Is screening bij longkanker zinvol? Dr. Karl Govaert 1
LUNG CANCER Leading cause of cancer-related death among men and women in US Second leading cause of cancer in US 75% present with advanced local or metastatic disease 5-year survival 16% 60% in stage I 5% in stage IV Cigarette smoking is causal in 90% of lung cancer 2
Voorwaarden tot screenen Hoge morbiditeit en mortaliteit Significante prevalentie (0.5-2.2%) Gekende risicofactoren waardoor risicopopulaties (rokers) kunnen geïdentificeerd worden die moeten gescreend worden Lange preklinische fase Aanwijzingen dat behandeling meer efficiënt is in vroeg ziektestadium 3
BIAS Lead time bias Overlevingsduur die wordt berekend vanaf het moment van de diagnose, is langer als de diagnose vroegtijdig wordt gesteld, maar dat vertaalt zich niet noodzakelijk in een langere globale overleving Length time bias Screening detecteert bij voorkeur traag groeiende tumoren, die de beste prognose bieden Overdiagnosis Screening detecteert tumoren die nooit klinisch aan het licht zouden zijn gekomen Volunteer bias 4
Lead-time bias 17/10/2006 5
Length-time bias
Overdiagnosis in cancer screening
Screening with X-ray/sputum cytology (1) John Hopkins Study n > 20,000 Annual chest x-rays, half randomly + sputum cytology Follow up 5-8 y No difference in lung cancer incidence or mortality Frost e.a., Am Rev Resp Dis 1984; 130:549 8
Screening with X-ray/sputum cytology (2) Mayo Lung Project (NCI) N = 10,993 6y-program of chest x-ray + sputum cytology every 4m control group: usual care + annual chest x-ray 20y follow up Baseline: 91 ca (0.8%) After 6y: 206 new ca in screening group, 160 in control group More early cancers in screened cohort but no reduction in late stage cancers Death rates significantly higher in screened group Follow up data suggests overdiagnosis Flehinger ea., Cancer 1993; 72:1573 9
Screening with CT: observational cohort studies (1) ELCAP 1000 asymptomatic pat. with >10 pack-y X-ray + low dose CT CT detected more malignant nodules (2.7 vs 0.7%) CT detected more benign nodules (20.6 vs 6.1%) 27 lung cancer, 85% stage I Henschke ea., Lancet 1999; 354:99 I-ELCAP 5 countries, 38 centers 31,567 asymptomatic pat Annual CT 484 lung cancer, 85% stage I 10-y survival 80% overall, 88% in stage I Henschke ea., N Engl J Med 2006; 355:1763 10
Screening with CT: observational cohort studies (2) Conclusion CT did increase the number of lung cancers diagnosed and the number of chirurgical resections Lung cancer detected through CT are very early stage and highly treatable No similar reduction in detection of late stage cancers Biology of small cancers detected through screening may not be the same as clinically detected lung cancers (lenght time bias) CT did not decrease predicted lung cancer death 11
Screening - van Klaveren ea.,n Engl J Med 2009; 361:2221
Ongoing randomised trials (1) NELSON (1) 7,557 current or former smokers Tot 2015 in Nl en B Volume CT baseline, 1y, and 3y later >< no screening Negative: <50 mm3 or 50-500 mm3 but no growth after 3m or doubling time >400 days Positive: >500 mm3 (>9.8mm diam) Growth: increase in volume at least 25% Powered to 25% mortality reduction van Iersel ea., Int J Cancer 2007; 120:868 van Klaveren ea., N Engl J Med 2009; 361:2221 14
Ongoing randomised trials (2) NELSON (2) 50% noncalcified pulmonary nodules Baseline positive CT: 1.6% 0.9% detection rate lung cancer 63.9% stage I Sensitivity 94,6%, specificity 98.3%, PPV 35,7%, NPV 99,9% 27.2% of invasive procedures revealed benign disease Among 7361 negative scans, 20 lung ca detected during 2 years of follow up) 1/1000 1y after negative base line CT 3/1000 2y after negative base line CT Van Iersel ea., Int J Cancer 2007; 120:868 Van Klaveren ea., N Engl J Med 2009; 361:2221 15
Ongoing randomised trials (3) NLST (National Lung Screening Trial) 50.000 current or former smokers VS Annual CT >< annual chest x-ray for 3y Powered to 20% mortality reduction Results to be published in 2011 Gohagan ea., Chest 2004; 126:114 16
Ongoing randomised trials (4) Dante Italy 2472 male smokers 60-74 y Lung cancer-specific mortality over 10y 5y annual CT Control group baseline screening with chest x-ray and sputum cytology Initial evaluation: Lung cancer in 2.2% (71% stage I) of CT group vs 0.67% in control group (50% stage I) 4% invasive procedures, benign pulmonary lesions in 19% 33.7m follow up: lung cancer in 4.7% (54% stage I) vs 2.8% (34% stage I) More stage I in CT group, but no difference in advanced lung cancer or mortality Infante ea., Am J resp Crit Care Med 2009; 180:445 17
Problemen bij screenen Vals positieven (benigne nodules) Aanvullende (invasieve) procedures zoals bv. bronchoscopie, transthoracale punctie houden morbiditeitsrisico in Screening enkel bij operabele patiënten (COPD)? Minderheid van vrouwen in screeningsstudies Meer adeno s Traaggroeiende tumoren Perifeer, dus makkelijker en sneller zichtbaar 18
OTHER FUTURE TECHNOLOGIES Immunostaining/molecular analysis of sputum for tumor markers (p53) Image cytometry of sputum Fluorescence bronchoscopy Exhaled breath analysis of volatile organic compounds Genomic or proteomic analysis of bronchoscopic samples Serum protein microarrays for detecting molecular markers 19
BESLUIT In afwachting van resultaten van gerandomiseerde studies, kan screening van longkanker NIET verdedigd worden buiten kader van klinische studies Preventie (rookstop) is belangrijker dan screening om incidentie van longkanker te verminderen 20
The reason for not screening on lung cancer today is.....that so far no randomized controlled study has demonstrated that screening leads to a reduction in lung cancer mortality