• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.12
• /
• pp.4816-4834
• /
• 2020

This paper proposes transfer learning and fine-tuning techniques for a deep learning model to detect three distinct brain tumors from Magnetic Resonance Imaging (MRI) scans. In this work, the recent YOLOv4 model trained using a collection of 3064 T1-weighted Contrast-Enhanced (CE)-MRI scans that were pre-processed and labeled for the task. This work trained with the partial 29-layer YOLOv4-Tiny and fine-tuned to work optimally and run efficiently in most platforms with reliable performance. With the help of transfer learning, the model had initial leverage to train faster with pre-trained weights from the COCO dataset, generating a robust set of features required for brain tumor detection. The results yielded the highest mean average precision of 93.14%, a 90.34% precision, 88.58% recall, and 89.45% F1-Score outperforming other previous versions of the YOLO detection models and other studies that used bounding box detections for the same task like Faster R-CNN. As concluded, the YOLOv4-Tiny can work efficiently to detect brain tumors automatically at a rapid phase with the help of proper fine-tuning and transfer learning. This work contributes mainly to assist medical experts in the diagnostic process of brain tumors.

• Journal of Korean Neurosurgical Society
• /
• v.46 no.4
• /
• pp.355-359
• /
• 2009

Objective : The definition of empty sella syndrome is 'an anatomical entity in which the pituitary fossa is partially or completely filled with cerebrospinal fluid, while the pituitary gland is compressed against the posterior rim of the fossa'. Reports of this entities relating to the brain tumors not situated in the pituitary fossa, have rarely been reported. Methods : In order to analyze the incidence and relationship of empty sella in patients having brain tumors, the authors reviewed preoperative magnetic resonance imaging (MRI) of 72 patients with brain tumor regardless of pathology except the pituitary tumors. The patients were operated in single institute by one surgeon. There were 25 males and 47 females and mean patient age was 53 years old (range from 5 years to 84 years). Tumor volume was ranged from 2 cc to 238 cc. Results : The overall incidence of empty sella was positive in 57/72 cases (79.2%). Sorted by the pathology, empty sella was highest in meningioma (88.9%, p=0.042). The empty sella was correlated with patient's increasing age (p=0.003) and increasing tumor volume (p=0.016). Conclusion : Careful review of brain MRI with periodic follow up is necessary for the detection of secondary empty sella in patients with brain tumors. In patients with confirmed empty sella, follow up is mandatory for the management of hypopituitarism, cerebrospinal fluid (CSF) rhinorrhea, visual disturbance and increased intracranial pressure.

• The Korean Journal of Nuclear Medicine
• /
• v.26 no.1
• /
• pp.14-25
• /
• 1992

Thallium-201 $(^{201}T1)$ SPECT studies were performed on a normal volunteer and 12 patients with intracerebral lesions: 3 patients with gliomas, 3 patients with meningiomas, 1 patient each with metastatic tumor, brain abscess, and cerebral infarction, and 3 postirradiation patients. (2 with metastatic tumors, 1 with lymphoma). A $^{201}T1$ index, based on the ratio of $^{201}T1$ uptake in the brain lesion versus the homologous contralateral brain, was calculated and compared with tumor histology and CT/MRI findings. The SPECT $^{201}T1$ scan showed minimal uptake of tracer in a normal brain. There was substantial uptake of $^{201}T1$ in high-grade gliomas (index>1.5) with little uptake in low-grade lesions. A previously irradiated patient with recurrent astrocytoma, in whom MRI study was unable to distinguish tumor recurrence from necrosis, showed the lesions with high $^{201}T1$ indices in both hemispheric regions (2.50/1.93), suggesting tumor recurrence. Two meningiomas and a metastatic tumor showed varying degrees of $^{201}T1$ uptake (index $1.71\sim8.15$), revealing that $^{201}T1$ uptake is not exclusive to high-grade gliomas. In 2 postirradiation patients with metastatic tumors, no abnormal $^{201}T1$ uptake was found in the cerebral lesions, shortly after the initiation of radiation therapy or despite the persistence of enhancing lesions-though improved-on MR images, suggesting that $^{201}T1$ uptake may reflect the metabolic and possibly clonogenic activities of tumors and the brain $^{201}T1$ SPECT imaging might be valuable for the evaluation of tumor responsiveness to the therapy and for early detection of tumor recurrence. A patient with brain abscess on antibiotic treatment, showig increased uptake of $^{201}T1$ in the resolving lesions (index 2.87/1.52) is discussed. In a patient with cerebral infarction, there was no abnormal uptake of $^{201}T1$ in infarcted tissue. When using a threshold index of 1.5, correlation rate between $^{201}T1$ uptake and contrast enhancement of the cerebral lesions on CT/MRI was 73% (8/11). In conclusion, the brain $^{201}T1$ SPECT imaging may be useful for assessment of tumor response to the therapy and to predict low-or high-grade lesions.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.9
• /
• pp.2589-2604
• /
• 2024

