Microcalcification crystallography as a potential marker of DCIS recurrence

Ductal carcinoma in-situ (DCIS) accounts for 20-25% of all new breast cancer diagnoses. DCIS has an uncertain risk of progression to invasive breast cancer and a lack of predictive biomarkers may result in relatively high levels (~ 75%) of overtreatment. To identify unique prognostic biomarkers of invasive progression, crystallographic and chemical features of DCIS microcalcifications have been explored. Samples from patients with at least 5-years of follow up and no known recurrence (174 calcifications in 67 patients) or ipsilateral invasive breast cancer recurrence (179 microcalcifications in 57 patients) were studied. Significant differences were noted between the two groups including whitlockite relative mass, hydroxyapatite and whitlockite crystal maturity and, elementally, sodium to calcium ion ratio. A preliminary predictive model for DCIS to invasive cancer progression was developed from these parameters with an AUC of 0.797. These results provide insights into the differing DCIS tissue microenvironments, and how these impact microcalcification formation.

Abstract P4-02-22: Breast microcalcification chemistry predicts DCIS prognosis

Introduction: Microcalcifications are a common feature in mammographic detection of ductal carcinoma in situ (DCIS), and occur in >80% of cases. Known to be present as type I (calcium oxalate-CaO) and type II (carbonated calcium hydroxyapatite-CHAP) microcalcifications, their association with DCIS and their role in the progression of DCIS to invasive breast cancer (IBC) remains unexplored. In an effort to understand the factors involved in DCIS prognosis, it is hypothesized that changes in the chemical composition of calcifications, in tandem with molecular changes in the surrounding soft tissue, will define patients with DCIS who will progress to develop IBC from those who remain with a stable DCIS phenotype. To this end, a novel label-free approach of hyperspectral imaging using mid-infrared (mid-IR) and Raman spectroscopy was used to probe calcification chemistry and molecular composition of the surrounding ductal and stromal soft tissue. The main aim of the work is to identify biomarkers for DCIS prognosis, based on chemical and molecular compositional changes of calcifications and the surrounding soft tissue. It is anticipated that the spectral biomarkers will provide patients and clinicians an informed risk assessment whether to undertake treatment for DCIS or to be placed under active surveillance. Methods: Tissue samples from 422 patient have been obtained and include (i) ‘pure DCIS’ (DCIS without recurrence) (n=193), (ii) ‘DCIS with invasive recurrence’ (DCIS from patients who subsequently were known to develop invasive disease) (n=123), (iii) ‘DCIS plus contemporaneous invasive cancer’ (n=44) and ‘benign’ (n=62) samples. Serial tissue sections were measured using mid-IR and Raman hyperspectral imaging approaches targeting the same calcification and soft tissue regions from specific DCIS ducts. Hyperspectral imaging data was initially pre-processed to digitally remove paraffin and unintended spectral interferences. The pre-processed data was subjected to cluster analysis followed by unsupervised and supervised machine learning classification models to identify spectral features associated with DCIS and its progression to IBC. Results: Cluster analysis based segmentation of hyperspectral images revealed histopathological features including calcifications, epithelium, necrotic areas, connective tissue and stroma. Spectra were extracted from each of the histopathological features using image coordinates, and biomodelling analysis was performed. Initial analysis of 314 calcification images from 170 patients with (i) ‘pure DCIS’ (n=118) and (ii) ‘DCIS with invasive recurrence’ (n=52) showed an area under the receiver operating characteristic (AUROC) mean value of 85% in distinguishing pure DCIS from DCIS that later recurred as IBC. The calcification features appeared to indicate pathology specific changes in phosphate and carbonate content as well as changes in magnesium whitlockite content. Similar analysis of the surrounding soft tissue spectral features showed an AUROC mean value of 85% (necrotic regions surrounding calcifications) and 76% (epithelium) respectively. The epithelial features showed changes in protein secondary structure and content, which together with the calcification changes indicate structural remodelling in DCIS that progresses to IBC, from those that do not. Perspectives: In the ongoing analyses of imaging data from 422 patients, it is anticipated that molecular/structural features from calcification and soft tissue imaging data will provide important cues in understanding DCIS prognosis and could be a promising way forward in determining management of DCIS risk and treatment underpinned by the identification of specific discriminatory spectral markers. Acknowledgments: This work was supported by Cancer Research UK and by KWF Kankerbestrijding (ref. C38317/A24043).

