AI can read and diagnose mammograms just as well as humans

Breast cancer screening, normally through mammograms, is a critical tool in early detection. However, it’s not always flawless.

Mammographic screenings, while imperative, sometimes miss their mark, leading to false-positive interpretations. Such inaccuracies can put women who don’t have cancer through unnecessary, and often stressful, imaging and biopsies.

One solution that has been proposed over the years to enhance the accuracy of these screenings is the so-called double reading, where two human experts interpret each mammogram. However, with a challenging backdrop of limited resources and occasional reader shortages, the method isn’t always feasible.

AI does a good job of reading mammograms

Enter the world of artificial intelligence (AI). AI algorithms have increasingly been seen as the answer to many such medical challenges, and breast cancer screening is no exception. But how does AI fare against human experts in this delicate and critical domain?

A comprehensive study led by Prof. Yan Chen, Ph.D., a professor of digital screening at the University of Nottingham, sought to answer this very question. Their findings, recently published in Radiology, the journal of the Radiological Society of North America (RSNA), provide illuminating insights into the potential of AI in mammographic screenings.

According to the researchers, employing two readers can amplify cancer detection rates by 6 to 15%, all the while maintaining low recall rates. Yet, this approach isn’t without its challenges. Prof. Chen stated, “There is a lot of pressure to deploy AI quickly to solve these problems, but we need to get it right to protect women’s health.”

How the study was conducted

The research framework centered around the Personal Performance in Mammographic Screening (PERFORMS) assessment. This is a quality assurance test utilized by the UK’s National Health Service Breast Screening Program (NHSBSP). This test comprises 60 challenging exams that encompass abnormal, benign, and normal findings.

 Using these test sets, the research team juxtaposed the performance of human readers with that of an AI algorithm. For each mammogram, scores from human readers were matched against the “ground truth” outcomes determined by the AI.

“The 552 readers in our study represent 68% of readers in the NHSBSP, so this provides a robust performance comparison between human readers and AI,” emphasized Prof. Chen.

In a meticulous classification, the breasts examined were categorized as 67% normal, 29% with malignancies, and 4% benign. The malignancies exhibited certain recurring patterns. Masses were predominant, followed by calcifications, asymmetries, and architectural distortions. On average, malignant lesions measured at 15.5 mm.

What the researchers learned

The crux of the study’s findings revealed a compelling parallel in performance between AI and human experts. The average sensitivity and specificity showcased by human readers were 90% and 76% respectively. Meanwhile, the AI algorithm mirrored a nearly identical performance, registering a sensitivity of 91% and specificity of 77%.

Prof. Chen concluded, “The results of this study provide strong supporting evidence that AI for breast cancer screening can perform as well as human readers.”

However, the road ahead for AI in this realm isn’t entirely devoid of challenges. Highlighting the importance of continuous research and monitoring, Prof. Chen shared her views on AI’s role in breast screening.

She stated that the ongoing large prospective clinical trials would provide more clarity. She further emphasized the importance of consistent performance monitoring for AI, especially as algorithms might experience performance drifts due to evolving operating environments.

Closing her remarks, Prof. Chen asserted, “It’s vital that imaging centers have a process in place to provide ongoing monitoring of AI once it becomes part of clinical practice.”

This study may well pave the way for an AI-driven future in breast cancer screenings, reinforcing the symbiotic relationship between human expertise and artificial intelligence.

More about mammograms

Mammograms serve as a front-line defense in breast cancer detection. This breast imaging test uses low-dose X-rays to visualize the internal structure of breasts, allowing doctors to spot tumors that may not be palpable and diagnose changes in the breast.

So, how does a mammogram work? When you go for a mammogram, a technician places your breast between two plates. These plates then compress the breast, ensuring a clear view of the breast tissue. An X-ray machine captures images of the breast from multiple angles. Though the compression might cause some discomfort, it lasts only a few seconds.

There are two types of mammograms:

Screening Mammogram: Routinely used for women who show no symptoms or signs of breast abnormalities. It typically captures two images of each breast.

Diagnostic Mammogram: Used when a woman shows symptoms like a lump, pain, nipple thickening or discharge, or if a potential issue arises in a screening mammogram. It captures more images from various angles.

The key benefits of mammograms include:

Early Detection: Mammograms can detect tumors that are too tiny to feel and identify cancers before symptoms develop, greatly improving the chances of successful treatment.

Decreasing Mortality: Regular mammograms reduce the risk of dying from breast cancer as early detection often means simpler, less invasive treatments.

However, it’s crucial to note that while mammograms are effective, they’re not perfect. They can miss some cancers, and sometimes, they may raise false alarms, leading to additional tests that reveal no cancer.

To get the most accurate results, women should:

Choose a Reputable Facility: Ensure you get your mammogram at a center with a strong reputation and certified technicians.

Maintain Regularity: Schedule regular mammograms. If you’re 40 or older, or have a family history of breast cancer, discuss the best timeline with your doctor.

Stay Consistent: If possible, go to the same facility every time, so your mammograms are easy to compare year-to-year.

Report Any Changes: Always inform your doctor about any changes in your breasts or any potential issues.

In summary, mammograms play a vital role in breast health. They provide a powerful tool for early detection of breast cancer, helping save countless lives every year. Ensure you prioritize this crucial test and take proactive steps for your health.

—

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.

—–

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—