1. Editorials
Radiology
Heber MacMahon, MB, Guidelines for Management of
BCh, BAO
John H. M. Austin, MD Small Pulmonary Nodules
Gordon Gamsu, MD
Christian J. Herold, MD
James R. Jett, MD
Detected on CT Scans:
David P. Naidich, MD
Edward F. Patz, Jr, MD
A Statement from the
Stephen J. Swensen, MD
Fleischner Society1
Published online
10.1148/radiol.2372041887
Radiology 2005; 237:395– 400
Lung nodules are detected very commonly on computed tomographic (CT) scans of
the chest, and the ability to detect very small nodules improves with each new
1
From the Department of Radiology, generation of CT scanner. In reported studies, up to 51% of smokers aged 50 years
University of Chicago, 5841 S Mary-
land Ave, MC2026, Chicago, IL 60637 or older have pulmonary nodules on CT scans. However, the existing guidelines for
(H.M.); Department of Radiology, Co- follow-up and management of noncalcified nodules detected on nonscreening CT
lumbia University Medical Center, scans were developed before widespread use of multi– detector row CT and still
New York, NY (J.H.M.A.); Department
of Radiology, New York Hospital, Cor-
indicate that every indeterminate nodule should be followed with serial CT for a
nell Medical Center, New York, NY minimum of 2 years. This policy, which requires large numbers of studies to be
(G.G.); Department of Radiology, Uni- performed at considerable expense and with substantial radiation exposure for the
versity of Vienna Medical School, Vi- affected population, has not proved to be beneficial or cost-effective. During the
enna, Austria (C.J.H.); Departments of
Medicine (J.R.J.) and Diagnostic Radi- past 5 years, new information regarding prevalence, biologic characteristics, and
ology (S.J.S.), Mayo Clinic College of growth rates of small lung cancers has become available; thus, the authors believe
Medicine, Rochester, Minn; Depart- that the time-honored requirement to follow every small indeterminate nodule with
ment of Radiology, NYU Medical Cen-
ter, New York, NY (D.P.N.); and De- serial CT should be revised. In this statement, which has been approved by the
partment of Radiology, Duke Univer- Fleischner Society, the pertinent data are reviewed, the authors’ conclusions are
sity Medical Center, Durham, NC summarized, and new guidelines are proposed for follow-up and management of
(E.F.P.). Received November 5, 2004;
small pulmonary nodules detected on CT scans.
revision requested January 5, 2005; re-
©
vision received March 3; accepted May RSNA, 2005
2. Address correspondence to H.M.
(e-mail: macm@midway.uchicago.edu).
© RSNA, 2005
Until now, it has been the accepted standard of practice to regard all noncalcified pulmo-
nary nodules as potentially malignant lesions that require close monitoring until proved
stable over a period of 2 years (1,2). This approach was adopted prior to the widespread use
of computed tomography (CT) and was based on the observation that a substantial
proportion of noncalcified nodules that were detected at chest radiography turned out to
be lung cancers. These nodules were almost all larger than 5 mm in diameter, and most
were in the 1–3-cm range.
Since the introduction of helical CT in the early 1990s and multi– detector row CT in the
late 1990s, the detection of focal rounded pulmonary opacities (“nodules”) as small as 1–2
mm in diameter has become routine. In fact, the majority of smokers who undergo
thin-section CT have been found to have small lung nodules, most of which are smaller
than 7 mm in diameter (3). However, the clinical importance of these extremely small
nodules differs substantially from that of larger nodules detected on chest radiographs, in
that the vast majority are benign. This issue has been highlighted in several recent
publications on CT screening for lung cancer, and the positive relationship of lesion size
to likelihood of malignancy has been clearly demonstrated (4,5).
The only current guidelines in the radiology literature for management of small nodules
are those that have been developed in the context of lung cancer screening programs
(6 – 8). However, subjects who undergo lung cancer screening in most countries are
selected on the basis of age, substantial smoking history, absence of serious comorbid
disease, and willingness to participate in all necessary follow-up imaging and intervention.
Also, these programs tend to take an aggressive approach to follow-up and early interven-
395

2. tion, with a view to achieving the highest nodule that would qualify for this proto- CT. If the nodule did not exhibit un-
possible cure rate while gaining further col (2). equivocally benign characteristics (be-
insight into the behavior and character- The rationale for recommending serial nign pattern of calcification, smooth
istics of small malignant lesions. There- follow-up studies for all indeterminate margins, and size less than 20 mm), it
fore, patients whose nodules are detected small nodules is that some of them will was followed up with thin-section CT at
Radiology
incidentally during the course of CT per- turn out to be cancers and that early in- 3 months and subsequently at 6, 12, and
formed for other reasons should not nec- tervention will provide an opportunity 24 months in the absence of change.