Citation Format: Jayakrupakar Nallala, Doriana Calabrese, Sarah Gosling, Esther Lips, Rachel Factor, Allison Hall, Sarah E. Pinder, Ihssane Bouybayoune, Lorraine King, Jeffrey Marks, Thomas Lynch, Donna Pinto, Jelle Wesseling, E Shelley Hwang, Keith Rogers, Nick Stone. Breast microcalcification chemistry predicts DCIS prognosis [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P4-02-22.

Unsupervised segmentation of biomedical hyperspectral image data: tackling high dimensionality with convolutional autoencoders

Information about the structure and composition of biopsy specimens can assist in disease monitoring and diagnosis. In principle, this can be acquired from Raman and infrared (IR) hyperspectral images (HSIs) that encode information about how a sample's constituent molecules are arranged in space. Each tissue section/component is defined by a unique combination of spatial and spectral features, but given the high dimensionality of HSI datasets, extracting and utilising them to segment images is non-trivial. Here, we show how networks based on deep convolutional autoencoders (CAEs) can perform this task in an end-to-end fashion by first detecting and compressing relevant features from patches of the HSI into low-dimensional latent vectors, and then performing a clustering step that groups patches containing similar spatio-spectral features together. We showcase the advantages of using this end-to-end spatio-spectral segmentation approach compared to i) the same spatio-spectral technique not trained in an end-to-end manner, and ii) a method that only utilises spectral features (spectral k-means) using simulated HSIs of porcine tissue as test examples. Secondly, we describe the potential advantages/limitations of using three different CAE architectures: a generic 2D CAE, a generic 3D CAE, and a 2D convolutional encoder-decoder architecture inspired by the recently proposed UwU-net that is specialised for extracting features from HSI data. We assess their performance on IR HSIs of real colon samples. We find that all architectures are capable of producing segmentations that show good correspondence with HE stained adjacent tissue slices used as approximate ground truths, indicating the robustness of the CAE-driven spatio-spectral clustering approach for segmenting biomedical HSI data. Additionally, we stress the need for more accurate ground truth information to enable a precise comparison of the advantages offered by each architecture.

A multi-modal exploration of heterogeneous physico-chemical properties of DCIS breast microcalcifications

Ductal carcinoma in situ (DCIS) is frequently associated with breast calcification. This study combines multiple analytical techniques to investigate the heterogeneity of these calcifications at the micrometre scale. X-ray diffraction, scanning electron microscopy and Raman and Fourier-transform infrared spectroscopy were used to determine the physicochemical and crystallographic properties of type II breast calcifications located in formalin fixed paraffin embedded DCIS breast tissue samples. Multiple calcium phosphate phases were identified across the calcifications, distributed in different patterns. Hydroxyapatite was the dominant mineral, with magnesium whitlockite found at the calcification edge. Amorphous calcium phosphate and octacalcium phosphate were also identified close to the calcification edge at the apparent mineral/matrix barrier. Crystallographic features of hydroxyapatite also varied across the calcifications, with higher crystallinity centrally, and highest carbonate substitution at the calcification edge. Protein was also differentially distributed across the calcification and the surrounding soft tissue, with collagen and β-pleated protein features present to differing extents. Combination of analytical techniques in this study was essential to understand the heterogeneity of breast calcifications and how this may link crystallographic and physicochemical properties of calcifications to the surrounding tissue microenvironment.

Abstract P2-08-07: Predicting DCIS prognosis using infrared and raman spectroscopy of breast calcifications and soft-tissue microstructure

