Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

AI-driven feature descriptor design using CNNs and transfer learning for enhanced image retrieval accuracy and efficiency.

Scientific reports·2026

Same author

PI-HydroGNN: a physics-informed spatiotemporal graph neural network framework for hydraulic reliability, leakage detection, and energy-efficient operation in water distribution systems.

Scientific reports·2026

Same author

A voyage on computer aided intelligent algorithms for the segmentation of brain tissues for neurodisorder diagnosis.

The International journal of neuroscience·2026

Same author

Temporally consistent longitudinal brain tumor segmentation using a temporal spatial transformer network.

Scientific reports·2026

Same author

Adaptive fuzzy deep learning with multimodal sensor fusion for enhanced plant disease detection.

Scientific reports·2026

Same author

GNN-ML-FRL: a graph-enhanced meta-adaptive federated learning framework for scalable pest identification and ernvironmental modeling.

Scientific reports·2026

Search research articles

Related Experiment Video

Updated: May 27, 2026

Lesion Explorer: A Video-guided, Standardized Protocol for Accurate and Reliable MRI-derived Volumetrics in Alzheimer's Disease and Normal Elderly

Lesion Explorer: A Video-guided, Standardized Protocol for Accurate and Reliable MRI-derived Volumetrics in Alzheimer's Disease and Normal Elderly

Published on: April 14, 2014

Progression-guided spatiotemporal memory transformers for accurate and consistent longitudinal brain tumor

Sandeep Kumar Mathivanan¹, Shamala K Subramaniam², Dafik³

¹School of Computing Science and Engineering, Galgotias University, Greater Noida, Uttar Pradesh, 203201, India.

Scientific Reports

|May 25, 2026

Summary

A new deep learning model, Progression-Guided Spatiotemporal Memory Transformer (PGSMT), accurately segments brain tumors over time. This model improves tumor progression monitoring and treatment efficacy by preserving shape consistency and distinguishing actual changes from noise.

Keywords:

Boundary-enhanced transformer Cross-time structural alignment Deep learning Dice score Longitudinal brain tumor segmentation MRI Progression-aware temporal modeling Spatiotemporal memory Temporal consistency Tumor evolution dynamics

More Related Videos

A Comprehensive Protocol for Manual Segmentation of the Medial Temporal Lobe Structures

A Comprehensive Protocol for Manual Segmentation of the Medial Temporal Lobe Structures

Published on: July 2, 2014

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

Related Experiment Videos

Last Updated: May 27, 2026

Lesion Explorer: A Video-guided, Standardized Protocol for Accurate and Reliable MRI-derived Volumetrics in Alzheimer's Disease and Normal Elderly

Lesion Explorer: A Video-guided, Standardized Protocol for Accurate and Reliable MRI-derived Volumetrics in Alzheimer's Disease and Normal Elderly

Published on: April 14, 2014

A Comprehensive Protocol for Manual Segmentation of the Medial Temporal Lobe Structures

A Comprehensive Protocol for Manual Segmentation of the Medial Temporal Lobe Structures

Published on: July 2, 2014

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

Area of Science:

Medical Imaging
Artificial Intelligence
Neuro-oncology

Background:

Accurate brain tumor segmentation is crucial for monitoring disease progression and treatment efficacy.
Current deep learning models often fail to account for the continuous nature of tumor growth, leading to inaccurate segmentations in longitudinal scans.
Existing methods struggle with preserving tumor morphology and distinguishing true progression from noise.

Purpose of the Study:

To propose a novel deep learning model, the Progression-Guided Spatiotemporal Memory Transformer (PGSMT), for accurate and consistent brain tumor segmentation across multiple time points.
To address the limitations of existing models in handling the temporal dynamics and morphological changes of brain tumors.

Main Methods:

Developed PGSMT, a framework incorporating a progression-aware temporal memory module, a cross-time structural alignment mechanism, and a boundary-enhanced transformer encoder.
The model learns temporal weighting to differentiate actual tumor progression from noise.
Utilized the BraTS longitudinal benchmark dataset for evaluation.

Main Results:

PGSMT significantly outperformed Convolutional Neural Network (CNN), hybrid CNN-Transformer, and Transformer models on the BraTS longitudinal dataset.
Achieved Dice scores of 88.1% for enhancing tumor, 90.2% for tumor core, and 93.0% for total tumor.
Demonstrated statistically significant improvements (p < 0.05) and reduced inter-scan volumetric inconsistencies, indicating enhanced temporal stability.

Conclusions:

PGSMT offers a robust solution for longitudinal brain tumor segmentation, improving accuracy and consistency.
The model's ability to capture temporal dynamics and preserve morphology enhances its utility for clinical monitoring and treatment assessment.
PGSMT represents a significant advancement in applying deep learning to dynamic medical image analysis.