Effective Targeted Therapy Needed to Establish Clinical Utility of a Molecular Diagnostic Test

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 has been considerable discussion and debate within the gynecologic oncology community for more than two decades regarding a role for second-look surgery (open or laparoscopic) in the management of advanced ovarian cancer. The argument centers on the specific question of the influence of this procedure on the outcome for patients with this malignancy.

Critics of the approach have appropriately concluded that there is no evidence that performance of this surgery improves survival, although there is certainly strong agreement that findings at the time of the second-look operation have strong prognostic relevance. Patients found to have no macroscopic disease experience a population-based overall statistically significant superior survival compared with women observed to have residual macroscopic cancer following the completion of frontline chemotherapy.

One large prospective analysis of a group of patients who either did or did not undergo a second-look procedure (physician/patient choice, no randomization) found no evidence that the surgery itself favorably impacted outcome.¹

Yet would it be correct to conclude that the performance of a second-look surgical procedure in the management of ovarian cancer could never be of clinical value?

In order to evaluate this question, one needs a clear understanding of the specific indication or indications for the actual performance of this surgery.

If the goal was the removal of all residual cancer (a therapeutic objective), one should reasonably expect patients who underwent the procedure to experience superior survival compared with patients not undergoing a second-look surgery. This outcome has not been observed,¹ strongly suggesting that the natural history of the disease is not significantly influenced by gross tumor removal at this point in the illness.

But if the goal of the second-look surgery was to define a unique patient population for whom a subsequently administered therapy based on the surgical findings would improve outcome, proof of the utility of this diagnostic role for the procedure requires the existence of such an effective therapeutic to test this hypothesis.

For example, what if a highly effective antineoplastic agent were developed that needed to be locally delivered but was only beneficial when employed in the setting of microscopic residual cancer due to its very limited ability to directly penetrate into tumor masses?

In this situation, the second-look surgical procedure would be a required diagnostic test designed to find patients who might be helped by utilizing a clinically active regional strategy and would exclude patients who would not experience benefit.

The very specific point to be made here is that in the absence of the effective antineoplastic drug, there would be no utility associated with obtaining the diagnostic data. Unfortunately, while there is no therapeutic answer at this point for the patients in this ovarian cancer setting, this framework of analysis can be applied to other diagnostics.

Evaluating Molecular Tests

One observes an analogous situation when considering the clinical utility of molecular diagnostics developed to define the utility of targeted therapeutic agents. In the absence of an effective therapeutic, the benefits of employing the diagnostic strategy will remain theoretical—not disproved, but also not proved.

An excellent example of the impact of the availability of effective target therapeutics is provided by comparing two reports from the phase I clinical trials program at The University of Texas MD Anderson Cancer Center. Two years ago, this program published a highly provocative analysis that revealed a substantially higher overall response rate among patients treated in phase I trials if the investigative drug selected for treatment was a match for a demonstrated target within the individual’s tumor compared with patients treated with an agent that was not matched to a demonstrated molecular target (27% vs 5%, respectively).²

However, in a recent report, although a higher objective response rate for the patient population able to receive therapy matched to a molecular target within their tumor was observed compared with the “unmatched” population (12% vs 5%, respectively) and the overall survival of this population was also substantially longer (median 11.4 months vs 8.6 months [P = .04], respectively), the overall response rate among the matched population in this analysis was clearly considerably lower than in the prior report (12% vs 27%)^.2,3

What accounts for this outcome? Of course, it is always difficult to provide definitive answers to such questions, particularly when dealing with small patient numbers and cross-study comparisons with known major differences within the populations evaluated such as age, sex, baseline stage, rate of disease progression, tumor histology, prior therapies, and presence of relevant comorbidities.

However, a highly logical explanation for the gap in results is the fact that different investigative therapies were available to these phase I investigators for use within the specific treated populations during the different time intervals. As a result, the potential for a given specific target (eg, a PI3K mutation) to be favorably impacted by a clinically active therapeutic agent was almost certainly very different in the two analyzed groups.

Thus, just as with second-look surgical assessments in ovarian cancer, proving the clinical utility of a diagnostic testing strategy requires the availability of an effective therapeutic whose benefit can be accurately predicted by the diagnostic test. In the absence of such a therapeutic to explore in a given setting, the utility of the diagnostic approach or, conversely, the lack of utility (as remains the case with second-look surgical reassessments in ovarian cancer), will unfortunately remain an unanswered question.

References

1. Greer BE, Bundy BN, Ozols RF, et al. Implications of second-look laparotomy in the context of optimally resected stage III ovarian cancer: a non-randomized comparison using an explanatory analysis: a Gynecologic Oncology Group study. Gynecol Oncol. 2005;99(1):71-79.
2. Tsimberidou AM, Iskander NG, Hong DS, et al. Personalized medicine in a phase I clinical trials program: the MD Anderson Cancer Center initiative [published online September 10, 2012]. Clin Cancer Res. 2012;18(22):6373-6383.
3. Tsimberidou AM, Wen S, Hong DS, et al. Personalized medicine for patients with advanced cancer in the Phase I program at MD Anderson:validation and landmark analyses [published online July 1, 2014]. Clin Cancer Res. 2014; 20(18):4827-4836.