Introduction: Ductal carcinoma in situ (DCIS) is a potential precursor of invasive breast cancer. The uncertain trajectory of DCIS, to progress to invasive disease or to remain in situ, currently drives treatment, despite lack of proven benefit. Therefore, understanding the molecular features of the DCIS trajectory may prevent overtreatment of this disease. Breast calcifications are a common feature in DCIS and are seen mammographically in over 80% of cases. Calcifications have been largely characterised based only on x-ray morphology; their chemical composition, their association with the surrounding soft tissue and role in DCIS and invasive breast cancer biology is largely unexplored. In this regard, a bio-photonic approach, based on infrared (IR) and Raman spectroscopy in combination with machine learning, was used to study DCIS by probing the chemical composition of calcifications and the surrounding soft tissue in breast lesions. The main aim of the work is to identify molecular compositional changes in calcifications and in soft tissue that potentially accompany or drive the progression of DCIS to invasive breast cancer, or indicates a stable DCIS phenotype. Methods: Serial tissue sections from 303 patients with (i) ‘pure DCIS’ (DCIS without recurrence) (n=158), (ii) ‘DCIS with invasive recurrence’ (DCIS from a patient who subsequently was known to develop invasive disease) (n=123) and (iii) ‘DCIS plus invasive cancer contemporaneously’ (n=22), were measured using mid-IR imaging and Raman mapping. The same calcifications and soft tissue regions from specific DCIS ducts were targeted across the techniques on the serial sections. Spectral images were analysed using cluster analysis followed by unsupervised and supervised machine learning classification models to identify spectral features associated with the progression of DCIS to invasive breast cancer. Results: Segmentation of IR and Raman spectral images based on cluster analysis identified important histopathological features including calcifications, epithelium, necrotic areas, connective tissue and stroma based on the spectral heterogeneity. Based on analysis of Raman calcification data from 145 patients with (i) ‘pure DCIS’ (n=90) and (ii) ‘DCIS with invasive recurrence’ (n=55), an area under the receiver operating characteristic (AUROC) mean value of 85% was obtained in distinguish pure DCIS from DCIS that later recurred as invasive cancer. The calcification features appeared to indicate pathology specific changes in phosphate and carbonate content and appearance of magnesium whitlockite. Similar analysis of the surrounding soft tissue spectral features showed an AUROC mean value of 76%, which showed changes in protein secondary structure and content, particularly in the necrotic regions surrounding calcifications. In addition, classification models are being developed and refined from the IR spectral data, the initial results of which have shown an AUROC value of only 54% from the same patients’ data. Perspectives: It is anticipated that the current novel approaches allowing label-free measurement of calcifications and soft tissue will provide important cues in understanding DCIS prognosis and could be a promising way forward in determining DCIS management. Current and future efforts include identification of specific discriminatory spectral features for molecular and pathological correlation. Acknowledgments: This work was supported by Cancer Research UK and by KWF Kankerbestrijding (ref. C38317/A24043).

Citation Format: Jayakrupakar Nallala, Doriana Calabrese, Sarah Gosling, Allison Hall, Sarah Pinder, Ihssane Bouybayoune, Lorraine King, Jeffrey Marks, Esther Lips, Thomas Lynch, Donna Pinto, Jelle Wesseling, Shelley Hwang, Keith Rogers, Nick Stone, on behalf of the Grand Challenge PRECISION consortium. Predicting DCIS prognosis using infrared and raman spectroscopy of breast calcifications and soft-tissue microstructure [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P2-08-07.

Infrared Spectroscopic Analysis in the Differentiation of Epithelial Misplacement From Adenocarcinoma in Sigmoid Colonic Adenomatous Polyps

Purpose

The differential diagnosis of epithelial misplacement from invasive cancer in the colon is a challenging endeavour, augmented by the introduction of bowel cancer population screening. The main aim of the work is to test, as a proof-of concept study, the ability of the infrared spectroscopic imaging approach to differentiate epithelial misplacement from adenocarcinoma in sigmoid colonic adenomatous polyps.

Methods

Ten samples from each of the four diagnostic groups, normal colonic mucosa, adenomatous polyps with low grade dysplasia, epithelial misplacement in adenomatous polyps and adenocarcinoma, were analysed using IR spectroscopic imaging and data processing methods. IR spectral images were subjected to data pre-processing and cluster analysis based segmentation to identify epithelial, connective tissue and stromal regions. Statistical analysis was carried out using principal component analysis and linear discriminant analysis based cross validation, to classify spectral features according to the pathology, and the diagnostic attributes were compared.

