Early Tumor Detection Research Articles

Development and Validation of a Deep Learning Model for Histopathological Slide Analysis in Lung Cancer Diagnosis.

Lung cancer is the leading cause of cancer-related deaths worldwide. Two of the crucial factors contributing to these fatalities are delayed diagnosis and suboptimal prognosis. The rapid advancement of deep learning (DL) approaches provides a significant opportunity for medical imaging techniques to play a pivotal role in the early detection of lung tumors and subsequent monitoring during treatment. This study presents a DL-based model for efficient lung cancer detection using whole-slide images. Our methodology combines convolutional neural networks (CNNs) and separable CNNs with residual blocks, thereby improving classification performance. Our model improves accuracy (96% to 98%) and robustness in distinguishing between cancerous and non-cancerous lung cell images in less than 10 s. Moreover, the model's overall performance surpassed that of active pathologists, with an accuracy of 100% vs. 79%. There was a significant linear correlation between pathologists' accuracy and years of experience (r Pearson = 0.71, 95% CI 0.14 to 0.93, p = 0.022). We conclude that this model enhances the accuracy of cancer detection and can be used to train junior pathologists.

Cancer Screening and Detection: Can AI Change The Game?

“We recognise that cancer detection is one of the main pillars of how to take care of the population today,” remarked Luis Marti-Bonmati, Le Fe Polytechnic and University Hospital, Valencia, Spain, who chaired a session at the European Congress of Radiology (ECR). In a timely conversation, the role of artificial intelligence (AI) in the screening, early detection, and depiction of tumours was explored at the annual ECR congress, which took place in Vienna, Austria, from the 28th February–3rd March 2024.

Deep transfer learning with fuzzy ensemble approach for the early detection of breast cancer

Breast Cancer is a significant global health challenge, particularly affecting women with higher mortality compared with other cancer types. Timely detection of such cancer types is crucial, and recent research, employing deep learning techniques, shows promise in earlier detection. The research focuses on the early detection of such tumors using mammogram images with deep-learning models. The paper utilized four public databases where a similar amount of 986 mammograms each for three classes (normal, benign, malignant) are taken for evaluation. Herein, three deep CNN models such as VGG-11, Inception v3, and ResNet50 are employed as base classifiers. The research adopts an ensemble method where the proposed approach makes use of the modified Gompertz function for building a fuzzy ranking of the base classification models and their decision scores are integrated in an adaptive manner for constructing the final prediction of results. The classification results of the proposed fuzzy ensemble approach outperform transfer learning models and other ensemble approaches such as weighted average and Sugeno integral techniques. The proposed ResNet50 ensemble network using the modified Gompertz function-based fuzzy ranking approach provides a superior classification accuracy of 98.986%.

Brain Tumor Detection Enhanced with Transfer Learning using SqueezeNet

The study introduces the Brain Tumor Detection Transfer Learning Algorithm (BTDTLA), a novel model that employs transfer learning and a comprehensive dataset of brain images. The algorithm makes a significant breakthrough in the precise detection of brain tumors, particularly critical for cases requiring swift intervention. Development and testing of BTDTLA are conducted on MATLAB 2018. The evaluation metrics, including sensitivity, specificity, precision, accuracy, and the Matthews correlation coefficient, highlight the robust performance of BTDTLA, positioning it as a valuable tool for medical practitioners. This underscores the algorithm's potential to advance practices for early and accurate brain tumor detection. The study emphasizes BTDTLA's pivotal role in contributing to the field, underscoring its significance in enhancing medical practices related to brain tumor diagnosis.

An Enhance CNN Model for Brain Tumor Detection from MRI Images

A Convolutional Neural Network (CNN) model for the precise classification and detection of brain tumors from Magnetic Resonance (MR) images is presented in this research. Leveraging the power of deep learning, the proposed CNN model demonstrates remarkable accuracy in distinguishing between different tumor types, aiding in effective diagnosis. The network is trained on a diverse dataset, enabling robust generalization to unseen cases. The proposed method not only enhances the classification performance but also facilitates early tumor detection. Operating on 256 x 256 image inputs, the proposed model boasts a complex architecture featuring six Conv2D layers, batch normalization, five MaxPooling2D layers, six dense layers, and an input layer. Throughout these layers, the ReLU activation function dominates with the final dense layer utilizing the sigmoid function. The potential of sophisticated CNN architectures in medical imaging is highlighted by this study. The outcomes validate the effectiveness of the CNN-based method, advancing computer-aided neuroimaging diagnostic tools and offering a useful tool for medical practitioners. The increased accuracy values demonstrate how well the sophisticated CNN model recognizes brain tumors and point to its potential as a more complex and dependable technique. It also presents a viable avenue for accurate and timely brain tumor identification.

