Estimating diagnostic uncertainty in artificial intelligence assisted pathology using conformal prediction
Open accessAbstract
Unreliable predictions can occur when an artificial intelligence (AI) system is presented with data it has not been exposed to during training. We demonstrate the use of conformal prediction to detect unreliable predictions, using histopathological diagnosis and grading of prostate biopsies as example. We digitized 7788 prostate biopsies from 1192 men in the STHLM3 diagnostic study, used for training, and 3059 biopsies from 676 men used for testing. With conformal prediction, 1 in 794 (0.1%) predictions is incorrect for cancer diagnosis (compared to 14 errors [2%] without conformal prediction) while 175 (22%) of the predictions are flagged as unreliable when the AI-system is presented with new data from the same lab and scanner that it was trained on. Conformal prediction could with small samples (N = 49 for external scanner, N = 10 for external lab and scanner, and N = 12 for external lab, scanner and pathology assessment) detect systematic differences in external data leading to worse predictive performance. The AI-system with conformal prediction commits 3 (2%) errors for cancer detection in cases of atypical prostate tissue compared to 44 (25%) without conformal prediction, while the system flags 143 (80%) unreliable predictions. We conclude that conformal prediction can increase patient safety of AI-systems.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Confidence and uncertainty estimation are recognized as critical components of AI systems, allowing algorithms to detect cases that may require expert review.
“Confidence and uncertainty estimation are also recognized as critical components of AI systems, allowing algorithms to detect cases that may require expert review [34,74].”
From Artificial Intelligence in Dermatopathology: An Update and Review of the Current Literature (2026) · cites this paper as [74]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · see citing works
XGBoost: A Scalable Tree Boosting System
Chen, T., & Guestrin, C. (2016) · Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2939672.2939785 · Cited by 54,569 · see citing works
Resistance to Medical Artificial Intelligence
Longoni, C., Bonezzi, A., & Morewedge, C. K. (2019) · Journal of Consumer Research · doi:10.1093/jcr/ucz013 · Cited by 2,015 · see citing works
Collins, G. S., Reitsma, J. B., Altman, D. G., & Moons, K. (2015) · BMC Medicine · doi:10.1186/s12916-014-0241-z · Cited by 1,927 · see citing works
Algorithmic Learning in a Random World
(2005) · doi:10.1007/b106715 · Cited by 1,119 · see citing works
Bulten, W., Pinckaers, H., van Boven, H., Vink, R., de Bel, T., van Ginneken, B., van der Laak, J., Hulsbergen-van de Kaa, C., & Litjens, G. (2020) · The Lancet Oncology · doi:10.1016/s1470-2045(19)30739-9 · Cited by 705 · see citing works
Deep Learning—A Technology With the Potential to Transform Health Care
Hinton, G. (2018) · JAMA · doi:10.1001/jama.2018.11100 · Cited by 667 · see citing works
Adapting to Artificial Intelligence: Radiologists and Pathologists as Information Specialists
Jha, S., & Topol, E. J. (2016) · JAMA · doi:10.1001/jama.2016.17438 · Cited by 622 · see citing works
Ström, P., Kartasalo, K., Olsson, H., Solorzano, L., Delahunt, B., Berney, D. M., Bostwick, D. G., Evans, A. J., Grignon, D. J., Humphrey, P. A., et al. (2020) · The Lancet Oncology · doi:10.1016/s1470-2045(19)30738-7 · Cited by 620 · see citing works
Self-citationArtificial intelligence for diagnosis and Gleason grading of prostate cancer: the PANDA challenge
Bulten, W., Kartasalo, K., Chen, P.-H. C., Ström, P., Pinckaers, H., Nagpal, K., Cai, Y., Steiner, D. F., van Boven, H., Vink, R., et al. (2022) · Nature Medicine · doi:10.1038/s41591-021-01620-2 · Cited by 559 · see citing works
