Tumour trials: Real-time boundary demarcation by ultrasonic aspiration and ambient mass spectrometry

Surgical certainty essential

The job of a surgeon removing a tumour is made much easier if the margins of the malignant growth can be differentiated from healthy tissue. Easy in principal, not so easy in practice. When the tumour is exposed in the operating theatre, it can be difficult to see where the tumour ends and healthy tissue begins, so other means must be employed.

The conventional way is for surgeons to refer to images taken before surgery using techniques such as magnetic resonance imaging or computerised tomography. More recently, histological examination has been carried out during surgery for near real-time analysis but, at 20-40 minutes duration, it is still regarded as too long.

The use of fluorescent or radioactive labels has also been employed but this approach is complicated by the requirement for specific labelling as well as surgical instruments that can detect the labels.

A more direct approach has been promised in recent years by the development of ambient mass spectrometry. Numerous examples of tissue analysis have been published which suggest that mass spectrometric analysis can identify tumour margins, but transfer to the operating theatre use is difficult. For instance, some of the ambient ionisation techniques require high voltage which is not compatible for in vivo deployment.

A safer solution has been demonstrated recently by the use of an ultrasonic aspirator that disintegrates tissue samples and continuously withdraws the debris from the surgical site. The particles were frozen for transport to a mass spectrometric lab for cryosectioning and imaging mass spectrometry using desorption ionisation.

Now, this offline procedure has been expanded by one of the leading researchers in ambient mass spectrometry methods who has developed an online process in real time for tissue type identification.

Ultrasonic aspiration for online analysis

Zoltan Takats who is jointly affiliated to Justus-Liebig-University, Giessen, Germany, and Semmelweis University, Budapest, Hungary, collaborated with colleagues from the University Hospital Giessen and Marburg, Medimass Ltd., Budapest, and the University of Debrecen.

They employed a cavitron ultrasonic surgical aspirator to break up the tissue and the drainage debris was led to a Venturi easy ambient sonic spray ionisation (V-EASI) source attached to the mass spectrometer.

If conventional spray ionisation were used the capillaries would become blocked by the macroscopic tissue particles but V-EASI uses larger capillaries with diameters greater than1 mm that allow smooth passage. In addition, the technique provides sufficient suction to draw up the liquid for transfer.

When the fluid produced from porcine brain cortex was ionised and analysed, signals were observed from different classes of endogenous compounds including lipids, carbohydrates, peptides and metabolites. This is a broader range than has been reported for other ambient mass spectrometric techniques.

For a sample of human brain tumour, analysed in negative-ion mode, a series of multiply charged negative ions were detected, one being assigned to calcitone gene related peptide, CGRP, an abundant brain peptide that has been linked with lung cancer. A second peptide was tentatively identified as thymosin beta4 peptide, which has been proposed as a prognostic marker of nonsmall cell lung cancer.

These conclusions were consistent with the fact that the tumour was a brain metastasis of nonsmall cell lung cancer, although Takats admitted that they need to be validated by further study.

Tissue differentiation in seconds

Having established that ultrasonic nebulisation combined with V-EASI mass spectrometry works in practice, the researchers studied the effects of the instrument settings on the signal intensities.

By varying the ultrasonic power, it was confirmed that ultrasonic disruption is required for transferring high concentrations of cell components to the draining fluid. Conventional surgical aspirators used in neurosurgery would not have the desired effect.

The pressure of nitrogen gas driving the Venturi inlet is also critical. Low pressures produced no signal but pressures above 10 bar led to signal saturation, so the value was optimised at 10 bar. The ionisation efficiency was also found to increase with the temperature of the capillary inlet.

The experimental set up and these signal dependencies suggested that ionisation occurs via a process similar to sonic spray ionisation, a conclusion supported by spectral comparison with sonic spray ionisation of a filtered tissue homogenate.

The optimised system was used to distinguish between the white and grey areas of healthy human brain tissue by comparing the full mass spectra from each tissue type. The mass spectra were searched against a database of 284 spectra of human brain tissue and tumours by principal components analysis and linear discriminant analysis.

The spectra were highly specific to tissue type, an essential prerequisite if the technique is to be used to identify tumour margins. Samples of glioblastoma multiforme were also clearly distinguished.

Under the optimum conditions, the system could operate for at least four hours. However, each tissue spectrum is measured and analysed within 2-3 seconds, a healthy improvement on the 30-40 minutes currently experienced for intraoperative histology during surgery or offline mass spectrometric analysis.

All of these experiments were conducted on removed tissues but Takats declared "Although the method has not yet been tested for in vivo tissue identification, the method is expected to perform similarly, especially since ultrasonic tissue ablation does not induce extensive bleeding, thus the drain sample does not contain an excessive amount of blood."

It has the potential to be used during surgery in real time on other organs such as the liver, kidney and pancreas and, if proved successful, will be a unique and invaluable aid to surgeons to ensure a successful outcome.

The views represented in this article are solely those of the author and do not necessarily represent those of John Wiley and Sons, Ltd.