Thursday March 21, 2019

New AI system can diagnose prostate cancer just good as pathologists

"The system was programmed to learn and gradually improve how it interpreted the samples. Our result show that the diagnosis the AI reported was at a level comparable to that of a pathologist"

0
//
Chronic diseases can cause cancer as well.
Chronic diseases can cause cancer as well.
  • This new AI can identify cancer just as accurately as a pathology lab
  • This app can overcome the shortage of pathologists
  • The app is very useful and revolutionary

Researchers have developed a new learning artificial intelligence (AI) system which can diagnose and identify cancerous prostate samples as accurately as any pathologist.

Chronic diseases are not yet included in cancer prevention schemes.
New mobile app can detect cancer.

According to the researchers, this holds out the possibility of streamlining and eliminating variation in the process of cancer diagnosis. It may also help overcome any local shortage of trained pathologists.

“This is not going to replace a human pathologist. We still need an experienced pathologist to take responsibility for the final diagnosis,” said lead author Hongqian Guo from the Nanjing University in China.

Also Read: Girls may inherit ovarian cancer gene from fathers

“What it will do is help pathologists make better, faster diagnosis, as well as eliminating the day-to-day variation in judgement which can creep into human evaluations,” Guo added.

For the study, presented at the 33rd European Association of Urology Congress in Copenhagen, researchers took 918 prostate whole mount pathology section samples from 283 patients, and ran these through the analysis system, with the software gradually learning and improving diagnosis.

These pathology images were subdivided into 40,000 smaller samples; 30,000 of these samples were used to ‘train’ the software, the remaining 10,000 were used to test accuracy.

The results showed an accurate diagnosis in 99.38 percent of cases (using a human pathologist as a ‘gold standard’), which is effectively as accurate as the human pathologist.

The app is just as effective as a pathologist. Pixabay

They were also able to identify different Gleason Grades in the pathology sections using AI; ten whole mount prostate pathology sections have been tested so far, with similar Gleason Grade in the AI and human pathologist’s diagnosis. The group has not started testing the system with human patients.

“The system was programmed to learn and gradually improve how it interpreted the samples. Our result show that the diagnosis the AI reported was at a level comparable to that of a pathologist. “Furthermore, it could accurately classify the malignant levels of prostate cancer,” Guo added. IANS

Next Story

Researchers Discover Balance of Two Enzymes That May Help Treat Pancreatic Cancer

While still in the earliest stages, Newton hoped this information might one day aid pancreatic diagnostics and treatment

0
Cancer
Cancer Ribbon. Pixabay

A new research has set the stage for clinicians to potentially use levels of a pancreatic cancer patient’s PHLPP1 and PKC enzymes as a prognostic and for researchers to develop new therapeutic drugs that change the balance of the two enzymes as a means to treat the disease.

The study, published on Wednesday in Molecular Cell, was led by Alexandra Newton, professor in the Department of Pharmacology at the University of California, San Diego, School of Medicine, and Timothy Baffi, a graduate student in her lab, Xinhua news agency reported.

The new study built on the team’s work in 2015 that found the enzyme PKC, which was believed in previous studies to promote tumour growth, actually suppressed it.

The latest study took the investigation a step further by uncovering how cells regulate PKC activity and discovered that any time an over-active PKC is inadvertently produced, the PHLPP1 “proofreader” tags it for destruction.

Cancer patient
Cancer patient.

“That means the amount of PHLPP1 in your cells determines your amount of PKC,” Newton said. “And it turns out those enzyme levels are especially important in pancreatic cancer.”

The team observed 105 pancreatic cancer tumours to analyze the enzyme levels in each one. About 50 per cent of patients with low PHLPP1/high PKC lived longer than five-and-a-half years.

Also Read- A Brain Circuit Can Help Reverse Craving for Liquor, Says Study

While still in the earliest stages, Newton hoped this information might one day aid pancreatic diagnostics and treatment.

Pancreatic cancer is caused by the abnormal and uncontrolled growth of cells in the pancreas, a large gland in the digestive system. It typically doesn’t show symptoms in the early stages. Sufferers tend to develop signs, such as back pain and jaundice, when it has spread to other organs. (IANS)