MS-MambaNet: A lightweight Multi-Scale Mamba architecture for lung cancer classification with explainability

Accurate noninvasive diagnosis of retinal disorders is required for appropriate treatment or precision medicine. This work proposes a multi-stage classification network built on a multi-scale (pyramidal) feature ensemble architecture for retinal image classification using optical coherence tomography (OCT) images. First, a scale-adaptive neural network is developed to produce multi-scale inputs for feature extraction and ensemble learning. The larger input sizes yield more global information, while the smaller input sizes focus on local details. Then, a feature-rich pyramidal architecture is designed to extract multi-scale features as inputs using DenseNet as the backbone. The advantage of the hierarchical structure is that it allows the system to extract multi-scale, information-rich features for the accurate classification of retinal disorders. Evaluation on two public OCT datasets containing normal and abnormal retinas (e.g., diabetic macular edema (DME), choroidal neovascularization (CNV), age-related macular degeneration (AMD), and Drusen) and comparison against recent networks demonstrates the advantages of the proposed architecture's ability to produce feature-rich classification with average accuracy of 97.78%, 96.83%, and 94.26% for the first (binary) stage, second (three-class) stage, and all-at-once (four-class) classification, respectively, using cross-validation experiments using the first dataset. In the second dataset, our system showed an overall accuracy, sensitivity, and specificity of 99.69%, 99.71%, and 99.87%, respectively. Overall, the tangible advantages of the proposed network for enhanced feature learning might be used in various medical image classification tasks where scale-invariant features are crucial for precise diagnosis.

The use of molecular diagnostics in the diagnosis and management of patients with advanced lung cancer has become widespread. Although molecular classification has increasingly been incorporated in the pathologic classification of certain types of human tumors (particularly within the hematologic, glial, and bone/soft tissue malignancies), genetic findings have not been formally incorporated into the pathologic classification of lung cancer, which presently relies solely on the assessment of histologic and immunophenotypic characteristics. Whether molecular classification should be adopted in lung cancer would depend on the diagnostic, prognostic, and predictive impacts of such classification—and whether these impacts confer significant values additive to those derived from the routine histologic and immunophenotypic assessment. We provide a brief overview on the genetics of lung cancer, including adenocarcinoma, squamous cell carcinoma, and neuroendocrine tumors (small cell carcinoma, large cell neuroendocrine carcinoma, and carcinoid tumors). We consider the values of molecular information with some examples, in terms of the current diagnostic, prognostic, and predictive impacts. Finally, we discuss the conceptual and technical challenges of adopting a molecular classification for lung cancer in clinical management for patients. While there are conceptual and technical hurdles to tackle in implementing molecular classification in the pathologic classification of lung cancer, such integrated histologic-molecular diagnosis may allow one to personalize and optimize therapy for patients with advanced lung cancer.

ES 01.02 New Histological Classification

The examination and treatment of lung cancer in its early stages rely heavily on diagnostic imaging technologies. Various imaging procedures used in healthcare, such as computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), chest x-ray, and molecular imaging (MI) approaches, have been extensively studied for the detection and classification of lung cancer. However, some of these systems have limitations, such as their inability to accurately classify lung cancer images, rendering them unsuitable for use in individuals with other illnesses. There is a critical need for the development of a robust and precise method for the early detection of lung cancer. One of the fastest-growing areas in health imaging is deep learning (DL), which involves rapidly developing techniques that utilize texture and clinical-image-based information. DL-based clinical imaging technologies may assist physicians in recognizing and classifying lung cancers more easily and accurately. In this paper, a deep ensemble model is proposed for detecting and classifying lung cancer based on three DL classifiers: MobileNet, ResNet50, and XceptionNet. These DL models are trained on a publicly available lung cancer dataset containing normal, benign, and malignant images. The performance of these baseline classifiers can be measured in terms of accuracy, precision, and recall. The proposed ensemble model accurately detects and classifies lung cancer images. The final results show classification accuracy scores of 28.00%, 13.00%, and 36.00% for the baseline classifiers MobileNet, ResNet50, and XceptionNet, respectively, over the dataset. The proposed ensemble model outperforms the baseline classifiers, achieving a classification accuracy score of 99.00%. Finally, the performance of the proposed ensemble model is compared with those of existing state-of-the-art methods.

Evolving Paradigms in Lung Cancer: Latest Trends in Diagnosis, Management, and Radiopharmaceuticals.

The International Association for the Study of Lung Cancer Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groups in the Forthcoming (Ninth) Edition of the TNM Classification for Lung Cancer

The 8th tumour, node and metastasis (TNM) revision of the staging classification for lung cancer has attempted to clarify the different properties of four distinct clinical presentations: synchronous multiple primary lung cancers; primary tumour with intrapulmonary metastases; pneumonic-type adenocarcinoma and multiple ground glass/lepidic lesions. The respective biological profiles determine different surgical strategies for each category. Accurate preoperative histological confirmation is required to identify synchronous primary tumours which should then be managed as two separate episodes within the confines of patient characteristics. Equally important is the confirmation of intrapulmonary metastasis which indicates a more conservative resection of the smaller lesion. In contrast, pneumonic-type adenocarcinoma may require larger parenchymal resection to achieve anatomical clearance but also symptom relief. Multiple subsolid lesions require a carefully planned strategy of parenchymal sparing and possible observation of less malignant abnormalities. In all the above situations careful clinical, radiological and pathological assessment is imperative to avoid one embarking on inappropriate invasive management. Either when the widespread nature of the disease renders extensive surgery futile or the relative benign or pre-malignant nature of the multiple lesions imply that the risks of surgery outweigh the prognostic benefits.

