Oncology Live®
March 2013
Volume 14
Issue 3

Rethinking Overall Survival: Still an Acceptable Sole Primary Endpoint in Cancer Clinical Trials?

A clinical commentary by Maurie Markman, MD, addressing whether overall survival can still be considered the sole acceptable primary endpoint in studies of novel anticancer therapeutic strategies.

Maurie Markman, MD

So that there is no confusion on this point: A major goal of all antineoplastic therapies and trials designed to evaluate them should be to improve survival—period. However, the question that this clinical commentary will address is not whether improved overall survival is a valid aim of such trials, but whether overall survival can still be considered the sole acceptable primary endpoint in studies of novel anticancer therapeutic strategies.

This question is critical to the development of effective strategies in nonresearch-based cancer management, and can be illustrated by the following phase III trial results involving entirely different cancers (ovarian vs lung cancer) and types of antineoplastic strategies (standard cytotoxic chemotherapy vs molecularly targeted drugs), and conducted more than a decade apart.1,2

Consider, first, a trial by Muggia et al, reported in 2000, in which patients with ovarian cancer were randomized to single-agent cisplatin, single-agent paclitaxel, or the combination of cisplatin and paclitaxel (Table 1).1 The study found a substantially higher objective response rate and superior time to disease progression in favor of the platinum-containing regimens (both the single-agent and combination regimens) compared with single-agent paclitaxel. However, there was no difference in overall survival between the study arms, and specifically between the platinum-containing and non-platinum-containing primary chemotherapy strategies.

This outcome almost certainly occurred because patients in both of the single-agent treatment arms crossed over to the alternative single agent or had the second drug added to their treatment, in some cases prior to the time of documented disease progression. As a result, by the time the study was completed and analyzed, the majority of patients had actually received both cisplatin and paclitaxel.

Table 1. Single Agent Cisplatin Versus Single Agent Paclitaxel Versus Cisplatin Plus Paclitaxel as Primary Chemotherapy in Advanced Ovarian Cancer1

Response Rate

Progression-Free Survival (median)

Overall Survival (median)



16.4 months

30.2 months



11.2 months

26 months

Cisplatin + Paclitaxel


14.0 months

26.6 months

P <.001, 2 cisplatin-containing regimens compared with paclitaxel

P <.001, 2 cisplatin-containing regimens compared with paclitaxel

P =.31, between the 3 treatment groups

Since there was no overall survival advantage observed in this trial, does this outcome indicate there was no advantage associated with employing both cisplatin and paclitaxel in ovarian cancer management, and that single-agent treatment with either of these therapies is of equivalent clinical utility? The answer to this question is unequivocally no.

In fact, a previous study by the same research group clearly demonstrated superior overall survival results in advanced ovarian cancer for cisplatin/paclitaxel versus cisplatin/cyclophosphamide.3 However, in that study, few patients randomized to the non-taxane-containing regimen (cisplatin/cyclophosphamide) ever subsequently received paclitaxel prior to their death from progressive cancer.

Thus, in the study by Muggia et al, the absence of an overall survival benefit within the trial population does not indicate a lack of benefit associated with the use of the two novel drugs.1 Rather, the data demonstrate that delivering a second clinically effective agent at some later point in the course of the illness may be as useful (in a statistical sense within a given patient population) as administering the drug as a component of the primary treatment regimen.

The bottom line is that all patients who participated in the study by Muggia et al were given the opportunity to potentially benefit from the delivery of the second active agent.1,4-6 Under such circumstances, there may be no difference in the measurement of survival between the study arms.

One could argue that a survival advantage might have been demonstrated if cisplatin had not been offered in the paclitaxel-only treatment arm, or if paclitaxel had not been offered in the cisplatin-only arm. However, such a decision would have been ethically unacceptable or even unconscionable, since it would have denied patients a treatment documented to be of genuine clinical utility in ovarian cancer.

Table 2. Primary Therapy With Gefitinib Versus Carboplatin/ Paclitaxel in Non-Small Cell Lung Cancer2,7

Progression-Free Survival (median)

Overall Survival (median)


10.8 months

27.7 months


5.4 months

26.6 months

P <.001

(P = .483)

The same points can be illustrated in a study of patients with advanced non-small cell lung cancer (NSCLC) and a documented sensitizing EGFR mutation who were randomized to receive, as primary therapy, either single-agent gefitinib or the combination of carboplatin plus paclitaxel (Table 2).2,7 Initial outcome data for this trial were published in 2010, and follow-up data were published in 2013. The trial demonstrated a highly statistically significant difference in time to disease progression in favor of treatment with gefitinib.7 However, overall survival was not different between the two study arms.2

As in the ovarian cancer trial, crossover from chemotherapy to gefitinib at time of disease progression was ethically mandated due to the known biological and clinical activity of this class of agents in this specific clinical setting. As a result, 98% of patients who were treated with primary chemotherapy ultimately received gefitinib.2 Thus, essentially all of the trial participants were given the opportunity to benefit from the targeted therapeutic.

