Maurie Markman, MD
Editor-in-Chief of OncologyLive
Senior vice president for Clinical Affairs and National Director for Medical Oncology
Cancer Treatment Centers of America, Eastern Regional Medical Center
There is a tension in the air these days when it comes to the manner in which evidence-based oncology research should be conducted in this revolutionary era, and that tension is evident on the pages of peer-reviewed articles describing novel findings in clinical molecular oncology, in editorials discussing these studies, and in commentaries related to future approaches to discovery.
The “academic purists” continue to maintain that except in the most unusual of situations randomized phase III trials should be required, and in the majority of circumstances the primary endpoint should be a statistically significant improvement in overall survival. Reluctantly, some may concede that progression-free survival is an acceptable study endpoint but generally only if it can be documented that such data serve as a reliable surrogate for overall survival.
The magnitude of the disconnect between the mantra of evidence-based phase III randomized trials and the reality that many such studies are initiated but never completed or reported in the peer-reviewed medical literature is testimony to the need for a far more rational discourse on the nature of evidence required in oncology for a new “standard-of-care” to gain general acceptance within the medical community.1
Aspirin Issue Illustrates Dilemma
A poignant example of the failure of rhetoric regarding randomized phase III clinical trials and the need to consider alternative approaches to determine clinical utility is provided by a recently published commentary discussing the relevance of the administration of aspirin to patients with colon cancer whose malignancy contains a PIK3CA mutation.2
In this paper, Tougeron et al declare that the results of a phase III randomized trial should be required before aspirin is considered a possible “standard-of-care option.”
The genesis of the specific discussion relates to a previously reported landmark study that retrospectively examined more than 600 colon cancers where detailed information was available regarding the individual patient’s aspirin intake.3
The analysis revealed a striking 83% reduction in the risk of subsequent death following the diagnosis of the malignancy with subsequent aspirin use but only in those individuals whose cancer possessed a PIK3CA abnormality, with no benefit being observed in the presence of wild-type PIK3CA.3
At least one additional retrospective analysis has provided confirmation of these highly provocative and potentially paradigm-changing data.4
Now how realistic is it that a definitive phase III trial would be conducted to address the specific relevance of the utility of aspirin in patients with colon cancer whose malignancy contains this molecular abnormality? And assuming such a study is ever completed, how many years from its initiation would the population of patients who might potentially benefit from employing this approach have to wait before the survival outcomes are available?
Although specific plans to explore aspirin use in PIK3CA-mutated colon cancer have not been announced, Tougeron et al suggest that the large international phase III ASCOLT trial5
examining this therapy for patients independent of mutational status may be able to address the molecularly targeted question. However, it is important to acknowledge the timeline of the ASCOLT trial. Launched in early 2009, the trial is still accruing participants; it has an estimated completion date of June 2021.6
Thus, from initiation of accrual until the final trial results, a time of interval of approximately 12 years will have elapsed, assuming that accrual on the study progresses at a prospectively defined rate that painful experience shows is frequently difficult to achieve.
Therefore, recognizing that PIK3CA mutations are found in approximately 15% to 20% of all colon cancers,3 and assuming aspirin intake only favorably impacts colon cancer–specific survival in the presence of such a mutation, how many patients would have to be entered into such a study and over how many years—or should one ask how many decades—before a so-called “definitive evidence-based answer” to the question of the impact of this pharmaceutical agent on outcome is known? The answer to this question might reasonably be described as unrealistic, distressing, or unacceptable.