AI predicts 38% of breast cancers up to 2 years prior diagnosis

PLUS: Breast cancer, Sarcoma, Lung cancer, Liver cancer

Hey all!

Welcome to another edition of the Cancer Innovations newsletter.

I am sorry I did get time to send one last week. The death of my grandmother and everything around it took a lot more time than I expected. 

For those new here, my grandmother died (of age/cancer), and last week was, therefore, quite hectic and out of the ordinary. 

And if you missed the last newsletter/first edition, where I introduced myself, you can check it out here: https://cancerinnovations.beehiiv.com/p/ai-detects-skin-cancer-100-accuracy 

Alright, here is the latest on cancer innovations and AI. 

- Thomas

Today’s topics

  1. AI grades the severity of rare cancer in 82% of cases

  2. Roche uses AI to find lung cancer patiens with ALK

  3. AI predicts 38% of breast cancers up to 2 years prior diagnosis

  4. AI used in liver cancer treatments in the UK

Tags: Sarcoma, Stomach cancer, Rare cancer, Early detection
AI grades the severity of rare cancer in 82% of cases 

Researchers at Royal Marsden Hospital and the Institute of Cancer Research in London, UK, have undertaken a retrospective multicohort analysis** in retroperitoneal sarcoma**. 

I know that’s a mouthful… Nonetheless, their findings are both interesting and promising. 

In short, they studied 170 patients and used AI to analyze CT scans of these patients with confirmed cancer. The AI was 82% accurate in their cancer detection, vs. the traditional lab analysis, which was only 42% accurate. In other words, the AI was twice as effective in diagnosing this rare and hard-to-detect cancer form. Why? Because AI can recognize details that are invisible to the naked eye.

This study also gives optimism about AI’s potential in the early detection of other types of cancers. 

[You find links to the article and research paper below]

** Retrospective multicohort analysis involves looking back at data from multiple groups to understand patterns, relationships, or outcomes that have already occurred. This type of study can be useful for identifying trends, comparing different populations, or understanding the impact of a particular exposure or intervention across various groups.

** Retroperitoneal sarcoma is a type of cancer that arises in the retroperitoneum, which is the space in the abdomen behind the peritoneum (the lining of the abdominal cavity).

** The retroperitoneum is an anatomical space in the abdominal cavity that lies behind (retro) the peritoneum. The peritoneum is the thin membrane that lines the inside of the abdomen and covers most of the abdominal organs.

Link to research paper: The Lancet 
News coverage: BBC

Doctors

Institutions mentioned: 

Tags: Lung cancer, ALK, Early detection
Roche uses AI to find lung cancer patiens with ALK

A recent study found that Roche’s drug Alecensa cut the risk of cancer recurrence or death by 76% compared with standard chemotherapy after lung tumor surgery. The study examined the effects on people with an error in a gene called ALK, which is found only in about 4% to 5% of patients. However, it is hard to identify this subset of lung cancer patients and the cancer goes undetected. Roche hopes that with the help of AI this will change and people at risk can get help earlier. 

Link to research paper: Clinicaltrails.gov
Link to press release: Roche
News coverage: FortuneWell

Doctor’s mentioned

Institutions mentioned: 

Tags: Breast cancer, early detection
AI predicts 38% of breast cancers up to 2 years prior diagnosis

A retrospective study by Norwegian scientists find that AI have high potential in early breast cancer detection. The research revealed that AI could predict one-third of breast cancer cases up to two years before diagnosis. The study analyzed imaging data and screening information from BreastScreen Norway exams conducted between January 2004 and December 2019. Women later diagnosed with breast cancer were assigned an AI risk score using a commercially available AI system. The scores ranged from 1-7 for low-risk malignancy, 8-9 for intermediate risk, and 10 for high-risk malignancy. The AI scores, along with mammographic features like calcifications and breast density, were evaluated in 2,787 screening exams from 1,602 women, averaging 59 years of age. The results showed that over 38% of detected cancers had a high AI risk score before a breast cancer diagnosis. 

Studies like these gives hope for developing processes that can help with early detection of breast cancer, which again can lead to less agressive treatment. 

Link to research paper: Radiology
News coverage: Fox News

Doctor’s mentioned:

Institutions mentioned: 

Tags: Liver cancer, thermal ablation
AI used in liver cancer treatments in the UK

In Cambridge, UK, doctors have used an AI system to help with a thermal ablation procedure** for liver cancer. The AI is enriched with data from thousands of patients and enhances the accuracy of outlining the tumor and the surrounding healthy tissue. The benefits of the system is its precision and the ability to immediately determine if enough tissue has been taken. AI in thermal ablationfor kidney and lung tumors  is already employed across Europe. This is a first in the UK and it will be discussed and observed as a treatment form moving forward. 

** Thermal ablation is a method where a needle or probe is used to destroy small tumors with heat, offering a less invasive alternative to traditional surgery.

Link to article: BBC

Doctor’s mentioned

Institutions mentioned: