NEW YORK – Researchers at the University of Texas and Baylor College of Medicine have presented the first data on the use of their MasSpec Pen device for cancer tissue margin analysis in actual surgery patients.
In a study published this week in the Proceedings of the National Academy of Sciences, the researchers used the device to collect tissue margin data during 18 pancreatic cancer surgeries, finding that their measurements achieved a high level of agreement with the final post-operative pathology reports for those surgeries.
Livia Eberlin, assistant professor of diagnostic medicine at UT and lead inventor of the MasSpec Pen, said that she and several colleagues have launched a company, Genio Technologies, that aims to commercialize the device. She serves as the Tulsa, Oklahoma-based firm's chief scientific officer. The company has raised $1.25 million in seed funding from investors including i2E Management and the Oklahoma Seed Capital Fund.
Accurately assessing tumor margins is a key part of many cancer surgeries where the goal is to remove all cancerous tissue while leaving as much healthy tissue as possible. Margins are typically determined using pathology, which involves freezing and staining tissue and then interpreting the results during surgery.
This adds time to surgery, however. Additionally, the accuracy of intra-operative pathology assays can vary with the facility where the surgery is being performed and the experience of the pathologist. Post-operative pathology studies will in some cases identify cancer missed during the procedure, requiring a second surgery.
These challenges have led researchers and clinicians to investigate methods for more rapidly and objectively assessing tissue margins during surgery. One area of focus has been ambient ionization mass spec, in which ions are generated directly from the target sample and injected directly into the mass spec, making for a streamlined, rapid assay well suited to an application like intra-operative tissue assessment.
The MasSpec Pen is based on ambient ionization, using drops of water to extract lipids and metabolites from tissue that are then analyzed via mass spectrometry to generate molecular profiles that distinguish between different tissue types.
Eberlin and her team first published on the device in 2017 and have used it to analyze a range of cancer tissues ex vivo, including breast, lung, thyroid, and ovary. They have also analyzed mouse tumors in vivo. The recent PNAS paper, however, represents the first publication of data collected during actual cancer surgery.
The researchers developed the classifier used in the study by analyzing 157 banked human tissues, including pancreatic ductal adenocarcinoma, pancreatic, and bile duct tissues, developing a model that matched conventional pathology results in 92 percent of cases with a sensitivity of 96 percent and specificity of 90 percent. They also developed a classifier that distinguished between bile duct cancer and pancreatic cancer with an accuracy of 95 percent.
They then gathered intra-operative data, collecting measurements from 78 intra-operative analyses of pancreatic or bile duct tissue made over the course of 18 surgeries. Of those 78 analyses, 14 did not provide sufficient signal to use and were excluded, leaving the researchers with data from 64 analyses made across 16 patients. In the case of those 64 analyses, the MasSpec Pen results agreed with the post-operative pathology results in 94 percent of cases. They found that when they incorporated the data from the banked tissues into their training sets, the MasSpec Pen analyses showed 100 percent agreement with the post-operative pathology results.
Beyond demonstrating the performance of the device in actual surgical patients, the study allowed the researchers to investigate how it could be integrated into surgical workflows.
"We really wanted to test it in the context of the surgical environment and also have the intended user (a surgeon) test it and handle it," Eberlin said, noting that this process made visible a number of areas where the device could be made more amenable to surgical workflows.
For instance, some surgeons said that they would prefer audio rather than visual cues alerting them to whether a portion of tissue was cancerous or not ,so that they could work without taking their eyes off the patient. They also realized they would need to make the tubing connecting the MasSpec Pen to the mass spectrometry adjustable to account for the different sizes and shapes of operating rooms. Clinicians also indicated that the noise produced by the mass spectrometer was distracting, an issue that Eberlin said could potentially be addressed by moving to smaller instruments.
"These were little things that you wouldn't necessarily be able to identify [unless] you were in the clinical environment and interacting with the surgeons," she said.
The researchers used in the study a Thermo Fisher Scientific Q Exactive mass spectrometer. Eberlin said, though, that ultimately a more compact and easier-to-use instrument would be necessary for implementing the device in surgeries. A more compact device would also allow clinicians to move the system between operating rooms, she noted.
"Our goal is to make it user-friendly, so that a technician in the hospital could service it and use it routinely," she said, citing the examples of imaging tools like MRI that are typically run by technicians despite being based on highly complex technologies.
She said that she believed the shift within the mass spec space toward smaller and more streamlined machines would produce instruments that would fit more easily into surgical workflows than existing systems.
"I think mass spec is just lagging a bit behind [in terms of ease of use] where we know there is a real need, but it just isn't quite there yet," she said. "It can make a huge impact, but it just really needs to be packaged in a way where we can put it in hospitals. I think all the mass spec vendors we are talking to understand that, but making that happen just takes time."
Eberlin said she and her colleagues are now putting together a paper covering their work using the system in breast cancer surgeries. In total, the researchers have used it for almost 200 surgeries with a focus thus far on breast and pancreatic cancer surgery where, she noted, the question of identifying tumor margins is particularly pressing.
Eberlin said she and Genio are also working to prepare data for the US Food and Drug Administration that would allow the company to start clinical trials for the device. In the PNAS work, the researchers were not allowed to provide their mass spec analyses to the surgeons in real time for fear of influencing their decision making, but would be allowed to do so in the context of a clinical trial.
"We've gotten to the point where we know it works in surgery, that it can detect surgical margins accurately," she said. "Now we're ready to actually move forward with the FDA steps."