TimelyScholarFree Tools › Kaplan-Meier Curve Digitizer
Extract survival data from a published Kaplan-Meier figure. Upload the image, calibrate the axes, click along each curve and export clean coordinates, ready for individual participant data reconstruction.
FreeNo account neededYour data stays in your browser
The CSV files (time, survival) are the coordinate input that reconstruction methods need. The Guyot algorithm, implemented in the IPDfromKM R package, uses them together with the numbers at risk to estimate individual participant data. The IPDfromKM authors suggest collecting 80 to 100 points per curve, including the points where survival drops.
Enter two known values on each axis and click them on the image.
Include the top and bottom of every step; add a curve for each arm.
Download a time-survival CSV per curve and copy R code for IPDfromKM, adding the numbers at risk if the figure reports them.
Many trials report survival outcomes only as Kaplan-Meier figures. Meta-analyses of time-to-event outcomes need hazard ratios or individual participant data, which are often not reported in full.
Guyot and colleagues developed an algorithm that uses the digitized curve coordinates together with the numbers at risk to reconstruct approximate individual participant data. The IPDfromKM R package implements this method; its authors suggest collecting 80 to 100 points per curve, including the points where survival drops.
Either. Enter the y-axis values as they appear on the figure (for example 0 and 1, or 0 and 100); the R code sets maxy to match.
Reconstruction is less accurate. Leave the numbers at risk blank and enter the total number of participants in the arm (totalpts) in the R code.