essarily be treated in the same way as for cure. The “downside” of this policy Only one cancer was less than 5 mm in
subjects in a screening program. Nonethe- includes potential morbidity and mor- diameter at the time of detection in the
less, although CT screening has not as yet tality from surgery for benign nodules baseline (10-mm-collimation) study. The
been proved to help reduce mortality from and other false-positive findings, poor results of the repeat screening were re-
lung cancer, these programs provide an utilization of limited resources, in- ported in 2001. Seven new cancers were
important source of information for deter- creased health care costs, unnecessary detected at the repeat screening; three of
mining the optimal management of “inci- patient anxiety, loss of credibility for these were 5 mm in size (defined as the
dental” nodules detected in other situa- radiologists who may seem to recom- average of the length and width in that
tions. mend excessive numbers of CT scans study); all the other proved cancers were
Our intent in this position statement is with little benefit, and increased radia- larger (8).
to provide practical guidelines for the tion burden for the affected popula- Swensen et al (7) reported the results of
management of small pulmonary nod- tion. The radiation issue is particularly the Mayo Clinic Lung Cancer Screening
ules that are detected during the course important in younger patients and Trial after completion of three annual
of CT examinations performed for pur- must be taken into account in deter- low-dose CT examinations in 1520 smok-
poses other than lung cancer screening. mining appropriate follow-up strategies ers. The subjects were aged 50 years or
We recognize that this issue is complex (10 –13). older and had a smoking history of 20
and difficult to reduce to a simple algo- There remains a growing need to reex- pack-years or more. Two years after base-
rithm. We also realize that the definition amine the radiologic approach to small line screening, 2832 noncalcified pulmo-
of a pulmonary nodule is itself elusive, nodules, particularly when CT is per-
nary nodules had been identified in 1049
and that not all focal opacities qualify as formed for indications other than screen-
(69%) participants. Thirty-six lung can-
nodules. Yet, we believe that there is a ing. Several fundamental issues need to
cers were diagnosed with the aid of CT
practical need in the medical community be considered before an imaging strategy
(2.6% of participants, 1.4% of nodules):
for guidance in this area. is recommended. Since the decision to
26 at baseline (prevalence) and 10 at sub-
perform follow-up studies relies on size,
sequent annual (incidence) CT examina-
lesion characteristics (eg, morphology),
BACKGROUND tions. Thirty-two (80%) of these cancers
and growth rates (typically described as
were larger than 8 mm, and only one was
doubling time), an understanding of
smaller than 5 mm at the time of detec-
Available data indicate that fewer than these features and their relationship to
tion. Two cancers were detected by using
1% of very small ( 5-mm) nodules in malignancy should dictate further evalu-
sputum cytology, and there were two in-
patients without a history of cancer will ation. In addition, the patient’s risk pro-
terval cancers.
demonstrate malignant behavior (ie, de- file, including age and smoking history,
tectable growth or metastases over a pe- needs to be integrated into the diagnostic Benjamin et al (14) reported the results
riod of 2 or more years) (5,7,9). Nonethe- algorithm. of a retrospective review of 334 non-
less, multiple follow-up examinations screening cases in which a lung nodule or
over a 2-year period are commonly per- nodules were detected that were less than
Nodules Detected on CT Scans 10 mm in their long axis and for which
formed when such nodules are detected
incidentally. Guidelines for the manage- During the past 5 years, new informa- follow-up CT was recommended. These
ment of the solitary pulmonary nodule tion regarding the morphology, biologic nodules were identified from 3446 con-
were published in 2003 by the American characteristics, and growth rates of small secutive chest CT studies performed at
College of Chest Physicians (2) and in a lung cancers has become available from their institution. Among patients with a
review in by Ost and colleagues in the CT lung cancer screening programs nodule, 87 patients had “definitive” fol-
New England Journal of Medicine (1). How- throughout the world. Selected studies low-up results (2 years without change or
ever, neither of these articles specifically are summarized below as representative biopsy). Nodules in 10 (11%) of these
addressed the issue of the very small nod- of the present state of knowledge in this patients were malignant. Nine of these
ule that is detected as an incidental find- area. cases were metastases from known pri-
ing on a CT scan. Ost and colleagues con- Henschke et al (9) published the initial mary tumors, and one turned out to be a
cluded by recommending CT follow-up results from the Early Lung Cancer Ac- metastasis from an occult primary tumor.
at 3, 6, 12, 18, and 24 months for all tion Project CT screening project in 1999. This high incidence of metastases is a
“low-probability” indeterminate nodules They enrolled 1000 asymptomatic smok- reflection of the patient population that
(ie, nodules with a low likelihood of be- ers or ex-smokers with at least 10 pack- was used for this study, in that 56% of
ing cancer), regardless of size (1). The years of cigarette smoking who were aged the included patients had a known pri-
American College of Chest Physicians 60 years or older. Screening CT studies mary tumor. The authors of that article
recommendation for management of in- were performed with 10-mm collima- suggested that the malignancy rate in
determinate solitary nodules was similar, tion. The investigators found noncalci- small nodules in patients without a
with the proposal for 3-, 6-, 12-, and 24- fied nodules in 23% of subjects and ma- known neoplasm may be as low as 1%
month CT follow-up intervals, also with- lignant nodules in 2.7%. All detected and that follow-up may not be necessary
out any specified lower size limit for a nodules were subjected to thin-section in such cases.
