DOI: 10.1016/j.athoracsur.2007.04.058 2007;84:959-966 Ann Thorac Surg

Cristiano Rampinelli, Giuseppe Trifirò, Angelica Sonzogni and Lorenzo Spaggiari Maisonneuve, Paolo Scanagatta, Francesco Leo, Giuseppe Pelosi, Laura Travaini,

Giulia Veronesi, Massimo Bellomi, Umberto Veronesi, Giovanni Paganelli, Patrick Nodules Detected at Baseline Computed Tomography Screening

Role of Positron Emission Tomography Scanning in the Management of Lung

http://ats.ctsnetjournals.org/cgi/content/full/84/3/959located on the World Wide Web at:

The online version of this article, along with updated information and services, is

Print ISSN: 0003-4975; eISSN: 1552-6259. Southern Thoracic Surgical Association. Copyright © 2007 by The Society of Thoracic Surgeons.

is the official journal of The Society of Thoracic Surgeons and theThe Annals of Thoracic Surgery

by on June 13, 2013 ats.ctsnetjournals.orgDownloaded from

RiaGGFCaDS

arbcsntt

faumCp

ub

LnbbmhdCtc

4

A

PS

AEe

©P

GEN

ERA

LT

HO

RA

CIC

ole of Positron Emission Tomography Scanningn the Management of Lung Nodules Detectedt Baseline Computed Tomography Screeningiulia Veronesi, MD, Massimo Bellomi, MD, Umberto Veronesi, MD,iovanni Paganelli, MD, Patrick Maisonneuve, ScD, Paolo Scanagatta, MD,

rancesco Leo, MD, Giuseppe Pelosi, MD, Laura Travaini, MD,ristiano Rampinelli, MD, Giuseppe Trifirò, MD, Angelica Sonzogni, MD,nd Lorenzo Spaggiari, MDepartments of Thoracic Surgery, Radiology, Nuclear Medicine, Epidemiology and Biostatistics, Division of Pathology, and

cientific Direction, European Institute of Oncology, Milan; and School of Medicine, University of Milan, Milan, Italy

5wwnbi

fifpnufpl

Background. Indeterminate noncalcified lung nodulesre a frequent finding when low-dose computed tomog-aphy (LD-CT) is used for lung cancer screening. Theest clinical management for such nodules remains un-ertain. We present results using positron tomographycanning (CT-PET) to evaluate LD-CT–detected lungodules during the first year of the Continuing Observa-

ion of Smoking Subjects (COSMOS) early detectionrial for lung cancer.

Methods. A total of 5200 asymptomatic current orormer smokers (>20 pack-years) older than 50 years ofge were enrolled in a single-institution screening trialsing annual LD-CT. Growing nodules and those with aaximum diameter exceeding 8 mm were studied withT-PET. Transthoracic needle biopsy was not a routineart of the protocol.Results. During the first year of study, 157 subjects

nderwent CT-PET, 66 of whom underwent surgical

iopsy. Of the 58 lung cancers found on surgical biopsy,

bbaetat

aTpan

2fihgo

uropean Institute of Oncology, Via Ripamonti 435, Milan, I-20141, Italy;- mail: giulia.veronesi@ieo.it.

2007 by The Society of Thoracic Surgeonsublished by Elsevier Inc

ats.ctsnetjournDownloaded from

1 were positive (standard uptake value >2.0) and sevenere negative for malignancy by CT-PET. Sensitivityas 88% overall, but 100% in the subgroup with solidodules of 10 mm or more. Among the 8 patients withenign disease at surgical biopsy, CT-PET was positive

n 6 and negative in 2.Conclusions. CT-PET is a highly promising modality

or identifying potentially malignant lesions in screen-ng-detected lung nodules and appears particularly use-ul as an alternative, in the screening setting, to invasiverocedures for the further investigation of uncertainodules. Our findings also indicate that the standardptake value threshold for positivity should be loweredor small nodules (<10 mm). Longer follow-up and largerrospective studies are necessary to confirm these pre-

iminary findings.

(Ann Thorac Surg 2007;84:959–66)

© 2007 by The Society of Thoracic Surgeons

ung cancer is the leading cause of cancer deathworldwide [1]. A major aim of research is to find

oninvasive methods of reliably diagnosing early diseaseecause surgery is effective in 80% of such cases [2]. Theest screening tool currently available is spiral low-doseultidetector computed tomography (CT). Pilot studies

ave shown that it is sensitive in detecting early stageisease [3–6], and the recent International Early Lungancer Action Program Investigators (I-ELCAP) publica-

ion [7] provided evidence that screening-detected lungancers have good long-term prognoses.

Noncalcified nodules are detected by low-dose CT in0% to 70% of screened subjects [6, 8, 9]. It has recently

ccepted for publication April 16, 2007.

resented at the Forty-third Annual Meeting of The Society of Thoracicurgeons, San Diego, CA, Jan 29–31, 2007.

ddress correspondence to Dr Veronesi, Division of Thoracic Surgery,

een shown that micronodules (�5 mm in diameter) cane safely checked once a year [10, 11], thereby consider-bly reducing the recall rate and screening costs. How-ver, nodules �5 mm, which require further investiga-ion, occur in 10% to 15% of screened subjects [7, 8, 12],nd there is no agreement on the best way of managinghem.

