Critical appraisal:Pastorino U, Rossi M, Rosato V, Marchianò A, Sverzellati N, Morosi C, et al 2012 1
- Pastorino U, Rossi M, Rosato V, Marchianò A, Sverzellati N, Morosi C, et al. Annual or biennial CT screening versus observation in heavy smokers: 5-year results of the MILD trial. Eur J Cancer Prev 2012 May;21(3):308-15 Available from: http://www.ncbi.nlm.nih.gov/pubmed/22465911.
- Assigned to
- Topic area
- Guidelines:Lung cancer/Screening and early detection
- Clinical question
- Study design
- randomised controlled trial
- Level of Evidence
Section below only relevant for Cancer Council Project Officer
Pastorino U, Rossi M, Rosato V, Marchianò A, Sverzellati N, Morosi C, et al. Annual or biennial CT screening versus observation in heavy smokers: 5-year results of the MILD trial. Eur J Cancer Prev 2012 May;21(3):308-15 Available from: http://www.ncbi.nlm.nih.gov/pubmed/22465911.
randomised controlled trial
the Multicentric Italian Lung Detection (MILD) study aims to evaluate the impact on mortality of early lung cancer detection through LDCT at annual or biennial intervals versus no screening.
control group=1723; annual LDCT=1190; biennial LDCT = 1186. (CT=2376).
Age and gender were comparable in the three arms but the proportion of current smokers was significantly higher in the control arm (89.7%) than in the LDCT arms (68.6%). Lung function (FEV1) appeared better in control group (19% with FEV1<90% in control compared to 27-28% in LDCT groups).
the national program faced many difficulties as a result of lack of funding, limited support from local authorities, and cultural prejudice. Volunteers were reluctant to enter the control arm of any randomized trial. Thus, we had initially to propose a randomized comparison between annual versus biennial LDCT. Once this study had
been approved and funded an observational control arm was added; this explains the lower number in the control group.
Volunteers were recruited from among respondents to advertisements and articles published in the lay press and
in television broadcasts.
Eligibility criteria: age >=49 years, current or former smokers (having quit smoking within 10 years of recruit-
ment) with at least 20 pack-years of smoking, and no history of cancer within the previous 5 years.
Centralized stratified randomization was accomplished by the use of blocks of variable size. The list of
randomization was stratified by reference center, age (up to 65 years or older), and duration of smoking (more
or less than 40 years). The group randomized to receive LDCT was further randomized to receive LDCT every 12
months (annual) or every 24 months (biennial).
statistical power: planned sample size of 10,000 individuals, a screening period of 10 years, and a total
follow-up of 100 000 person-years. Such a sample size would be adequate to detect a 30% reduction in lung
lung cancer mortality rates
stage of cancer
20 lung cancers were diagnosed in the control group, 25 in the biennial and 34 in the annual LDCT groups.
5-yr cumulative lung cancer incidence rate was 310.9/100 000 in the control group, 457.0 in the biennial, and 620.2 in the annual LDCT group (P=0.036).
Lung cancer mortality rates were 108.5/100 000 in the control, 108.8 in the biennial, and 216.0 in the annual LDCT groups. After adjustment for age and smoking, the HR was 1.64 (95% CI, 0.67–4.01) when the two LDCT arms combined were compared with the control group(P=0.21).
All-cause mortality rate was 310.1/100 000 in the control group, 362.5 in the biennial LDCT, and 557.9 in the annual LDCT. After adjustment for age and smoking, the HR was 1.40 (95% CI, 0.82–2.38)when comparing the two LDCT arms together with the control group.
|High risk of bias||Comments: Please replace this text and include any additional comments in regards to your quality rating|
Risk of bias assessment: randomised controlled trial
- Outcomes not blinded, substantial side-effects, or not reported.
- Adequately concealed (e.g. central randomisation, numbered or coded bottles, drugs prepared by pharmacy).
- No exclusions or survival analysis used with all subjects included (>95% follow-up for all groups).
The field below is not considered when calculating the risk of bias rating
|3||Reason for decision: Please replace this text and briefly describe the reasons for your rating|
|1||Additional comments: Please replace this text and briefly describe the reasons for your rating|