Imaging analysis method can reduce incorrect cancer diagnoses

Researchers out of Norway say they have developed a computerize method of analyzing images that can cut in half the number of patients incorrectly diagnosed with cancer.

Håvard Danielsen, head of research in the Institute for Medical Informatics at Radiumhospitalet, and Fritz Albregtsen of the department of informatics, say that many cancer patients either are under- or over-treated because pathologists don't have enough information.

"Pathology is a vulnerable scientific part of medicine. When pathologists look at a section they must compare the results using their own memory. The results can vary from pathologist to pathologist. The results can also vary from the same pathologist between Monday and Friday. Though pathologists can diagnose cancer with 100 percent certainty, the assessment of how serious the cancer is and how far it has advanced is very uncertain," said Danielsen in an article in Apollon, a research magazine published by the University of Oslo. "There are no absolute 'yes' or 'no' answers. The answer determines whether the patient receives treatment A or B. Pathologists disagree because the methods are not good enough. Their analyses are based on their own best judgment."

Cancer prognoses today have about a 60 percent certainty, according to Danielsen and Albregtsen. Their new computerized method can increase that percentage to 80 percent, they say, which means the number of incorrect prognoses can be halved.

The researchers were able to access the pathological sections of 5,000 patients with colon and rectal cancer for 10 years after their diagnosis and evaluated 300 of them--half of whom performed well, and half of who did not. Using the image analysis, they were able to study how the DNA strings and other proteins in the cell nuclei of the patients were organized, eventually distinguishing between good and incorrect prognoses.

Danielsen said that he and his colleagues--along with researchers from Karolinska University Hospital in Stockholm, Oxford University and Harvard University--are not testing this method on other pathological cancer sections. He hopes it can be used in clinical practice within three years.

To learn more:
- read the article in Apollon