Subscribe to our Newsletters !!
Fermentation is a natural method which is applied
In the world of Biochemistry, Microbiology and Pha
The human metapneumovirus, more simply referred to
Indegene, a digital-first life sciences commercial
Alembic Pharmaceuticals Limited (Alembic) announce
'Simplicity' seems to be the perfect word to descr
Dear Readers, Welcome to the latest issue of The Magazine
A study in the University of Pittsburgh shows the highest accuracy thus far in understanding and characterizing prostate cancer using an artificial intelligence (AI) application.1
“Humans are great at recognizing anomalies, but they have their particular biases or previous experience,” says senior writer Rajiv Dhir, MD, MBA, chief pathologist and vice chair of pathology at University of Pittsburgh Medical Center, Shadyside and professor of biomedical informatics in Pitt. “Machines are detached from the entire story. There’s certainly an element of standardizing care”
To prepare the AI to recognize prostate cancer, Dhir and his coworkers supplied images from more than a million components of stained tissue slides taken from individual biopsies. Each image was labeled by specialist pathologists to instruct the AI the way to discriminate between healthy and abnormal tissue. The plan was subsequently tested on a different group of 1,600 slides taken from 100 consecutive patients observed in UPMC for suspected prostate cancer.
In addition, this will be the first algorithm to expand beyond cancer detection, reporting high performance for tumor evaluation, sizing, along with invasion of the surrounding nerves. These all are clinically significant attributes required as part of the pathology report.
AI also flagged six slides that weren’t noted by the expert pathologists.
However, Dhir explained that this does not automatically mean that the system is superior to humans. By way of instance, in the course of evaluating these instances, the pathologist might have simply seen enough evidence of malignancy elsewhere in that patient’s samples to advocate treatment. For less experienced pathologists, though, the algorithm could act as a failsafe to capture cases which may otherwise be overlooked.
“Algorithms such as this are particularly beneficial in lesions which are atypical,” Dhir states. “A nonspecialized individual might not be in a position to make the correct assessment. That is a major benefit of this kind of system.”
When these results are promising, Dhir cautions that new algorithms will have to be trained to detect different kinds of cancer. The pathology markers are not universal across all tissue forms. But he didn’t see why that could not be done to accommodate this technology to work with breast cancer, for instance.
Read more from the University of Pittsburgh.
Reference
1. Pantoanowitz L, Quiroga-Garza GM, Bien L, et al. An artificial intelligence algorithm for prostate cancer diagnosis in whole slide images of core needle biopsies: a blinded clinical validation and deployment study. Lancet Digital Health. 2020;2(8):E407-E416. doi: 10.1016/S2589-7500(20)30159-X.
Featured image: Prostate biopsy with cancer probability (blue is low, red is high). This case was originally diagnosed as benign but changed to cancer upon further review. The AI accurately detected cancer in this tricky case. Credit: Ibex Medical Analytic.
