KRAS A146 Mutations Are Associated With Distinct Clinical Behavior in Patients With Colorectal Liver Metastases

PURPOSE Somatic KRAS mutations occur in approximately half of the patients with metastatic colorectal cancer (mCRC). Biologic tumor characteristics differ on the basis of the KRAS mutation variant. KRAS mutations are known to influence patient prognosis and are used as predictive biomarker for treatment decisions. This study examined clinical features of patients with mCRC with a somatic mutation in KRAS G12, G13, Q61, K117, or A146. METHODS A total of 419 patients with colorectal cancer with initially unresectable liver-limited metastases, who participated in a multicenter prospective trial, were evaluated for tumor tissue KRAS mutation status. For the subgroup of patients who carried a KRAS mutation and were treated with bevacizumab and doublet or triplet chemotherapy (N = 156), pretreatment circulating tumor DNA levels were analyzed, and total tumor volume (TTV) was quantified on the pretreatment computed tomography images. RESULTS Most patients carried a KRAS G12 mutation (N = 112), followed by mutations in G13 (N = 15), A146 (N = 12), Q61 (N = 9), and K117 (N = 5). High plasma circulating tumor DNA levels were observed for patients carrying a KRAS A146 mutation versus those with a KRAS G12 mutation, with median mutant allele frequencies of 48% versus 19%, respectively. Radiologic TTV revealed this difference to be associated with a higher tumor load in patients harboring a KRAS A146 mutation (median TTV 672 cm3 [A146] v 74 cm3 [G12], P = .036). Moreover, KRAS A146 mutation carriers showed inferior overall survival compared with patients with mutations in KRAS G12 (median 10.7 v 26.4 months; hazard ratio = 2.5; P = .003). CONCLUSION Patients with mCRC with a KRAS A146 mutation represent a distinct molecular subgroup of patients with higher tumor burden and worse clinical outcomes, who might benefit from more intensive treatments. These results highlight the importance of testing colorectal cancer for all KRAS mutations in routine clinical care.


INTRODUCTION
Oncogenic KRAS mutations are highly prevalent in multiple cancers and drive cell differentiation and proliferation. 1 KRAS mutations stimulate KRAS to stay in its active state, thereby triggering the oncogenic signaling pathway. 2 Around 40%-50% of the patients with metastatic colorectal cancer (mCRC) harbor a somatic KRAS mutation. [3][4][5] In general, patients with a KRAS wild-type tumor have a better prognosis than patients carrying a KRAS-mutated tumor. 6,7 Moreover, KRAS mutation status is a predictive marker for poor response to anti-epidermal growth factor receptor (EGFR) monoclonal antibody therapy, 8 one of the options for systemic treatment for patients with mCRC. 9 Therefore, analysis of KRAS mutation status has been widely adopted in routine clinical practice. 10 It is known that the biologic characteristics of tumors, like cellular phenotypes and metabolomic characteristics, differ on the basis of the KRAS mutation variant and amino acid substitution. [11][12][13] In a substantial part of routine diagnostic KRAS tissue panels, only the most common driver mutations in KRAS codons G12 and G13 are tested, which are affected in 28% and 8% of all patients with mCRC, respectively. However, mutations are also commonly present in KRAS Q61 (2%), K117 (1%), and A146 (4%). 4,5,14 Here, we investigated clinical features like tumor load and overall survival of patients with mCRC with a somatic mutation in KRAS G12, G13, Q61, K117, or A146.

Patient Characteristics
Liquid biopsies of patients with histologically proven colorectal cancer (CRC) with isolated, previously untreated, initially unresectable colorectal liver metastases (CRLM) were collected in the ongoing multicenter ASSOCIATED CONTENT

Data Supplement
Author affiliations and support information (if applicable) appear at the end of this article.
Accepted on October 13, 2021 and published at ascopubs.org/journal/ po on November 17, 2021: DOI https://doi. org/10.1200/PO.21. 00223 phase III CAIRO5 trial (NCT02162563). 15 A total of 419 patients with CRLM, enrolled between November 2014 and July 2019, were evaluated in this study. For all patients, tissue KRAS mutation analyses were performed in the participating hospitals before randomization following routine clinical practice. Only those patients who were randomized for treatment with bevacizumab and chemotherapy consisting of 5-fluorouracil, leucovorin, oxaliplatin, and/or irinotecan were selected for the current study. Clinical followup was performed according to standard of care, including clinical review every three months as well as computed tomography (CT) imaging every six months, for patients with resectable disease and every two months for patients with unresectable disease, using the RECIST 1.1 for reporting. Follow-up was recorded until March 23, 2020. The study was performed in accordance with the Declaration of Helsinki and a medical ethical committee approved the trial, and all patients signed written informed consent for study participation and liquid biopsy collection.

