Subsolid Pulmonary Nodules and the Spectrum of Peripheral

Subsolid Pulmonary Nodulesand the Spectrum of PeripheralAdenocarcinomas of the Lung:Recommended Interim Guidelines forAssessment and Management1

Myrna C. B. Godoy, MDDavid P. Naidich, MD Pulmonary nodule characterization is currently being re-

defined as new clinical, radiologic, and pathologic data arereported, necessitating a reevaluation of the clinical man-agement, especially of subsolid nodules. These are nowknown to frequently, although not invariably, fall into thespectrum of peripheral adenocarcinomas of the lung.Strong correlation between the Noguchi histologic classifi-cation and computed tomographic (CT) appearances ofthese lesions, in particular, has been reported. Serial CTfindings have further documented that stepwise progres-sion of lesions with ground-glass opacity, manifested as anincrease in size or the appearance and/or subsequent in-crease of solid components, does occur in a select subset ofpatients. As a consequence, recognition of the potentialassociation between subsolid nodules and peripheral ade-nocarcinomas requires a review of current guidelines forthe management of these lesions, further necessitated by adifferential diagnosis that includes benign lesions such asfocal inflammation, focal fibrosis, and organizing pneumo-nia. Specific issues that need to be addressed are the needfor consensus regarding an appropriate CT classification,methods for precise measurement of subsolid nodules,including the extent of both ground-glass and solid compo-nents, as well as accurate assessment of the growth ratesas means for predicting malignancy and prognosis. It isanticipated that interim guidelines may serve to standard-ize our current management of these lesions, pending fur-ther clarification of their natural history.

� RSNA, 2009

1 From the Department of Radiology, New York University-Langone Medical Center, 560 First Ave, IRM 236, NewYork, NY 10016. Received January 29, 2009; revisionrequested March 16; revision received May 27; acceptedJune 4. Address correspondence to D.P.N. (e-mail:[email protected] ).

� RSNA, 2009

REVI

EWS

AND

COM

MEN

TARY

�ST

ATE

OFTH

EAR

T

606 radiology.rsna.org ▪ Radiology: Volume 253: Number 3—December 2009

Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

Recent advances in technology, in-cluding widespread availability ofmultidetector computed tomo-

graphic (CT) scanners associated with anabundance of new information obtainedespecially from low-dose CT lung cancerscreening programs, have increased ourunderstanding of the varieties of small pe-ripheral lung nodules encountered in dailyclinical practice, in particular, the impor-tance and prevalence of subsolid pulmo-nary nodules (1,2).

Subsolid nodules are now known tofrequently represent the histologic spec-trum of peripheral adenocarcinomas.

Pending potential revisions in pathologicclassification of these lesions, this in-cludes premalignant atypical adenoma-tous hyperplasia (AAH), bronchioloalveo-lar carcinoma (BAC), and mixed subtypeadenocarcinoma (3).

In this review we focus on clinical,radiologic, and pathologic aspects ofsubsolid pulmonary nodules, with theintention to propose new interim man-agement guidelines.

Epidemiology of Adenocarcinomaof the Lung

Lung cancer continues to be the lead-ing cause of cancer-related death inboth men and women in the UnitedStates (4). Adenocarcinoma is themost common type of lung cancer,representing 30%–35% of all primarylung tumors, and its incidence has in-creased over the past few decades(5,6). Included in the spectrum of ad-enocarcinomas of the lung (3,7), BACconsists of a peripheral tumor withdistinct epidemiology and clinicopath-ologic and radiologic features, ac-counting for 2%–6% of all non–smallcell lung cancers (8). When BAC isadded to mixed subtype adenocarci-noma with a BAC component, theircombined incidence reaches up tonearly 20% of all lung cancers (5,6,8).

BAC presents a unique growth pat-tern along alveolar septa without stro-mal invasion and has an indolentcourse (8). It is less commonly associ-ated with smoking, when comparedwith other non–small cell lung cancers(although smokers are at increasedrisk of all types of lung cancer, regard-less of histology), has a higher inci-dence in women, and has a youngerage distribution (8–10). The presenceof other pulmonary diseases, such asfibrotic disorders, increases the risk ofdeveloping BAC (8). BAC also con-trasts with other forms of lung cancerby exhibiting a relatively high inci-dence of multifocality (25% vs 5%)(5). More recently, it has also beenshown that adenocarcinomas, includ-ing those with BAC features, dispro-portionately respond to treatmentwith tyrosine kinase inhibitors (11).

The incidence of BAC is higher inJapan than in other parts of the world,including the United States or Europe(3). In the United States, there is evi-dence that the incidence of BAC hasincreased four-fold from 1955 to 1990,with much of the increase in BAC oc-curring in women (5).

Histopathologic Classification ofAdenocarcinoma of the Lung

In 1995, Noguchi et al (12) proposed ahistologic classification for small pe-ripheral adenocarcinomas, includingsix subtypes, on the basis of the pat-terns of tumor growth. Type A corre-sponds to localized BAC; type B, local-ized BAC with foci of collapsed alveo-lar structures; type C, localized BACwith foci of active fibroblastic prolifer-ation; type D, poorly differentiatedadenocarcinoma; type E, tubular ade-nocarcinoma; and type F, papillary ad-enocarcinoma with evidence of com-pressive and destructive growth.Types A, B, and C represent a distinctgrouping, as they show in common apredominant growth pattern involving“replacement” of alveolar lining cells.Type C, although denominated BACaccording to prior criteria, is differen-tiated from types A and B by the pres-ence of foci of active fibroblastic pro-liferation, indicating a more advancedform of the disease, with stromal in-vasion (12). In distinction, types Dthrough F represent “non-replace-ment” invasive adenocarcinomas and,as initially proposed by Noguchi et al,likely arise de novo, distinct fromtypes A through C. In this seminalstudy, types A, B, and C representedthe most common forms of adenocar-cinoma (74%).

