Dorith Shaham, M.D. Department of Radiology Hadassah Hebrew Uviversity Medical Center
LUNG CANCER The most common fatal malignancy in the Western world Estimated 1.3 million deaths worldwide in 2000 In the US 222,520 new cases, 157, 300 deaths in 2010 more deaths than breast, prostate, cervix and colon cancer combined In Israel 1,897 new cases and 1,561 deaths in 2008
Lung cancer: Leading cause of death in the Western world Overall survival: 13% Virtually no improvement over the last 40 years Stage I lung cancer: Cure rate of 70%
Overview CXRs and their role in lung cancer screening Prospective RCTs performed in the 1980 s Principles of LDCT screening for lung cancer Results of LDCT trials I ELCAP Hadassah NLST Cost effectiveness of LDCT screening for lung cancer Future of LDCT screening for lung cancer Research Clinical
CXR Inexpensive, readily available In Hadassah: 6500/month Lesions =>1 cm are usually detected (smaller lesions if calcified) Improve diagnostic accuracy PA + LAT films Comparison with old films
1999
CAD: Experimental Results
Algorithm Overview Lung Segmentation Preprocessing Candidate Generation Feature Extraction / Selection Classification Chest CAD + X-Ray AP view Detect and tag anatomical areas Filter Identify ROI For the probable suspicious region Extract features Apply rule logic to select features Classify Nodule and Non Nodules irrelevant Nodules + suspicious region
53 year-old heavy smoker (37 pack years) with fever, productive cough, left chest pain. CXR 9/5/10: LLL consolidation, loculated left pleural effusion.
Spiculated RUL nodule, 12X9mm
Prospective RCTs on lung cancer screening Used CXR and sputum cytology Mayo lung project Memorial Sloan Kettering Lung Project Johns Hopkins Lung Project Czechoslovakian study Only the Mayo and Czech studies evaluated CXR No statistically significant difference in mortality in a population screened by CXR compared to control Fontana et al., Am Rev Respir Dis 1984 Melamed et al., Chest 1984 Tockman et al., Chest 1986 Kubic et al., Cancer 1986
International Early Lung Cancer Action Program (I ELCAP) Initially 2 institutions ; now >50 institutions world wide Investigator team at each institution Radiologists, pulmonologists, oncologists, thoracic surgeons, pathologists, epidemiologists and a computer science team Screening for lung cancer using the same protocol Share knowledge about lung cancer screening with the goal of early detection and reducing mortality worldwide
ELCAP:Baseline Findings (started: 1993) 1000 high risk participants underwent baseline LDCT and chest radiograph, and one annual repeat LDCT. Age: 60, smoking: 10 pack years 27 lung cancers diagnosed at baseline 23 (83%) stage I median diameter: 15 mm Henschke et al, Lancet 1999; 354:99-105
ELCAP:Baseline Findings (started: 1993) Of the 23 stage I lung cancers 19 (83%) were missed on chest radiographs. 23% had positive result of the initial LDCT Only one patient who underwent recommended biopsy had a benign lesion Henschke et al, Lancet 1999; 354:99-105
ELCAP:Annual repeat findings Smaller lung cancers Median diameter: 8 mm 85.7% were stage I < 3% with positive test result Henschke et al, Cancer 2001; 92:153 159
CT Screening: Baseline & Annual Repeat Baseline Year 1 Year 2 Year 3.. Prevalence Annual incidence, usually pooled in screening
Screening for lung cancer: The I ELCAP approach Study design Baseline/ annual repeat screening Regimen of screening
I ELCAP Approach: Diagnostic Prognostic Trial Screen all Stage I Stage II Stage III Stage IV Early Rx Delayed Rx Early Rx Delayed Rx Early Rx Delayed Rx Early Rx Delayed Rx Fatality Rate Early Rx by Stage Fatality Rate Delayed Rx by Stage Diagnostic Mission Prognostic Mission Each, specific to stage and size
The regimen of screening Population to be screened (age, smoking history) Screening positives Baseline Annual repeat Work up
Prevalence of malignancy Dependent on risk profile of participants Original ELCAP: 60 years of age and older with a smoking history of 10 pack years or more I ELCAP: 40 years of age and older with any smoking history This needs to be considered when making recommendations for biopsy
