Evolution of RAS Mutational Status in Liquid Biopsies During First-Line Chemotherapy for Metastatic Colorectal Cancer

Abstract

Treatment options for patients with metastatic colorectal cancer (mCRC) are limited. This particularly affects the largest group of patients with RAS mutations, who are considered ineligible for therapy with antiEGFR antibodies. In this liquid biopsy-based study, we performed the first in-depth analysis of the RAS mutational status in initially RAS-mutated patients during first-line therapy. RAS status of twelve patients with initially RAS-mutated mCRC was monitored longitudinally in 69 liquid biopsy samples. We focused on patients with stable disease (SD) or partial remission (PR) during first-line therapy (11 patients). Detection of fragmented RAS-mutated circulating cell-free tumor DNA (ctDNA) in plasma was performed by digital-droplet PCR (ddPCR) and BEAMing. Patients' total tumor masses were determined by measuring the tumor volumes using CT scan data. All patients with PR or SD at first follow-up showed a significant decrease of RAS mutational load. In ten patients (91%), the ctDNA-based RAS mutational status converted to wild-type in ddPCR and BEAMing. Remarkably, conversions were observed early after the first cycle of chemotherapy. Plasma concentration of ctDNA was controlled by determination of methylated WIF1-promotor ctDNA burden as a second tumor marker for mCRC. Persistent presence of methylated WIF1-promotor fragments confirmed the ongoing release of ctDNA during treatment. In patients with initially RAS-mutated mCRC, RAS mutations rapidly disappeared during first-line therapy in liquid biopsy, independent of type and intensity of chemotherapy and irrespective of anti-VEGF treatments. Following our results demonstrating conversion of RAS-mutational status, potential effectiveness of anti-EGFR antibodies in selected patients becomes an attractive hypothesis for future studies.

Keywords: BEAMing; circulating tumor DNA; clonal selection; colorectal cancer; digital droplet PCR; liquid biopsy; methylation WIF1 promotor; neoRAS wild-type.

Copyright © 2020 Klein-Scory, Wahner, Maslova, Al-Sewaidi, Pohl, Mika, Ladigan, Schroers and Baraniskin.

Figures

Figure 1

Dynamics of KRAS mutant clones measured in plasma samples. (A)Patient 3: RAS mutations disappeared already after 3 cycles FOLFOXIRI and remained not detectable for 9 months. PR after 6 cycles FOLFOXIRI made the patient eligible for tumor resection (arrow). After the following chemotherapy break PD occurred simultaneously with renewed rise of RAS mutation load. Already after 1 cycle FOLFIRI and bevacizumab the RAS mutations disappeared again and rose only after the next chemotherapy break due to TACE treatment. This case conclusively demonstrated that even multiple conversions of RAS status are possible with appropriate chemotherapy. (B)Patient 4: Already after 2 cycles FOLFIRI + bevacizumab the RAS mutations disappeared and remained not detectable for 6 months. Due to PR after 8 further cycles, the treatment was deescalated to 5-FU and bevacizumab. During this period, RAS mutation load increased again. Next, PD was diagnosed and the subsequent treatment change to FOLFIRI and aflibercept failed to achieve tumor response. However, RAS mutation load did not increase. This case points out that a renewed rise of RAS mutation load may be an early indicator for a lack of response. Furthermore, it can decrease during PD indicating that PD is caused by a RAS wild-type clone. (C)Patient 12: Neither FOLFOX nor FOLFIRI or addition of anti-VEGF antibody therapy led to a decrease of RAS mutation load or tumor response.

Figure 2

The dramatic decrease of RAS mutational load was detectable after 2 cycles (A,B,D) and the disappearance of RAS mutations occurs earliest after 2 cycles of therapy (B,D). In other cases (A,C,E–H) the disappearance is obvious after 4–12 cycles of therapy. The tumor loads, measured by CT scans, were strongly reduced by 75–90% in two cases (F–H), and more moderately reduced by 20–60% in other cases (A–E). In case (I), despite the dramatic decrease of RAS mutational load the Ras mutations remained detectable. (A) Patient 7, (B) patient 8, (C) patient 2, (D) patient 6, (E) patient 10, (F) patient 9, (G) patient 1, (H) patient 5, and (I) patient 11.

Figure 3

Change of WIF1 promotor methylation proportion vs. change of RAS mutational allele frequency. Samples of patients at the diagnosis time points vs. samples of the time point of disappearance of RAS mutations were measured by methylation specific ddPCR (see Supplementary File). WIF1 promotor methylation proportion remained detectable in samples with massive RAS MAF% reduction. (A–F) (solid line with circle, change of RAS mutant allele frequency; dashed line with cross, change of WIF1 promotor methylation proportion).