Published online before print10.1148/radiol.2533090179

Radiology 2009; 253:606–622

Abbreviations:AAH � atypical adenomatous hyperplasiaBAC � bronchioloalveolar carcinomaFDG � fluorine 18 fluorodeoxyglucoseGGO � ground-glass opacity

Authors stated no financial relationship to disclose.

Essentials

� Subsolid nodules are now knownto frequently represent the histo-logic spectrum of peripheral ade-nocarcinomas, including premalig-nant atypical adenomatous hyper-plasia (AAH), bronchioloalveolarcarcinoma (BAC), and mixed sub-type adenocarcinoma; on the ba-sis of current knowledge, newguidelines are proposed for fol-low-up and management of sub-solid nodules on CT scans.

� Isolated lesions with pure ground-glass opacity (GGO) that are lessthan 5 mm in size do not neces-sarily require follow-up CT stud-ies since they nearly always repre-sent foci of AAH; for those be-tween 5 and 10 mm, follow-up isrequisite pending better definitionof their true nature.

� Nodules with pure GGO that arelarger than 1 cm in size should beassumed as BAC or invasive adeno-carcinoma provided persistence forat least 3 months, although 20%-25% will prove to be benign at re-section; surgery should be consid-ered especially if the nodule is en-larging or if there is an increase inattenuation or development of asolid component.

� Lesions with mixed solid compo-nent and GGO should also be pre-sumed malignant and surgical re-section should be considered,again provided lack of intervalchange over at least 3 months.

STATE OF THE ART: Subsolid Pulmonary Nodules and Peripheral Adenocarcinomas of the Lung Godoy and Naidich

Radiology: Volume 253: Number 3—December 2009 ▪ radiology.rsna.org 607

Since Noguchi et al first proposedthis classification, substantial changeshave been made in the World HealthOrganization subclassification of lungadenocarcinomas, reflecting betterunderstanding of the histopathologicfeatures of this disease (3). A majorchange made in the 1999 classificationand maintained in the 2004 classifica-tion was the addition of AAH (Fig 1) asa premalignant lesion (3). As docu-mented at histologic evaluation of re-sected lung cancers, these lesionshave been identified accompanying ad-enocarcinomas in up to 60% of cases(13–15).

Three subtypes of BAC are nowrecognized: mucinous (Fig 2a), non-mucinous (Fig 2b), and mixed muci-nous and nonmucinous, or indetermi-nate. Importantly, in accordance withthe 2004 World Health Organizationclassification, to be classified as BAC,tumors must show pure lepidicgrowth, defined as growth of neoplas-tic cells along pre-existing structures,without stromal, pleural, or vascularinvasion (3,16). As defined, this cor-responds to Noguchi type A and B le-sions only. In terms of distinction, anytumor that manifests as BAC with aninvasive component (Fig 3) is termedadenocarcinoma, mixed subtype (16),corresponding to Noguchi type C le-sions. It is now documented that mostadenocarcinomas are heterogeneousand consist of more than one subtype,with mixed subtype adenocarcinomabeing most frequently diagnosed (3).

Peripheral Adenocarcinoma:CT-Pathologic Correlations

Currently, pulmonary nodules arecharacterized at CT as either solid orsubsolid. Solid nodules are defined asthose that completely obscure the lungparenchyma. In distinction, subsolidnodules include both pure ground-glass opacities (GGOs) and mixedsolid component and GGO. GGOs aredefined as focal nodular areas of in-creased lung attenuation throughwhich normal parenchymal structuressuch as airways, vessels, and interlob-ular septa can be defined. Nodules

with mixed solid component andGGO, or part-solid nodules, by defini-tion present both ground-glass andsolid components (1).

To date, most studies correlatingpathologic with CT findings in patientswith peripheral adenocarcinomas ofthe lung have focused on the Noguchiclassification, for which excellent al-though imperfect correlations havebeen well documented (Fig 4) (17–20); hence, the rationale for maintain-ing familiarity with this classification(Table). Following this classification,foci of AAH appear as lesions withGGO that typically measure less than5 mm in size, although these lesionsmay be as large as 1–2 cm (13,14). Inturn, Noguchi type A and B lesions,which demonstrate purely lepidicgrowth pattern, typically manifest asnodules greater than 5 mm in diame-ter with pure GGO. In distinction,Noguchi type B lesions showing evi-dence of structural collapse may alsomanifest as predominantly subsolid le-sions with GGO with a small solidcomponent, while Noguchi type C le-sions showing fibroblastic prolifera-tion with stromal invasion correlatewith lesions with mixed solid compo-nent and GGO, with still more exten-sive solid components than seen inNoguchi type B lesions (9,17). As willbe discussed later, progression of le-sions from those with pure GGO to

those with mixed solid component andGGO has been shown to occur in se-lect cases, correlating as predicted tostepwise progression of Noguchi re-placement-type adenocarcinomas(21,22).

Unlike Noguchi types A, B, and Clesions, Noguchi types D, E, and Flesions histologically correspond topurely invasive adenocarcinomas andtypically manifest either as solid ornear completely solid pulmonary nod-

Figure 1

Figure 1: Photomicrograph demonstratesAAH. Alveolar wall is slightly thickened andlined by a single layer of cuboidal cells withmild to moderate nuclear atypia. It is apparentwhy these lesions present as subtle, often diffi-cult to identify, small GGOs. (Hematoxylin-eosin stain; original magnification, �400.)

Figure 2

Figure 2: (a) Photomicrograph demonstrates BAC, mucinous type. Columnar cells show lepidic growth pattern,with tumor cells mantling pre-existing alveolar septa. These cells contain abundant cytoplasmic mucin. No stromalinvasion is seen. (Hematoxylin-eosin stain; original magnification, �200.) (b) Photomicrograph demonstrates BAC,nonmucinous type. Alveolar wall is thickened and lined by a single layer of atypical cuboidal to columnar cells. Nostromal invasion is seen. (Hematoxylin-eosin stain; original magnification, �200.)



ules that range from well-defined to ir-regular, spiculated lesions at CT (17).

