Early prediction of lung cancer based on the combination of trace element analysis in urine and an Adaboost algorithm

Warburg Effect Revisited: Embodiment of Classical Biochemistry and Organic Chemistry. Current State and Prospects

The Nobel Prize Winner (1931) Dr. Otto H. Warburg had established that the primary energy source of the cancer cell is aerobic glycolysis (the Warburg effect). He also postulated the hypothesis about "the prime cause of cancer", which is a matter of debate nowadays. Contrary to the hypothesis, his discovery was recognized entirely. However, the discovery had almost vanished in the heat of battle about the hypothesis. The prime cause of cancer is essential for the prevention and diagnosis, yet the effects that influence tumor growth are more important for cancer treatment. Due to the Warburg effect, a large amount of data has been accumulated on biochemical changes in the cell and the organism as a whole. Due to the Warburg effect, the recovery of normal biochemistry and oxygen respiration and the restoration of the work of mitochondria of cancer cells can inhibit tumor growth and lead to remission. Here, we review the current knowledge on the inhibition of abnormal glycolysis, neutralization of its consequences, and normalization of biochemical parameters, as well as recovery of oxygen respiration of a cancer cell and mitochondrial function from the point of view of classical biochemistry and organic chemistry.

LightGBM: an efficient and accurate method for predicting pregnancy diseases

As machine learning is becoming the fashion in disease prediction while no prediction model has performed very efficiently and accurately on predicting pregnancy diseases up to now, it's necessary to compare several common machine learning methods' performance on pregnancy diseases prediction and select out the best one. The data of two common pregnancy complications, pregnancy-induced hypertension (PIH) and Intrahepatic cholestasis of pregnancy (ICP), based on various maternal characteristics measured in patients' routine blood examination in 10-19 weeks of gestation are considered to be suitable to be learned. This is a retrospective study of 320 healthy pregnancies in 10-19 weeks, with 149 patients who subsequently developed PIH and 250 patients who subsequently developed ICP. Nine machine learning methods were used to predict PIH and ICP and their performance was compared via 8 evaluation indexes. Finally, the light Gradient Boosting Machine (lightGBM) is considered to be the best method to predict gestational diseases.Impact statementWhat is already known on this subject? As a kind of commonly used method in disease prediction, machine learning could be applied to clinical data for developing robust risk models and many achievements have been made. Also, machine learning can be used to predict pregnancy diseases. Although some machine learning methods have been used for screening gestational diseases, methods based on simple theories, such as logistic regression and decision tree, are frequently used. They don't always have a very satisfactory prediction results. Besides, only a few types of pregnancy diseases can be predicted.What do the results of this study add? LightGBM has the best prediction results of PIH and ICP among 9 machine learning methods in this study. It can predict PIH (AUC = 81.72%) with a sensitivity of 70.59%, and ICP (AUC = 95.91%) with a sensitivity of 97.91%.What are the implications of these findings for clinical practice and/or further research? A new model has been developed for effective first-trimester screening for two common pregnancy diseases, PIH and ICP. This lightGBM model can be used in relative hospitals and population of the research, and provide references for doctors' diagnosis and treatment of pregnant women. In further research, the predicted effect of lightGBM on daily practice and other pregnancy diseases such as pregnancy diabetes, will be verified.

Advances in lung cancer biomarkers: The role of (metal-) metabolites and selenoproteins

Lung cancer (LC) is the second most common cause of death in men after prostate cancer, and the third most recurrent type of tumor in women after breast and colon cancers. Unfortunately, when LC symptoms begin to appear, the disease is already in an advanced stage and the survival rate only reaches 2%. Thus, there is an urgent need for early diagnosis of LC using specific biomarkers, as well as effective therapies and strategies against LC. On the other hand, the influence of metals on more than 50% of proteins is responsible for their catalytic properties or structure, and their presence in molecules is determined in many cases by the genome. Research has shown that redox metal dysregulation could be the basis for the onset and progression of LC disease. Moreover, metals can interact between them through antagonistic, synergistic and competitive mechanisms, and for this reason metals ratios and correlations in LC should be explored. One of the most studied antagonists against the toxic action of metals is selenium, which plays key roles in medicine, especially related to selenoproteins. The study of potential biomarkers able to diagnose the disease in early stage is conditioned by the development of new analytical methodologies. In this sense, omic methodologies like metallomics, proteomics and metabolomics can greatly assist in the discovery of biomarkers for LC early diagnosis.

