Deborah Cunningham
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Deborah Cunningham has a total of 46 co-authors. Here are the top co-authors by number of shared publications:
Amandeep Hujan
St George's University Hospitals NHS Foundation Trust, London, UK
2 shared publicationsAminata Sy
Imperial College London, London, UK
2 shared publicationsAra Darzi
Imperial College London, London, UK
2 shared publicationsCharlotte Purdy
AIMS Public Engagement Group, London, UK
2 shared publications- Christopher Joseph Kelly
Google Research, Mountain View, USA
2 shared publications - Daniel I. Golden
Google, Mountain View, USA
2 shared publications
Latest Publications
- Top Cited PapersOpen Access
- 10 March 2026
- Published by Springer Nature in Nature Cancer
- Vol. 7 (3) , 494-506
Diagnostic accuracy, fairness and clinical implementation of AI for breast cancer screening: results of multicenter retrospective and prospective technical feasibility studies
Artificial intelligence (AI) promises to enhance breast cancer screening. Here we evaluated Google’s mammography AI system (version 1.2) across two phases: a retrospective study using 115,973 mammograms from five National Health Service screening services with 39-month follow-up and prospective noninterventional feasibility deployment at 12 sites (9,266 cases). The primary endpoint was AI sensitivity and specificity versus first reader using a 5% noninferiority margin. The secondary endpoints were performance versus second or consensus readers and breast-level analyses. Retrospectively, AI achieved superior sensitivity (0.541 versus 0.437 for first reader, P < 0.001) and noninferior specificity (0.943 versus 0.952, P < 0.001). Cancer detection rate increased from 7.54 to 9.33 per 1,000 women, with AI detecting 25.0% of interval cancers. Performance was particularly strong for first screens (39.3% fewer recalls, 8.8% higher detection) and invasive cancers. No systematic demographic disparities were observed. Simulated second-reader replacement reduced reading time by 32% while increasing detection by 17.7%. Prospective deployment confirmed technical feasibility but revealed a distribution shift requiring threshold recalibration. Implementation requires adaptive calibration and continuous monitoring to ensure safety and equity.
- 10 March 2026
- Published by Springer Nature in Nature Cancer
- Vol. 7 (3) , 494-506
Top Cited PapersOpen AccessDiagnostic accuracy, fairness and clinical implementation of AI for breast cancer screening: results of multicenter retrospective and prospective technical feasibility studies
Artificial intelligence (AI) promises to enhance breast cancer screening. Here we evaluated Google’s mammography AI system (version 1.2) across two phases: a retrospective study using 115,973 mammograms from five National Health Service screening services with 39-month follow-up and prospective noninterventional feasibility deployment at 12 sites (9,266 cases). The primary endpoint was AI sensitivity and specificity versus first reader using a 5% noninferiority margin. The secondary endpoints were performance versus second or consensus readers and breast-level analyses. Retrospectively, AI achieved superior sensitivity (0.541 versus 0.437 for first reader, P < 0.001) and noninferior specificity (0.943 versus 0.952, P < 0.001). Cancer detection rate increased from 7.54 to 9.33 per 1,000 women, with AI detecting 25.0% of interval cancers. Performance was particularly strong for first screens (39.3% fewer recalls, 8.8% higher detection) and invasive cancers. No systematic demographic disparities were observed. Simulated second-reader replacement reduced reading time by 32% while increasing detection by 17.7%. Prospective deployment confirmed technical feasibility but revealed a distribution shift requiring threshold recalibration. Implementation requires adaptive calibration and continuous monitoring to ensure safety and equity.
- Top Cited PapersOpen Access
- 10 March 2026
- Published by Springer Nature in Nature Cancer
- Vol. 7 (3) , 507-521
Impact of using artificial intelligence as a second reader in breast screening including arbitration
The impact of incorporating artificial intelligence (AI) into a double-read breast-screening workflow, including arbitration, is unclear. This retrospective study included 50,000 representative women from two NHS breast-screening centers. All the women had long-term follow-up, allowing us to determine whether use of AI leads to earlier cancer detection. Cases requiring arbitration (8,732 cases) were read by 22 readers in a reader study, following their normal arbitration workflow. Overall, after arbitration, replacing the second reader with AI was noninferior (5% margin) to two human readers in terms of sensitivity and specificity ( P < 0.001) while offering a workload benefit. Arbitration improved the specificity of the AI arm by overruling cases incorrectly recalled by the AI tool; however, it also overruled the AI tool recall decision for some interval and next-round cancers. Further development of the AI tool alongside improvement in its explainability could lead to the earlier detection of cancers.