Results

The combined 4-group classification model on an average showed a sensitivity of 64%, a specificity of 88% and an accuracy of 76% for prediction based on a 'single spectrum', whilst a 'majority-vote' prediction on an average showed a sensitivity of 73%, a specificity of 90% and an accuracy of 82%. The 2-group model, for the differential diagnosis of epithelial misplacement versus adenocarcinoma, showed an average sensitivity and specificity of 82.5% for prediction based on a 'single spectrum' whilst a 'majority-vote' classification showed an average sensitivity and specificity of 90%. A 92% area under the curve (AUC) value was obtained when evaluating the classifier using the Receiver Operating Characteristics (ROC) curves.

Conclusions

IR spectroscopy shows promise in its ability to differentiate epithelial misplacement from adenocarcinoma in tissue sections, considered as one of the most challenging endeavours in population-wide diagnostic histopathology. Further studies with larger series, including cases with challenging diagnostic features are required to ascertain the capability of this modern digital pathology approach. In the long-term, IR spectroscopy based pathology which is relatively low-cost and rapid, could be a promising endeavour to consider for integration into the existing histopathology pathway, in particular for population based screening programmes where large number of samples are scrutinised.

Characterization of colorectal mucus using infrared spectroscopy: a potential target for bowel cancer screening and diagnosis

Biological materials presenting early signs of cancer would be beneficial for cancer screening/diagnosis. In this respect, the suitability of potentially exploiting mucus in colorectal cancer was tested using infrared spectroscopy in combination with statistical modeling. Twenty-six paraffinized colon tissue biopsy sections containing mucus regions from 20 individuals (10 normal and 16 cancerous) were measured using mid-infrared spectroscopic imaging. A digital de-paraffinization, followed by cluster analysis driven digital color-coded multi-staining segmented the infrared images into various histopathological features such as epithelium, connective tissue, stroma, and mucus regions within the tissue sections. Principal component analysis followed by supervised linear discriminant analysis was carried out on pure mucus and epithelial spectra from normal and cancerous regions of the tissue. For the mucus-based classification, a sensitivity of 96%, a specificity of 83%, and an area under the curve performance of 95% was obtained. For the epithelial tissue-based classification, a sensitivity of 72%, a specificity of 88%, and an area under the curve performance of 89% was obtained. The mucus spectral profiles further showed contributions indicative of glycans including that of sialic acid changes between these pathology groups. The study demonstrates that infrared spectroscopic analysis of mucus discriminates colorectal cancers with high sensitivity. This concept could be exploited to develop screening/diagnostic approaches complementary to histopathology.

Calcification Microstructure Reflects Breast Tissue Microenvironment

Microcalcifications are important diagnostic indicators of disease in breast tissue. Tissue microenvironments differ in many aspects between normal and cancerous cells, notably extracellular pH and glycolytic respiration. Hydroxyapatite microcalcification microstructure is also found to differ between tissue pathologies, including differential ion substitutions and the presence of additional crystallographic phases. Distinguishing between tissue pathologies at an early stage is essential to improve patient experience and diagnostic accuracy, leading to better disease outcome. This study explores the hypothesis that microenvironment features may become immortalised within calcification crystallite characteristics thus becoming indicators of tissue pathology. In total, 55 breast calcifications incorporating 3 tissue pathologies (benign - B2, ductal carcinoma in-situ - B5a and invasive malignancy - B5b) from archive formalin-fixed paraffin-embedded core needle breast biopsies were analysed using X-ray diffraction. Crystallite size and strain were determined from 548 diffractograms using Williamson-Hall analysis. There was an increased crystallinity of hydroxyapatite with tissue malignancy compared to benign tissue. Coherence length was significantly correlated with pathology grade in all basis crystallographic directions (P < 0.01), with a greater difference between benign and in situ disease compared to in-situ disease and invasive malignancy. Crystallite size and non-uniform strain contributed to peak broadening in all three pathologies. Furthermore, crystallite size and non-uniform strain normal to the basal planes increased significantly with malignancy (P < 0.05). Our findings support the view that tissue microenvironments can influence differing formation mechanisms of hydroxyapatite through acidic precursors, leading to differential substitution of carbonate into the hydroxide and phosphate sites, causing significant changes in crystallite size and non-uniform strain.

Detection of Aβ plaque-associated astrogliosis in Alzheimer's disease brain by spectroscopic imaging and immunohistochemistry

Correlative vibrational spectroscopy and immunohistochemistry reveal astroglial processes co-localised with the lipid-rich shell of Aβ plaques.

