Nicolas Wagneur
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Nicolas Wagneur has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Denise Silber
Basil Strategies, Paris, France
2 shared publications - Fabrice Denis
Institut Inter-régional de Cancérologie Jean Bernard, Le Mans, France
2 shared publications - Florent Le Borgne
Data Factory & Analytics Department, Institut de Cancérologie de l'Ouest, Nantes-Angers, France
2 shared publications - François Bocquet
Data Factory & Analytics Department, Institut de Cancérologie de l'Ouest, Nantes-Angers, France
2 shared publications - Jean-David Zeitoun
Proctologie Médico-Interventionnelle, groupe hospitalier Diaconesses Croix Saint-Simon, site Avron, 125 rue d’Avron, F-75020, Paris
2 shared publications Mario Campone
Oncology Department, Institut de Cancérologie de l'Ouest, Nantes-Angers, France
2 shared publications
Latest Publications
- Open Access
- 1 January 2026
- Published by American Society of Clinical Oncology (ASCO) in JCO Clinical Cancer Informatics
- Vol. 10 (1) , e2500133
Development and Assessment of a Pipeline for Extracting Structured Data From Free-Text Medical Reports Using a Large Language Model
PURPOSE: Medical free texts such as pathology reports contain valuable clinical data but are challenging to structure at scale. Traditional natural language processing approaches require extensive annotated data and training. We investigate the use of large language model (LLM) like Mistral to automatically extract three breast cancer (BC) biomarkers from pathology reports. MATERIALS AND METHODS: We developed and evaluated a pipeline combining Mistral Large LLM and a postprocessing phase. The pipeline's performance was assessed both at document and patient levels. For evaluation, two data sets were used: a data set of 1,152 pathology reports associated with 150 patients with BC focused solely on biomarker values and a gold standard database containing 101 patients with metastatic BC, enriched with detailed patient and tumor characteristics and double-blind validated by clinical research assistants. We also explored the pipeline's performance according to the use of a confidence prompt (CP), a chain of thought (CoT), and few-shot examples. RESULTS: Our extraction pipeline achieved F1 scores of more than 95% and both recall and precision of more than 94% for each biomarker of interest (ie, estrogen receptor, progesterone receptor and human epidermal growth factor receptor 2 status and score) at the document level. At the patient level, the F1 score decreased between 87% and 90% with a greater drop in recall (ranging between 83% and 87%) compared with precision, which remained >90%. The results were similar whether the pipeline included a CP, CoT, or few-shot examples. CONCLUSION: Our study provides strong evidence of the potential of LLMs like Mistral Large for extracting structured BC biomarker data from pathology reports and the potential of such methods for broader digital transformation of health care documents.
- 1 January 2026
- Published by American Society of Clinical Oncology (ASCO) in JCO Clinical Cancer Informatics
- Vol. 10 (1) , e2500133
Open AccessDevelopment and Assessment of a Pipeline for Extracting Structured Data From Free-Text Medical Reports Using a Large Language Model
PURPOSE: Medical free texts such as pathology reports contain valuable clinical data but are challenging to structure at scale. Traditional natural language processing approaches require extensive annotated data and training. We investigate the use of large language model (LLM) like Mistral to automatically extract three breast cancer (BC) biomarkers from pathology reports. MATERIALS AND METHODS: We developed and evaluated a pipeline combining Mistral Large LLM and a postprocessing phase. The pipeline's performance was assessed both at document and patient levels. For evaluation, two data sets were used: a data set of 1,152 pathology reports associated with 150 patients with BC focused solely on biomarker values and a gold standard database containing 101 patients with metastatic BC, enriched with detailed patient and tumor characteristics and double-blind validated by clinical research assistants. We also explored the pipeline's performance according to the use of a confidence prompt (CP), a chain of thought (CoT), and few-shot examples. RESULTS: Our extraction pipeline achieved F1 scores of more than 95% and both recall and precision of more than 94% for each biomarker of interest (ie, estrogen receptor, progesterone receptor and human epidermal growth factor receptor 2 status and score) at the document level. At the patient level, the F1 score decreased between 87% and 90% with a greater drop in recall (ranging between 83% and 87%) compared with precision, which remained >90%. The results were similar whether the pipeline included a CP, CoT, or few-shot examples. CONCLUSION: Our study provides strong evidence of the potential of LLMs like Mistral Large for extracting structured BC biomarker data from pathology reports and the potential of such methods for broader digital transformation of health care documents.