Cáncer de mama: valor pronóstico del estado «TN» versus perfil de expresión génica tumoral

ObjectivesTo analyze the prognostic value of the “TNM” system, and specifically the “TN” status, compared to the tumor gene expression profile, when evaluating the risk of metastatic disease. Materials and methodsRetrospective review of 224 cases of infiltrating breast cancer. Genomic risk was established using the 70-gene gene expression platform (MammaPrint®). The result obtained with the genomic risk assessment has been considered “gold standard”. Results126 cases (56.25%) showed a “low” risk genomic profile and 98 cases (43.75%) showed a “high” risk. Among those with low genomic risk, 26.19% of tumors classified “T2”, “T3” and “T4” were observed. On the contrary, among those with “high” risk, 67.35% of tumors classified as “T1” were observed. Regarding the “N” status, among the “low” risk tumors, 26.98% of cases classified as “N1”, “N2” and “N3” were observed. On the other hand, among those with “high” genomic risk, 77.14% of cases classified as “N0” were observed. Among a total of 131 “T1N0” tumors, 57 (43.51%) showed “high” genomic risk. No significant differences have been observed in terms of risk assessment of metastatic disease (high vs. low) between the various “TN” categories studied. ConclusionsThe evaluation of the risk of metastatic disease using the “TN” system is not accurate. The detection of early tumors does not necessarily mean a good prognosis. On the contrary, the detection of “T2” or “N1” tumors does not imply a poor prognosis in all cases.

A Review of Brain Tumor Detection Techniques Using YOLOv8

Abstract: Brain tumors, whether benign or malignant, pose significant risks, especially in children and young adults, impacting both life expectancy and quality of life. This review addresses the critical importance of early detection and treatment of brain tumors. With a focus on artificial intelligence, particularly object detection using the YOLOv8 model, the paper explores the potential for accurate and efficient brain tumor detection from magnetic resonance images (MRI). The YOLOv8 model is known for its real-time performance, efficiency, and high accuracy, making it a promising tool in the field of medical image analysis. The paper presents a method for brain cancer detection and localization, discusses experimental results, reviews the state-of-theart literature, and outlines future research directions.[1-22]

Brain Tumour Detection and Multi‐Classification Using Advanced Deep Learning Techniques

The early detection and accurate classification of brain tumors are pivotal in enhancing treatment efficacy and patient survival rates. Traditional diagnostic methods, while effective to a degree, are often invasive and reliant on subjective interpretations. This paper introduces a novel approach using advanced deep learning techniques to automate the detection and multi-classification of brain tumors from medical imaging data. Leveraging a comprehensive dataset, we employed state-of-the-art convolutional neural networks (CNNs), incorporating innovative mechanisms such as transfer learning and attention models to refine the accuracy and interpretability of tumor identification and classification. Our methodology encompasses rigorous preprocessing, data augmentation, and a multi-faceted evaluation framework to assess model performance comprehensively. The results indicate a significant improvement over conventional methods and existing machine learning models, showcasing high precision, recall, and F1 scores across multiple tumor types. This research not only contributes to the body of knowledge in medical image analysis but also presents practical implications for integrating advanced AI technologies into clinical diagnostics, thereby potentially transforming patient outcomes through earlier and more accurate diagnoses. The discussion extends to the challenges faced, including dataset imbalances and model deployment in healthcare settings, and proposes directions for future research to further enhance model effectiveness and applicability.

