Leveraging EfficientNetB3 in a Deep Learning Framework for High-Accuracy MRI Tumor Classification

Brain tumor is a global issue due to which several people suffer,and its early diagnosis can help in the treatment in a more efficient manner.Identifying different types of brain tumors,including gliomas,meningiomas,pituitary tumors,as well as confirming the absence of tumors,poses a significant challenge using MRI images.Current approaches predominantly rely on traditional machine learning and basic deep learning methods for image classification.These methods often rely on manual feature extraction and basic convolutional neural networks(CNNs).The limitations include inadequate accuracy,poor generalization of new data,and limited ability to manage the high variability in MRI images.Utilizing the EfficientNetB3 architecture,this study presents a groundbreaking approach in the computational engineering domain,enhancing MRI-based brain tumor classification.Our approach highlights a major advancement in employing sophisticated machine learning techniques within Computer Science and Engineering,showcasing a highly accurate framework with significant potential for healthcare technologies.The model achieves an outstanding 99%accuracy,exhibiting balanced precision,recall,and F1-scores across all tumor types,as detailed in the classification report.This successful implementation demonstrates the model’s potential as an essential tool for diagnosing and classifying brain tumors,marking a notable improvement over current methods.The integration of such advanced computational techniques in medical diagnostics can significantly enhance accuracy and efficiency,paving the way for wider application.This research highlights the revolutionary impact of deep learning technologies in improving diagnostic processes and patient outcomes in neuro-oncology.

ARGA-Unet:Advanced U-net segmentation model using residual grouped convolution and attention mechanism for brain tumor MRI image segmentation

Background Magnetic resonance imaging(MRI)has played an important role in the rapid growth of medical imaging diagnostic technology,especially in the diagnosis and treatment of brain tumors owing to its non invasive characteristics and superior soft tissue contrast.However,brain tumors are characterized by high non uniformity and non-obvious boundaries in MRI images because of their invasive and highly heterogeneous nature.In addition,the labeling of tumor areas is time-consuming and laborious.Methods To address these issues,this study uses a residual grouped convolution module,convolutional block attention module,and bilinear interpolation upsampling method to improve the classical segmentation network U-net.The influence of network normalization,loss function,and network depth on segmentation performance is further considered.Results In the experiments,the Dice score of the proposed segmentation model reached 97.581%,which is 12.438%higher than that of traditional U-net,demonstrating the effective segmentation of MRI brain tumor images.Conclusions In conclusion,we use the improved U-net network to achieve a good segmentation effect of brain tumor MRI images.

Intelligent Machine Learning Based Brain Tumor Segmentation through Multi-Layer Hybrid U-Net with CNN Feature Integration

Brain tumors are a pressing public health concern, characterized by their high mortality and morbidity rates.Nevertheless, the manual segmentation of brain tumors remains a laborious and error-prone task, necessitatingthe development of more precise and efficient methodologies. To address this formidable challenge, we proposean advanced approach for segmenting brain tumorMagnetic Resonance Imaging (MRI) images that harnesses theformidable capabilities of deep learning and convolutional neural networks (CNNs). While CNN-based methodshave displayed promise in the realm of brain tumor segmentation, the intricate nature of these tumors, markedby irregular shapes, varying sizes, uneven distribution, and limited available data, poses substantial obstacles toachieving accurate semantic segmentation. In our study, we introduce a pioneering Hybrid U-Net framework thatseamlessly integrates the U-Net and CNN architectures to surmount these challenges. Our proposed approachencompasses preprocessing steps that enhance image visualization, a customized layered U-Net model tailoredfor precise segmentation, and the inclusion of dropout layers to mitigate overfitting during the training process.Additionally, we leverage the CNN mechanism to exploit contextual information within brain tumorMRI images,resulting in a substantial enhancement in segmentation accuracy.Our experimental results attest to the exceptionalperformance of our framework, with accuracy rates surpassing 97% across diverse datasets, showcasing therobustness and effectiveness of our approach. Furthermore, we conduct a comprehensive assessment of ourmethod’s capabilities by evaluating various performance measures, including the sensitivity, Jaccard-index, andspecificity. Our proposed model achieved 99% accuracy. The implications of our findings are profound. Theproposed Hybrid U-Net model emerges as a highly promising diagnostic tool, poised to revolutionize brain tumorimage segmentation for radiologists and clinicians.