- 1 September 2025
- Published by American Society of Clinical Oncology (ASCO) in JCO Clinical Cancer Informatics
- Vol. 9 (9) , e2500130
Hybrid ReGex and Natural Language Inference Model as a Zero-Shot Classifier for Extracting Data From Medical Reports
PURPOSE: This study presents a new method based on regular expressions (ReGex) and artificial intelligence for extracting relevant medical data from clinical reports. This hybrid approach is designed to address the limitations of each technique. The pipeline is evaluated for its effectiveness in extracting key clinical information from prostate cancer medical reports. METHODS: We developed a hybrid pipeline that combines ReGex for initial data extraction with a Natural Language Inference model for classification. This approach was retrospectively applied to 1,000 reports randomly selected among all consultation reports of patients with prostate cancer treated at the institute, focusing on identifying key clinical information such as rectal bleeding, dysuria, pollakiuria, and hematuria. The model's performance was evaluated using precision, recall, accuracy, F1-score, and Cohen's kappa coefficient. RESULTS: The pipeline demonstrated high performance, with precision scores ranging from 0.778 to 0.954 and recall consistently high at 0.920 to 1.00. F1-scores indicated balanced accuracy across symptoms, and Cohen's kappa values (0.871 to 0.951) reflected strong agreement with physician-labeled data. CONCLUSION: The proposed pipeline is both efficient and fast while being computationally lightweight. It achieves high accuracy in extracting medical data from clinical reports, making it an effective and practical tool for clinical research and health care applications.
- 1 September 2025
- Published by American Society of Clinical Oncology (ASCO) in JCO Clinical Cancer Informatics
- Vol. 9 (9) , e2500130
Hybrid ReGex and Natural Language Inference Model as a Zero-Shot Classifier for Extracting Data From Medical Reports
PURPOSE: This study presents a new method based on regular expressions (ReGex) and artificial intelligence for extracting relevant medical data from clinical reports. This hybrid approach is designed to address the limitations of each technique. The pipeline is evaluated for its effectiveness in extracting key clinical information from prostate cancer medical reports. METHODS: We developed a hybrid pipeline that combines ReGex for initial data extraction with a Natural Language Inference model for classification. This approach was retrospectively applied to 1,000 reports randomly selected among all consultation reports of patients with prostate cancer treated at the institute, focusing on identifying key clinical information such as rectal bleeding, dysuria, pollakiuria, and hematuria. The model's performance was evaluated using precision, recall, accuracy, F1-score, and Cohen's kappa coefficient. RESULTS: The pipeline demonstrated high performance, with precision scores ranging from 0.778 to 0.954 and recall consistently high at 0.920 to 1.00. F1-scores indicated balanced accuracy across symptoms, and Cohen's kappa values (0.871 to 0.951) reflected strong agreement with physician-labeled data. CONCLUSION: The proposed pipeline is both efficient and fast while being computationally lightweight. It achieves high accuracy in extracting medical data from clinical reports, making it an effective and practical tool for clinical research and health care applications.
- Open Access
- 5 July 2022
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 24 (7) , e39590
Assessing a New Prescreening Score for the Simplified Evaluation of the Clinical Quality and Relevance of eHealth Apps: Instrument Validation Study
Background: In 2020, more than 250 eHealth solutions were added to app stores each day, or 90,000 in the year; however, the vast majority of these solutions have not undergone clinical validation, their quality is unknown, and the user does not know if they are effective and safe. We sought to develop a simple prescreening scoring method that would assess the quality and clinical relevance of each app. We designed this tool with 3 health care stakeholder groups in mind: eHealth solution designers seeking to evaluate a potential competitor or their own tool, investors considering a fundraising candidate, and a hospital clinician or IT department wishing to evaluate a current or potential eHealth solution. Objective: We built and tested a novel prescreening scoring tool (the Medical Digital Solution scoring tool). The tool, which consists of 26 questions that enable the quick assessment and comparison of the clinical relevance and quality of eHealth apps, was tested on 68 eHealth solutions. Methods: The Medical Digital Solution scoring tool is based on the 2021 evaluation criteria of the French National Health Authority, the 2022 European Society of Medical Oncology recommendations, and other provided scores. We built the scoring tool with patient association and eHealth experts and submitted it to eHealth app creators, who evaluated their apps via the web-based form in January 2022. After completing the evaluation criteria, their apps obtained an overall score and 4 categories of subscores. These criteria evaluated the type of solution and domain, the solution’s targeted population size, the level of clinical assessment, and information about the provider. Results: In total, 68 eHealth solutions were evaluated with the scoring tool. Oncology apps (22%, 20/90) and general health solutions (23%, 21/90) were the most represented. Of the 68 apps, 32 (47%) were involved in remote monitoring by health professionals. Regarding clinical outcomes, 5% (9/169) of the apps assessed overall survival. Randomized studies had been conducted for 21% (23/110) of the apps to assess their benefit. Of the 68 providers, 38 (56%) declared the objective of obtaining reimbursement, and 7 (18%) out of the 38 solutions seeking reimbursement were assessed as having a high probability of reimbursement. The median global score was 11.2 (range 4.7-17.4) out of 20 and the distribution of the scores followed a normal distribution pattern (Shapiro-Wilk test: P=.33). Conclusions: This multidomain prescreening scoring tool is simple, fast, and can be deployed on a large scale to initiate an assessment of the clinical relevance and quality of a clinical eHealth app. This simple tool can help a decision-maker determine which aspects of the app require further analysis and improvement.
