🧠 AI-Powered Cancer Detection: Effective and Economical

Subscribe | Contact | Meet The Team

Welcome back AI prodigies!

In today’s Sunday Special:

• 🔬Cancer Detection 101

• 🦠Metabolites > Biomarkers

• 🧪Machine Learning Madness

• 🔑Key Takeaway

Read Time: 8 minutes

🎓Key Terms

• Biomarker: a measure of what’s happening in a cell or an organism at a given moment.

• MicroRNAs: molecules that manage gene expression to regulate cellular processes such as development, metabolism, and immune function.

• Glycolysis: when glucose (i.e., sugar) is partially broken down by cells to produce energy.

• Dried Serum Spots (DSS): a yellowish fluid called “serum” that appears after blood is clotted. This “serum” contains compounds used to diagnose cancer.

🩺 PULSE CHECK

🔬CANCER DETECTION 101

With over 2 million diagnoses per year, cancer is the second leading cause of death in the United States (U.S.). The average person knows about 600 people. In a country of 340 million citizens, you almost certainly know someone with cancer. In a given year, males and females under the age of 49 have a 3.5% and 5.9% chance of developing invasive cancer, respectively. This discrepancy results from higher rates of female-specific cancers like breast cancer in younger women and higher rates of male-specific cancers like prostate cancer in older men. In recent years, early-onset cancers have increased by 79% for those under the age of 50, making early detection even more critical.

Conventional cancer detection methods rely on a variety of diagnostic tools and techniques. These methods can be broadly categorized into Imaging Tests, Laboratory Tests, Biopsies, and Endoscopic Examinations. Here’s a universal breakdown of each method:

1. Imaging Tests: An assessment that leverages various forms of energy to make detailed pictures of areas inside the body. For example, ultrasounds use sound waves to monitor fetal development in pregnant women.

2. Laboratory Tests: A medical procedure that involves testing a sample of a patient’s blood, urine, or other bodily tissue to determine if their health is normal.

3. Biopsies: A medical procedure that involves removing a piece of tissue or a sample of cells from the body to examine it for abnormalities.

4. Endoscopic Examinations: A procedure that requires inserting a thin, flexible tube with a light and a camera (i.e., an endoscope) through a natural opening (e.g., mouth, anus, or urethra) to examine the inside of organs.

Many techniques, especially Biopsies and Endoscopic Examinations, are invasive and can be uncomfortable or painful. Repeated Imaging Tests like Computed Tomography (CT) scans can expose patients to significant levels of radiation, which carries its own risks. Consequently, there’s often a tradeoff between invasiveness or convenience and efficacy. But there doesn’t have to be. 

🦠METABOLITES > BIOMARKERS

💉Blood Samples?

Though cheap, effective, and widely accepted, taking various blood samples to detect cancer isn’t optimal. It’s estimated that 20% to 30% of young adults fear needles to the point of skipping routine vaccinations. Additionally, contaminated needles transmit infections, resulting in 1.3 million global deaths per year. Furthermore, current blood tests are unreliable for diagnosing certain cancers due to the lack of approved biomarkers. For instance, traditional biomarkers like microRNAs are unstable during transportation or with slight temperature changes, inhibiting test accuracy.

🧫Blood Spot Tests?

On the other hand, blood spot tests are far less invasive, only requiring a few drops of blood. You may recall a biomedical devices startup named Theranos, which promised to perform hundreds of tests using a single drop of blood from a patient’s fingertip. Although Theranos’s offerings were fraudulent, blood spot tests aren’t. By taking a few drops of blood from the heels of newborns, healthcare providers known as midwives can test for over 60 rare conditions. However, blood spot tests often fall short when used to detect cancer-specific biomarkers because they can be interfered with by other substances present in the blood.

🦠Metabolites?

Metabolites, or small molecules produced during metabolism to create energy for growth, reproduction, and health, have emerged as promising biomarkers for cancer detection. Metabolites remain stable on dried blood spots without degrading over time, enhancing diagnostic accuracy. Metabolites also offer a glimpse into the physiological state of a cell or tissue.

