The University of California, San Francisco (UCSF) and digital pathology company Proscia have announced a partnership to advance artificial intelligence (AI) in interpreting pathology results in patients. The goal of this partnership is to increase the speed and accuracy at which cancer can be diagnosed.
As an early adopter of digital pathology, UCSF has amassed volumes of data that can be used by Proscia to ensure its computational application surmounts the obstacles involved in creating AI software in a clinical, commercial application.
Initial trials testing out the clinical efficacy of Proscia’s prostate AI application software will begin in the field of prostate cancer, the second leading cause of cancer deaths among men in the United States. If this trial goes as planned, it will validate this application for expansion into additional cancer types, particularly high-impact pathology subspecialties.
Unlike other forms of cancer diagnosis, conventional methods of prostate cancer diagnosis rely heavily on human interpretation, and require pathologists to review a large amount of prostate cellular material from each patient. This standard of care for diagnosing cancer relies on the pathologist’s assessment of tissue biopsies viewed under a microscope.
After reviewing a large volume of cellular material, the pathologist must assign a qualitative grade to indicate the severity of the patient’s disease. This scoring will be used to diagnosis and provide a prognosis for the patient.
This 150-year-old manual and subjective practice cannot keep pace with the rising cancer burden amid decreasing pathologist workforce. It is very problematic in terms of turnaround time for the patient, and can lead to delays and/or a lack of confidence in treatment decisions.
Prostate cancer diagnosis is especially problematic given its high slide-per-case volume, complex reporting requirements, and qualitative grading system, which often lead to delayed turnaround times, increased use of ancillary tests, and reduced confidence in treatment decisions.
If Proscia can achieve a successful trial, it will imply that the AI application can accurately account for the variability that exists across a wide range of tissue diagnoses, the method of biopsy that is used, how the tissue was prepared for analysis, tissue fixation procedures, and the digital scanning processes. It will also validate the use of computational pathology applications in driving a much-needed quality and efficiency gain in clinical laboratories using digital pathology.
As one of the earliest adopters of digital pathology for primary diagnosis, UCSF has amassed volumes of diverse, high-quality digitized data. This data is initially being used to ensure that Proscia’s computational pathology application for prostate cancer accurately accounts for that variability.
“As prostate cancer impacts millions of patients each year, it is critical that we improve productivity and confidence in this high-impact specialty,” said Mike Bonham,M.D.,Ph.D, Proscia’s chief medical officer. “Through our partnership with UCSF, an institution that achieves the highest standards in patient care, research, and education, we are gaining the data and experience required to ensure that our AI delivers meaningful benefits in practice, where so many other solutions have struggled to perform.”
Beyond their initial work in prostate cancer, Proscia and UCSF will expand their focus to accelerate the introduction of similar solutions that advance the practice of pathology for subspecialty-specific processes, which make up the majority of pathology case work.
As digital pathology continues to gain traction and use in the laboratory, these deep learning-enabled applications hope to increase the drive of adopting more modern methods of diagnosis. It is also a hope that these new methods may unlock new and additional diagnostic information to further cancer discovery and improve patient outcomes.
“UCSF prides itself as being an institution in the intersection of research and clinical practice of medicine, continuously working to translate new findings into more effective prevention, diagnosis, and treatment,” said Zoltan Laszik, professor of Pathology at UCSF. “Proscia’s focus on delivering practical AI solutions strongly aligns with our efforts, and we are pleased to work together to improve the routine pathology workflow.”
This partnership adds to the growing list of leading academic and commercial labs with which Proscia is collaborating with to bring about computational pathology applications to market. In December 2019, the company announced a data collaboration with Johns Hopkins School of Medicine, and last week Proscia released the results of the largest AI validation study in pathology, conducted in collaboration with the Dermatopathology Laboratory of Central States, the University of Florida, and the Thomas Jefferson University Hospital. This collaboration was in support of the June 2019 release of another AI software application in dermatopathology.
