select.mtd.kb {Keyboard}

Maximum Tolerated Dose (MTD) Selection for Single-agent Trials

Description

This function selects the maximum tolerated dose (MTD) after the single-agent trial is completed.

Usage

select.mtd.kb(
  target,
  npts,
  ntox,
  cutoff.eli = 0.95,
  extrasafe = FALSE,
  offset = 0.05
)

Arguments

target

The target dose-limiting toxicity (DLT) rate.

npts

A vector containing the number of patients treated at each dose level.

ntox

A vector containing the number of patients at each dose level who experienced a DLT at each dose level.

cutoff.eli

The cutoff to eliminate an overly toxic dose and all higher doses for safety.
The default value is 0.95.

extrasafe

Set extrasafe=TRUE to impose a stricter stopping rule.
The default is FALSE.

offset

A small positive number (between 0 and 0.5) to control how strict the stopping rule is when extrasafe=TRUE. A larger value leads to a stricter stopping rule.
The default value is 0.05.

Details

The Keyboard design starts by specifying a target toxicity interval (referred to as the "target key") such that any dose with a toxicity probability within that interval can be practically viewed as the MTD. Based on this interval's width, the Keyboard design forms a series of equally wide keys that span the rest of range from 0 to 1.

This function selects the MTD based on isotonic estimates of toxicity probabilities, selecting dose level j* for which the isotonic estimate of the DLT rate is closest to the target. If there are ties, then we select from the ties the highest dose level when the estimate of the DLT rate is smaller than the target, or the lowest dose level when the estimate of the DLT rate is greater than the target. The isotonic estimates are obtained by applying the pooled-adjacent-violators algorithm (PAVA) [Barlow, 1972].

For some applications, investigators may prefer a stricter stopping rule to ensure the lowest dose is not overly toxic. This can be achieved by setting extrasafe=TRUE, which imposes the following stricter safety stopping rule:
Stop the trial if
(i) the number of patients treated at the lowest dose \ge 3, and
(ii)

Pr((toxicity rate of the lowest dose > target) | data) > cutoff.eli - offset

As a tradeoff, the strong stopping rule will decrease the MTD selection percentage when the lowest dose actually is the MTD.

Value

The function returns a list with:

1. the target toxicity probability ($target),
   

2. the selected MTD ($MTD),
   

3. the isotonic estimates of the DLT probability at each dose and corresponding 95% credible interval ($p_est),
   

4. the probability of overdosing defined as
   Pr(toxicity > target | data) ($p_overdose).

Note

The MTD selection and dose escalation/de-escalation rules are two independent components of the trial design. When appropriate, another dose selection procedure (e.g., one based on a fitted logistic model) can be used to select the MTD after completing the trial using the Keyboard design.

Author(s)

Xiaomeng Yuan, Chen Li, Hongying Sun, Li Tang and Haitao Pan

References

Yan F, Mandrekar SJ, Yuan Y. Keyboard: A Novel Bayesian Toxicity Probability Interval Design for Phase I Clinical Trials. Clinical Cancer Research. 2017; 23:3994-4003. http://clincancerres.aacrjournals.org/content/23/15/3994.full-text.pdf

See Also

Other single-agent functions: get.boundary.kb(), get.oc.kb()

Examples

### Single-agent trial ###

n <- c(3, 3, 15, 9, 0)
y <- c(0, 0, 4, 4, 0)

selmtd <- select.mtd.kb(target=0.3, npts=n, ntox=y)

selmtd