Related Experiment Video
Updated: Aug 27, 2025
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Deep Learning-Based Nuclear Morphometry Reveals an Independent Prognostic Factor in Mantle Cell Lymphoma
Wen-Yu Chuang1, Wei-Hsiang Yu2, Yen-Chen Lee3
1Department of Pathology, Chang Gung Memorial Hospital and Chang Gung University, Taoyuan, Taiwan; School of Medicine, Chang Gung University, Taoyuan, Taiwan; Chang Gung Molecular Medicine Research Center, Chang Gung University, Taoyuan, Taiwan; Center for Vascularized Composite Allotransplantation, Chang Gung Memorial Hospital, Taoyuan, Taiwan.
A new deep learning model objectively measures nuclear morphometric features in mantle cell lymphoma (MCL). This nuclear morphometric score, derived from nuclear irregularity, serves as an independent prognostic factor for MCL patients.
Area of Science:
- Hematologic Oncology
- Computational Pathology
- Artificial Intelligence in Medicine
Background:
- Mantle cell lymphoma (MCL) prognosis is influenced by blastoid/pleomorphic morphology, but Ki-67 often overrides its value.
- Objective and robust prognostic markers are needed for MCL patient stratification.
Purpose of the Study:
- To develop and validate a deep learning-based nuclear segmentation model for extracting quantitative nuclear morphometric features in MCL.
- To assess the prognostic value of these morphometric parameters, individually and as a score, in MCL.
Main Methods:
- A deep learning model was used to segment tumor cell nuclei in 103 MCL cases.
- Eight nuclear morphometric attributes were extracted, and their mean, variance, skewness, and kurtosis were calculated, yielding 32 parameters.
- Univariate and multivariate analyses were performed to identify independent prognostic factors.
Main Results:
- 17 morphometric parameters significantly differed between classic and blastoid/pleomorphic MCL.
- Low skewness and high mean of nuclear irregularity emerged as independent adverse prognostic factors.
- A novel nuclear morphometric score, combined with the Biologic MCL International Prognostic Index (bMIPI), improved risk stratification.
Conclusions:
- Nuclear morphometric analysis using deep learning provides an objective and robust prognostic tool for MCL.
- The developed morphometric score is a more reliable prognostic indicator than blastoid/pleomorphic morphology.
- This approach enhances risk stratification for patients with mantle cell lymphoma.