396 Radiology November 2005 MacMahon et al

3. Nodule Size pers, or software that performs auto- diameter of 20.3 mm in 12 months,
mated volumetric measurement of the whereas a similar nodule with a doubling
On the basis of analysis of information
nodule (15). The minified display used time of 240 days would reach a diameter
from the ongoing Mayo Clinic CT
for hard copy makes accurate measure- of only 7.1 mm in the same period.
Screening Trial, Midthun et al (4) re-
ment of subcentimeter nodules virtually
ported that fewer than 1% of very small
Radiology
impossible. Even with a state-of-the-art
( 5-mm) nodules in patients without a Relative Risk
soft-copy display and electronic calipers,
history of cancer were malignant. They
substantial inter- and intrareader varia- Sone et al (22) reported the results of a
indicated a likelihood of malignancy of
tions in two-dimensional measurements CT screening program carried out in Ja-
0.2% for nodules smaller than 3 mm,
have been documented (16). Revel and pan in 1996 –1998 that enrolled 5483
0.9% for those 4 –7 mm, 18% for those colleagues (16) determined that two-di- subjects. They found an equal prevalence
8 –20 mm, and 50% for those larger than mensional measurements obtained with (0.5%) of lung cancer in smokers and
20 mm. electronic calipers were unreliable as a nonsmokers, although the death rate
Henschke and colleagues (5) recently basis for distinguishing benign from ma- from lung cancer in smokers in Japan is
addressed the issue of optimal follow-up lignant solid nodules in the 5–15-mm approximately four times higher than
intervals for nodules smaller than 5 mm size range. Limited studies (15,17) that that in nonsmokers (23). This discrep-
in diameter. They performed a retrospec- have been performed with automated ancy may be explained by an overdiag-
tive review of a total of 2897 baseline three-dimensional analysis suggest that nosis bias, as many of these slow-growing
screening studies performed between such systems can be considerably more cancers in nonsmokers might never have
1993 and 2002 in order to identify those accurate and consistent than unaided been detected or become symptomatic if
subjects with noncalcified nodules that, radiologists in determining interval the subjects had not been screened. It is
at initial detection, were (a) smaller than changes in small nodules. However, Kos- unclear how many of the nonsmokers in
5 mm in diameter and (b) 5–9 mm in tis and colleagues (18) have shown that this study may have been affected by sec-
diameter. In accordance with the screen- even sophisticated automated volume- ond-hand smoke and whether this phe-
ing program guidelines, indeterminate measurement methods are susceptible to nomenon, or underreporting of smoking
nodules were rescanned at 3-, 6-, and 12- errors due to motion artifacts or segmen- by screening subjects, might account for
month intervals, while some in the tation problems, particularly in the case the relatively high lung cancer mortality
5–9-mm range that had particularly sus- of very small nodules. in nonsmokers in Japan compared with
picious morphology were recommended that in the United States (24).
for biopsy. On the basis of the results of The relative risk for developing lung
these follow-up studies and biopsies, the Growth Rate
carcinoma in male smokers was about 10
authors determined that when the largest Hasegawa et al (19) reported an analy- times that in nonsmokers in the eight
noncalcified nodule was smaller than 5 sis of the growth rates of small lung can- prospective studies reviewed for the 1982
mm in diameter (378 patients), a fol- cers detected during a 3-year mass screen- report of the Surgeon General on “The
low-up study in 12 months would have ing program. They classified nodules as Health Consequences of Smoking” (25).
resulted in no case of delayed diagnosis, ground-glass opacity, as ground-glass For heavy smokers, the risk was 15–35
compared with more aggressive short- opacity with a solid component, or as times greater (25,26). Despite initial evi-
term follow-up. However, when the larg- solid. Mean volume doubling times were dence suggesting an increased risk of
est nodule was 5–9 mm in diameter, ap- 813 days, 457 days, and 149 days, respec- lung cancer in women compared with
proximately 6% of cases (all of which tively, for these three types, all of which that in men with an equal smoking his-
were malignant) showed interval nodule were significantly different. In addition, tory, this has not been confirmed in
growth detectable on 4 – 8-month fol- the mean volume doubling time for can- more recent studies (27–30).
low-up scans. Therefore, they recom- cerous nodules in nonsmokers was signif- A history of lung cancer in first-degree
mended that patients with nodules no icantly longer than that for cancerous relatives is also a notable risk factor, and
larger than 5 mm in diameter on a base- nodules in smokers. The mean volume strong evidence for a specific lung cancer
line screening CT scan should be referred doubling time was also significantly susceptibility gene has been discovered
for repeat annual screening in 12 months longer for nodules not visible on chest recently (31,32). Other established risk
time, with no interval scans. This recom- radiographs (presumably a function of factors include exposure to asbestos, ura-
mendation differs only slightly from the their smaller average size and/or their nium, and radon (33–35). However, cig-
current protocol for the ongoing Na- lesser average opacity). These data further arette smoke remains the overwhelm-
tional Lung Screening Trial, in which pa- support the use of extended follow-up ingly dominant culprit.