Computational Approaches in Theranostics: Mining and Predicting Cancer Data

The ability to understand the complexity of cancer-related data has been prompted by the applications of (1) computer and data sciences, including data mining, predictive analytics, machine learning, and artificial intelligence, and (2) advances in imaging technology and probe development. Computational modelling and simulation are systematic and cost-effective tools able to identify important temporal/spatial patterns (and relationships), characterize distinct molecular features of cancer states, and address other relevant aspects, including tumor detection and heterogeneity, progression and metastasis, and drug resistance. These approaches have provided invaluable insights for improving the experimental design of therapeutic delivery systems and for increasing the translational value of the results obtained from early and preclinical studies. The big question is: Could cancer theranostics be determined and controlled in silico? This review describes the recent progress in the development of computational models and methods used to facilitate research on the molecular basis of cancer and on the respective diagnosis and optimized treatment, with particular emphasis on the design and optimization of theranostic systems. The current role of computational approaches is providing innovative, incremental, and complementary data-driven solutions for the prediction, simplification, and characterization of cancer and intrinsic mechanisms, and to promote new data-intensive, accurate diagnostics and therapeutics.

The Metallome of Lung Cancer and its Potential Use as Biomarker

Carcinogenesis is a very complex process in which metals have been found to be critically involved. In this sense, a disturbed redox status and metal dyshomeostasis take place during the onset and progression of cancer, and it is well-known that trace elements participate in the activation or inhibition of enzymatic reactions and metalloproteins, in which they usually participate as cofactors. Until now, the role of metals in cancer have been studied as an effect, establishing that cancer onset and progression affects the disturbance of the natural chemical form of the essential elements in the metabolism. However, it has also been studied as a cause, giving insights related to the high exposure of metals giving a place to the carcinogenic process. On the other hand, the chemical species of the metal or metallobiomolecule is very important, since it finally affects the biological activity or the toxicological potential of the element and their mobility across different biological compartments. Moreover, the importance of metal homeostasis and metals interactions in biology has also been demonstrated, and the ratios between some elements were found to be different in cancer patients; however, the interplay of elements is rarely reported. This review focuses on the critical role of metals in lung cancer, which is one of the most insidious forms of cancer, with special attention to the analytical approaches and pitfalls to extract metals and their species from tissues and biofluids, determining the ratios of metals, obtaining classification profiles, and finally defining the metallome of lung cancer.

Metal dyshomeostasis based biomarkers of lung cancer using human biofluids

Metals, ratios, interactions and species in serum, urine and bronchoalveolar lavage fluid as biomarkers of lung cancer.

Applying the Temporal Abstraction Technique to the Prediction of Chronic Kidney Disease Progression

Chronic kidney disease (CKD) has attracted considerable attention in the public health domain in recent years. Researchers have exerted considerable effort in attempting to identify critical factors that may affect the deterioration of CKD. In clinical practice, the physical conditions of CKD patients are regularly recorded. The data of CKD patients are recorded as a high-dimensional time-series. Therefore, how to analyze these time-series data for identifying the factors affecting CKD deterioration becomes an interesting topic. This study aims at developing prediction models for stage 4 CKD patients to determine whether their eGFR level decreased to less than 15 ml/min/1.73m² (end-stage renal disease, ESRD) 6 months after collecting their final laboratory test information by evaluating time-related features. A total of 463 CKD patients collected from January 2004 to December 2013 at one of the biggest dialysis centers in southern Taiwan were included in the experimental evaluation. We integrated the temporal abstraction (TA) technique with data mining methods to develop CKD progression prediction models. Specifically, the TA technique was used to extract vital features (TA-related features) from high-dimensional time-series data, after which several data mining techniques, including C4.5, classification and regression tree (CART), support vector machine, and adaptive boosting (AdaBoost), were applied to develop CKD progression prediction models. The results revealed that incorporating temporal information into the prediction models increased the efficiency of the models. The AdaBoost+CART model exhibited the most accurate prediction among the constructed models (Accuracy: 0.662, Sensitivity: 0.620, Specificity: 0.704, and AUC: 0.715). A number of TA-related features were found to be associated with the deterioration of renal function. These features can provide further clinical information to explain the progression of CKD. TA-related features extracted by long-term tracking of changes in laboratory test values can enable early diagnosis of ESRD. The developed models using these features can facilitate medical personnel in making clinical decisions to provide appropriate diagnoses and improved care quality to patients with CKD.

