Circulating tumor DNA (ctDNA) can help differentiate what is known as pseudoprogression from true progression of disease in patients with melanoma who are treated with programmed death 1 (PD-1) inhibitors, according to a new study. It also showed that ctDNA could be used as a powerful biomarker to predict long-term outcomes.
“Pseudoprogression, a response that occurs after the initial development of new lesions or an increase in the size of target lesions, occurs in up to 10% of patients treated with PD-1 antibodies,” wrote study authors led by Jenny H. Lee, MBBS, of Macquarie University in Sydney, Australia. “Confirmation of pseudoprogression requires subsequent imaging, imposing an ongoing challenge.”
Previous research has shown that ctDNA levels at baseline could be used to predict response to PD-1 therapy, so investigators examined whether ctDNA could be used to aid in differentiating pseudoprogression from true progression. The study included 125 patients with stage IV melanoma and established BRAF or NRAS mutations receiving PD-1 antibodies alone or in combination with ipilimumab. The results were published in JAMA Oncology.
Pseudoprogression was defined as a more than 25% increase in tumor burden at imaging assessment 1 at week 12, and not confirmed as progressive disease on the subsequent assessment. Patients were considered to have a favorable ctDNA profile if they had undetectable ctDNA at baseline, or if they had detectable ctDNA at baseline that became undetectable or decreased by at least 10-fold. Those with detectable ctDNA at baseline that remained stable or increased during treatment were considered unfavorable.
Of the 125 patients, 29 had progressive disease at the first restaging. Of those 29, 9 patients (31%) had pseudoprogression, and 20 (69%) had true progression. Eleven of the 29 patients had a favorable ctDNA profile; 2 had undetectable levels at baseline, 5 had detectable levels at baseline that became undetectable within 6 weeks, and 4 had detectable levels that dropped by 10-fold or more by week 12.
All of the nine patients with pseudoprogression had a favorable ctDNA profile. Three of the patients subsequently progressed; two of them, whose disease progressed at 30 and 43 weeks, respectively, had ctDNA that remained detectable despite the 10-fold decrease within 12 weeks. Of the 20 patients with true progression, 90% had an unfavorable ctDNA profile.
The sensitivity of ctDNA for predicting pseudoprogression from true progression was 90%, and the specificity was 100%. This biomarker had a better sensitivity and specificity, as well as positive and negative predictive values, than lactate dehydrogenase.
At 1 year, the survival rate for patients with progressive disease and favorable ctDNA was 82%, compared with 39% for unfavorable ctDNA, for a hazard ratio of 4.8 (95% CI, 1.6–14.3; P = .02).
“Early recognition of [pseudoprogression] has important implications for patient management,” the authors wrote. “Recognition of this response pattern prevents premature termination of potentially effective treatment and may influence the selection and administration of optimal sequential therapy in a time-sensitive manner. Within 12 weeks of therapy, ctDNA profiles accurately identify pseudoprogression from true progression.”