BARCELONA, Spain—A novel symptom-based objective response criterion was used in a phase III trial to rapidly evaluate the efficacy of a new cancer drug, the anti-Interleukin-1α agent MABp1, as treatment for advanced metastatic colorectal cancer (NCT02138422).
Lead investigator Tamas Hickish, MD, Consultant Medical Oncologist, Dorset Cancer Centre, Bournemouth University, Bournemouth, U.K., said that after only 8 weeks of treatment, “health status can be used to measure efficacy of anti-tumor therapy, and clinical responses based on health status can be a predictor of overall survival benefit.”
Hickish reported a 33 percent clinical response rate based on health status among 207 patients treated with MABp1, compared with 19 percent of 102 patients on placebo.
“This represents a 76 percent increased clinical response rate of treatment versus placebo,” Hickish said, in a presentation at the recent European Society of Medical Oncology World Congress on Gastrointestinal Cancer.
The overall survival rate among patients who showed a clinical response was 11.5 months versus 4.2 months for non-responders.
“That survival among these heavily pretreated patients is remarkable,” Hickish said.
But whether these data have any clinical consequences, or can be relied on at all, was questioned by the discussant for the paper, Dirk Arnold, MD, PhD, Director of Oncology Services, CUF Institute of Oncology, Lisbon, Portugal.
Arnold asked whether the report of a “76 percent increased clinical response rate” was reliable.
“My calculation is that 33 percent minus 19 percent equals 14 percent, the difference of the number of patients with symptom improvement,” Arnold said. The claim of a “76 percent increase in improvement” represents only the relative improvement rate, he explained.
And in a statement issued after the meeting, Arnold noted that the “responders” in the trial included patients in the placebo arm who had a clinical response as well as those in the treatment arm.
“Are the results reliable?” Arnold said at the conclusion of his discussion. “Not really.”
“And is there a benefit [to treatment with MABp1] which is worth the effort? The answer is not yet clear,” Arnold noted. “Will the results have clinical consequences? Not now.”
IL-1 Abundant in Tumors
Interleukin-1α (IL-1α, also known as hematopoietin-1) is a proinflammatory protein encoded by the IL-1A gene. IL-1 is abundant at tumor sites, where it may affect the process of carcinogenesis, tumor growth, and invasiveness, and also the patterns of tumor-host interactions (Cancer Metastasis Rev 2006;25:387-408).
Hickish said measures of improved health status among patients in the trial correlated with improvement in other self-reported and laboratory-based measures of health, including improved control of tumor-related white blood cell activity and reduced systemic inflammation.
The objective response criteria used here were developed in collaboration with European Medicines Agency's (EMA) Scientific Advice Working Group, he said.
The trial included 309 patients with refractory disease and a constellation of symptoms/functional impairment: wasting, fatigue, pain, anorexia, and ECOG Performance Status 1 or 2. They were randomly assigned to MABp1 (207 patients) or placebo (102 patients), both arms receiving best supportive care.
The overall survival rate among patients who responded to MABp1 (87 patients) was 11.5 months versus 4.2 months for non responders (222 patients).
But no statistical comparison of overall survival could be made between the treatment and placebo arms, Hickish said, because in an open-label extension of the study placebo patients were allowed to receive MABp1 after the primary endpoint was met.
The primary trial endpoint was maintenance or improvement in both lean body mass and two of three symptoms: pain, fatigue, or anorexia. Objective response criteria used dual-energy X-ray absorptiometry and the EORTC QLQ-C30 questionnaire to assess disease control, both prior to initiation of therapy and after eight weeks of therapy.
Secondary endpoints were safety, measures of functional performance, and quality of life, as well as pharmacodynamic measures known to correlate with survival such as control of paraneoplastic thrombocytosis and reduced systemic inflammation (IL-6).
Hickish said earlier observations of MABp1 therapy in advanced cancer patients suggested physical recovery that strongly correlated with significant improvement in survival in colorectal cancer (Lancet Oncology 2014;15:656-666).
“The clinical response endpoint offers the potential to identify new agents based on their ability to improve health status of patients while prolonging life,” Hickish concluded. “Clinical response [in this trial] translated into substantial benefit in overall survival.”
Discussant Differs on Data Interpretation
In his discussion, Arnold said the trial used a novel patient-relevant endpoint with a new score of symptom improvement parameters, which was written by the MABp1 manufacturer and developed together with an EMA working group.
“This is not entirely new, as a similar approach focusing on improvement of disease-related symptoms was used in the mid-1990s resulting in the registration of gemcitabine,” Arnold said, citing a publication by Howard Burris III, MD, and colleagues (J Clin Oncol 1997;15:2403-2413).
Arnold called out the claim of a “76 percent increase in improvement,” saying it was “misleading” because it represents only the relative improvement rate.
Furthermore, Arnold said, no reference was made to when the onset of symptom relief was noted during the 8-week period.
In an unusual development, Arnold issued a follow-up to his comments as discussant via a post-meeting press release from the ESMO Press Office.
This was prompted by a discussion after Hickish's presentation in which it was brought out that the number of responders in the trial—as assessed by improvements in tumor-related symptoms—included treatment patients as well as patients in the placebo group.
“In my discussion about the abstract, I emphasized that the ‘detailed analysis of clinical responders’ with the new symptom-improvement score included all 309 patients,” Arnold said in the press release. “However, the ‘correlations with disease stabilization’ and ‘numbers of patients with serious adverse events’ had only been done for ‘responders’ versus ‘non-responders’—independent from treatment, so that ‘responders’ to placebo were also included in this analysis.”
The same applies to the “overall survival” curve included, he noted, for which, only a subset of 175 patients was registered.
“In conclusion, the jury is still out on whether the study made sense, but as the results do not bear close inspection they also cannot be relied upon,” Arnold said.
Robert H. Carlson is a contributing writer.