乌鸦传媒

Tumor-informed circulating tumor DNA (ctDNA) testing for residual disease was able to distinguish patients with colorectal cancer (CRC) at risk for recurrence after surgery, as well as those who could safely avoid adjuvant chemotherapy, according to a large prospective study from Japan.

Over a median follow-up of 16.7 months, 61% of patients who were ctDNA-positive 4 weeks after surgery had recurrences, compared with 10% of those who were ctDNA-negative (HR 10.0, 95% CI 7.7-14.0, P<0.0001), reported Eiji Oki, MD, PhD, of Kyushu University in Fukuoka, and colleagues.

The study included over 1,000 patients, 18% of whom were ctDNA-positive at 4 weeks after surgery, the researchers detailed in .

At 18 months, disease-free survival (DFS) rates in the two groups were 38.4% and 90.5%, respectively -- a trend that was observed across all pathological stages.

In multivariate analysis for DFS in patients with pathological stage II-III disease, ctDNA positivity 4 weeks after surgery was the most significant prognostic factor associated with increased risk for recurrence (HR 10.82, 95% CI 7.07-16.60, P<0.001).

As for the association of adjuvant chemotherapy with postsurgical ctDNA status, the investigators found that patients with stage II-III CRC with postsurgical ctDNA positivity significantly benefited from adjuvant chemotherapy (HR 6.59, 95% CI 3.53-12.30, P<0.0001), and this trend was observed across all pathological stages:

• High-risk stage II: HR 5.84 (95% CI 1.36-25.10)
• Stage III: HR 7.02 (95% CI 3.46-14.20)
• Stage IV: HR 4.00 (95% CI 1.85-8.80)

Furthermore, a multivariate analysis of ctDNA-positive patients with stages II-IV disease showed that a lack of adjuvant chemotherapy was a significantly negative prognostic factor (HR 5.03, 95% CI 3.17-8.90, P=0.001), the team reported. On the other hand, there was no statistically significant benefit associated with adjuvant chemotherapy in ctDNA-negative patients (HR 1.71, 95% CI 0.80-3.70, P=0.167).

Results of the analysis showed that postsurgical ctDNA status is a more significant prognostic biomarker than the currently used high-risk clinicopathological features and can potentially be predictive of adjuvant chemotherapy benefit, said Oki and co-authors.

The findings are "potentially practice changing and clearly in the direction we want to go," said Aasma Shaukat, MD, MPH, director of Outcomes Research in the Division of Gastroenterology and Hepatology at NYU Grossman School of Medicine in New York City, who was not involved with the study.

"We've been trying to move toward precision medicine for decades, and in oncology the way it applies is understanding which patients are most likely to benefit from chemotherapy, with the goal of increasing disease-free survival and overall survival," she told 乌鸦传媒. "We have some markers, but they are not very targeted, and as a result a lot of people who won't benefit end up suffering the toxicity and cost of chemotherapy, while individuals who are likely to benefit sometimes get skipped."

The commercial assay used in the study (Signatera) is "all we would need to offer individuals 4 weeks after surgery," Shaukat continued. "We could very confidently make decisions about who would benefit from additional chemotherapy, and perhaps keep them disease free for a longer period of time."

For the research, the team used whole exome sequencing, and the results were analyzed to design a tumor-informed, personalized ctDNA assay. A total of 8,374 genes were selected for 1,039 patients, with TP53 (25.6%) and APC (17.5%) the most frequently selected. More than half of the genes were unique to each patient -- suggesting, the investigators explained, that there is "large variability in the mutational landscape of CRC outside of known hotspot regions."

"This highlights the importance of personalized ctDNA analysis based on patient-specific somatic tumor mutations," Oki and co-authors added.

"The question now," Shaukat said, "is how this goes from this research to actually developing into an assay. And I think we'll probably see several different iterations. It's very difficult to do an over-8,000 gene exome every time, so what may happen is that they may refine it and narrow it down to the genes of interest so that it still gives us that detailed information -- it is able to discern and distinguish -- yet is more streamlined."

Comment