If Only That Phase III Oncology Trial Had Been Designed Differently

Maurie Markman, MD, editor in chief, is president of Medicine and Science at Cancer Treatment Centers of America and clinical professor of medicine, Drexel University College of Medicine

Maurie Markman, MD

The cancer treatment community's ultimate perception of a successfully completed phase III randomized trial depends in large part on how well the trial was conceived and structured. To permit adequate accrual in a timely manner and optimize the chances for a study to achieve success, the question it poses must be relevant to ensure interest by clinical investigators, referring clinicians, and potential research subjects. Further, the initiative must have adequate funding for data collection and analysis, translational laboratory investigations, and other trial components.

Even with essential funding and interest, a successfully completed trial must comprise other critical components. These include compliance with eligibility and ineligibility criteria; satisfactory collection and documentation of required data; careful and objective analysis of the study results; and subsequent, timely publication in the peerreviewed literature.

In the development of a randomized trial examining a novel therapeutic approach to cancer management, no element is more important than the study design itself. As a result, investigators and study sponsors will spend considerable time and effort to ensure that a contemplated study will appropriately address the critical questions being posed regarding efficacy and toxicity. Depending on the preexisting experience with the drug or regimen, the study design will incorporate knowledge of prior trials employing the strategy as well as preclinical data that might provide insight into potential effectiveness and adverse effects.

Further, in trials designed for potential registration purposes, trial investigators must work to obtain the very specific outcome measures required by agencies for evaluation of efficacy. According to the study or clinical setting, investigators will include a control arm, for comparison with the novel strategy, and establish the degree of superiority or noninferiority that the trial should demonstrate upon completion.

One aspect of clinical trials that has perhaps received somewhat less attention than it warrants is the risk that study results may lead investigators and others to question the study’s design in retrospect.

Consider, for example, the recently reported phase III randomized trial in the primary chemotherapeutic management of small-volume residual advanced ovarian cancer that examined the clinical utility of the regional (intraperitoneal) delivery of platinum-based therapy versus systemic administration of the agents.¹ Of note, several previously conducted phase III trials had revealed the overall survival advantage of intraperitoneal cisplatin (compared with systemic delivery) administered at a 100 mg/m² dose.² Unfortunately, this dose level was found to result in considerable toxicity (eg, emesis, fatigue, risk of neuropathy), although, importantly, investigators observed no increase in treatment-related mortality.

Because of the adverse effect profile, the investigators decided to further explore in a subsequent randomized trial the potential of regional therapy employing a lower dose of cisplatin (75 mg/m²) versus intraperitoneal carboplatin.¹ This new study failed to reveal a survival advantage (progression-free or overall) for the regional approach. However, a study arm employing the original cisplatin dose level (100 mg/m²), which had demonstrated the overall survival benefit, was not included in this trial.

This raises the obvious question of whether the previous study results were in error (“statistical flukes”), or, perhaps, the 25-mg/m² reduction in the regional cisplatin dose level resulted in the elimination of the survival advantage. Unfortunately, in the absence of a higher cisplatin dose control arm, the answer to this question is simply unknown.

Further complicating matters, investigators in this trial decided, interestingly, to include bevacizumab (Avastin) in each study arm (both systemic and intraperitoneal), resulting in the absence of a non—bevacizumab-containing control arm. So, it must also be asked whether the use of bevacizumab somehow eliminated the benefits of regional cisplatin delivery entirely or, alternatively, the lower dose of cisplatin or substituted intraperitoneal carboplatin did it. Again, in the absence of a control arm for comparison, the answer to these questions will remain unanswered. And unfortunately, more than 1500 patients participated in this study with little or no resolution to the relevant question of the role of regional drug delivery in the management of ovarian cancer.

Of course, it is not difficult to raise similar concerns with the designs of other reported, clinically relevant trials published in high-impact medical journals. For example, sticking with the theme of advanced ovarian cancer studies, and at the risk of quite appropriately being accused of sounding very much like a Monday morning quarterback, I can ask the following questions related to 2 landmark, paradigm-changing studies.

The SOLO1 trial (NCT01844986) demonstrated a major benefit from the delivery of 2 years of olaparib (Lynparza) as frontline maintenance therapy in BRCA mutation—positive advanced ovarian cancer, but would the trial have demonstrated even greater superiority in progression- free survival if the agent had been continued (in the absence of disease progression or excessive toxicity) for 3 or even 4 years?³

In the phase III Gynecologic Oncology Group trial (NCT00262847) of frontline maintenance bevacizumab, would the positive progression- free survival advantage associated with delivery of this agent have been even longer if the antiangiogenic drug had been continued (in the absence of disease progression or excessive toxicity) for more than 22 cycles?⁴

To be clear, my intent in writing this commentary has not been to criticize the investigators or pharmaceutical sponsors of the highlighted studies. Rather, the goal has been to emphasize both the critical relevance of very careful consideration of the impact of trial design on the interpretation of possible study results and the legitimate risk that individual trials, no matter how meaningful or even “positive” their outcomes may be, may raise as many clinically relevant questions as they answer.

References

1. Walker JL, Brady MF, Wenzel L, et al. Randomized trial of intravenous versus intraperitoneal chemotherapy plus bevacizumab in advanced ovarian carcinoma : an NRG Oncology/Gynecologic Oncology Group study. J Clin Oncol. 2019;37(16):1380-1390. doi: 10.1200/JCO.18.01568.
2. Elit L, Oliver TK, Covens A, et al. Intraperitoneal chemotherapy in the first-line treatment of women with stage III epithelial ovarian cancer: a systematic review with metaanalyses. Cancer. 2007;109(4):692-702. doi: 10.1002/cncr.22466.
3. Moore K, Colombo N, Scambia G, et al. Maintenance olaparib in patients with newly diagnosed advanced ovarian cancer. N Engl J Med. 2018;379(26):2495-2505. doi: 10.1056/NEJMoa1810858.
4. Burger RA, Brady MF, Bookman MA, et al; Gynecologic Oncology Group. Incorporation of bevacizumab in the primary treatment of ovarian cancer. N Engl J Med. 2011;365(26):2473-2483. doi: 10.1056/NEJMoa1104390.