Cystic Degeneration of Uterine Myoma Simulating Ovarian Cancer in Postpartum: A Case Report at the Teaching Hospital of Angre

The transformation of uterine fibroids is common in relation to their development. Giant forms of cystic degeneration are rare. They raise diagnostic difficulties with other pelvic tumors, such as ovarian tumors and leiomyosarcomas. Magnetic resonance imaging specifies the original organ, the volume and the main relationships of fibromyoma with adjacent structures. The diagnosis of certainty is based on laparotomy coupled with histology. The authors illustrate these difficulties by observing a giant cystic degenerative fibroma in a 26-year-old G1P1 woman in the postpartum period.

An Adapted Convolutional Neural Network for Brain Tumor Detection

In medical imaging, particularly for analyzing brain tumor MRIs, the expertise of skilled neurosurgeons or radiologists is often essential. However, many developing countries face a significant shortage of these specialists, which impedes the accurate identification and analysis of tumors. This shortage exacerbates the challenge of delivering precise and timely diagnoses and delays the production of comprehensive MRI reports. Such delays can critically affect treatment outcomes, especially for conditions requiring immediate intervention, potentially leading to higher mortality rates. In this study, we introduced an adapted convolutional neural network designed to automate brain tumor diagnosis. Our model features fewer layers, each optimized with carefully selected hyperparameters. As a result, it significantly reduced both execution time and memory usage compared to other models. Specifically, its execution time was 10 times shorter than that of the referenced models, and its memory consumption was 3 times lower than that of ResNet. In terms of accuracy, our model outperformed all other architectures presented in the study, except for ResNet, which showed similar performance with an accuracy of around 90%.

Machine Learning-Based Models for Magnetic Resonance Imaging(MRI)-Based Brain Tumor Classification

In the medical profession,recent technological advancements play an essential role in the early detection and categorization of many diseases that cause mortality.The technique rising on daily basis for detecting illness in magnetic resonance through pictures is the inspection of humans.Automatic(computerized)illness detection in medical imaging has found you the emergent region in several medical diagnostic applications.Various diseases that cause death need to be identified through such techniques and technologies to overcome the mortality ratio.The brain tumor is one of the most common causes of death.Researchers have already proposed various models for the classification and detection of tumors,each with its strengths and weaknesses,but there is still a need to improve the classification process with improved efficiency.However,in this study,we give an in-depth analysis of six distinct machine learning(ML)algorithms,including Random Forest(RF),Naïve Bayes(NB),Neural Networks(NN),CN2 Rule Induction(CN2),Support Vector Machine(SVM),and Decision Tree(Tree),to address this gap in improving accuracy.On the Kaggle dataset,these strategies are tested using classification accuracy,the area under the Receiver Operating Characteristic(ROC)curve,precision,recall,and F1 Score(F1).The training and testing process is strengthened by using a 10-fold cross-validation technique.The results show that SVM outperforms other algorithms,with 95.3%accuracy.

Glomus Tumor of the Kidney: A Case Report with CT, MRI, and Histopathological Findings

We describe the computed tomographic (CT) and magnetic resonance imaging (MRI) features of a very rare renal neoplasm, a glomus tumor. Our patient was a 68-year-old woman with a history of high grade T1 stage bladder cancer, status post intravesical Bacillus Calmette-Guérin (BCG) therapy and left ureteral stent placement, who presented for routine follow-up imaging evaluation of the urothelial tract. Computed tomographic urography (CTU) incidentally demonstrated a 1.7 cm well-circumscribed, non-calcified, non-fat containing lesion in the left renal cortex with arterial phase continuous peripheral rim enhancement and central hypoattenuation relative to enhanced renal parenchyma. Subsequent MRI showed the lesion to be isointense in signal intensity relative to the renal parenchyma on T1-weighted imaging and hyperintense on T2-weighted imaging. No macroscopic fat or microscopic lipid was seen within the lesion, and there were no foci of susceptibility artifact on T1-weighted images. Diffusion-weighted and apparent diffusion coefficient images demonstrated no restricted diffusion. Contrast-enhanced images demonstrated continuous peripheral rim enhancement in the arterial phase and persistent rim enhancement with partial centripetal fill in of enhancement on venous phase images, similar to the pattern seen on CT. Partial left nephrectomy was performed for the suspected solid renal neoplasm. Histopathological assessment was diagnostic of a renal glomus tumor.

