The Sleepless Sherlock Holmes: How can artificial intelligence help unravel the mysteries of the cellular world?

In the world of cells, there is a technology called "cervical liquid-based thin layer cytology" (LBC), which is like a cell detective, mainly responsible for identifying potential health risks. Today, we are going to reveal the latest assistant of this cell detective - artificial intelligence (AI)!

Imagine AI, a super assistant that not only speeds things up, but also makes detective work more accurate and more fun!

1. Cervical liquid-based thin layer cytology - the stage of cells under the microscope

So what is cervical liquid-based thin-layer cytology examination technology?

Simply put, it is a technology that detects whether cervical cells are healthy. Unlike traditional scraping technology, it converts cervical cell samples into liquid form and examines them under a microscope to look for potential abnormal cells.

"Cervical liquid-based thin layer cytology" (LBC) does not refer to a specific medical testing method, but a general term for a class of methods. In actual application, our doctors or laboratories may use different types of LBC methods, such as the TCT that we are more familiar with.

How does this technology work?

Similar to high-definition lenses in movies, it can provide high-definition "close-ups" of cells, helping doctors see subtle changes in cells more clearly and thus diagnose the type of pathology.

△ Cervical exfoliated cells under Papanicolaou staining

2. Artificial Intelligence: A New Partner for Cell Detectives

• The role of AI: “Sleepless Sherlock Holmes”

AI is a "super detective" who is proficient in the cellular world and can analyze thousands of cell images without resting or drinking water. It is much faster and makes fewer mistakes than human doctors.

Among 1,500 cervical liquid-based thin-layer cytology smears collected from the First Affiliated Hospital of a Medical University from February to November 2023, the sensitivity, specificity, and accuracy of pathologists' single reading, artificial intelligence (AI)'s single reading, and artificial intelligence (AI)-assisted physician reading were analyzed and compared, and the reading time was also compared.

The results showed that artificial intelligence (AI)-assisted physician reading has the advantages of high sensitivity, specificity and accuracy compared with artificial intelligence (AI) alone and pathologists reading films.

Therefore, using artificial intelligence (AI)-assisted diagnosis system as a "super assistant" to read films can improve screening efficiency, reduce the workload of cytopathologists, and better become a "capable helper" of cell detectives.

How does AI work?

Using deep learning algorithms, AI can identify abnormal patterns in cell images like a highly trained detective, and gradually improve its accuracy through continuous learning. Imagine that when a pathologist gets a slice, he only needs to use a microscope to observe whether the cell is cancerous.

However, if the diseased tissue is unevenly distributed on the slice, the doctor needs to move the slice repeatedly to observe. This is not only time-consuming and laborious, but may also lead to missed diagnosis. AI can solve this problem very well.

First, it can divide large images into blocks, then automatically track and identify the lesion area, zoom in and analyze it, and finally complete the case-level interpretation and output the results. In this way, doctors can complete the film reading work more efficiently and improve the accuracy of diagnosis.

3. AI’s “superpower”: more meticulous than doctors?

• The “main force” for improving efficiency

If you have a super efficient copy machine that can copy thousands of documents in a few minutes, such a "workhorse" is also indispensable in cell inspection. AI is like this "copy machine". It can process thousands of cell images in a very short time, greatly reducing the workload of doctors and making the inspection much faster.

Through continuous training and learning, AI is like an extremely meticulous "explorer" that can examine thousands of cell images in a few minutes and discover those tiny changes, just like carefully examining every detail with a magnifying glass. Its level of detail exceeds that of traditional methods and can accurately identify tiny differences in cells, which are often difficult for doctors to detect. AI has demonstrated a more meticulous "super detective" quality than doctors in cell examinations.

△Analysis of the results of three film reading methods

Of course, AI also has its "little flaws". It performs quite well in positive detection, just like a detector that can accurately find "hidden treasures". But AI occasionally makes mistakes in false positives (falsely judging healthy cells as abnormal) and false negatives (falsely judging abnormal cells as healthy). It's like a "super detective" sometimes mixes up some small details. However, when AI and pathologists join forces, it is a powerful combination, the number of errors is significantly reduced, and the performance is even better.

4. The “golden partnership” of AI and doctors: a perfect combination to fight together

AI cannot completely replace doctors

Although AI is powerful, it is more of an auxiliary tool. The professional judgment and experience of doctors are still the key, and the cooperation between AI and doctors can achieve the best results. Now, the cooperation between AI and pathologists is like a superhero team.

Studies have shown that when AI and doctors work together, their performance is almost impeccable. AI is particularly good at identifying tiny abnormalities in cells, like a "magnifying glass" in the cell world, able to see details that traditional methods cannot see.

The perfect partner for the future

In the future, as AI technology continues to advance, its collaboration with pathologists will become even closer. AI will not only continue to improve its performance in positive tests, but will also become more accurate in false positives and false negatives. Together with his doctor partners, this "super detective" will continue to explore the world of cells and promote the charm of medicine.

5. A new journey after the revelation

The introduction of artificial intelligence (AI) technology has ushered in a new era for cervical liquid-based thin-layer cytology. 90% of AI pathology applications are concentrated on cervical cancer. The incidence and mortality of cervical cancer in China ranks sixth. Cervical cancer is a common gynecological malignancy and the only cancer that can be eliminated through early prevention and treatment.

The incidence of cervical cancer in China is increasing. In recent years, the country has promoted the screening of "two cancers". AI pathology-assisted cervical cancer screening can effectively improve the screening efficiency. With the assistance of AI, the efficiency and accuracy of screening have been greatly improved.

Let us look forward to the future, when this detective and the "Sleepless Sherlock Holmes" will continue to uncover more secrets of health. This is also a big step towards a more efficient and effective medical system, safeguarding the future of health for us humans.

References:

[1] Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J]. CA: a cancer journal for clinicians, 2021, 71(3): 209-249.

[2] Cohen PA, Jhingran A, Oaknin A, et al. Cervical cancer[J]. The Lancet, 2019, 393(10167): 169-182.

[3] Canfell K, Kim JJ, Brisson M, et al. Mortality impact of achieving WHO cervical cancer elimination targets: a comparative modeling analysis in 78 low-income and lower-middle-income countries[J]. The Lancet, 2020, 395(10224): 591-603.

Author: Ni Bingyan, Lanjiang Branch of Lanxi People's Hospital;

Xue Shuman, Renji College, Wenzhou Medical University.

Instructor: Ma Jun, The First Affiliated Hospital of Wenzhou Medical University.