Determination of urinary 5-hydroxyindoleacetic acid as a metabolomics in gastric cancer

PURPOSE

The aim of this paper is to study urinary5-hydroxyindoleacetic acid (5-HIAA) in gastric cancer patients with a biochemical method and compare this metabolite with normal control and individuals with chronic gastritis.

MATERIALS AND METHODS

The subjects were 48 histologically proven gastric adenocarcinoma patients. They were 10 women and 38 men with mean age of 63.73 years. For determination of urinary excretion of 5-HIAA, a biochemical method was applied. According to kit protocol, the patients' fresh urine was added to the reagent material, and the color of the sediment that was the result of interaction between 5-HIAA and the mercury salt was compared with the standard colorimetric plate of the kit. The same method was also performed for a group of 47 patients with chronic gastritis and also a group of 50 normal individuals (age and sex matched).

RESULTS

Urinary 5-HIAA was significantly higher in gastric cancer patients compared to individuals with chronic gastritis and normal controls (P value <0.001), but no association was detected in urinary 5-HIAA based on age, sex, or site of tumor and tumor grade in gastric cancer patients group. Also, no significant difference was noted in 5-HIAA excretion between chronic gastritis and normal control groups.

CONCLUSION

Urinary excretion of 5-HIAA is significantly higher in the gastric cancer patients in comparison with that of chronic gastritis patients or normal individuals. So, this test could be regarded as a tumor marker in conjunction with other modalities in diagnosis of gastric cancer.

Continuous wavelet transform-based feature selection applied to near-infrared spectral diagnosis of cancer

Spectrum is inherently local in nature since it can be thought of as a signal being composed of various frequency components. Wavelet transform (WT) is a powerful tool that partitions a signal into components with different frequency. The property of multi-resolution enables WT a very effective and natural tool for analyzing spectrum-like signal. In this study, a continuous wavelet transform (CWT)-based variable selection procedure was proposed to search for a set of informative wavelet coefficients for constructing a near-infrared (NIR) spectral diagnosis model of cancer. The CWT provided a fine multi-resolution feature space for selecting best predictors. A measure of discriminating power (DP) was defined to evaluate the coefficients. Partial least squares-discriminant analysis (PLS-DA) was used as the classification algorithm. A NIR spectral dataset associated to cancer diagnosis was used for experiment. The optimal results obtained correspond to the wavelet of db2. It revealed that on condition of having better performance on the training set, the optimal PLS-DA model using only 40 wavelet coefficients in 10 scales achieved the same performance as the one using all the variables in the original space on the test set: an overall accuracy of 93.8%, sensitivity of 92.5% and specificity of 96.3%. It confirms that the CWT-based feature selection coupled with PLS-DA is feasible and effective for constructing models of diagnostic cancer by NIR spectroscopy.

Metabolomics specificity of tuberculosis plasma revealed by (1)H NMR spectroscopy

Tuberculosis (TB) is a communicable disease of major global importance and causes metabolic disorder of the patients. In a previous study, we found that the plasma metabolite profile of TB patients differs from that of healthy control subjects based on nuclear magnetic resonance (NMR) spectroscopy. In order to evaluate the TB specificity of the metabolite profile, a total of 110 patients, including 40 with diabetes, 40 with malignancy, and 30 with community-acquired pneumonia (CAP), assessed by NMR spectroscopy, and compared to those of patients with TB. Based on the orthogonal partial least-squares discriminant analysis (OPLS-DA), the metabolic profiles of these diseases were significant different, as compared to the healthy controls and TB patients, respectively. The score plots of the OPLS-DA model demonstrated that TB was easily distinguishable from diabetes, CAP and malignancy. Plasma levels of ketone bodies, lactate, and pyruvate were increased in TB patient compared to healthy control, but lower than CAP and malignancy. We conclude that the metabolic profiles were TB-specific and reflected MTB infection. Our results strongly support the NMR spectroscopy-based metabolomics could contribute to an improved understanding of disease mechanisms and may offer clues to new TB clinic diagnosis and therapies.

Classification models based on the level of metals in hair and nails of laryngeal cancer patients: diagnosis support or rather speculation?

Several larynx cancer prediction models were built and each was weighted according to their performance.

