Survey on recent cancer classification systems for cancer diagnosis

Breast cancer is a significant factor in female mortality. An early cancer diagnosis leads to a reduction in the breast cancer death rate. With the help of a computer-aided diagnosis system, the efficiency increased, and the cost was reduced for the cancer diagnosis. Traditional breast cancer classification techniques are based on handcrafted features techniques, and their performance relies upon the chosen features. They also are very sensitive to different sizes and complex shapes. However, histopathological breast cancer images are very complex in shape. Currently, deep learning models have become an alternative solution for diagnosis, and have overcome the drawbacks of classical classification techniques. Although deep learning has performed well in various tasks of computer vision and pattern recognition, it still has some challenges. One of the main challenges is the lack of training data. To address this challenge and optimize the performance, we have utilized a transfer learning technique which is where the deep learning models train on a task, and then fine-tune the models for another task. We have employed transfer learning in two ways: Training our proposed model first on the same domain dataset, then on the target dataset, and training our model on a different domain dataset, then on the target dataset. We have empirically proven that the same domain transfer learning optimized the performance. Our hybrid model of parallel convolutional layers and residual links is utilized to classify hematoxylin–eosin-stained breast biopsy images into four classes: invasive carcinoma, in-situ carcinoma, benign tumor and normal tissue. To reduce the effect of overfitting, we have augmented the images with different image processing techniques. The proposed model achieved state-of-the-art performance, and it outperformed the latest methods by achieving a patch-wise classification accuracy of 90.5%, and an image-wise classification accuracy of 97.4% on the validation set. Moreover, we have achieved an image-wise classification accuracy of 96.1% on the test set of the microscopy ICIAR-2018 dataset.

Download Full-text

Combining two grading systems: the clinical validity and inter-observer variability of the 1973 and 2004 WHO bladder cancer classification systems assessed in a UK cohort with 15 years of prospective follow-up

World Journal of Urology ◽

10.1007/s00345-020-03180-5 ◽

2020 ◽

Author(s):

Paramananthan Mariappan ◽

Paul Fineron ◽

Marie O’Donnell ◽

Ruth M. Gailer ◽

David J. Watson ◽

...

Keyword(s):

Bladder Cancer ◽

Classification Systems ◽

Cancer Classification ◽

Observer Variability ◽

Clinical Validity ◽

Inter Observer Variability ◽

Grading Systems

Download Full-text

Molecular signatures for inflammation vary across cancer types and correlate significantly with tumor stage, gender and vital status of patients

10.1101/732867 ◽

2019 ◽

Author(s):

Alexandra R. So ◽

Jeong Min Si ◽

David Lopez ◽

Matteo Pellegrini

Keyword(s):

Expression Profiles ◽

Molecular Classification ◽

Classification Systems ◽

Cancer Classification ◽

P Value ◽

Molecular Heterogeneity ◽

Molecular Signatures ◽

Systematic Analysis ◽

Vital Status ◽

Cancer Types

AbstractCancer affects millions of individuals worldwide. One shortcoming of traditional cancer classification systems is that, even for tumors affecting a single organ, there is significant molecular heterogeneity. Precise molecular classification of tumors could be beneficial in personalizing patients’ therapy and predicting prognosis. To this end, here we propose to use molecular signatures to further refine cancer classification. Molecular signatures are collections of genes characterizing particular cell types, tissues or disease. Signatures can be used to interpret expression profiles from heterogeneous samples. Large collections of gene signatures have previously been cataloged in the MSigDB database. We have developed a web-based Signature Visualization Tool (SaVanT) to display signature scores in user-generated expression data. Here we have undertaken a systematic analysis of correlations between inflammatory signatures and cancer samples, to test whether inflammation can differentiate cancer types. Inflammatory response signatures were obtained from MsigDB and SaVanT and a signature score was computed for samples associated with 7 different cancer types. We first identified types of cancers that had high inflammation levels as measured by these signatures. The correlation between signature scores and metadata of these patients (gender, age at initial cancer diagnosis, cancer stage, and vital status) was then computed. We sought to evaluate correlations between inflammation with other clinical parameters and identified four cancer types that had statistically significant association (p-value < 0.05) with at least one clinical characteristic: pancreas adenocarcinoma (PAAD), cholangiocarcinoma (CHOL), kidney chromophobe (KICH), and uveal melanoma (UVM). These results may allow future studies to use these approaches to further refine cancer subtyping and ultimately treatment.

Download Full-text

Outcomes in cancer patients (pts) admitted to a surgical (SICU) or medical ICU (MICU) compared to a newly opened cancer specific ICU (HICU).

