In the era of precision oncology, targeted therapy and immunotherapy have finally brought a new hope to patients harboring KRAS G12C-mutant cancers. G12C is the most common subtype of KRAS mutations in NSCLC, and is a frequently detected driver oncogene in clinical practice.13 Indeed, G12C mutations were present in 12.5% of patients in the current cohort, which is among the highest in the literature. In the Genomics Evidence Neoplasia Information Exchange (GENIE) project, KRAS G12C was found in 1,443 of 10,444 NSCLC patients (13.8%).14 However, lower rates have been detected in other databases, varying from 6.9% in the Catalog of Somatic Mutations in Cancer (COSMIC) to 10.5% in Spira et al.7,9 The average mutation frequency of KRAS G12C is higher than most common druggable oncogenes in the clinic, highlighting the relevance of routine testing.14
As opposed to other common driver mutations, such as EGFR, ALK, ROS1 and RET, which are predominant in never smokers, KRAS G12C more often affects current or past smokers.8,9 In the current study, 88.0% of G12C-positive cases were in smokers, leaving only 12.0% in patients without a past smoking history. Other authors have found significantly lower rates of never smokers among KRAS G12C-mutant cases. For instance, Arbour et al. described only 5 never smokers (1.4%) among 352 G12C-mutants in a US database from 2014 to 2018.15 Similarly, the rate of never smokers among G12C cases was 3.2% in Spira et al. and 2.3% in Dogan et al.8,9 Interestingly, 30% of patients in our database were never smokers, which is higher than the usual norm in lung cancer – approximately 85% of patients are smokers.9 This finding suggests an anticipated bias toward never smokers in the selection of patients that were submitted to lung cancer genotyping. During the study period, targeted therapy was only available for EGFR-mutants, reinforcing the greater propension to test never smokers. It is relevant to mention that genomic testing is recommended for all advanced NSCLC cases that present with an adenocarcinoma component, independently of smoking history.9,16 In the case of KRAS, the current data reinforces that a significant amount of never smokers may still develop KRAS G12C-driven NSCLC, further supporting that smoking status should not be a criterion for molecular testing.
In the current study, KRAS G12C mutations were more frequent in females (60.0% of cases), which is in line with most trends in the literature.8,9,15 KRAS G12C has been more often described in women, representing from 61.1% of cases in Spira et al. to 69.3% in Dogan et al.8,9 In another database from a genomics laboratory in Brazil, however, KRAS G12C occurred in 173 women among 346 cases (50.0%), in a similar manner to other KRAS mutation subtypes.17 Altogether, the current data comes to stress that sex per se is a poor predictor of mutational profile in NSCLC, and likely exerts limited effect on lung carcinogenesis.
The pattern of distant metastases has been rarely described in the literature of KRAS G12C NSCLC. In the present study, a higher frequency of CNS involvement was described in KRAS G12C-mutant cases. Nine out of 25 cases had metastasis to the CNS (36.0%). Spira et al. also assessed brain metastasis, and found a slightly higher rate in G12C mutants (23.4%).9 In other studies, brain metastases were found in 37.2% and 40.3% of cases, in Lamberti et al. and Sabari et al., respectively.18,19 As opposed to the current study, the incidence of brain metastasis was similar among G12C and non-G12C mutants in both studies.18,19 These findings have relevant clinical implications since targeted agents are not universally active in the CNS.20 Although specific KRAS G12C inhibitors have shown CNS penetration and clinical activity, data are limited to a handful of cases or exploratory analysis.4 Moreover, untreated CNS metastases were exclusion criteria in clinical trials to date.
The prognostic impact of KRAS G12C has been explored in the literature. For instance, Spira et al. did not find a significant difference in OS between KRAS G12C mutants and KRAS/EGFR/ALK wild types – named triple wild types.9 Similarly, OS was similar in G12C and non-G12C KRAS subtypes in Arbour et al.15 In the current study, OS was not significantly different according to KRAS status, in line with the aforementioned studies. However, it is important to mention that immunotherapy was not available for these patients in the current study, therefore our data should not be extrapolated to scenarios where ICB is a standard. In that setting, G12C mutants have shown similar outcomes on pembrolizumab, in comparison to non-G12C or wild-type KRAS.15
The current study provides real-world data on patients treated in a routine standard. Among the potential limitations are lack of novel therapies, including specific KRAS G12C inhibitors and immunotherapy. As discussed above, G12C alone have failed to predict responses to immunotherapy, however concomitant mutations in genes such as STK11 and KEAP1 are associated with worse prognosis and poor responses to ICB.9,21 A customized NGS panel was used in all patient cases herein, which provided data on 4 genes for the whole cohort. Unfortunately, a more comprehensive genomic panel was not available, which could have added information on other driver oncogenes, as well as on co-mutant genes that may influence prognosis and predict benefit from novel therapies. Nonetheless, patients in the current cohort were mostly treated exclusively with cytotoxic chemotherapy. Since targeted therapy was only available for EGFR, and immunotherapy was only available in clinical trials, comprehensive NGS was not clinically justified. The retrospective nature of this study could represent an additional source of bias. On the other hand, this was conducted in a reference cancer center in the country, and provides a fully-annotated clinical dataset. Oncology routine was standardized, and no major changes occurred in patient care during the study period. Data presented could be reproducible of other scenarios worldwide where limited access to novel agents is a norm.