Metabolomic studies of human gastric cancer: Review

Metabolomics is a field of study in systems biology that involves the identification and quantification of metabolites present in a biological system. Analyzing metabolic differences between unperturbed and perturbed networks, such as cancerous and non-cancerous samples, can provide insight into underlying disease pathology, disease prognosis and diagnosis. Despite the large number of review articles concerning metabolomics and its application in cancer research, biomarker and drug discovery, these reviews do not focus on a specific type of cancer. Metabolomics may provide biomarkers useful for identification of early stage gastric cancer, potentially addressing an important clinical need. Here, we present a short review on metabolomics as a tool for biomarker discovery in human gastric cancer, with a primary focus on its use as a predictor of anticancer drug chemosensitivity, diagnosis, prognosis, and metastasis.

The diagnostics of diabetes mellitus based on ensemble modeling and hair/urine element level analysis

The aim of the present work focuses on exploring the feasibility of analyzing the relationship between diabetes mellitus and several element levels in hair/urine specimens by chemometrics. A dataset involving 211 specimens and eight element concentrations was used. The control group was divided into three age subsets in order to analyze the influence of age. It was found that the most obvious difference was the effect of age on the level of zinc and iron. The decline of iron concentration with age in hair was exactly consistent with the opposite trend in urine. Principal component analysis (PCA) was used as a tool for a preliminary evaluation of the data. Both ensemble and single support vector machine (SVM) algorithms were used as the classification tools. On average, the accuracy, sensitivity and specificity of ensemble SVM models were 99%, 100%, 99% and 97%, 89%, 99% for hair and urine samples, respectively. The findings indicate that hair samples are superior to urine samples. Even so, it can provide more valuable information for prevention, diagnostics, treatment and research of diabetes by simultaneously analyzing the hair and urine samples.

Confronting deep uncertainties in risk analysis

How can risk analysts help to improve policy and decision making when the correct probabilistic relation between alternative acts and their probable consequences is unknown? This practical challenge of risk management with model uncertainty arises in problems from preparing for climate change to managing emerging diseases to operating complex and hazardous facilities safely. We review constructive methods for robust and adaptive risk analysis under deep uncertainty. These methods are not yet as familiar to many risk analysts as older statistical and model-based methods, such as the paradigm of identifying a single "best-fitting" model and performing sensitivity analyses for its conclusions. They provide genuine breakthroughs for improving predictions and decisions when the correct model is highly uncertain. We demonstrate their potential by summarizing a variety of practical risk management applications.

Prediction of type-2 diabetes based on several element levels in blood and chemometrics

The present study was designed to evaluate the levels of eight elements including lithium, zinc, chromium, copper, iron, manganese, nickel and vanadium in whole blood of type-2 diabetes patients, to compare them with age-matched healthy controls and to investigate the feasibility of combining them with an ensemble model for diagnosing purpose. A dataset involving 158 samples, among which 105 were taken from healthy adults and the remaining 53 from patients with type-2 diabetes, was collected. All samples were split into the training set and the test set with the equal size. Based on a simple variable selection, two elements, i.e., chromium and iron, are also picked out as the most important elements. Three kinds of algorithms, i.e., fisher linear discriminate analysis (FLDA), support vector machine (SVM) and decision tree (DT), were used for constructing member models. The best ensemble classifiers constructed on the training set were validated on the independent test set, and the prediction results were compared with those from clinical diagnostics on the same subjects. The results reveal that almost all ensemble classifiers exhibit similar performance, implying that these elements coupled with an appropriate ensemble classifier can serve as a valuable tool of diagnosing diabetes type-2.

Effects of chronic chromium(vi) exposure on blood element homeostasis: an epidemiological study

One hundred chromate production workers chronically exposed to low-level of hexavalent chromium [Cr(vi)] and eighty healthy individuals free from Cr exposure were recruited to the study. Personal sampling of airborne Cr was conducted and Cr content was quantified by Flame Atomic Absorption Spectrometry (FAAS). At the end of the sampling shift, blood samples were collected and element concentrations were measured by inductively coupled plasma mass spectrometry (ICP-MS) for Cr, Cd, Cu, Mo and Se and inductively coupled plasma atomic emission spectrometry (ICP-AES) for Ca, Fe, Mg and Zn. According to our results, 90% of the chromate production workers were exposed to airborne Cr in a concentration lower than 50 μg m(-3), which is the threshold limit value recommended by the American Conference of Governmental Industrial Hygienists and Chinese Ministry of Health. After Cr(vi) exposure, a significant increase in blood Cr, Cd, Fe, Mg, Mo, Se and Zn concentrations was observed, as well as a significant decrease in Ca concentration. A decrease in blood Cu was only observed among female workers. Blood Cr concentrations of the exposed workers (median = 15.68 ng mL(-1)) was four times higher than that of the controls (median = 3.03 ng mL(-1)), and significantly correlated with airborne Cr (r = 0.568, P<0.001). In addition, the inter-element correlations exhibited significant differences between the two groups. Our findings of the related health effects suggested that the underlying mechanisms of chronic Cr(vi) exposure on blood element homeostasis might be partly explained by oxidative stress in the body, dysfunction of Fe metabolism and renal injury.

