Grand Rounds: Clinical Development of Anti-Cancer Drugs: Succeed Slowly or Fail Fast

October 3, 2008

Bellevue AmphitheaterCommentary by Daniel Egan MD, PGY2

Please also see the clinical vignette presented during last week’sgrand rounds

On Wednesday, September 24, 2008, the NYU Department of Medicine Grand Rounds featured guest speaker Hilary Calvert, MD, Professor of Medical Oncology at the University of Newcastle upon Tyne, England, with a lecture entitled “Clinical Development of Anti-Cancer Drugs: Succeed Slowly or Fail Fast.”

Dr. Calvert first oriented the audience by reviewing the traditional method in which new chemotherapeutic drugs have been developed and introduced to the market.  Traditionally, the process begins with animal trials, in which toxicity data is collected for various species, and is later extrapolated to humans after adjustments for body surface area, as demonstrated by Freireich et al. (Cancer Chemotherapy Reports, 1966).   Phase 1 studies are then designed to examine pharmacokinetics, find dose-limiting toxicities in humans, and document any observed side effects, with the goal of recommending a starting dose for subsequent Phase 2 studies, which then begin to assess drug efficacy.

Next, variations on the early developmental process were presented, along with their advantages and disadvantages.  Pharmacologic-guided Phase 1 trials, in which a target area under the curve (AUC) is established from pre-clinical studies, and doses are escalated in large increments to achieve the AUC, were described as requiring less time and money but with accuracy affected by inter-patient variability.  Continual reassessment stochastic models were depicted as potentially more accurate and also requiring less time, but with the drawback of being exceedingly complicated in design.  Accelerated-phase trials, in which rapid drug dose escalations are made within even smaller numbers of patients, are extremely fast and informative, yet can be hazardous to study participants and suffer from little statistical confidence.

Dr. Calvert then spoke about two drugs “that nearly failed” during development, but which were ultimately salvaged and are successfully used to treat patients today.  Carboplatin was introduced in the 1980s as a potentially safer alternative to its predecessor, cisplatin; yet the drug was limited by resulting thrombocytopenia.  However, with the employment of AUC-based dosing by glomerular filtration rate (GFR), a significant improvement in the drug’s safety profile led to widespread clinical and commercial success for the drug.  Similarly, pemetrexed, an antifolate drug, was initially plagued by a high incidence of treatment-related deaths, mainly due to neutropenia and infection.  The drug was almost abandoned.  However, it was later discovered that elevated homocysteine levels, as a marker for antifolate activity, were associated with pemetrexed toxicity.  Subsequent trials of B12 and folate supplementation prior to administration of pemetrexed have resulted in decreased toxicity observed with this drug, and it is now widely used as a therapeutic agent, being licensed for pleural mesothelioma and lung cancer.

Lastly, Dr. Calvert discussed the special challenges in developing a class of targeted agents, focusing on the PARP inhibitors.  These drugs inhibit PARP-mediated single-stranded DNA repair, and are expected to potentiate the effects of monomethylating agents, such as temozolamide.  They are also expected to be most active against cells already defective in DNA repair, especially those arising in patients who are carriers for the BRCA1 or BRCA2 gene.  While dosing of traditional, systemically cytotoxic agents is more correlated with preclinical and Phase 1 toxicity profiles, Dr. Calvert suggested the proper dosing of newer, targeted therapies remains difficult to arrive at.  Although it is frequently possible to measure the drug effect on the target, it is frequently unclear what effects these agents will have on target homologues, and how data from animal studies will be able to be extrapolated into humans.  Furthermore, the optimal level and duration of target inhibition to achieve both efficacy and safety may be hard to define.