Who Should Be Screened for Lung Cancer? Michael K. Gould, MD, MS Senior Scientist and Leader, Care Improvement Research Team Director for Health Services Research and Implementation Science Department of Research and Evaluation Kaiser Permanente Southern California
DEPARTMENT OF RESEARCH AND EVALUATION
Disclosures Salary support from Archimedes, Inc. to help develop computer models of lung cancer screening Grant support from NIH/NCI to evaluate computerized decision support for lung nodule evaluation
Member, ACCP/ASCO lung cancer guidelines panel Guest member, MEDCAC
Objectives Discuss potential harms and benefits of LDCT screening and how they may be different for individual patients Provide information about existing lung cancer risk assessment models
Preview Lung cancer screening with low-dose CT (LDCT) prevents a substantial fraction of deaths from lung cancer in high-risk smokers and former smokers Like other screening interventions, LDCT has potential harms, including false positive test results and downstream invasive testing Balance of pros and cons likely differs across individuals; decision-making should be tailored Both elderly (≥65) and non-elderly candidates for screening face similar tradeoffs
Case History Healthy, asymptomatic 76 year-old woman
Former smoker – 1 pack per day for 25 years – Quit ~30 years ago
Underwent screening with LDCT after seeing advertisement for lung cancer screening at her community hospital Did she meet NLST or USPSTF criteria?
Case History Healthy, asymptomatic 76 year-old woman
Found to have a 11 mm, solid NCN in LUL PET scan showed hypermetabolic focus Underwent VATS wedge resection Cured
How do we help each individual make the right decision about screening FOR HIM or HERSELF?
Potential Benefits of Screening Reduced lung cancer mortality Reduced all-cause mortality Others – Reduced morbidity from lung cancer treatment – Reduced morbidity from other illnesses – Teachable moment for smoking cessation
– Reassurance if results are negative
Potential Harms False positive test results – Unnecessary follow-up imaging – Unnecessary invasive biopsy or surgery – Complications from unnecessary procedures
Over-diagnosis Exposure to ionizing radiation
Other – Inconvenience – Anxiety/emotional distress
– False reassurance if results are negative
History of Lung Cancer Screening 1950s and 1960s: – Lung cancer epidemic first recognized – First uncontrolled studies of CXR and sputum cytology
1970s: 4 large randomized controlled trials of CXR and sputum in US and Czechoslovakia – Mayo Lung Project compared intensive screening (CXR and sputum every 4 months) with “usual Mayo advice”
No evidence that these screening interventions reduced premature death from lung cancer
Uncontrolled study of screening with low radiation-dose CT
23% of 1,000 volunteers found to have at least 1 nodule 2.7% found to have cancer, most stage I Low risks of invasive testing, complications
NLST Methods Design: multi-site, unblinded RCT Participants: >53,000 high risk smokers and former smokers, age 55-74 – ≥30 pack-years, quit within 15 years
Intervention: annual LDCT for 3 years Comparator: annual CXR for 3 years Outcomes: mortality (lung cancer-specific and all-cause) over 6.5 years Settings: 33 U.S. centers, mostly academic Powered to detect 20% reduction in LCspecific mortality NLST Research Team. NEJM 2011;365:395-409.
NLST Results: Benefits Outcome
CT
CXR
Difference
N=26,722
N=26,732
Per 1,000
Lung cancer death
1.33%
1.66%
3 to 4 fewer*
Any death
7.0%
7.5%
4 to 5 fewer*
Non-lung cancer death
5.7%
5.8%
1 fewer
Lung cancer cases Other positive finding Smoking cessation Relapses
4.0% 7.5%
3.5% 2.1%
4 to 5 more* 54 more NR NR
7.2% 4.7%
NLST Results: Harms Outcome
CT
CXR
Difference
N=26,722
N=26,732
Per 1,000
One or more positive screening results
39.1%
16.0%
231 more*
Positive result on baseline screening exam
27.3%
9.2%
181 more*
NLST Results: Follow-Up Testing in CT Arm Outcome
Lung Cancer Confirmed
Not confirmed
N=649
N=26,073
618 (95%)
457 (1.8%)
Percutaneous biopsy
≥33
≥66
Bronchoscopy
≥76
≥227
509 (82%)
164 (0.6%)
170 (26%) Death