Recent work using micro-Fourier transform infrared (μFTIR) imaging has revealed that a lipid-rich layer surrounds many plaques in post-mortem Alzheimer's brain. However, the origin of this lipid layer is not known, nor is its role in the pathogenesis of Alzheimer's disease (AD). Here, we studied the biochemistry of plaques in situ using a model of AD. We combined FTIR, Raman and immunofluorescence images, showing that astrocyte processes co-localise with the lipid ring surrounding many plaques. We used μFTIR imaging to rapidly measure chemical signatures of plaques over large fields of view, and selected plaques for higher resolution analysis with Raman microscopy. Raman maps showed similar lipid rings and dense protein cores as in FTIR images, but also revealed cell bodies. We confirmed the presence of plaques using amylo-glo staining, and detected astrocytes using immunohistochemistry, revealing astrocyte co-localisation with lipid rings. This work is important because it correlates biochemical changes surrounding the plaque with the biological process of astrogliosis.

Rapid infrared mapping for highly accurate automated histology in Barrett's oesophagus

Barrett's oesophagus (BE) is a premalignant condition that can progress to oesophageal adenocarcinoma. Endoscopic surveillance aims to identify potential progression at an early, treatable stage, but generates large numbers of tissue biopsies. Fourier transform infrared (FTIR) mapping was used to develop an automated histology tool for detection of BE and Barrett's neoplasia in tissue biopsies. 22 oesophageal tissue samples were collected from 19 patients. Contiguous frozen tissue sections were taken for pathology review and FTIR imaging. 45 mid-IR images were measured on an Agilent 620 FTIR microscope with an Agilent 670 spectrometer. Each image covering a 140 μm × 140 μm region was measured in 5 minutes, using a 1.1 μm² pixel size and 64 scans per pixel. Principal component fed linear discriminant analysis was used to build classification models based on spectral differences, which were then tested using leave-one-sample-out cross validation. Key biochemical differences were identified by their spectral signatures: high glycogen content was seen in normal squamous (NSQ) tissue, high glycoprotein content was observed in glandular BE tissue, and high DNA content in dysplasia/adenocarcinoma samples. Classification of normal squamous samples versus 'abnormal' samples (any stage of Barrett's) was performed with 100% sensitivity and specificity. Neoplastic Barrett's (dysplasia or adenocarcinoma) was identified with 95.6% sensitivity and 86.4% specificity. Highly accurate pathology classification can be achieved with FTIR measurement of frozen tissue sections in a clinically applicable timeframe.

Mid-infrared multispectral tissue imaging using a chalcogenide fiber supercontinuum source

We present, to the best of our knowledge, the first demonstration of mid-infrared supercontinuum (SC) tissue imaging at wavelengths beyond 5 μm using a fiber-coupled SC source spanning 2-7.5 μm. The SC was generated in a tapered large-mode-area chalcogenide photonic crystal fiber in order to obtain broad bandwidth, high average power, and single-mode output for diffraction-limited imaging performance. Tissue imaging was demonstrated in transmission at selected wavelengths between 5.7 (1754  cm^-1) and 7.3 μm (1370  cm^-1) by point scanning over a sub-millimeter region of colon tissue, and the results were compared to images obtained from a commercial instrument.

Chemico-mechanical imaging of Barrett's oesophagus

Barrett's oesophagus is a condition characterized by a change in the lining of the oesophagus that markedly increases the risk of adenocarcinoma. We demonstrate the first site-matched application of Brillouin microscopy, Raman microscopy and FTIR micro-spectroscopic imaging to ex-vivo epithelial tissue - Barrett's oesophagus. The mechanical and chemical characters of the epithelium were assessed in histological sections from a patient subjected to endoscopic oesophageal biopsy. Previous studies have shown that both these properties change within the oesophageal wall, owing to the presence of distinct cellular and extracellular constituents which are putatively affected by oesophageal cancer. Brillouin microscopy enables maps of elasticity of the epithelium to be obtained, whilst Raman and FTIR imaging provide 'chemical images' without the need for labelling or staining. This site-matched approach provides a valuable platform for investigating the structure, biomechanics and composition of complex heterogeneous systems. A combined Brillouin-Raman device has potential for in-vivo diagnosis of pathology. First application of site-matched micro Brillouin, Raman and FTIR spectroscopic imaging to epithelial tissue in Barrett's oesophagus.

