The Illusion of Progress in Cancer

Two investigators from the MD Anderson Cancer Centre in Houston, an internationally renowned research institution, recently published a startling dressing-down of the Medical-Pharmaceutical Complex for creating the illusion of progress in the treatment of cancer. Their writing is direct, harshly critical, and blisteringly accurate, and as such I can not improve on it so present a few lengthy passages below. Their commentary was published in the January 20, 2009 issue of the Journal of Clinical Oncology.

Their main thesis: the initial impressive double-digit survival gains of past decades can no longer be achieved. Instead we are left to produce 'advances' with diminishing returns, at staggering expense, and of dubious benefit. We may even be discarding solutions of true benefit because our research systems are too blunt or cumbersome to detect an advantage in small groups of patients.

Low thresholds for judging effectiveness "benefit many stakeholders. Investigator careers are built on "positive" studies, statisticians become indispensable when advances are so small that only they can detect them, National Cancer Institute programs demonstrate that they help identify "active" new drugs, the US Food and Drug Administration demonstrates that it fosters safe access to "useful" new therapies, companies can market new agents, and providers can bill for new therapeutic options delivered. Drug sales yield advertising dollars to fuel publications and campaign dollars to fuel elections. Insurers simply pass on the higher costs by raising premiums. Almost everyone benefits. Unfortunately, the gain for patients is often exceptionally modest and at exorbitant cost."

"Large armies of dedicated people are fighting very hard in this war on cancer, but maladjusted research goals, governance, economics, politics, legalities, accepted clinical research practices, and dogma unintentionally impede progress."

"a drug that leads to dramatic responses, but only in 5% to 10% of patients, might be predicted to fail in randomized trials involving unselected patients." [...] "It is time to stop investigating efficacy of drugs against tumors defined solely by histopathologic type and to instead investigate efficacy against tumors defined by a combination of histopathology and molecular and genetic profiles." [Note: This has been done to some degree, e.g. for her2/neu-positive breast cancer and trastuzumab/Herceptin, for c-kit-positive gastrointestinal stromal tumours and imatinib/Gleevec, and others]

"Like amassed troops targeting only a nearby trench, large randomized trials aiming to increase survival by a few weeks are a huge expense likely to achieve only modest gains, because modest gains are precisely what they are aiming for. We instead need a plethora of highly mobile smaller trials each with the goal of achieving much more ambitious outcomes."

"The current regulatory burden in the conduct of clinical trials is to the war on cancer what World War I mud was to trench warfare. The thigh-deep, sticky Flanders mud jammed rifles, entrapped vehicles, weighed down massively caked clothing, pulled like glue at legs and boots, and swallowed and drowned those who stumbled and fell. This regulatory burden is onerous, misguided, and expensive, with little value added. For example, one author's (D.J.S.) department receives around 600 external serious adverse event (SAE) reports per month for agents that it is studying clinically. A single event may generate a dozen or more SAE reports (one for the original event, one for each follow-up, and a separate copy of each of these for each protocol using the agent). In addition to the SAE reports being highly repetitive, the overwhelming majority report events that are more likely related to tumor progression or comorbidity, are well-known toxicities of the agent, or are irrelevant for other reasons. Despite this, purportedly to raise the safety bar, each copy of each form received required a separate document be prep ared, submitted, reviewed, corrected (to properly conform to the shifting regulatory interpretation du jour) and resubmitted to the institutional review board (IRB). This does not improve patient safety but does magnify research costs and investigator frustration. Similarly, Humphreys et al documented that 16.8% of the total costs of an eight-site observational trial were devoted to IRB interactions, desp ite no visible effect on human subject protection, and the essential procedures of their study never changed substantially despite exchanges of more than 15,000 pages of material."

I recently submitted a clinical trial application for review by my IRB. They reviewed the consent document, a document meant to assist a patient in understanding a trial before making a decision to participate. The trial tests high doses of radiation therapy together with chemotherapy and a new targeted agent in this setting. The patients enrolled to the study have stage III lung cancer, a disease with a five-year survival of about 15%.

The treatments are potentially life-threatening in a variety of ways that I enunciated ad nauseum in the consent document I submitted. Nevertheless, the IRB objected to the omission of a statement that patients may find items on a questionnaire about their quality of life "upsetting." When someone's treatment or disease has a substantial risk of killing them, a likelihood of making them seriously ill, and a certainty of making them feel like hell for about four months, the possible eyebrow-raising discomfiture provoked by a questionnaire item is not a make-or-break feature of the trial.

The document runs 13 pages and contains redundant assurances about protection of privacy, superfluous statements about expenses such as specifying that parking costs will not be covered, and the like. It is a Frankensteinian concatenation of every paternal impulse ponderous regulatory institutions can muster, ostensibly for patient protection, more likely to limit liability.

I applaud the courage of Stewart and Kurzrock for making a case that is long overdue. Their piece is an electrifying manifesto, and should be required reading for every researcher and regulator. I have no doubt that they are incredibly frustrated and inhibited by the regulatory and granting environment they find themselves in. Who knows what great advances we have overlooked, what novel ideas have been turned away, how many eager researchers have been stonewalled by a process that is "perfectly designed to get exactly the results it gets."