Brain Tumor Detection and Classification Using Deep Learning Models on MRI Scans

INTRODUCTION: The primary goal of artificial intelligence (AI) is to develop computers that exhibit human-like behavior and functionality. Computer-based activities employing artificial intelligence encompass a variety of extra features beyond only pattern detection, planning, and problem resolution. METHODOLOGY: Machines use a set of techniques collectively called "deep learning." Magnetic resonance imaging (MRI) is employed with the use of deep learning methods to develop models that can effectively identify and classify brain cancers. This technique facilitates the rapid and straightforward detection of brain cancers. Brain problems mainly arise from the abnormal multiplication of brain cells, leading to detrimental alterations in brain structure and finally culminating in the development of cancer in the brain, malignant. Early detection of brain tumors along with following effective intervention can reduce mortality rates. This paper proposes convolutional neural network (CNN) architecture to effectively detect brain cancers using magnetic resonance (MR) images. RESULTS: This research further examines several models, including ResNet-50, VGG16, and Inception V3, and compares the proposed architecture and these models. For the efficacy of the models, many measures were evaluated, including accuracy, recall, loss, and area under the curve (AUC). After analyzing several models and comparing them with the suggested model using the specified metrics, it was determined that the proposed model exhibited superior performance compared to the alternative models. Based on an analysis conducted on data from 3265 MR images. CONCLUSION: It was seen that the CNN model exhibited a classification precision of 93.3%. Additionally, the area under the receiver operating characteristic curve (AUC) was determined to be 98.43%, while the recall rate was 91.19%. Furthermore, the model's loss function yielded a value of 0.25. Based on a comparative analysis with other models, it can be inferred that the suggested model is highly reliable in detecting various types of brain cancers at an early stage.

Abstract 4844: Epidemiology of bone and soft-tissue sarcomas: A two-decade analysis of the surveillance, epidemiology, and end results program

Abstract In the United States, ~15000 cases of bone and soft-tissue sarcomas are diagnosed annually, representing roughly 1% of all incident cancers. While several studies have analyzed independent predictors of outcomes for these sarcomas, small samples and confounding bias have limited their ability to establish reliable conclusions. The present analysis provides a comprehensive epidemiological description of patients with any of 16 distinct bone and soft-tissue sarcomas by querying the Surveillance, Epidemiology, and End Results database between 2000 and 2020. Incidence, prevalence, incidence-based mortality, survival rates, and case-fatality ratios were computed. Using patient age, sex, race, and ethnicity, as well as tumor size, grade, and stage as covariates, multivariable survival Cox proportional hazards regression analyses were performed to model hazard ratios (HRs) and identify independent predictors of worse prognosis. Only patients with myosarcoma and fibrous histiocytoma witnessed a decrease in the prevalence of their disease over time (both, p<0.05). The incidence of chondrosarcoma and synovial sarcoma increased in White and Asian patients but decreased in Black individuals (both, p<0.001), while their case-fatality rates remained constant and increased, respectively, regardless of race and ethnicity (all, p<0.05). Case-fatality rates only declined significantly for White and Non-Hispanic patients with solitary fibrous tumor (both, p<0.05). Aside from alveolar soft-part sarcoma and clear cell sarcoma, a 10-year increase in age of patients with bone and soft-tissue sarcomas resulted in 18% to 61% higher likelihood of death (all, p<0.05). For most malignancies, having a tumor size greater than 5 cm yielded a 37% to 189% higher likelihood of death. While incidence trends suggest improvements in diagnostic techniques and early tumor detection, survival and case-fatality trends indicate potential for advancement in treatment for most bone and soft-tissue sarcomas. Citation Format: Marc El Beaino, Rachel Baum, Katherine M. Connors, Karim Masrouha, Santiago A. Lozano-Calderon, Patrick P. Lin. Epidemiology of bone and soft-tissue sarcomas: A two-decade analysis of the surveillance, epidemiology, and end results program [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4844.

Abstract 6376: Establishment of the Consortium for Childhood Cancer Predisposition and the Childhood Cancer Predisposition Study

