SAN FRANCISCO – Everolimus improves progression-free survival in all subgroups of patients with progressive advanced neuroendocrine tumors, and does so regardless of prior use of somatostatin analogs, according to updated results of the randomized RADIANT-2 and RADIANT-3 trials.
Dr. James C. Yao
Both trials showed that the reduced risk of progression seen with everolimus in their study populations overall also was present to some extent in subgroups stratified by patient and disease characteristics, and by prior use of somatostatin analogs.
The updates were reported at a meeting on gastrointestinal cancers sponsored by the American Society of Clinical Oncology.
"There is a significant unmet medical need for patients with advanced NETs [neuroendocrine tumors]," asserted RADIANT-2 lead investigator Dr. James C. Yao of the department of gastrointestinal medical oncology at the University of Texas M.D. Anderson Cancer Center in Houston. In particular, "there are currently no drugs approved by the health authorities for the oncologic therapy of advanced carcinoid tumors."
RADIANT-2: Unselected NETs
The RADIANT-2 trial was conducted among 429 patients with progressive advanced NETs of diverse primary sites who had carcinoid syndrome. It compared everolimus (Afinitor) plus octreotide LAR (Sandostatin LAR, or long-acting release) with placebo plus octreotide LAR.
Everolimus, an inhibitor of the mammalian target of rapamycin (mTOR), is approved by the Food and Drug Administration for the treatment of certain subependymal giant cell astrocytomas and advanced renal cell carcinoma.
Octreotide LAR is a somatostatin analog approved for symptom control in patients with metastatic carcinoid tumors, a subset of NETs that cause secretory symptoms.
The trial found a 23% reduction in the risk of progression, as assessed by independent central review, with everolimus vs. placebo (hazard ratio [HR] 0.77, P = .026), but it fell just short of its predefined statistical boundary for significance (P = .0246), according to Dr. Yao.
Updated analyses showed that the benefit seen in the entire study population was also evident in subgroups stratified by patient and disease factors, and by prior treatment, although most of the hazard ratios overlapped unity.
About 80% of patients in both treatment groups had previously used long-acting somatostatin analogs. In exploratory analyses, everolimus was associated a reduced risk of progression in previous users (HR 0.81, P = .077) and in previous nonusers (HR 0.63, P = .054). "However, the absolute and relative benefits were perhaps greater among patients who have not had prior somatostatin [analog] exposure," he observed.
Dr. Yao also reported new analyses of the trial’s main results, showing that, when progression was assessed by local investigators instead of central reviewers, everolimus was associated with a 22% reduction in risk and the study met its statistical boundary (P = .018). The difference was due to 51 more events tallied by the local investigators vs. the central reviewers.
Additionally, the progression-free survival estimates for both treatment groups were longer with central review than with local investigator assessment. "These findings are actually hallmarks of informative censoring," he said. "The main source of informative censoring lies in the fact that patient management is based on investigator review, while the end point assessment is based on central review."
Issues arise when the local investigator calls progression before the central reviewer does. "In these cases, because of the crossover and change in therapy, the cases are censored. In fact, the central radiologist is effectively prevented from seeing the progression event," Dr. Yao explained. "This results in an inflation of PFS [progression-free survival] and reduced power in the study."
When the trial’s main analysis was repeated with correction for informative censoring, reduced power, and imbalances in randomization, the median progression-free survival was 13.8 months with everolimus and 8.3 months with placebo. That difference corresponds to a 40% reduction in risk and far exceeds the predefined boundary for significance (HR 0.60, P = .0014).
Commenting on the discrepancy between local and central review, discussant Dr. Mary F. Mulcahy said, "I think what this tells us ... is that progression-free survival may not be the right end point to be looking at when evaluating these diseases with secretory symptoms," because the symptoms may prompt investigators to switch treatment before radiographic progression occurs.
"So I think more so than these fancy calculations to try and rectify this, we need to maybe consider what our end point is for these very clinical syndromes that we see," she commented.
She called attention to the fact that the rate of discontinuation for progressive disease or adverse events was similarly high in the everolimus and placebo groups (70% and 75%), and median duration of treatment exposure was the same.