- 5 July 2022
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 24 (7) , e39590
Open AccessAssessing a New Prescreening Score for the Simplified Evaluation of the Clinical Quality and Relevance of eHealth Apps: Instrument Validation Study
Background: In 2020, more than 250 eHealth solutions were added to app stores each day, or 90,000 in the year; however, the vast majority of these solutions have not undergone clinical validation, their quality is unknown, and the user does not know if they are effective and safe. We sought to develop a simple prescreening scoring method that would assess the quality and clinical relevance of each app. We designed this tool with 3 health care stakeholder groups in mind: eHealth solution designers seeking to evaluate a potential competitor or their own tool, investors considering a fundraising candidate, and a hospital clinician or IT department wishing to evaluate a current or potential eHealth solution. Objective: We built and tested a novel prescreening scoring tool (the Medical Digital Solution scoring tool). The tool, which consists of 26 questions that enable the quick assessment and comparison of the clinical relevance and quality of eHealth apps, was tested on 68 eHealth solutions. Methods: The Medical Digital Solution scoring tool is based on the 2021 evaluation criteria of the French National Health Authority, the 2022 European Society of Medical Oncology recommendations, and other provided scores. We built the scoring tool with patient association and eHealth experts and submitted it to eHealth app creators, who evaluated their apps via the web-based form in January 2022. After completing the evaluation criteria, their apps obtained an overall score and 4 categories of subscores. These criteria evaluated the type of solution and domain, the solution’s targeted population size, the level of clinical assessment, and information about the provider. Results: In total, 68 eHealth solutions were evaluated with the scoring tool. Oncology apps (22%, 20/90) and general health solutions (23%, 21/90) were the most represented. Of the 68 apps, 32 (47%) were involved in remote monitoring by health professionals. Regarding clinical outcomes, 5% (9/169) of the apps assessed overall survival. Randomized studies had been conducted for 21% (23/110) of the apps to assess their benefit. Of the 68 providers, 38 (56%) declared the objective of obtaining reimbursement, and 7 (18%) out of the 38 solutions seeking reimbursement were assessed as having a high probability of reimbursement. The median global score was 11.2 (range 4.7-17.4) out of 20 and the distribution of the scores followed a normal distribution pattern (Shapiro-Wilk test: P=.33). Conclusions: This multidomain prescreening scoring tool is simple, fast, and can be deployed on a large scale to initiate an assessment of the clinical relevance and quality of a clinical eHealth app. This simple tool can help a decision-maker determine which aspects of the app require further analysis and improvement.
