Photoacoustic imaging shows accurate tumor margins

24 May 2017

Using ultraviolet wavelengths reveals cell architecture more rapidly than standard histology.

The importance of complete tumor removal during breast cancer surgery is highlighted by the significant number of patients who need to return for subsequent procedures, to excise parts of the tumor initially left behind.

At present, the gold-standard technique for assessing the cut margins of tissues and determining whether surgeons have removed all of a tumor involves post-operative analysis of the excised lump by microscopic examination and histology, which can take days to be completed.

An alternative histology procedure using frozen tissue samples has been under development as a route to producing results more rapidly, although the freezing procedure is difficult and large samples still take too long to analyze.

A project at Washington University in St Louis (WUSTL) and Caltech has now shown that a photoacoustic (PA) technique might offer a more rapid alternative. The results were published in Science Advances.

Photoacoustic microscopy using infra-red illumination to image internal organs and clusters of blood vessels is an active area of current research, but the new method employs ultraviolet illumination at 266 nanometers, exploiting the ability of the PA technique to image different targets with appropriate wavelengths.

In its paper, the team notes that the use of UV allows the cell nuclei to be highlighted, since the DNA and RNA in those nuclei absorb the 266 nm wavelength much more strongly than the surrounding biological components. Hence UV photoacoustic microscopy (UV-PAM) can provide the the same contrast as the hematoxylin labeling used in conventional histology, but without the need for labels - and crucially produce results in a shorter time.

Lihong Wang, a pioneer of PA techniques and co-author of the new study, said that using UV illumination to spot the margins of tumors was a further indication of what makes the technology so powerful.

"All molecules absorb light at some wavelength," he said. "Essentially, you can see any molecule, provided you have the ability to produce light of any wavelength. None of the other imaging technologies can do that."

Towards clinical use
The researchers tested the UV-PAM technique by scanning ex vivo slices of tumors removed from three breast cancer patients, and compared the images it created to conventionally stained specimens. According to the paper, the UV-PAM images matched the stained samples in all key features, revealing the key cell parameters such as growth pattern and size that can be used to differentiate normal tissue from a malignancy.

Although improvements to the scanning time and rate of image throughput will be needed before UV-PAM can realize its likely potential, these proof-of-concept results hold out the promise of using the technique in clinical workflows, and potentially for in vivo tissue analysis.

"In vivo intraoperative guidance would be a natural next step," said Rebecca Aft, a WUSTL professor of surgery and a co-senior author on the study, speaking to Optics.org.

"We envision that this device would allow us to evaluate surgical margins intra-operatively for any cancer cells, and to act on the results while we have the patient in the operating room. The main challenges to achieving this are timely evaluation of specimens, and refining the software to accommodate the many histological variations observed in breast tissue."

Other enhancements to the platform could include the incorporation of a system to physically mark the tissue at the margins of the tumor, as an indication to pathologists of the area to focus on when performing histological evaluation.

Aft echoed Lihong Wang's comment about the range of wavelengths that could be of use in a PA platform. It may be that visible, rather than UV, wavelengths could be be used to image particular proteins called cytochromes which absorb in the visible range, opening another possible route to obtaining histology-like image information in an in vivo scenario.

"One day we think we will be able to take a specimen straight from the patient, put it into a machine in the operating room and know in minutes whether we have got all the tumor out or not," Aft said. "That is the goal."

About the Author

Tim Hayes is a contributor to Optics.org.