximately 1 month after the genetic analysis due to carcinomatous meningitis. The results of FoundationOne Liquid arrived about one and a half months after sample submission, whereas the LCCP results were available in about 1 week.

Discussion

Although liquid biopsy is a minimally invasive test, its sensitivity does not surpass that of genetic analysis using tissue from the primary tumor [5]. Furthermore, when lesions are confined to the central nervous system (CNS), the sensitivity of cfDNA obtained from peripheral blood may decrease even further [6, 8]. In such scenarios, using cerebrospinal fluid (CSF) instead of peripheral blood cfDNA may overcome issues related to sensitivity. In this case, we successfully detected driver mutations by utilizing nucleic acids from CSF rather than from peripheral blood.

Recent studies have demonstrated the clinical utility of CSF ctDNA analysis in both primary CNS tumors and CNS metastases from non-CNS primary tumors. CSF ctDNA analysis provides a more comprehensive and accurate representation of the tumor’s genetic landscape within the CNS compared to plasma cfDNA [7, 9]. This has important implications for guiding targeted therapies and improving patient outcomes. Our case contributes novel information by highlighting the detection of a KRAS G12V mutation in the CSF of a patient with CNS-dominant metastatic lung adenocarcinoma—a mutation less commonly reported in CSF studies. To our knowledge, this is among the few reported cases where a KRAS G12V mutation was detected in the CSF when plasma cfDNA analysis was negative and tumor tissue was unavailable for genetic analysis. This emphasizes the potential of CSF ctDNA analysis in uncovering actionable mutations that might be missed by plasma cfDNA testing.

In the Lung Cancer Compact Panel (LCCP) used in this study [10], liquid specimens containing cancer cells were directly added into GM tubes, up to a maximum volume of 5 mL, from which nucleic acids were extracted. Therefore, nucleic acids were extracted not from cfDNA in CSF but from liquid specimens containing malignant cells. This approach may have been advantageous for detecting tumor cell-derived driver mutations compared to cfDNA from peripheral blood. Currently, among gene panels approved for clinical use, none are authorized for the use of CSF specimens. Even in LCCP, the use of CSF specimens is largely experimental; however, there have been reports where the same driver mutations as in the primary tumor were detected from CSF specimens containing malignant cells [10]. Recent studies support the established clinical utility of CSF ctDNA analysis in CNS tumors [7, 9, 11]. Therefore, we acknowledge that CSF ctDNA testing is a valuable and clinically relevant tool in the management of patients with CNS involvement, and not merely experimental.

In diagnosing meningeal dissemination and carcinomatous meningitis, the gold standard is examining the presence or absence of malignant cells in the CSF obtained via lumbar puncture [12]. Meningeal dissemination is classified as class I when malignant cells are detected in the CSF, and as class II when malignant cells are not found in the CSF but are suspected based solely on imaging studies, primarily contrast-enhanced MRI of the brain [12]. In gene panel testing using CSF, the detection of driver mutations is expected in cases of class I meningeal dissemination. Indeed, in our previous study [10], the same driver mutations as those in the primary lesion were detected in all class I cases. In this case, the patient exhibited symptoms of dizziness, and meningitis due to the spread of malignant cells into the cerebrospinal cavity was suspected. Therefore, gene panel testing utilizing CSF may be valuable in class I meningeal dissemination cases, including those with such symptoms.

KRAS mutations are among the most common oncogenic drivers in non-small cell lung cancer (NSCLC) [13]. The KRAS gene encodes a GTPase involved in cell signaling pathways that regulate cell proliferation, differentiation, and survival. Mutations in codon 12, such as KRAS G12V—which results from a glycine-to-valine substitution—lead to constitutive activation of the KRAS protein, promoting oncogenesis by continuously stimulating downstream signaling pathways [14]. Historically, KRAS mutations have been considered “undruggable” due to the protein's high affinity for GTP/GDP and the lack of suitable binding pockets for inhibitors, making it challenging to develop effective targeted therapies [15]. Unlike other driver mutations in NSCLC, such as EGFR mutations or ALK rearrangements, KRAS mutations have lacked effective targeted treatments for many years. Recent advancements have led to the development of KRAS G12C inhibitors like sotorasib and adagrasib, which have shown promising results in clinical trials [16, 17]. However, these agents are specific to the KRAS G12C mutation and are not effective against other variants like KRAS G12V.

Patients with KRAS G12V mutations currently have limited treatment options and often rely on conventional chemotherapy or immunotherapy, which may have suboptimal efficacy [18]. The identification of a KRAS G12V mutation in our patient underscores the aggressive nature of this mutation and the unmet need for effective therapies targeting KRAS G12V. This case highlights the importance of ongoing research to develop inhibitors for KRAS mutations beyond G12C. The detection of the KRAS G12V mutation in the CSF emphasizes the potential utility of CSF ctDNA analysis in uncovering actionable mutations in patients with CNS metastases when plasma cfDNA analysis is negative. Given the challenges associated with the BBB and the lower sensitivity of plasma cfDNA in detecting CNS tumor-derived mutations, CSF analysis may provide a more accurate representation of the tumor's genetic landscape in such patients.A limitation of this report is that we could not analyze peripheral blood cfDNA and CSF using the same gene panel, making it impossible to accurately compare the sensitivity of each panel. This is because FoundationOne Liquid is not suitable for analysis of CSF, and LCCP is not suitable for analysis of peripheral blood cfDNA, making such comparisons challenging even in research settings. Furthermore, since the extraction methods for cfDNA from peripheral blood and DNA from CSF differ, this represents a limitation that makes direct comparison difficult. Second, although genetic analysis of the primary tumor could not be performed due to depletion of tumor tissue samples, KRAS G12V is a well-established driver mutation in lung cancer [19], and this was thought to be the driver mutation in the primary tumor.

Conclusion

Our case demonstrates that CSF ctDNA analysis can effectively detect actionable genetic mutations, such as KRAS G12V, in patients with CNS metastases when plasma cfDNA analysis is negative. This adds to the current literature supporting the clinical utility of CSF ctDNA testing. Incorporating CSF analysis into the diagnostic workflow may enhance personalized treatment strategies and improve outcomes for patients with CNS involvement.

Author contributions

Kei Kunimasa: conceptualization, investigation, methodology, writing—original draft, writing—review & editing. Daisuke Aohara: investigation, writing—review & editing. kazumi nishino: writing—review & editing.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the author(s) used Chat-GPT 4o for English proofreading and spell check. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Ethics approval and consent to participate

The protocol was approved by Institutional Review Board in Osaka International Cancer Institute in accordance with the Constitution of Japan, the Act on the Protection of Personal Information. The institutional ethics committee approved the acquisition of all samples and all methods conducted. Informed consent statement for case reporting and publishing was obtained from the participants and their families. The patient and their family provided consent for the publication of anonymized data in an open-access journal. All data have been anonymized.

Competing interests

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Dr. Kunimasa reports honoraria for lecture from AstraZeneca, Chugai Pharma and Novartis; Dr. Nishino reports receiving grants from Ono, TAIHO, MSD, AbbVie, DAIICHI SANKYO, Amgen, Eisai, Sanofi, Janssen, Novartis, Pfizer, Eli Lilly, Merck, Takeda, Chugai, Merus and personal fees from AstraZeneca, Chugai, Nippon Boehringer Ingerheim, Eli Lilly, Roche, Novartis, Pfizer, Merck, Janssen, Bristol Myers Squibb and Nippon Kayaku. Dr. Aohara declares no conflict of interest.