Lung Cancer Screening in 2013

Cheryl Czerlanis, MD

Loyola University Medical Center

March 02, 2013

The scope of the problem

An estimated 29% of the global population older than 15 years smokes tobacco.

75% of patients with lung cancer present with symptoms due to locally advanced or metastatic disease.

Overall five-year survival for lung cancer is 16%.

Fry WA. Cancer 1999;86:1867-76.

Ten Leading Cancer Types for the Estimated New Cancer Cases and Deaths by Sex, 2013

From Seigel,R et al. CA Cancer J Clin 2013; 63(1): 14.

Five-year overall survival by stage (percent)

Goldstraw, P, et al. J Thorac Oncol 2007; 2:706.

Screening Test Principles

1. The disease must be prevalent2. There must be an advantage to early

detection3. Screening must be safe, sensitive, and

have an acceptable rate of false positives

4. Cost to society must be acceptable

Henschke CI. Radiol Clinic No Amer 2000;38(3):287-95.

Reduce mortality, improve quality of life, or both.

Screening Rationale: Early diagnosis

5–year Survival

Stage Distribution

Wagner H. Cancer Screening 1996;118-49.

Mountain CF. Chest 1997;111:1710-17.

Inoue K. Thorac Cardiovasc Surg 1998;116:407-11.

Outcomes to be assessed in lung cancer screening trials

Cancer detection rates Stage at detection Disease–specific mortality Overall survival

Mortality rate is the number of persons who die of a certain cause in a time period per population.

Survival rates calculate the percentage of persons with a disease who are still alive a set amount of time after diagnosis. 

Methodological biases

Bias related to apparent effects of early diagnosis and treatment that may inflate the benefit seen by screening Measured in terms of how screen-detected

cases compare to cases detected by signs and symptoms Lead time bias Length time bias Overdiagnosis bias Volunteer bias

Prior Attempts

Four nonrandomized uncontrolled studies Philadelphia Pulmonary Neoplasm Research

Project VA Trial Tokyo Metropolitan Government Study South London Lung Cancer Study

Two nonrandomized but controlled studies North London Cancer Study Erfurt County Study

Patz, EF Jr. NEJM 2000;343(22):1627-33.

1950’s

Four randomized trials of CXR and Sputum Cytology (SC)

Mayo Lung Project (MLP)

Czechoslovakian Study

Memorial Sloan-Kettering Lung Project

Johns Hopkins Study

1970’s

Results

Increased Number of early stage cancers Number of resectable cancers

NO Reduction in Lung Cancer Mortality

Why? Screened patients had a higher likelihood of

being diagnosed and living longer from the time of diagnosis.

But equal numbers of patients in both groups ultimately died of cancer.

Patz, EF Jr. NEJM 2000;343(22):1627-33.

Low-dose screening CT

No Contrast Low Radiation Dose

Average effective dose varies between 0.6 mSv and 1.5 mSv

Equivalent to 2 to 5 chest x-rays or a transatlantic flight

Less than a minute $150-400

Rationale for CT screening

Onset of tumor

Detectableby CT Onset of

symptoms

Detectableby CXR

Increased window of opportunity

Early Lung Cancer Action Program (ELCAP)

Cornell and NYU 1000 patients Entry Criteria

Age > 60 ≥ 10 pack year smoking history

Actual Mean Patient Age = 67 45 pack years

Plain film and CT imaging

Henschke CI. Lancet 1999;354(173):99-105.

ELCAP Protocol

If Negative Recommend annual repeat CT

If Diffuse disease Recommend Pulmonary Evaluation

If Positive <5mm – Follow-up CT scan at 3, 6, 12, 24

months 6 – 10 mm – Investigator discretion > 10 mm – Biopsy or Surgery

Henschke CI. Lancet 1999;354(173):99-105.

ELCAP Data

233 “Positive” Scans (23.3%) 33 also seen on CXR Conversely, less than half of the “nodules”

seen on CXR were confirmed on CT 27 Malignant NCN (2.7%)

CXR missed 20 of these Mainly adenocarcinoma (Plus 4 other malignancies: 2

endobronchial lesions and 2 mediastinal adenopathy)

Henschke CI. Lancet 1999;354(173):99-105.