Prevalence of malignancy on baseline low dose CT in ELCAP ( 60 yo, 10 pack years) <5 5-9 10-14 15+ Total # people 99 46 9 5 159 # malignant 1 11 3 4 19 %malignant 1 24 33 80 12 12% of all participants with nodules had a malignancy
Prevalence of malignancy on baseline low dose CT in I ELCAP ( 40 yo, any smoking) <5 5-9 10-14 15+ Total # people 5344 3221 576 445 9586 # malignant 20 118 118 186 442 %malignant 0 4 20 42 5 5% of all participants with nodules had a malignancy
Prognostic significance of nodule type Solid Nonsolid Part solid
Prognostic significance of nodule type Three fold malignancy rate in part solid nodules compared to solid/nonsolid nodules (Henschke et al, AJR Am J Roentgenol. 2002; 178: 1053 7)
Positive result : Baseline screening Any solid or part solid NCN > 5.0 mm in diameter or any nonsolid nodule > 8 mm in diameter Nodules <5 mm: highly unlikely to present with malignancy during the first cycle of screening (Henschke et al, Radiology 2004; 231: 164 8) Between 12 15% had a positive test result on baseline screening at any given institution
Positive result: Annual Repeat Newly detected or growing non calcified nodule Based on comparison with previous CT Approximately 6%
Diagnostic work up of screening positives Repeat LDCT Growth? Antibiotics followed by CT at 1 month PET scanning Biopsy I-ELCAP protocol available at http://ielcap.org
I ELCAP results (N Engl J Med 2006;355:1763 71) 31,567 asymptomatic persons at risk for lung cancer screened using low dose CT (1993 2005) Stage I lung cancer diagnosed in 412/484 (85%) 10 year survival in stage I lung cancer Overall: 88% Surgical resection in 1 month: 92%
HM ELCAP results (started 1998) Total Enrollment as of Jan 31 st, 2011: 1080 56% males, 44% females. Mean age: 57±12 Mean pack years of smoking: 39±27.05 Smoking Hx Current smokers: 71% Former smokers: 26% Never smokers: 3% Positive studies Baseline: 12.1% Annual: 5.4%
HM ELCAP results: Detected cancers Clinical stage I II III IV Total Baseline 11 0 1 1 13 Annual 3 0 0 1 4 Two additional patients discontinued screening and were diagnosed with advanced lung cancer elsewhere
National Lung Screening Trial (NLST) (Started in 2002) Press release: Nov. 4, 2010 >53,000 current and former heavy smokers, ages 55 to 74 compared the effects of two screening procedures for lung cancer low dose helical computed tomography (CT) standard chest X ray 20% fewer lung cancer deaths among trial participants screened with low dose helical CT Lung cancer deaths in CT screened: 354, in CXR screened: 442 (p=0.0041)
Cost effectiveness of LDCT screening in Israel Screening arm: 842 smokers and past smokers, 45 years screened at Hadassah in 1998 2004 Usual care arm Stage distribution and stage specific life expectancy 2,906 patients diagnosed in 1994 2006 (NCI) Lifetime stage specific costs medical records of 146 patients diagnosed and treated at Hadassah in 2003 2004 The analysis took into consideration possible screening biases such as lead time, overdiagnosis, and self selection
Cost effectiveness of LDCT screening in Israel The cost per LY or QALY gained by screening was about $20 The results of all the sensitivity checks confirmed the low cost per LY or per QALY and, in some cases, the dominance of screening LDCT screening for lung cancer in Israel provides a good economic value under the common standards of health technology assessments.
Future Research Imaging Lowering radiation exposure Software: Detection and characterization of nodules Volumetric measuring
Future Research Blood/sputum biomarkers Increased risk Low OCG (DNA repair enzyme 8 oxoguanine DNA N glycosylase) activity is associated with increased risk of lung cancer (Paz Elizur et al, J Natl Cancer Inst 2003; 95:1263 5)
Is LDCT screening ready for clinical use in Israel? Effectiveness was confirmed: LDCT screening saves lives! Cost effectiveness in Israel was confirmed Well established protocol prevents unnecessary interventional work up Radiation exposure in the same range as mammography
To be determined Who should be screened? Who will pay? Guidelines for performing and interpreting screening LDCTs should be prepared and distributed