SoylentNews

19/11/20/0318226 story

A Computer Model Has Learned To Detect Prostate Cancer

posted by Fnord666 on Wednesday November 20 2019, @12:43PM   
from the positive-results dept.

Arthur T Knackerbracket has found the following story:

Scientists at the TSU Laboratory of Biophotonics, working with Tomsk National Research Medical Center (TNIMC) oncologists, have developed a new approach to the diagnosis of adenocarcinoma, a malignant tumor of the prostate gland, that uses artificial intelligence to identify oncopathology and determine the stage of the disease. Using machine learning, a computer model was taught to distinguish between healthy tissues and pathology with 100 percent accuracy.

The gold standard for the diagnosis of cancer is histology, during which tissue from a patient is examined for malignant changes. So that the samples can be stored for a long time, they are dehydrated and packed in paraffin. Then experts make thin sections and examine these slides under a microscope.

"Usually, several people work with prostate biopsy samples, and after studying the sections, they make a collegial decision," says Yuri Kistenev, executive director of the TSU Institute of Biomedicine. "The human factor has not been eliminated, therefore, due to subjective assessment, there are erroneous conclusions. We tried to solve this problem using IT technologies—we developed a computer model and, through machine learning, taught it how to detect abnormal areas using a tool such as terahertz spectroscopy."

---

Original Submission

• GIGOGIGO (Score: 0) by Anonymous Coward on Wednesday November 20 2019, @03:23PM (2 children)

  by Anonymous Coward on Wednesday November 20 2019, @03:23PM (#922393)

  "Usually, several people work with prostate biopsy samples, and after studying the sections, they make a collegial decision," says Yuri Kistenev, executive director of the TSU Institute of Biomedicine. "The human factor has not been eliminated, therefore, due to subjective assessment, there are erroneous conclusions. We tried to solve this problem using IT technologies—we developed a computer model and, through machine learning, taught it how to detect abnormal areas using a tool such as terahertz spectroscopy."

  So what did they use for the gold standard, ground truth? The erroneous human judgment?

  Then I see they didn't use a real holdout so they overfit to the (erroneous) target... There is something seriously wrong with medical training, it isn't normal to make so many obvious mistakes.

  On two occasions I have been asked, "Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?" ... I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.
  — Charles Babbage, Passages from the Life of a Philosopher[5]

  • Re:GIGORe:GIGO (Score: 0) by Anonymous Coward on Wednesday November 20 2019, @07:09PM (1 child)

    by Anonymous Coward on Wednesday November 20 2019, @07:09PM (#922541)

    There is something seriously wrong with medical training, it isn't normal to make so many obvious mistakes.

    Are you asserting that the mistakes made were medical in nature? Because I see plenty of reasons to assume otherwise.

    Parent

    • Re:GIGO(Score: 0) by Anonymous Coward on Wednesday November 20 2019, @07:27PM

      by Anonymous Coward on Wednesday November 20 2019, @07:27PM (#922552)

      I'm asserting that they had doctors decide whether the images indicated cancer or not, then trained an algorithm using those as the gold standard even though their motivation for doing so was the doctor decisions were often inaccurate.

      Parent

• No thanks(Score: 2) by The Mighty Buzzard on Wednesday November 20 2019, @03:30PM

  by The Mighty Buzzard (18) <themightybuzzard@proton.me> on Wednesday November 20 2019, @03:30PM (#922398) Homepage Journal

  If I'm going to find out I have cancer, I'm not making things pleasant for the doctor either. We're talking Taco Bell here.

  --
  My rights don't end where your fear begins.

• BiospyBiospy (Score: 1, Insightful) by Anonymous Coward on Wednesday November 20 2019, @06:35PM (1 child)

  by Anonymous Coward on Wednesday November 20 2019, @06:35PM (#922501)

  If I read it correctly, this "technology" still requires my prostate to be poked with needles, or perhaps the Terrahertz Technology device gets inserted? It's not exactly clear. Wake me up when The Technology can predict prostate cancer without chopping it open and taking bits of it out.

  • Re:Biospy(Score: 0) by Anonymous Coward on Wednesday November 20 2019, @06:41PM

    by Anonymous Coward on Wednesday November 20 2019, @06:41PM (#922510)

    It is impossible for it to even predict any better than the current method, since they trained it to mimic whatever the doctors decided (which they call "error prone").

    What it could be useful for is saving time/money since doctors don't have to meet to discuss the images.

    Parent