High-resolution FTIR imaging of colon tissues for elucidation of individual cellular and histopathological features

Comparison of spectral-histopathological features of a colon tissue measured using a conventional (5.5 μm × 5.5 μm, left) and a high-magnification (1.1 μm × 1.1 μm, right) FTIR imaging system with respect to HE stained tissue (middle).

Evaluation of different tissue de-paraffinization procedures for infrared spectral imaging

Differential distribution of paraffin in a normal colon tissue section after various de-Waxing procedures in comparison to a paraffinized tissue.

Infrared and Raman imaging for characterizing complex biological materials: a comparative morpho-spectroscopic study of colon tissue

Complementary diagnostic methods to conventional histopathology are currently being investigated for developing rapid and objective molecular-level understanding of various disorders, especially cancers. Spectral histopathology using vibrational spectroscopic imaging has been put in the frontline as potentially promising in this regard as it provides a "spectral fingerprint" of the biochemical composition of cells and tissues. In order to ascertain the feasible conditions of vibrational spectroscopic methods for tissue-imaging analysis, vibrational multimodal imaging (infrared transmission, infrared-attenuated total reflection, and Raman imaging) of the same colon tissue has been implemented. The spectral images acquired were subjected to multivariate clustering analysis in order to identify on a molecular level the constituent histological organization of the colon tissue such as the epithelium, connective tissue, etc., by comparing the cluster images with the histological reference images. Based on this study, a comparative analysis of important factors involved in the vibrational multimodal imaging approaches such as image resolution, time constraints, their advantages and limitations, and their applicability to biological tissues has been carried out. Out of the three different vibrational imaging modalities tested, infrared-attenuated total reflection mode of imaging appears to provide a good compromise between the tissue histology and the time constraints in achieving similar image contrast to that of Raman imaging at an approximately 33-fold faster measurement time. The present study demonstrates the advantages, the limitations of the important parameters involved in vibrational multimodal imaging approaches, and their potential application toward imaging of biological tissues.

Infrared spectral histopathology for cancer diagnosis: a novel approach for automated pattern recognition of colon adenocarcinoma

Automated and label-free colon cancer diagnosis and identification of tumor-associated features using FTIR spectral histopathology directly on paraffinized tissue arrays.

Abstract 733: Infrared spectral imaging: a new automated diagnostic tool for colon cancer

Purpose: Novel methods are the need of the hour that could complement the ‘gold standard’ histopathology for cancer diagnosis. In this perspective, biophotonic approach of infrared (IR) spectral micro-imaging is one of the candidates, as it provides spectral fingerprint of cell and tissue biochemistry in a non-destructive and label-free manner. This ability has been exploited: 1) to develop a new concept of spectral bar-coding for rapid characterization of biochemical alterations between normal and tumoral epithelial components of colonic tissues, and 2) to identify spectral signatures for colon histology in order to develop a prediction model comprising potential diagnostic markers for rapid and automated colorectal cancer diagnosis.

Experimental procedure: Ten frozen colon tissue samples (five tumoral and non-tumoral pairs from five patients), and sixty-eight colonic samples (39 tumoral and 29 non-tumoral) from 32 patients in the form of paraffinized tissue arrays were imaged using IR spectral micro-imaging in a non-destructive manner. In case of paraffinized tissues, in order to avoid chemical deparaffinization, a mathematical deparaffinization based on extended multiplicative signal correction (EMSC) was implemented to neutralize the spectral interferences from paraffin. The spectral images were processed by a multivariate clustering method to identify the histological organization in a label-free manner.

Summary: In the first part, the spectral information from the epithelial components of the frozen tissues was automatically recovered on the basis of the intrinsic biochemical composition, and compared using a statistical method (Mann-Whitney U test) to construct spectral barcodes specific to each patient. In the case of paraffinized tissue arrays, an LDA based robust prediction model (comprising 86802 spectra, constructed from 9 samples, and tested on 59 unknown samples involving a huge bank of 3620287 spectra) showed 100 % sensitivity for malignancy, while 10 out of 29 non-tumoral samples were identified as having tumor pixels. Further tests are under way to analyze these false positive samples as they were either present in the peri-tumoral regions, or appear having an inflammatory signature. Important features difficult to discern by conventional histopathology like tumor budding, tumor-stroma association, and inflammation, were easily identified by this methodology.