Comparative Analysis of MRI Diagnosis and Multislice CT in the Diagnosis of Soft Tissue Tumors in Limbs

The purpose of this study was to compare and analyze the diagnostic applications of multisliecs spiral computed tomography(MSCT)and diffusively-weighted magnetic resonance imaging(MRI)in soft tissue tumors of extremities.A total of 104 patients with primary soft tissue tumors of limbs were selected for MSCT and MRI examination.MSCT values of various tumor types were compared during CT examination.In MRI detection,the diffusion sensitivity factor(b)of diffuse-weighted MRI was 0.500 s/mm2 to avoid as much as possible the bleeding,necrosis,scar and calcification heterogeneity components during operation.The maximum interest point was selected to observe the apparent diffusion coefficient(ADC)between benign and malignant tumors and muscle tissue.Results showed that among 104 patients,36 of patients were malignant and 68 of patients were benign.MSCT examination was consistent with pathology in 45 cases and MRI in 87 cases.In addition,the 59 of patients in MSCT examination and the 17 of patients in MRI were qualitative error and uncertainty.Soft tissue tumors on MSCT showed a low-density mass,and lipoma and cyst were specific.In MRI examination,T1WI and T2WI were significantly different in different types of tumors,while ADC values of malignant tumors were significantly lower than those of benign tumors and muscle tissues(P<0.01).However,there was no significant difference in ADC between benign tumor and muscle tissue(P>0.05).Therefore,MSCT can clearly show the soft tissue tumor lesions of the limbs and identify the relationship between them and the surrounding tissues.However,MSCT cannot make accurate characterization.MRI diffusivity weighted imaging can better differentiate benign from malignant and infer the histological origin of lesions.The detection level of MRI was significantly higher than that of CT,which was more consistent with pathology.Therefore,in the preoperative diagnosis of soft tissue tumors in limbs,MRI diffused-weighted imaging should be the first choice.

基于瘤体及瘤周多参数MRI对乳腺病变良恶性诊断列线图预测模型的构建与评价

目的:建立基于瘤体及瘤周多参数MRI的乳腺病变良恶性鉴别诊断的列线图模型,并验证其预测效能。方法:纳入经病理学检查明确乳腺病变性质的100例患者作为研究对象,所有患者均行核磁共振(MRI)检查和病理检查,根据病理检查结果分为乳腺良性病变组(n=62)和乳腺恶性病变组(n=38)。收集患者临床资料、瘤体各参数、瘤周各参数以及乳腺病变良恶性情况。多因素Logistic回归分析筛选乳腺恶性病变的危险因素并构建列线图预测模型,采用受试者工作特征(ROC)曲线和Hosmer-Lemeshow拟合优度检验验证模型的预测效能及拟合优度;内部验证采用Bootstrap。结果:乳腺恶性病变组病灶直径、平均扩散峰度(MK)、MDp/t、瘤周与瘤体MKp/n高于乳腺良性病变组(均P<0.05);乳腺恶性病变组表观扩散系数(ADC)值、平均扩散率(MD)、非对称磁化转移率(MTRasym)、MKp/t、MDp/n低于乳腺良性病变组(均P<0.05)。多因素分析结果显示,病灶直径、MK、MDp/t、MKp/n升高,ADC值、MD、MTRasym、MKp/t、MDp/n降低是乳腺恶性病变的独立影响因素(均P<0.05)。基于上述独立影响因素构建乳腺恶性病变的列线图预测模型,曲线下面积(AUC)为0.827。Hosmer-Lemeshow拟合优度检验显示P值为0.004。采用Bootstrap法,生成的校准曲线拟合良好。结论:瘤体及瘤周多参数MRI对乳腺病变良恶性鉴别诊断具有重要预测价值,基于乳腺恶性病变的独立影响因素构建的列线图预测效果良好,能直观预测乳腺发生恶性病变的概率。