Screening of prostate cancer by analyzing trace elements in hair and chemometrics

Prostate cancer is the most common non-cutaneous malignancy and second leading cause of cancer mortality in men. The principle goal of this study was explore the feasibility of applying boosting coupled with trace element analysis of hair, for accurately distinguishing prostate cancer from healthy person. A total of 113 subjects containing 55 healthy men and 58 prostate cancers were collected. Based on a special index of variable importance and a forward selection scheme, only nine elements (i.e., Zn, Cr, Mg, Ca, Al, P, Cd, Fe, and Mo) were picked out from 20 candidate elements for modeling the relationship. As a result, an ensemble classifier consisting of only eight decision stumps achieved an overall accuracy of 98.2%, a sensitivity of 100%, and a specificity of 96.4% on the independent test set while all subjects on the training set are classified correctly. It seems that integrating boosting and element analysis of hair can serve as a valuable tool of diagnosing prostate cancer in practice.

Ensemble modeling coupled with six element concentrations in human blood for cancer diagnosis

Six important metal contents (i.e., zinc, barium, magnesium, calcium, copper, and selenium) in blood samples coupled with an ensemble classification algorithm have been used for the classification of normal people and cancer patients. A dataset containing 42 healthy samples and 32 cancer samples was used for experiment. The prediction results from this method outperformed those from the newly developed support vector machine, i.e., a sensitivity of 100%, a specificity of 95.2%, and an overall accuracy of 98.6%. It seems that ELDA coupled with blood element analysis can serve as a valuable tool for diagnosing cancer in clinical practice.

Magnesium, calcium, and trace elements excretion in 24-h urine

Urine is a clinical specimen often used in medical diagnostics for monitoring of elements concentrations and kidneys function. We determined the contents of magnesium (Mg), calcium (Ca), zinc (Zn), copper (Cu), iron (Fe), lead (Pb), and cadmium (Cd) in 74 samples of 24-h urine (from 46 women and 28 men). The measurements were realized by the atomic absorption spectrometry (AAS) with atomization in the flame (FAAS) and in the graphite furnace (GFAAS). The received results were the subject of statistical analysis including the sex and age of volunteers. Moreover, correlations between the elements and the relationships between age and amounts of excreted elements with urine were tested. We found the statistically significant higher content of Zn in men's urine than in women(')s one. Moreover, both adult women and men (>18 years) excreted much more Ca in urine in comparison to young subjects. Only in case of Pb the significant positive correlation between its amount in 24-h urine of all donors and age was stated. The correlation analysis has shown the significant positive relationships between Ca-Mg, Ca-Fe, Mg-Fe, Cu-Fe, Cu-Cd, Fe-Cd, and Pb-Cd in total samples of urine. Basing on our results, we concluded that the gender and age of donors may impact on the elemental status of 24-h urine.

Modeling the relationship between cervical cancer mortality and trace elements based on genetic algorithm-partial least squares and support vector machines

The relationship between the mortality of cervical cancer and soil trace elements of 23 regions of China was investigated. A total of 25 elements (i.e., Na, K, Mg, Ca, Sr, Hg, Pb, B, Tm, Th, U, Sn, Hf, Bi, Ta, Te, Mo, Br, I, As, Cr, Cu, Fe, Zn, and Se) were considered. First, 23 samples were split into the training set with 12 samples and the test set with 11 samples. Then, a combination strategy called genetic algorithm-partial least squares (GA-PLS) was used to pick out five important elements. i.e., Br, Ta, Pb, Cr, and As. Afterwards, the classic partial least squares (PLS) model and least square support vector machine (LSSVM) model were developed and compared. The results revealed that the SVM model significantly outperforms the PLS model, indicating that the combination of GA-PLS and LSSVM can serve as a potential tool for predicting the mortality of cancer based on trace elements.

