Can incorrect artificial intelligence (AI) results impact radiologists, and if so, what can we do about it? A multi-reader pilot study of lung cancer detection with chest radiography
© 2023. The Author(s)..
OBJECTIVE: To examine whether incorrect AI results impact radiologist performance, and if so, whether human factors can be optimized to reduce error.
METHODS: Multi-reader design, 6 radiologists interpreted 90 identical chest radiographs (follow-up CT needed: yes/no) on four occasions (09/20-01/22). No AI result was provided for session 1. Sham AI results were provided for sessions 2-4, and AI for 12 cases were manipulated to be incorrect (8 false positives (FP), 4 false negatives (FN)) (0.87 ROC-AUC). In the Delete AI (No Box) condition, radiologists were told AI results would not be saved for the evaluation. In Keep AI (No Box) and Keep AI (Box), radiologists were told results would be saved. In Keep AI (Box), the ostensible AI program visually outlined the region of suspicion. AI results were constant between conditions.
RESULTS: Relative to the No AI condition (FN = 2.7%, FP = 51.4%), FN and FPs were higher in the Keep AI (No Box) (FN = 33.0%, FP = 86.0%), Delete AI (No Box) (FN = 26.7%, FP = 80.5%), and Keep AI (Box) (FN = to 20.7%, FP = 80.5%) conditions (all ps < 0.05). FNs were higher in the Keep AI (No Box) condition (33.0%) than in the Keep AI (Box) condition (20.7%) (p = 0.04). FPs were higher in the Keep AI (No Box) (86.0%) condition than in the Delete AI (No Box) condition (80.5%) (p = 0.03).
CONCLUSION: Incorrect AI causes radiologists to make incorrect follow-up decisions when they were correct without AI. This effect is mitigated when radiologists believe AI will be deleted from the patient's file or a box is provided around the region of interest.
CLINICAL RELEVANCE STATEMENT: When AI is wrong, radiologists make more errors than they would have without AI. Based on human factors psychology, our manuscript provides evidence for two AI implementation strategies that reduce the deleterious effects of incorrect AI.
KEY POINTS: • When AI provided incorrect results, false negative and false positive rates among the radiologists increased. • False positives decreased when AI results were deleted, versus kept, in the patient's record. • False negatives and false positives decreased when AI visually outlined the region of suspicion.
Medienart: |
E-Artikel |
---|
Erscheinungsjahr: |
2023 |
---|---|
Erschienen: |
2023 |
Enthalten in: |
Zur Gesamtaufnahme - volume:33 |
---|---|
Enthalten in: |
European radiology - 33(2023), 11 vom: 01. Nov., Seite 8263-8269 |
Sprache: |
Englisch |
---|
Beteiligte Personen: |
Bernstein, Michael H [VerfasserIn] |
---|
Links: |
---|
Themen: |
Artificial intelligence |
---|
Anmerkungen: |
Date Completed 27.10.2023 Date Revised 09.12.2023 published: Print-Electronic Citation Status MEDLINE |
---|
doi: |
10.1007/s00330-023-09747-1 |
---|
funding: |
|
---|---|
Förderinstitution / Projekttitel: |
|
PPN (Katalog-ID): |
NLM357686950 |
---|
LEADER | 01000caa a22002652 4500 | ||
---|---|---|---|
001 | NLM357686950 | ||
003 | DE-627 | ||
005 | 20231227131308.0 | ||
007 | cr uuu---uuuuu | ||
008 | 231226s2023 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.1007/s00330-023-09747-1 |2 doi | |
028 | 5 | 2 | |a pubmed24n1225.xml |
035 | |a (DE-627)NLM357686950 | ||
035 | |a (NLM)37266657 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
100 | 1 | |a Bernstein, Michael H |e verfasserin |4 aut | |
245 | 1 | 0 | |a Can incorrect artificial intelligence (AI) results impact radiologists, and if so, what can we do about it? A multi-reader pilot study of lung cancer detection with chest radiography |