tients with nodules smaller than 4 mm in intervals for small nonsolid or partly Several large screening programs are
diameter are recommended to return for solid nodules, even in high-risk patients. continuing, and we will learn more from
screening after 12 months, without inter- Authors of a number of other series these studies in the next several years.
val scans or other work-up. Note that (20,21) have confirmed similar findings Although the available data are still in-
these recommendations refer to high-risk and have estimated the median tumor complete, certain tentative conclusions
subjects who have enrolled in a screening doubling times, assuming a constant can be drawn at the present:
program. growth rate to be in the 160 –180-day 1. Approximately half of all smokers
A further practical issue arises in the range. Authors of all of these reports, over 50 years of age have at least one lung
follow-up to detect interval growth of however, recognize wide variations, and nodule at the time of an initial screening
very small nodules. Comparison of cur- in one study 22% of tumors had a vol- examination. In addition, approximately
rent and previous scans can be performed ume doubling time of 465 days or more 10% of screening subjects develop a new
by using hard copy with manual mea- (21). Note that a 5-mm nodule with a nodule during a 1-year period (36).
surement, soft copy with electronic cali- doubling time of 60 days will reach a 2. The probability that a given nodule
Volume 237 Number 2 Fleischner Society Statement on CT of Small Pulmonary Nodules 397

4. is malignant increases according to its
size (4,5). Even in smokers, the percent- Recommendations for Follow-up and Management of Nodules Smaller than
8 mm Detected Incidentally at Nonscreening CT
age of all nodules smaller than 4 mm that
will eventually turn into lethal cancers is Nodule Size
very low ( 1%), whereas for those in the (mm)* Low-Risk Patient† High-Risk Patient‡
Radiology
8-mm range the percentage is approxi- 4 No follow-up needed §
Follow-up CT at 12 mo; if
mately 10%–20% (4,7,8,37). unchanged, no further follow-up
3. Cigarette smokers are at greater risk 4–6 Follow-up CT at 12 mo; if Initial follow-up CT at 6–12 mo then
for lethal cancers, and malignant nodules unchanged, no further follow-up at 18–24 mo if no change
6–8 Initial follow-up CT at 6–12 mo then Initial follow-up CT at 3–6 mo then
in smokers grow faster, on average, than at 18–24 mo if no change at 9–12 and 24 mo if no change
do those in nonsmokers (19,25,26). Also, 8 Follow-up CT at around 3, 9, and 24 Same as for low-risk patient
the cancer risk for smokers increases in mo, dynamic contrast-enhanced
proportion to the degree and duration of CT, PET, and/or biopsy
exposure to cigarette smoke (38). Note.—Newly detected indeterminate nodule in persons 35 years of age or older.
4. Certain features of nodules correlate * Average of length and width.
† Minimal or absent history of smoking and of other known risk factors.
with likelihood of malignancy, cell type, ‡ History of smoking or of other known risk factors.
and growth rate. For instance, small §
The risk of malignancy in this category ( 1%) is substantially less than that in a baseline CT scan
purely ground-glass opacity (nonsolid) of an asymptomatic smoker.
nodules that have malignant histopatho- Nonsolid (ground-glass) or partly solid nodules may require longer follow-up to exclude
logic features tend to grow very slowly, indolent adenocarcinoma.
with a mean volume doubling time on
the order of 2 years (19). Solid cancers, on
the other hand, tend to grow more rap-
idly, with a mean volume doubling time be subjected to a minimum of four or five this regard. Rather, we have elected to
on the order of 6 months. The growth follow-up CT examinations before being focus on the issue of follow-up imaging
rate of partly solid nodules tends to fall designated benign and the patient being of smaller nodules. Specifically, for what
between these extremes, and this partic- reassured (1,2). As summarized above, kinds of lesions is it appropriate to fol-
ular morphologic pattern is highly pre- data from ongoing CT screening pro- low, and if followed, at what intervals?
dictive of adenocarcinoma (39 – 41). grams with multidetector CT with 5-mm Therefore, we propose a set of guidelines,
5. Increasing patient age generally cor- collimation indicate that approximately summarized in the Table, for the man-
relates with increasing likelihood of ma- half of all smokers over 50 years of age agement of small pulmonary nodules de-
lignancy. Lung cancer is uncommon in have at least one lung nodule on the ini- tected on CT scans.
patients younger than 40 years and is rare tial scan (36). In addition, approximately Note that the recommendations
in those younger than 35 years (42). At 10% of screening subjects develop a new shown in the Table apply only to adult
the other end of the age scale, although nodule over a 1-year period, and about patients with nodules that are “inciden-
the likelihood of cancer increases, surgi- 12% can be expected to have one or more tal” in the sense that they are unrelated
cal intervention carries greater risks. Also, additional nodules that were missed on to known underlying disease. The follow-
the likelihood of a small nodule evolving the original scan (36). Assuming similar ing examples describe patients to whom
into a cancer that will cause premature demographics, approximately 20% of pa- the above guidelines would not apply.