Abstract Background: Genetic predisposition contributes to a much higher proportion of cancer in children than previously appreciated. This creates a unique opportunity to develop improved methods to identify children at increased risk for cancer, enable early detection or prevention of tumors, and improve cure rates with decreased morbidity. However, several challenges have impeded successful research and systematic approaches to the care of children with cancer predisposition. Methods: We established the Consortium for Childhood Cancer Predisposition (C3P), currently comprised of seven large pediatric institutions in North America. The long-term goal of C3P is to improve outcomes for children with genetic susceptibility to cancer through collaborative research and data sharing. To this end, we have developed the Childhood Cancer Predisposition Study (CCPS), a multicenter registry and biorepository, to achieve three main aims: 1) Establish and maintain the infrastructure for recruitment, participation, collection and sharing of clinical data and biological samples for children with cancer predisposition syndromes (CPS); 2) Define the spectrum of disease in children with CPS and explore the unique characteristics and management approaches of CPS-related tumors compared to their sporadic counterparts; and 3) Evaluate the clinical impact and effectiveness of standard and emerging tumor surveillance strategies. Results: The CCPS opened to enrollment in May, 2021 and is currently recruiting at 3 member sites. To date, 276 primary subjects and 44 family members have been enrolled, 211 in the last year. In total, 37 different cancer predisposition diagnoses are represented among participants, including those that are more prevalent, e.g., Li-Fraumeni syndrome (n=42), DICER1 syndrome (n=27), and PTEN Hamartoma Syndrome (n=21), and those which are very rare, e.g., Rothmund-Thomson syndrome (n=2). There is excellent uptake on obtaining samples, with the CCPS Biorepository at Emory University receiving biospecimens from 261 subjects to date, including constitutional DNA (n=261), frozen peripheral blood mononuclear cells (n=32), and plasma samples for ctDNA analysis (n= 26). Conclusions: C3P has established the infrastructure to facilitate comprehensive collaborative work to improve outcomes for children with cancer predisposition, which represent a wide variety of disorders. Future efforts will include the development of innovative strategies to promote recruitment beyond the primary C3P sites and the support of additional clinical and biological studies of specific CPS. Citation Format: Anita Villani, Garrett M. Brodeur, Lisa R. Diller, Chloe Edgerton, Junne Kamihara, Wendy Kohlmann, Suzanne P. MacFarland, Luke Maese, David Malkin, Kim E. Nichols, Melissa R. Perrino, Sharon E. Plon, Surya Rednam, Christopher C. Porter. Establishment of the Consortium for Childhood Cancer Predisposition and the Childhood Cancer Predisposition Study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6376.

Abstract 6373: Effectiveness of tumor surveillance across pediatric cancer predisposition syndromes

Abstract Background: Pediatric oncology patients are increasingly recognized as having an underlying cancer predisposing syndrome (CPS). For individuals diagnosed with a CPS, surveillance is often recommended to detect new tumors at their earliest and most curable stages. Screening is heavily based on radiologic imaging, however data on the effectiveness and outcomes of surveillance are limited. Methods: Patients with a clinical and/or molecular CPS diagnosis, including 274 children and young adults with 35 CPS at a single specialty pediatric oncology center, were monitored using standardized surveillance protocols over an average of 3 years (range 1 month - 12 years). Patients with hereditary retinoblastoma were excluded as surveillance was completed by a separate team. The cohort included 139 patients (50.7%) with a prior tumor history while 135 (49.3%) had no prior cancer. This study reviewed tumors diagnosed, method of detection, clinical course, and patient outcomes. Individual reports from 1,533 surveillance tests including diagnostic imaging, colonoscopy, and upper endoscopy, were assessed and classified for accuracy. Sensitivity, specificity, positive predictive value (PPV), and negative predictive values (NPV) were calculated for selected surveillance methods, and themes of common strengths or weaknesses of specific imaging modalities were identified. Results: During the study period from January 1, 2009 - September 31, 2021, 35 asymptomatic tumors were detected in 27 patients through surveillance (9.9% of the cohort), while five symptomatic tumors were detected in five patients (1.8%) outside of surveillance. Ten tumors (28.6%) discovered through surveillance were identified on first surveillance imaging. Among these 35 new tumors, 19 (54.3%) were diagnosed in children with a previously treated cancer, six (17.1%) were diagnosed in patients still being treated for cancer, and 10 (28.6%) were diagnosed in those with no prior cancer history. Malignant solid and brain tumors identified through surveillance were more often localized (20 of 24; 83.3%) than similar tumors detected before CPS diagnosis (71 of 125; 56.8%, p ≤0.0001). Of the 24 tumors identified through surveillance and surgically resected, 17 (70.8%) had completely negative margins. When analyzed across all CPS, the sensitivity (96.4%), specificity (99.6%), PPV (95.7%), and NPV (99.6%) of surveillance testing was high, with few false positive (n=6/1,533, 0.4%) or false negative (n=5/1,533, 0.3%) reports. No invasive interventions were performed due to false positive results. Conclusion: Patients with CPS are at increased risk of primary and secondary tumor development with 10% of our cohort developing a tumor in a short time span. Surveillance effectively enabled early detection of new brain and solid tumors across a wide spectrum of CPS which allowed for complete surgical resection and successful treatment in most cases. Citation Format: Melissa R. Perrino, Alise Blake, Cara E. Morin, Leslie Taylor, Rose B. McGee, Sara Lewis, Stacy Hines-Dowell, Arti Pandey, Paige Turner, Manish Kubal, Yin Su, Li Tang, Laura Howell, Lynn W. Harrison, Zachary Abramson, Ann Schechter, Noah D. Sabin, Kim E. Nichols. Effectiveness of tumor surveillance across pediatric cancer predisposition syndromes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6373.

