Background and Objective: Primary lung cancer is a lethal and rapidly-developing cancer type and is one of the most leading causes of cancer deaths. Materials and Methods: Statistical methods such as Cox regression are usually used to detect the prognosis factors of a disease. This study investigated survival prediction using machine learning algorithms. The clinical data of 28,458 patients with primary lung cancers were collected from the Surveillance, Epidemiology, and End Results (SEER) database. Results: This study indicated that the survival rate of women with primary lung cancer was often higher than that of men (p < 0.001). Seven popular machine learning algorithms were utilized to evaluate one-year, three-year, and five-year survival prediction The two classifiers extreme gradient boosting (XGB) and logistic regression (LR) achieved the best prediction accuracies. The importance variable of the trained XGB models suggested that surgical removal (feature “Surgery”) made the largest contribution to the one-year survival prediction models, while the metastatic status (feature “N” stage) of the regional lymph nodes was the most important contributor to three-year and five-year survival prediction. The female patients’ three-year prognosis model achieved a prediction accuracy of 0.8297 on the independent future samples, while the male model only achieved the accuracy 0.7329. Conclusions: This data suggested that male patients may have more complicated factors in lung cancer than females, and it is necessary to develop gender-specific diagnosis and prognosis models.
Breast cancer is one of main causes of death for women. Most of the existing survival analyses focus on the features' associations with whether the patients may survive five years or not. The personalized question remains largely unresolved about how long a breast cancer patient will live. This study aims to predict the patient-specific survival time of breast cancer patients. It formulates the personalized question into two machine learning problems. The first problem is the binary classification of whether a patient will live longer than five years or not. The second one is to build a regression model to predict the patient's survival time within five years. The methylome of a breast cancer patient is used for the prediction. A new algorithm Crystall is presented to find the methylomic features for this regression model. Our models perform well in the above two problems, and achieve the mean absolute error (MAE) of about 1 month for predicting how long a breast cancer patient will live within five years. The detected biomarker genes demonstrate close connections with breast cancers.
Breast cancer is one of the most frequently onset female cancer types and represents a major cause of death among women worldwide. Breast cancer is heterogeneous in both molecular characteristics and clinical outcomes for its different molecular subtypes. High-throughput technologies facilitated the fast accumulations of the multiple Omic data for cancer patients. These data sources posed a computational challenge for the efficient integrated multi-Omic analysis. The existing studies usually investigated the differential representation or machine learning problems using single type of Omic data. This study hypothesized that different Omic types contributed complementary information to each other and their integrated analysis may improve the single-Omic models. An efficient logistic regression-based multi-Omic integrated analysis method (ELMO) was proposed to integrate the RNA-seq and DNA methylation data to detect the breast cancer intrinsic subtypes. ELMO achieved the highest accuracy with a smaller number of features, compared with the existing filter and wrapper feature selection methods in this study. The experimental data supported our hypothesis that multi-Omic models outperformed the single-Omic ones.
Melanoma is a lethal skin disease that develops from moles. This study aimed to integrate multimodal data to predict metastatic melanoma, which is highly aggressive and difficult to treat. The proposed EnsembleSKCM method evaluated the prediction performances of long noncoding RNAs (lncRNAs), protein-coding messenger genes (mRNAs) and pathology images (images) for metastatic melanoma. Feature selection was used to screen for metastatic biomarkers in the lncRNA and mRNA datasets. The integrated EnsembleSKCM model was built based on the weighted results of the lncRNA-, mRNA- and image-based models. EnsembleSKCM achieved 0.9444 in the prediction accuracy of metastatic melanoma and outperformed the single-modal prediction models based on the lncRNA, mRNA and image data. The experimental data suggest the importance of integrating the complementary information from the three data modalities. WGCNA was used to analyze the relationship of molecular-level features and image features, and the results show connections between them. Another cohort was used to validate our prediction.
The neurological disorder mild cognitive impairment (MCI) demonstrates minor impacts on the patient’s daily activities and may be ignored as the status of normal aging. But some of the MCI patients may further develop into severe statuses like Alzheimer’s disease (AD). The brain functional connectivity network (BFCN) was usually constructed from the resting-state functional magnetic resonance imaging (rs-fMRI) data. This technology has been widely used to detect the neurodegenerative dementia and to reveal the intrinsic mechanism of neural activities. The BFCN edge was usually determined by the pairwise correlation between the brain regions. This study proposed a weighted voting model of multi-source connectivity networks (MuscNet) by integrating multiple BFCNs of different correlation coefficients. Our model was further improved by removing redundant features. The experimental data demonstrated that different BFCNs contributed complementary information to each other and MuscNet outperformed the existing models on detecting MCI patients. The previous study suggested the existence of multiple solutions with similarly good performance for a machine learning problem. The proposed model MuscNet utilized a weighted voting strategy to slightly outperform the existing studies, suggesting an effective way to fuse multiple base models. The reason may need further theoretical investigations about why different base models contribute to each other for the MCI prediction.
Motivation. The worldwide incidence and mortality rates of melanoma are on the rise recently. Melanoma may develop from benign lesions like skin moles. Easy-to-use mole detection software will help find the malignant skin lesions at the early stage. Results. This study developed mole detection and segmentation software DiaMole using mobile phone images. DiaMole utilized multiple deep learning algorithms for the object detection problem and mole segmentation problem. An object detection algorithm generated a rectangle tightly surrounding a mole in the mobile phone image. Moreover, the segmentation algorithm detected the precise boundary of that mole. Three deep learning algorithms were evaluated for their object detection performance. The popular performance metric mean average precision (mAP) was used to evaluate the algorithms. Among the utilized algorithms, the Faster R-CNN could achieve the best mAP=0.835, and the integrated algorithm could achieve the mAP=0.4228. Although the integrated algorithm could not achieve the best mAP, it can avoid the missing of detecting the moles. A popular Unet model was utilized to find the precise mole boundary. Clinical users may annotate the detected moles based on their experiences. Conclusions. DiaMole is user-friendly software for researchers focusing on skin lesions. DiaMole may automatically detect and segment the moles from the mobile phone skin images. The users may also annotate each candidate mole according to their own experiences. The automatically calculated mole image masks and the annotations may be saved for further investigations.