Framework for 2D-3D image fusion of infrared thermography with preoperative MRI

Artificial Intelligence in IR Thermal Imaging and Sensing for Medical Applications

The state of the art in IR thermal imaging methods for applications in medical diagnostics is discussed. A review of advances in IR thermal imaging technology in the years 1960-2024 is presented. Recently used artificial intelligence (AI) methods in the analysis of thermal images are the main interest. IR thermography is discussed in view of novel applications of machine learning methods for improved diagnostic analysis and medical treatment. The AI approach aims to improve image quality by denoising thermal images, using applications of AI super-resolution algorithms, removing artifacts, object detection, face and characteristic features localization, complex matching of diagnostic symptoms, etc.

Application of Preoperative Multimodal Image Fusion Technique in Microvascular Decompression Surgery via Suboccipital Retrosigmoid Approach

OBJECTIVE

To explore the application value of preoperative multimodal image fusion technique in microvascular decompression (MVD) surgery via the suboccipital retrosigmoid approach.

METHODS

Comprehensive data of 13 patients with primary trigeminal neuralgia (TN) and 13 patients with hemifacial spasm (HFS) treated by MVD surgery via the suboccipital retrosigmoid approach at the Department of Neurosurgery in Zhuhai People's Hospital from January 2021 to December 2021 were retrospectively analyzed. Preoperatively, all patients underwent cranial thin-section computed tomography and magnetic resonance examinations. Three-dimensional (3D) digital images of the skull, brainstem, nerves, and blood vessels were constructed by the 3D-slicer software or RadiAnt DICOM Viewer, which were then applied to design the surgical approach and surgical plan. The multimodal image fusion results, clinical characteristics, intraoperative data, surgical outcomes, and complications of all patients were summarized.

RESULTS

The 3D digital images after fusion reconstruction can vividly show the anatomical relationship between the skull, brainstem, nerves, and blood vessels and was helpful to tailor the surgical strategy. All 26 patients underwent a smooth surgery. During the surgery, the key points were accurately located, the corners of the transverse sinus and sigmoid sinus were completely exposed, and no venous sinus injury occurred in all 26 patients. The key point was approximately located at the top point of the digastric groove, 12.3 ± 0.46 mm vertically above and 6.3 ± 0.6 mm laterally to the Frankfurt horizontal plane. The average cranial opening time was 30.4 (±3.6) min, and the mean operating time was 104.7 (±12.1) min. The diameter of the bone window was about 2.0 cm-3.0 cm, and the bone flap was restored. Among the 13 patients with primary TN, 12 (92.3%) exhibited complete relief of pain and 1 had significant relief. Complications of surgery included facial sensory numbness in 1 case, vertigo in 2 cases, and herpes at the corners of the mouth in 1 case. Of the 13 patients with HFS, 12 (92.3%) had complete relief of facial twitching symptoms and 1 had significant relief, and the complications included mild facial palsy in 2 (15.4%) cases and facial sensory numbness in another 2 (15.4%) cases. The mean follow-up time after surgery ranged from 6-16 months, and 1 of 26 patients experienced recurrence of HFS during the follow-up period.

CONCLUSIONS

Preoperative multimodal image fusion technology can provide adequate preoperative assessment for patients and assistance in designing surgical approaches, which is an important guideline for MVD surgery via the suboccipital retrosigmoid approach for primary TN and facial muscle spasm.

Towards an Accurate MRI Acute Ischemic Stroke Lesion Segmentation Based on Bioheat Equation and U-Net Model

Acute ischemic stroke represents a cerebrovascular disease, for which it is practical, albeit challenging to segment and differentiate infarct core from salvageable penumbra brain tissue. Ischemic stroke causes the variation of cerebral blood flow and heat generation due to metabolism. Therefore, the temperature is modified in the ischemic stroke region. In this paper, we incorporate acute ischemic stroke temperature profile to reinforce segmentation accuracy in MRI. Pennes bioheat equation was used to generate brain thermal images that may provide rich information regarding the temperature change in acute ischemic stroke lesions. The thermal images were generated by calculating the temperature of the brain with acute ischemic stroke. Then, U-Net was used in this paper for the segmentation of acute ischemic stroke. A dataset of 3192 images was created to train U-Net using k-fold crossvalidation. The training time was about 10 hours and 35 minutes in NVIDIA GPU. Next, the obtained trained model was compared with recent methods to analyze the effect of the ischemic stroke temperature profile in segmentation. The obtained results show that significant parts of acute ischemic stroke and background areas are segmented only in thermal images, which proves the importance of using thermal information to improve the segmentation outcomes in MRI diagnosis.

Infrared thermography mapping plus neuronavigation target location in an eloquent area cavernoma resection

Background

Safety and efficacy are irrebuttable goals in neurosurgery.

Methods

We performed a subcortical cavernoma resection in an eloquent area, where we recorded and compared the maximal and minimal brain temperature measured by an infrared thermographic camera and thermometer with the neuronavigation (NN) target location and real anatomical lesion location.

Results

The hottest cortical point correlated to the subcortical cavernoma location. The NN located the target at 10 mm away from the hottest point.

Conclusion

More studies are needed to better understand the thermic radiation of the brain in health and in disease, but we believe that evaluating brain temperature, it could be possible to improve accuracy in neurosurgery and generate more knowledge about brain metabolism in vivo.

Thermography mapping patterns in temporal lobe epilepsy surgery

Background:

In several epilepsy etiologies, the macroscopic appearance of the epileptogenic tissue is identical to the normal, which makes it hard to balance between how much cytoreduction or disconnection and brain tissue preservation must be done. A strategy to tackle this situation is by evaluating brain metabolism during surgery using infrared thermography mapping (IrTM).

Methods:

In 12 epilepsy surgery cases that involved the temporal lobe, we correlated the IrTM, electrocorticography, and neuropathology results.

Results:

Irritative zones (IZ) had a lower temperature in comparison to the surrounding cortex with normal electric activity (difference in temperature (ΔT) from 1.2 to 7.1, mean 3.40°C standard deviation ± 1.61). The coldest zones correlated exactly with IZ in 9/10 cortical dysplasia (CD) cases. In case 3, the coldest area was at 1 cm away from the IZ. In 10/10 dysplasia cases (cases 1–4, 6–11), there was a radial heating pattern originating from the coldest cortical point. In 2/2 neoplasia cases, the temporal lobe cortical temperature was more homogeneous than in the CD cases, with no radial heating pattern, and there were no IZ detected. In case 8, we found the coldest IrTM recording in the hippocampus, which correlated to the maximal irritative activity recorded by strip electrodes. The ΔT is inversely proportional to epilepsy chronicity.

Conclusion:

IrTM could be useful in detecting hypothermic IZ in CD cases. As the ΔT is inversely proportional to epilepsy chronicity, this variable could affect the metabolic thermic patterns of the human brain.

Registration and Fusion of Thermographic and Visual-Light Images in Neurosurgery

An intraoperative imaging system facilitating the localisation and characterisation of functional areas, pathological tissue, or perfusion disorders, could enormously support medical decisions during neurosurgical interventions and, thus, reduce the risk for the patients. To provide both structural and functional information of the brain tissue to the surgeon, a novel multimodal approach based on the measurement of long-wave infrared radiation and visual-light imaging is very promising. In this contribution, we discuss various methods for the registration and fusion of thermographic and visual-light images. The methods are evaluated quantitatively and qualitatively regarding their practicability during surgery. Furthermore, we introduce appropriate architectures for a digital hardware implementation of the registration and fusion algorithms. The designs are implemented on our reconfigurable intraoperative imaging system, revealing real-time processing performance.