动态增强MRI定量参数及肿瘤标记物对乳腺癌新辅助化疗疗效的评估价值分析

目的:探索动态增强磁共振成像(DCE-MRI)定量参数及肿瘤标记物对乳腺癌新辅助化疗疗效的评估价值。方法:选取2019年5月至2022年5月在济宁市第一人民医院接受新辅助化疗联合手术干预的75例乳腺癌患者,根据实体瘤疗效评价标准(RECIST)将其分为有效组(54例)和无效组(21例),比较化疗前和化疗后两组患者DCE-MRI定量参数血管外细胞外间隙容积比(V_(e))、速率常数(K_(ep))及容积转换常数(K^(trans))指标与肿瘤标志物癌胚抗原(CEA)、糖类抗原(CA125)及糖类抗原15-3(CA15-3)水平,采用受试者工作特征(ROC)曲线分析各项诊断方式预测效能。结果:化疗后,有效组患者DCEMRI定量参数Ve、K_(ep)及K^(trans)与无效组比较,差异有统计学意义(t=7.237、51.695、16.879,P<0.05)。有效组患者肿瘤标志物CEA、CA125及CA15-3与无效组比较,差异有统计学意义(t=44.201、6.736、6.885,P<0.05)。V_(e)、K_(ep)、K^(trans)、CEA、CA125及CA15-3的6项指标联合预测乳腺癌新辅助化疗疗效ROC曲线下面积(AUC)值为0.979,显著高于各项指标单独检测的AUC值,差异有统计学意义(Z=2.993、2.679、2.510、2.731、3.215、3.071,P<0.05)。结论:肿瘤标记物联合DCE-MRI定量参数可更好预测乳腺癌新辅助化疗疗效情况,间接评估预后。

MRI多模态联合CT对早期卵巢囊(实)性肿瘤的诊断价值

目的探讨早期卵巢囊(实)性肿瘤应用3.0TMRI多模态联合螺旋CT扫描的诊断价值。方法选取我院2019年6月至2021年6月收治早期卵巢囊(实)性肿瘤56例,针对3.0TMRI多模态和螺旋CT表现,并与病理学组织学对照,统计MRI和CT联合检查与病理检查组织学诊断结果,分析不同检查组别诊断的符合率。结果根据病理检查结果,56例早期卵巢(囊)实性肿瘤中,恶性病变25例,良性病变31例,MRI多序列联合CT检查组的诊断率和病理检查结果较为接近。两者联合检查组早期卵巢囊(实)性肿瘤的确诊率96.43%,明显高于单一螺旋CT组83.93%的诊断率和单一MRI组的85.71%诊断率(P<0.05),但单一CT组和MRI组的正确诊断率对比无差异(P>0.05)。两者联合检查早期卵巢囊(实)性肿瘤特异性(96.77%)、敏感性(96.00%)、阴性似然比(96.77%)、阳性似然比(96.00%)高于单一CT(64.52%、64.00%、68.97%、59.26%)与MRI多序列检查(70.97%、72.00%、75.86%、66.67%),差异明显(P<0.05),但CT与MRI多序列检查的特异性、敏感性、阴性与阳性似然比对比无差异(P>0.05)。结论MRI多序列联合螺旋CT检查对于早期卵巢囊(实)性肿瘤比单一检查的确诊率更高,可作为早期卵巢囊(实)性肿瘤诊断比较准确的方法,并且可明确病变累及范围,以指导后续治疗。

增强MRI全域直方图鉴别儿童星形细胞瘤和室管膜瘤的价值

目的深入探究T_(1)全域灰度直方图分析在儿童后颅窝星形细胞瘤与室管膜瘤鉴别诊断中的实用价值。方法本研究对我院数据库中记录的51例经MRI检查及病理确诊的儿童后颅窝肿瘤病例进行了详尽的回顾性剖析。在这51例病例中,星形细胞瘤占26例(男性12例,女性14例),室管膜细胞瘤占25例(男性13例,女性12例)。患者年龄分布在1至12岁之间,平均年龄为(5.5±2.3)岁。我们运用Mazda软件,在两组MR增强T_(1)矢状位图像的每一层肿瘤层面上,精确地勾画出感兴趣区域,并进行全面的灰度全域直方图分析。接着,我们对这两组直方图参数特征进行了详尽的统计学对比,旨在深入剖析各参数在鉴别诊断中的统计意义,以期为儿童后颅窝肿瘤的精确诊断提供更为有效的辅助手段。结果在深入分析通过增强T_(1)全域灰度直方图所提取的九个参数时,我们发现其中四个参数——变异度(Variance)、偏度(Skewness)、第一百分位数(Perc.01%)以及第10百分位数(Perc.10%)在统计学上具有显著差异(P均<0.05),这一发现为我们的研究提供了强有力的数据支撑。特别是Variance这一参数,其灵敏度高达73.3%,特异度也达到了61.9%,且其曲线下面积(AUC)为0.731,展现出优秀的鉴别效能。进一步地,我们确定了Variance的最佳临界值为740.71,这一具体数值为我们的诊断提供了明确的参考标准。结论增强的T_(1)全域灰度直方图分析在儿童后颅窝星形细胞瘤和室管膜瘤的鉴别诊断中,展现出了不容忽视的价值。这一方法不仅能够为医生提供新的诊断视角,而且有望成为一种针对这两种儿童后颅窝肿瘤的有效辅助诊断工具,为我们的医疗实践带来积极的影响。