Think of metabolites as advanced basketball statistics like True Shooting Percentage (TS%) or Player Efficiency Rating (PER). These advanced basketball statistics reveal deeper insights into a team’s overall performance during a basketball game beyond the final score.

Similarly, metabolites provide a more comprehensive understanding of a person’s health beyond traditional diagnostic methods. Cancer cells exhibit unique metabolic signatures that differentiate between various cancer types. For example, elevated levels of specific lipid metabolites (i.e., break down fats in cells for energy storage) might indicate breast cancer, which requires more lipid metabolites. This metabolic specificity offers a promising avenue for identifying the type and origin of cancer, leading to more targeted and effective treatments.

🧪MACHINE LEARNING MADNESS

In a Chinese study published in Nature, researchers developed a new technique that uses Nanoparticle-Enhanced Laser Desorption/Ionization Mass Spectrometry (NPELDIMS), Dried Serum Spots (DSS), and AI models to diagnose colorectal, pancreatic, and gastric cancers.

NPELDIMS uses DSS to measure a patient’s metabolic levels and diagnose multiple cancers within minutes at an affordable cost.

NPELDIMS leverages AI models to generate a forest of decision trees using a balanced Random Forest Approach: a Machine Learning (ML) algorithm that combines multiple decision trees to reach a single result.

Here’s How It Works:

1. Sample Preparation: DSS are collected from patients and mixed with microscopic particles from materials like gold, silver, and silicon with unique surface properties.

2. NPELDIMS Analysis: The patient’s DSS samples are analyzed using NPELDIMS, which involves laser ionization and mass spectrometry to separate the target molecule from other molecules.

3. Data Generation: The NPELDIMS Analysis is curated into a dataset.

4. AI Models Application: A balanced Random Forest Approach analyzes the NPELDIMS Analysis dataset, identifying patterns and correlations between the patient’s molecular profiles and previously analyzed cancer cases.

5. Cancer Diagnosis: This process provides a probabilistic prediction of the likelihood of different cancer types in patients.

If this new technique were implemented in rural China, the researchers estimated that the rate of undiagnosed cancer cases would fall from:

• 84.30% to 29.29% for Colorectal Cancer

• 34.56% to 9.30% for Pancreatic Cancer

• 77.57% to 57.22% for Gastric Cancer

Even though this new technique offers significant promise for cancer diagnosis, it’s crucial to acknowledge its potential limitations:

1. Data Quality: The accuracy of NPELDIMS relies on the quality of the collected data. Factors such as sample storage, handling, and the sensitivity of the mass spectrometry instrument can affect the results.

2. AI Training: The AI models used in NPELDIMS must be trained on a large and diverse dataset of cancer samples, which can be challenging, especially for rare cancer types.

🔑KEY TAKEAWAY

Chinese researchers have designed a simple, cheap, new technique for diagnosing cancer, minimizing invasiveness, cost, and convenience while maximizing efficacy.

Though testing on thousands of real-world patients is required before adoption, their early findings have surprised oncology experts around the world. With hundreds of thousands of undiagnosed cases of colorectal, pancreatic, and gastric cancer per year, a significant improvement in detection rates would likely save tens of thousands of lives.

📒FINAL NOTE

If you found this useful, follow us on Twitter or provide honest feedback below. It helps us improve our content.

How was today’s newsletter?

• ⭐️⭐️⭐️⭐️⭐️Nailed it!

• ⭐️⭐️⭐️Average

• ⭐️Epic fail.

❤️TAIP Review of the Week

REFER & EARN

🎉Your Friends Learn, You Earn!

You currently have 0 referrals, only 1 away from receiving ⚙️Ultimate Prompt Engineering Guide.

Refer 5 friends to enter 🎰October’s $200 Gift Card Giveaway.

Copy and paste this link to others: https://theaipulse.beehiiv.com/subscribe?ref=PLACEHOLDER