Related Experiment Video
Updated: Jun 5, 2025
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Development and validation of a deep learning model for morphological assessment of myeloproliferative neoplasms
Rong Wang1, Zhongxun Shi1, Yuan Zhang2,3
1Department of Haematology, Collaborative Innovation Center for Cancer Personalized Medicine, Jiangsu Province Hospital, The First Affiliated Hospital of Nanjing Medical University, Nanjing, China.
Abstract:
The subjectivity of morphological assessment and the overlapping pathological features of different subtypes of myeloproliferative neoplasms (MPNs) make accurate diagnosis challenging. To improve the pathological assessment of MPNs, we developed a diagnosis model (fusion model) based on the combination of bone marrow whole-slide images (deep learning [DL] model) and clinical parameters (clinical model). Thousand and fifty-one MPN and non-MPN patients were divided into the training, internal testing and one internal and two external validation cohorts (the combined validation cohort). In the combined validation cohort, fusion model achieved higher areas under curve (AUCs) than clinical or DL model or both for MPNs and subtype identification. Compared with haematopathologists with different experience, clinical model achieved AUC which was comparable to seniors and higher than juniors (p = 0.0208) for polycythaemia vera. The AUCs of fusion model were comparable to seniors and higher than juniors for essential thrombocytosis (p = 0.0141), prefibrotic primary myelofibrosis (p = 0.0085) and overt primary myelofibrosis (p = 0.0330) identification. In conclusion, the performances of our proposed models are equivalent to senior haematopathologists and better than juniors, providing a new perspective on the utilization of DL algorithms in MPN morphological assessment.
Insights
Diagnosing myeloproliferative neoplasms (MPNs) is challenging. A new fusion model combining deep learning (DL) and clinical data shows diagnostic performance comparable to senior hematopathologists for MPN subtypes.
Area of Science:
- Hematology
- Computational Pathology
- Artificial Intelligence in Medicine
Background:
- Morphological assessment of myeloproliferative neoplasms (MPNs) faces challenges due to subjectivity and overlapping features.
- Accurate diagnosis of MPNs is crucial for effective treatment and patient outcomes.
- Existing diagnostic methods may lack the precision needed for complex MPN subtypes.
Purpose of the Study:
- To develop and validate a novel diagnostic model for MPNs.
- To improve the accuracy of pathological assessment in MPN diagnosis.
- To evaluate the performance of a combined deep learning and clinical data model against human experts.
Main Methods:
- A fusion model was developed integrating bone marrow whole-slide images analyzed by a deep learning (DL) model with clinical parameters.
- A total of 1051 MPN and non-MPN patients were included in training, internal testing, and validation cohorts.
- The fusion model's diagnostic performance was compared against a DL-only model, a clinical-only model, and hematopathologists.
Main Results:
- The fusion model demonstrated superior areas under the curve (AUCs) for MPN and subtype identification compared to individual models in the validation cohort.
- The clinical model's AUC for polycythemia vera identification was comparable to senior and higher than junior hematopathologists.
- The fusion model's AUCs for essential thrombocytosis and primary myelofibrosis subtypes were comparable to senior and higher than junior hematopathologists.
Conclusions:
- The developed fusion model achieves diagnostic performance equivalent to senior hematopathologists and superior to junior ones.
- This study offers a new perspective on using deep learning algorithms for MPN morphological assessment.
- The fusion model shows promise in enhancing the accuracy and consistency of MPN diagnosis.
