Using Targeted Deep Sequencing to Detect ctDNA in Pancreatic Cancer

Study shows that ctDNA levels measured by targeted deep sequencing sensitively indicate the presence of cancer.

Using targeted deep DNA sequencing, researchers in South Korea were able to improve the detection of pancreatic ductal adenocarcinoma (PDAC) circulating tumor DNA (ctDNA). The technique might also allow for monitoring tumor responses to treatment.

“We demonstrated that ctDNA levels measured by targeted deep sequencing sensitively indicate the presence of cancer and correlate well with clinical responses to therapy and disease progression in PDAC patients,” wrote the authors of the feasibility study published in Scientific Reports.

The authors improved on the sequencing technology used in previous studies, and their findings were consistent with previous work using different technologies.

However, the study was based on samples from only 17 patients, cautioned Anton Wellstein, MD, PhD, professor of oncology and pharmacology at Georgetown University Medical School and associate director for basic and translational science at the Lombardi Comprehensive Cancer Center, in Washington, DC.

“This is clearly the way to go forward, to test the makeup of a cancer and follow changes in the makeup of the cancer,” Wellstein told Cancer Network. “It shows that patients can be followed and there are hints at getting a different profile during treatment. But, it’s a very small study.”

“You can get a footprint of the molecular changes in a cancer, almost in real time,” he noted. “Not the same day, but quickly.”

PDAC is one of the deadliest forms of cancer around the world, partly because there is no noninvasive biomarker for early sensitive and specific detection, the authors reported.

Several techniques of detecting ctDNA have been explored by different research teams in recent years, most focusing on the detection of KRAS, EGFR, and PIK3CA mutations. More than 90% of PDAC tumors harbor KRAS mutations, and these mutations appear to occur in most cells in PDAC tumors, providing a rationale for targeted digital polymerase chain reactions (PCR) for ctDNA. However, PCR has “often fallen short of high expectations, as the ctDNA detection rate has averaged as low as 50%,” the authors wrote.

The researchers used targeted deep sequencing to detect mutations across a larger swath of the tumor genome. Previous work with other cancer types has demonstrated that its use with ctDNA can yield insights into tumor burden, intratumoral heterogeneity, and the evolution and clonal expansion of treatment-resistance mutations.

Researchers sequenced 83 target genes, comparing mutation detection using tumor biopsy vs mutations detected from circulating cell-free tumor DNA (cfDNA), from 17 patients diagnosed with PDAC. KRAS mutations were identified in 10 of the 17 pretreatment plasma samples, and cfDNA-detected mutations matched primary tumor biopsy–detected mutations in 15 of 17 samples, leading to a detection sensitivity of 88%.

Comparing plasma ctDNA mutation detection from blood samples collected longitudinally over time, the authors found that somatic gene mutations were least frequent when chemotherapy achieved complete or partial tumor responses.

“I would not put too much stock into the precision of the sensitivity figure,” cautioned Wellstein. “That doesn’t diminish the study. It’s just a question of numbers. With small numbers and a very heterogeneous population, you’d have a pretty big error rate on that [88% sensitivity] figure.”

The findings, if validated in a larger cohort study, suggest that targeted deep sequencing of “broad regions” of ctDNA genomes will have diagnostic advantages over traditional biopsy in pancreatic cancer, according to Xin Yi, PhD, and Xuefeng Xia, MD, of the Geneplus-Beijing Institute in China.

The biggest challenge for the field has less to do with the tools for tracking pancreatic tumors’ evolution with circulating DNA than with limited targeted treatment options, Wellstein said.

“In pancreatic cancer, we can follow the changes, but it doesn’t give you an actionable read-out yet, such as, to see if a mutation emerges, you would use a different drug-it doesn’t yet impact treatment,” Wellstein said. “But over the next 5 or 10 years, from these measurements, we can start making decisions about changing treatment. This is the first step in that direction.”