To prevent one death from lung cancer, 320 people need to be screened annually. With that number as a starting point, I came up with this (very rough) calculation of the cascade of benefit and harm from screening for lung cancer using low-dose spiral computed tomography, or CT. It’s an imperfect screening test that raises so many problems that it’s probably not worthwhile unless you are at very high risk of developing lung cancer, i.e., you are 55 to 80 years old and have been smoking the equivalent of a pack of cigarettes a day for 30 years.
320 people must be screened annually to prevent one death from lung cancer. 125 of them will test positive and require further diagnostic imaging and in some cases biopsies or even surgery. Each dot in the following illustration represents one person:
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • • •
120 positive screenings will prove to be false alarms.
• • • • • 5 will be diagnosed with lung cancer.
• • • 3 will die despite early detection.
• 1 will be overdiagnosed and treated unnecessarily.
• 1 life will be saved.
__________
Notes:
Keep in mind that this infographic is a work in progress. The false positive and other rates come from The National Lung Screening Trial, except for the estimates of the share of screening-detected cancers that are ‘overdiagnosed.’ (Some tumors are destined to grow slowly or regress and people would be better off if they didn’t treat them, but it’s not yet possible to predict if a screening-detected tumor will become dangerous so doctors treat every case as potentially life threatening.) The ‘number needed to screen’ helped me think about the tradeoffs as I worked on a piece about the conundrum of screening for lung cancer.
Among 320 people screened annually for three years, you could expect:
125 people will get a positive screening test (39.1 percent of those screened in NLST)
24 will have scans that reveal ‘incidentalomas’ – abnormalities other than lung cancer that trigger further cascades of diagnostic testing with unknown benefit (7.5 percent of those screened in NLST).
Among the 125 with positive screening tests:
• Nearly all will undergo repeated CT scans for 1-2 years; 3 will undergo a biopsy (2.2 percent), and 5 will undergo a surgical procedure (4.2 percent) to find out if they have cancer.
• 120 people will find out that the screening result was a false positive (96.4 percent)
• 2 will suffer a serious complication from the diagnostic workup (1.4 percent)
• 5 will be diagnosed with cancer
Among the 5 diagnosed with cancer:
• 1 person will avoid death from cancer
• 3 people will die from their lung cancer despite screening
• 1 person will be overdiagnosed and be treated for a disease that would never have caused a problem had it not been detected by screening. (20 to 25 percent in Harris et. al.)
Sources:
Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening, The National Lung Screening Trial Research Team, N Engl J Med 2011; 365:395-409
The Harms of Screening: A Proposed Taxonomy and Application to Lung Cancer Screening. Harris RP, Sheridan SL, Lewis CL, et al. JAMA Intern Med. 2014;174(2):281-286
joe rojas-burke 