Risk Benefit Analysis of X-Ray Exposure Associated with Lung Cancer Screening in the Italung-CT Trial

Transcription

1 Risk Benefit Analysis of Lung Cancer Screening Chest Imaging Original Research A C D E M N E U T R Y L I A M C A I G O F I N G Mario Mascalchi 1 Giacomo Belli 2 Marco Zappa 3 Giulia Picozzi 1 Massimo Falchini 1 Riccardo Della Nave 1 Germana Allescia 1 Andrea Masi 4 Andrea Lopes Pegna 5 Natale Villari 1 Eugenio Paci 3 Mascalchi M, Belli G, Zappa M, et al. Keywords: cancer screening, CT, lung cancer, MDCT, radiation exposure, single-detector CT DOI: /AJR Received January 18, 2005; accepted after revision June 24, The Italung-CT Trial is supported by the Health Department of the Tuscany Region, Italy; and the Ministry of Instruction, University and Scientific Research of Italy (grant ). 1 Sezione di Radiodiagnostica, Dipartimento di Fisiopatologia Clinica, Università di Firenze, Viale Morgagni 85, Firenze, Italia. Address correspondence to M. Mascalchi (m.mascalchi@dfc.unifi.it). 2 Fisica Sanitaria, Azienda Ospedaliera Careggi, Firenze, Italia. 3 Centro di Studio e Prevenzione Oncologica, Firenze, Italia. 4 U.O. Radiologia Diagnostica, Azienda Ospedaliera Careggi, Firenze, Italia. 5 U.O. Pneumologia, Azienda Ospedaliera Careggi, Firenze, Italia. AJR 2006; 187: X/06/ American Roentgen Ray Society Risk Benefit Analysis of X-Ray Exposure Associated with Lung Cancer Screening in the Italung-CT Trial OBJECTIVE. Prior analyses of X-ray exposures in lung cancer screening with CT considered the basic acquisition technique in single-detector scanners and the effects of a lifetime screening regimen, whereas the potential benefit in terms of lives saved was not addressed. MATERIALS AND METHODS. We determined the total-body effective dose of different acquisition techniques for one single-detector and one MDCT scanner and made projections about the cumulative radiation exposure to smokers undergoing four annual CT examinations on the same scanners in the Italung-CT Trial. Combining these data with estimates of radiation-induced fatal cancer and of the benefit of screening, we calculated the risk benefit ratio for participants in the trial, ex-smokers, and never-smokers. RESULTS. The cumulative effective doses per 1,000 subjects were 3.3 Sv using an MDCT scanner and 5.8 or 7.1 Sv using a single-detector scanner. Potential fatal cancers associated with radiation exposure were 0.11 per 1,000 subjects for MDCT scanners and 0.20 or 0.24 for singledetector scanners, which is about times lower than the number of expected lives saved by screening assuming a 20 30% lung cancer specific mortality reduction in current smokers. They were, however, of similar magnitude to the lives saved by screening in never-smokers and former smokers assuming a 10% efficacy of screening. CONCLUSION. MDCT is associated with lower radiation doses than single-detector CT technology. The risk of radiation dose in the Italung-CT Trial is compensated for by the expected benefit. CT screening for lung cancer should not be offered to never-smokers, whereas its recommendation in former smokers is debatable. T is capable of revealing peripheral C lung tumors in the early stages [1, 2]. Observational studies [3 9] and randomized trials [10 13] involving thousands of individuals are in progress worldwide to assess whether CT screening is effective in reducing mortality due to lung cancer. In the available accounts of the dose exposure associated with lung cancer screening using CT, only the basic low-dose techniques with single-detector helical scanners similar to that originally used in the Early Lung Cancer Action Project (ELCAP) study [1] were considered [14, 15], whereas the dose associated with additional full-dose high-resolution acquisitions and repeated low-dose examinations recommended in the same study [1, 16] for characterization and follow-up of suspicious noncalcified nodules was not addressed. As well, only the dose associated with the basic low-dose technique was used in the analyses of the risk of radiation-induced lung cancer associated with CT screening [17, 18]. The advent of MDCT scanners and updates to the protocol [15, 19] prompted us to investigate the cumulative dose currently delivered to the screened population. For this purpose, we calculated the radiation dose in a pilot study and made projections of the radiation exposure to smokers undergoing four annual CT examinations in a randomized clinical trial named Italung-CT [13], currently in progress. Moreover, we performed a risk benefit analysis of radiation exposure associated with lung cancer screening for participants in the Italung-CT Trial and for exsmokers and never-smokers. Materials and Methods The Pilot Study Between November 2000 and November 2003, we performed a pilot observational study of 60 smokers that was approved by the local ethics committee; the results of that study are reported elsewhere [13]. Twenty-four subjects were examined on a single-detector scanner (Somatom Plus, Siemens AJR:187, August