death becomes a lesser concern as comor- tients who have a nodule detected on CT Patients known to have or suspected of
bidity increases in a person and predicted scans can be expected to have at least one having malignant disease.—Patients with a
survival decreases with advancing years. new nodule detected during the cur- cancer that may be a cause of lung me-
rently recommended 2-year minimum tastases should be cared for according to
follow-up period, which will in turn the relevant protocol or specific clinical
Management Approach
mandate another series of follow-up CT situation. Pertinent factors will include
A number of investigators have already studies with similar opportunities for the site, cell type, and stage of the pri-
raised serious concerns about current new nodules to be detected during the mary tumor and whether early detection
management strategies for the indeter- additional follow-up period. Therefore, of lung metastases will affect care. In this
minate small nodule, particularly after strict application of the existing recom- setting, frequent follow-up CT may be
encountering an overwhelming number mendations would result in multiple fol- indicated.
of such abnormalities on CT scans low-up studies over 2 or more years for a Young patients.—Primary lung cancer is
(7,14,36). In the study by Henschke et al large proportion of all patients who un- rare in persons under 35 years of age
(5) described earlier, the authors found dergo thoracic CT. ( 1% of all cases), and the risks from
no cancers in patients in whom the larg- In the case of nodules larger than 8 radiation exposure are greater than in the
est noncalcified nodule was less than 5 mm, additional options such as contrast older population. Therefore, unless there
mm in diameter on the initial scan (zero material– enhanced CT, positron emis- is a known primary cancer, multiple fol-
of 378 patients). Thus there was no ad- sion tomography (PET), percutaneous low-up CT studies for small incidentally
vantage in performing short-interval fol- needle biopsy, and thoracoscopic resec- detected nodules should be avoided in
low-up for nodules smaller than 5 mm in tion can be considered (43– 46). Because young patients. In such cases, a single
their study, even in high-risk patients. these approaches depend greatly on low-dose follow-up CT scan in 6 –12
Therefore, we recommend altering the available expertise and equipment and months should be considered.
existing recommendations, which indi- have limited applicability to nodules in Patients with unexplained fever.—In cer-
cate that every indeterminate nodule, re- the subcentimeter range, we have chosen tain clinical settings, such as a patient
gardless of size and morphology, should not to offer detailed recommendations in presenting with neutropenic fever, the
398 Radiology November 2005 MacMahon et al

5. presence of a nodule may indicate active conclusive evidence, as yet, that serial CT malignancy and should be managed ac-
infection, and short-term imaging fol- studies with early intervention for de- cordingly (39). Depending on the cir-
low-up or intervention may be appropri- tected cancers can reduce disease-specific cumstances, follow-up imaging studies or
ate. mortality, even in high-risk patients (7). intervention may be appropriate. Con-
Previous CT scans, chest radiographs, Therefore, we do not recommend fol- servative management is generally ap-
and other pertinent imaging studies low-up CT for every small indeterminate
Radiology
propriate for nodules in very elderly pa-
should be obtained for comparison nodule. tients or in those with major comorbid
whenever possible, as they may serve to Management decisions should not be disease. Interval growth of any nodule
demonstrate either stability or interval based on nodule size alone. While any suggests an active process, and further
growth of the nodule in question. calcification in a small nodule favors a evaluation or intervention should be
A low-dose, thin-section, unenhanced benign cause, central, laminar, or dense considered in such cases.
technique should be used, with limited diffuse patterns of calcification are reli- It is impossible to ignore medicolegal
longitudinal coverage, when follow-up of able evidence of benignancy (49). Fat considerations when discussing manage-
a lung nodule is the only indication for content suggests a hamartoma or occa- ment of pulmonary nodules. The current
the CT examination. sionally a lipoid granuloma or lipoma practice in the United States of recom-
(50). Solid versus nonsolid appearance, mending follow-up studies for all inde-
spiculation, or other characteristics influ- terminate opacities is partly related to
DISCUSSION ence the likelihood of malignancy and perceived liability if a cancer should de-
probable growth rate in any given case velop (59). When the medical commu-
Not every focal opacity qualifies as a nod- (19,39,49,51,52). Longer follow-up in- nity has preached the importance of
ule. Unfortunately, the familiar lung tervals are appropriate for nonsolid early detection of cancer for so long, it
nodule has eluded all efforts at precise (ground-glass opacity) and very small may prove difficult to convince physi-
definition. A committee of the Fleischner opacities (19,40). For instance, even if cians and the public that follow-up CT of
Society on CT nomenclature described malignant, a nonsolid nodule that is every nodule in every patient is unneces-
the pathologic definition of a nodule as a smaller than 6 mm will probably not sary. Nonetheless, it is our hope that the
“small, approximately spherical, circum- grow perceptibly in much less than 12 guidelines presented here will support a
scribed focus of abnormal tissue” and the months (19,40). Also, as discussed earlier, practical and medically appropriate ap-
radiologic definition as a “round opacity, it has been established that accurate mea- proach to the management of inciden-
at least moderately well marginated and surement of growth in subcentimeter tally detected small pulmonary nodules.
