Hi,

Thanks it is interesting to add Survival plots
The phenoTest menu may has in part what do you expect (without curve).  It is possible to cluster a list of genes significantly
regulated (gene expression) at a range of tumor size (continuous variable) or tumoral stage (ordinal variable, Figure 12) for recurred or DiseaseFree cases (survival and categorical).

The Figure 14  associate Overall Survival Status of patients depending on the fold change of gene expression and the Level of PSA in prostate cancer.  The two last columns show opposite sign between the PSA fold change and the living status which is expect in medicine

YES, It is easy to add this plots. In the next release I will add Survival function in plot menu :)

here two examples from survival package:

##fit a Kaplan-Meier and plot it
fit <- survfit(Surv(OS_MONTHS,OS_

STATUS)~DFS_STATUS, data=ClinicData)
plot(fit, lty=2:3, xlab="OS_MONTHS", ylab="OS_STATUS: Survival", main="Lung Squamous Cell Carcinoma (TCGA)")
legend(60, 1.1, c("DiseaseFree", "Recurred/Progressed"), lty = 2:3)

# fit a Cox proportional hazards model and plot the 
#predicted survival for a 60 year old

fit<- coxph(Surv(OS_MONTHS, OS_STATUS)~ AGE, data= ClinicData)
plot(survfit(fit, newdata=data.frame(AGE=60)), xlab = "OS_MONTHS", ylab="OS_STATUS: Survival", main="Lung Squamous Cell Carcinoma (TCGA)")
text(90,0.95,"Cox: Predicted survival for a 60 year old", cex=1)

 

Please tell me which others Survival models are useful for biologists.

Thanks,

Karim

On Fri, Mar 27, 2015 at 4:00 AM, Dario Strbenac <dstr7320@uni.sydney.edu.au> wrote:
An analysis that biologists would be interested in is a Kaplan-Meier analysis with a survival time plot. Can the package do this ? It's the most obvious way to integrate gene expression (allow the samples to be put into a low expression and high expression group, based on a threshold) and clinical data, but I don't see how it could be done with canceR.

--------------------------------------
Dario Strbenac
PhD Student
University of Sydney
Camperdown NSW 2050
Australia

_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

Dear All,
 Survival Plots  In medical research is often useful to estimate the survival of patient amount time at tumor stage, or
after treatment or particular event. The OS MONTHS is the overall Survival duration of patient after
the first surgery. The OS STATUS is the event at OS MONTHS. The survival plot works only with
studies that have non empty OS MONTHS and OS STATUS.

 The Kaplan-Meier estimator measures the fraction of patients in life for a
certain amount of time after the first surgery. The survival curve can be created assuming various
situations. This can be calculated for two groups (DiseaseFree, Recurred) or more of subjects. When
the shapes of the curves are similar, the Survival would not be dependent of the groups. The Figure 1
shows two plots of survival curves depending, in left, to Tumor stage of patients grouped into 4 stages.
The curves show that patients with advanced stage (4) have less survive than early ones (stage 1). in
the same way, the right plot shows more survival patients without disease than patients with reccured
disease.

​Figure 1: Kaplan-Meier Curves, left, Tumor Stage; right, DFS STATUS

Cox proportional Hazards Model: A Cox model is a statistical technique for exploring relationship
between the survival of a patient and several explanatory variables. In survival analysis the Cox model
is preferred to a logistic model, since the latter one ignores survival times and censoring information.
The Figure 2 shows an example of Cox model of survival patients with 62 years old. If selected variable
has NA value (Censoring data), two curves were plotted (censored and variable).


Figure 2: left, Cox Proportional Hazard model; right, plot of residuals and/or test of proportional
hazards assumptions for Cox model.

Karim

Thanks for Tim Triche

How to install canceR-devel:
library("devtools")
install_github("canceR")