超声、MRI及肿瘤标志物联合检查对卵巢良恶性肿瘤的鉴别诊断价值

目的探讨超声、MRI及肿瘤标志物联合检查对卵巢良恶性肿瘤的鉴别诊断价值。方法选取80例卵巢肿瘤患者,所有患者均进行超声、MRI及肿瘤标志物检查。根据术后病理检查结果将患者分为良性组(n=50)和恶性组(n=30)。比较两组患者的超声图像特征、MRI影像学表现及肿瘤标志物[糖类抗原19-9(CA19-9)、甲胎蛋白(AFP)、糖类抗原125(CA125)及癌胚抗原(CEA)]水平。分析超声、MRI、肿瘤标志物单独及联合检查对卵巢恶性肿瘤的诊断价值。结果超声检查显示,恶性组中肿瘤形态不规则、回声不均匀、无包膜比例均明显高于良性组,恶性组患者血流阻力指数明显低于良性组,差异均有统计学意义(P﹤0.01)。MRI检查显示,恶性组中肿瘤形态不规则、双侧分布、实性或囊实性成分、边界不清晰比例均明显高于良性组,差异均有统计学意义(P﹤0.01)。恶性组患者CA125、CA19-9、AFP、CEA水平均明显高于良性组,差异均有统计学意义(P﹤0.01)。超声、MRI、肿瘤标志物联合检查诊断卵巢恶性肿瘤的灵敏度、特异度、准确度、阳性预测值和阴性预测值均高于三者单独检查。结论超声、MRI、肿瘤标志物联合检查对卵巢恶性肿瘤具有较高的诊断价值。

合成MRI定量弛豫参数对不同病理类型腮腺肿瘤的鉴别诊断

目的:探讨合成MRI定量弛豫参数对于腮腺常见肿瘤的鉴别诊断。方法:选择2022年5月至2023年1月郑州大学第一附属医院收治的腮腺肿瘤患者59例,均于治疗前行常规MRI和合成MRI,测量肿瘤全域弛豫参数[纵向弛豫时间(T1)、横向弛豫时间(T2)和质子密度(PD)]。比较腮腺多形性腺瘤组、Warthin瘤组和恶性肿瘤组T1、T2和PD。应用Bayes判别分析法对3种腮腺肿瘤进行分类。结果:腮腺多形性腺瘤、Warthin瘤和恶性肿瘤组T1、T2和PD的差异均有统计学意义(P均<0.001)。Warthin瘤组的T1和PD小于腮腺恶性肿瘤和多形性腺瘤组(P<0.05);腮腺多形性腺瘤组的T2大于恶性肿瘤和Warthin瘤组(P<0.05)。采用Bayes判别分析建立的诊断模型自身验证的准确率为61.0%,交叉验证的准确率为57.6%。结论:合成MRI定量弛豫参数对腮腺多形性腺瘤、Warthin瘤和恶性肿瘤的鉴别诊断有一定价值。

Differential study of DCE-MRI parameters in spinal metastatic tumors, brucellar spondylitis and spinal tuberculosis

