Development of a memetic clustering algorithm for optimal spectral histology: application to FTIR images of normal human colon.

The coupling between Fourier-transform infrared (FTIR) imaging and unsupervised classification is effective in revealing the different structures of human tissues based on their specific biomolecular IR signatures; thus the spectral histology of the studied samples is achieved. However, the most widely applied clustering methods in spectral histology are local search algorithms, which converge to a local optimum, depending on initialization. Multiple runs of the techniques estimate multiple different solutions. Here, we propose a memetic algorithm, based on a genetic algorithm and a k-means clustering refinement, to perform optimal clustering. In addition, this approach was applied to the acquired FTIR images of normal human colon tissues originating from five patients. The results show the efficiency of the proposed memetic algorithm to achieve the optimal spectral histology of these samples, contrary to k-means.

Development of a hierarchical double application of crisp cluster validity indices: a proof-of-concept study for automated FTIR spectral histology.

Fourier-transform infrared (FTIR) spectral imaging is currently used as a non-destructive and label-free method for analyzing biological specimens. However, to highlight the different tissue regions, unsupervised clustering methods are commonly used leading to a subjective choice of the number of clusters. Here, we develop a hierarchical double application of 9 selected crisp cluster validity indices (CCVIs) using K-Means clustering. This approach when tested first on an artificial dataset showed that the indices Pakhira-Bandyopadhyay-Maulik (PBM) and Sym-Index (SI) perfectly estimated the expected 9 sub-clusters. Then, the concept was applied to a real dataset consisting of FTIR spectral images of normal human colon tissue samples originating from 5 patients. PBM and SI were revealed to be the most efficient indices that correctly identified the different colon histological components including crypts, lamina propria, muscularis mucosae, submucosa, and lymphoid aggregates. In conclusion, these results strongly suggest that the hierarchical double CCVI application is a promising method for automated and informative spectral histology.

Predictors of social risk for post-ischemic stroke reintegration.

After stroke rehabilitation, patients need to reintegrate back into their daily life, workplace and society. Reintegration involves complex processes depending on age, sex, stroke severity, cognitive, physical, as well as socioeconomic factors that impact long-term outcomes post-stroke. Moreover, post-stroke quality of life can be impacted by social risks of inadequate family, social, economic, housing and other supports needed by the patients. Social risks and barriers to successful reintegration are poorly understood yet critical for informing clinical or social interventions. Therefore, the aim of this work is to predict social risk at rehabilitation discharge using sociodemographic and clinical variables at rehabilitation admission and identify factors that contribute to this risk. A Gradient Boosting modelling methodology based on decision trees was applied to a Catalan 217-patient cohort of mostly young (mean age 52.7), male (66.4%), ischemic stroke survivors. The modelling task was to predict an individual's social risk upon discharge from rehabilitation based on 16 different demographic, diagnostic and social risk variables (family support, social support, economic status, cohabitation and home accessibility at admission). To correct for imbalance in patient sample numbers with high and low-risk levels (prediction target), five different datasets were prepared by varying the data subsampling methodology. For each of the five datasets a prediction model was trained and the analysis involves a comparison across these models. The training and validation results indicated that the models corrected for prediction target imbalance have similarly good performance (AUC 0.831-0.843) and validation (AUC 0.881 - 0.909). Furthermore, predictor variable importance ranked social support and economic status as the most important variables with the greatest contribution to social risk prediction, however, sex and age had a lesser, but still important, contribution. Due to the complex and multifactorial nature of social risk, factors in combination, including social support and economic status, drive social risk for individuals.

Discrimination of immune cell activation using Raman micro-spectroscopy in an in-vitro & ex-vivo model.

Activation and proliferation of immune cells such as lymphocytes and monocytes are appropriate inflammatory responses to invading pathogens and are key to overcoming an infection. In contrast, uncontrolled and prolonged activation of these cellular signalling pathways can be deleterious to the body and result in the development of autoimmune conditions. The understanding of cellular activatory status therefore plays a significant role in disease diagnosis and progression. Conventional automated approaches such as enzyme linked immunosorbent assays (ELISA) and immune-labelling techniques are time-consuming and expensive, relying on a commercially available and specific antibody to identify cell activation. Developing a label-free method for assessing molecular changes would therefore offer a quick and cost-efficient alternative in biomedical research. Here Raman spectroscopy is presented as an effective spectroscopic method for the identification of activated immune cells using both cell lines and primary cells (including purified monocyte and lymphocyte subgroups and mixed peripheral blood mononuclear cell (PBMC) populations) obtained from healthy donors. All cell lines and primary cells were exposed to different stimulants and cellular responses confirmed by flow cytometry or ELISA. Machine learning models of cell discrimination using Raman spectra were developed and compared to reference flow-cytometry, with spectral discrimination levels comparing favourably with the reference method. Spectral signatures of molecular expression after activation were also extracted with results demonstrating alignment with expected profiles. High performance classification models constructed in these in-vitro and ex-vivo studies enabled identification of the spectroscopic discrimination of immune cell subtypes in their resting and activated state. Further spectral fitting analysis identified a number of potential spectral biomarkers that elucidate the spectral classification.

Vibrational spectroscopy of liquid biopsies for prostate cancer diagnosis.

