The photoacoustic technique developed by the Washington University team produced images (right) comparable in diagnostic features to those from traditional histology (left), but in far less time. The researchers are working to make the technique fast enough to be used to assess tumor removal during cancer surgery, and thereby avoid the need for second surgeries. [Image: Terence T.W. Wong]
A few weeks back, we reported on a multimodal fiber-based imaging probe that could help cancer surgeons assess tumor tissue in the operating room rather than the biopsy lab, and thereby avoid the need for repeat surgeries to get at cancerous tissue missed the first time around. U.S. researchers have now proposed an approach using another up-and-coming technique, photoacoustic microscopy (PAM), to attack the same problem (Sci. Adv., doi: 10.1126/sciadv.1602168). In a study of human breast cancer, the research team found that PAM could provide “multilayered histology-like imaging” of cancer tissues, comparable in key diagnostic features to the gold-standard images that currently come from biopsies—but in far less time.
Addressing a time-consuming process
In the United States each year, roughly 180,000 women undergo breast cancer lumpectomies—and as many as 25 percent of those patients end up back in the operating room for a second surgery, to get cancerous tissue missed in the first operation. The reason for the high repeat-surgery incidence is that it’s quite difficult to tell in the surgical suite where tumor tissue ends and healthy tissue begins. Thus, to be sure the surgery has gotten to all of the cancer, the tissue removed needs to be sent to a biopsy lab for histologic analysis, a process that can take days to complete.
The research team—headed by Rebecca Aft and Deborah Novack of the Washington University School of Medicine, USA, and by OSA Fellow Lihong Wang, formerly at Washington University and now at the California Institute of Technology—decided to look at the potential of optical imaging to provide a faster view for the cancer surgeon. They focused on PAM, a technique in which an optical pulse from a laser is absorbed by a specimen and partly converted into acoustic, or sound, energy.
Those sound pulses, picked up by an ultrasonic transducer, can be converted into images, much as with conventional ultrasound. Because different tissues or features preferentially absorb light at different wavelengths, photoacoustic microscopy does not require that the tissue be labeled or stained. Thus, in principle, PAM can be used to do direct imaging, without the intermediate steps required in a detailed biopsy.
The Washington University team zeroed in on imaging the specific set of features—the size variation and packing densities of cell nuclei—that are important in conventional histology to identify cancer cell clusters. To get at these features, it turns out that UV light, at a wavelength of 266 nm, is the best laser source to excite the right vibrations in the DNA and RNA molecules that characterize the nuclei. The short-wavelength UV light has the extra advantage of offering a high spatial resolution for these fine cellular features.
The team built a UV-PAM system that fired a UV laser through a series of optical elements to focus it on a tissue sample, placed in a water tank. An acoustic lens focuses the sound signal from the irradiated sample onto an ultrasound transducer, which translates the sound into an electric signal that is computer-amplified, recorded and projected into an image. In testing on specimens of various thicknesses, the team reported that the system “was able to image fixed, unprocessed breast tumors with an image quality comparable to that of conventional histology,” a much more involved and time-consuming technique.
Speeding things up
While considerably faster than the biopsy route, the PAM system is still not quite ready for prime time: imaging a 1-cm2 sample with the system at histology-level resolution took some seven hours in the tests. But the team thinks that several improvements could bring that time down by at least an order of magnitude. One tweak, says Wang, would be to use a microlens array to focus the laser on 80 spots at once rather than just one. “For this study we had only a single channel for emitting light,” he noted in a press release. “If you have multiple channels, you can scan in parallel, and that reduces the imaging time.”
He adds that increasing the laser pulse rate could also improve throughput. Aft, Novack and Wang are now applying for a grant to build such a faster system that could someday find its way into the operating room—and into the cancer surgeon’s toolkit.