Objective： In the present study, spinal metastatic tumors, brucellar spondylitis and spinal tuberculosis werequantitatively analyzed using dynamic contrast-enhanced magnetic resonance imaging （DCE-MRI） to assess thevalue of DCE-MRI in the differential diagnosis of these diseases.Methods： Patients with brucellar spondylitis, spinal tuberculosis or a spinal metastatic tumor （30 cases of each）received conventional MRI and DCE-MRI examination. The volume transfer constant （Ktrans）, rate constant （Kep）,extravascular extracellular volume fraction （Ve） and plasma volume fraction （Vp） of the diseased vertebral bodieswere measured on the perfusion parameter map, and the differences in these parameters between the patients werecompared.Results： For pathological vertebrae in cases of spinal metastatic tumor, brucellar spondylitis and spinaltuberculosis, respectively, the Ktrans values （median ＋ quartile pitch） were 0.989±0.014, 0.720±0.011 and0.317±0.005 min-1; the Kep values were 2.898±0.055, 1.327±0.017 and 0.748±0.006 min-1; the Ve values were0.339±0.008, 0.542±0.013 and 0.428±0.018; the Vp values were 0.048±0.008, 0.035±0.004 and 0.028±0.009; thecorresponding H values were 50.25 （for Ktrans）, 52.47 （for Kep）, 48.33 （for Ve） and 46.56 （for Vp）, and all differenceswere statistically significant （two-sided P〈0.05）.Conclusions： The quantitative analysis of DCE-MRI has a certain value in the differential diagnosis of spinalmetastatic tumor, brucellar spondylitis and spinal tuberculosis.

MRI Brain Tumor Segmentation Using 3D U-Net with Dense Encoder Blocks and Residual Decoder Blocks

The main task of magnetic resonance imaging (MRI) automatic brain tumor segmentation is to automaticallysegment the brain tumor edema, peritumoral edema, endoscopic core, enhancing tumor core and nonenhancingtumor core from 3D MR images. Because the location, size, shape and intensity of brain tumors vary greatly, itis very difficult to segment these brain tumor regions automatically. In this paper, by combining the advantagesof DenseNet and ResNet, we proposed a new 3D U-Net with dense encoder blocks and residual decoder blocks.We used dense blocks in the encoder part and residual blocks in the decoder part. The number of output featuremaps increases with the network layers in contracting path of encoder, which is consistent with the characteristicsof dense blocks. Using dense blocks can decrease the number of network parameters, deepen network layers,strengthen feature propagation, alleviate vanishing-gradient and enlarge receptive fields. The residual blockswere used in the decoder to replace the convolution neural block of original U-Net, which made the networkperformance better. Our proposed approach was trained and validated on the BraTS2019 training and validationdata set. We obtained dice scores of 0.901, 0.815 and 0.766 for whole tumor, tumor core and enhancing tumorcore respectively on the BraTS2019 validation data set. Our method has the better performance than the original3D U-Net. The results of our experiment demonstrate that compared with some state-of-the-art methods, ourapproach is a competitive automatic brain tumor segmentation method.

In vivo tumor detection with com bined MR-Photoacoustic-Thermoacoustic imaging

Here,we report a new method using combined magnetic resonance(MR)-Photoacoustic(PA)-Thermoacoustic(TA)imaging techmiques,and demonstrate its unique ability for in vrivo cancer detection using tumor-bearing mice.Circular scanning TA and PA imaging systems were used to recover the dielectric and optical property dist ributions of three colon carcinoma bearing mice While a 7.0-T magnetic resonance imaging(MRI)unit with a mouse body volume coil was utilized for high resolution structural imaging of the same mice.Three plastic tubes flled with soybean sauce were used as fiducial markers for the co-registration of MR,PA and TA images.The resulting fused images provided both enhanced tumor margin and contrast relative to the surrounding normal tissues.In particular,some finger-like protrusions extending into the surrounding tissues were revealed in the MR/TA infused images.These results show that the tissue functional optical and dielectric properties provided by PA and TA images along with the anatomical structure by MRI in one picture make accurate tumor identification easier.This combined MR-PA-TA-imaging strategy has the potential to offer a dinically useful triple-modality tool for accurate cancer detection and for intraoper ative surgical navigation.

Classification of Brain Tumors Using Hybrid Feature Extraction Based on Modified Deep Learning Techniques