BACKGROUND: Screening for prostate cancer with prostate specific antigen and digital rectal examination allows early diagnosis of prostate malignancy but has been associated with poor sensitivity and specificity. There is also a considerable risk of over-diagnosis and over-treatment, which highlights the need for better tools for diagnosis of prostate cancer. This study investigates the potential of high throughput Raman and Fourier Transform Infrared (FTIR) spectroscopy of liquid biopsies for rapid and accurate diagnosis of prostate cancer. METHODS: Blood samples (plasma and lymphocytes) were obtained from healthy control subjects and prostate cancer patients. FTIR and Raman spectra were recorded from plasma samples, while Raman spectra were recorded from the lymphocytes. The acquired spectral data was analysed with various multivariate statistical methods, principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA) and classical least squares (CLS) fitting analysis. RESULTS: Discrimination was observed between the infrared and Raman spectra of plasma and lymphocytes from healthy donors and prostate cancer patients using PCA. In addition, plasma and lymphocytes displayed differentiating signatures in patients exhibiting different Gleason scores. A PLS-DA model was able to discriminate these groups with sensitivity and specificity rates ranging from 90% to 99%. CLS fitting analysis identified key analytes that are involved in the development and progression of prostate cancer. CONCLUSIONS: This technology may have potential as an alternative first stage diagnostic triage for prostate cancer. This technology can be easily adaptable to many other bodily fluids and could be useful for translation of liquid biopsy-based diagnostics into the clinic.

Correction: Discrimination of breast cancer from benign tumours using Raman spectroscopy.

[This corrects the article DOI: 10.1371/journal.pone.0212376.].

Discrimination of breast cancer from benign tumours using Raman spectroscopy.

Breast cancer is the most common cancer among women worldwide, with an estimated 1.7 million cases and 522,000 deaths in 2012. Breast cancer is diagnosed by histopathological examination of breast biopsy material but this is subjective and relies on morphological changes in the tissue. Raman spectroscopy uses incident radiation to induce vibrations in the molecules of a sample and the scattered radiation can be used to characterise the sample. This technique is rapid and non-destructive and is sensitive to subtle biochemical changes occurring at the molecular level. This allows spectral variations corresponding to disease onset to be detected. The aim of this work was to use Raman spectroscopy to discriminate between benign lesions (fibrocystic, fibroadenoma, intraductal papilloma) and cancer (invasive ductal carcinoma and lobular carcinoma) using formalin fixed paraffin preserved (FFPP) tissue. Haematoxylin and Eosin stained sections from the patient biopsies were marked by a pathologist. Raman maps were recorded from parallel unstained tissue sections. Immunohistochemical staining for estrogen receptor (ER) and human epidermal growth factor receptor 2 (HER2/neu) was performed on a further set of parallel sections. Both benign and cancer cases were positive for ER while only the cancer cases were positive for HER2. Significant spectral differences were observed between the benign and cancer cases and the benign cases could be differentiated from the cancer cases with good sensitivity and specificity. This study has shown the potential of Raman spectroscopy as an aid to histopathological diagnosis of breast cancer, in particular in the discrimination between benign and malignant tumours.

Monitoring Radiotherapeutic Response in Prostate Cancer Patients Using High Throughput FTIR Spectroscopy of Liquid Biopsies.

Radiation therapy (RT) is used to treat approximately 50% of all cancer patients. However, RT causes a wide range of adverse late effects that can affect a patient's quality of life. There are currently no predictive assays in clinical use to identify patients at risk of normal tissue radiation toxicity. This study aimed to investigate the potential of Fourier transform infrared (FTIR) spectroscopy for monitoring radiotherapeutic response. Blood plasma was acquired from 53 prostate cancer patients at five different time points: prior to treatment, after hormone treatment, at the end of radiotherapy, two months post radiotherapy and eight months post radiotherapy. FTIR spectra were recorded from plasma samples at all time points and the data was analysed using MATLAB software. Discrimination was observed between spectra recorded at baseline versus follow up time points, as well as between spectra from patients showing minimal and severe acute and late toxicity using principal component analysis. A partial least squares discriminant analysis model achieved sensitivity and specificity rates ranging from 80% to 99%. This technology may have potential to monitor radiotherapeutic response in prostate cancer patients using non-invasive blood plasma samples and could lead to individualised patient radiotherapy.

Fully unsupervised inter-individual IR spectral histology of paraffinized tissue sections of normal colon.

In label-free Fourier-transform infrared histology, spectral images are individually recorded from tissue sections, pre-processed and clustered. Each single resulting color-coded image is annotated by a pathologist to obtain the best possible match with tissue structures revealed after Hematoxylin-Eosin staining. However, the main limitations of this approach are the empirical choice of the number of clusters in unsupervised classification, and the marked color heterogeneity between the clustered spectral images. Here, using normal murine and human colon tissues, we developed an automatic multi-image spectral histology to simultaneously analyze a set of spectral images (8 images mice samples and 72 images human ones). This procedure consisted of a joint Extended Multiplicative Signal Correction (EMSC) to numerically deparaffinize the tissue sections, followed by an automated joint K-Means (KM) clustering using the hierarchical double application of Pakhira-Bandyopadhyay-Maulik (PBM) validity index. Using this procedure, the main murine and human colon histological structures were correctly identified at both the intra- and the inter-individual levels, especially the crypts, secreted mucus, lamina propria and submucosa. Here, we show that batched multi-image spectral histology procedure is insensitive to the reference spectrum but highly sensitive to the paraffin model of joint EMSC. In conclusion, combining joint EMSC and joint KM clustering by double PBM application allows to achieve objective and automated batched multi-image spectral histology.