264 | 1 | |c 2023 | |
336 | |a Text |b txt |2 rdacontent | ||
337 | |a ƒaComputermedien |b c |2 rdamedia | ||
338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
500 | |a Date Completed 27.10.2023 | ||
500 | |a Date Revised 09.12.2023 | ||
500 | |a published: Print-Electronic | ||
500 | |a Citation Status MEDLINE | ||
520 | |a © 2023. The Author(s). | ||
520 | |a OBJECTIVE: To examine whether incorrect AI results impact radiologist performance, and if so, whether human factors can be optimized to reduce error | ||
520 | |a METHODS: Multi-reader design, 6 radiologists interpreted 90 identical chest radiographs (follow-up CT needed: yes/no) on four occasions (09/20-01/22). No AI result was provided for session 1. Sham AI results were provided for sessions 2-4, and AI for 12 cases were manipulated to be incorrect (8 false positives (FP), 4 false negatives (FN)) (0.87 ROC-AUC). In the Delete AI (No Box) condition, radiologists were told AI results would not be saved for the evaluation. In Keep AI (No Box) and Keep AI (Box), radiologists were told results would be saved. In Keep AI (Box), the ostensible AI program visually outlined the region of suspicion. AI results were constant between conditions | ||
520 | |a RESULTS: Relative to the No AI condition (FN = 2.7%, FP = 51.4%), FN and FPs were higher in the Keep AI (No Box) (FN = 33.0%, FP = 86.0%), Delete AI (No Box) (FN = 26.7%, FP = 80.5%), and Keep AI (Box) (FN = to 20.7%, FP = 80.5%) conditions (all ps < 0.05). FNs were higher in the Keep AI (No Box) condition (33.0%) than in the Keep AI (Box) condition (20.7%) (p = 0.04). FPs were higher in the Keep AI (No Box) (86.0%) condition than in the Delete AI (No Box) condition (80.5%) (p = 0.03) | ||
520 | |a CONCLUSION: Incorrect AI causes radiologists to make incorrect follow-up decisions when they were correct without AI. This effect is mitigated when radiologists believe AI will be deleted from the patient's file or a box is provided around the region of interest | ||
520 | |a CLINICAL RELEVANCE STATEMENT: When AI is wrong, radiologists make more errors than they would have without AI. Based on human factors psychology, our manuscript provides evidence for two AI implementation strategies that reduce the deleterious effects of incorrect AI | ||
520 | |a KEY POINTS: • When AI provided incorrect results, false negative and false positive rates among the radiologists increased. • False positives decreased when AI results were deleted, versus kept, in the patient's record. • False negatives and false positives decreased when AI visually outlined the region of suspicion | ||
650 | 4 | |a Journal Article | |
650 | 4 | |a Artificial intelligence | |
650 | 4 | |a Cognitive science | |
650 | 4 | |a Psychology | |
700 | 1 | |a Atalay, Michael K |e verfasserin |4 aut | |
700 | 1 | |a Dibble, Elizabeth H |e verfasserin |4 aut | |
700 | 1 | |a Maxwell, Aaron W P |e verfasserin |4 aut | |
700 | 1 | |a Karam, Adib R |e verfasserin |4 aut | |
700 | 1 | |a Agarwal, Saurabh |e verfasserin |4 aut | |
700 | 1 | |a Ward, Robert C |e verfasserin |4 aut | |
700 | 1 | |a Healey, Terrance T |e verfasserin |4 aut | |
700 | 1 | |a Baird, Grayson L |e verfasserin |4 aut | |
773 | 0 | 8 | |i Enthalten in |t European radiology |d 1991 |g 33(2023), 11 vom: 01. Nov., Seite 8263-8269 |w (DE-627)NLM087691310 |x 1432-1084 |7 nnns |
773 | 1 | 8 | |g volume:33 |g year:2023 |g number:11 |g day:01 |g month:11 |g pages:8263-8269 |
856 | 4 | 0 | |u http://dx.doi.org/10.1007/s00330-023-09747-1 |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a GBV_NLM | ||
951 | |a AR | ||
952 | |d 33 |j 2023 |e 11 |b 01 |c 11 |h 8263-8269 |