Journal of Clinical Oncology ◽

10.1200/jco.2014.32.30_suppl.279 ◽

2014 ◽

Vol 32 (30_suppl) ◽

pp. 279-279

Author(s):

David D. Stenehjem ◽

Meredith Bannon ◽

Jonathan Boltax

Keyword(s):

Overall Survival ◽

Length Of Stay ◽

Cancer Patients ◽

Cancer Diagnosis ◽

Quality Metrics ◽

Primary Objective ◽

Cancer Classification ◽

Icu Mortality ◽

University Of Utah ◽

The University

279 Background: The University of Utah opened a cancer specific ICU (HICU) in 2011 admitting medical and surgical cancer pts. Prior to this, medical cancer pts were admitted to the MICU while the SICU admitted both and medical and surgical pts. The primary objective of this study was to compare the quality metrics of mortality and length of stay (LOS). Methods: Pts with a cancer diagnosis and admitted to the MICU or SICU from 2009-2011 or the HICU from 2011-2013 were evaluated. Pts were stratified by ICU type and the HICU was also analyzed by excluding post-operative pts (HICU-MED). Survival from admission, hospital and ICU mortality, and LOS was assessed. Results: A total of 1,375 pts were included with 259 (19%), 416 (30%), and 700 (51%) pts admitted to the MICU, SICU, and HICU, respectively. The median age was 62 years (range 18-91 years) and 57% (n = 787) were male; no significant differences in age (p = 0.1975) and sex (p = 0.7681) were observed between ICU’s. Cancer classification was oncology (65%, n = 888), hematology (32%, n = 435), and BMT (4%, n = 52). Of all pts admitted to the HICU, 44% (n = 308) were post-operative and excluded from HICU-MED (n = 392). See table for survival and LOS comparisons. Conclusions: No differences in overall survival and hospital morality (ICU and out-of-ICU) were observed comparing pts admitted to the HICU vs SICU and HICU-MED vs MICU. ICU-free days were significantly shorter resulting in shorter hospitalizations for the HICU vs SICU, which may have contributed to the reduced 30-day mortality in pts admitted to the HICU. [Table: see text]

Download Full-text

DA-Based Parameter Optimization of Combined Kernel Support Vector Machine for Cancer Diagnosis

Processes ◽

10.3390/pr7050263 ◽

2019 ◽

Vol 7 (5) ◽

pp. 263 ◽

Cited By ~ 3

Author(s):

Tao Xie ◽

Jun Yao ◽

Zhiwei Zhou

Keyword(s):

Support Vector Machine ◽

Kernel Function ◽

Cancer Diagnosis ◽

Classification Accuracy ◽

Cancer Classification ◽

Support Vector ◽

Svm Classifier ◽

Data Sets ◽

Dragonfly Algorithm ◽

Cancer Data

As is well known, the correct diagnosis for cancer is critical to save patients’ lives. Support vector machine (SVM) has already made an important contribution to the field of cancer classification. However, different kernel function configurations and their parameters will significantly affect the performance of SVM classifier. To improve the classification accuracy of SVM classifier for cancer diagnosis, this paper proposed a novel cancer classification algorithm based on the dragonfly algorithm and SVM with a combined kernel function (DA-CKSVM) which was constructed from a radial basis function (RBF) kernel and a polynomial kernel. Experiments were performed on six cancer data sets from University of California, Irvine (UCI) machine learning repository and two cancer data sets from Cancer Program Legacy Publication Resources to evaluate the validity of the proposed algorithm. Compared with four well-known algorithms: dragonfly algorithm-SVM (DA-SVM), particle swarm optimization-SVM (PSO-SVM), bat algorithm-SVM (BA-SVM), and genetic algorithm-SVM (GA-SVM), the proposed algorithm was able to find the optimal parameters of SVM classifier and achieved better classification accuracy on cancer datasets.

Download Full-text

Cancer classification in the genomic era: five contemporary problems

10.1101/023127 ◽

2015 ◽

Author(s):

Qingxuan Song ◽

Sofia D Merajver ◽

Jun Z Li

Keyword(s):

Biological Diversity ◽

Scientific Discipline ◽

Classification Systems ◽

Cancer Classification ◽

Disease Classification ◽

Major Type ◽

Intratumor Heterogeneity ◽

Tumor Evolution ◽

Large Tumor ◽

Cancer Subtypes

Classification is an everyday instinct as well as a full-fledged scientific discipline. Throughout the history of medicine, disease classification is central to how we develop knowledge, make diagnosis, and assign treatment. Here we discuss the classification of cancer, the process of categorizing cancer subtypes based on their observed clinical and biological features. Traditionally, cancer nomenclature is primarily based on organ location, e.g., "lung cancer" designates a tumor originating in lung structures. Within each organ-specific major type, finer subgroups can be defined based on patient age, cell type, histological grades, and sometimes molecular markers, e.g., hormonal receptor status in breast cancer, or microsatellite instability in colorectal cancer. In the past 15+ years, high-throughput technologies have generated rich new data regarding somatic variations in DNA, RNA, protein, or epigenomic features for many cancers. These data, collected for increasingly large tumor collections, have provided not only new insights into the biological diversity of human cancers, but also exciting opportunities to discover previously unrecognized cancer subtypes. Meanwhile, the unprecedented volume and complexity of these data pose significant challenges for biostatisticians, cancer biologists, and clinicians alike. Here we review five related issues that represent contemporary problems in cancer taxonomy and interpretation. 1. How many cancer subtypes are there? 2. How can we evaluate the robustness of a new classification system? 3. How are classification systems affected by intratumor heterogeneity and tumor evolution? 4. How should we interpret cancer subtypes? 5. Can multiple classification systems coexist? While related issues have existed for a long time, we will focus on those aspects that have been magnified by the recent influx of complex multi-omics data. Ongoing exploration of these problems is essential for data-driven refinement of cancer classification and the successful application of these concepts in precision medicine.

Download Full-text