MIT researchers develop an AI model that may detect future lung cancer risk


The name Sybil has its origins within the oracles of Ancient Greece, also often known as sibyls: feminine figures who were relied upon to relay divine knowledge of the unseen and the omnipotent past, present, and future. Now, the name has been excavated from antiquity and bestowed on a man-made intelligence tool for lung cancer risk assessment being developed by researchers at MIT’s Abdul Latif Jameel Clinic for Machine Learning in Health, Mass General Cancer Center (MGCC), and Chang Gung Memorial Hospital (CGMH).

Lung cancer is the No. 1 deadliest cancer on the earth, leading to 1.7 million deaths worldwide in 2020, killing more people than the following three deadliest cancers combined. 

“It’s the largest cancer killer since it’s relatively common and comparatively hard to treat, especially once it has reached a sophisticated stage,” says Florian Fintelmann, MGCC thoracic interventional radiologist and co-author on the brand new work. “On this case, it’s essential to know that should you detect lung cancer early, the long-term final result is significantly higher. Your five-year survival rate is closer to 70 percent, whereas should you detect it when it’s advanced, the five-year survival rate is just in need of 10 percent.” 

Although there was a surge in recent therapies introduced to combat lung cancer in recent times, nearly all of patients with lung cancer still succumb to the disease. Low-dose computed tomography (LDCT) scans of the lung are currently essentially the most common way patients are screened for lung cancer with the hope of finding it within the earliest stages, when it could still be surgically removed. Sybil takes the screening a step further, analyzing the LDCT image data without the help of a radiologist to predict the chance of a patient developing a future lung cancer inside six years.

Of their recent paper published within the , Jameel Clinic, MGCC, and CGMH researchers demonstrated that Sybil obtained C-indices of 0.75, 0.81, and 0.80 over the course of six years from diverse sets of lung LDCT scans taken from the National Lung Cancer Screening Trial (NLST), Mass General Hospital (MGH), and CGMH, respectively — models achieving a C-index rating over 0.7 are considered good and over 0.8 is taken into account strong. The ROC-AUCs for one-year prediction using Sybil scored even higher, starting from 0.86 to 0.94, with 1.00 being the very best rating possible. 

Despite its success, the 3D nature of lung CT scans made Sybil a challenge to construct. Co-author Peter Mikhael, an MIT PhD student in electrical engineering and computer science, and affiliate of Jameel Clinic and the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), likened the method to “trying to seek out a needle in a haystack.” The imaging data used to coach Sybil was largely absent of any signs of cancer because early-stage lung cancer occupies small portions of the lung — only a fraction of the a whole lot of hundreds of pixels making up each CT scan. Denser portions of lung tissue are often known as lung nodules, and while they’ve the potential to be cancerous, most usually are not, and might occur from healed infections or airborne irritants.  

To be certain that Sybil would have the option to accurately assess cancer risk, Fintelmann and his team labeled a whole lot of CT scans with visible cancerous tumors that might be used to coach Sybil before testing the model on CT scans without discernible signs of cancer. 

MIT electrical engineering and computer science PhD student Jeremy Wohlwend, co-author of the paper and Jameel Clinic and CSAIL affiliate, was surprised by how highly Sybil scored despite the dearth of any visible cancer. “We found that while we [as humans] couldn’t quite see where the cancer was, the model could still have some predictive power as to which lung would eventually develop cancer,” he recalls. “Knowing [Sybil] was able to spotlight which side was the most certainly side was really interesting to us.” 

Co-author Lecia V. Sequist, a medical oncologist, lung cancer expert, and director of the Center for Innovation in Early Cancer Detection at MGH, says the outcomes the team achieved with Sybil are essential “because lung cancer screening just isn’t being deployed to its fullest potential within the U.S. or globally, and Sybil may have the option to assist us bridge this gap.”

Lung cancer screening programs are underdeveloped in regions of the USA hardest hit by lung cancer attributable to quite a lot of aspects. These range from stigma against smokers to political and policy landscape aspects like Medicaid expansion, which varies from state to state.

Furthermore, many patients diagnosed with lung cancer today have either never smoked or are former smokers who quit over 15 ago — traits that make each groups ineligible for lung cancer CT screening in the USA. 

“Our training data consisted only of smokers because this was a crucial criterion for enrolling within the NLST,” Mikhael says. “In Taiwan, they screen nonsmokers, so our validation data is anticipated to contain individuals who didn’t smoke, and it was exciting to see Sybil generalize well to that population.” 

“An exciting next step within the research will probably be testing Sybil prospectively on people in danger for lung cancer who haven’t smoked or who quit many years ago,” says Sequist. “I treat such patients daily in my lung cancer clinic and it’s understandably hard for them to reconcile that they’d not have been candidates to undergo screening. Perhaps that may change in the long run.”

There’s a growing population of patients with lung cancer who’re categorized as nonsmokers. Women nonsmokers usually tend to be diagnosed with lung cancer than men who’re nonsmokers. Globally, over 50 percent of ladies diagnosed with lung cancer are nonsmokers, in comparison with 15 to twenty percent of men.

MIT Professor Regina Barzilay, a paper co-author and the Jameel Clinic AI faculty lead, who can also be a member of the Koch Institute for Integrative Cancer Research, credits MIT and MGH’s joint efforts on Sybil to Sylvia, the sister to a detailed friend of Barzilay and one in every of Sequist’s patients. “Sylvia was young, healthy and athletic — she never smoked,” Barzilay recalls. “When she began coughing, neither her doctors nor her family initially suspected that the cause could possibly be lung cancer. When Sylvia was finally diagnosed and met Dr. Sequist, the disease was too advanced to revert its course. When mourning Sylvia’s death, we couldn’t stop pondering what number of other patients have similar trajectories.”

This work was supported by the Bridge Project, a partnership between the Koch Institute at MIT and the Dana-Farber/Harvard Cancer Center; the MIT Jameel Clinic; Quanta Computer; Stand Up To Cancer; the MGH Center for Innovation in Early Cancer Detection; the Bralower and Landry Families; Upstage Lung Cancer; and the Eric and Wendy Schmidt Center on the Broad Institute of MIT and Harvard. The Cancer Center of Linkou CGMH under Chang Gung Medical Foundation provided assistance with data collection and R. Yang, J. Song and their team (Quanta Computer Inc.) provided technical and computing support for analyzing the CGMH dataset. The authors thank the National Cancer Institute for access to NCI’s data collected by the National Lung Screening Trial, in addition to patients who participated within the trial.


What are your thoughts on this topic?
Let us know in the comments below.


0 0 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments

Share this article

Recent posts

Would love your thoughts, please comment.x