According to the World Health Organization(WHO),Brain Tumors(BrT)have a high rate of mortality across the world.The mortality rate,however,decreases with early diagnosis.Brain images,Computed Tomography(CT)scans,Magnetic Resonance Imaging scans(MRIs),segmentation,analysis,and evaluation make up the critical tools and steps used to diagnose brain cancer in its early stages.For physicians,diagnosis can be challenging and time-consuming,especially for those with little expertise.As technology advances,Artificial Intelligence(AI)has been used in various domains as a diagnostic tool and offers promising outcomes.Deep-learning techniques are especially useful and have achieved exquisite results.This study proposes a new Computer-Aided Diagnosis(CAD)system to recognize and distinguish between tumors and non-tumor tissues using a newly developed middleware to integrate two deep-learning technologies to segment brain MRI scans and classify any discovered tumors.The segmentation mechanism is used to determine the shape,area,diameter,and outline of any tumors,while the classification mechanism categorizes the type of cancer as slow-growing or aggressive.The main goal is to diagnose tumors early and to support the work of physicians.The proposed system integrates a Convolutional Neural Network(CNN),VGG-19,and Long Short-Term Memory Networks(LSTMs).A middleware framework is developed to perform the integration process and allow the system to collect the required data for the classification of tumors.Numerous experiments have been conducted on different five datasets to evaluate the presented system.These experiments reveal that the system achieves 97.98%average accuracy when the segmentation and classification functions were utilized,demonstrating that the proposed system is a powerful and valuable method to diagnose BrT early using MRI images.In addition,the system can be deployed in medical facilities to support and assist physicians to provide an early diagnosis to save patients’lives and avoid the high cost of treatments.

Brain Cancer Tumor Classification from Motion-Corrected MRI Images Using Convolutional Neural Network

Detection of brain tumors in MRI images is the first step in brain cancer diagnosis.The accuracy of the diagnosis depends highly on the expertise of radiologists.Therefore,automated diagnosis of brain cancer from MRI is receiving a large amount of attention.Also,MRI tumor detection is usually followed by a biopsy(an invasive procedure),which is a medical procedure for brain tumor classification.It is of high importance to devise automated methods to aid radiologists in brain cancer tumor diagnosis without resorting to invasive procedures.Convolutional neural network(CNN)is deemed to be one of the best machine learning algorithms to achieve high-accuracy results in tumor identification and classification.In this paper,a CNN-based technique for brain tumor classification has been developed.The proposed CNN can distinguish between normal(no-cancer),astrocytoma tumors,gliomatosis cerebri tumors,and glioblastoma tumors.The implemented CNN was tested on MRI images that underwent a motion-correction procedure.The CNN was evaluated using two performance measurement procedures.The first one is a k-fold cross-validation testing method,in which we tested the dataset using k=8,10,12,and 14.The best accuracy for this procedure was 96.26%when k=10.To overcome the over-fitting problem that could be occurred in the k-fold testing method,we used a hold-out testing method as a second evaluation procedure.The results of this procedure succeeded in attaining 97.8%accuracy,with a specificity of 99.2%and a sensitivity of 97.32%.With this high accuracy,the developed CNN architecture could be considered an effective automated diagnosis method for the classification of brain tumors from MRI images.

Advancing Brain Tumor Analysis through Dynamic Hierarchical Attention for Improved Segmentation and Survival Prognosis

Gliomas,the most prevalent primary brain tumors,require accurate segmentation for diagnosis and risk assess-ment.In this paper,we develop a novel deep learning-based method,the Dynamic Hierarchical Attention for Improved Segmentation and Survival Prognosis(DHA-ISSP)model.The DHA-ISSP model combines a three-band 3D convolutional neural network(CNN)U-Net architecture with dynamic hierarchical attention mechanisms,enabling precise tumor segmentation and survival prediction.The DHA-ISSP model captures fine-grained details and contextual information by leveraging attention mechanisms at multiple levels,enhancing segmentation accuracy.By achieving remarkable results,our approach surpasses 369 competing teams in the 2020 Multimodal Brain Tumor Segmentation Challenge.With a Dice similarity coefficient of 0.89 and a Hausdorff distance of 4.8 mm,the DHA-ISSP model demonstrates its effectiveness in accurately segmenting brain tumors.We also extract radio mic characteristics from the segmented tumor areas using the DHA-ISSP model.By applying cross-validation of decision trees to the selected features,we identify crucial predictors for glioma survival,enabling personalized treatment strategies.Utilizing the DHA-ISSP model and the desired features,we assess patients’overall survival and categorize survivors into short,mid,in addition to long survivors.The proposed work achieved impressive performance metrics,including the highest accuracy of 0.91,precision of 0.84,recall of 0.92,F1 score of 0.88,specificity of 0.94,sensitivity of 0.92,area under the curve(AUC)value of 0.96,and the lowest mean absolute error value of 0.09 and mean squared error value of 0.18.These results clearly demonstrate the superiority of the proposed system in accurately segmenting brain tumors and predicting survival outcomes,highlighting its significant merit and potential for clinical applications.