Table 2 Proportions of Disease Groups Obtained from the Algorithm’s Top-1 Prediction by Continent and Countries

From: Planet-wide performance of a skin disease AI algorithm validated in Korea

Location

Total Queries

Benigns

Allergies

Infections

Malignancies

Estimated FP

Korea, Republic of

382,602

249,395 (65.2%, CI 65.0–65.3%)

27,069 (7.1%, CI 7.0–7.2%)

27,570 (7.2%, CI 7.1–7.3%)

7957 (2.1%, CI 2.0–2.1%)

3.1% (CI 3.0–3.2%)

Italy

151,574

79,350 (52.4%, CI 52.1–52.6%)

16,232 (10.7%, CI 10.6–10.9%)

15,258 (10.1%, CI 9.9–10.2%)

3550 (2.3%, CI 2.3–2.4%)

4.3% (CI 4.1–4.4%)

United States

105,233

40,957 (38.9%, CI 38.6–39.2%)

14,888 (14.1%, CI 13.9–14.4%)

13,252 (12.6%, CI 12.4–12.8%)

2953 (2.8%, CI 2.7–2.9%)

6.7% (CI 6.5–7.0%)

Germany

100,285

51,000 (50.9%, CI 50.5–51.2%)

12,027 (12.0%, CI 11.8–12.2%)

10,413 (10.4%, CI 10.2–10.6%)

1935 (1.9%, CI 1.8–2.0%)

3.7% (CI 3.5–3.8%)

France

68,408

30,430 (44.5%, CI 44.1–44.9%)

10,399 (15.2%, CI 14.9–15.5%)

8,538 (12.5%, CI 12.2–12.7%)

1,393 (2.0%, CI 1.9–2.1%)

4.4% (CI 4.2–4.6%)

Russian Federation

64,997

17,925 (27.6%, CI 27.2–27.9%)

13,723 (21.1%, CI 20.8–21.4%)

10,653 (16.4%, CI 16.1–16.7%)

619 (1.0%, CI 0.9–1.0%)

3.3% (CI 3.1–3.6%)

Spain

56,347

27,667 (49.1%, CI 48.7–49.5%)

6210 (11.0%, CI 10.8–11.3%)

5383 (9.6%, CI 9.3–9.8%)

1142 (2.0%, CI 1.9–2.1%)

4.0% (CI 3.7–4.2%)

Brazil

54,974

18,305 (33.3%, CI 32.9–33.7%)

8063 (14.7%, CI 14.4–15.0%)

7642 (13.9%, CI 13.6–14.2%)

841 (1.5%, CI 1.4–1.6%)

4.4% (CI 4.1–4.7%)

United Kingdom

54,028

23,692 (43.9%, CI 43.4–44.3%)

6939 (12.8%, CI 12.6–13.1%)

6127 (11.3%, CI 11.1–11.6%)

1479 (2.7%, CI 2.6–2.9%)

5.9% (CI 5.6–6.2%)

Hungary

47,937

21,224 (44.3%, CI 43.8–44.7%)

7418 (15.5%, CI 15.2–15.8%)

6168 (12.9%, CI 12.6–13.2%)

759 (1.6%, CI 1.5–1.7%)

3.5% (CI 3.2–3.7%)

Europe

810,181

376,874 (46.5%, CI 46.4–46.6%)

108,783 (13.4%, CI 13.4–13.5%)

93,324 (11.5%, CI 11.4–11.6%)

15,073 (1.9%, CI 1.8–1.9%)

3.8% (CI 3.8–3.9%)

Asia

533,106

295,687 (55.5%, CI 55.3–55.6%)

54,081 (10.1%, CI 10.1–10.2%)

53,415 (10.0%, CI 9.9–10.1%)

9,714 (1.8%, CI 1.8–1.9%)

3.2% (CI 3.1–3.2%)

North America

170,948

65,039 (38.0%, CI 37.8–38.3%)

24,904 (14.6%, CI 14.4–14.7%)

21,972 (12.9%, CI 12.7–13.0%)

4466 (2.6%, CI 2.5–2.7%)

6.4% (CI 6.2–6.6%)

South America

96,373

31,100 (32.3%, CI 32.0–32.6%)

15,339 (15.9%, CI 15.7–16.1%)

14,274 (14.8%, CI 14.6–15.0%)

1,344 (1.4%, CI 1.3–1.5%)

4.1% (CI 3.9–4.4%)

Africa

48,483

10,858 (22.4%, CI 22.0–22.8%)

7485 (15.4%, CI 15.1–15.8%)

8284 (17.1%, CI 16.8–17.4%)

460 (0.9%, CI 0.9–1.0%)

4.1% (CI 3.7–4.4%)

Oceania

31,941

14,943 (46.8%, CI 46.2–47.3%)

3611 (11.3%, CI 11.0–11.7%)

2739 (8.6%, CI 8.3–8.9%)

549 (1.7%, CI 1.6–1.9%)

3.5% (CI 3.3–3.8%)

Total

1,691,032

794,501 (47.0%, CI 46.9–47.1%)

214,203 (12.7%, CI 12.6–12.7%)

194,008 (11.5%, CI 11.4–11.5%)

31,606 (1.9%, CI 1.8–1.9%)

3.8% (CI 3.8–3.9%)

  1. A 95% confidence interval (CI) was reported using the binomial distribution method. This table shows statistics for the 10 countries with the highest usage and 6 continents. The proportions of benign tumors, malignant tumors, allergic disorders, and infectious disorders among the diagnoses provided by the algorithm were calculated. The estimated false positive rate (Estimated FP) was calculated using the formula: False Positive Rate (FP) = (cases of malignant tumors)/(cases of malignant tumors + cases of benign tumors).
  2. The results calculated using the Top-3 predictions are shown in Supplementary Table 4.