Skipper and colleagues proposed a “steep” dose-response relationship in cancer chemotherapy, referring to a linear relationship between drug doses and fractional cell kill. This concept dated back to the 1960s when Skipper and Schabel predicted a log cell kill model for antineoplastic drugs. In this model, the relationship between tumor cell kill and drug dose was exponential, with the number of cells killed by a given dose of drug being proportional to both the dose of the drug and the number of cells exposed to the drug.
The steepness of the dose-response curve implies that a disproportionately high number of cancer cells are killed when drug doses are minimally increased. Thus, the higher the dose, the more effective the drug. This is in contrast to most other classes of therapeutic agents, which exhibit a sigmoidal dose-response relationship, with a linear relationship between dose and response over a relatively narrow range of drug doses. The plateau in this traditional curve suggests that after a certain threshold dose, further increments do not lead to an increased response.
Toxicity and Efficacy In addition to the relationship between dose and anti-tumor response, cytotoxic agents also exhibit a dose-toxicity relationship. Thus, dose-related toxicity is regarded, as a surrogate for efficacy. The highest safe dose is assumed to be the one most likely to be efficacious. These assumptions underlie the design of phase I studies. Doses of drugs are escalated until the achievement of significant, but reversible, toxicity. These toxicities that preclude further dose escalation, represent the dose-limiting toxicity (DLT).
DLTs are defined in advance in phase I trials. The maximum-tolerated dose (MTD) is defined as that dose producing a certain frequency of DLTs within the treated patient population. The MTD is either taken as one dose level below the highest toxic dose (ie the DLT dose), in which case the MTD is the recommended phase II dose; or the MTD is sometimes taken as the DLT dose, in which case the dose level below the MTD is the recommended phase II dose. How is the starting phase I dose selected ?
Initial toxicologic studies are performed in rodents (typically mice) and another large species (typically canines). In the mice studies, a dose at which approximately 10% of the mice die (the murine LD10), is defined. One tenth of this murine equivalent LD10 (0. 1 MELD10), expressed in milligrams per meters squared, has historically been a safe starting dose in humans when toxicologic studies in a second species (eg, rat, dog) do not show substantial differences in the dose-toxicity relationship.
Patients are accrued in cohorts of three, and escalating the dose according to a modified Fibonacci sequence in which dose increments become gradually lower as the MTD is approached(example, dose increases of 100%, 65%, 50%, 40%, and 30% to 35% thereafter). The dose escalation is continued in cohorts of three patients until typically, 2 DLTs are seen in a maximum of 6 patients. Ethical Issues i) Substantial numbers of patients are treated at non-therapeutic doses. It has been calculated that the majority responses in phase I studies occur within 80% to 120% of the recommended phase II dose.
These considerations raise ethical pressures to treat fewer patients at the initial dose levels in the absence of toxicity. ii) Informed consent. Oftentimes patients are desperate and cannot make truly informed consent iii) Discordance between investigator objectives (define pharmacology of agent) and patient objective (therapeutic benefit) Alternate Designs The modified Fibonacci escalation scheme can lead to multiple dose escalations utilizing dozens of patients before the MTD is defined. There are now multiple agents for phase I testing, and more efficient phase I designs are needed.
The ideal design should accomplish the following: i) Minimize the number of patients treated at sub-therapeutic doses ii) Minimum number of patients treated per dose level iii) Minimize unnecessary toxicity iv) Rapid dose-escalation leading to rapid completion of trial Several statistical methods have been developed for novel Phase I dose escalation designs. The primary goal of all these designs is to shorten the duration of phase I trials and to enhance the precision of the phase II dose recommendation.
These methods are typically based on the concept using toxicity, as the end point of the trial. A mathematical function is created that describes the hypothesized relationship (curve) between the incidence of DLT and dose. This curve is reasonably predicted to assume a sigmoid shape for which the MTD must be estimated first. As information regarding the occurrence or absence of toxicity accumulates from the trial, the original estimate of the MTD is updated to more accurately fit the hypothesized curve to the actual data.
Under these types of trial designs, the occurrence of toxicity results in an adjustment of the curve to match the probability that one is now approaching the MTD. Conversely, the absence of toxicity results in adjustments of the curve to match the probability that one is not yet at the MTD. Therefore, the occurrence of no DLT in several sequential patients results in a statistical prediction that the dose can be more rapidly escalated in a safe manner. Designs based on these paradigms include :
i)Pharmacokinetically guided dose escalation ii)Accelerated titration designs iii)Modified continual reassessment methods iv)Intrapatient dose escalation The Concept of Optimal Biologic Dose for Novel, Non-Toxic Agents Advances in molecular biology have led to a new generation of anticancer agents that inhibit aberrant and cancer specific proliferative and anti-apoptotic pathways. These agents may be cytostatic and may produce relatively minimal organ toxicity, compared to standard cytotoxic agents.
This has fueled interest in alternatives to toxicity as a surrogate endpoint in phase I trials. The concept of an “optimal biologic dose” defined as a dose that reliably inhibits a drug target or achieves a target plasma concentration, is seen as desirable and appropriate for the phase I study of mechanism-based, relatively non-toxic novel agents. This idea is appropriate, if certain inherent problems can be resolved. In the case of a pharmacokinetic end point, it has to be shown that the target concentration chosen can inhibit the drug target in patient tumors.
This requires accounting for plasma protein binding that determines the amount of free drug available to interact with the target, as well as inter-individual variations in drug absorption and metabolism. When target modulation is chosen as the end point, the drug target as well as the magnitude of inhibition necessary for clinical benefit have to be known. Finally, while target inhibition in normal tissue may provide important supplementary information, critical drug development decisions will need to be made with information gleaned from target suppression in tumor samples.
An optimal biologic dose should inhibit the target in patient tumors. Most importantly, there should be absolute certainty of the drug target, and there should be evidence that modulating the target in tumors consistently leads to growth inhibition. The selected dose should incorporate the fact that there will be wide variations in steady-state drug levels in patients. Having outlined these issues, how does one select a phase II dose of a drug with minimal dose-dependent organ toxicities? Apart from immunotherapeutic agents, whose development has been recently reviewed, it is debatable if many such agents currently exist.
Most small molecule inhibitors of cellular proteins will demonstrate chronic low grade toxicities at high doses that preclude continuous dosing. Thus, the concept of MTD will need to be re-defined as a dose that can be safely administered chronically. With this definition, an MTD can be established for most drugs. In the rare case of truly minimally toxic agents, the optimal dose may be defined by saturation in absorption, quantity of tablets to be ingested or volume of drug to be infused, and other practical issues that would preclude dose escalation.
In a review of 60 Phase I studies of targeted agents, Parulekar and Eisenhauer found that the optimal biologic dose rarely formed the basis of dose selection. Ultimately, until we are able to incorporate modulation of validated drug targets in tumors into studies and utilize information on allelic variants of polymorphic genes responsible for drug transport and metabolism to select doses for individual patients, efforts have to be made to define the MTD of phase I agents based on classical definitions of cycle I toxicity, as well as toxicities from, and the feasibility of chronic administration.