2 Medical Solutions) with a 1-second rotation time and 36 subjects, on an MDCT scanner (Somatom Plus 4 VZ, Siemens) with a 0.5-second rotation time and 4 rows of detectors. The study included one baseline and two annual repeat low-dose CT examinations for a total of three screening rounds. In case of suspicious nodules, additional follow-up CT examinations were performed according to the initial ELCAP protocol [1]. The CT acquisition techniques were those recommended in the same study [1]. Italung-CT Trial The Italung-CT Trial is a multicenter randomized controlled study, which was approved by the local ethics committees of the participating centers, aiming to evaluate reduction of mortality from lung cancer with CT screening [13]. It started in 2004 and will enroll 3,000 smokers years old who will be randomized in an active arm (1,500 subjects) undergoing annual low-dose CT for 4 years and a control arm (1,500 subjects) who will receive usual care [13]. Risk Benefit Analysis The risk benefit analysis was performed by two epidemiologists after approval of the local ethics committee. Experimental Dose Measurements Dose measurements were obtained on the single-detector and MDCT scanners used in the pilot study and in the Italung-CT Trial. The dose at the isocenter of the scanners was measured using a pencil ionization chamber (10 cm long) and a readout multimeter (NEROmAX, Victoreen), and the air CT dose index (CTDI) was calculated. Several CTDIs at different beam collimations were measured, and the dose length product (DLP) involved in each acquisition technique was obtained. Finally, the dose was computed based on the air CTDI measured at the isocenter of the scanner using the software, CT-Dose, which was developed by the Department of Biomedical Engineering, County of Aarhus and the National Board of Health, Denmark ( This software provides Monte Carlo simulation based on an anthropomorphic (Adam and Eve) phantom. The effective (whole-body) dose, as defined by the International Commission on Radiological Protection (ICRP) 60 [20], and the lung dose were determined. Three basic CT techniques were assessed: first, low-dose at thick and thin collimations; second, full-dose at a thin collimation; and, third, full-dose at a thick collimation. In addition, we estimated the dose of the preliminary scout anteroposterior view with 120 kvp, 40 mas, and 2- mm beam collimation, which can be considered computed projection radiography. Dose Estimates Dose in the pilot study Using the previously described measurements of the dose associated with the single-detector and MDCT acquisition techniques, we retrospectively estimated the dose radiation that was actually delivered to the 60 subjects participating in the pilot study. Dose projections in the screened arm of the Italung-CT Trial The dose projections in subjects undergoing screening for lung cancer in the active arm of the Italung-CT Trial were computed by, first, considering the dose associated with the MDCT and single-detector scanners and protocols used in the Italung-CT Trial [13]; and, second, summing up the dose to subjects with negative tests, the dose to subjects requiring additional follow-up CT examinations, and the dose to subjects requiring intervention. The acquisition techniques for single-detector and MDCT scanners in the Italung-CT Trial are essentially the same as those recommended in the last available EL- CAP protocol (icscreen.med.cornell.edu) with some minor differences. In particular, the ELCAP protocol recommends supplemental acquisition of a package of thin-collimation slices at full dose centered on indeterminate nodules when these are found using lowdose, thick-collimation acquisition on a single-detector scanner. Actually, we observed that even low-dose 3-mm-collimation acquisitions obtained on a singledetector scanner with 1.5-step reconstructions provide images with sufficient spatial resolution, and we adopted this technique for lung cancer screening with a single-detector scanner. Nonetheless, for the singledetector scanner, we computed also the dose associated with the previously mentioned ELCAP recommendation. For that purpose, we considered the dose associated with one 20-mm-thick package of thincollimation slices at full dose centered on a nodule 8 mm in diameter. Additional follow-up CT examinations are recommended in subjects with noncalcified nodules at baseline screening test that are 5 mm or more in mean diameter and new nodules at annual repeat test of 3 mm or more in diameter. The number and the time schedule of such follow-up examinations vary according to the size and the intervening size change of the nodule. We assumed one follow-up examination at 3 months after the initial examination for an indeterminate nodule at baseline and two follow-up examinations at 1, 3, or 6 months for an indeterminate nodule at annual repeat screening examinations. The frequency of noncalcified nodules at baseline and hence the proportions of subjects with a negative screening test and of subjects requiring follow-up reflect the selection criteria of the screened population and can vary as a function of several factors, including race, age, smoking habits, incidence of granulomatous diseases, and so on [21]. In the CT series reported to date, the frequency of noncalcified nodules at baseline ranged from 5.1% in a Japanese study including nonsmokers [2] to 51% in a Mayo Clinic study [6]. More importantly, this frequency is reduced up to more than one half if a cutoff of 5 mm in mean diameter is used for the baseline screening examination [4, 6, 9, 13, 22]. The incidence of new nodules at annual repeat examinations (all sizes) ranged from 5% in the ELCAP study [3] to 13% in a Mayo Clinic study [6]. For computation of the dose projections in the active arm of the Italung-CT Trial, we used arbitrary values of 10% frequency of subjects showing indeterminate nodules requiring follow-up at baseline and annual repeat screening rounds. Also, the percentage of subjects enrolled in CT screening for lung cancer requiring intervention varies [12, 21]. In the studies of heavy smokers in Western countries, the percentage is consistently below 3% at baseline and repeat screening rounds and is usually lower in the latter [1, 3, 8, 13]. We used an arbitrary value of 1% for the baseline test and 0.5% for the annual repeat test. Before intervention, these subjects usually undergo examinations with additional radiation exposure, including fulldose chest CT before and after IV contrast