- Open Access
- 16 May 2022
- Published by JMIR Publications Inc. in JMIR Preprints
Assessing a New Prescreening Score for the Simplified Evaluation of the Clinical Quality and Relevance of eHealth Apps: Instrument Validation Study (Preprint)
BACKGROUND In 2020, more than 250 eHealth solutions were added to app stores each day, or 90,000 in the year; however, the vast majority of these solutions have not undergone clinical validation, their quality is unknown, and the user does not know if they are effective and safe. We sought to develop a simple prescreening scoring method that would assess the quality and clinical relevance of each app. We designed this tool with 3 health care stakeholder groups in mind: eHealth solution designers seeking to evaluate a potential competitor or their own tool, investors considering a fundraising candidate, and a hospital clinician or IT department wishing to evaluate a current or potential eHealth solution. OBJECTIVE We built and tested a novel prescreening scoring tool (the Medical Digital Solution scoring tool). The tool, which consists of 26 questions that enable the quick assessment and comparison of the clinical relevance and quality of eHealth apps, was tested on 68 eHealth solutions. METHODS The Medical Digital Solution scoring tool is based on the 2021 evaluation criteria of the French National Health Authority, the 2022 European Society of Medical Oncology recommendations, and other provided scores. We built the scoring tool with patient association and eHealth experts and submitted it to eHealth app creators, who evaluated their apps via the web-based form in January 2022. After completing the evaluation criteria, their apps obtained an overall score and 4 categories of subscores. These criteria evaluated the type of solution and domain, the solution’s targeted population size, the level of clinical assessment, and information about the provider. RESULTS In total, 68 eHealth solutions were evaluated with the scoring tool. Oncology apps (22%, 20/90) and general health solutions (23%, 21/90) were the most represented. Of the 68 apps, 32 (47%) were involved in remote monitoring by health professionals. Regarding clinical outcomes, 5% (9/169) of the apps assessed overall survival. Randomized studies had been conducted for 21% (23/110) of the apps to assess their benefit. Of the 68 providers, 38 (56%) declared the objective of obtaining reimbursement, and 7 (18%) out of the 38 solutions seeking reimbursement were assessed as having a high probability of reimbursement. The median global score was 11.2 (range 4.7-17.4) out of 20 and the distribution of the scores followed a normal distribution pattern (Shapiro-Wilk test: P=.33). CONCLUSIONS This multidomain prescreening scoring tool is simple, fast, and can be deployed on a large scale to initiate an assessment of the clinical relevance and quality of a clinical eHealth app. This simple tool can help a decision-maker determine which aspects of the app require further analysis and improvement.
- 16 May 2022
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessAssessing a New Prescreening Score for the Simplified Evaluation of the Clinical Quality and Relevance of eHealth Apps: Instrument Validation Study (Preprint)
BACKGROUND In 2020, more than 250 eHealth solutions were added to app stores each day, or 90,000 in the year; however, the vast majority of these solutions have not undergone clinical validation, their quality is unknown, and the user does not know if they are effective and safe. We sought to develop a simple prescreening scoring method that would assess the quality and clinical relevance of each app. We designed this tool with 3 health care stakeholder groups in mind: eHealth solution designers seeking to evaluate a potential competitor or their own tool, investors considering a fundraising candidate, and a hospital clinician or IT department wishing to evaluate a current or potential eHealth solution. OBJECTIVE We built and tested a novel prescreening scoring tool (the Medical Digital Solution scoring tool). The tool, which consists of 26 questions that enable the quick assessment and comparison of the clinical relevance and quality of eHealth apps, was tested on 68 eHealth solutions. METHODS The Medical Digital Solution scoring tool is based on the 2021 evaluation criteria of the French National Health Authority, the 2022 European Society of Medical Oncology recommendations, and other provided scores. We built the scoring tool with patient association and eHealth experts and submitted it to eHealth app creators, who evaluated their apps via the web-based form in January 2022. After completing the evaluation criteria, their apps obtained an overall score and 4 categories of subscores. These criteria evaluated the type of solution and domain, the solution’s targeted population size, the level of clinical assessment, and information about the provider. RESULTS In total, 68 eHealth solutions were evaluated with the scoring tool. Oncology apps (22%, 20/90) and general health solutions (23%, 21/90) were the most represented. Of the 68 apps, 32 (47%) were involved in remote monitoring by health professionals. Regarding clinical outcomes, 5% (9/169) of the apps assessed overall survival. Randomized studies had been conducted for 21% (23/110) of the apps to assess their benefit. Of the 68 providers, 38 (56%) declared the objective of obtaining reimbursement, and 7 (18%) out of the 38 solutions seeking reimbursement were assessed as having a high probability of reimbursement. The median global score was 11.2 (range 4.7-17.4) out of 20 and the distribution of the scores followed a normal distribution pattern (Shapiro-Wilk test: P=.33). CONCLUSIONS This multidomain prescreening scoring tool is simple, fast, and can be deployed on a large scale to initiate an assessment of the clinical relevance and quality of a clinical eHealth app. This simple tool can help a decision-maker determine which aspects of the app require further analysis and improvement.