Using XLSTAT, researchers evaluated PSMA expression in breast cancer samples, applying Fisher’s exact test and logistic regression to explore potential links with tumor characteristics. While they found some association with triple-negative breast cancer, overall PSMA expression levels were low compared to prostate tumors, limiting its theranostic potential. Even so, PSMA may still provide valuable insights for tracking treatment effectiveness and advancing future breast cancer research.
Medical researchers are continually pushing to improve how cancer is detected and treated—and advanced statistical analysis plays a critical role in that progress. These techniques are saving lives: according to Cancer Research UK, the mortality risk for early-stage breast cancer patients fell by 66% between the 1990s and the 2010s.
One interesting avenue of research is in theranostics, defined by the University of Texas MD Anderson Cancer Center as a combination of “therapy” and “diagnostics.” One candidate for theranostics in breast cancer is prostate-specific membrane antigen (PSMA). According to an article in Reviews in Urology, PSMA is a naturally occurring protein originally identified in the male prostate gland that has since been discovered in other tissues in male and female bodies—including breast tissue.
A French research team from UNICANCER at Centre François Baclesse in Caen investigated the potential for PSMA to be used for both diagnostic imaging and as a target for nuclear medical therapies. Their study, “Potential of PSMA for breast cancer in nuclear medicine: digital quantitative immunohistochemical analysis and implications for a theranostic approach” used Lumivero’s statistical analysis software, XLSTAT, to analyze imaging data and uncover statistical patterns in tumor samples.
PSMA is much more prevalent in cancerous tissue than in normal tissue, making it an important biomarker for clinicians trying to diagnose prostate cancer via immunohistochemistry—a microscope technique that can identify proteins and other large molecules in tissue samples. Using immunohistochemistry to find PSMA, then targeting it with radiotherapy, has resulted in better outcomes for treating prostate cancer.
What the research team wanted to understand was whether PSMA could be used as a biomarker and therapy target in breast cancer, too. Previous research had indicated that PSMA was especially prevalent with prostate tumors that had undergone neovascularization—that is, the tumors had created their own blood vessels.
The team decided to conduct imaging that would help them confirm whether PSMA was also more prevalent in neovascularized breast tumors as well. This could have promising therapeutic implications for some of the most difficult-to-treat types of breast cancer.
The team selected 58 random patients from their clinical database. All patients had undergone breast cancer surgery at UNICANCER between 2012–2017, and all had given a specimen of their tumor to the clinic’s tumor library. Samples of the tumors were split onto slides and subject to immunohistochemistry testing.
A second enzyme, CD31, was also imaged to highlight blood vessels, helping pathologists differentiate between PSMA expressed within the tumor and PSMA expressed within the newly created blood vessels. The slides were then digitally scanned and evaluated by pathologists to determine the level of PSMA expression.
Alongside the immunohistochemistry testing, the research team also conducted DNA sequencing of each patient who was included in the study to understand whether there are also associations between PSMA expression and certain types of genes, such as the BRCA (for BReast CAncer) gene.
PSMA slides were given scores based on how prevalent the protein was in the samples and where it was found. A score of 0 meant that patients showed no PSMA expression, 1 meant that patients had up to five PMSA cells in the image field, and 2 meant that there were more than five PSMA cells.
The genetic data and the imaging findings were analyzed with XLSTAT in two ways:
The statistical analysis found that about 31.6% of patients had no PSMA present in their sample images. 50.9% had a PMSA score of 1—low expression, while 17.5% had PMSA scores of 2. The logistic regression found some association between triple-negative breast cancer and PSMA expression, but none for age, sex, or other factors.
Finally, the low level of PSMA expression in breast tumors compared to prostate tumors means that there is likely “a lack of specific targets for delivering the radioactive substance, which will clearly hamper the effectiveness of this approach in breast cancer treatment.”
However, the researchers note that PSMA expression could be tracked throughout breast cancer treatment to determine the effectiveness of other therapies, especially those that have had variable success. “Future research,” they conclude, “should aim to correlate PSMA expression with clinical outcomes.”
While using PSMA as a theranostic may not be practicable, there is still plenty to learn about what PSMA expression can tell us about how breast cancer develops.
Turn complex datasets into clear insights with XLSTAT—directly in Microsoft Excel. From imaging studies to genetic associations, XLSTAT gives researchers the statistical power to uncover meaningful patterns and accelerate discovery.