Yang et al (18) in a seminal studycorrelated Noguchi types A, B, C, and Dlesions with four thin-section CT pat-terns: pure GGOs; heterogeneous GGOs,a subset of lesions with pure GGO inwhich either reticular opacities or air al-veolograms or bronchiolograms can beidentified (Fig 5); mixed nodules with acentral solid component and a ground-glass halo (Fig 6); and solid nodules, re-spectively. On the basis of this CT classi-fication, pure GGOs could be categorizedas type A in 94% of cases, heterogeneousGGOs proved to represent type B lesionsin 71% of cases. Type C lesions proved tobe those with mixed solid component andGGO in 29% of cases and solid nodules in50%, while all type D lesions proved to besolid nodules. Not unexpectedly, as mightbe predicted on the basis of the underly-ing histologic features, among theso-called replacement-pattern lesions(Noguchi types A to C), type C lesionswere substantially larger and had higherattenuation than type A and B lesions,with less extensive ground-glass compo-nent and absent air alveolograms or bron-chiolograms (18).

Despite reported close correlations,however, it is apparent that considerableoverlap exists between CT appearances(24,25), with differentiation betweenAAH, especially when lesions are greaterthan 5 mm in size, BAC, and minimally

invasive adenocarcinoma especially prob-lematic.

Oda et al (26), for example, evaluated52 nonsolid nodules that included 35 BACsand 17 AAHs. According to multivariateanalysis, nodular sphericity was shown tobe substantially more often associated withAAH, whereas internal air bronchogramsproved to be highly suggestive of BAC. Al-though AAH has been reported character-istically to be smaller than 10 mm in size(19), in this study, size was shown to be aninaccurate predictor differentiating AAHfrom BAC, as 47% of AAHs measuredmore than 10 mm and 14% of BAC lesionsmeasured less than 10 mm (26).

Perhaps most important, it has alsobeen reported that although nonsolidnodules with pure GGO are likely to rep-resent either AAH or BAC, they mayrarely represent invasive adenocarcino-mas (Fig 7). Nakata et al (19), for exam-ple, have reported that 7% of nonsolidnodules measuring less than 1 cm provedto be mixed subtype adenocarcinomas. Itis also worth emphasizing that the distri-bution of histologic subtypes and CT ap-pearances varies when solitary lesionsare compared with multiple lesions, al-though the implications of these differ-ences remain unclear. Kim et al (27), forexample, in a retrospective study compar-ing multiple (n � 105) with solitary (n �31) subsolid nodules found that both AAH(P � .001) and BAC (P � .029) provedsignificantly more likely to be present in

patients with multiple lesions, whereas ad-enocarcinomas proved more frequent insolitary subsolid nodules (P � .001).

Given the differences among these re-ports, not surprisingly, additional ap-proaches emphasizing use of quantitativedensitometric methodologies to differen-tiate between AAH, BAC, and mixed typeadenocarcinoma have been proposed(28,29). Although suggestive, these find-ings require further validation prior toroutine acceptance.

CT-Pathologic Correlations: Prognosis

Equally important to the observation of aclose correspondence between CT ap-pearances and the Noguchi classificationis the demonstration of close correlationwith prognosis. As documented by Nogu-chi, type A and B lesions are rarely asso-ciated with lymph node metastasis or vas-cular or pleural invasion and demonstratelow mitotic rates, resulting in 5-year sur-vival rates of nearly 100% (12). Type Clesions typically show a higher rate oflymph node metastasis, vascular andpleural invasion, and a higher mitotic ratecorresponding to a lower 5-year survivalrate of 74.8% (P � .01) (12). In distinc-tion, types D to F adenocarcinomasproved to have a significantly less favor-able prognosis, resulting in a 5-year sur-vival of approximately 50%. More recenthistopathologic studies have further doc-umented these initial observations (30).

Further validation of the prognosticsignificance of the percentage of BAC andnon-BAC components of adenocarcino-mas has recently been shown, even whencompared with the presence of epidermalgrowth factor receptor gene mutations,which are known to be associated withBAC component and to correlate with tu-mor sensitivity to tyrosine kinase inhibi-tors (31,32). Kobayashi et al (32), in astudy that evaluated 127 patients withpathologic stage 1A adenocarcinomas ofthe lung (�20 mm), reported that tumorswith high percentage of non-BAC compo-nent are a sufficient risk factor for bothrecurrence and poor prognosis to benefitfrom postoperative adjuvant chemother-apy, despite absence of epidermal growthfactor receptor mutations.

Given these findings, it is not surpris-

Figure 3

Figure 3: Photomicrographs demonstrate mixed subtype adenocarcinoma. (a) BAC component. A single layerof cuboidal to columnar atypical cells grows along alveolar wall without stromal invasion. (Hematoxylin-eosin stain;original magnification, �200.) (b) Invasive acinar adenocarcinoma. Stromal invasion is associated with a substan-tial increase of nuclear atypia and fibroblastic proliferation. (Hematoxylin-eosin stain; original magnification, �200.)



ing that all studies to date have showngood correlation between CT featuresand prognosis in patients with peripheraladenocarcinomas. Most important, it hasbeen shown that the percentage of GGOwithin lesions at CT correlates with theBAC (lepidic) component of the lesions inpatients with histologically proved adeno-carcinomas (18,20,24,33,34).

In the largest study to date evaluatingthe relationship between CT findings andprognosis, Vazquez et al (35) evaluated338 patients with a diagnosis of adenocar-cinoma in a low-dose CT screening pro-gram and showed that the proportion ofBAC component within adenocarcinomasrepresented a positive prognostic factorand correlated with the proportion ofGGO at CT. In keeping with prior re-ports, as the percentage of BAC histolog-ically decreased from 100% to 0%, theproportion of GGO at CT similarly de-creased, while the number of cases show-ing pleural, angiolymphatic, and bron-chial invasion proportionately increased(35). They further reported that the 10-year Kaplan-Meier lung cancer-specificsurvival rate was 100% for patients withlesions that proved to have 90%–99%BAC component (similar to survival in pa-tients with 100%). Patients with lesionswith less than 90% BAC component had apoorer rate of survival (95%), but stillbetter than those without a BAC compo-nent (90%).