administration, CT-guided biopsy, and 18 F-FDG PET. Although the dose associated with those examinations differs from center to center and from subject to subject, we considered an average effective dose of 15 msv for subjects examined with MDCT and 20 msv for subjects examined with single-detector CT, including 7 msv for the chest PET examination [23]. Computation of the dose in this small fraction of subjects is justified by the fact that benign lesions are definitely found in up to 25% of the subjects undergoing intervention in the context of CT lung cancer screening [8]. The Risk The radiation-induced risk associated with the CT screening procedure was assessed combining the experimental dose measurements, the theoretic dose projections, and the estimates of the radiation-induced cancer deaths calculated from the English National Radiological Protection Board (NRPB) data for different age groups [24]. In the same report, sex differences were minor and were neglected here. Estimate of the Benefit from CT Screening for Lung Cancer The estimate of the benefit from screening was derived from the expected incidence of lung cancer in the absence of screening and from the available predictions of lung cancer specific reduction of mortality associated with CT screening [25, 26]. The following assumptions were made. First, we considered a population of 100,000 subjects with one third ranging in age from 55 to 59 years; one third, years; and one third, years. Second, we performed separate analyses for men and women. Third, we considered a screening program of 4 years 422 AJR:187, August 2006

3 Risk Benefit Analysis of Lung Cancer Screening with four rounds (one CT screening examination per year), as in the Italung-CT Trial. Fourth, we assumed that the benefit of early diagnosis was limited to all the cancers that would have developed during the years of the screening tests and over the ensuing 4 years, whereas no benefit is assumed after that period. Fifth, we estimated the incidence in our population by smoking attitude (current smokers, ex-smokers, and neversmokers) taking into account the following: first, the age and sex incidence rate of lung cancer in the general population as derived from the local tumor registry for the period [27] (Table 1); second, the proportion of smokers, ex-smokers, and never-smokers in the Tuscany population (Table 2) [28]; and third, the relative risk for lung cancer of smokers and former smokers in comparison with never-smokers [29] (Table 3). Former smokers were defined as individuals TABLE 1: Lung Cancer Annual Incidence Rates (per 100,000) in Florence, Tuscany, Italy [27], Annual Incidence Rate Age (y) Men TABLE 2: Proportion of Smokers, Ex-Smokers, and Never-Smokers in the Tuscany Population [28]: Estimates for % in Tuscany Population Risk Category Men Never-smokers Ex-smokers a Smokers a Defined as individuals who smoked at least 100 cigarettes over their lives and who now never smoke at all [29]. TABLE 3: Relative Risk of Lung Cancer Among Current Smokers, Ex-Smokers, and Never-Smokers [35] Relative Risk Risk Category Men Never smokers Ex-smokers a Smokers a Defined as individuals who smoked at least 100 cigarettes over their lives and who now never smoke at all [29]. who smoked at least 100 cigarettes over their lives and who now never smoke at all [29]. To minimize the healthy screening effect for which people undergoing screening have a lower risk in the first years, we assumed that in the first year the expected incidence is one third of that in the general population and in the second year, two thirds. We considered that, without screening, the patients with lung cancer diagnosed in the 8-year period since the beginning of the screening program would have died according to the observed relative survival at 15 years. In this way, we estimated the cumulative number of deaths in the population cohort. We arbitrarily applied to the cumulative number of deaths expected in the population cohort (in the absence of screening) a reduction of mortality of 30%, 20%, 10%, and 0% associated with early diagnosis due to active screening with CT. Our predictions extend the most favorable and unfavorable sets assumed in a cost-effective analysis by Mahadevia et al. [25]. Those researchers hypothesized a 50% stage shift associated with lung cancer screening and, by weighting the possible influence of confounding factors such as variable adherence to the annual screening regimen and the degree of length bias and overdiagnosis bias, constructed three fundamental scenarios namely, a base case scenario implying a 13% reduction of mortality, a favorable estimate scenario with a 16% reduction, and an unfavorable estimate scenario with a 4% reduction. Otherwise, based on an analysis of the stages of all cancers detected in the ELCAP and Mayo Clinic observational studies, Patz et al. [26] anticipated that no lung cancer specific mortality change with screening is expected. It is noteworthy that if CT screening does not modify mortality, the specific lung cancer mortality reduction is 0%. The number of deaths potentially prevented by screening was estimated taking into account patient sex and smoking habits. The net benefit of screening was computed as the difference of the number of lives saved minus the number of estimated radiation-induced fatal cancers. This was expressed as a risk benefit ratio, where a value of 1 implies that the number of lives saved with screening is equal to the number of radiation-induced deaths; a value of 0.1, that the number of lives saved exceeds 10 times the number of radiation-induced deaths; and so on. Results Experimental Dose Measurements Table 4 reports the estimates of the effective dose to the whole body and of the dose to the lung for