Conclusion: The discriminant infrared spectral wavenumbers enabled characterization of some of the malignancy associated biochemical alterations associated with mucin, nucleotides, carbohydrates and protein regions. This study constituting a label-free and non-destructive approach demonstrates the potential of IR spectral micro-imaging, combined with multivariate statistical image analysis, as a complementary tool to conventional histopathology for an automated and objective cancer diagnosis.

Citation Format: Jayakrupakar Nallala, Marie-Danielle Diebold, Cyril Gobinet, Olivier Bouché, Ganesh-Dhruvananda Sockalingum, Olivier Piot, Michel Manfait. Infrared spectral imaging: a new automated diagnostic tool for colon cancer. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 733. doi:10.1158/1538-7445.AM2013-733

Note: This abstract was not presented at the AACR Annual Meeting 2013 because the presenter was unable to attend.

Infrared imaging as a cancer diagnostic tool: introducing a new concept of spectral barcodes for identifying molecular changes in colon tumors

Complementary diagnostic methods to conventional histopathology are under scrutiny for various types of cancers for rapid and molecular level diagnostics. In this perspective, a biophotonic approach based on infrared spectral micro-imaging combined with multivariate statistical analysis has been implemented on colon tissues. The ability of infrared imaging to investigate the intrinsic biochemical features of cells and tissues has been exploited to develop a new concept of spectral bar coding. To implement this concept, 10 frozen colon tissue samples (five nontumoral and tumoral pairs from five patients) were imaged using infrared spectral micro-imaging in a nondestructive manner. The spectral images were processed by a multivariate clustering method to identify the histological organization in a label-free manner. Spectral information from the epithelial components was then automatically recovered on the basis of their intrinsic biochemical composition, and compared using a statistical method (Mann-Whitney U-test) to construct spectral barcodes specific to each patient. The spectral barcodes representing the discriminant infrared spectral wavenumbers (900-1,800 cm(-1) ) enabled characterization of some of the malignancy-associated biochemical alterations associated with mucin, nucleotides, carbohydrates, and protein regions. This approach not only allowed the identification of common biochemical alterations among all the colon cancer patients, but also revealed a difference of gradient within individual patients. This new concept of spectral bar coding gives insight into the potential of infrared spectral micro-imaging as a complementary diagnostic tool to conventional histopathology, for biochemical level understanding of malignancy in colon cancers in an objective and label-free manner.

Infrared spectral imaging as a novel approach for histopathological recognition in colon cancer diagnosis

Innovative diagnostic methods are the need of the hour that could complement conventional histopathology for cancer diagnosis. In this perspective, we propose a new concept based on spectral histopathology, using IR spectral micro-imaging, directly applied to paraffinized colon tissue array stabilized in an agarose matrix without any chemical pre-treatment. In order to correct spectral interferences from paraffin and agarose, a mathematical procedure is implemented. The corrected spectral images are then processed by a multivariate clustering method to automatically recover, on the basis of their intrinsic molecular composition, the main histological classes of the normal and the tumoral colon tissue. The spectral signatures from different histological classes of the colonic tissues are analyzed using statistical methods (Kruskal-Wallis test and principal component analysis) to identify the most discriminant IR features. These features allow characterizing some of the biomolecular alterations associated with malignancy. Thus, via a single analysis, in a label-free and nondestructive manner, main changes associated with nucleotide, carbohydrates, and collagen features can be identified simultaneously between the compared normal and the cancerous tissues. The present study demonstrates the potential of IR spectral imaging as a complementary modern tool, to conventional histopathology, for an objective cancer diagnosis directly from paraffin-embedded tissue arrays.