Again, rather than failing to demonstrate the favorable impact of gefitinib on overall survival, this study could rationally and quite appropriately be interpreted to have revealed the ability of this class of agents to “rescue” patients from a less-than-optimal approach to primary treatment.

This commentary has highlighted only two examples of many that could be advanced to reveal the inappropriateness of continuing to insist that studies demonstrate an improvement in overall survival when effective therapy is available to “rescue” patients assigned to a less-effective study arm. Instead, trials are increasingly allowing for all patients (including individuals receiving the control drug) to have the chance to benefit from the best available treatments. Further, as cancer increasingly becomes a chronic illness, where multiple strategies are available to favorably impact time to subsequent disease progression, quality of life, and overall survival, it is clear that strategy will be even more common.

In summary, it is important to remember that the absence of a statistically significant survival benefit in phase III, randomized trials of antineoplastic drugs may be due to trial designs that allow for crossover of treatment arms, with the aim of achieving superior survival benefits. This is the only ethically correct way to treat patients who agree to become research subjects in a cancer clinical trial, and it is the way that physicians and investigators would want their own family members to be treated if they were participants in these research studies.


  1. Muggia FM, Braly PS, Brady MF, et al. Phase III randomized study of cisplatin versus paclitaxel versus cisplatin and paclitaxel in patients with suboptimal stage III or IV ovarian cancer: A Gynecologic Oncology Group study. J Clin Oncol. 2000;18:106-115.
  2. Inoue A, Kobayashi K, Maemondo M, et al. Updated overall survival results from a randomized phase III trial comparing gefitinib with carboplatin-paclitaxel for chemo-naïve non-small cell lung cancer with sensitive EGFR gene mutations (NEJ002). Ann Oncol. 2013;24:54-59.
  3. McGuire WP, Hoskins WJ, Brady MF, et al. Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med. 1996;334:1-6.
  4. McGuire WP, Rowinsky EK, Rosenshein NB, et al. Taxol: a unique antineoplastic agent with significant activity in advanced ovarian epithelial neoplasms. Ann Intern Med. 1989;111:273-279.
  5. Thigpen JT, Blessing JA, Ball H, Hummel SJ, Barrett RJ. Phase II trial of paclitaxel in patients with progressive ovarian carcinoma after platinumbased chemotherapy: A Gynecologic Oncology Group study. J Clin Oncol. 1994;12:1748-1753.
  6. Thigpen T, Vance R, Lambuth B, et al. Chemotherapy for advanced or recurrent gynecologic cancer. Cancer. 1987;60(8 suppl):2104-2116.
  7. Maemondo M, Inoue A, Kobayashi K, et al. Gefitinib or chemotherapy for non-small-cell lung cancer with mutated EGFR. N Engl J Med. 2010;362:2380-2388.

Related Videos
Julia Rotow, MD, clinical director, Lowe Center for Thoracic Oncology, Dana-Farber Cancer Institute; assistant professor, medicine, Harvard Medical School
Joshua K. Sabari, MD, assistant professor, Department of Medicine, New York University Grossman School of Medicine; director, High Reliability Organization Initiatives, Perlmutter Cancer Center
Alastair Thompson, BSc, MBChB, MD, FRCS
C. Ola Landgren, MD, PhD
Sara M. Tolaney, MD, MPH
Adam M. Brufsky, MD, PhD, FACP
Justin M. Watts, MD
Sara M. Tolaney, MD, MPH
Leah Backhus, MD, MPH, FACS, professor, University Medical Line, Cardiothoracic Surgery, co-director, Thoracic Surgery Clinical Research Program, associate program director, Thoracic Track, CT Surgery Residency Training Program, Thelma and Henry Doelger Professor of Cardiovascular Surgery, Stanford Medicine; chief, Thoracic Surgery, VA Palo Alto
Roy S. Herbst, MD, PhD, Ensign Professor of Medicine (Medical Oncology), professor, pharmacology, deputy director, Yale Cancer Center; chief, Medical Oncology, director, Center for Thoracic Cancers, Yale Cancer Center and Smilow Cancer Hospital; assistant dean, Translational Research, Yale School of Medicine