Abstract 1024: Differential gene expression of colorectal cancer biomarkers in Hispanic individuals

Abstract Colorectal cancer (CRC) is a leading cause of cancer death among men and women in Puerto Rico and the United States. About 60-70% of CRC cases are found at advanced stages (III/IV), suggesting limited adherence to available screening methods. Several molecular biomarkers have been identified as discriminatory for CRC (e.g., KRAS, NRAS, MSI), improving CRC management and patient outcomes. This study evaluates the differential expression patterns of DKK3, PKM2, IGFBP2, and SEPT9 genes between matched CRC tumors and adjacent mucosa samples. The secondary outcome was to examine the association between the presence of the selected biomarkers in tissue compared to blood samples. Using a retrospective study design, we examined 112 tissue samples of Hispanic individuals with a pathology-confirmed CRC diagnosis recruited throughout the Puerto Rico Familial Colorectal Cancer Registry (PURIFICAR). We quantified expression levels of target genes by RT-qPCR normalized with endogenous control GAPDH, using TaqMan Gene Expression Assay in the QuantStudio 3 (Applied Biosystems). Wilcoxon signed-rank test was used to calculate differences between gene expression from tumor and adjacent mucosal tissue samples. Additionally, 39 paired blood and tissue samples were analyzed to explore the association between biomarkers in these two types of samples. Protein biomarkers were measured using a commercially available ELISA kit according to manufacturer protocols (Thermo Scientific Federick, MD and Biosource San Diego, CA). mSEPT9 was measured using the Epi proColon® 2.0 CE test. The association between biomarker presence in tissue and blood samples was assessed using the Chi-square test of independence, with biomarker concentrations dichotomized at the median value. Statistical analyses were performed using Stata 17 and R version 4.3.1. DKK3 and IGFBP2 expression levels were higher among CRC tumor samples. However, expression levels of PKM2 and SEPT9 genes were slightly lower. DKK3 was the only target gene significantly different (p<0.05) when comparing overall gene expression between the matched samples. Upon comparison between early and advanced CRC stages, DKK3 expression levels were higher in advanced-stage tumor samples when compared to adjacent mucosal tissue samples (p<0.05). Moreover, a significant association between IGFBG2 presence in tissue compared to blood was detected at a 90% confidence level (p-value<0.10). This association was evident in early-stage CRC cases; there is a relationship between IGFBG2 biomarker presence in tissue and IGFBG2 biomarker presence in blood (p<0.05). Our findings indicate DKK3 and IGFBP2 as potential biomarkers for CRC progression and early detection. The study’s insights into the differential expression patterns of these biomarkers can inform targeted screening strategies. Further research is needed to assess their potential to enhance early tumor detection in Hispanics. Citation Format: Elba V. Caraballo-Rivera, Hilmaris Centeno-Girona, Gabriel De Jesus-Astacio, Camille Zenon-Melendez, Brenda Torres-Velasquez, Madeline Martir-Ocasio, Marcia R. Cruz-Correa. Differential gene expression of colorectal cancer biomarkers in Hispanic individuals [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1024.

Advancement in the treatment, genetic understanding, and diagnostic approaches for retinoblastoma

The most common intraocular tumor in children is retinoblastoma. In developed countries, there has been an impressive rise in the survival rate and visual outcome of patients with retinoblastoma. This is clarified by developments in retinoblastoma treatment and early detection of tumors. The primary therapy remedy for intraocular retinoblastoma is now chemotherapy along with adjuvant consolidative treatment, instead of external beam radiation. Likewise, prophylactic chemotherapy is now allowed to treat potential micro metastases in enucleated eyes based on the identification of high-risk histopathological factors. Extra ocular retinoblastoma still has a poor future outcome; chances of survival have been reported to be between 50% and 70%. Retinoblastoma patients' overall survival is still struggling in developing countries, mostly because of delayed presentation and greater risk of extra ocular illness while compared to with the developed world, where intraocular disease contributes to majority of cases. To enhance the survival rate for those with retinoblastoma in developing countries, greater priority must be given to early detection of tumors. We provide a summary of the latest clinical management of retinoblastoma in this article.

