Table 2

Performance of DL0 considering a fivefold cross-validation

Predicted 0sPredicted 1s
True 0s63918
True 1s5645

  • Sensitivity 44.5%; specificity 97%.

  • Note that it performs very well with rejecting unsuitable patients accurately, but it does not perform well with predicting suitable participants (the true 1s).

  • DL, deep learning paradigm.