Lung cancer is the main cause of all cancer diagnoses which reaches 13 percent in the world. According to WHO, lung cancer is the most common type of cancer in men in Indonesia and the fifth most common for all types of cancer in women. This is because the majority of smoking is experienced by men, causing lung cancer. Until now there is no suitable screening method for lung cancer in general. Screening methods that have been recommended for the detection of lung cancer are limited to high-risk patient groups. Whereas the risk of the severity of lung cancer will be even greater if it is not detected early, so as a result lung cancer patients are increasingly difficult to treat. The development of this biomedical technology can be used to assist the early detection process in patients suffering from lung cancer regardless of the criteria for the high risk group first, because if cancer is detected early, the death rate decreases. So in this study the researchers segmented and classified lung cancer using one of the architectures of the CNN algorithm, namely Mobile-Net to facilitate the classification and detection process of lung cancer. As the results of the research conducted by the author on the image of lung cancer with the K-Means segmentation process and classification using the CNN method with the Mobile-Net model, the accuracy is 96.70% and the validation accuracy is 90.45%. This shows that the classification using Mobile-Net with the segmentation process first on the lung cancer image can properly classify the type of disease well

Comprehensive Computational Pathological Image Analysis Predicts Lung Cancer Prognosis

The cancer diagnosis is currently experiencing an alteration of molecular biomarkers using a paradigm for the diagnostic panel. One of the most important genomic datasets showing genetic sequences is the classification of lung cancer. Many research studies have been conducted in this area, but none of them yields satisfied outcomes on account of lower classification accuracy. At first, the data are collected through the lung cancer dataset. Afterward, data are fed into pre-processing. The pre-processing segment removes the noise and enhances the input images utilising Dynamic Context-Sensitive Filtering. The pre-processing output is fed to the feature extraction segment. Here, four statistical features are extracted based on the Force Invariant Improved Feature Extraction Model. The extracting features are fed to the feature segmentation segment. Lung cancer images can segment the ROI region using Adaptive Density-Based Spatial Clustering. After that, the segmentation features are given to ACWGAN for effectively categorizing cancer and non-cancer of the lung cancer Hence, the Hunter–prey optimisation algorithm is employed to optimise the ACWGAN classifier. It classifies the lung cancer images accurately. The LCC-ACWGAN-HPOA-CTI algorithm is activated in MATLAB under the lung cancer database. The performance of the LCC-ACWGAN-HPOA-CTI approach shows higher accuracy and lower computation time than those of the existing methods.

Lung cancer remains a pervasive global health challenge, necessitating precise and efficient classification methods to inform optimal treatment strategies. In this study, we investigate the potential of machine learning algorithms in developing a computer-aided diagnostic tool for early and accurate lung cancer classification, leveraging essential protein biomarkers as predictive features. Our investigation reveals the superiority of Deep Neural Networks (DNNs) in lung cancer classification, achieving an impressive accuracy of 96.91%. This highlights the transformative potential of DNNs in facilitating real-world clinical applications, offering clinicians a powerful tool for accurate diagnosis and prognosis. Additionally, we analyze the performance of Ensemble Methods, including Voting (91.75%) and Bagging (93.81%), as viable alternatives to DNNs. These ensemble techniques demonstrate robust performance, further underlining their utility in lung cancer classification tasks. Furthermore, our study underscores the critical role of hyperparameter tuning, particularly in adjusting parameters such as min child weights and learning rates, in mitigating overfitting and enhancing the generalizability of diverse machine learning models. For instance, Support Vector Machines (SVM) exhibit varying accuracies ranging from 74.23% to 95.88% across different hyperparameter configurations, emphasizing the importance of hyperparameter optimization in maximizing model performance. In summary, our comprehensive comparative analysis sheds light on the efficacy of different machine learning approaches in lung cancer classification. By leveraging advanced algorithms and optimizing model parameters, researchers and clinicians can harness the power of machine learning to advance early detection and personalized treatment strategies for lung cancer patients.

