Detection of circulating tumor DNA (ctDNA) after surgery for colorectal cancer (CRC) identified patients with a high risk of relapse, which could be modified by adjuvant chemotherapy in some cases, data from a prospective study showed.
Among 218 patients with postoperative ctDNA results, 20 had detectable ctDNA, and 15 (75%) subsequently had disease recurrence. The five patients who did not relapse all received adjuvant chemotherapy. By comparison, 198 patients who tested negative for ctDNA had a relapse rate of 13.6%.
In a subgroup of 155 patients with postoperative measurement of ctDNA and carcinoembryonic antigen (CEA), positive ctDNA results had a significant association with relapse-free survival (RFS) whereas CEA did not, reported Tenna V. Henriksen, a PhD candidate at Aarhus University in Denmark, at the Gastrointestinal Cancers Symposium virtual meeting.
“We saw that patients with ctDNA detected immediately after surgery had a very high risk of recurrence,” Henriksen said. “We also saw that longitudinal monitoring increased the predictive power of ctDNA. Molecular recurrence by ctDNA was detected a median of 8 months before radiological detection of recurrence. Using longitudinal testing with ctDNA outperforms CEA in recurrence-free survival prediction.”
Randomized trials of ctDNA in different clinical settings will be required to move ctDNA into clinical practice, and several studies have already begun, she added.
Despite improvements in curative-intent treatment for stages I-III CRC, 20%-30% of patients relapse. Better detection of minimal residual disease (MRD) could improve postoperative risk assessment, and earlier detection of recurrence would allow more patients to receive curative-intent therapy after recurrence and lead to better survival, said Henriksen. Among potential strategies to detect MRD, ctDNA has produced promising results in several studies.
Investigators organized a clinical study to test the hypothesis that postoperative ctDNA measurement could identify patients with MRD and stratify patients into high- and low-risk groups. Henriksen and colleagues also wanted to assess post-treatment relapse risk in ctDNA-positive patients and determine the lead time from ctDNA detection to radiographic recurrence.
Data analysis included 260 patients with stages I-III CRC, 48 of whom relapsed after curative-intent treatment. The cohort consisted of four patients with stage I disease, 90 with stage II, and 166 with stage III. Henriksen reported that 165 patients received adjuvant therapy, and relapse-free patients had a median follow-up of 29.9 months.
Assessment of ctDNA was performed in 218 patients with the Natera (Signatera) assay, which identified 20 patients with detectable ctDNA (MRD positive) and 198 with no detectable ctDNA. A positive ctDNA test was associated with a recurrence hazard ratio of 11.0 (95% CI 5.9-21, P<0.0001). Among patients who received adjuvant chemotherapy, a positive ctDNA test at the end of treatment was associated with a recurrence rate of 83.3% as compared with 12.5% for those who had negative ctDNA tests (HR 12, 95% CI 4.9-27, P<0.0001).
Longitudinal assessment of ctDNA showed that the risk of recurrence increased over time in ctDNA-positive patients and decreased in ctDNA-negative patients (89.3% vs 3.4%, HR 51, 95% CI 20-125, P<0.0001).
In a subgroup of 29 patients with clinical recurrence detected by CT scan, ctDNA detection occurred a median of 8.1 months earlier, said Henriksen.
Investigators compared the performance of ctDNA with the tumor-associated protein CEA. Measurement of postoperative CEA (n=175) and after adjuvant chemotherapy (n=99) did not have a significant association with the risk of recurrence. Longitudinal assessment of CEA (n=197) did predict an increased risk of recurrence (HR 4.9, 95% CI 3.2-15, P<0.0001) but not as well as longitudinal ctDNA (n=197, HR 95.7, 95% CI 28-322, P<0.0001).
Two clinical trials of ctDNA-guided clinical management have already begun, said Henriksen. The IMPROVE-IT study involves patients with stage I or low-risk stage II CRC, a group who usually does not receive adjuvant chemotherapy. Studies have suggested that 10%-15% of such patients are undertreated. In the trial, ctDNA-positive patients will receive adjuvant chemotherapy, and ctDNA-negative patients will receive no adjuvant therapy but will be followed with serial ctDNA testing.
IMPROVE-IT2 compares ctDNA and radiologic surveillance for recurrence. Currently, only 20% of patients with radiologically detected relapse are eligible for curative-intent therapy. In the clinical trial, patients with ctDNA-positive tests will undergo intensified radiologic surveillance and intensified treatment, whereas those with negative ctDNA tests will be followed with serial ctDNA testing only.
The study reported by Henriksen adds to a growing body of evidence for the “convincing prognostic ability of circulating tumor DNA following surgical resection,” said invited discussant Michael Overman, MD, of the MD Anderson Cancer Center in Houston. However, the assay used in the study is one of multiple approaches to ctDNA detection under investigation.
“The key for us is to understand how to use circulating tumor DNA to guide therapy,” he said. “It must be recognized that the availability of a test is not the same as the actionability. To answer this question, I encourage all of us to support the enrollment of ongoing randomized clinical trials using circulating tumor DNA to help optimize our use of adjuvant therapy for colorectal cancer patients.”
The study was supported by the Danish Council for Independent Research, the Novo Nordisk Foundation, the Danish Cancer Society, and Natera. Some co-authors are Natera employees.
Henriksen disclosed no relevant relationships with industry. Co-authors disclosed relevant relationships with Natera, Mission Bio, Bayer, Pierre Fabre, Merck Serono, Roche, Servier, AbbVie, Amcure, Array BioPharma, Astellas, AstraZeneca, BeiGene, Bristol-Myers Squibb, FibroGen, Genentech, Johnson & Johnson, Lilly, MedImmune, Merck, Novartis, Sierra Oncology, Takeda, Tesaro, and Theradex.