Artificial intelligence (AI) has potential to transform healthcare. One area where AI is being employed is in lung cancer screenings using CT scans. Lung cancer is the number-one cancer killer in the United States, so methods to improve the screening process hold a lot of promise. However, AI technology in this area is not without its challenges.
In this episode of Critical Point, Milliman consultant Bruce Pyensen and Rush University Medical Center’s Jim Mulsine discuss the benefits and the challenges of using AI for lung cancer detection.
More people die from lung cancer globally than any other form of cancer. The disease is expected to kill over 154,000 people in the United States alone in 2018. Most recent reports about treating lung cancer have focused on innovative treatments around immunotherapy. But an alternative form of disease management exists that has been getting press attention: catching the disease early via CT scan.
Jim Mulshine, a thoracic medical oncologist by training who spent 25 years at the National Cancer Institute, and Bruce Pyenson, a consulting actuary at Milliman, sat down for a Q&A to discuss lung cancer and provide a medical and actuarial take around finding lung cancer early via CT scan.
In this Q&A, Mulshine, now at Rush University Medical Center, and Pyenson discuss the prognosis and progression of lung cancer, various treatments for the disease, including CT screening, and what future treatment could look like.
Milliman’s actuarial research has played a key role in the establishment of an evidence-based case for lung cancer screening. The announcement by the U.S. Preventive Services Task Force (USPSTF) stating it will now recommend lung cancer screening for certain high-risk populations follows a series of actuarial studies that helped make the value case for such screenings.
The following reading list highlights the actuarial research and also provides links to other lung cancer-related content.
• An actuarial approach to comparing early stage and late stage lung cancer mortality and survival (subscription required)
This actuarial analysis of lung cancer mortality published in Population Health Management provides evidence that early detection of lung cancer generates genuine mortality reductions not associated with lead time bias, and therefore could reduce late stage deaths by over 70,000 people in the United States each year.
• An actuarial analysis shows that offering lung cancer screening as an insurance benefit would save lives at relatively low cost
Using actuarial models, this study published in Health Affairs estimates the costs and benefits of annual lung cancer screening if offered as a commercial insurance benefit in the high-risk U.S. population, ages 50 to 64.
• Improved lung cancer screening could lead to earlier detection
In this interview, the authors and sponsors of the first actuarial analysis of lung cancer mortality discuss the broader implications their research may have in the effort to reduce deaths associated with the disease.
• An actuarial analysis of lung cancer screening
This blog post highlights Bruce Pyenson’s presentation on lung cancer screening at Health Affairs’ “Value in Cancer Care” briefing in 2012.
Bruce Pyenson presented his study on lung cancer screening at Health Affair’s “Value in Cancer Care” briefing on April 12 in Washington DC. Watch his presentation at HealthAffairs.org – scroll down to “The Case For Lung Cancer Screening of High-Risk Patients & Improving Payment And Delivery Of Cancer Care.”
Also, here is the study’s abstract:
Some of you may remember this interview from last May, which explained the results of a 2009 study for the Lung Cancer Alliance that used actuarial methods to clarify the opportunity posed by lung cancer screening.
The case for increased lung cancer screening gained more momentum last week when the Los Angeles Times reported on the results of an analysis by the National Cancer Institute. Here is an excerpt from the Times article:
Advanced CT imaging can reduce deaths from lung cancer by 20% among heavy smokers by detecting tumors at an earlier stage when they are more treatable, according to results released Thursday from the first study to compare the value of CT scans and regular chest X-rays for lung cancer screening.
The long-awaited results of the trial involving more than 53,000 former and current heavy smokers were so conclusive that the study was terminated ahead of schedule last week and letters were sent to all the participants advising them of the results.
The findings are considered a major step forward in fighting the most deadly form of cancer — which is expected to kill an estimated 157,000 Americans this year — because chest X-rays have never been proven to be an effective tool for identifying tumors. CT scans are more powerful and provide a much clearer picture of the lungs.
The Lung Cancer Alliance issued a press release today announcing results of a study on early-stage screening for lung cancer. Here is an excerpt:
The first ever actuarial analysis of lung cancer mortality, published today in Population Health Management Journal, provides strong evidence that earlier detection could reduce the number of late stage lung cancer deaths by over 70,000 people each year in the US. Calling the number “profound,” Lung Cancer Alliance (LCA) President Laurie Fenton-Ambrose said, “This would be the equivalent of eliminating all deaths from breast and prostate cancer each year. It clearly demonstrates why we must make research and development of earlier detection tools for lung cancer a public health priority.”
The study was carried out by Milliman Inc., an internationally renowned actuarial firm, and commissioned by Lung Cancer Alliance, the American Legacy Foundation, the Bonnie J. Addario Lung Cancer Foundation, Joan’s Legacy Foundation, Lungevity Foundation, the Prevent Cancer Foundation and the Thomas G. LaBrecque Foundation.
Bruce S. Pyenson, FSA, one of the co-authors of the study said, “We found that higher stage at diagnosis was profoundly associated with higher all-cause mortality and lower stage at diagnosis had profoundly lower all-cause mortality.”
“Our reporting all-cause mortality is perhaps more relevant to patients than the more common disease-specific survival or 5-year survival, as patients probably are more concerned about overall survival, not whether they face death from cancer, treatment side-effects, or something else,” he noted.
The study analyzed detailed records of over 241,000 lung cancer patients diagnosed and treated between 1988 and 2003 from the Surveillance, Epidemiology and End Results (SEER) database of the National Cancer Institute.
Mortality rates from those records were compared to demographically- and year-adjusted standard national mortality rates to develop “load” mortality ratios. These show the added mortality burden that treated lung cancer brings to patients, and how that burden dramatically increases by stage.
Read the full press release here.