Serial comprehensive genomic profiling by next-generation sequencing for patients with metastatic colorectal cancer

To The Editor,

With the increasing number of potentially actionable genomic alterations to be assessed, comprehensive genomic profiling (CGP) using next-generation sequencing (NGS) has become more common for patients with advanced-stage cancers. International practice guidelines recommend CGP over single-gene testing, particularly when tissue is difficult to obtain or when therapies are available for multiple genomic targets.^{1, 2}

In Japan, CGP for advanced cancer is covered by national health insurance only after completion of standard therapy and only once in a patient's lifetime. Recent data on the use of first-line CGP for advanced solid tumors in Japan has prompted interest in earlier use of such technology.³

If CGP were to move to earlier lines of testing in Japan, what might be the clinical value for testing patients again by CGP after disease progression? We considered metastatic colorectal cancer (mCRC) as a model for two reasons: (1) It is the most registered tumor type with the Japanese Center for Cancer Genomics and Therapeutics (C-CAT),⁴ and (2) international clinical practice guidelines recommend testing for the presence of multiple genomic biomarkers prior to treatment for patients with mCRC. These include mutations in KRAS and NRAS, BRAF V600E, ERBB2 amplification, NTRK and RET fusions, and microsatellite instability-high (MSI-high) or mismatch repair deficiency.^{1, 2, 5}

We published findings from a real-world claims database of 1064 mCRC patients tested by a plasma-based NGS CGP assay after completion of first- or second-line therapy in the United States.⁶ In this setting, plasma-based NGS testing of circulating tumor DNA (ctDNA) performs similarly to standard tumor tissue testing⁷ and tissue-based NGS⁸ with a high negative predictive value for the absence of RAS mutations, particularly in plasma samples having a non-RAS mutation with variant allelic frequency (VAF) ≥1%.⁹ Therefore, we considered that this cohort could be useful for exploring the role of serial CGP testing for patients with mCRC.

From the same dataset used for our previous manuscript we identified 82 patients who were tested by CGP prior to two different lines of systemic anticancer therapy within the first three lines of treatment. Among these, 34 were tested prior to first- and second-line therapy, 12 prior to first- and third-line therapy, and 36 prior to second- and third-line therapy.

For this post hoc analysis, we employed the definition of actionability and matched therapy as previously described⁶ but with a more conservative approach for determination of actionable RAS wild-type status; namely, absence of KRAS or NRAS mutations in the presence of a non-RAS mutation with VAF ≥1% and restricted to left-sided primary tumors. Actionable profiles were identified for 29 patients (35.4%) with the initial test and for 30 patients (36.6%) with the subsequent test. There were 23 patients with actionable profiles from both tests; 6 had actionable profiles only with the first test; 7 had actionable profiles only with the second test; and 46 did not have an actionable profile with either test.

Results from initial and subsequent testing are summarized in the Table 1. The most common actionable profile was RAS/BRAF wild-type without ERBB2 amplification, 19 (23.1%) with the first test and 20 (24.4%) with the second test; 15.9% of patients had RAS/BRAF wild-type profiles with both tests. Among 47 patients with tumors harboring RAS or BRAF mutations initially, the second test showed reversion to wild-type status for 5 (10.6%). Other changes identified with the second test included acquisition of RAS mutations (n = 5) or BRAF V600E (n = 2) for those with RAS/BRAF wild-type status initially and loss of ERBB2 amplification (n = 2). Concurrent MSI-high was detected in a patient with acquired BRAF V600E.

Therapy was matched to the CGP result at least once for 68 of the 82 patients (82.9%), including 20 (24.4%) who received targeted treatment directed against an actionable molecular profile. Matched therapy was administered to 52 patients after initial testing (63.4%; 17.1% targeted) and to 49 after subsequent profiling (59.8%; 12.2% targeted); 33 patients (40.2%) received matched therapy after both lines of testing, including 4 (4.9%) with targeted treatment in both lines.

For patients with ctDNA detected at both timepoints (n = 69), the second test identified a change in actionability for 20.3% (n = 14) and led to a change in matched targeted therapy for 8.7% (n = 6).

These findings suggest a role for CGP prior to more than one line of mCRC therapy. Actionable alterations were detected for approximately one-third of the patients with both initial and repeat tests, and about 60% of these were treated with therapy consistent with the CGP findings at each line of treatment. Given the nature of real-world clinical practice and the limited sample size of the cohort, the application of matched targeted therapy had no impact on overall survival, consistent with our earlier findings.⁶

Despite the relatively small sample size in this retrospective exploratory analysis, these data support the concept that CGP testing may inform treatment decisions in more than one line of therapy for patients with mCRC.

Whether matched targeted therapy in multiple lines of therapy might improve overall survival remains an open question, although clinical advantages have been reported when matched targeted therapy has been applied in the first-¹⁰ and second-line¹¹ settings. While these results are not definitive and do not address comparisons of CGP over single-gene testing, we hope that they encourage investigators in Japan and elsewhere to conduct well-powered studies to address the potential role of serial CGP testing for patients with mCRC or other advanced cancers expected to have a high proportion of clinically informative genomic alterations.

This research was funded by Guardant Health.

The authors contributed equally to the conceptualization, data analysis and interpretation, and writing of this letter.

Dr Olsen is an employee of Guardant Health. Dr Nakamura reports lecture fees from Chugai, Guardant Health, and Merck Biopharma and research funding from Chugai, Daiichi Sankyo, Genomedia, Guardant Health, Roche Diagnostics, Seagen, and Taiho.

Ethics approval and participant consent were not necessary because this study used a deidentified database in accordance with US patient confidentiality requirements set forth in Sections 164.514 (a)–(b)1ii of the Health Insurance Portability and Accountability Act (HIPAA) regarding the determination and documentation of statistically deidentified data.

Approval of the research protocol by an Institutional Review Board: N/A.

Informed consent: N/A.

Registry and the Registration No. of the study/trial: N/A.

Animal Studies: N/A.