- 10 March 2026
- Published by Springer Nature in Nature Cancer
- Vol. 7 (3) , 507-521
Top Cited PapersOpen AccessImpact of using artificial intelligence as a second reader in breast screening including arbitration
The impact of incorporating artificial intelligence (AI) into a double-read breast-screening workflow, including arbitration, is unclear. This retrospective study included 50,000 representative women from two NHS breast-screening centers. All the women had long-term follow-up, allowing us to determine whether use of AI leads to earlier cancer detection. Cases requiring arbitration (8,732 cases) were read by 22 readers in a reader study, following their normal arbitration workflow. Overall, after arbitration, replacing the second reader with AI was noninferior (5% margin) to two human readers in terms of sensitivity and specificity ( P < 0.001) while offering a workload benefit. Arbitration improved the specificity of the AI arm by overruling cases incorrectly recalled by the AI tool; however, it also overruled the AI tool recall decision for some interval and next-round cancers. Further development of the AI tool alongside improvement in its explainability could lead to the earlier detection of cancers.
- Open Access
- 14 February 2025
- Published by MDPI AG in Behavioral Sciences
- Vol. 15 (2) , 209
A Population Survey on Barriers and Facilitators to Breast Cancer Screening Based on the Theoretical Domains Framework
Background: In the UK, breast cancer affects 1 in 8 women, accounting for 31% of cancers and 15% of cancer-related deaths in women. In 2023, London’s breast screening coverage was 56%, falling below the NHS target of 70%. This survey assesses psychological and behavioural factors impacting screening attendance. Methods: Using the Theoretical Domains Framework, an online survey including 15 behavioural factors was distributed through a market research company to women eligible for breast screening (aged 47–73) across London and Southeast and West England. Logistic regression was used to predict the impact of behavioural variables on history of attendance and intention to attend future screening opportunities. Results: Of the 922 respondents who returned the full survey, 88.6% intended to attend future screenings, and 88.1% reported previously attending screenings regularly. ‘Behavioural regulation’ had the strongest influence on past attendance (OR = 1.92, p < 0.001) and future intent (OR = 1.56, p = 0.003). Not intending to attend was linked to emotional consequences (OR = 0.68, p = 0.003) and environmental barriers (OR = 0.66, p < 0.001), where OR—Odds Ratio. Conclusions: This survey identifies behavioural factors influencing breast screening participation and screening intention, providing insights that may help design interventions to increase attendance rates.
- 14 February 2025
- Published by MDPI AG in Behavioral Sciences
- Vol. 15 (2) , 209
Open AccessA Population Survey on Barriers and Facilitators to Breast Cancer Screening Based on the Theoretical Domains Framework
Background: In the UK, breast cancer affects 1 in 8 women, accounting for 31% of cancers and 15% of cancer-related deaths in women. In 2023, London’s breast screening coverage was 56%, falling below the NHS target of 70%. This survey assesses psychological and behavioural factors impacting screening attendance. Methods: Using the Theoretical Domains Framework, an online survey including 15 behavioural factors was distributed through a market research company to women eligible for breast screening (aged 47–73) across London and Southeast and West England. Logistic regression was used to predict the impact of behavioural variables on history of attendance and intention to attend future screening opportunities. Results: Of the 922 respondents who returned the full survey, 88.6% intended to attend future screenings, and 88.1% reported previously attending screenings regularly. ‘Behavioural regulation’ had the strongest influence on past attendance (OR = 1.92, p < 0.001) and future intent (OR = 1.56, p = 0.003). Not intending to attend was linked to emotional consequences (OR = 0.68, p = 0.003) and environmental barriers (OR = 0.66, p < 0.001), where OR—Odds Ratio. Conclusions: This survey identifies behavioural factors influencing breast screening participation and screening intention, providing insights that may help design interventions to increase attendance rates.