The U.S. Food and Drug Administration (FDA) has designated 4D Path’s software a breakthrough device as a possible way to achieve a faster and more accurate breast cancer diagnosis based on tissue taken in a breast biopsy or resection.
Named 4D Q-plasia OncoReader Breast, the software is intended to aid doctors in evaluating biopsies and resections (tumors that are surgically removed), and in achieving a more accurate diagnosis.
“We are thrilled that the FDA has recognized the potential of our technology to offer significant advantages over existing approved or cleared alternatives to establish long-term clinical efficiencies. Our device acts on par with a diagnostic histopathologist in the identification of invasive cancer,” Nic Orsi, PhD, chief pathologist for 4D Path, said in a press release.
Breakthrough device status is given to medical technologies that may lead to more effective diagnosis or treatment for life-threatening or debilitating conditions. It fosters closer and more frequent meetings with the FDA to speed access to a promising product by supporting its development and review, and grants prioritized review for subsequent submissions.
While in a breast biopsy a small piece of tissue is collected for clinical evaluation and diagnosis, in a resection — a more invasive procedure — doctors surgically remove as much as possible of a tumor to evaluate.
Traditionally, images obtained via these methods are analyzed under a microscope by a pathologist, who assesses the different clinical features of tumor cells, like aggressiveness and cancer grade.
The variability in the visual inspection of these images, including discordance among physicians, has been reported as a major cause of diagnostic inaccuracies. Moreover, tumors are made up of heterogeneous populations of cells, not all of which are easily recognized.
4D’s software is a platform based on artificial intelligence (AI) that can directly analyze a scanned biopsy image without human input, an approach called digital pathology.
This software uses statistical physics, cancer cell biology and mathematical principles to analyze tumor cells present in a biopsy or resection, but also the tumor cell’s surroundings, called tumor microenvironment. Information gathered is then integrated with knowledge from known perturbed pathways in cancer to provide a potentially more accurate diagnosis of the tumor grade, how abnormal tumor cells are, and how aggressive a tumor is likely to be.
A capacity to analyze the tumor microenvironment and pin it to disturbed cancer pathways is this platform’s key ability, according to Satabhisa Mukhopadhyay, PhD, founder and chief scientist at 4D Path.
“This is totally new in digital pathology, where other AI-based diagnostic algorithms have been limited to learning morphology [shape]-based cell type identification and classification. This makes them inherently vulnerable to patient-to-patient variability,” Mukhopadhyay said. “We tackle this problem from an entirely novel perspective by unveiling hidden data in [histological] images that enables us to determine both clinical diagnosis and tumor molecular profile in a single step.”
The improved diagnostic performance of 4D Q-plasia OncoReader Breast may reduce the need for additional tests, shorten the time to diagnosis, and lower healthcare costs.
It could also reduce the error rate on biopsies obtained before surgery “from 20% to less than 5%,” Orsi said. “This is important for patients who have less invasive treatment; where diagnostic tissue is limited; or for those who receive chemotherapy before surgery, where tumor appearance is altered by the treatment.”
The device’s features were welcomed by the cancer-specific Patient and Public Involvement (PPI), a patient advocacy group in Leeds.
“By accelerating the process and overcoming supplementary testing, both diagnostic costs and turnaround times would be significantly decreased, thereby reducing the financial burden on patients and healthcare providers. These developments are timely given the recognized shortage of diagnostic histopathologists nationally and globally, coupled with the ever-growing demand for diagnostic services,” said Andrew Hanby, a professor of breast cancer pathology at Leeds Teaching Hospitals Trust, and a long-time collaborator with 4D Path.
“Unlike current rival AI methodologies, 4D Path’s technology significantly augments and revolutionizes the goal of achieving high levels of accuracy in diagnostics. It promises to enhance both the quality and efficacy of the process,” Hanby added.