no greater than 3 cm in maximum diam- nodules is problematic (16). The recommendations presented here
eter” (47). Therefore, a linear or essen- Other features such as clustering of are based on our current understanding
tially two-dimensional opacity that does multiple nodules in a single location in of pulmonary nodules, and we expect
not have an approximately spherical the lung tend to favor an infectious pro- that they will continue to evolve as more
component is not a nodule. In general, cess, although a dominant nodule with information becomes available.
purely linear or sheetlike lung opacities adjacent small satellite nodules can be
are unlikely to represent neoplasms and seen in primary lung cancer (53,54). For a
do not require follow-up, even when the single nodule, upper lobe location in- References
maximum dimension exceeds 8 mm creases the likelihood of malignancy, be- 1. Ost D, Fein AM, Feinsilver SH. The solitary
(48). cause primary lung cancers are more pulmonary nodule. N Engl J Med 2003;
348:2535–2542.
Depending on their appearance and ra- common in the upper lobes (55). On the
2. Tan BB, Flaherty KR, Kazerooni EA, Iannet-
diologic context, certain nodular opaci- other hand, small, irregular, benign sub- toni MD; American College of Chest Phy-
ties may be judged sufficiently typical of pleural opacities, presumably due to scar- sicians. The solitary pulmonary nodule.
scarring that follow-up is not warranted. ring, are extremely common in the apical Chest 2003;123(suppl 1):89S–96S.
Because at least 99% of all nodules 4 areas in older patients, whereas triangu- 3. Swensen SJ, Silverstein MD, Ilstrup DM,
Schleck CD, Edell ES. The probability of
mm or smaller are benign and because lar or ovoid circumscribed nodules 3–9 malignancy in solitary pulmonary nod-
such small opacities are extremely com- mm in diameter adjacent to pleural fis- ules: application to small radiologically in-
mon on thin-section CT scans, we do not sures commonly represent intrapulmo- determinate nodules. Arch Intern Med
recommend follow-up CT in every such nary lymph nodes (56). A history of can- 1997;157:849 – 855.
4. Midthun DE, Swensen SJ, Jett JR, Hartman
case; in selected cases with suspicious cer can greatly increase the likelihood of TE. Evaluation of nodules detected by
morphology or in high-risk subjects, a a nodule being malignant, depending on screening for lung cancer with low dose
single follow-up scan in 12 months the nature and stage of the primary neo- spiral computed tomography. Lung Can-
should be considered. We accept that this plasm (57). cer 2003;41(suppl 2):S40.
5. Henschke CI, Yankelevitz DF, Naidich DP,
protocol could result in a few indolent The patient’s age and the presence of
et al. CT screening for lung cancer: suspi-
cancers being missed, but we believe that comorbid conditions should influence ciousness of nodules according to size on
the number of such instances would be management recommendations (58). For baseline scans. Radiology 2004;231(1):164 –
extremely small relative to the reduction instance, an incidentally detected 5-mm 168.
in the number of unnecessary studies. indeterminate nodule in an 85-year-old 6. Aberle DR, Gamsu G, Henschke CI,
Naidich DP, Swensen SJ. A consensus state-
Given the high prevalence of small be- patient with comorbid conditions is rel- ment of the Society of Thoracic Radiology:
nign nodules, the requirement to follow atively unlikely to develop into a symp- screening for lung cancer with helical
every such case for a period of 2 years tomatic lung cancer during the patient’s computed tomography. J Thorac Imaging
requires vast resources. The radiation lifetime and, therefore, does not neces- 2001;16(1):65– 68.
7. Swensen SJ, Jett JR, Hartman T, et al.
burden for the affected population is also sarily require follow-up (5). However, Screening for lung cancer with CT: Mayo
substantial, and this could be a cause of noncalcified nodules larger than 8 mm Clinic experience. Radiology 2003;226(3):
cancer in itself (10). Finally, there is no diameter can bear a substantial risk of 756 –761.
Volume 237 Number 2 Fleischner Society Statement on CT of Small Pulmonary Nodules 399

6. 8. Henschke CI, Naidich DP, Yankelevitz DF, and Human Services, Centers for Disease 43. Swensen SJ, Viggiano RW, Midthun DE, et
et al. Early Lung Cancer Action Project: Control and Prevention, National Center al. Lung nodule enhancement at CT: mul-
initial findings on repeat screening. Can- for Chronic Disease Prevention and ticenter study. Radiology 2000;214:73– 80.
cer 2001;92:153–159. Health Promotion, Office on Smoking 44. Rohren EM, Turkington TG, Coleman RE.
