library(escalation)## Loading required package: magrittr
p1 <- get_three_plus_three(num_doses=1, allow_deescalate=TRUE) %>%
get_dose_paths(cohort_sizes=rep(3, 2)) # D=1 trial enrolls at most 2 cohorts
g1 <- graph_paths(p1)
g1$width <- g1$height <- 250
g1
Dose transition paths computed by package escalation v0.1.4, for the trivial 3+3 trial with a single dose, under the allow_deescalate=TRUE regime which implements the “more common” variant described in Korn et al. (1994, p1801), in which the “requirement is to have 6 patients treated at the MTD (if it is higher than level 0).” The top, right yellow leaf node of this graph suggests that the case ‘1NNN’ proceeds immediately to declare \(MTD=1\) without enrolling a second cohort of 3.
p2 <- get_three_plus_three(num_doses=2, allow_deescalate=TRUE) %>%
get_dose_paths(cohort_sizes=rep(3, 4)) # D=2 trial enrolls at most 4 cohorts## You have requested 341 model evaluations. Be patient.
graph_paths(p2)
Dose transition paths computed by package escalation v0.1.4, for the 3+3 trial with 2 doses, under the allow_deescalate=TRUE regime which implements the “more common” variant described in Korn et al. (1994, p1801), in which the “requirement is to have 6 patients treated at the MTD (if it is higher than level 0).” Paths ‘1NNN, 2NNN, 2’ and ‘1NNT, 1NNN, 2NNN, 2’ appear to declare \(MTD=2\) after enrolling only 3 patients at dose level 2.
set.seed(2020)
true_prob_tox <- c(0, 0, 0) # drive simulation to top dose with zero toxicity
num_sims <- 1
sims <- get_three_plus_three(num_doses = 3, allow_deescalate = TRUE) %>%
simulate_trials(num_sims = num_sims, true_prob_tox = true_prob_tox)
sims[[1]][[1]][[1]]$fit## Patient-level data:
## # A tibble: 9 × 4
## Patient Cohort Dose Tox
## <int> <int> <int> <int>
## 1 1 1 1 0
## 2 2 1 1 0
## 3 3 1 1 0
## 4 4 2 2 0
## 5 5 2 2 0
## 6 6 2 2 0
## 7 7 3 3 0
## 8 8 3 3 0
## 9 9 3 3 0
##
## Dose-level data:
## # A tibble: 4 × 8
## dose tox n empiric_tox_rate mean_prob_tox median_prob_tox admissible
## <ord> <dbl> <dbl> <dbl> <dbl> <dbl> <lgl>
## 1 NoDose 0 0 0 0 0 TRUE
## 2 1 0 3 0 NA NA TRUE
## 3 2 0 3 0 NA NA TRUE
## 4 3 0 3 0 NA NA TRUE
## # … with 1 more variable: recommended <lgl>
##
## The model advocates stopping and recommending dose 3.Korn, Edward L., Douglas Midthune, T. Timothy Chen, Lawrence V. Rubinstein, Michaele C. Christian, and Richard M. Simon. 1994. “A Comparison of Two Phase I Trial Designs.” Statistics in Medicine 13 (18): 1799–1806. https://doi.org/10.1002/sim.4780131802.