Modality Comparison

• modality is imaging methods e.g.,

  MRI, CT, PET,…

• Which modality is better to detect lesions in radiographs?

• Create FROC data or

  Convert from Jafroc formulation

• Fitting models

• Draw Curves

• Comparison of Modalities

For more details of Jafroc please visit the following Chakraborty page:

http://www.devchakraborty.com/.

Make a data

 dat <- list(

c = c(3,2,1),    #Confidence level (c.l.)
h = c(97,32,31), #Number of hits for each c.l.
f = c(1,14,74),  #Number of false alarms for each c.l.

NL= 259,       #Number of lesions
NI= 57,        #Number of images
 C= 3
)          #Number of confidence level

BayesianFROC::viewdata(dat)

.                     Confidence.Level   False.Positives   True.Positives
-------------------  -----------------  ----------------  ---------------
Obviouly present                     3                 1               97
Relatively obvious                   2                14               32
Subtle                               1                74               31

Fit the FROC model for the above data.

 fit <-   BayesianFROC::fit_Bayesian_FROC(dat, summary = F)

BayesianFROC:::summary_EAP_CI_srsc(fit)

From this, you can see the following:

• the Expected A Posterior (EAP) estimates
• 95% Credible Intervals

Note that this function BayesianFROC:::summary_EAP_CI_srsc() run automatically when you run BayesianFROC::fit_Bayesian_FROC().

By running the following, we get a estimates as a return value, FROC curves.

 fit <-   BayesianFROC::fit_Bayesian_FROC(dat)

So, Single reader and single modality case, it very simple.

From here, we show the case of single reader and single modality.

Next, we show the multiple readers and multiple modalities.

• Make a data of MRMC ———— create_dataset()
• Fitting ( comparison of modalities) — fit_Bayesian_FROC()

• Draw the FROC curve ———– DrawCurves_MRMC_pairwise()

  # Make a data
  data <- create_dataset()
  
  # Fitting
  fit <- fit_Bayesian_FROC(dat)
  
  #Draw curves for the 1st modality and 2nd reader
  DrawCurves(
  
    #Estimates
    fit,
  
    #Specify   modality IDs to draw curves
    modalityID =c(1,2),
  
    #Specify  Reader IDs  to draw curves
    readerID   =c(2,3,4)