Brain tumors, characterized by uncontrollable cellular growths, are a significant global health challenge. Navigating the complexities of tumor identification due to their varied dimensions and positions, our research introduces enhanced methods for precise detection. Utilizing advanced learning techniques, we've improved early identification by preprocessing clinical dataset-derived images, augmenting them via a Generative Adversarial Network, and applying an Improved Privacy-Preserving Collaborative Convolutional Neural Network (IPC-CNN) for segmentation. Recognizing the critical importance of data security in today's digital era, our framework emphasizes the preservation of patient privacy. We evaluated the performance of our proposed model on the Figshare and BRATS 2018 datasets. By facilitating a collaborative model training environment across multiple healthcare institutions, we harness the power of distributed computing to securely aggregate model updates, ensuring individual data protection while leveraging collective expertise. Our IPC-CNN model achieved an accuracy of 99.40%, marking a notable advancement in brain tumor classification and offering invaluable insights for both the medical imaging and machine learning communities.

• The Journal of the Korean bone and joint tumor society
• /
• v.9 no.2
• /
• pp.223-232
• /
• 2003

As we know, alveolar soft part sarcoma is usually found at the head region in children and thigh in adults. It is very rare tumor that has poor prognosis due to its late detection after distant metastasis in spite of its slow growth rate. It is histologically characterized by pseudoalveolar pattern tumor cells. And metastasis usually occur in the site of lung, brain and skeleton in order lately. We have managed two cases of the sarcoma, one which took place in relatively rare part, pelvic bone and has spread to the brain, the other which primarily occured in the calf. For its varity, we report these two cases with reviewing of the literatures.

• Korean Journal of Head & Neck Oncology
• /
• v.20 no.1
• /
• pp.7-12
• /
• 2004

Background and Objectives: Brain metastasis from well-differentiated thyroid carcinoma is uncommon, and the treatment remains controversial. In this study, we analyzed cases of brain metastasis from well differentiated thyroid carcinoma to determine optimal treatment. Material and Methods: From March 1986 to May 2003, we experienced 13 cases of brain metastases from well-differentiated thyroid carcinomas. There were 6 men and 7 women with a mean age of 62.3 years. The time interval detecting the brain metastasis, treatment methods, outcomes, and the prognostic factors were analyzed. Results: In 11 patients, brain metastasis was found at a mean of 96 months after thyroid surgery (range, 1-204months). In 1 patient, brain metastasis was found simultaneously with the thyroid cancer. In the remaining one, brain metastasis was presented as the first sign. Treatment included gamma knife therapy in 2, external radiation therapy (ExRT) in 2, resection and ExRT in 2, gamma knife therapy and ExRT in 2 and intracranial holmium injection in 1. Three patients received no treatment. The mean survival was 20.5months (1-119 months). Only 2 patients with early detection remained alive after treatment (23months and 119months). The age, sex, primary tumor size, combined organs of distant metastasis and treatment methods did not affect the prognosis. There were noticeable symptomatic improvements in patients who received any kind of treatment for brain metastasis. Conclusion: Early detection and treatment of brain metastasis appears to bring symptomatic relief and improve survival. Therefore, early screening for brain metastasis is recommended for patients with symptoms. However, further study is needed to determine the optimal method of treatment.