233 Patients with non-calcified nodules

0

10

20

30

40

50

60

70

2-5 6-10 11-20 >20

%

mm

Size of Largest NCN

0

10

20

30

40

50

60

70

1 2-3 4-6

Number of NCN’s/Scan

number

%

Henschke CI. Lancet 1999;354(173):99-105.

Chance of Malignancy

0

10

20

30

40

50

60

70

2-5 6-10 11-20 >20

%

mm

Size of Largest NCN

0

10

20

30

40

50

60

70

1 2-3 4-6

Number of NCN’s/Scan

number

%

12%

10%

14%

1%

24%

33%

80%

Henschke CI. Lancet 1999;354(173):99-105.

Characteristics of the 27 lung cancers

I

IIIII

85% were stage I 83% missed by

CXR 96% were

resectable

Stage distribution Results

Henschke CI. Lancet 1999;354(173):99-105.

CT scan versus CXR

More sensitive Non-calcified nodules (23 versus 7 percent) Malignancy (2.7 versus 0.7 percent) Benign nodules (20.6 versus 6.1 percent)

BUT no stage shift 3.0/1000 versus 2.1to 3.8/1000 with

advanced disease

Patz, EF Jr. NEJM 2000;343(22):1627-33.

Other results

The rate of biopsy for benign disease

was 1.7% 4 of 233 patients

No patient underwent a thoracotomy for a benign lesion in the ELCAP study

ELCAP had no control group Mortality inferences are subject to bias

Meta-analysis of baseline findings of randomized, controlled trials

NameScreening Duration

Control Arm

Age

Garg/Colorado University

2001 50-80

DANISH 2004-2006 49-74

NELSON 2004- 50-75

ITALUNG 2004-2006 55-69

LSS 2000-2004 CXR 55-77

DEPISCAN 2002-2004 CXR 50-75

DANTE 2001-2006 CXR 60-74

CT Screening for lung cancerOdds Ratio 95% CI

Higher number of stage 1 cancers

3.9 2.0 – 7.4

Higher total lung cancers

4.1 2.4 – 7.1

Detection of false-positive nodules

3.1 2.6 – 3.7

Thoracotomies for benign lesions

Event rate 3.7 per 1000

3.5 – 3.8

Gopal M. J Thorac Oncol. 2010;5:1233-1239.

National lung cancer screening trial More than 53,000 current or former

smokers were enrolled in NLST at more than 30 study sites across the country (2002–2004)

Examined the risks and benefits of spiral CT scans compared to chest X-rays. Repeated at one and two years after the

first scan. Powered to detect a 20% reduction in

mortality due to screening.

NLST design NLST design

Prospective, randomized trial comparing low-dose helical CT screening to chest x-ray screening with the endpoint of lung cancer specific mortality in high risk participants

Eligibility Age 55-74 Asymptomatic current or former smoker; 30 pack year smoking history Former smokers: quit within preceding 15 years No prior lung cancer diagnosis No evidence of other cancer within preceding 5 years

http://radiology.rsna.org/content/early/2010/10/28/radiol.10091808.full Slides courtesy of Christine Berg, MD (NIH)

NLST primary endpoint

http://radiology.rsna.org/content/early/2010/10/28/radiol.10091808.full

Helical CT vs. CXR

Lung cancer-specific mortality 20% difference

α 5%

Power 90%

Compliance 85% CT | 80% CXR

Contamination 5% CT | 10% CXR

Size 25,000 / arm

Secondary endpoints All cause mortality Lung cancer: prevalence | incidence | interval cancers Stage distribution Screening test performance Medical resource utilization for [+] screen

NLST secondary endpoints

http://radiology.rsna.org/content/early/2010/10/28/radiol.10091808.full Slides courtesy of Christine Berg, MD (NIH)

53,454 participants NLST US Census

Male (%) 59.0 58.5

Age

55-59 (%) 42.8 35.2

60-64 (%) 30.6 29.3

65-69 (%) 17.8 20.8

70-74 (%) 8.8 14.7

Race | Ethnicity

Black (%) 4.4 5.5 Hispanic (%) 1.7 2.4

Comparing NLST with eligible US census population

Aberle DR, et al. Natl Cancer Inst (2010) 102 (23): 1771-1779.