the single-detector and MDCT scanners and the different techniques. Dose Delivered in the Pilot Study Two hundred ten CT examinations were performed in the pilot study over 3 years and the low-dose 3-mm-collimation technique was used in 14 examinations performed on a single-detector scanner. The cumulative 3-year doses were 53.9 msv in the 36 subjects screened with the MDCT system (mean exposure per subject, 1.49 msv/3 y and 0.49 msv/y) and msv in the 24 subjects screened with the single-detector system (mean exposure per subject, 5.9 msv/3 y and 1.9 msv/y). Dose Projections in the Active Arm of the Italung- CT Trial The dose projections are detailed in Table 5. The cumulative effective dose to the active arm of the Italung-CT Trial was 3.35 Sv per 1,000 subjects over 4 years (0.83 msv per subject/y) using the MDCT scanner (low-dose 4-mm collimation, yielding four 1-mm-thick sections) and 5.87 Sv (1.46 msv/y) (low-dose 3-mm collimation only) or 7.12 Sv (1.78 msv/y) (lowdose 10-mm collimation and one full-dose thincollimation package) per 1,000 subjects over 4 years using the single-detector scanner. The Risk With the previously described cumulative doses and the estimate of fatal cancers for each Sievert for subjects ranging in age from 50 to 70 years (men and women altogether) [24], the numbers of lifetime fatal cancers associated with 4 years of CT screening programs for lung cancer for subjects years old were 11.7 per 100,000 (0.11 per 1,000) for the MDCT scanner and 20.5 or 24.9 per 100,000 (0.20 or 0.24 per 1,000) for the single-detector scanner. The Benefit Table 6 details the lung cancer incidence rates in our area estimated for current smokers, ex-smokers, and never-smokers and different age groups and sex. Table 7 reports the expected number of lung cancers in 8 years in a cohort of 100,000 subjects. Based on a probability of dying within 15 years after the diagnosis of lung cancer equal to 91% in men and 90% in women [27], Table 8 reports the number of expected deaths from lung cancer in the same area and the same period in the absence of screening. Table 9 shows the estimates of deaths from lung cancer saved assuming a 10%, 20%, and 30% reduction of specific lung cancer mortality with CT screening according to sex and smoking habits. AJR:187, August

4 TABLE 4: Doses of Radiation for Scanners and Techniques Used in CT Screening for Lung Cancer in the Italung-CT Trial [13] Table Feed (mm) CT Parameter Scan Length (mm) Beam Collimation (mm) Slice Thickness (mm) Rotations CTDI DLP Lung Dose Technique kvp mas (no.) (mgy/mas) (mgy cm) (mgy) 4-MDCT (rotation time, 0.5 s) Low-dose Full-dose, thin-collimation a a a 0.33 Full-dose, thick-collimation Single-detector (rotation time, 1 s) Low-dose Two breath-holds Two breath-holds Full-dose, thin-collimation a a a 0.73 Full-dose, thick-collimation, two breath-holds Note CTDI = CT dose index in air at the isocenter, DLP = dose length product in air at the isocenter. a The dose is related to partially irradiated lung. Effective Dose (msv) TABLE 5: Cumulative Doses for 1,000 Subjects Screened for Lung Cancer with Low-Dose CT in the Italung-CT Trial [13] % of Subjects at Baseline Dose from Baseline Examination (msv) % of Subjects at Annual Repeat Examination Dose from Annual Repeat Examination 3 (msv) Cumulative 4-y Dose (Sv) Technique MDCT scanner Low-dose, 4 1 mm collimation Negative findings (D1) (D1) 2.28 Follow-up a (D2) (D2) 0.70 Intervention a 1 15 (D3) (D3) 0.37 Single-detector scanner Low-dose, 3-mm collimation Negative findings (D1) (D1) 4.11 Follow-up a (D2) (D2) 1.26 Intervention a 1 20 (D3) (D3) 0.50 Low-dose, 10-mm collimation low-dose; and one full-dose, thin-collimation package Negative findings (D1) (D1) 4.46 Follow-up a (D2) (D2) 2.18 Intervention a 1 20 (D3) (D3) 0.48 Note The 0.05-mSv dose associated with scout acquisition obtained using low milliamperage setting (ma) and 2-mm collimation is included in dose estimate for each examination. D1 = dose for subjects with negative findings; D2 = dose for subjects with noncalcified nodules 5 mm in mean diameter requiring one follow-up at 3 months for initial detection at baseline screening or two follow-up examinations at 1, 3, or 6 months for initial detection at annual repeat examinations; D3 = dose for intervention, which includes radiation exposure associated with full-dose diagnostic chest CT before and after IV contrast administration, CT-guided biopsy or fine-needle aspiration, and 18 F-FDG PET. a Arbitrary frequency assumed for the cumulative calculation. 424 AJR:187, August 2006

5 Risk Benefit Analysis of Lung Cancer Screening TABLE 6: Estimated Annual Lung Cancer Incidence for 100,000 in Florence, Tuscany, Italy, by Age, Sex, and Smoking Habits Men Age (y) Smokers Ex-Smokers a Never-Smokers Smokers Ex-Smokers a Never-Smokers , , Note Equation for estimate of rates is as follows: L(T) = L(0) p(0) + p(1) r(1) + L(2) p(2) r(2), where L(T) = lung cancer rate for whole population; L(0) = lung cancer rate for never-smokers (unknown); p = proportion of never-smokers (0), ex-smokers (1), and current smokers (2) in general population; and r = relative risk of exsmokers (1) and current smokers (2) as compared with never-smokers. a Defined as individuals who smoked at least 100 cigarettes over their lives and who now never smoke at all [29]. TABLE 7: Estimated Incidence of Lung Cancer for 100,000 Subjects in 8 Years in Florence, Italy Age at Men Entry (y) Smokers Ex-Smokers a Never-Smokers Smokers Ex-Smokers a Never-Smokers , , , Total 5, a Defined as individuals who smoked at least 100 cigarettes over their lives and who now never smoke at all [29]. TABLE 8: Numbers of Expected Deaths from Lung Cancer for 100,000 Subjects in 8 Years in Florence, Italy Age at Men Entry (y) Smokers