In most screening studies, percutaneous fine-needlespiration is used to investigate suggestive nodules [5].his technique requires considerable expertise, however,articularly to biopsy small deep lesions, and is associ-ted with low sensitivity for benign lesions [13, 14] andonnegligible morbidity [15].Positron emission tomography (PET) with F-18 fluoro-

-deoxyglucose (FDG) is a noninvasive technique usedor identifying malignant lung lesions [16] with sensitiv-ty of 80% to 100% in many series [17–21]. PET, however,as poor sensitivity for small lesions and those withround glass opacity (GGO), which are often bronchi-

loalveolar carcinoma, on CT. Nomori and colleagues

0003-4975/07/$32.00doi:10.1016/j.athoracsur.2007.04.058

by on June 13, 2013 als.org

[wn

tbueaalrvs

P

SAelpbdcsssntrt

CimisaN(

ud

Casti

wrbsiid

LTEswmwrwautwdgc

CTPwaws

T1

A

S

PN

a

i

As

960 VERONESI ET AL Ann Thorac SurgCT-PET AND LUNG CANCER SCREENING 2007;84:959–66

GEN

ERA

LT

HO

RA

CIC

22] found that sensitivity and specificity for GGO lesionsere only 10% and 20%, respectively, and for malignantodules smaller than 1 cm, PET had a sensitivity of 0%.Integrated PET and CT (CT-PET) is increasingly used

o image and characterize lesions in cancer patientsecause of its improved instrument resolution [23]. Thetility of CT-PET in a screening setting has not beenxtensively investigated, however [9, 24]. We conductedprospective study to assess the sensitivity, specificity,

nd accuracy of CT-PET in the differential diagnosis ofung nodules detected by low-dose spiral CT in high-isk individuals at baseline in the Continuing Obser-ation of Smoking Subjects (COSMOS) lung cancercreening trial [12].

atients and Methods

creening Protocolsymptomatic volunteers aged 50 years or older wereligible for recruitment to the COSMOS single-instituteung cancer screening trial if they were smokers (�20ack-years) or had stopped smoking in the 10 yearsefore study entry and did not have any malignantisease except adequately treated nonmelanoma skinancer. The protocol required once yearly low-dose CTcans for 5 years, baseline spirometry, compilation of twoelf-reporting questionnaires investigating respiratoryymptoms (British Medical Research Council question-aire, 1960) [25] and smoking habits, and blood sampling

o perform genetic marker and proteomic studies inelation to lung cancer risk. The study was approved byhe Ethics Committee of European Institute of Oncology.

All recruited volunteers gave written consent to annualT for 5 consecutive years and agreed they would not be

nformed of detected lesions with a maximum diameter of 5m or less, although such lesions would be specifically

nvestigated on the next scan. Nodules 5 mm or less werecheduled for repeat CT a year later. Nodules between 5.1nd 8 mm were scheduled for repeat CT 3 months later.oncalcified nodules of maximum size exceeding 8.1 mm

or growing lesions �8 mm after repeat scan) were sched-

able 1. Comparison of 157 Computed Tomography-Positron-Year Follow-upa

ALL TP TN FP

ll nodules 157 51 92 7�10 mmb 112 41 60 6�10 mmb 44 10 32 0

olid nodules 121 41 73 5�10 mm 83 33 45 5�10 mm 38 8 28 0

artially solid 30 10 18 2onsolid 6 0 1 0

Computed tomography-positron emission tomography findings were cos missing for 1 patient.

cc � accuracy; FN � false negative; FP � false positive; NPV � neensitivity; Spec � specificity; TP � true positive; TN � true negati

ats.ctsnetjournDownloaded from

led for PET-CT unless they appeared benign, as will beescribed in subsequent sections.Small lesions suggestive of malignancy (growing or

T-PET–positive) were scheduled for surgical biopsy anddditional interventions. In addition to PET scan, nonre-ectable lesions (advanced stage) received CT with con-rast of the upper abdomen, chest, and brain, and furthernterventions as determined by imaging outcomes.

Lesions of any size considered to be from infectionere treated with oral antibiotics for 10 days and CT was

epeated a month later. Other lesions were defined asenign-based on morphologic CT features [26]—ovalhape (maximum axial diameter greater than twice min-mum diameter), thickening of fissures (liquid density,nside apical scar)—and were scheduled for repeat lowose CT 3 months later.

ow-Dose Computed Tomographyhe CT equipment was a High Speed Advantage (Generallectric, Milwaukee, WI) 8-slice or 16-slice multidetectorcanner. No contrast was used for screening scans, whichere taken in a single breath (machine setting: 140 kVp, 30A, 1.75:1 pitch ratio, and 2.5-mm slice thickness). Imagesere retro-reconstructed using standard and lung algo-

ithms at 2.5-mm intervals. The dose equivalent per patientas estimated at 0.81 mSv. A radiologist read the images onworkstation (Advantage Windows 4.2, General Electric),sing lung parenchyma windows with maximum in-

ensity projection reconstruction and mediastinumindows. Findings of noncalcified lung nodules wereiscussed at meetings with radiologists, thoracic sur-eons, and nuclear medicine physicians. Nodules werelassified as solid, partially solid, or nonsolid [27].

omputed Tomography-Positron Emissionomographyatients were fasted 6 hours. After their blood glucoseas confirmed to be less than 150 mg/dL, they were

dministered 5 MBq/kg of FDG intravenously. They thenaited in calm conditions (minimum movement, no

peaking) for 50 to 60 minutes. Images were acquired

sion Tomography Investigations and Biopsy Findings, and

Sens Spec PPV NPV Acc

88% 93% 88% 93% 91%89% 91% 87% 92% 90%83% 100% 100% 94% 95%95% 94% 89% 97% 94%

100% 90% 87% 100% 94%80% 100% 100% 93% 95%

100% 90% 83% 100% 93%0% 100% — 17% 17%

red positive for maximum standard uptake value �2.0. b Nodule size

Emis

FN

75220205

nside

gative predictive value; PPV � positive predictive value; Sens �ve.