9. Henschke CI, McCauley DI, Yankelevitz D, and Health, 1982. Clinical applications of PET in oncology.
et al. Early Lung Cancer Action Project: 26. Guyatt GH, Newhouse MD. Are active and Radiology 2004;231(2):305–332.
overall design and findings from baseline passive smoking harmful? determination 45. Yamagami T, Iida S, Kato T, Tanaka O,
Radiology
screening. Lancet 1999;354:99 –105. of causation. Chest 1985;88(3):445– 451. Nishimura T. Combining fine-needle aspi-
10. Brenner DJ. Radiation risks potentially as- 27. Risch HA, Howe G, Jain M, Burch JD, Ho- ration and core biopsy under CT fluoros-
sociated with low-dose CT screening of lowaty EJ, Miller AB. Are female smokers at copy guidance: a better way to treat pa-
adult smokers for lung cancer. Radiology higher risk for lung cancer than male tients with lung nodules? AJR Am J Roent-
2004;231(2):440 – 444. smokers? a case-control analysis by histo- genol 2003;180:811– 815.
11. Mayo JR, Alrich J, Muller NL. Radiation
¨ logic type. Am J Epidemiol 1993;138(5): 46. Suzuki K, Nagai K, Yoshida J, et al. Video-
exposure at chest CT: a statement of the 281–293. assisted thoracoscopic surgery for small in-
Fleischner Society. Radiology 2003;228(1): 28. Blot WJ, McLaughlin JK. Are women more determinate pulmonary nodules. Chest
15–21. susceptible to lung cancer? J Natl Cancer 1999;115:563–568.
12. Imhof H, Schibany N, Ba-Ssalamah A, et al. Inst 2004;96(11):812– 813. 47. Austin JH, Muller N, Friedman PJ, et al.
¨
Spiral CT and radiation dose. Eur J Radiol 29. Zang EA, Wynder EL. Differences in lung Glossary of terms for CT of the lungs: rec-
2003;47(1):29 –37. cancer risk between men and women: ex- ommendations of the nomenclature com-
13. Kalra MK, Maher MM, Toth TL, et al. Strat- amination of the evidence. J Natl Cancer mittee of the Fleischner Society. Radiology
egies for CT radiation dose optimization. Inst 1996;88:183–192. 1996;200:327–331.
Radiology 2004;230(3):619 – 628. 30. Bain C, Feskanich D, Speizer FE, et al. Lung 48. Takashima S, Sone S, Li F, et al. Small sol-
14. Benjamin MS, Drucker EA, McLoud T, cancer rates in men and women with com- itary pulmonary nodules ( 1 cm) detected
Shepard JO. Small pulmonary nodules: de- parable histories of smoking. J Natl Cancer at population-based CT screening for lung
tection at chest CT and outcome. Radiol- Inst 2004;96:826 – 834. cancer: reliable high-resolution CT fea-
ogy 2003;226(2):489 – 493. 31. Mayne ST, Buenconsejo J, Janerich D. Fa- tures of benign lesions. AJR Am J Roentge-
15. Revel MP, Lefort C, Bissery A, et al. Pulmo- milial cancer history and lung cancer risk nol 2003;180:955–964.
nary nodules: preliminary experience with in United States nonsmoking men and 49. Erasmus JJ, Connolly JE, McAdams HP,
three-dimensional evaluation. Radiology women. Cancer Epidemiol Biomarkers Roggli VL. Solitary pulmonary nodules. I.
2004;231(2):459 – 466. Prev 1999;8:1065–1069. Morphologic evaluation for differentia-
16. Revel MP, Bissery A, Bienvenu M, Aycard 32. Bailey-Wilson JE, Amos CI, Pinney SM, et tion of benign and malignant lesions. Ra-
L, Lefort C, Frija G. Are two-dimensional al. A major lung cancer susceptibility locus dioGraphics 2000;20:43–58.
CT measurements of small noncalcified maps to chromosome 6q2325. Am J Hum 50. Gaerte SC, Meyer CA, Winer-Muram HT,
pulmonary nodules reliable? Radiology Genet 2004;75:460 – 474.
Tarver RD, Conces DJ. Fat-containing le-
2004;231(2):453– 458. sions of the chest. RadioGraphics 2002;
33. Lee PN. Relation between exposure to as-
17. Yankelevitz DF, Gupta R, Zhao B, Hen- 22(Spec Issue):S61–S78.
bestos and smoking jointly and the risk of
schke CI. Small pulmonary nodules: eval- 51. Wang JC, Sone S, Feng L, et al. Rapidly grow-
lung cancer. Occup Environ Med 2001;58:
uation with repeat CT—preliminary expe- ing small peripheral lung cancers detected
145–153.
rience. Radiology 1999;212:561–566. by screening CT: correlation between radio-
34. Gottlieb LS, Husen LA. Lung cancer among
18. Kostis W, Yankelevitz DF, Reeves AP, Flu- logical appearance and pathological fea-
Navajo uranium miners. Chest 1982;81:
ture SC, Henschke CI. Small pulmonary tures. Br J Radiol 2000;73:930 –937.
