Merja Rossi (Presenter)
Imperial College London
Bio: DPhil from University of Oxford on studying cell type specific metabolism and metabolic flux analysis, currently working in the iKnife project at Imperial College London.
Authorship: Rossi MT, Gurung D, St John ER, Doria, L, Balog J, Gildea L, Speller A, Ramakrishnan R, Golf O , Veselkov K, Mckenzie J, Takats Z
Imperial College London
| Short Abstract Rapid Evaporative Ionisation Mass Spectrometry (REIMS) is a technique used to analyse ions produced in electrosurgical smoke. This technique has recently been developed for near real time identification of normal and cancerous tissue in the form of the iKnife technology. The technology uses multivariate statistics for computational analysis of the whole lipid profile detected in the spectra derived from tissue. Here we describe how REIMS lipid profiles obtained from ex vivo breast specimens can be matched to other datasets to provide information on differences in lipid composition of breast tissue when compared with data obtained through DESI-imaging of the same clinical samples with confirmed histology. |
Long Abstract
Introduction:
Rapid Evaporative Ionisation Mass Spectrometry (REIMS) is a mass spectrometry technique that uses the surgical smoke produced by the electrosurgical knife routinely used in many types of cancer surgery including breast conserving surgery. The smoke produced by the diathermy hand piece is led into a mass spectrometer and the analysis produces a lipid profile that has been found to often be tissue type specific. This allows differentiating between normal and cancerous tissue in a variety of human tissues including normal and cancerous breast tissue (Balog et al. 2013). The REIMS technique has recently been used to develop the intelligent knife (iKnife) technology towards near real time tissue identification in the operating theatre.
In order to distinguish between cancerous and normal tissue, the lipidomic profiles observed in REIMS spectra are used to compare tissue types through use of multivariate statistical analysis. These lipid profiles are therefore different enough between the tissue types to allow successful classification into the correct classes as confirmed by histology. However, the REIMS lipid profile arises from a non-targeted approach and therefore further information is needed to elucidate the details of what information these spectra contain. A combination of techniques may be required to provide further understanding on the changes in lipid metabolism between normal and cancer tissue that give rise to the observed lipid profile.
Other mass spectrometric techniques have been used to detect changes that arise in the lipidome in cancer and studies have reported on the significant difference in proportions of phospholipids and triglycerides in these tissues. Previously changes in lipid species have been confirmed through DESI-imaging (Guenther et al. 2015), leading to established protocols and a lipid database that was used in this work. As a first step for identifying significant changes in the lipidome in breast tissue samples, data from DESI-imaging has been compared to REIMS data measured from the same tissue samples.
Method:
REIMS data was obtained from 17 normal, 10 tumour and 5 fibroadenoma samples with on average 3 burns on the tissue analysed for each sample. Samples for DESI-imaging were cut from the same tissue and stored frozen until processing. The REIMS data was obtained using an electrosurgical knife (iKnife) and a set up previously validated for analysing breast tissue. Both types of data were preprocessed following standard techniques used for REIMS and DESI data. The data for each mass peak was then averaged across samples of the same type. The sample type was confirmed through traditional histopathology.
The DESI-imaging data was acquired for frozen sections processed from 6 normal and 6 tumour samples from this same set of samples. This data from tissue with confirmed histology was also averaged. The mass charge peaks present in both average datasets were then matched to create a third set of data, consisting of lipids consistently detected through both techniques. For each of the matching peaks, the average data from tumour and normal tissue were compared to highlight differences between the two tissue types. These data were matched against an in house lipid database for the DESI-imaging data to provide an identity for each of the matches. Boxplots were used to compare all the datapoints between normal and tumour tissue.
Results
The analysis of the sample data suggests there are differences in individual lipid species covered by our dataset between the data derived from normal tissue and that derived from tumour tissue. These differences are mainly detected in lipids belonging to the group of glycerophospholipids. For example, PG 36:7 is detected at higher levels using REIMS in tumour than normal tissue and at higher levels from fibroadenoma samples than from normal tissue. However, the level detected from fibroadenomas is on average lower than that detected from tumour tissue. On the other hand, many lipids show no significant differences between tumour and normal data. Further analysis of these results will allow us to map these data against pathways in the lipid metabolism and start developing further understanding on the biology behind the differences that drive the classification of REIMS data.
Conclusions:
The iKnife technology is a promising novel tool for breast conserving surgery and several other types of cancer surgery. Its ability to distinguish between normal and cancerous tissue is based on the changes in the lipidomic profile. Average REIMS spectra follow the pattern reported by Sakai and group, where phospholipid content is higher and triacylglycerol content significantly lower in the human breast cancer tissue than normal tissue (Sakai et al. 1992). Some of the changes we detected, such as higher level of PE(32:0) in tumour tissue have also been previously described (Guenther et al. 2015).
However, the workflow and results described here will form the basis for further work on understanding the tissue type specific changes that occur in the lipidomic profile in cancerous tissue. Using a combination of techniques will allow us to better understand what lipids are detected in surgical smoke and how the lipid profile differs in cancerous and normal tissue. Future work will focus on adding to this dataset by increasing the number of samples and using other techniques to add new kinds of data.
References & Acknowledgements:
Balog J, Sasi-Szabó L, Kinross J, Lewis M, Muirhead L, Veselkov K, Mirnezami R, Dezső B, Damjanovich L, Darzi A, Nicholson J, and Takáts Z. Intraoperative Tissue Identification Using Rapid Evaporative Ionization Mass Spectrometry, Science Translational Medicine 17 Jul 2013: 5; pp.194ra93.
Guenther S, Muirhead L, Speller A, Golf O, Strittmatter N, Ramakrishnan R, Goldin RD, Jones E, Veselkov K, Nicholson J, Darzi A, and Takats Z. Spatially Resolved Metabolic Phenotyping of Breast Cancer by Desorption Electrospray Ionization Mass Spectrometry, Cancer Research 1 May 2015: 75; 1828.
Sakai K, Okuyama H, Yura J, Takeyama H, Shinaqawa N, Tsuruga N, Kato K, Miura K, Kawase K, Tsujimura T, et al. Composition and turnover of phospholipids and neutral lipids in human breast cancer and reference tissues, Carcinogenesis Apr 1992: 13: 579-84.
| Description | Y/N | Source |
| Grants | no | |
| Salary | no | |
| Board Member | no | |
| Stock | no | |
| Expenses | no |
IP Royalty: no
| Planning to mention or discuss specific products or technology of the company(ies) listed above: | no |