The Use of Liquid Biopsies to Discover Recurrent Colon Cancer

A liquid biopsy taken just 1 month following surgery was able to detect “ultra-low” circulating tumor DNA (ctDNA) signals of recurrent colorectal cancer, even flagging distant metastases like lung cancer, according to interim findings presented during the annual meeting of the American Association for Cancer Research (AACR) Annual Meeting 2025. The novel assay was also able to detect cancer recurrence among these patients at a median lead time of 198 days before standard-of-care imaging, providing earlier prognostic value for clinical management and care. “After surgery, ctDNA-based liquid biopsies may help identify patients who would benefit most from additional treatment,” said Jonathan Loree, MD, MS, a medical oncologist at BC Cancer in Vancouver, Canada, and senior investigator of the study. “Alternatively, this may help patients with good prognosis avoid toxicities from unnecessary chemotherapy. By monitoring patients for recurrences, liquid biopsies can continue to support clinical care and allow more patients to undergo second curative intent surgeries to remove early recurrences.” Early detection of ctDNA after treatment is a strong indicator of recurrence in colorectal cancer, but often goes undetected because traces of these blood-based DNA particles can be very low. And many liquid biopsies lack the sensitivity required to detect these early-stage signs of cancers. “It has also been shown that treatment intervention can limit detection of ctDNA,” said Emma Titmuss, MS, Bioinformatician in Medical Oncology at BC Cancer in Vancouver, Canada, and the study's lead author. “Interventions like surgery and chemotherapy can increase shedding of DNA from normal cells—cell-free DNA (cfDNA)—which dilutes the signal from ctDNA making it harder to detect. This raises the question: is there an optimal timepoint post-intervention, like surgery, to detect ctDNA?” Study Details With this question in mind, Titmuss and colleagues turned to findings from VICTORI, an observational and ongoing study investigating the prognostic value and best time to administer an “ultrasensitive” liquid biopsy—used to detect ctDNA at levels of less than 100 parts per million (ppm)—for patients with colorectal cancer. A liquid biopsy is a test that detects signs of cancerous tumors by identifying microscopic bits of DNA released into the bloodstream by dying cancer cells. Information gleaned from these minute DNA particles—referred to as minimal residual disease (MRD)—can be used following surgery as a prognostic indicator, with results pointing to optimal pathways for treating and/or managing their disease. As described during AACR, the VICTORI study has enrolled 71 patients with resectable disease, including 52 with Stage I-III colorectal cancer and 19 with Stage IV disease, with 530 blood samples available up to 2 years post-surgery. The investigators built a tumor tissue-derived panel with up to 1,800 high-quality genetic variants to create a personalized MRD signature for each patient. Liquid biopsies were analyzed with the NeXT Personal assay prior to surgery every 2 weeks for 8 weeks post-surgery and every 3 months for a potential of 3 years. Currently, all 33 patients enrolled with treatment-naive disease greater than Stage I had detectable disease prior to surgery, with 40 percent of patients having detection levels of less than 100 ppm—the ultrasensitive level of detection. Among 65 patients now evaluable for clinical outcomes, some 23 experienced disease recurrence. Of these, some 87 percent tested ctDNA-positive within the 8-week post-surgical period—the “landmark window” of time when patients are given adjuvant chemotherapy. Of the three detected after the landmark window, one (Stage IV) had neoadjuvant therapy before surgery, and was detected at the first follow-up timepoint month (Month 3). Another (Stage III) was detected at Month 6, and the third (Stage II) at Month 9. All patients with clinical recurrence were ctDNA positive before recurrence was detected via reflex imaging, by a median of 198 days earlier, up to 416 days for one patient. According to Titmuss, ctDNA was detected as low as 2 ppm, with a median ctDNA level at first detection within the ultrasensitive range. the median ctDNA level at first detection was 24.4 ppm, and the highest was 111,120 ppm. Higher DNA levels at first detection were linked to shorter times to clinical response and poorer prognosis. Results further suggest that the “ideal” timepoint for ctDNA testing—when blood-based signals from cfDNA are sufficiently diluted following surgery—is 4 weeks. “As I mentioned, this is a very preliminary result,” Titmuss cautioned. “We only have 71 patients enrolled at this point. We are planning to enroll up to 160 patients, and the current follow-up is still relatively short. We only have 16 months [of follow-up], but we are taking positive samples from patients up to 3 years and then following them for another couple of years after that, so we do expect to be able to get more information and make firmer conclusions when we have a larger cohort.” One limitation of the study is the lack of intervention after ctDNA detection, as this was an observational study. More randomized trials are needed to determine the most effective use of this technology in clinical management. However, these early findings are nevertheless impressive, said Jayesh Desai, MD, Associate Director of Clinical Research with the Peter MacCallum Cancer Centre in Victoria, Australia, and moderator for the AACR press briefing. “I know this is an observational study, but I was particularly impressed by that high rate of negative results correlating no recurrence,” he said. “That would give you confidence in that patient to think about not giving adjuvant chemotherapy. Obviously, those things need to be confirmed with prospective trials, but that potentially could change practice and avoid someone having 3-5 months of adjuvant chemotherapy. So, it's a huge step forward in this space.” Warren Froelich is a contributing writer.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.