Ex-Smokers a Never-Smokers Smokers Ex-Smokers a Never-Smokers , , , Total 5, a Defined as individuals who smoked at least 100 cigarettes over their lives and who now never smoke at all [29]. TABLE 9: Estimates of Deaths from Lung Cancer Potentially Saved by CT Screening in Florence, Italy in a Cohort of 100,000 Subjects Undergoing Annual Screening Examination for 4 Years Assumed Men Reduction of Mortality (%) Smokers Ex-Smokers a Never-Smokers Smokers Ex-Smokers a Never-Smokers 30 1, , Note One third of subjects were years old, one third were years old, and one third were years old. a Defined as individuals who smoked at least 100 cigarettes over their lives and who now never smoke at all [29]. The Risk-to-Benefit Ratio Different scenarios may be drawn according to efficacy of screening, smoking habits, sex of the screened subject, use of single-detector or MDCT technology, and patient age at the commencement of screening. In Figure 1, the risk benefit ratios associated with undergoing four annual screening examinations with the MDCT and single-detector scanners (low-dose plus full-dose highresolution CT) used in the Italung-CT Trial assuming 10%, 20%, and 30% lung cancer mortality reduction are reported for male and female never-smokers, ex-smokers, and current smokers. Assuming a 10% reduction of mortality, the risk benefit ratio is higher than 1 for male and female never-smokers examined with a single-detector scanner. At the same level of efficacy, the ratio is between 1 and 0.1 for male and female never-smokers examined with MDCT and for male or female former smokers examined with a single-detector or MDCT scanner. The risk-to-benefit ratio for current smokers, assuming a 10% screening efficacy, ranges between 0.32 (females examined on a single-detector scanner) and 0.02 (males examined on an MDCTscanner). Assuming 20% or 30% screening efficacy, the ratio is around or below 0.1 for male and female current smokers examined with either a single-detector or an MDCT scanner. In Figure 2, the variation of the risk benefit ratios as a function of patient age at the commencement of screening is displayed assuming a 20% efficacy of screening. The riskto-benefit ratio decreases with advancing age and is lower in men. Discussion The increase in diagnostic radiology activity is emerging as a main concern in medicine due to the risk of radiation-induced fatal cancers [30]. The estimation of this radiation-induced cancer risk is difficult. In fact, it is fundamentally based on observational studies of the atomic bomb survivors and radiation workers chronically exposed to low radiation doses and implies several assumptions. In particular, the cancer risk from low-level radiation such as that used for diagnostic radiology procedures has been extrapolated by observations obtained at moderate and high doses using a linear no-threshold relation between the radiation dose and the risk of cancer [31]. However, other scenarios including the possibility that the linear no-threshold relation underestimates or overestimates the cancer AJR:187, August

6 risk from a low radiation dose, the presence of a dose threshold below which the risk is 0, and the possible protective effect of low-radiation dose against cancer (so-called hormetic response) can be hypothesized based on observational, experimental, and radiobiologic data. Even adopting a linear no-threshold dose risk relation, which appears to be the most reasonable and prudent assumption, other variables that can modify the Risk Benefit Ratio % 20% 30% Estimated Mortality Reduction from Screening MD never-smokers SD never-smokers MD ex-smokers SD ex-smokers MD current smokers SD current smokers dose risk relation, such as a theoretic relationship, need to be considered. First is the age of the patient at exposure. In fact, the low-dose radiation-induced cancer incidence generally decreases with advancing age. It is noteworthy, however, that the risk of radiation-induced lung cancer does not seem to show this pattern and peaks at the age of years [18]. Second is the temporal profile of the dose exposure. In fact, protracted exposures are associated with A lower risks of cancer than those of an acute exposure to the same total dose [18]. This is particularly relevant, but to date unsettled, in the context of screening procedures using X-rays in which a series of low-dose examinations is performed over many years. Third is that the latent period between radiation exposure and cancer death increases with decreasing exposure, and it is possible that for low doses the latent period exceeds the normal life span [31]. Fig. 1 Risk-to-benefit ratios for study participants. A and B, Risk-to-benefit ratio (log units) histograms for male (A) and female (B) never-smokers, ex-smokers, and current smokers associated with four annual screening examinations with MDCT (MD) and single-detector (SD) (low-dose thick-collimation plus one full-dose thin-collimation package) scanners in the Italung-CT Trial. Three different levels of expected benefit namely, 10%, 20%, and 30% reduction of lung cancer mortality are considered. Assuming 10% reduction of mortality, risk benefit ratio is over the unit that is, the number of radiation-induced deaths overcomes the estimated number of lives saved, for male (1.13) and female (1.66) never- smokers examined with the single-detector scanner. At the same level of screening efficacy, the ratio is between 1 and 0.1 and close to the critic value of 0.5, implying only two lives saved for one radiation-induced death, for male (0.53) or female (0.78) never-smokers examined with an MDCT scanner, for female former smokers examined with a single-detector (0.87) or MDCT (0.41) scanner, and for male former smokers examined with a single-detector scanner (0.53). Assuming 20% or 30% screening efficacy, the ratio is around or below 0.1 for male and female current smokers examined with either a single-detector or MDCT scanner. Risk Benefit Ratio Age (y) MD never-smokers SD never-smokers MD ex-smokers SD ex-smokers MD current smokers SD current smokers A MD never-smokers SD never-smokers MD ex-smokers SD ex-smokers MD current smokers SD current smokers Fig. 2 Risk-to-benefit histograms for men and women stratified by age. A and B, Risk-to-benefit ratio (log units) histograms for males (A) and females (B) commencing 4 years of annual screening rounds on an MDCT (MD) or single-detector (SD) (low-dose thick-collimation plus one full-dose thin-collimation package) scanner in different age categories: 55 59, 60 64, and years. Efficacy of screening is assumed to be 20% in reducing mortality in 8 years after start of program. For current smokers, risk benefit ratio ranges between 0.21 in women who are years old examined on a single-detector scanner and in men who are years old examined on an MDCT scanner. Risk Benefit Ratio Risk Benefit Ratio % 20% 30% Estimated Mortality Reduction from Screening Age (y) MD never-smokers SD never-smokers MD ex-smokers SD ex-smokers MD current smokers SD current smokers B B 426 AJR:187, August 2006

7 Risk Benefit Analysis of Lung Cancer Screening To perform a balanced analysis of the harm and gain of diagnostic radiology related radiation exposure, studies evaluating also the benefit in terms of reduced mortality are needed [32, 33]. There is no proof that CT screening is effective in reducing mortality of lung cancer, but there is consensus that randomized clinical trials should be performed to evaluate this possibility [15]. Because there is no indication supporting the extension of CT screening examinations for life, even for heavy smokers [15, 26], we focused on the risk-to-benefit ratio of radiation exposure in the active arm of the Italung-CT Trial, which is a 4-year clinical trial and is currently in progress [13]. Accordingly, on the scanners used for the CT screening program, we experimentally measured the radiation dose related to the CT acquisition techniques and protocols currently used in this trial. The cumulative dose delivered to subjects undergoing lung cancer screening was calculated in retrospect for a pilot study and as projections for the Italung- CT Trial. For the calculation of the risk of fatal cancer, we used the effective dose [20] and NRPB data [24], whereas for the estimation of the benefit we assessed the number of expected cancers in our area based on the data from the local tumor registry [27]. A range of efficacy of the screening procedure in reducing the mortality of lung cancer of between 0% and 30% was assumed. For the basic low-dose CT technique, we measured a dose of 0.36 msv on the MDCT scanner using a 1-mm slice collimation, 120 kvp, and 20 mas and a dose of 1.1 or 1.2 msv on the single-detector scanner using a 3- or 10-mm collimation, 140 kvp, and 43 ma. Although a small increase in the peak kilovoltage for instance, from 120 to 140 kvp implies an increase in the dose of about 30% on the single-detector scanner we used, this change was necessary to improve the image quality, taking into consideration the characteristics of the detectors and time of rotation of the single-detector scanner of our study. Nishizawa et al. [17] experimentally calculated an effective dose of 3.6 msv for the basic CT examination technique. However, they used a single-detector mobile scanner with 100 mas and a table speed of 10 mm that is, pitch of 1 and 10-mm-collimation slices. Diederich and Lenzen [14] reported an effective dose of 0.3 msv for men and 0.55 msv for women on a single-detector scanner (10- or 5-mm collimation, pitch of 2, 25 mas). Our data indicate that MDCT technology enables considerable dose savings as compared with single-detector CT technology. This is especially true if one considers that low-dose thin-collimation acquisitions with 1- to 1.5-mm reconstruction can obviate supplemental full-dose thin-collimation packages. A similar dose saving can be obtained in the case of a single-detector scanner if lowdose 3-mm-collimation acquisitions are used. At variance with prior studies [14, 17, 18] in the calculation of the radiation dose associated with lung cancer screening, we included the additional exposure related to follow-up and interventional CT examinations required in the management of suspected nodules. The considerable dose associated with such additional examinations was previously emphasized [26, 34]. To calculate dose projections for participants in the Italung-CT Trial, we considered the dose exposures with current single-detector and multidetector technology, the percentages of nodules requiring followup and intervention reported in the literature, and the last ELCAP protocol and management recommendations. In particular, we adopted a cutoff value of 5 mm in mean diameter for an indeterminate nodule at baseline requiring follow-up and a restricted number of follow-up CT examinations namely, one examination for nodules detected at baseline screening and two examinations for nodules initially detected at annual repeat screening rounds. The mean cumulative 4-year doses for 1,000 screened subjects examined on an MDCT scanner were 3.3 Sv (0.83 msv/y per subject) and 5.8 Sv (1.46 msv per subject); the dose was 7.1 Sv (1.78 msv per subject) for those examined on a single-detector scanner. The lower dose exposure associated with the adoption of MDCT technology for lung cancer screening programs was confirmed in the pilot study, in which the mean annual dose per subject was 0.49 msv for MDCT and 1.9 msv for single-detector CT. Although no definite studies are yet available, preliminary