NEW YORK (GenomeWeb) – Levels of circulating tumor DNA can predict ovarian cancer patients' response to treatment or time to disease progression, according to a new study from University of Cambridge researchers.
In a retrospective study published today in PLOS Medicine, researchers led by Cambridge's James Brenton used patient-specific digital PCR TP53 mutation assays to quantify the amount of circulating tumor DNA present in serial blood samples obtained from 40 high-grade serous ovarian cancer patients. The researchers found that the fraction of ctDNA levels in patients' blood reflected the effects of beginning chemotherapy sooner than CA-125, the current disease response marker, and correlated with time to disease progression.
"In this study, we showed the potential of ctDNA to identify, after one [cycle to] two cycles of treatment, ovarian cancer patients with an expected poor response to chemotherapy," Brenton and his colleagues wrote in their paper. "These findings have strong potential for clinical utility owing to the ease of assaying DNA in plasma and the low cost and speed of ctDNA testing."
Using TaqMan assays, the researchers identified 31 unique TP53 mutations in FFPE samples from 40 patients — more than 99 percent of ovarian cancer patients have TP53 mutations, they noted. The most common mutations were q.1245 8G>A (p.R175H) and g.13744G>A (p.R273H), and both were present in four patients.
Based on these mutations, Brenton and his colleagues generated personalized digital PCR assays to analyze patients' ctDNA levels. Of the 318 plasma samples they had, 261 were collected during treatment of relapsed disease and 57 were collected during first-line chemotherapy treatment.
The fraction of ctDNA the patients had in their samples correlated with disease volume, as gauged on a CT scan, the researchers reported, and that link was stronger among patients without ascites.
The researchers also compared levels of tumor DNA and CA-125 in blood samples taken from the patients before and as they underwent chemotherapy. In a multivariate analysis, they found that only their TP53 mutant allele fraction and platinum sensitivity prior to starting treatment were prognostic factors; CA-125 and total cell-free DNA levels were not.
As the patients' chemotherapy progressed, the researchers noted changes in their ctDNA and CA-125 levels, with the TP53 mutant allele fraction exhibiting a more rapid decrease and greater dynamic range. Indeed, they reported that the mean time to nadirwas shorter for the TP53 mutant allele fraction than it was for CA-125, possibly because of the longer half-life of CA-125. This suggested to the researchers that chemotherapy response could be determined more quickly if ctDNA were used as a biomarker.
A decrease of 60 percent or greater in TP53 mutant allele fraction after one chemotherapy cycle could further point out patients with a six-month time to progression, with 88 percent specificity, the researchers reported. The predictive value of the TP53 mutant allele fraction remained significant even after a second chemotherapy cycle and with an 80 percent threshold, they added, noting that the response classification after one and after two cycles were consistent.
Though the researchers cautioned that their study was a small retrospective that could be swayed by variability among the patients and their treatments, Brenton and his colleagues said that it underscored the potential of circulating tumor DNA as a biomarker. In future work, they noted that next-generation sequencing assays could replace the patient-specific assays they used.
"These results provide evidence that ctDNA has the potential to be a highly specific early molecular response marker in [high-grade serous ovarian cancer] and warrants further investigation in larger cohorts receiving uniform treatment," they wrote.