The prognostic significance of the highpercentage of BAC components in theselesions raises the question of the need for anew subclassification of adenocarcinoma,denominated “minimally invasive BAC” or“BAC with minimal invasion,” as a separatecategory from mixed type adenocarcinomawith BAC features. On the basis of thesedata, some have recommended that adeno-carcinomas with greater than 90% BACcomponent should be considered minimallyinvasive (35). Alternatively, it has also beensuggested that lesions confirmed as pureBAC should be reclassified as adenocarci-noma in situ. Although suggestive, to date,proposed new subclassifications of these le-sions have not yet been incorporated intothe World Health Organization classifica-tion. It should be also noted that currentstaging TNM criteria for lung cancer doesnot take into account the above-discussed

CT and/or histologic characteristics of BACand mixed subtype adenocarcinoma withBAC features. To date, both these lesionsare similarly staged as T1, although thenewly proposed staging system for non–small cell lung cancer does discriminate be-tween lesions smaller than 2 cm and thosebetween 2 and 3 cm in size (36).

Benign versus Malignant SubsolidNodules: CT Evaluation

Given the close correlation betweenpathologic and CT findings in patients

with documented peripheral adenocarci-nomas, the sensitivity and specificity ofsubsolid nodules identified at CT has alsobeen evaluated, albeit most extensively inlow-dose screening studies. To date, al-though most studies have confirmedclose, although imperfect, correlation be-tween the spectrum of adenocarcinomasand CT appearances, benign conditions,including organizing pneumonia, focal fi-brosis, and focal inflammation, may alsopresent as subsolid nodules (Fig 8)(9,37,38). Unfortunately, the value of CTin distinguishing benign from malignant

Figure 4

Figure 4: Illustration of the relationship between the Noguchi histologic classification of adenocarcinomaof the lung (Noguchi types A though F) and corresponding CT appearances of these lesions. As denoted by thelarge arrow on the right, there is also good correlation between CT appearances and worsening prognosisprogressing through the Noguchi classification A to F.



subsolid nodules, to date, has been re-ported to vary considerably (2,37–39).

As reported by Henschke et al (2), ina population of high-risk individualsscreened with low-dose CT, while therate of malignancy proved greater forpart-solid (63%) than nonsolid (pureGGO) (18%) nodules, only 34% of sub-solid nodules proved to be malignantcompared with 7% for solid nodules. Sim-ilar findings have been reported by Kim etal (38), who retrospectively found in a

nonscreened population that while 81%of persistent nonsolid nodules proved tobe either AAH, BAC, or adenocarcinomawith BAC features, the remaining 19%proved histologically to represent eitherorganizing pneumonia or nonspecific fi-brosis (Fig 9). More disturbingly, in thisstudy, there were no differences betweenbenign and malignant lesions when as-sessed by shape, marginal characteris-tics, or the presence of pleural tags.Polygonal shape, nodule lobulation, or

presence of irregular margins or spicula-tion was found in both malignant andbenign lesions. Irregular margins, inparticular, proved nondiagnostic,caused by interstitial fibrosis or infil-trative tumor growth in malignant le-sions and granulation tissue in benignlesions, respectively.

In distinction, Lee et al (37), in arecent study evaluating 80 subsolidnodules (47 malignant and 33 benign),showed that CT predictors of malig-nancy for GGOs included size greaterthan 8 mm (P � .04) and lobulatedborder (P � .01), with lobulation as-sociated with higher risk of malig-nancy (P � .02) for part-solid nodules.Most important, in this study all pureGGOs measuring 4 mm or less provedbenign.

Interestingly, while not all subsolidnodules prove to be malignant, theyrarely represent metastases, even inpatients with documented extratho-racic tumors. In one study of 59 patho-logically proved GGOs in 34 patientswith a history of extrapulmonary ma-lignancy, althought 28 (82.4%) pa-tients proved to have lesions (includ-ing 24 adenocarcinomas, 16 BACs, 14AAHs, four focal fibrotic foci, and oneinflammatory nodule), there were nocases of metastases despite the factthat primary sites included adenocar-cinomas of the breast, stomach, andcolon (40).

Current Concepts in the Diagnosis andManagement of Subsolid Nodules

Despite close correlation between CTfindings and the spectrum of periph-eral adenocarcinoma, the diagnosisand management of these lesions re-main problematic given that not allsubsolid nodules (and in particular le-sions with pure GGO) prove to be ma-lignant. Given limitations in sensitivityand specificity described above, alter-nate approaches to diagnosis andmanagement necessarily have been in-vestigated, including follow-up surveil-lance CT, alternate imaging studiesthat include positron emission tomog-raphy (PET) and/or combined PET/CT, and biopsy.

Figure 5

Figure 5: CT scans (1-mm section) of BAC (Noguchi type B lesion), heterogeneous GGO, show a nodulewith GGO with (a) superimposed reticulation and (b) air bronchiolograms.

Correlation between the Noguchi Classification of Adenocarcinoma and World HealthOrganization 2004 Classifications and CT Findings

Noguchi Type World Health Organization 2004 CT Features

. . . AAH GGOA, Localized BAC BAC (mucinous, nonmucinous and

mixed mucinous andnonmucinous or indeterminate)

GGO

B, Localized BAC withalveolar collapse

BAC (mucinous, nonmucinous andmixed mucinous andnonmucinous or indeterminate)

GGO with possible solidcomponent

C, Localized BAC with activefibroblastic proliferation

Adenocarcinoma, mixed subtype(with BAC component

Part-solid nodule with increase insolid component or solid nodule

D, Poorly differentiated Solid adenocarcinoma with mucin Part-solid nodule with greaterincrease in solid component orsolid nodule

E, Tubular Acinar adenocarcinoma Part-solid nodule with greaterincrease in solid component orsolid nodule

F, Papillary Papillary adenocarcinoma Part-solid nodule with greaterincrease in solid component orsolid nodule

Note.—Adapted, with permission, from reference 9.