by on June 13, 2013 als.org

wGs

tdCuPii

aPpmtt

tectr

cor

SPwuttzppl

STficp

Fut

Fsmowo

T

S

�

�

�

�

�

�

�

a

tm

As

961Ann Thorac Surg VERONESI ET AL2007;84:959–66 CT-PET AND LUNG CANCER SCREENING

GEN

ERA

LT

HO

RA

CIC

ith a combined CT-PET in-line system (Discovery LS,E Medical Systems) consisting of an Advance NXi PET

canner and an 8-slice Light Speed Plus CT scanner.Patients were positioned headfirst supine and moved

o the CT scanning position. A scout scan was acquired toefine the axial imaging range, which for whole-bodyT-PET typically extended from the lower jaw to thepper thighs. CT settings were 140 kVp and 80 mA.atients were instructed to breathe normally. The CT

mage was matched to the section thickness of the PETmage, which covered the same axial field as the CT.

After completion of CT, the examination table wasdvanced by 60 cm into the PET gantry (axes of CT andET gantries in line). PET acquisition time was 4 minuteser table position, with a complete scan time of about 30inutes. PET image data sets were reconstructed itera-

ively, with segmented correction for attenuation usinghe CT data [28].

The CT-PET scan results were considered positive ifhe maximum standard uptake value (max-SUV) of FDGxceeded 2.0 in a region of interest (ROI) automaticallyalculated on lesions identified by CT. We used this lowhreshold because the consequences of a false-negativeesult (treatment delay) were more undesirable than the

ig 1. Receiver operating curve (line) for maximum standardizedptake value as a predictor of nodule malignancy. The area underhe curve is 0.95.

able 2. Prognostic Value of Maximum Standard Uptake Val

UV cutoff ALLb TPb TN FP

1.5 154 51 77 2210 mmc 111 41 45 2110 mmc 42 10 32 02.0 154 48 92 72.5 154 40 96 33.0 154 35 98 13.5 154 33 99 0

The reference standard was biopsy findings and 1-year follow-up.omography not performed at our institute so standard uptake value w

issing for 1 patient.

cc � accuracy; FN � false negative; FP � false positive; NPV � neensitivity; Spec � specificity; SUV � standard uptake value; TN �

ats.ctsnetjournDownloaded from

onsequences of a false-positive result (unnecessary bi-psy or surgery). We also assessed the diagnostic accu-acy of other thresholds.

urgical Biopsy and Treatmentatients with a suggestive lesion underwent diagnosticedge resection. A video thoracoscopic approach wassed whenever lesion site and size allowed, otherwise

he approach was through a standard muscle-sparinghoracotomy. The wedge underwent intraoperative fro-en section examination. If the pathology result wasositive, anatomic resection with curative intent waserformed immediately, usually associated with radical

ymph node dissection.

tatistical Analysishe CT-PET results were compared with histopathologicndings on surgical biopsy and with at least 12 months oflinical follow-up in the other cases. SUV distributions inatient subgroups are presented using box and whisker

ig 2. Maximum standardized uptake value in relation to noduleize in patients with malignant nodules. The horizontal line in theiddle of each box indicates the median; the top and bottom borders

f the box mark the 75th and 25th percentiles, respectively. Thehiskers mark the 90th and 10th percentiles. (� � mean; Œ �

utliers.)

Relation to Various Cutoff Valuesa

Sens Spec PPV NPV Acc

93% 78% 70% 95% 83%91% 68% 66% 92% 77%

100% 100% 100% 100% 100%87% 93% 87% 93% 91%73% 97% 93% 86% 88%64% 99% 97% 83% 86%60% 100% 100% 82% 86%

Excluding 3 patients with computed tomography-positron emissiont available (but the result was considered positive). c Nodule size is

ue in

FN

4407

152022

b

as no

gative predictive value; PPV � positive predictive value; Sens �true negative; TP � true positive.

by on June 13, 2013 als.org

pwdrSpttdppwN

R

Br5st

msbCfealboa

wib

lTm(sass

tetmCssooTtrtwfs

tvsfdt2

Frm2p

Fhltts

962 VERONESI ET AL Ann Thorac SurgCT-PET AND LUNG CANCER SCREENING 2007;84:959–66

GEN

ERA

LT

HO

RA

CIC

lots. The differences in median SUV between subgroupsere assessed using the nonparametric two-samples me-ian test. The diagnostic value of CT-PET was assessed byeceiver operating characteristics (ROC) curves of max-UV. Nonparametric ROC curves were generated bylotting sensitivity versus one-specificity, so that an ideal

est has sensitivity and specificity of one. The area underhe ROC curve (AUROC) was used as a measure of theiagnostic efficiency of a test. Sensitivity, specificity,ositive and negative predictive values, and accuracy areresented for several cutoff values of SUV. The analysesere performed with SAS software (SAS Institute, Cary,C). All p values are two-sided.

esults

etween October 2004 and October 2005, the study en-olled of 5200 volunteers (76% men) with a median age of7 years (range, 50 to 85 years), and 4816 subjects pre-ented for the second screening a year later. Medianobacco consumption was 44 pack-years.

At the baseline screening, 2197 volunteers had one orore noncalcified nodules of 5 mm or less and were

cheduled for annual follow-up, and 354 had a noduleetween 5 and 8 mm and were scheduled for low-doseT 3 months later, and 9 subsequently received CT-PET

or a growing nodule. Another 205 volunteers had a nodulexceeding 8 mm at baseline, 148 of whom received CT-PETnd 65 did not. Those who did not receive CT-PET had aesion compatible with infection and were prescribed anti-iotics and underwent repeat low-dose CT 3 months later,r had a lesion with clearly benign morphologic featuresnd were scheduled for repeat low-dose CT.