Prediction of gastric cancer metastasis through urinary metabolomic investigation using GC/MS

AIM: To gain new insights into tumor metabolism and to identify possible biomarkers with potential diagnostic values to predict tumor metastasis.

METHODS: Human gastric cancer SGC-7901 cells were implanted into 24 severe combined immune deficiency (SCID) mice, which were randomly divided into metastasis group (n = 8), non-metastasis group (n = 8), and normal group (n = 8). Urinary metabolomic information was obtained by gas chromatography/mass spectrometry (GC/MS).

RESULTS: There were significant metabolic differences among the three groups (t test, P < 0.05). Ten selected metabolites were different between normal and cancer groups (non-metastasis and metastasis groups), and seven metabolites were also different between non-metastasis and metastasis groups. Two diagnostic models for gastric cancer and metastasis were constructed respectively by the principal component analysis (PCA). These PCA models were confirmed by corresponding receiver operating characteristic analysis (area under the curve = 1.00).

CONCLUSION: The urinary metabolomic profile is different, and the selected metabolites might be instructive to clinical diagnosis or screening metastasis for gastric cancer.

Analysis of the relationship between leukemia mortality and soil trace elements using chemometrics

The relationship between the mortality of leukemia and the contents of trace elements in the soils of 29 regions of China was investigated. A total of 27 elements were determined for each region. Considering that an efficient variable selection can be highly beneficial both to improve the predictive ability of the model and to greatly reduce its complexity, genetic algorithm-partial least squares was used to screen out 13 qualified elements. As a result, only 13 elements, i.e., As, Hg, Mn, Sr, Ba, Cu, Ti, Co, K, Ca, Rb, Zn, and Mg, were picked out and, a partial least squares model with three latent variables was obtained, whose prediction exhibited a correlative coefficient of 0.874 with actual mortality. Especially, it showed a high negative correlation between the content of soil As and the mortality of leukemia. Such a fact can be explained by the apoptotic effect of cancerous cells by trace-amount arsenic trioxide. Furthermore, according to whether the mortality was larger than two out of 100,000 (2 × 10⁻⁵), all the 29 regions were divided into 21 high-mortality regions and eight low-mortality regions and were assigned the label -1 or 1, respectively. Using the same 13 elements, a Fisher's discriminant analysis model was developed, which can successfully discriminate low- and high-mortality groups.

Application of boosting classification and regression to modeling the relationships between trace elements and diseases

The study on the relationship between trace elements and diseases often need to build a classification/regression model. Furthermore, the accuracy of such a model is of particular importance and directly decides its applicability. The goal of this study is to explore the feasibility of applying boosting, i.e., a new strategy from machine learning, to model the relationship between trace elements and diseases. Two examples are employed to illustrate the technique in the applications of classification and regression, respectively. The first example involves the diagnosis of anorexia according to the concentrations of six elements (i.e. classification task). Decision stump and support vector machine are used as the weak/base algorithm and reference algorithm, respectively. The second example involves the prediction of breast cancer mortality based on the intake of trace elements (i.e. a regression task). In this regard, partial least squares is not only used as the weak/base algorithm, but also the reference algorithm. The results from both examples confirm the potential of boosting in modeling the relationship between trace elements and diseases.

Chemometrics in metabolomics--a review in human disease diagnosis

Metabolomics is a post genomic research field concerned with developing methods for analysis of low molecular weight compounds in biological systems, such as cells, organs or organisms. Analyzing metabolic differences between unperturbed and perturbed systems, such as healthy volunteers and patients with a disease, can lead to insights into the underlying pathology. In metabolomics analysis, large amounts of data are routinely produced in order to characterize samples. The use of multivariate data analysis techniques and chemometrics is a commonly used strategy for obtaining reliable results. Metabolomics have been applied in different fields such as disease diagnosis, toxicology, plant science and pharmaceutical and environmental research. In this review we take a closer look at the chemometric methods used and the available results within the field of disease diagnosis. We will first present some current strategies for performing metabolomics studies, especially regarding disease diagnosis. The main focus will be on data analysis strategies and validation of multivariate models, since there are many pitfalls in this regard. Further, we highlight the most interesting metabolomics publications and discuss these in detail; additional studies are mentioned as a reference for the interested reader. A general trend is an increased focus on biological interpretation rather than merely the ability to classify samples. In the conclusions, the general trends and some recommendations for improving metabolomics data analysis are provided.