data (personal unpublished observation) indicate that with the newer MDCT scanners (12, 16, > 16 detector rows) low-dose thin-collimation acquisition techniques recommended for lung cancer screening will provide dose exposures similar to those of the 4-MDCT scanner of our study. To calculate the risk of radiation associated with lung cancer screening with CT, we used the total-body (effective) dose and adopted the estimates of lifetime radiationinduced fatal cancer from English NRPB publications, in which data are stratified according to age at exposure and in which the effect of sex is minor [24]. Pending results of randomized trials, the estimate of the potential benefit of lung cancer screening with CT in reducing the specific mortality rate is speculative and controversial [1 3, 9, 19, 26]. In our analysis of the benefit of screening in terms of reduction of specific lung cancer mortality, we covered a range of between 30% and 0% screening efficacy [25, 26], the most favorable estimate being similar to that of mammographic screening for breast cancer [35]. The additional assumption was made that the 4-year mortality rate of lung cancer outside a screening program is so high that it can be considered equivalent to the incidence. In the present analysis, we applied the assumptions discussed earlier to the expected number of lung cancers based on current incidence rates in our area provided by the local tumor registry and on the Italian estimate on relative risk of current smokers in comparison to ex-smokers and nonsmokers. In calculating the risk benefit ratio related to the radiation exposure in lung cancer screening with CT, besides the estimated efficacy of the screening procedure in reducing lung cancer specific mortality, the interaction of several variables related to risk and benefit must be taken into account. In particular, because the incidence of lung cancer and the potential benefit of screening differ according to smoking habits, age, and sex, the risk-to-benefit ratio of radiation exposure related to lung cancer screening with CT varies accordingly. For male and female current smokers participating in a 4-year program, such as the Italung-CT Trial, our data indicate that assuming lung cancer specific mortality reduction of 10 30% with CT screening, the benefit of screening overcomes the risk associated with the radiation exposure. Our data also indicate that this favorable ratio, ranging from 0.32 (in women examined on a single-detector scanner with a 10% screening efficacy) to (in men examined on an MDCT scanner with a 30% screening efficacy), is more pronounced in elderly smokers who show the highest incidence of lung cancer. Conversely, our data indicate that even a 4-year screening program with CT is not indicated for subjects with a low risk of lung cancer such as never-smokers, with a risk benefit ratio ranging from 1.66 (women examined on a single-detector scanner with a 10% screening efficacy) to 0.08 (men exam- AJR:187, August

8 ined on an MDCT scanner with a 30% screening efficacy). In former smokers, the situation is intermediate, but it is noteworthy that for a screening efficacy of 10% the ratio is between 0.25 and 0.87 with either CT technology for women and men. The estimate of lung cancer risk in former smokers is difficult to calculate owing to the controversial definition of this category and its presumable heterogeneity and to the uncertainties about the change over time of the lung cancer risk in subjects who quit smoking. Our data seem to indicate that some caution should be exercised in initiating lung cancer screening programs in former smokers, and that in such a case a keen understanding of the individual risk and of the specific risk categories is necessary. Although it is possible that extension of CT screening to more than 4 years will increase the benefit more than the risk in heavy smokers, we did apply our analysis to the typical regimen of screening used in ongoing randomized clinical trials namely, annual CT examination for a few years [11]. If CT screening does not affect mortality, subjects participating in CT screening programs have no benefit of radiation exposure and only a possible detrimental effect due to radiation-induced cancers. This assumption enables us to assess the risk associated with lung cancer screening in a simpler way than when some degree of efficacy is assumed. Several considerations are needed to properly weigh the risk of radiation-induced cancers in subjects participating in annual CT screening programs for lung cancer. In fact, one has to consider the interaction between radiation exposure for the lung and other factors in modifying the risk of lung cancer. In particular, it was reported that the interaction between X-ray radiation and smoking could imply a multiplicative effect [18]. Moreover, recent data showed that the risk of radiation-induced lung cancer does not decrease with increasing age at exposure [18]. On the other hand, there is considerable uncertainty about the time needed for development of radiation-induced cancer after X-ray exposure for lung cancer screening, but it can reasonably be estimated in terms of many years. It is conceivable that most subjects will receive in the meantime considerably higher doses of radiation for diagnostic or therapeutic procedures due to current diseases [30]. In conclusion, MDCT technology is associated with lower radiation doses than singledetector technology and is recommended for lung cancer screening with CT. However, even a 4-year annual screening with an MDCT scanner implies a nonnegligible dose exposure and a risk of radiation-induced fatal cancer. The amount of this risk in consideration of the uncertainties of the benefit of screening in terms of