Thus, 157 patients underwent CT-PET. Their nodulesere solid in 121 cases, partially solid in 30, and not solid

n 6. A total of 58 lung cancers were diagnosed: 55 afteraseline screening and three others at screening 1 year

ig 3. Maximum standardized uptake value in relation to the finalesult. The horizontal line in the middle of each box indicates theedian; the top and bottom borders of the box mark the 75th and

5th percentiles, respectively. The whiskers mark the 90th and 10th

fercentiles. (� � mean; Œ � outliers.)

ats.ctsnetjournDownloaded from

ater. The mean size was 20.5 mm (range, 6 to 60 mm).here were 41 (71%) adenocarcinomas, eight (14%) squa-ous cell carcinomas, three (5%) typical carcinoids, three

5%) small cell lung cancers, and three (5%) not otherwisepecified non-small cell lung cancers. Three patients withdenocarcinoma had multifocal disease. Forty-one (69%)ubjects were stage I; 3 (5.2%) were stage II, 11 (19%) weretage III, and 3 (5.2%) were stage IV.

The max-SUV was available for 154 of 157 examina-ions; in the remaining three, FDG-PET was performedlsewhere and the SUV was not available, although allhree were reported suggestive of cancer. Table 1 sum-

arizes the variables of diagnostic accuracy for theT-PET examinations, with max-SUV exceeding 2.0 con-

idered a positive result. Of the 58 lung cancers found onurgical biopsy, 51 were positive and seven were negativen CT-PET. Overall sensitivity, specificity, and accuracyf the examination were 88%, 93%, and 91%, respectively.able 1 also shows results according to nodule size and

ype. For solid nodules, sensitivity, specificity, and accu-acy of CT-PET were 95%, 94%, and 94%, respectively;he corresponding values for nodules of 10 mm or moreere 89%, 91%, and 90%. CT-PET performance was best

or solid nodules of 10 mm or more (sensitivity 100%,pecificity 90%).

Figure 1 shows the ROC curve for max-SUV as predic-or of malignant nodules. The AUC was 95%, which isery close to 100%, indicating an ideal test result (100%ensitivity; 100% specificity); thus max-SUV is a good testor malignancy. Table 2 summarizes the variations iniagnostic accuracy with max-SUV threshold and illus-

rates that when the threshold exceeded 1.5, there were2 false-positive and 4 false-negative results. All of the

ig 4. Maximum standardized uptake value expression according toistology in the 58 patients with malignant tumors. The horizontal

ine in the middle of each box indicates the median; the top and bot-om borders of the box mark the 75th and 25th percentiles, respec-ively. The whiskers mark the 90th and 10th percentiles. (SCLC �mall cell lung cancer.) (� � mean; Œ � outliers.)

alse findings were confined to nodules 10 mm or larger,

by on June 13, 2013 als.org

amshfsipfi

1i10m

M20

5m3h

m(swfidmcscgnas

wPa

C

TCilvtit5

aiclhCpsmnS

m

TP

1234567

a

M mus

TM

12345678

a

t

963Ann Thorac Surg VERONESI ET AL2007;84:959–66 CT-PET AND LUNG CANCER SCREENING

GEN

ERA

LT

HO

RA

CIC

nd for nodules smaller than 10 mm, a threshold ofax-SUV of 1.5 was associated with 100% sensitivity,

pecificity, and accuracy. When the threshold exceeded 2,owever, there were seven false-negative and seven

alse-positive results; furthermore, the false-negative re-ults increased progressively as the positivity thresholdncreased. Thus, the max-SUV threshold of more than 2.0roduced the best combination of sensitivity and speci-city for all nodules, irrespective of size.Median max-SUV in patients with nodules of less than

0 mm was 2.8 (range, 1.7 to 16.2), and median max-SUVn patients with nodules of 10 mm or more was 4.8 (range,.0 to 21.1; p � 0.044). Figure 2 shows a significant (p �.008) association between max-SUV and nodule size inalignant nodules.Figure 3 shows max-SUV in relation to the final result.edian max-SUV in the seven false-positive nodules was

.5 (range, 2.1 to 3.0) and was 4.9 (range, 2.0 to 21.1; p �

.006) in the 51 true-positive nodules.Figure 4 shows max-SUV according to histology in the

5 malignant cases for which it was available. Medianax-SUV in squamous cell carcinoma was 12.7 (range,

.7 to 21.1) and median max SUV in patients with otheristology types was 3.8 (range, 1.0 to 17.1; p � 0.040).Table 3 summarizes the nodule size and type andax-SUV in the seven false-negative CT-PET results

SUV positivity threshold �2.0). Five patients had earlytage (N0) disease and underwent radical surgery, not-ithstanding the PET result (surgery indicated from CTndings). In 1 patient, the false-negative result delayediagnosis until the tumor was locally advanced (singleetastatic mediastinal lymph node). This patient re-

eived induction chemotherapy and radical surgical re-ection. In the other patient, who had severe cardiacomorbidity, the falsely negative nodule was a slow-rowing GGO. The patient refused treatment for thisodule, judged suggestive on baseline CT, and a moreggressive contralateral malignancy and distant metasta-is later developed.

