Skip to main content

Table 1 Comparison between actual and AI-predicted results of full and parsimonious models after 12, 24, 36 and 48 months

From: Prediction of cognitive conversion within the Alzheimer’s disease continuum using deep learning

 

Actual reassignment status

Predicted reassignment status

Full model

Parsimonious model

12-month

Overall, n (%)

459/538 (85.32)

485/538 (90.15)

485/538 (90.15)

Observation

312/380 (82.11)

334/380 (87.89)

334/380 (87.89)

Monotherapy-ChEI

93/101 (92.08)

97/101 (96.04)

95/101 (94.06)

Monotherapy-Memantine

49/50 (98.00)

48/50 (96.00)

49/50 (98.00)

Combined therapy

5/7 (71.43)

6/7 (85.71)

7/7 (100.00)

24-month

Overall, n (%)

384/482 (79.67)

435/482 (90.25)

430/482 (89.21)

Observation

304/386 (78.76)

341/386 (88.34)

339/386 (87.82)

Monotherapy-ChEI

54/62 (87.10)

61/62 (98.39)

59/62 (95.16)

Monotherapy-Memantine

25/31 (80.65)

31/31 (100.00)

31/31 (100.00)

Combined therapy

1/3 (33.33)

2/3 (66.67)

1/3 (33.33)

36-month

Overall, n (%)

202/268 (75.37)

219/268 (81.72)

221/268 (82.46)

Observation

157/198 (79.29)

170/198 (85.86)

169/198 (85.35)

Monotherapy-ChEI

36/50 (72.00)

39/50 (78.00)

38/50 (76.00)

Monotherapy-Memantine

7/17 (41.18)

10/17 (58.82)

12/17 (70.59)

Combined therapy

2/3 (66.67)

0/3 (0.00)

2/3 (66.67)

48-month

Overall, n (%)

214/280 (76.43)

232/280 (82.86)

222/280 (79.29)

 

Observation

183/233 (78.54)

199/233 (85.41)

187/233 (80.26)

 

Monotherapy-ChEI

26/35 (74.29)

26/35 (74.29)

27/35 (77.14)

 

Monotherapy-Memantine

5/10 (50.00)

7/10 (70.00)

7/10 (70.00)

 

Combined therapy

0/2 (0.00)

0/2 (0.00)

1/2 (50.00)