The ratio 1660/1690 cm(-1) measured by infrared microspectroscopy is not specific of enzymatic collagen cross-links in bone tissue

In postmenopausal osteoporosis, an impairment in enzymatic cross-links (ECL) occurs, leading in part to a decline in bone biomechanical properties. Biochemical methods by high performance liquid chromatography (HPLC) are currently used to measure ECL. Another method has been proposed, by Fourier Transform InfraRed Imaging (FTIRI), to measure a mature PYD/immature DHLNL cross-links ratio, using the 1660/1690 cm⁻¹ area ratio in the amide I band. However, in bone, the amide I band composition is complex (collagens, non-collagenous proteins, water vibrations) and the 1660/1690 cm⁻¹ by FTIRI has never been directly correlated with the PYD/DHLNL by HPLC. A study design using lathyritic rats, characterized by a decrease in the formation of ECL due to the inhibition of lysyl oxidase, was used in order to determine the evolution of 1660/1690 cm⁻¹ by FTIR Microspectroscopy in bone tissue and compare to the ECL quantified by HPLC. The actual amount of ECL was quantified by HPLC on cortical bone from control and lathyritic rats. The lathyritic group exhibited a decrease of 78% of pyridinoline content compared to the control group. The 1660/1690 cm⁻¹ area ratio was increased within center bone compared to inner bone, and this was also correlated with an increase in both mineral maturity and mineralization index. However, no difference in the 1660/1690 cm⁻¹ ratio was found between control and lathyritic rats. Those results were confirmed by principal component analysis performed on multispectral infrared images. In bovine bone, in which PYD was physically destructed by UV-photolysis, the PYD/DHLNL (measured by HPLC) was strongly decreased, whereas the 1660/1690 cm⁻¹ was unmodified. In conclusion, the 1660/1690 cm⁻¹ is not related to the PYD/DHLNL ratio, but increased with age of bone mineral, suggesting that a modification of this ratio could be mainly due to a modification of the collagen secondary structure related to the mineralization process.

Abstract 2741: Spectral histopathology of paraffinised colon tissue microarrays: a new approach by infrared spectral imaging for colon cancer diagnosis

Purpose: Colorectal cancers are the third most common type and second leading cause of cancer-related deaths in Western world. Histopathology is the gold standard method for diagnosis of colon cancers. To improve cancer diagnosis, ensure an effective treatment and better efficacy of treatment modalities, it is important to develop effective diagnostic method more reliable in terms of biochemical changes that can be detected in a cancerous tissue. To this end, we have developed spectral histopathology based on infrared (IR) imaging. The aims of this study are therefore to: (a) examine, using FTIR spectroscopy, the molecular changes between normal and tumoral colon tissues, (b) exploit the potentials of IR imaging to identify new diagnostic markers, and (c) develop an algorithm which could be applied in routine as a diagnostic tool directly and automatically on new unknown samples.

Experimental procedure: IR imaging is a promising technique which has the potential to reveal intrinsic bio-molecular information in a tissue by probing vibrational motions of chemical bonds, thus giving a fingerprint of the composition and the structures. An IR imaging system (Spotlight 300, Perkin Elmer, Les Ulys, France) equipped with nitrogen-cooled 16-element MCT detector was calibrated to acquire images at 6.25 µm spatial and 4 cm−1 spectral resolutions averaged to 16 accumulations from 10µm thick sections of colon tissues placed on a calcium fluoride IR transparent window. These sections were embedded in paraffin block and stabilized in agarose matrix of a tissue micro array (TMA) slide consisting of 13 spots, each 3 mm in diameter. Adjacent 10 µm sections from the same TMA block were H&E stained for histological analysis. A modified Extended Multiplicative Signal Correction (EMSC) algorithm was applied to the IR images of colon tissues to neutralize spectral variability from paraffin and agarose in order to conserve spectral variability originating only from the biological tissue. For that, spectra collected from paraffin and agarose with same parameters were modelled by their first ten principal components.

Summary: Image analysis using K-means clustering on comparison with H&E stained images revealed diverse spectral features representing the biochemical make up of the tissues. Selected cluster images from normal tissue were then used as reference. To detect the presence of abnormal chemical features (cancerous) in new unknown tissues, we developed an innovative algorithm permitting to construct automatically high-contrast color-coded images based on the correlation coefficients between tissue spectral signatures.

Conclusion: IR imaging allowed differentiating and detecting underlying differences between normal and tumoral colon tissues supporting the capability of exploiting this innovative spectral histopathological method in colon cancer diagnosis.

Note: This abstract was not presented at the AACR 101st Annual Meeting 2010 because the presenter was unable to attend.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 2741.