449 – 452.
nodules: reproducibility of three-dimen- 52. Takashima S, Sone S, Li F, Maruyama Y,
35. Field RW, Steck DJ, Smith BJ, et al. Resi-
sional volumetric measurement and esti- Hasegawa M, Kadoya M. Indeterminate
mation of time to follow-up CT. Radiology dential radon gas exposure and lung can- solitary pulmonary nodules revealed at
2004;231(2):446 – 452. cer: the Iowa Radon Lung Cancer Study. populations-based CT screening of the
19. Hasegawa M, Sone S, Takashima S, et al. Am J Epidemiol 2000;151(11):1091–1102. lung: using first follow-up diagnostic CT to
Growth rate of small lung cancers detected 36. Swensen SJ. CT. Screening for lung cancer. differentiate benign and malignant le-
on mass CT screening. Br J Radiol 2000;73: AJR Am J Roentgenol 2002;179:833– 836. sions. AJR Am J Roentgenol 2003;180:
1252–1259. 37. Pastorino U, Bellomi M, Landoni C, et al. 1255–1263.
20. Usuda K, Sato Y, Sagawa M, et al. Tumor Early lung-cancer detection with spiral CT 53. Heitzman ER, Markarian B, Raasch BN, et al.
doubling time and prognostic assessment and positron emission tomography in Pathways of tumor spread through the lung:
of patients with primary lung cancer. Can- heavy smokers: 2-year results. Lancet 2003; radiologic correlations with anatomy and
cer 1994;74(8):2239 –2244. 362:593–597. pathology. Radiology 1982;144(1):3–14.
21. Winer-Muram HT, Jennings SG, Tarver 38. Bach PB, Kattan M, Thornquist MD, et al. 54. Yano M, Arai T, Inagaki K, Morita T, No-
RD, et al. Volumetric growth rate of stage Variations in lung cancer risk among mura T, Ito H. Intrapulmonary satellite
I lung cancer prior to treatment: serial smokers. J Natl Cancer Inst 2003;95(6): nodule of lung cancer as a T factor. Chest
CT scanning. Radiology 2002;223(3):798 – 470 – 478. 1998;114:1305–1308.
805. 39. Henschke CI, Yankelevitz DF, Mirtcheva R, 55. Byers TE, Vena JE, Rzepka TF. Predilection
22. Sone S, Li F, Yang ZG, et al. Results of et al. CT screening for lung cancer: fre- of lung cancer for the upper lobes: an ep-
three-year mass screening programme for quency and significance of part-solid and idemiologic inquiry. J Natl Cancer Inst
lung cancer using mobile low-dose spiral nonsolid nodules. AJR Am J Roentgenol 1984;72:1271–1275.
computed tomography scanner. Br J Can- 2002;178:1053–1057. 56. Hyodo T, Kanazawa S, Dendo S, et al. In-
cer 2001;84(1):25–32. 40. Aoki T, Nakata H, Watanabe H, et al. Evo- trapulmonary lymph nodes: thin-section
23. Cancer statistics in Japan ’99: index of ta- lution of peripheral lung adenocarcino- CT findings, pathological findings, and CT
bles—standardized cancer mortality ratio mas: CT findings correlated with histology differential diagnosis from pulmonary
among daily smokers in Japan (cohort and tumor doubling time. AJR Am J Roent- metastatic nodules. Acta Med Okayama
study 1966-1982). National Cancer Cen- genol 2000;174:763–768. 2004;58(5):235–240.
ter Web site. http://www.ncc.go.jp/en 41. Li F, Sone S, Abe H, MacMahon H, Doi K. 57. Quint LE, Park CH, Iannettoni MD. Solitary
/statistics/1999/tables.html. Updated De- Comparison of thin section CT findings in pulmonary nodules in patients with ex-
cember 1, 2004. malignant and benign nodules in CT trapulmonary neoplasms. Radiology 2000;
24. Hirayama T. Non-smoking wives of heavy screening for lung cancer. Radiology 2004; 217:257–261.
smokers have a higher risk of lung cancer: 233(3):793–798. 58. Read WL, Tierney RM, Page NC, et al. Dif-
a study from Japan. Br Med J (Clin Res Ed) 42. Gadgeel SM, Ramalingam S, Cummings G, ferential prognostic impact of comorbid-
1981;282(6259):183–185. Kraut MJ, Wozniak AJ, Gaspar LE. Lung ity. J Clin Oncol 2004;22:3099 –3103.
25. U.S. Department of Health and Human cancer in patients 50 years of age: the 59. Berlin L. Malpractice issues in radiology:
Services. The health consequences of experience of an academic multidisci- failure to diagnose lung cancer—anatomy
smoking: a report of the surgeon general. plinary program. Chest 1999;115(5):1232– of a malpractice trial. AJR Am J Roentgenol
Atlanta, Ga: U.S. Department of Health 1236. 2003;180:37– 45.
400 Radiology November 2005 MacMahon et al