NLST US CensusMarried 66.6 60.9

Education

< HS 6.1 21.3 ≥ College 31.5 14.4

Current smoker 48.2 57.1

Median pack yrs 48.0 47.0

Comparing NLST with US census population

Aberle DR, et al. Natl Cancer Inst (2010) 102 (23): 1771-1779.

Compared with similar US population, NLST cohort has similar gender distribution and smoking exposure

However, NLST participants Younger Better educated Less likely to be current smokers

Comparing NLST with US census population

Aberle DR, et al. Natl Cancer Inst (2010) 102 (23): 1771-1779.

Screening exam compliance

Study Year

Helical CT Chest X-ray TotalExpected Screened Expected Screened Expected Screened

T0 26,713 98.5% 26,722 97.5% 53,435 98.0%

T1 26,282 94.0% 26,398 91.3% 52,680 92.6%

T2 25,935 92.9% 26,097 89.5% 52,032 91.2%

Screen positivity rate by screening round & arm

Low dose helical CT CXR

Number screened

Number positive

% Positive

Number screened

Number positive

% Positive

Screen 1 26,314 7,193 27.3 26,049 2,387 9.2Screen 2 24,718 6,902 27.9 24,097 1,482 6.2Screen 3 24,104 4,054 16.8** 23,353 1,175 5.0**

All screens 75,136 18,149 24.2 73,499 5,044 6.9

* Positive screen: nodule ≥ 4 mm or other findings potentially related to lung cancer.** Abnormality stable for 3 rounds could be called negative by protocol.

True and false positive screens

Screening Result

Low Dose Helical CT CXR

Screen 1N (%)

Round 2N (%)

Round 3N (%)

Round 1N (%)

Round 2N (%)

Round 3N (%)

Total Positives

Lung cancerNo lung cancer

7,193 (100)

270 (4)6,923 (96)

6,902 (100)

168 (2)6,734 (98)

4,054 (100)

211 (5)3,843 (95)

2,387 (100)

136 (6)2,251 (94)

1,482 (100)

65 (4)1,417 (96)

1,175 (100)

78 (7)1,097 (93)

Data reflect the final interpretation, including benefit of historical comparison exams.

Interim analysis: lung cancer mortality 10-20-2010

Arm Person Years (py)

Lung cancer deaths

Lung cancer mortality per

100,000 py

Reduction in lung cancer mortality (%)

Value of test

statisticEfficacy

boundary

CT 144,097.6 354 245 20.3 –3.21 –2.02

CXR 143,363.5 442 308

Deficit of lung cancer deaths in CT arm exceeds that expected by chance, even allowing for multiple looks at the data.

p = 0.0041

Interim analysis: all-cause mortality 10-20-2010

Lung cancer: 25% of all deaths in NLST Lung cancer: 56% of 126 excess deaths in CXR arm

p = 0.023

Arm Person Years (py) Deaths

All-cause mortality per

100,000 py

Reduction in all cause

mortality (%)

Value of test

statistic

Value for significanc

e

CT 167,389.9 1870 1117 6.9 –2.27 –1.96

CXR 166,328.2 1996 1200

Kaplan-Meier curves for lung cancer mortality

1.00

0.99

0.98

0.97

0.96

0.95

0.94

0.93

0.92

0.91

0.90

0 1 2 3 4 5 6 7

Prob

abili

ty o

f sur

viva

l: A

LL p

artic

ipan

ts

CT armCXR arm

Years from randomization

Kaplan-Meier curves for all-cause mortality

1.00

0.99

0.98

0.97

0.96

0.95

0.94

0.93

0.92

0.91

0.90

0 1 2 3 4 5 6 7 8Years from randomization

Prob

abili

ty o

f sur

viva

l: A

LL p

artic

ipan

ts

CT armCXR arm

Results

At the time the DSMB held its final meeting on October 20, 2010: 356 deaths from lung cancer had occurred

among participants in the CT arm of the study

443 lung cancer deaths had occurred among those in the chest X-ray group

The DSMB concluded that this 20.3 percent reduction in lung cancer mortality met the standard for statistical significance and recommended ending the study.