Table 4 presents the pathologic results in the 8 patientsith no malignancy who underwent surgical biopsy. CT-ET was performed during the first year of screening,

able 3. Characteristics of Patients and Nodules in the Sevenositron Emission Tomographya

Patient Sex Age Max-SUVNodule Si

(mm)

M 59 �1 18M 68 1.7 7M 61 1.7 21F 58 1.4 17M 61 1.7 7F 50 �1 14M 60 1.3 15

One patient (stage IIIA, N2 disease) had a delayed diagnosis as a resul

ax SUV � maximum level of standardized uptake value; MBA �pecified; NSCLC � non- small cell lung cancer.

nd 6 had a positive result by this examination. M

ats.ctsnetjournDownloaded from

omment

he main findings of this prospective study are thatT-PET has satisfactorily high sensitivity (88%) and spec-

ficity (93%) for diagnosing lung nodules detected byow-dose CT screening. This result is particularly good iniew of the small size (median, 14 mm; mean, 18 mm) ofhe nodules investigated and is similar to findings emerg-ng in the nonscreening setting, where the sensitivity ofhe CT-PET examination is 83% to 100%, specificity is2% to 100%, and accuracy is 86% to 100% [29–31].For nodules of less than 10 mm, our diagnostic vari-

bles were somewhat better (83% sensitivity, 100% spec-ficity, and 95% accuracy) than overall, and in apparentontrast with Nomori and colleagues’ [22] finding that allesions up to 1 cm were negative on PET, regardless ofistology. Likely explanations are improved sensitivity ofT-PET compared with PET alone and a possibly higherroportion of difficult-to-diagnose GGO nodules in theireries. In our series, 10 of 12 solid lesions smaller than 10m were positive on CT-PET; the remaining 2 were false

egatives with max-SUVs of 1.6 and 1.7, higher thanUVs of true negative lesions.From our analysis of the prognostic values of differentax SUV thresholds, we found that adopting a max-SUV

e Negative Cases on Low-Dose Computed Tomography-

Nodule Type Tumor Type Final Stage

Nonsolid Adenocarcinoma T1 N0Solid Adenocarcinoma T1 N2Nonsolid NSCLC (NOS) T1 N0Nonsolid Adenocarcinoma T1 N0Solid Adenocarcinoma T1 N0Nonsolid Adenocarcinoma T1 N0 (MBA)Nonsolid Adenocarcinoma T1 N0

e false- negative result.

ltifocal bronchioloalveolar adenocarcinoma; NOS � not otherwise

able 4. Pathologic Finding in Eight Cases Negative foralignancy After Surgical Biopsya

PatientNodule Size

(mm) Max-SUV Definitive Pathology

15.5 2.06 Fibrosis34.2 2.47 Lymphoid hyperplasia17.2 2.21 Hamartochondroma14 3.01 Lymphoid hyperplasia33 2.96 Hamartoma12.8 2.9 Chronic inflammation7.9 1.15 Chronic inflammation

24.6 1.27 Hamartochondroma

Six cases were positive on computed tomography-positron emissionomography.

Fals

ze

t of th

ax-SUV � maximum level of standardized uptake value.

by on June 13, 2013 als.org

owicww

Ppdtwwmoa

Ftsanoediicss

rmhTaun

asa

dclwacNbtlo

smaat

piap

eddawfscptoPpn

eacof

adliw

aCCsptete

adpicl

R

964 VERONESI ET AL Ann Thorac SurgCT-PET AND LUNG CANCER SCREENING 2007;84:959–66

GEN

ERA

LT

HO

RA

CIC

f 1.5 for lesions smaller than 10 mm increased sensitivityithout decreasing specificity. We therefore suggest us-

ng 1.5 as the max-SUV cutoff for small lesions. Byontrast, for solid nodules of 10 mm or more, sensitivityas 100% and specificity was 90% when a max-SUV of 2.0as used as the cutoff.The 2005 study of Port and colleagues [32] found that

ET had a sensitivity of only 45% for the diagnosis ofatients with lung cancers of 2 cm or less and was of “noemonstrable benefit.” It is important to note, however,

hat Port and colleagues used a higher SUV cutoff thane did (2.5 versus 2.0), and sensitivity dropped to 73%hen we adopted a 2.5 SUV cutoff. Furthermore, theirean tumor size was smaller than the mean tumor size in

ur study (14 mm versus 20.5 mm), and this may havelso have contributed to sensitivity reduction.We found that nonsolid nodules do not have altered

DG metabolism, so PET is nondiagnostic; however lesshan 10% of malignant lesions in our study were non-olid. It is also noteworthy that most of nonsolid nodulesssessed by CT-PET and biopsied surgically were malig-ant. The implication is that the radiologic characteristicsf nonsolid nodules suggestive for bronchioalveolar ad-nocarcinoma are so typical that CT alone is enough toiagnose malignancy. Thus, no diagnosis delays occurred

n nonsolid nodules due to false negative CT-PET find-ngs. We also found that mixed nodules with a solidomponent (partially solid) had CT-PET characteristicsimilar to solid nodules, and the examination had highensitivity, specificity, and accuracy for these nodules.

Comparison of false-positive and true-positive nodulesevealed they differed significantly in terms of medianax-SUV, and that false-positive (benign) lesions never

ad a max-SUV above 3.01, even if the lesion was large.his finding may be useful for interpreting PET results,nd perhaps justifying postponement of surgical biopsyntil a subsequent follow-up CT suggests that it is aecessity.Collateral findings included a significant direct associ-

tion between nodule size and SUV, and that SUV wasignificantly higher in squamous cell carcinoma than inll other tumor types combined.It is important to assess CT-PET in relation to other

iagnostic modalities in the screening setting. Bronchos-opy with brush cytology must be excluded because of itsow diagnostic yield for lesions smaller than 2 cm [33],

hereas transthoracic needle aspiration biopsy is char-cterized by false-negative levels of up to 30% [13] and aomplication rate of up to 30% in some series [14].odule-enhancement CT has also been investigated [34],ut the authors concluded that FDG-PET was preferable