Self-citationLeveraging uncertainty information from deep neural networks for disease detection
Leibig, C., Allken, V., Ayhan, M. S., Berens, P., & Wahl, S. (2017) · Scientific Reports · doi:10.1038/s41598-017-17876-z · Cited by 510 · see citing works
The “inconvenient truth” about AI in healthcare
Panch, T., Mattie, H., & Celi, L. A. (2019) · npj Digital Medicine · doi:10.1038/s41746-019-0155-4 · Cited by 473 · see citing works
Grönberg, H., Adolfsson, J., Aly, M., Nordström, T., Wiklund, P. N., Brandberg, Y., Thompson, J., Wiklund, F. E., Lindberg, J., Clements, M., et al. (2015) · The Lancet Oncology · doi:10.1016/s1470-2045(15)00361-7 · Cited by 410 · see citing works
Self-citationArtificial intelligence for diagnosis and Gleason grading of prostate cancer: the PANDA challenge
Wouter, B., Kimmo, K., Cameron, C. P.-H., Peter, S., Hans, P., Kunal, N., Yuannan, C., Steiner, D. F., van Boven, H., Robert, V., et al. (2022) · UTUPub (University of Turku) · Cited by 375 · see citing works
Self-citationNagpal, K., Foote, D., Tan, F. E., Liu, Y., Chen, P.-H. C., Steiner, D. F., Manoj, N. S., Olson, N. H., Smith, J. L., Mohtashamian, A., et al. (2020) · JAMA Oncology · doi:10.1001/jamaoncol.2020.2485 · Cited by 223 · see citing works
Inductive Conformal Prediction: Theory and Application to Neural Networks
Papadopoulos, H. (2008) · InTech eBooks · doi:10.5772/6078 · Cited by 195 · see citing works
Standardization of Gleason grading among 337 European pathologists
Egevad, L., Ahmad, A. S., Ferran Algaba, B., Berney, D. M., Boccon-Gibod, L. A., Compérat, É., Evans, A. J., Griffiths, D. J., Grobholz, R., Kristiansen, G., et al. (2012) · Histopathology · doi:10.1111/his.12008 · Cited by 187 · see citing works
Self-citationHedging Predictions in Machine Learning
Gammerman, A., & Vovk, V. (2007) · The Computer Journal · doi:10.1093/comjnl/bxl065 · Cited by 135 · see citing works
Predicting With Confidence: Using Conformal Prediction in Drug Discovery
Alvarsson, J., Arvidsson McShane, S., Norinder, U., & Spjuth, O. (2021) · Journal of Pharmaceutical Sciences · doi:10.1016/j.xphs.2020.09.055 · Cited by 116 · see citing works
Self-citationKartasalo, K., Bulten, W., Delahunt, B., Chen, P.-H. C., Pinckaers, H., Olsson, H., Ji, X., Mulliqi, N., Samaratunga, H., Tsuzuki, T., et al. (2021) · European Urology Focus · doi:10.1016/j.euf.2021.07.002 · Cited by 49 · see citing works
Self-citationUtility of Pathology Imagebase for standardisation of prostate cancer grading
Egevad, L., Delahunt, B., Berney, D. M., et al. (2018) · Histopathology · doi:10.1111/his.13471 · Cited by 48 · see citing works
Self-citationWieslander, H., Harrison, P. J., Skogberg, G., Jackson, S. G., Fridén, M., Karlsson, J., Spjuth, O., & Wählby, C. (2020) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2020.2996300 · Cited by 36 · see citing works
Self-citation(2018) · Nature · doi:10.1038/d41586-018-03067-x · Cited by 33 · see citing works
Histopathologic False-positive Diagnoses of Prostate Cancer in the Age of Immunohistochemistry
Beltrán, L., Ahmad, A. S., Sandu, H., Kudahetti, S. C., Soosay, G. N., Møller, H., Cuzick, J. M., & Berney, D. M. (2018) · The American Journal of Surgical Pathology · doi:10.1097/pas.0000000000001202 · Cited by 17 · see citing works
Error rates in reporting prostatic core biopsies
OXLEY, J. D., & Sen, C. (2011) · Histopathology · doi:10.1111/j.1365-2559.2011.03794.x · Cited by 13 · see citing works
Flagged first, then by citations
Research works printed but not linked: 4
- [17]Fedorova, V., Gammerman, A., Nouretdinov, I. & Vovk, V. Plug-in martingales for testing exchangeability on-line. in Proceedings of the 29th International Conference on Machine Learning, ICML 20122 (2012).