Conclusions

The vast majority of lung nodules detected by either CT or CXR are benign.

Most CT-detected nodules require some form of additional follow-up.

CT detects more lung cancers than does CXR.

Most of these excess cancers are early stage cancers.

Screening programs still uncover late-stage cancers on initial and interval scans.

Conclusions

Prior to the NLST trial, there had been no evidence of a mortality benefit for lung cancer-related mortality with CT screening.

The NLST trial showed a 20% relative risk reduction in lung-cancer specific mortality.

Cost-effectiveness ratio

Modeling algorithm Analysis based on a model designed prior

to completion of NSLT modeled cost-effectiveness of CT scan screening for six patient cohorts Different ages and smoking histories

CT screening may decrease lung cancer at 10 years by 18 to 25% at a cost ranging from $126,000 to $269,000 per quality adjusted life year (QALY) Cost-effective ratio for other screened cancers

Colorectal cancer was $47,700 per QALY Breast cancer was $13,300 to $32,000 per QALY

McMahon PM, et al. J Thorac Oncol 2011; 6:1841.

Number needed to screen

NNS represents the number of patients who must screened over a given period of time to prevent one death from the disease in question.

Reflects both the prevalence of the disease and the effectiveness of therapy.

Does not specifically account for the risks or the costs of screening.

Reciprocal of the frequency of the reduction in absolute risk.

Other screening modalitiesSCREENING MODALITY

NUMBER NEEDED TO SCREEN

Annual screening mammography of women ages 40-841

84

Biennial screening mammography of women ages 50-741

144

Guaiac-based FOBT screening2 1,173 to prevent 1 death over 10 years

Colonoscopy3 200 men with regular screening

Papanicolaou smear screening for cervical cancer4

1,140 women with regular screening to prevent 1 death over 10 years

Low-dose CT screening for lung cancer5

256 persons screened annually for three years to prevent 1 death over 6 years

1Hendrick RE, et al. Am J Roentgenol. 2012 Mar;198(3):723-8.2Hewitson P, et al. Cochrane Database Syst Rev. 2007;(1):CD001216.3Barrett B et al. Fam Med 2011;43(4):248-53.4Benedet JL, et al. Am J Obstet Gynecol. 1992;166:1254–9.5Berg CD, et al. N Eng J Med. 2011;365(5):395-409.

Recommendations for screening by expert groups

In 2012, a systematic review was commissioned by American Cancer Society (ACS), American College of Chest Physicians (ACCP), American Society of Clinical Oncology (ASCO), and the National Comprehensive Cancer Network (NCCN) Screening guidelines supporting low-dose

CT scans for high-risk groups were issued by the NCCN and ACCP/ASCO

Bach, PB, et al. JAMA. 2012: 307(22): 2418-2429.

Guidelines for lung cancer screeningAmerican Cancer Society

Recommends that clinicians with access to high-volume, high-quality lung cancer screening and treatment centers should initiate a discussion about screening with apparently healthy patients aged 55 years to 74 years who have at least a 30 pack-year smoking history and who currently smoke or have quit within the past 15 years.

2013

ACCP/ASCO Recommends annual low-dose CT screening for high-risk individuals (ages 55 to 74 years with 30 pack-year history of smoking and current smoker or quit within past 15 years).

2012

National Comprehensive Cancer Network (NCCN)

Recommends annual low-dose CT screening for high-risk individuals (ages 55 to 74 years with 30 pack-year history of smoking or 20 pack-year history with an additional risk factor).

2011

US Preventive Services Task Force

Evidence is insufficient to recommend for or against screening asymptomatic persons for lung cancer with CXR, low-dose CT, or a combination.

2004; undergoing review

Canadian Task Force on the Periodic Health Examination

Recommends against the use of chest x-ray in asymptomatic persons. Evidence is insufficient to recommend for or against screening with spiral CT in asymptomatic persons.

2003; undergoing review

Issues to be addressed

Optimal population for screening Cost-effectiveness of screening Optimal management of screen-detected

nodules Optimal screening interval and number of

screening rounds Reimbursement issues Importance of implementing screening

programs only in the setting of multidisciplinary programs with experience in evaluation and management of early lung cancers

A proven intervention