o nodule-enhancement CT in evaluating indeterminateung nodules because of its much higher specificity andnly slightly reduced sensitivity.In the future, it is likely that we will be able to use

erum or sputum biomarkers to define the risk of lungalignancy in individuals exposed to carcinogens, en-

bling the timing of follow-up CT scans to be optimizednd also suggesting the best type of invasive investiga-

ions. In the meantime, CT-PET appears to be the most

ats.ctsnetjournDownloaded from

romising noninvasive imaging procedure in the screen-ng setting because of its high specificity and sensitivitys well as its ability to reduce the number of invasiverocedures performed and their attendant complications.The important additional dimension of the cost-

ffectiveness of screening CT-PET has also been ad-ressed. The study of Lejeune and colleagues [35] used aecision analysis model to compare costs in three man-gement scenarios for solitary lung nodules: wait andatch, PET with anatomic CT, and CT plus PET. They

ound that CT plus PET was not only the most effectivetrategy of the three but also had the lowest incrementalost-effectiveness ratio. Gugiatti and colleagues [36] com-ared the addition of PET with traditional workup (CT,

ransthoracic needle aspiration biopsy, or surgical bi-psy) for solitary lung nodules. They found addition ofET was associated with a cost reduction of about €50 peratient because it reduced inappropriate invasive diag-ostic investigations and their complications.In the present study, we chose to use only instrumental

stimates of FDG uptake as an indicator of nodule naturend did not include operator (visual) estimates. Thishoice is appropriate in view of a hypothesized extensionf the study to a regional or national screening programor lung cancer.

A limitation of CT-PET is that it may not easily bepplicable to populations where histoplasmosis is en-emic, because the false-positive rate would be high. A

imitation of our study is the short 1-year follow-up, andt is possible that some very slow growing cancers tumorsere not recognized.The COSMOS study is ongoing, and second-year data

re being analyzed. We are expanding our assessment ofT-PET in this setting by investigating relations betweenT-PET findings and the growth rate of malignant le-

ions, evaluating the staging role of the technique (im-lications for minimally invasive surgery, resection ex-

ension and lymph node dissection), comparing CTnhancement and CT-PET findings, further evaluatinghe utility of SUV, and also performing a cost-ffectiveness analysis.CT-PET is characterized by high sensitivity, specificity,

nd accuracy in the diagnosis of malignant lung lesionsetected by low-dose CT in the screening setting. It isarticularly promising, because it reduces the use of

nvasive diagnostic procedures. Our findings also indi-ate that the SUV threshold for positivity should beowered for small nodules (less than 10 mm).

eferences

1. Jemal A, Thomas A, Murray T, et al. Cancer statistics, 2002.CA Cancer J Clin 2002;52:23–47.

2. National Cancer Institute. SEER Cancer Statistics Review,1996 –2002. Available at: http://seer.cancer.gov/statistics/.Accessed Apr 24, 2007.

3. Humphrey LL, Teutsch S, Johnson M. Lung cancer screeningwith sputum cytologic examination, chest radiography, andcomputed tomography: an update for the U.S. Preventive

Service Task Force. Ann Intern Med 2004;140:740–53.

by on June 13, 2013 als.org

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

D

DsD

Dbh

965Ann Thorac Surg VERONESI ET AL2007;84:959–66 CT-PET AND LUNG CANCER SCREENING

GEN

ERA

LT

HO

RA

CIC

4. Mulshine JL, Sullivan DC. Clinical practice. Lung cancerscreening. N Engl J Med 2005;352:2714–20.

5. Henschke CI, McCauley DI, Yankelevitz DF, et al. EarlyLung Cancer Action Project: overall design and findingsfrom baseline screening. Lancet 1999;354:99–105.

6. Diederich S, Thomas M, Semik M, et al. Screening for earlylung cancer with low dose spiral computed tomography:results of annual follow-up examinations in asymptomaticsmokers. Eur Radiol 2004;14:691–702.

7. The International Early Lung Cancer Action Program Inves-tigators. Survival of patients with stage I lung cancer de-tected on CT screening. N Engl J Med 2006;355:1763–71.

8. Swensen SJ, Jett JR, Hartman TE, et al. Lung cancer screen-ing with CT: Mayo Clinic experience. Radiology 2003;226:756–61.

9. Bastarrika G, Garcia-Velloso MJ, Lozano MD, et al. Earlylung cancer detection using spiral computed tomographyand positron emission tomography. Am J Respir Crit CareMed 2005;171:1378–85.

0. Henschke CI, Yankelevits DF, Naidich DP, et al. CT screen-ing for lung cancer. Suspiciousness of nodules accordingwith size on baseline scans. Radiology 2004;231:164–8.

1. Bellomi M, Rampinelli C, Veronesi G, Ferretti S, De Fiori E,Maisonneuve P. Evolution of pulmonary nodules �5 mm de-tected with low-dose CT in asymptomatic smokers. BJR [in press].

2. Veronesi G, Bellomi M, Spaggiari L, et al. Low dose spiralcomputed tomography for early diagnosis of lung cancer.Results of baseline screening in 5000 high-risk volunteers.2006 ASCO Annual Meeting Proceedings, Part I. J Clin Oncol2006;24(18S):7029.

3. Geraghty PR, Kee ST, McFarlane G, Razavi MK, Sze DY,Dake MD. CT-guided transthoracic needle aspiration biopsyof pulmonary nodules: needle size and pneumothorax rate.Radiology 2003;229:475–81.

4. Quint LE, Kretschmer M, Chang A, Nan B. CT-guidedthoracic core biopsies: value of a negative result. CancerImaging 2006;6:163–7.