Liquid Biopsy Collection
Liquid biopsies were collected before study treatment using a cell-stabilizing BCT tube (Streck, La Vista, NE) in the participating hospitals and shipped to the Netherlands Cancer Institute. Here, cell-free plasma was collected in a two-step centrifugation process: 10 minutes at 1.700 g followed by 10 minutes at 20.000 g, and stored at -80°C until further processing. Cell-free DNA (cfDNA) was isolated using the QIAsymphony (Qiagen, Hilden, Germany) with an elution volume set to 60 mL. The concentration of cfDNA was measured using the Qubit dsDNA High-Sensitivity Assay (TFS, Waltham, MA).

Liquid Biopsy Mutation Analyses
For patients with an established KRAS mutation on the basis of tumor analysis, liquid biopsy mutation analyses were performed using four droplet digital polymerase chain reaction (ddPCR; Bio-Rad, Hercules, CA) screening kits, namely ddPCR KRAS G12/G13 (#1863506), ddPCR KRAS Q61 (#12001626), ddPCR KRAS K117N (#10049047), and ddPCR KRAS A146T (#10049550). Table 1 in the Data Supplement shows the different amino acid variants detected by these assays. The ddPCR assays were performed according to the manufacturer's instruction, making use of 1 mL of the multiplex assay, 11 mL of the ddPCR supermix for probes (no dUTP), 9 mL of sample, and 1 mL H 2 O. All measurements were performed in duplicate and included a blank (nuclease-free water) and a positive control. Data were analyzed using the Quan-taSoft software version 1.6.6 (Bio-Rad, Hercules, CA). The number of mutant copies per mL plasma (MTc/mL) and mutant allele frequency (MAF) were used as outcome measures. For the cfDNA samples with a KRAS A146 mutation, orthogonal validation was performed using targeted deep sequencing, as described previously. 16 In brief, genomic libraries were prepared from 125 ng of cfDNA, following normalization, end-repair, A-tailing, adapter ligation, and PCR amplification. Target capture was performed using a panel consisting of 58 genes, covering 81 kb. Candidate somatic alterations across the region of interest were identified using VariantDx (Personal Genome Diagnostics, Baltimore, MD).

Radiologic Total Tumor Volume Quantification
For patients with an identified KRAS mutation on tumor tissue, pretreatment contrast-enhanced abdominal CT images were used for semiautomatic segmentation in the Tumor Tracking Modality of IntelliSpace Portal 9.0 (Philips, Eindhoven, the Netherlands). The liver itself and all metastases were segmented by two trained members of the research team and subsequently adjusted and verified by a radiologist specialized in abdominal pathology. All segmentations and related CT images were processed and analyzed with the SAS Viya analytical platform (SAS Institute Inc, Cary, NC) for volume quantification using the quantifyBioMed Images action. 17 This action calculates the total tumor volume (TTV) directly out of the segmentation from all tumors presented in the liver by determining the volume of one voxel and multiplying this volume with the number of voxels included in the tumor segmentation. A CT scan is built up by voxels, the three-dimensional equivalent of a pixel, and a voxel's volume depends on the pixel spacing and slice thickness attributes of the CT scan. The volume of the liver was calculated similarly, on the basis of the three-dimensional liver segmentation. In addition, the percentage TTV of the total liver volume, including TTV, was calculated. Furthermore, the radiologist registered the number of liver lesions.

Statistical Analyses
A Brown-Forsythe analysis of variance test using Dunnett's multiple comparisons was used for the liquid biopsy analyses. A one-way analysis of variance corrected for multiple comparisons using Tukey's multiple comparisons test was used for the volumetric analyses. A two-sided P-value of .05 was used as a cutoff for significance. A Mantel-Cox log-rank test using a Bonferroni-corrected threshold of P , .005 for significance was performed for the survival analyses. To determine the equivalence between ddPCR and sequencing circulating tumor DNA (ctDNA) levels, a Pearson correlation was used. Clinical patient characteristics were compared between carriers of different KRAS mutant variants using Fisher's exact tests. Univariate and multivariate Cox proportional hazards regression analyses were performed to analyze prognostic factors for overall survival, adjusted for potential confounders. Analyses were performed with Prism version 8 (GraphPad Software, Inc, San Diego, CA) and SPSS software version 27 (IBM, New York, NY).