- [19]Gal Y Ghahramani Z Dropout as a Bayesian approximation: representing model uncertainty in deep learning 33rd Int. Conf. Mach. Learn. 2016 3 1651 1660
- [20]Lakshminarayanan B Pritzel A Blundell C Simple and scalable predictive uncertainty estimation using deep ensembles Adv. Neural Inf. Process. Syst. 2017 2017 6403 6414
- [26]Szegedy C Vanhoucke V Ioffe S Shlens J Wojna Z Rethinking the inception architecture for computer vision Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit. 2016 2016 2818 2826
Cited by
Sorted by citations
Multimodal Artificial Intelligence for Precision Critical Care: A Scoping Review
Yang, M., Shi, N., Chen, H., Po, E. Y. P., Ng, P. Y., Lai, P. C. K., Wu, Y.-C., & Fong, D. Y. T. (2026) · Health Data Science · doi:10.34133/hds.0356 · Cited by 3 · see citing works
“Therefore, we recommend increased reporting of calibration metrics and consider using conformal prediction methods to enhance model reliability [ 122 , 123 ].”
Artificial Intelligence in Dermatopathology: An Update and Review of the Current Literature
Abu-Dayeh, A., Cazzato, G., & Giubellino, A. (2026) · Diagnostics · doi:10.3390/diagnostics16111702 · Cited by 1 · see citing works
“Confidence and uncertainty estimation are also recognized as critical components of AI systems, allowing algorithms to detect cases that may require expert review [34,74].”
- Authors
- Olsson, H., Kartasalo, K., Mulliqi, N., Capuccini, M., Ruusuvuori, P., Samaratunga, H., Delahunt, B., Lindskog, C., Janssen, E. A. M., Blilie, A., Panel, I. P. I. E., Egevad, L., Spjuth, O., & Eklund, M.
- Affiliations
- Karolinska Institutet (SE)
- Uppsala University (SE)
- University of Turku (FI)
- The University of Queensland (AU)
- University of Otago (NZ)
- Stavanger University Hospital (NO)
- Corresponding author
- Olsson, Henrik
- ORCID
- Venue
- Nature Communications
- Published
- 2022-12-15
- Volume
- 13
- Issue
- 1
- Pages
- 7761-7761
- ISSN
- 2041-1723 (ISSN-L)
- PMC
- PMC9755280
- OpenAlex
- W4311506764
- Cited by
- 111
- References
- 25
- Citations by year
- 2026: 39 · 2025: 33 · 2024: 27 · 2023: 12
- FWCI
- 5.92
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- Nature Communications (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- KTH Publication Database DiVA (KTH Royal Institute of Technology) (submittedVersion, other-oa)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Tampere University Institutional Repository (Tampere University) (submittedVersion, cc-by)
- NORA - Norwegian Open Research Archives (submittedVersion, cc-by)
- Funders
- David och Astrid Hageléns Stiftelse
- Orionin Tutkimussäätiö
- KAUTE-Säätiö
- Oskar Huttunen Foundation
- Vetenskapsrådet (2020-00692)
- Cancerfonden (21 1715 Pj)
- Magnus Bergvalls Stiftelse
- Region Stockholm, Svenska Druidorden, Åke Wibergs Stiftelse, Swedish e-Science Research Center (SeRC), the Swedish Prostate Cancer Foundation
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiomics and Machine Learning in Medical Imaging
- Keywords
- conformal prediction, diagnostic uncertainty, artificial intelligence in pathology, prostate cancer diagnosis, histopathological grading, prostate biopsy, external validation, patient safety
- MeSH terms
- Artificial Intelligence, Biopsy, Humans, Male, Neoplasms, Prostate, Uncertainty
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