5. Yeow KM, See LC, Lui KW, et al. Risk factors for pneumo-thorax and bleeding after CT-guided percutaneous coaxialcutting needle biopsy of lung lesions. J Vasc Interv Radiol2001;12:1305–12.

6. Gould MK, Maclean CC, Kuschner WG, Rydzak CE, OwensDK. Accuracy of positron emission tomography for diagno-sis of pulmonary nodules and mass lesions: a meta-analysis.JAMA 2001;285:914–24.

7. Christensen JA, Nathan MA, Mullan BP, Hartman TE,Swensen SJ, Lowe VJ. Characterization of the solitary pul-monary nodule: 18F-FDG PET versus nodule-enhancementCT. AJR Am J Roentgenol 2006;187:1361–7.

8. Hashimoto Y, Tsujikawa T, Kondo C, et al. Accuracy of PETfor diagnosis of solid pulmonary lesions with 18F-FDGuptake below the standardized uptake value of 2.5. J NuclMed 2006;47:426–31.

9. Yi CA, Lee KS, Kim BT, et al. Tissue characterization ofsolitary pulmonary nodule: comparative study between he-lical dynamic CT and integrated PET/CT. J Nucl Med 2006;47:443–50.

0. Lowe VJ, Fletcher JW, Gobar L, et al. Prospective investiga-tion of positron emission tomography in lung nodules. J Clin

Oncol 1998;16:1075–84.

ad grants for the CT screening but not for the diagnostic work-up

obrh

Dt

ats.ctsnetjournDownloaded from

1. Hickeson M, Yun M, Matthies A, et al. Use of a correctedstandardized uptake value based on the lesion size on CTpermits accurate characterization of lung nodules on FDG-PET. Eur J Nucl Med Mol Imaging 2002;29:1639–47.

2. Nomori H, Watanabe K, Ohtsuka T, Naruke T, Suemasu K,Uno K. Evaluation of F-18 fluorodeoxyglucose (FDG) PETscanning for pulmonary nodules less than 3 cm in diameter,with special reference to the CT images. Lung Cancer2004;45:19–27.

3. Devaraj A, Cook GJ, Hansell DM. PET/CT in non-small celllung cancer staging-promises and problems. Clin Radiol2007;62:97–108.

4. Pastorino U, Bellomi M, Landoni C, et al. Early lung-cancerdetection with spiral CT and positron emission tomographyin heavy smokers: 2-year results. Lancet 2003;362:593–7.

5. Fletcher CM (chairman). Standardised questionnaire onrespiratory symptoms: a statement prepared and ap-proved by the MRC Committee on the Aetiology ofChronic Bronchitis (MRC breathlessness score). Br Med J1960;2:1655.

6. Diederich S, Hansen J, Wormanns D. Resolving small pul-monary nodules: CT features. J Eur Radiol 2005:10;2064–9.

7. Henschke CI, Yankelevitz DF, Mirtcheva R, McGuinness G,McCauley DI, Miettinen OS. CT screening for lung cancer:frequency and significance of part-solid and nonsolid nod-ules. AJR Am J Roentgenol 2002;178:1053–7.

8. Kamel E, Hany TF, Burger C, et al. CT vs 68Ge attenuationcorrection in a combined PET/CT system: evaluation of theeffect of lowering the CT tube current. Eur J Nucl Med MolImaging 2002;29:346–50.

9. Vansteenkiste JF, Stroobants SG. The role of positron emis-sion tomography with 18F-fluoro-2-deoxy-D-glucose in re-spiratory oncology. Eur Respir J 2001;17:802–20.

0. Gould MK, Maclean CC, Kuschner WG, Rydzak CE, OwensDK. Accuracy of positron emission tomography for diagno-sis of pulmonary nodules and mass lesions: a meta-analysis.JAMA 2001;21;285:914–24.

1. Fischer BM, Mortensen J, Hojgaard L. Positron emissiontomography in the diagnosis and staging of lung cancer: asystematic, quantitative review. Lancet Oncol 2001;2:659–66.

2. Port JL, Andrade RS, Levin MA, et al. Positron emissiontomographic scanning in the diagnosis and staging of non-small cell lung cancer 2 cm in size or less. J Thorac Cardio-vasc Surg 2005;130:1611–5.

3. Wallace JM, Deutsch AL. Flexible fiberoptic bronchoscopyand percutaneous needle lung aspiration for evaluating thesolitary pulmonary nodule. Chest 1982;81:665–71.

4. Christensen JA, Nathan MA, Mullan BP, Hartman TE,Swensen SJ, Lowe VJ. Characterization of the solitary pul-monary nodule: 18F-FDG PET versus nodule-enhancementCT. AJR Am J Roentgenol 2006;187:1361–7.

5. Lejeune C, Al Zahouri K, Woronoff-Lemsi MC, et al. Use ofa decision analysis model to assess the medicoeconomicimplications of FDG PET imaging in diagnosing a solitarypulmonary nodule. Eur J Health Econ 2005;6:203–14.

6. Gugiatti A, Grimaldi A, Rossetti C, et al. Economic analyseson the use of positron emission tomography for the work-upof solitary pulmonary nodules and for staging patients withnon-small-cell-lung-cancer in Italy. J Nucl Med Mol Imaging

2004;48:49–61.

ISCUSSION

R SCOTT J. SWANSON (New York, NY): How do you pay fortudies like this? Is this something funded by your government?o you get grants? How are you paying for this?

R VERONESI: That is a good question. In fact there was a littleit of discussion with the region that had to pay for the CT-PET. We

f screening-detected nodules that in our opinion should haveeen paid by the national health system. In our experience, weeached a compromise: we paid half of the CT-PET and the otheralf was paid by the national health system.