High Plasma ctDNA Levels in Patients With KRAS A146 Mutant Tumors
We previously measured plasma ctDNA levels in 100 patients with CRLM and noticed remarkably high plasma ctDNA levels in patients harboring a KRAS A146-mutated tumor, an observation that warranted further investigation. 18 The current study investigated the liquid biopsy ctDNA levels for all 156 patients included.
Patients without a pretreatment liquid biopsy (N = 32) and patients carrying a tumor with a KRAS mutation that could not be detected by the ddPCR kits (N = 2) were excluded, leaving 122 ctDNA samples for liquid biopsy analyses (Data Supplement Figure 1). Liquid biopsy ddPCR analyses showed more MTc/mL plasma and a higher MAF for patients with KRAS A146-mutated tumors (N = 10, median MTc/mL = 35,338, median MAF = 48%) compared with patients carrying a different KRAS variant, for example, a KRAS G12 mutation (N = 92, median MTc/mL = 700, median MAF = 19%), see Figure 3A (MTc/mL) and Figure 3B (MAF). To ensure that these high plasma ctDNA levels were not because of the KRAS codon 146 ddPCR assay's test characteristics, we performed orthogonal testing using a targeted deep-sequencing approach. A strong confirmation of the high KRAS A146 ctDNA levels was observed, with a Pearson correlation (R 2 ) of 0.98 (95% CI, 0.96 to 1.00; P , .0001) between the ddPCR and sequencing MAF results ( Figure 2 in Data Supplement).
The high plasma ctDNA levels in KRAS A146 mutation carriers were not caused by DNA copy-number gains or focal amplification of the KRAS locus (see methods in Data Supplement). Moreover, the high plasma ctDNA levels in patients harboring a KRAS A146-mutated tumor were accompanied by high plasma ctDNA levels for other genes like TP53, TERT, and PIK3CA ( Figure 3 in Data Supplement), implying that high plasma ctDNA levels for KRAS A146-mutated tumors are associated with tumor burden.

Patients With KRAS A146-Mutated Tumors Have High TTV
As all patients in this study had liver-only metastases, total tumor burden could be assessed by measuring the pretreatment TTV. Since abdominal contrast-enhanced CT images could be used for segmentation, patients with a magnetic resonance imaging (N = 17) and positron emission tomography-CT (PET-CT) or non-contrastenhanced scans (N = 4) were excluded from the volumetric analysis. Other reasons for exclusion were technical errors in the segmentation software (N = 3), missing scans (N = 2), and incomplete scans (N = 4), leaving 126 patients for volumetric analysis. Figure 4A shows the absolute TTV and Figure 4B shows the relative TTV as percentage of the liver volume. Patients with a KRAS A146-mutated tumor have a significantly higher absolute and relative TTV (median TTV of 672 cm 3 and 24.5% of total liver volume) compared with patients with a KRAS G12 mutation (median TTV of 74 cm 3 and 4.1% of total liver volume; P = .036 and P = .053, respectively) and G13 mutation (median TTV of 55 cm 3 and 3.5% of total liver volume; P = .021 and P = .026, respectively). In addition, the median number of  Figure 4 in the Data Supplement). High TTV was also observed in patients with the less prevalent KRAS K117 mutation (median absolute TTV = 592 cm 3 , relative TTV = 24.1%). The volumetric results of the four most frequent G12-mutated residues (G12A, G12C, G12D, and G12V) did not differ significantly ( Figure 5 in Data Supplement).

KRAS A146-Mutated Tumors Are Associated With Poor Overall Survival
Patients with mCRC with a KRAS A146-mutated tumor showed a worse prognosis than patients with another KRAS mutation variant (Fig 5) 1). CRLM, colorectal liver metastases.  Table 3 in Data Supplement). After adjusting for these clinical characteristics (age, sex, sidedness, and performance status), the multivariable Cox regression analysis showed that only the KRAS alteration was an independent prognostic factor for worse overall survival. The reported contrast between the KRAS mutation variants showed that KRAS A146 was the only significant feature behind this observation (hazard ratio = 2.5; 95% CI, 1.4 to 4.6; log-rank P = .003). No indications of an association between the baseline patient characteristics and the KRAS mutation variants were found (Table 1). No significant differences were seen in overall survival between the four most frequently mutated G12 residues (G12A, G12C, G12D, and G12V; see Figure 6 in the Data Supplement). Furthermore, the location of disease progression showed similar patterns for the different KRAS mutation variants (Figure 7 in Data Supplement).