Growth Rate of Small PeripheralAdenocarcinomas in Low-Dose CTScreening for Lung Cancer

The mean doubling time of adenocar-cinomas is longer than that of othertumor cell types, as shown in previouschest radiographic studies (41). Asimilar pattern has been identifiedwith CT (Figs 10, 11) (39,42). In onestudy (39) of peripheral lung cancersidentified at low-dose screening CT,when stratified according to tumorhistologic type, the volume doublingtime was higher for the spectrum ofadenocarcinomas (988 days � 470 forAAH, 567 days � 168 for BAC, and384 days � 212 for mixed subtype ad-enocarcinoma with BAC features)compared with peripheral squamouscell carcinomas (122 days � 68), im-portantly emphasizing that subsolidnodules tend to present considerablyslower growth rates compared withsolid lesions (39). The clear implicationof these data is that the previous conceptthat lack of growth over a 2-year fol-low-up indicates a benign etiology doesnot apply for subsolid nodules.

Methods for Measuring Interval Changein the Appearance of Focal Nodules

By allowing for the slower growthrates of subsolid nodules, in particu-lar, emphasis has been placed onmeans for obtaining accurate mea-surements at follow-up studies. Unfor-tunately, the accuracy of measure-ments of small lung nodules, in partic-ular, is subject to high interobserverand intraobserver variability (43,44).Although three-dimensional volumemeasurements have been advocatedby some, especially for evaluation ofsolid nodules, to date, there are lesscompelling data supporting the use ofthis approach for subsolid nodules(Fig 12) (45–48).

A simple change in the attenuationof lesions in itself may also be indicativeof substantial interval change, makingreliance on either unidimensional or bi-dimensional measurements potentiallystill less meaningful (Figs 13, 14) (49).Furthermore, as much as 5%–10% of

Figure 6

Figure 6: CT scan of mixed subtype adenocar-cinoma with BAC component. Retrospective targetreconstruction, 1-mm section, shows two part-solid nodules in the middle lobe with a centralsolid component and a ground-glass halo, whichtypically corresponds to Noguchi type C adeno-carcinoma. (Reprinted, with permission, fromreference 23.)

Figure 7

Figure 7: CT scan (1-mm section) of mixedsubtype adenocarcinoma with BAC component(Noguchi type C lesion) shows a nodule with pureGGO, demonstrating that although nonsolid nod-ules are likely to represent AAH or BAC, an inva-sive component may rarely be present as in thiscase.

Figure 8

Figure 8: Focal inflammation mimicking adenocarcinoma. (a) Magnified 1-mm CT section through theright upper lobe shows nodules with GGO initially diagnosed as probable BAC. (b) Follow-up CT scan ob-tained 3 months later shows near complete resolution of the lesion (arrow), now presumed to represent focalnonspecific inflammation. (Reprinted, with permission, from reference 23.)



small malignant nodules may actuallyappear smaller with short-term fol-low-up intervals (50–52).

In an attempt to improve accuracy,Kakinuma et al (53) compared fourmethods for measuring GGOs: lesionlength, a modified measurement oflength that takes into account thelength of a solid component, lesionarea and its vanishing ratio, and thepercentage of a lesion’s area that isnot seen at thin-section CT when com-paring images reconstructed with me-diastinal versus lung window settings.Of these, the vanishing ratio method,although subjective, proved the mostaccurate predictor of 5-year relapse-free survival (53). Although the advan-tage of such an approach seems logicalon the basis of enhanced visualization

of the solid component of lesions (Fig 15),this technique requires further valida-tion prior to routine acceptance.

Regardless of approach, optimalevaluation of subsolid nodules, espe-cially lesions with GGO, requires thatappropriate CT technique be used.This includes most importantly the useof thin (1–3-mm) sections (54), as wellas appropriate exposure factors (55).Given the likelihood of performing nu-merous follow-up CT studies, espe-cially for smaller lesions, all effortshould be made to minimize radiationexposure (56,57).

Role of PET

The important role of PET in the diag-nosis and management of lung cancer

is well established. For lesions assmall as 8–10 mm in size, fluorine 18fluorodeoxyglucose (FDG) PET is ac-curate in differentiating benign frommalignant lesions, with an overall sen-sitivity, specificity, and accuracy of96%, 88%, and 94%, respectively(58). FDG PET, however, has a lowersensitivity for small (�10 mm) orslow-growing lesions, such as carci-noid tumors and BAC (58–64). In arecent study, Tsunezuka et al (62) cor-related the effectiveness of FDG PETto characterize adenocarcinomas (�2cm) with Noguchi classification andfound that the false-negative rate fortype A lesions was 100%, that for typeB lesions was 80%, and that for type Clesions was 47%, while the true-positive rate for types D, E, and Flesions was 67%, 100%, and 86%, re-spectively (Fig 16). Similar resultshave been reported by Yap and col-leagues (64).

It has been suggested that character-istic lower FDG uptake in BACs, whenassociated with CT findings that includenodule attenuation and size, can be ofvalue in differentiating this tumor frommixed subtype adenocarcinomas withBAC features (65). Moreover, PET hasbeen shown to also correlate with prog-nosis. Malignant nodules that have lowFDG uptake are more likely to have anindolent nature and lack intratumorallymphatic vessel invasion and lymph nodemetastasis (66), while high FDG uptakein an adenocarcinoma correlates withpoorer survival (67). Despite these data,the role of FDG PET imaging for assessingpredominantly lesions with GGO remainsto be established.