Ingestible Artificial Urinary Biomarker Probes for Urine Test of Gastrointestinal Cancer.

Although colorectal cancer diagnosed at an early stage shows high curability, methods simultaneously possessing point-of-care testing ability and high sensitivity are limited. Here, an orally deliverable biomarker-activatable probe (termed as HATS) for early detection of orthotopic tumors via remote urinalysis is presented. To enable its oral delivery to the colon, HATS is designed to have remarkable resistance to acidity and digestive enzymes in the stomach and small intestine and negligible intestinal absorption. Upon reaction with a cancer biomarker in the colon segment, HATS releases a small fragment of tetrazine that can transverse the intestinal barrier, enter blood circulation, and ultimately undergo renal clearance to urine. Subsequently, the urinary tetrazine fragment is detected by bioorthogonal reaction with trans-cyclooctene-caged resorufin (TCO-Reso) to afford a rapid and specific fluorescence enhancement of TCO-Reso. Such signal readout is correlated with the urinary tetrazine concentration and thus measures the level of cancer biomarkers in the colon. HATS-based optical urinalysis detects orthotopic colon tumors two weeks earlier than clinical serological tests and can be developed to a point-of-care paper test. Thereby, HATS-based urinalysis provides a non-invasive and sensitive approach to cancer screening at low-resource settings.

Advancing brain tumor classification through MTAP model: an innovative approach in medical diagnostics

The early diagnosis of brain tumors is critical in the area of healthcare, owing to the potentially life-threatening repercussions unstable growths within the brain can pose to individuals. The accurate and early diagnosis of brain tumors enables prompt medical intervention. In this context, we have established a new model called MTAP to enable a highly accurate diagnosis of brain tumors. The MTAP model addresses dataset class imbalance by utilizing the ADASYN method, employs a network pruning technique to reduce unnecessary weights and nodes in the neural network, and incorporates Avg-TopK pooling method for enhanced feature extraction. The primary goal of our research is to enhance the accuracy of brain tumor type detection, a critical aspect of medical imaging and diagnostics. The MTAP model introduces a novel classification strategy for brain tumors, leveraging the strength of deep learning methods and novel model refinement techniques. Following comprehensive experimental studies and meticulous design, the MTAP model has achieved a state-of-the-art accuracy of 99.69%. Our findings indicate that the use of deep learning and innovative model refinement techniques shows promise in facilitating the early detection of brain tumors. Analysis of the model’s heat map revealed a notable focus on regions encompassing the parietal and temporal lobes.Graphical Grad-CAM heat map visualization results

Adaptive Machine Learning Approach for Importance Evaluation of Multimodal Breast Cancer Radiomic Features

Breast cancer holds the highest diagnosis rate among female tumors and is the leading cause of death among women. Quantitative analysis of radiological images shows the potential to address several medical challenges, including the early detection and classification of breast tumors. In the P.I.N.K study, 66 women were enrolled. Their paired Automated Breast Volume Scanner (ABVS) and Digital Breast Tomosynthesis (DBT) images, annotated with cancerous lesions, populated the first ABVS+DBT dataset. This enabled not only a radiomic analysis for the malignant vs. benign breast cancer classification, but also the comparison of the two modalities. For this purpose, the models were trained using a leave-one-out nested cross-validation strategy combined with a proper threshold selection approach. This approach provides statistically significant results even with medium-sized data sets. Additionally it provides distributional variables of importance, thus identifying the most informative radiomic features. The analysis proved the predictive capacity of radiomic models even using a reduced number of features. Indeed, from tomography we achieved AUC-ROC 89.9%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$89.9\\%$$\\end{document} using 19 features and 92.1%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$92.1\\%$$\\end{document} using 7 of them; while from ABVS we attained an AUC-ROC of 72.3%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$72.3\\%$$\\end{document} using 22 features and 85.8%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$85.8\\%$$\\end{document} using only 3 features. Although the predictive power of DBT outperforms ABVS, when comparing the predictions at the patient level, only 8.7% of lesions are misclassified by both methods, suggesting a partial complementarity. Notably, promising results (AUC-ROC ABVS-DBT 71.8%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$71.8\\%$$\\end{document}-74.1%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$74.1\\%$$\\end{document}) were achieved using non-geometric features, thus opening the way to the integration of virtual biopsy in medical routine.