R GEORGE B. HAASLER (Milwaukee, WI): In the nodules

hat were PET-positive, how many of the mediastinal lymph

by on June 13, 2013 als.org

nstla

Dodsawc

Drltluetopsn

Dlwsrwap

DtaedP

Dyi

tptt

Dslttymatnyp

DllaissFnc

Doas

Datosc

Dt

966 VERONESI ET AL Ann Thorac SurgCT-PET AND LUNG CANCER SCREENING 2007;84:959–66

GEN

ERA

LT

HO

RA

CIC

odes were positive in that setting? In other words, what was theensitivity for the study for those lymph nodes for the noduleshat were positive? Also, did the degree or the likelihood ofymph node positivity in the mediastinum relate at all to howvid the nodules were?

R VERONESI: The staging accuracy of PET was not the mainbjective of this paper, so I have only preliminary data andetails will be analyzed in a near future. Anyway we had a highensitivity for nodal staging in the stage III disease that werebout 15% of the tumors found with screening. The cases thatere positive at PET scan underwent biopsy with mediastinos-

opy and induction chemotherapy.

R FREDERIC W. GRANNIS (Duarte, CA): Despite the recentesults of the ELCAP study, which showed excellent results inong-term survival, we’re having a hard time convincing CMS,he American Cancer Society, et cetera, in the United States thatung cancer screening is safe and effective, and I would congrat-late you on this study that appears to show what we havemphasized, that lung cancer screening is a process, not just aest, that we must consider the false positives within the contextf the protocol of diagnosis, and, finally, that good protocolsrevent unnecessary operations. How many patients in youreries had a surgical operation for a nonmalignant pulmonaryodule out of your 5000 screened patients?

R VERONESI: I presented these data at the ASCO meetingast June; the final result of our first year of screening concludedith 13% of surgical procedures for benign disease. We had

even cases of benign disease detected at surgery with minoresection out of 62 surgical procedures. We compared these dataith our standard practice in our division and found that it waslittle bit lower than our standard result in a nonscreened

opulation.

R ARA VAPORCIYAN (Houston, TX): I enjoyed your presen-ation very much. I wanted you to comment on the generaliz-bility of your findings to areas of the world that have morendemic infectious diseases, such as histoplasmosis and otheriseases, which really increase the rate of false positivity on aET.

R VERONESI: I think that the implication of this problem thatou have underlined could result in a higher percentage of

nvasive procedures for benign disease; however, the optimiza- D

ats.ctsnetjournDownloaded from

ion of the diagnostic protocol with repeated low-dose CTs inositive cases to see the development of the lesions associated to

he use of antibiotics before one CT scan and the other will limithe false positives.

R RICARDO S. SANTOS (Pittsburgh, PA): If we all do lung CTcreening programs, we’re going to find smaller and smalleresions, and then we present the case with a very small lesionhat you do a lobectomy with radical lymphadenectomy. I knowhat this is a little bit out of the scope, but I would like to hearour comment about which type of operation you would recom-end to patients in whom their lesions are less than 2 cm. I havelot of data from the literature, the Japanese literature and also

he American literature, that these patients probably would noteed a lobectomy, that segmentectomy could be enough. Couldou please comment on that? Congratulations on yourresentation.

R VERONESI: Thank you. My personal opinion is that for aesion smaller than 15 mm, in a peripheral position, and with aow level of SUV, probably a wide wedge resection would be anppropriate treatment, but this should be confirmed and studiedn randomized trials. For the moment, I think that as thiscreening process is in an experimental phase, we must use atandardized technique for treatment not to add other variables.or this reason, we still consider lobectomy with radical lymphode dissection the standard extension of resection for theseases.

R NASSER K. ALTORKI (New York, NY): Your results arebviously spectacular. I may have missed this, but did you drawdistinction between findings on the baseline and the repeat

creen?

R VERONESI: I didn’t bring those data in this setting, but were analyzing the results of the annual screening. My feeling ishat small lesions that have a very slow growth rate are moreften negative at PET scan. So maybe the data of the annualcreening may produce a lower level of sensitivity because aertain number of nodules have grown very, very slowly.

R ALTORKI: It would also improve your specificity comparedo the baseline.

R VERONESI: Probably yes.

by on June 13, 2013 als.org

DOI: 10.1016/j.athoracsur.2007.04.058 2007;84:959-966 Ann Thorac Surg

Cristiano Rampinelli, Giuseppe Trifirò, Angelica Sonzogni and Lorenzo Spaggiari Maisonneuve, Paolo Scanagatta, Francesco Leo, Giuseppe Pelosi, Laura Travaini,

Giulia Veronesi, Massimo Bellomi, Umberto Veronesi, Giovanni Paganelli, Patrick Nodules Detected at Baseline Computed Tomography Screening

Role of Positron Emission Tomography Scanning in the Management of Lung

& ServicesUpdated Information

http://ats.ctsnetjournals.org/cgi/content/full/84/3/959including high-resolution figures, can be found at:

References http://ats.ctsnetjournals.org/cgi/content/full/84/3/959#BIBL

This article cites 34 articles, 10 of which you can access for free at:

Citations http://ats.ctsnetjournals.org/cgi/content/full/84/3/959#otherarticles

This article has been cited by 7 HighWire-hosted articles:

Subspecialty Collections

http://ats.ctsnetjournals.org/cgi/collection/lung_cancer Lung - cancer

following collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

.healthpermissions@elsevier.comemail: orhttp://www.us.elsevierhealth.com/Licensing/permissions.jsp

in its entirety should be submitted to: Requests about reproducing this article in parts (figures, tables) or

Reprints reprints@elsevier.com

For information about ordering reprints, please email:

by on June 13, 2013 ats.ctsnetjournals.orgDownloaded from