Role of Transbronchial andTransthoracic Needle Biopsy forDiagnosis of BAC

Since a substantial percentage of sub-solid nodules prove to be benign, it isreasonable to question the role oftransbronchial needle aspiration bi-opsy to establish a diagnosis. A majorpotential limitation to transbronchialneedle aspiration biopsy is the factthat accurate differentiation betweenAAH, BAC, and mixed type adenocar-

Figure 9

Figure 9: Focal nonspecific interstitial pneumo-nia. (a, b) Thin-section CT scans at the level of the leftpulmonary artery and aortic arch, respectively, showthree foci of persistent GGO. (c) Histologic specimenshows thickening of the alveolar wall with chronicinflammatory infiltrates. No tumor was identified.(Hematoxylin-eosin stain.)



cinomas with bronchioloalveolar com-ponents may not be feasible on thebasis of limited cytologic or even his-tologic sampling.

The diagnostic yield of CT-guidedfine-needle aspiration biopsy for sub-solid nodules has been evaluated byShimizu et al (68), who found a diag-nostic yield in ground glass–dominantlesions (GGO ratio � 50%) and solid-dominant lesions (GGO ratio � 50%)of 51.2% and 75.6%, respectively.The overall diagnostic yield of CT-guided fine-needle aspiration biopsy,however, was only 64.6%, while forlesions smaller than 10 mm with pre-dominantly ground-glass appearance thediagnostic yield was as low as 35.2%.

In an attempt to improve diagnos-tic accuracy, it has been suggestedthat subsolid nodules should preferen-tially be evaluated with core needlebiopsy. Kim et al (69) evaluated theaccuracy of CT-guided core biopsy forthe diagnosis of subsolid lesions andfound that the overall concordancerate between core and surgical biop-sies in malignant and premalignant le-sions was 73%. However, in 28% ofcases diagnosed at surgical resectionas mixed subtype adenocarcinomawith BAC features, the core biopsyfailed to identify the area of invasion,incorrectly diagnosing the lesion asBAC. The authors concluded that,given the requirement for BAC toshow pure lepidic growth without in-vasion, correlation with thin-sectionCT images is necessary and that com-plete surgical removed is necessary toexclude invasion before a final diagno-sis of BAC can be determined (69).

In fact, difficulty in establishing adefinitive pathologic diagnosis even incases for which surgical biopsies areavailable has recently been reported.In a review of histologic diagnoses re-evaluated by a panel of expert pathol-ogists in a population of patients ac-crued in a low-dose CT lung cancerscreening study, 50 of 59 lesions ini-tially diagnosed in outside institutionsas BACs were revised to invasive ade-nocarcinoma, while an additional 10cases, all smaller than 5 mm in size,initially diagnosed as AAH in patients

with established lung cancers were re-interpreted as nine BACs and one in-vasive adenocarcinoma (15).

On the basis of these data, it is con-cluded that transbronchial needle aspira-tion biopsy ideally should only be per-formed in patients with subsolid nodulesat CT who are either nonsurgical candi-dates, surgical candidates for whom proofof malignancy is still considered neces-sary, or who present with multifocal dis-ease (70). In these cases, correlation oftransbronchial or transthoracic needle bi-opsy with CT results can be used for apresumptive diagnosis.

Surgical Resection of Small PeripheralAdenocarcinomas

In light of recent data documenting mark-edly improved 5-year survival of patientswith subsolid nodules, in particular BAC(77–74), despite previous consensus for theneed to perform lobectomy for stage 1 lungcancer (75), the potential role of limitedsurgical resections, including partial wedgeresections and segmentectomies, has comeunder renewed scrutiny (71–74,76–79).

Currently, most studies have fo-cused on limited surgical resectionsfor patients with BAC (71,72,74,78),

Figure 10

Figure 10: BAC. Sequential magnified 1-mm CT sections through the right upper lobe show minimal in-crease in size of a nodule with GGO over a 3-year period. The central area of higher attenuation represents avessel bifurcation and not a solid component, which was better characterized on sequential images.



while the value of limited resection forminimally invasive adenocarcinoma(24,74,76) and part-solid nodules withground-glass component greater than

50% have also been reported (24,76).In these studies, cases in which radicalsegmentectomies with hilar and medi-astinal lymph node dissections or sam-

pling only were performed, no postop-erative recurrences were identified,which is consistent with the high cur-ability rate of these cancers.

Figure 11

Figure 11: Sequential magnified 5-mm CT sections through the left upper lobe show two discrete GGOs ini-tially measuring 8 mm in size over a 3-year period. The nodule with GGO in the top row remained stable, whereasan initially similar-appearing nodule (bottom row) progressed, with increase in size and subsequent develop-ment of a solid component. Histologic analysis of the second lesion showed mixed subtype adenocarcinomacomposed of acinar adenocarcinoma (40%) and BAC (60%). There are no predictive CT features to aid in differ-entiating lesions likely to progress versus those that remain stable.



As documented in a recent report byVazquez et al (35), contrary to traditionalstaging predictions, cases of multiple,node-negative adenocarcinoma proved tohave the same excellent prognosis as sol-itary node-negative cases. In this study(35), the 10-year Kaplan-Meier survivalrate of 213 patients with pathologic node-negative solitary lung cancer was 97%(95% confidence interval: 95%, 100%)versus 100% for 44 patients with multiplenode-negative malignancy, with no signif-icant difference between these two rates(P � .29). These data suggests that mostof the small, node-negative multiple ade-nocarcinomas currently identified at CTmost likely represent multiple primarycancers rather than intrapulmonary me-tastasis.

Mun et al (73) have further confirmedthe benefits of limited resections forNoguchi types A to C lesions, including a70.3% overall 5-year disease-free survivalof patients with multifocal BAC. Similarly,Carreta et al (77) in a study of 26 patientswith multiple primary adenocarcinomasof the lung and a total of 52 tumors founda favorable result with surgical treatment.Additionally, they reported that sublobarresections, when feasible, provided ade-quate oncologic management (77).