• Asian Pacific Journal of Cancer Prevention
• /
• v.16 no.13
• /
• pp.5401-5406
• /
• 2015

Background: Brain metastases are the most common intra-cranial neoplasms. The incidence is on a rise due to advanced imaging techniques. Aims: The objective of the study was to analyse the clinical and demographic profile of patients with brain metastases from primary solid tumors. Materials and Methods: This is a retrospective single institutional study covering 130 consecutive patients with brain metastases from January 2007 to August 2014. Results: Some 64.6% of the patients were females. The majority were in the sixth decade of life. The site of the primary tumor was the lungs in 50.8% of the cases. The overall median time from the diagnosis of the primary malignancy to detection of brain metastases was 21.4 months. Survival was found to be significantly improved in patients with solitary brain lesions when compared to patients with multiple brain metastases, and in patients undergoing surgical excision with or without cranial irradiation when compared to whole brain irradiation alone. The majority of the cases belonged to the recursive partitioning analysis class II group. Whole brain radiation therapy was delivered to 79% of the patients. Conclusions: Most of the patients with brain metastases in the study belonged to recursive partitioning analysis classes II or III, and hence had poor prognosis. Most of the patients in the Indian context either do not satisfy the indications for surgical excision or are incapable of bearing the high cost associated with stereotactic radiosurgery. Treatment should be tailored on an individual basis to all these patients.

• Current Optics and Photonics
• /
• v.3 no.3
• /
• pp.256-262
• /
• 2019

Extensive tumor resection accompanied by radiotherapy and chemotherapy is the standard of care for malignant gliomas. However, there is a significant obstacle to the complete resection of the tumor due to the difficulty of distinguishing tumor and normal brain tissue with a conventional surgical microscope. Recently, multiple studies have shown the possibility of fluorescence-guided surgery in malignant gliomas. The most used fluorescence dyes for brain tumor surgery are 5-aminolevulinic acid (5-ALA) and indocyanine green (ICG). In this paper, a new fluorescence guided operation system, which can detect both 5-ALA and ICG fluorescent images simultaneously, is presented. This operation system consists of light emitting diodes (LEDs) which emits 410 nm and 740 nm wavelengths. We have performed experiments on rats in order to verify the operation of the newly developed operation system. Oral administration and imaging were performed to observe the fluorescence of 5-ALA and ICG fluorescence in rats. When LEDs at wavelengths of 410 nm and 740 nm were irradiated on rats, 628 nm wavelength with a violet fluorescence color and 825 nm wavelength with a red fluorescence color were expressed in 5-ALA and ICG fluorescent material, respectively, thus we were able to distinguish the tumor tissues easily. Previously, due to the poor resolution of the conventional surgical microscope and the fact that the color of the vein is similar to that of the tumor, the tumor resection margin was not easy to observe, thus increasing the likelihood for cancer recurrence. However, when the tumor is observed through the fluorescence guided operation system, it is possible to easily distinguish the color with the naked eye and it can be completely removed. Therefore, it is expected that surgical removal of cancerous tumors will be possible and surgical applications and surgical microscopes for cancer tumor removal surgery will be promising in the future.

• Journal of Korea Multimedia Society
• /
• v.9 no.9
• /
• pp.1150-1159
• /
• 2006

This paper proposed a curve progress control function of the area base instead of the existing edge indication function, in order to detect the brain ventricle area by utilizing a geodesic active contour model. The proposed curve progress control function is very effective in detecting the brain ventricle area and this function is based on the average brightness of the brain ventricle area which appears brighter in MRI images. Compared numerically by using various measures, the proposed method in this paper can detect brain ventricle areas better than the existing method. By examining images of normal and diseased brain's images by brain tumor, we compared the several brain ventricle detection algorithms with proposed method visually and verified the effectiveness of the proposed method.

• Korean Journal of Optics and Photonics
• /
• v.26 no.1
• /
• pp.23-29
• /
• 2015

In this paper, we propose a microscope system for detecting both a tumor and blood vessels in brain tumor surgery as fluorescence images by using multiple light sources and a beam-splitter module. The proposed method displays fluorescent images of the tumor and blood vessels on the same display device and also provides accurate information about them to the operator. To acquire a fluorescence image, we utilized 5-ALA (5-aminolevulinic acid) for the tumor and ICG (Indocyanine green) for blood vessels, and we used a beam-splitter module combined with a microscope for simultaneous detection of both. The beam-splitter module showed the best performance at 600 nm for 5-ALA and above 800 nm for ICG. The beam-splitter is flexible to enable diverse objective setups and designed to mount a filter easily, so beam-splitter and filter can be changed as needed, and other fluorescent dyes besides 5-ALA and ICG are available. The fluorescent images of the tumor and the blood vessels can be displayed on the same monitor through the beam-splitter module with a CCD camera. For ICG, a CCD that can detect the near-infrared region is needed. This system provides the acquired fluorescent image to an operator in real time, matching it to the original image through a similarity transform.