reduction in specific lung cancer mortality can equal or even dominate the spontaneous lung cancer incidence and mortality in nonsmokers in Western countries. Studies of lung cancer screening with CT should be restricted to current smokers, and their extension to former smokers is debatable. References 1. Henschke CI, McCauley DI, Yankelevitz DF, et al. Early Lung Cancer Action Project: overall design and findings from baseline screening. Lancet 1999; 354: Sone S, Li F, Yang Z-G, et al. Results of three-year mass screening programme for lung cancer using mobile low-dose spiral computed tomography scanner. Br J Cancer 2001; 84: Henschke CI, Naidich DP, Yankelevitz DF, et al. Early Lung Cancer Action Project: initial findings on repeat screening. Cancer 2001; 92: Diederich S, Wormanns D, Semik M, et al. Screening for early lung cancer with low-dose spinal CT: prevalence in 817 asymptomatic smokers. Radiology 2002; 222: Sobue T, Moriyama N, Kaneko M, et al. Screening for lung cancer with low-dose helical computed tomography: anti-lung cancer association project. J Clin Oncol 2002; 20: Swensen SJ, Jett AR, Sloan JA, et al. Screening for lung cancer with low-dose spiral computed tomography. Am J Respir Crit Care Med 2002; 165: Swensen SJ, Jett JR, Hartman TE, et al. Lung cancer screening with CT: Mayo Clinic experience. Radiology 2003; 226: Diederich S, Thomas M, Semik M, et al. Screening for lung cancer with low-dose spiral computed tomography: results of annual follow-up examinations in asymptomatic smokers. Eur Radiol 2004; 14: Pastorino U, Bellomi M, Landoni C, et al. Early lung-cancer detection with spiral CT and positron emission tomography in heavy smokers: 2-year results. Lancet 2003; 362: Garg K, Keith RL, Byers T, et al. Randomized controlled trial with low-dose spiral CT for lung cancer screening: feasibility study and preliminary results. Radiology 2002; 225: Gohagan J, Marcus J, Fagerstrom R, Pinsky P, Kramer B, Prorok P. Baseline findings of a randomized feasibility trial of lung cancer screening with spiral CT scan vs chest radiograph lung screening study of the National Cancer Institute. Chest 2004; 126: Diederich S, Wormanns D, Heindel W. Lung cancer screening with low-dose CT. Eur J Radiol 2003; 45: Picozzi G, Paci E, Lopez Pegna A, et al. Screening of lung cancer with low dose spiral CT: results of a three-year pilot study and design of the randomised clinical trial Italung-CT. Radiol Med (Torino) 2005;109: Diederich S, Lenzen H. Radiation exposure associated with imaging of the chest. Cancer 2000; 89: Aberle DR, Gamsu G, Henschke C, Naidich DP, Swensen SJ. A consensus statement of the Society of Thoracic Radiology: screening for lung cancer with helical computed tomography. J Thoracic Imaging 2001; 16: Libby DM, Smith JP, Altorki NK, Pasmantier MW, Yankelevitz D, Henschke CI. Managing the small pulmonary nodule discovered by CT. Chest 2004; 125: Nishizawa K, Iwai K, Matsumoto T, et al. Estimation of the exposure and a risk benefit analysis for a CT system designed for a lung cancer mass screening unit. Radiat Prot Dosimetry 1996; 67: Brenner DJ. Radiation risks potentially associated with low-dose CT screening of adult smokers for lung cancer. Radiology 2004; 231: Henschke CI, Yankelevitz DF, Smith JP, Miettinen OS; ELCAP group. Screening for lung cancer: the early lung cancer action approach. Lung Cancer 2002; 35: ICRP publication 60: 1990 Recommendations of the International Committee on Radiological Protection, 60. Ann ICRP 1990; vol 21: Bach PB, Kelley MJ, Tate RC, McCrory DC. Screening for lung cancer: a review of the current literature. Chest 2003; 123[1 suppl]:72s 82S 22. Henschke CI, Yankelevitz DF, Naidich DP, et al. CT screening for lung cancer: suspiciousness of nodules according to size on baseline scans. Radiology 2004; 231: Worsley DF, Celler A, Adam MJ, et al. Pulmonary nodules: differential diagnosis using 18 F-fluorodeoxyglucose single-photon emission computed tomography. AJR 1997; 168: Cox R, MacGibbon BH, Muirhead CR, et al. Estimates of the late radiation risks to the UK population. Documents of the National Radiological Protection Board (NRPB) 1993; 4(no. 4): Mahadevia PJ, Fleisher LA, Frick KD, Eng J, Goodman SN, Powe NR. Lung cancer screening with helical computed tomography in older adult smokers: a decision and cost-effectiveness analysis. JAMA 2003; 289: AJR:187, August 2006

9 Risk Benefit Analysis of Lung Cancer Screening 26. Patz EF, Swensen SJ, Herndon JE II. Estimate of lung cancer mortality from low-dose spiral computed tomography screening trials: implications for current mass screening recommendations. J Clin Oncol 2004; 22: Crocetti E, Paci E. Stage IA non small cell lung cancer: a small proportion of cases in the general population. Chest 2001; 119: Istituto Nazionale di Statistica. ISTAT indagine multiscopo: aspetti della vita quotidiana; Roma, Italy: ISTAT, Crispo A, Brennan P, Jockel KH, et al. The cumulative risk of lung cancer among current, ex- and never-smokers in European men. Br J Cancer 2004; 91: Berrington de Gonzalez A, Derby S. Risk of cancer from diagnostic X-rays: estimates for the UK and 14 other countries. Lancet 2004; 363: Cohen BL. Cancer risk from low-level radiation. AJR 2002; 179: Williams J. Is there a benefit in promoting the concept of radiation risk? Br J Radiol 2004; 77: Law J, Faulkner K. Cancers detected and induced, and associated risk and benefit, in a breast screening programme. Br J Radiol 2001; 74: Goodman PC. Radiation risk in CT screening for lung cancer. (abstr) Radiology 2002; 225[suppl]: Vaino H, Bianchini F, eds. International Agency for the Research on Cancer. Breast cancer screening handbook of cancer prevention, vol. 7. Lyon, France: International Agency for the Research on Cancer, 2002: AJR:187, August