Current Status and OngoingControversies in the Managementof Subsolid Lung Nodules

It is apparent from the preceding dis-cussions that pulmonary nodule charac-terization and indications for therapyare undergoing profound changes asnew clinical, radiologic, and pathologicdata are reported. Although advancedtechnology, in particular multidetectorCT, has provided important new in-sights, it is equally apparent that thesame technology has created an entirelynew set of problems requiring consider-ation.

Patients for Whom Follow-up Studies ArePrioritized in Place of Biopsy or SurgeryThe optimal time and especially dura-tion of follow-up imaging studies in pa-tients with subsolid nodules remain tobe determined (Fig 10). As notedabove, especially for lesions with pure

GGO, traditional 2-year follow-up pe-riods are insufficient to safely diag-nose benign disease.

The best method for measuring le-sions in cases in which follow-up im-aging studies are performed also re-quires continued evaluation. The needfor improved methods to accurately

measure lesions that are otherwisepoorly or irregularly marginated, inparticular, remains a clear priority(43,45,46,48,53). Perhaps most im-portant, the relationship betweenAAH, BAC, and invasive adenocarci-noma requires further clarification.Determining how often and how rap-

Figure 12

Figure 12: Automated three-dimensional segmentation. (a) Magnified 1-mm CT section through the rightupper lobe shows a nodule with GGO with indistinct margin. (b) Corresponding three-dimensional segmenta-tion provides automated estimation of its volume and cross-sectional dimensions in the x, y, and z planes, aswell as minimum and maximum diameters measured in off-axial planes.

Figure 13

Figure 13: Mixed subtype adenocarcinoma, progression of GGO to a nodule with mixed solid componentand GGO. (a) Magnified 1-mm CT section shows a discrete GGO (arrows). (b) Follow-up CT scan obtained 1year later shows clear progression of the disease, with the development of a central solid component, althoughthere is no appreciable enlargement of the lesion (arrows).



idly lesions with pure GGO enlarge orincrease in attenuation at present re-mains problematic when based solelyon CT morphologic grounds. It shouldbe noted that to date only limited dataare available that document the natureof disease progression in patients withsubsolid nodules (21,22). In one study(22) in which follow-up CT studies

were used to evaluate 48 subsolid le-sions diagnosed as either AAH orNoguchi types A-C adenocarcinomas,of those initially recognized as noduleswith pure GGO (56%), 75% subse-quently demonstrated an increase insize (Fig 10), with solid componentsfirst appearing in 17% (Fig 13) andsubsequently increasing in size in 23%

(Fig 14). While these data support theconcept proposed by Noguchi et al(12) that there is a stepwise progres-sion of replacement types A to C le-sions, these authors failed to provideany insight into any predictive CT fea-tures to aid in differentiating lesionslikely to progress (Fig 11).

Patients with Multiple Subsolid NodulesWhile the distribution of histologicsubtypes differs from that seen in sol-itary nodules, the implications for pa-tient treatment remain unclear. Forexample, treatment of patients withmultiple tiny (� 5 mm) GGOs presum-ably will differ from treatment of pa-tients with multiple, variable-sized le-sions, especially those with one or afew dominant lesions, including thoselarger than 10 mm (Fig 17) or thosethat are part-solid in appearance (Fig 18).In the latter case, the value of limitedsurgical resections in patients withmultiple lesions, one or a few of whichprove to be dominant either by size orattenuation, remains to be deter-mined.

Is There Overdiagnosis of Lung Cancerin Patients with Subsolid Nodules?Although a detailed evaluation of thiscomplex issue is outside the scope of thepresent report, this question does bearmentioning given the current intense in-terest toward the potential of low-doseCT lung cancer screening to decreaselung cancer mortality (15,80,81). Prelim-inary reports do seem to indicate that inselect cases identification of small sub-solid lesions at CT may in fact lead tooverdiagnosis and unnecessary treatment(82). Lindell et al (52), in one recentstudy, for example, evaluated 61 lungcancers and found 27% of them had amean volume doubling time longer than400 days, meeting criteria for overdiagno-sis according to those previously pro-posed (83). Furthermore, in this samestudy, the mean volume doubling timewas longer in women, raising concern ofeven higher likelihood of overdiagnosis inthis population (84).

Additional insight into the likelihoodof overdiagnosis in screened populationmay be inferred from a recent report by

Figure 14

Figure 14: Mixed subtype adenocarcinoma. (a) Magnified 1-mm CT section through the left lower lobeshows a nodule with mixed solid component and GGO. (b) Follow-up CT scan obtained 6 months later showsincrease in the extent of the solid component within the nodule.

Figure 15

Figure 15: Thin-section CT images of nodule with mixed solid component and GGO. The vanishing ratiorepresents the percentage of the area of the lesion that is not seen at thin-section CT when comparing imagesreconstructed with (a) lung versus (b) mediastinal window settings. The mediastinal window setting is used toenhance visualization of the delimitation of the solid component.



Carter et al (15), in which histologicallyproved lesions identified at 174 baseline(prevalence) studies were compared withthose from 37 annual (incidence) screen-ing studies. Of the baseline cancers, 142(82%) of 174 proved to be adenocarcino-mas, of which 84 (60%) proved to besubsolid lesions, nine proved to be BACs,and 75 proved to be mixed type adeno-carcinomas. Importantly, of these 84 sub-solid lesions, only 14 (17%) proved inva-sive at resection. In distinction, of 37 fol-low-up (incidence) cancers identified,only six (16%) proved to be subsolid le-sions versus 60% at baseline screenings(15). These data, while still inconclusive,are consistent with the notion of overdi-agnosis occurring in at least some per-centage of screening-identified lesions.

It is worth emphasizing that thesedata leave unanswered the question ofthe clinical significance of a histologic di-agnosis of “minimal invasion,” an issuestill unresolved with important implica-tions for clinical management. It may beanticipated that better understanding ofthe natural history of adenocarcinomasthat allows differentiation between indo-lent tumors from more aggressive lesionswill likely have to await the developmentof ancillary measures, including the use ofbiologic markers (84).

