A ConvLSTM-based model for predicting thermal damage during laser interstitial thermal therapy
Tingting Gao1, Libin Liang2, Hui Ding1
1Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing 100084, People's Republic of China.
Physics in Medicine and Biology
|February 7, 2025
Summary
This study introduces a new AI model using intraoperative temperature data to accurately predict thermal damage during brain laser interstitial thermal therapy (LITT). The model shows promise for real-time assessment of treatment-induced tissue changes.
Area of Science:
- Medical physics and biomedical engineering
- Artificial intelligence in medicine
- Neurosurgery and oncology
Background:
- Accurate prediction of thermal damage is critical for effective laser interstitial thermal therapy (LITT), particularly in brain treatments.
- Current methods using contrast-enhanced T1-weighted imaging (CE-T1WI) provide a post-operative assessment but rely on traditional models with limitations in predicting in vivo damage.
- Existing thermal damage models often use empirical parameters from in vitro studies and fail to capture the nuanced three-state tissue damage observed clinically.
Purpose of the Study:
- To develop and evaluate a novel convolutional long short-term memory (LSTM)-based model for predicting thermal damage extent during LITT.
- To utilize intraoperative magnetic resonance temperature imaging (MRTI) data to forecast the enhancing rim seen on postoperative CE-T1WI.
- To address the limitations of traditional models by capturing more detailed tissue states and enabling real-time damage assessment.
Main Methods:
- A convolutional LSTM deep learning model was designed to process intraoperative temperature distribution history data from MRTI during LITT.
- The model was trained and validated on retrospective data from 56 patients who underwent brain LITT.
- Performance was evaluated by comparing the model's predicted enhancing rim with actual postoperative CE-T1WI findings.
Main Results:
- The proposed model demonstrated strong performance in predicting the enhancing rim on postoperative images, achieving an average Dice similarity coefficient of 0.82 (±0.063) on the test dataset.
- The model generated real-time predicted thermal damage area variation trends that closely mirrored those of traditional thermal damage models.
- This indicates the potential for accurate, real-time visualization and assessment of intraoperative thermal damage.
Conclusions:
- The novel convolutional LSTM model effectively predicts thermal damage extent in LITT using intraoperative MRTI data.
- This approach offers a significant improvement over traditional models by providing more accurate and potentially real-time assessments of thermal injury.
- The developed method could serve as a valuable tool for enhancing the precision and safety of image-guided LITT procedures.
Related Concept Videos
01:09Thermal Stress
If the temperature of an object is changed while it is prevented from expanding or contracting, the object is subjected to stress. The stress is compressive if the object expands in the absence of constraint and tensile if it contracts. This stress resulting from temperature change is known as thermal stress. It can be quite large and can cause damage. To avoid this stress, engineers may design components so they can expand and contract freely. For instance, on highways, gaps are deliberately...
01:27Thermal expansion and Thermal stress: Problem Solving
San Francisco's Golden Gate Bridge is exposed to temperatures ranging from -15 °C to 40 °C. At its coldest, the main span of the bridge is 1275 m long. Assuming that the bridge is made entirely of steel, what is the change in its length between these temperatures?
To solve the problem, first, identify the known and unknown quantities. The initial length (L) of the bridge is 1275 m, the coefficient of linear expansion (α) for steel is 12 x 10-6/°C, and the change in temperature (ΔT) is 55 °C.
To solve the problem, first, identify the known and unknown quantities. The initial length (L) of the bridge is 1275 m, the coefficient of linear expansion (α) for steel is 12 x 10-6/°C, and the change in temperature (ΔT) is 55 °C.
01:19Thermal Strain
Thermal strain is a concept that arises when we consider how temperature changes affect structures. Unlike the conventional assumption that structures remain constant under load, real-world scenarios often involve temperature fluctuations that can significantly impact these structures. Consider a homogeneous rod with a uniform cross-section resting freely on a flat horizontal surface. If the rod's temperature increases, the rod elongates. This elongation is proportional to the temperature...