The International Agency for Research on Cancer (IARC) revealed staggering figures, with 19.3 million global cancer cases and 10 million related deaths in that year. Cancer, characterized by abnormal cell growth, can potentially be dangerous with the ability to metastasize. Notably, lung cancer is often detected in an advanced stage due to a lack of awareness and comprehensive medical assessment. Lung cancer usually presents with a late-stage diagnosis. From 60% to 85% of individuals diagnosed with lung cancer show a lack of awareness about their condition. Early diagnosis using an accurate classification method can significantly increase the success of lung cancer diagnosis. To improve predictions, Decision Tree Classifier method was used in lung cancer classification, resulting in a significant increase in accuracy. This study achieved a good level of accuracy, with an accuracy value of 95.16% at a max_depth model depth of 15, and tested in 40 experimental iterations. These results are expected to provide hope for progress in the classification of lung cancer. Keywords: Lung, Cancer, Classification, Decision Tree

For over 50 years, lung cancer has been viewed as an overwhelmingly fatal disease for which radiologic screening was of unproven utility and the pathologist had a limited role differentiating small cell from non-small cell carcinoma and staging resection specimens. The 21st century has given birth to a revolution in the treatment and classification of lung cancer that promises to radically enhance survival of patients. Particularly, the advent of molecular targeted therapies makes identification of the various histologic subtypes, especially adenocarcinoma and squamous cell carcinoma, more important. However, the majorities of lung carcinomas are not resected and are diagnosed using small biopsies or cytology specimens. Such a small biopsied tissue sample is likely to be restricted to only one or two histological tests and could take days because of the time taken for tissue processing, sectioning and staining. The ability to rapidly recognize lung cancer subtypes, with minimal tissue consumption, will thus not only facilitate the diagnostic process, but also enable maximum preservation of tissue samples for subsequent molecular testing for targeted therapy. Hereby, we focus on developing a label-free molecular imaging platform that enables diagnostic imaging of non-small cell lung carcinomas. Mouse lung cancer models were developed by injecting human lung cancer cell lines, including adenocarcinoma (A549) and squamous cell carcinoma (NCI-H226), into the lungs of nude mice (15 mice in each group). Coherent anti-Stokes Raman scattering (CARS) microscopy was used to acquire lung tissue images using intrinsic molecular contrast from symmetric CH2 bonds, thus avoiding tissue consumption or staining. CARS images showed cell nuclei as dark roughly ellipsoidal structures surrounded by brighter extracellular tissues richer in CH2 structures, and were in good correlation with H&E results. By stacking multiple image slices from one z-stack (multiple images acquired at the same field of view but different imaging depths) into a single data structure, 3-D volumes were reconstructed. A computer algorithm was subsequently developed to perform 3-D nuclear segmentation and measurements on these volumes of such disease-related features like nuclear volume, cell-cell distance, etc. The algorithm was developed using superpixel context with artificial neural networks. The measured features were used to develop a classification system for separation of adenocarcinoma from squamous cell carcinoma. Our results showed greater than 97% accuracy and specificity. Therefore, this study shows that the developed 3-D label-free molecular diagnostic platform can accurately delineate cellular structures for classification of non-small cell carcinomas, thus holding substantial potential to provide fast diagnosis while effectively preserving specimens for followed molecular tests. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4055. doi:1538-7445.AM2012-4055

The classification of lung cancer into small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC) is essential for disease prognosis and treatment. For this purpose, we have tried to optimize the use of three tumour markers determined on pleural effusions, to differentiate SCLC from NSCLC by means of a canonic variable, generated by discriminant analysis, including subjects with histologically proven lung cancer. Discriminant analysis was performed by using carcinoembryonic antigen, neuron-specific enolase and tissue polypeptide antigen pleural levels, determined in 65 consecutive and unselected patients, histologically classified as 49 NSCLC and 16 SCLC. To validate the formula generated, a control group of 37 lung cancer patients (10 SCLC and 27 NSCLC), enrolled subsequently, was employed. Applying the discriminant analysis to SCLC and NSCLC patients a good classification was obtained (92% rate of correct classification). The aforementioned formula, applied to the validation group, showed a 92% rate of correct classification. This method, which is rapid, inexpensive and routinely applicable to malignant pleural effusions, may be reliably used to classify lung cancer patients.

Lung cancer significantly contributes to cancer mortality worldwide, and prompt diagnosis is essential for enhancing patient outcomes. Identifying the condition at an early stage remains challenging, particularly in regions with limited medical facilities and experienced radiologists. The paper aims to introduce a fully automated DL system capable of identifying, segmenting, and classifying lung cancer at an early phase. This approach aims to enhance the precision and efficacy of lung cancer screening in resource‐constrained environments. The recommended architecture has three stages: (1) lung preprocessing, employing a bilateral filter to enhance image quality; (2) lung segmentation, utilizing Otsu's thresholding to delineate lung areas; and (3) lung cancer classification, implementing a modified CNN referred to as the P‐Model. The framework was assessed utilizing the publicly available IQ‐OTH/NCCD dataset. The proposed framework exhibited high performance metrics, with lung segmentation accuracy at 96%, classification precision at 97%, sensitivity at 96%, and an F1 score of 96%. Additionally, the Matthews correlation coefficient was recorded at 0.9406. The findings indicate that our paradigm surpasses previous studies across all pertinent evaluation metrics. The proposed deep learning approach shows considerable potential for precise detection and classification of lung cancer, particularly in areas with limited resources. This study significantly advances the field of lung cancer detection and has the capacity to enhance healthcare outcomes and patient care standards globally.