Do you know how long it takes for a pathologist

to diagnose one LBC case?

The average time is 3-5 min.

But what about “AI pathology”?

In 2021, a total of 150,000 cervical cancer screening cases were sent to the pathology department of Ningbo Pathology Center for cervical cancer screening, including 100,000 cases sent by local hospitals and 50,000 cases sent from other provinces by the means of the pathology remote consultation platform.

There are 30,000-50,000 cells on a gynecological liquid-based cytology slide. A pathologist can only diagnose 100 slides a day. 

Since the positive rate of cervical cancer pathology slides is less than 10%, a system to screen out a large number of negative pathology slides would be necessary to greatly reduce the workload of pathologists.

After 3 years of research, the artificial intelligence cervical cancer screening auxiliary diagnosis system developed by the Ningbo Pathology Center and KFBIO was officially applied to pathology diagnosis last year. 

Artificial intelligence cervical cancer screening auxiliary diagnosis system is the practical application of AI (artificial intelligence) in the field of pathology. Based on the analysis of mass data and learning ability, every cell in the slides is scanned and analyzed automatically. And thousands of characteristic data of each cell nucleus are collected for regular classification of cancerous cells. In this way, quantitative and objective diagnostic reports can be generated to assist in pathology diagnosis. 

Pathologists put the slides into the scanner after samples are received, prepared, stained and sealed. And then the AI system will automatically identify and label abnormal cells on the slides. Pathologists only need to diagnose the suspected positive slides screened by the system and check the abnormal cells. 

Currently, the system can screen 1,200 slides a day, each in just 90 seconds.

In the future, “AI+ medical” will be more influential and powerful. And AI will assist pathologists to improve the efficiency and reliability of pathology diagnoses. 

(Article from Ningbo Pathology Center)