New Prostate Cancer Test Uses Machine Learning to Efficiently Spot New Cancer Biomarkers Accurately and Non-Invasively

Researchers in Sweden develop urine test that more effectively screens for prostate cancer than standard PSA test

Clinical laboratories may soon have a new inexpensive, non-invasive urine test to screen for prostate cancer that produces superior results compared to the standard PSA test.

An international team of scientists led by researchers at the Karolinska Institutet in Sweden found they could use machine learning to not only accurately identify the presence of a new set cancer biomarkers in urine samples but also determine the stage or grade of the cancer.

“There are many advantages to measuring biomarkers in urine,” said Mikael Benson, principal researcher in the Department of Clinical Science, Intervention and Technology at Karolinska Institutet and senior investigator for the study, in a news release. “It’s non-invasive and painless and can potentially be done at home. The sample can then be analyzed using routine methods in clinical labs.”

The researchers published their findings in Cancer Research titled, “Combining Spatial Transcriptomics, Pseudotime, and Machine Learning Enables Discovery of Biomarkers for Prostate Cancer.”

“New, more precise biomarkers than PSA can lead to earlier diagnosis and better prognoses for men with prostate cancer,” said Mikael Benson, principal researcher at Karolinska Institutet and senior investigator for the study, in a news release. “Moreover, it can reduce the number of unnecessary prostate biopsies in healthy men.” (Photo copyright: Karolinska Institutet.)

New Prostate Cancer Biomarkers

According to the American Cancer Society, there will be approximately 313,780 new cases of prostate cancer diagnosed this year in the US with about 35,770 deaths due to the disease. About one in eight US men will be diagnosed with prostate cancer in their lifetime, and the lifetime risk of dying from prostate cancer is one in 44 men.

“Early cancer diagnosis is crucial but challenging owing to the lack of reliable biomarkers that can be measured using routine clinical methods. The identification of biomarkers for early detection is complicated by each tumor involving changes in the interactions between thousands of genes. In addition to this staggering complexity, these interactions can vary among patients with the same diagnosis as well as within the same tumor,” the researchers wrote in Cancer Research.

The scientists “hypothesized that reliable biomarkers that can be measured with routine methods could be identified by exploiting three facts:

• The same tumor can have multiple grades of malignant transformation;
• These grades and their molecular changes can be characterized using spatial transcriptomics; and,
• These changes can be integrated into models of malignant transformation using pseudotime models to prioritize the genes that were most correlated with malignant transformation.”

To perform their study, the scientists analyzed the mRNA activity of cells in prostate tumors to construct digital models of prostate cancer. These models were then examined using machine learning, a type of artificial intelligence (AI), to locate specific proteins that could be used as biomarkers.

The researchers evaluated these new biomarkers in urine, blood, and tissue samples from more than 2,000 prostate cancer patients along with a control group. The team’s final calculations found the results of the urine test surpassed the current PSA test traditionally used for diagnosing prostate cancer.

“Prostate cancer can be effectively identified by analyzing the expression of candidate biomarkers in urine,” lead study author Martin Smelik, PhD student at Karolinska Institutet, told Fox News. “This approach outperforms the current blood tests based on PSA, but at the same time keeps the advantages of being non-invasive, painless, and relatively cheap.”

Advancements over Traditional PSA Test

Although the prostate-specific antigen (PSA) test typically used by doctors to diagnose prostate cancer can screen for the disease and monitor its progression, it has limitations.

“While PSA is an incredibly sensitive tool for issues related to the prostate, it is not specific to prostate cancer,” Matthew Abramowitz, MD, associate professor in the Department of Radiation Oncology at the Sylvester Comprehensive Cancer Center, University of Miami Miller School of Medicine, told Fox News. “The techniques proposed in the current study suggest the promise of identifying specific cancer markers in the urine, minimizing some of the specificity concerns associated with PSA.”

“This study highlights the power of machine learning applied to patient data in identifying breakthroughs that can help us diagnose cancer earlier, when our treatments are most effective,” Timothy Showalter, MD, a radiation oncologist at UVA Health in Virginia, told Fox News. “Prostate cancer screening has not seen a transformative advance in decades, and current approaches still rely on the PSA blood test, which is known to have low specificity for clinically significant cancers.”

“Overall, this study demonstrates the diagnostic potential of combining spatial transcriptomics, pseudotime, and machine learning for prostate cancer, which should be further tested in prospective studies,” the researchers wrote.

The Karolinska Institutet team is planning large-scale clinical trials as the next phase of their exploration.

—JP Schlingman