Only about 5% of patients with cancer are recruited into oncology clinical trials, raising a concern that unless this improves, progress in the development new treatments for cancer may be delayed.
David E. Gerber, MD
Only about 5% of patients with cancer are recruited into oncology clinical trials,1,2raising a concern that unless this improves, progress in the development new treatments for cancer may be delayed. In 2013, Denicoff et al3published the updated summary and recommendations of the Cancer Trial Accrual Symposium. It considers aspects of barriers to recruitment, including patient eligibility, clinical trial awareness, provider-physician issues, and site and logistical support.David E. Gerber, MD, is a thoracic medical oncologist and associate professor of internal medicine in the division of hematology and oncology in the Harold C. Simmons Comprehensive Cancer Center at the University of Texas Southwestern Medical Center in Dallas. He and his colleagues have tackled the issue of a particular trial exclusion criterion, prior cancer, as influencing patient accrual, focusing specifically on lung cancer clinical trials.1,2,4
“We had noticed in our practice that increasingly numerous and stringent eligibility criteria limited enrollment in clinical trials,” said Gerber. “In particular, we noted the widespread and longstanding practice of excluding patients with prior cancer from participating in clinical trials for their current cancer,” and he added “…there was little evidence to support the assumption that patients with prior cancer will do worse than those without.”
As a first step, in an earlier paper, Gerber et al analyzed lung cancer clinical trials either sponsored or endorsed by the Eastern Oncology Cooperative Group, for exclusion criteria for a previous cancer diagnosis.4They used Surveillance Epidemiology and End Results (SEER)-Medicare-linked data for 210,509 patients to determine the prevalence, type, stage, and timing of prior cancer. They used the data to predict the proportion and absolute numbers of patients excluded from trials because of prior cancer.
“We found that up to 18% of patients are excluded from lung cancer clinical trials for this reason alone, translating to more than 200 patients for some large phase III trials,” explained Gerber.
Describing their latest paper,1he continued, “We determined the prognostic impact of a prior cancer diagnosis in patients with lung cancer. For this analysis we specifically focused on stage 4 lung cancer.” The researchers used the SEER program linked with Medicare data to obtain treatment and outcome data. Records retrieved were limited to patients older than 65 years who were diagnosed with stage 4 lung cancer between 1992 and 2009. Patients with previous primary lung cancer were excluded. Data regarding carrier claims and inpatient and outpatient treatments for comorbidities were also collected.
The primary outcomes were overall (all-cause) survival and lung cancer mortality. Survival was defined as the period between diagnosis and death.
The team analyzed data on 102,929 patients with stage 4 lung cancer (only non-small cell or small cell lung cancer), which included 15,170 (14.7%) patients with a prior cancer. The table shows the commonest prior cancers identified in men and women diagnosed with stage 4 lung cancer. The incidence of prior cancer was higher in older, male, non-Medicaid patients. Information was also available for the stage of the prior cancers. Interestingly, 76% of them had either localized or regional disease and only 6.3% were classified as distant stage. Most of them had occurred ≤5 years prior to the diagnosis of lung cancer.
“We had found that, ‘5 years prior to the lung cancer diagnosis,’ is the most commonly used restriction for prior cancer, suggesting that most patients with prior cancers are still being excluded for this reason,” stated Gerber.
Patients with a prior cancer were found to have statistically significant superior all-cause and lung cancer-specific survival on Kaplan-Meier analysis (P<.001) versus patients with no prior cancer. When the data were analyzed using multivariable covariate-adjusted Cox models, prior cancer was also found to be associated with favorable all-cause survival (hazard ratio [HR], 0.93; 95% confidence interval [CI], 0.91-0.94;P<.001) and lung cancer-specific survival (HR, 0.81; 95% CI, 0.79-0.82;P<.001). They also examined a subset of the patients who more closely resembled a
clinical trial population. They were younger than 75 years, had no comorbidities, and had received chemotherapy for stage IV lung cancer (n = 9024). Thirteen percent of these patients had prior cancer and, once again, in common with the main analysis, there was no adverse effect of prior cancer on outcomes versus no prior cancer.
Commenting on these unexpected findings, Gerber said: “Somewhat surprisingly, we found that patients with stage 4 lung cancer who had a history of prior cancer not only did as well as those without a prior cancer, but actually seemed to do a little better,” and further pointed out that, “Their overall survival was almost 10% better and their lung cancer-specific survival was almost 20% better than those stage 4 lung cancer patients without a prior cancer. Thus, we believe that these patients should not be excluded from clinical trial participation.” However, he cautioned that, “Importantly, we cannot generalize these findings (limited to stage 4 lung cancer) from this paper to other stages of lung cancer, or to other cancer types.”This study was not designed to reveal an underlying mechanism, but it is interesting to consider possible hypotheses. The Cancer Genome Atlas (TCGA) has published extensive analyses of lung cancer.5,6Considering whether TCGA could possibly contribute to explaining the results, Gerber said, “There are a number of potential reasons for our unexpected observation. The first is that patients who survive one cancer may do better when they get subsequent cancers. This would suggest a different biologic phenotype, a scenario in which molecular profiling data could be relevant.”
Another possible explanation is that the lung cancer was detected during clinical care for the earlier cancer, and not by symptoms. This would result in a longer survival interval measured from time of diagnosis to time of death. “One might consider this a form of lead-time bias,” said Gerber.
This study1,2makes the case for the inclusion of patients with a prior cancer in clinical trials for advanced lung cancer. Researchers hope that such a change would improve recruitment into these trials and possibly result in improved treatments.