DISCUSSION
Oncogenic KRAS mutations occur in approximately 50% of patients with mCRC and are known to be predictive for      ). a A significant difference was observed between KRAS G12 and A146 (log-rank P = .0045), by the Mantel-Cox log-rank test using a Bonferroni-corrected threshold for every combination of P , .005 for significance. mCRC, metastatic colorectal cancer.
Clinical Impact of Mutant KRAS in CRC Liver Metastases Biologically, KRAS mutation variants display distinct metabolic profiles. Oncogenic KRAS can dysregulate cell metabolism via glycolysis and the following tricarboxylic acid cycle. Enhanced glycolysis of cancer cells generating lactate even when exposed to abundant oxygen (Warburg effect) 29 is shown to be upregulated via oncogenic KRAS. [30][31][32] This Warburg effect is marked by low levels of ATP. However, a human CRC cell line study revealed distinct metabolic profiles of different KRAS mutation variants. Where for most KRAS variants the nucleotide imbalance is shifted toward a decrease in ATP and other nucleotides like guanosine triphosphate (GTP), cell lines harboring the KRAS A146 mutation displayed increased levels of these nucleotides. 12 The distinct metabolic profiles among KRAS mutation variants do not directly explain the observed differences in clinical outcome. Future research is needed to examine whether KRAS A146-mutated tumors are a distinct metabolic subgroup for which other therapeutic targets might be beneficial.
The differences in biology and consequently clinical outcome between KRAS mutation variants observed in this study might originate from the molecular mechanism of oncogenic activation. KRAS changes between two nucleotide-binding states, the inactive form (guanosine diphosphate-bound) and the active (GTP-bound) form, with the help of guanine nucleotide exchange factors (GEF) and GTPase-activating proteins (GAP). Somatic mutations stimulate KRAS to be in the active GTP-bound form, 33 but the impairment of GTP hydrolysis occurs via different mechanisms. 11,34 KRAS G12 and Q61 mutations mainly affect GAP-driven GTP hydrolysis, whereas mutations in G13 and K117 influence both GEF and GAP. 35,36 By contrast, mutations in KRAS A146 cause an increase in GEF-mediated nucleotide exchange without affecting GAP activity, 37 suggesting that tumors with a KRAS A146 mutation may be prone to respond to GEF inhibitors. Inhibition of the GEF Son Of Sevenless protein 1 (SOS1) reduces KRAS activation, especially when combined with an MEK inhibitor. 38 Likewise, RAS activation via GEFs was reduced by inhibition of the protein tyrosine phosphatase SHP2, 39 which was more effective in cells harboring KRAS G12C compared with cells harboring KRAS G12D. 40 Recently, AMG 510 (sotorasib), an antitumor agent targeting KRAS G12C mutant advanced solid tumors, has shown to improve the efficacy of (targeted) treatments in vivo. 41 AMG 510 is currently under investigation in a clinical trial (NCT03600883), including patients with CRC and non--small-cell lung cancer. 42 Another potential treatment strategy could be dual phosphatidylinositol 3-kinase (PI3K)/ mammalian target of rapamycin (mTOR) inhibition. Overexpression of the PI3K/Akt/mTOR signaling pathway is common in (m)CRC, resulting in enhanced tumor growth. Dual PI3K/mTOR inhibitors have shown to reduce cell proliferation of PIK3CA mutant tumors in mice 43 and phase I clinical studies. 44 However, this effect was not seen in cell lines where KRAS and PIK3CA mutations co-occurred. 45 When combining the dual PI3K/mTOR inhibitor with an MEK inhibitor, significant tumor reduction was seen in KRAS mutant tumors. 46,47  In conclusion, patients with mCRC with a KRAS A146 mutation represent a distinct molecular subtype of patients with poor survival who might benefit from more intensive treatments. Therefore, KRAS A146 mutation testing should be adopted in routine diagnostic testing.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs. org/po/author-center. Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).