In recent months, I have written about one of the more important emerging changes in health care, the changing focus from volume to value, and the inevitable need for us to change our stance and strategy to maintain quality, empathetic patient care and the ability to make progress.
Subliminally, I have been conscious of a rising concern, somewhere near the back of my mind — specifically, the lack of connection between hype, hope and reality in the clinical trial publications about pancreatic cancer, as a surrogate for a greater concern about oncology promises in general.
I have the greatest admiration for those who have spent large parts of their careers attempting to manage such a vexing and frustrating disease. We are all aware that pancreatic cancer is often a silent killer, presenting late — often with few symptoms other than vague abdominal discomfort and sudden onset of jaundice — and remarkably resistant to all of our therapeutic strategies.
Fortunately, our surgeons have improved the techniques of Whipple and equivalent procedures, reducing both morbidity and mortality, and managing to cure patients with localized disease.
Although I have some difficulty evaluating the significance of some of the combined-modality approaches to locally extensive disease, given the preponderance of nonrandomized phase 1 and phase 2 data, I’m still happy to see that our brave investigators and patients are trying to make progress in that domain and are perhaps shifting the needle, albeit partly as the product of meticulous case selection.
My concern is more focused on metastatic disease, and perhaps I’m influenced a little by the death of a close friend a couple of years ago who chose to try combination chemotherapy, based on what he had read, and essentially wasted the last months of his life in a toxic and fruitless quest for a remission that he thought might be likely to be sustained.
When I entered oncology, pancreatic cancer was feared as a relentless and dangerous killer — intractable to most systemic therapies — and incidence was slowly rising.
Despite the hype of the popular news media, a review of national cancer statistics reconfirms worrying details. In 1981, 24,200 new cases of pancreatic cancer were identified, and 24,000 deaths were recorded. In 2012, 43,920 new cases of pancreas cancer were associated with 37,390 deaths. In 2015, it is projected that there will be 48,960 new cases, with 40,560 estimated deaths.
It appears that most of the progress is associated with the management of localized disease, presumably through improvements in surgery and support. More disturbing is that the 5-year survival for advanced disease remains at only 2%. Somehow that just doesn’t correlate with the enthusiasm that I read in the journals these days.
I can imagine your furrowed brow, and you ask, “Give me an example!”
Well, I have always viewed The New England Journal of Medicine as one of the key repositories of dramatic medical progress, so I did a literature search, placing “pancreas cancer” and “N Engl J Med” into the PubMed search engine.
I found 343 listed papers between 1946 and 2015. Many were focused on neuroendocrine tumors or diagnostic considerations, on symptom control and palliative care. However, I was perturbed when I focused on the full manuscripts published about treatment of advanced pancreatic cancer during the past 5 years. There was a series of elegant papers, and some reporting well-designed and well-powered studies. However, the bottom line was that median survival was still well less than a year, and long-term survival was approximately 5% or less.
So why were they published in “Man’s Best Journal of Medicine”?
I believe in the publication of negative data from well-conducted studies, but not with verbiage implying significant progress, and not necessarily in one of the premier general medical journals.
Four years ago, in an elegant randomized trial that compared FOLFIRINOX (5-FU plus leucovorin, irinotecan and oxaliplatin) with gemcitabine, Conroy and colleagues reported a median survival of 11.1 months vs. 6.8 months (P < .001) with only a tiny proportion surviving beyond 2 years. They described FOLFIRINOX as “an effective first-line treatment” while emphasizing the increase in median survival of only 4 months!
Von Hoff and colleagues reported increased survival with nab-paclitaxel plus gemcitabine compared with gemcitabine alone, noting median survival figures of 8.5 vs. 6.7 months, respectively (HR = 0.72; P < .001), commenting that the combination significantly improved survival. Once again, the paper was adorned with a set of forest plots, which were meaningless unless attached to real survival numbers.
I’m struggling with the thought that an increment of 1.8 months of median survival, or a tiny difference in OS at 21 months, can be described as “significant.” What was the point of recruiting more than 800 patients to demonstrate such a paltry change?
Statistical significance vs. clinical importance
We seem to be content to demonstrate statistically significant but clinically unimportant differences in outcome.
For example, the CALGB executed a well-designed study that tested the impact of adding bevacizumab (Avastin, Genentech) to gemcitabine. The randomized study, which included 602 patients, showed with high confidence that median survival in both arms was about 5.8 months.
I can’t help wondering whether a smaller study would have shown the same result with less confidence but less expense of clinical, fiscal and patient resources. Perhaps someone could have built in an interim analysis of futility that would have allowed earlier closure when it was clear that no useful difference would be shown.
Nonetheless, this study made it to Journal of Clinical Oncology, our premier oncology journal, but questions about needless expenditure of resources were apparently not raised.
When David P. Ryan, MD, published a thoughtful and detailed review on progress in the field in The New England Journal of Medicine — a paper that identified “progress” and a new standard of care while citing similarly depressing survival statistics — only two letters to the editor were published, and they focused only on nuances of diagnosis or symptoms. Where was the concern about the lack of progress?
I do understand that progress can be difficult and laborious, but I have always been influenced by the investigators in the battle against advanced renal cell carcinoma, who spent decades writing factual (almost dull, pro forma) negative reports until they eventually hit upon agents that actually did make a difference, and then trumpeted them with enthusiasm. The driver was that progress was easily seen when they identified drugs that really worked.
I definitely don’t wish to pick a fight with the gastrointestinal oncology fraternity. I recognize it is doing important and difficult work, and being both creative and industrious. However, we are in an era when medical autonomy is increasingly being challenged. If we are unable to look dispassionately at what we do and criticize ourselves accurately — or at least avoid hype and rhetoric when discussing disappointing outcomes — our critics will supervene, and government will regulate us with even more enthusiasm, notwithstanding its lack of expertise in our domain.
Conroy T, et al. N Engl J Med. 2011;doi:10.1056/NEJMoa1011923.
Kindler HL, et al. J Clin Oncol. 2010;doi:10.1200/JCO.2010.28.1386.
Ryan DP, et al. N Engl J Med. 2014;doi:10.1056/NEJMra1404198.
Siegel R, et al. CA Cancer J Clin. 2012;doi:10.3322/caac.20138.
Siegel RL, et al. CA Cancer J Clin. 2015;doi:10.3322/caac.21254.
Silverberg E. CA Cancer J Clin. 1981;31:13-28.
Von Hoff DD, et al. N Engl J Med. 2013;doi:10.1056/NEJMoa1304369.
For more information:
Derek Raghavan, MD, PhD, FACP, FRACP, FASCO, is HemOnc Today’s Chief Medical Editor for Oncology. He also is president of Levine Cancer Institute at Carolinas HealthCare System. He can be reached at firstname.lastname@example.org.
Disclosure: Raghavan reports no relevant financial disclosures.