Development and validation of a platform based on 3D stochastic sensors for fast screening of whole blood and brain tissue for brain cancer

A novel stochastic platform was assembled by integrating a sensor based on single-wall carbon nanotube paste and ZnO modified with (Z)-N-[2-(4-hydroxyphenyl) ethyl] octadec-9-enamide for the enantioanalysis of two amino acids in whole blood and brain tissue tumor samples. The stochastic platform was utilized to obtain wide concentration ranges for the L-tryptophan, D-tryptophan, L-glutamine, and D-glutamine. The reported sensitivities for both isomers of glutamine exhibited a comparable magnitude order of 107. In contrast, among the four enantiomers, D-Trp displayed the highest sensitivity, measuring at 4.43 × 108. The limits of quantification were in the order of 1.00 × 10−18 mol•L−1. The efficacy of the platform was demonstrated through successful testing on biological samples. The advantages of using this type of platform for such analysis are the possibility of identifying and quantifying in one run different types of biomarkers as well as other substances of biological importance usually requested by medical doctors to be quantified, which can take the test from a fast-screening method to a fast diagnostic method. The analysis is performed within minutes, which is very important for illnesses such as brain cancer that can develop quickly in the body. The early detection of brain tumors can save lives. Furthermore, the portability of the developed detection technology makes it suitable for on-site analysis.

The unique risk factor profile of triple-negative breast cancer: a comprehensive meta-analysis.

Triple-negative breast cancer (TNBC) has a poor prognosis compared with other breast cancer subtypes. This systematic review and meta-analysis examines whether known risk factors for breast cancer are also associated with TNBC in adult women. EMBASE, Medline, SCOPUS, and gray literature were queried with no limit on the date or language of publication. The exposures of interest included parity, breastfeeding, duration of breastfeeding, age at menarche, age at first live birth, oral contraceptive (OC) use, duration of OC use, use of menopausal hormone therapy (MHT), family history, body mass index (BMI), alcohol use, smoking, and breast density. The main outcome of interest was TNBC. Study quality was determined using the Newcastle-Ottawa scale for case control studies and cohort studies. We estimated weighted odds ratios from random effects models to study the exposure-outcome associations. Protocol was registered under the number: PROSPERO 2021 CRD42021254594. Thirty-three studies were included. Family history, longer duration of OC use, and higher breast density were significantly associated with increased risk for TNBC, whereas later age at menarche, later age at first birth, and breastfeeding were protective against TNBC. Parity, MHT, alcohol, smoking, and BMI were not significantly associated with TNBC overall, but higher parity was associated with higher risk among Black women. Our findings highlight that TNBC has a distinct risk factor profile compared with overall breast cancer. This can be the foundational work in identification of actionable TNBC risk factors to improve prevention and early detection of these poor prognosis breast tumors.

Brain Tumor Detection: Transfer Learning Approach with VGG-16

Brain tumors have a profound impact on individuals worldwide, necessitating early detection to improve patient outcomes. This study explores the potential of deep learning in facilitating early tumor detection to aid medical professionals in timely intervention and personalized treatment strategies. While current tumor detection methodologies rely on traditional imaging techniques and manual analysis, they may suffer from limitations in accuracy and efficiency. To address these challenges, we propose a novel method using the VGG-16 architecture, a powerful Convolutional Neural Network (CNN) model, with pre-trained weights for feature extraction. By harnessing the capability of VGG-16, our model effectively discerns tumor patterns from brain MRI scans. The dataset utilized for training and evaluation consists of Brain MRI Images for Brain Tumor Detection from Kaggle, featuring a diverse range of annotated scans. During the evaluation, we focus on the primary performance metric of accuracy, which measures the model's ability to precisely classify tumor and non-tumor cases. Through extensive experimentation and analysis, we critically assess the model's performance and its potential clinical applicability