Table 1 Results comparing deep learning model with expert Surgeons.

From: Expert surgeons and deep learning models can predict the outcome of surgical hemorrhage from 1 min of video

 

Accuracy (SN %, SP %)

RMSE (R2)

M-S agreement:a success/failure

M-S agreement:b blood loss

Ground truth

11 success

9 failures

–

–

Avg blood loss: 568 (range:20–1640)

Model

17/20 (85%)

(100, 66)

295 (0.74)

–

–

Expert cohort

55/80 (68.75)

(79, 56)

351 (0.70)

0.43‡

0.73c

Surgeon 1

13/20 (65%)

(73, 55)

306 (0.73)

0.34

0.74

Surgeon 2

14/20 (65%)

(81, 55)

335 (0.66)

0.43

0.66

Surgeon 3

14/20 (65%)

(81, 55)

423 (0.65)

0.43

0.65

Surgeon 4

14/20 (65%)

(81, 55)

329 (0.74)

0.43

0.72

  1. SN: sensitivity; SP: specificity; M-S: model-surgeon.
  2. aKappa coefficient.
  3. bInter-class coefficient.
  4. cInter-Surgeon Agreement: Success/Failure = 0.95, Blood-Loss: 0.72.
Back to article page