Suggested Guidelines in theManagement of Subsolid Nodules

Incidentally identified, isolated pureGGOs smaller than 5 mm in size repre-sent foci of AAH sufficiently often toobviate routine follow-up CT studies,especially in the elderly. Although it isacknowledged that some of these le-sions may prove to be BAC, the extremerarity of invasive adenocarcinomas inthis subgroup coupled with their ex-tremely prolonged doubling times sug-gest that there is no reason to undergoeither the added expense or radiationexposure necessary to follow these le-sions presumably over prolonged timeintervals measured in years. An excep-tion is patients enrolled in low-dose lungcancer screening programs for whichfollow-up imaging is presumably dic-tated by protocol.

Given that the uncertainty regarding

the above issues will likely persist for theforeseeable future, by using currentlyavailable data, the following interimguidelines are proposed. These do not dif-ferentiate between low- and high-riskgroup as per Fleischner criteria (85) dueto the increased incidence of adenocarci-nomas in younger and nonsmoking pa-tients. It cannot be overemphasized thatthese guidelines need to be interpreted inthe light of individual clinical history.

Lesions Smaller than 10 mm with Pure GGOIsolated lesions smaller than 5 mm in sizewith pure GGO represent foci of AAHsufficiently often to obviate follow-up CTstudies.

Conservative management of nodulesbetween 5 and 10 mm in size with pureGGO requires at least an initial follow-upexamination in 3–6 months to documentthat lesions have not resolved spontane-ously (or following antibiotic therapy). Al-though a small percentage of these lesionswill prove to be invasive adenocarcino-mas, the role of surgical biopsy in thesecases remains problematic with appropri-ate management, necessitating case bycase evaluation. For most of these le-sions, continued long-term follow-up islikely preferable to surgical resection.

When opted for, follow-up surveil-lance should extend for more than 2years. As the optimal duration of fol-

Figure 16

Figure 16: Subsolid nodule characterization with PET/CT. (a) Coronal 5-mm CT section shows multicen-tric BAC characterized by GGOs in the left upper and lower lobes (arrows), corresponding to Noguchi types Aand B, respectively. (b) Correlated coronal PET image shows no FDG uptake by the lesions (false-negativestudy). (c) Axial magnified 1-mm CT section in another patient shows mixed subtype adenocarcinoma (Nogu-chi type C lesion), characterized by mixed solid and ground-glass appearance. (d) Correlated coronal PET im-age shows FDG uptake (arrow) by the lesion in the left upper lobe (maximum standardized uptake value, 3.4).



low-up of lesions with pure GGO has yetto be determined, at least three consecu-tive annual studies is a minimum require-ment to document stability. The need forfurther follow-up examinations remainsspeculative: While the number of lesionsthat enlarge or increase in attenuationlater than 3 years is small, a sufficient

number evolve to justify extended surveil-lance. In our experience, it was for as longas 5 years, although it cannot be overem-phasized that the risk of subsequent de-velopment of cancer clearly needs to bebalanced against the risks of unnecessaryradiation exposure and especially surgicalintervention (86).

Accurate assessment of interval changeis best accomplished by comparing thin-section CT scans, allowing as precise com-parison as possible to minimize both inter-and intraobserver variability in measure-ments, as well as close monitoring of anychange in the attenuation of lesions, as ei-ther change should be interpreted as indic-ative of possible malignancy, in most casesnecessitating surgical resection.

PET or PET/CT scans are of question-able diagnostic value or may even be mis-leading in the majority of cases of lesionssmaller than 1 cm in diameter and shouldnot be routinely obtained, especially giventhe small likelihood of associated metastaticdisease rendering the potential use of PETfor staging less useful.

Transbronchial aspiration should bediscouraged as unlikely to provide sufficientdata to either accurately assess malignantpotential or establish a benign diagnosis.Instead, core needle biopsy should be pref-erentially performed, especially in cases forwhich surgery is contraindicated and histo-logic evaluation is deemed necessary.

In all cases in which follow-up surveil-lance is undertaken, subsequent CT studiesshould be performed with the lowest possi-ble exposure techniques to minimize radia-tion exposure in these patients. In our ex-

Figure 17

Figure 17: Variable-sized subsolid nodules. (a) Magnified 1-mm CT section through the right upper lobeshows multiple small lesions with GGO and one dominant larger nodule with GGO (arrow). (b) CT scan at4-year follow-up shows no substanital interval change (arrow) and the lesions were presumed to representAAH and BAC (dominant lesion).

Figure 18

Figure 18: Multiple subsolid nodules with variable size and appearance. (a, b) Sequential axial 1-mm CT sections through the mid thorax in the same patient showmultiple subsolid pulmonary nodules differing in size and attenuation located in the superior segment of the right and left lower lobes (arrows), including nodules withmixed solid component and GGO, nodules with GGO, and a solid nodule. In such cases, selective limited resection of the dominant lesions may be acceptable, contrary tostandard treatment of patients with multiple foci of BAC and/or adenocarcinoma.



perience, studies performed with as low as80 mAs (depending on body habitus) are ofsufficient quality to allow accurate follow-upassessment of these lesions.

Solitary Lesions 10 mm or Larger in Sizewith GGOAs a general rule, solitary lesions 10 mm orlarger in size with pure GGO should beresected, provided that persistence orgrowth of the lesion is again establishedover at least a 3–6-month period. In thissetting, indications for percutaneous needlebiopsy are still limited as the likelihood of adefinitive histologic diagnosis remains prob-lematic given substantial sampling error.Similarly, indications for PET or PET/CTremain doubtful as PET-negative studies donot exclude the possibility of invasive ade-nocarcinoma, while these lesions are alsostill unlikely to be associated with distantmetastases.

Lesions with Mixed Solid Componentand GGOSimilar to solitary lesions 10 mm or largerin size with GGO, any lesion with mixedsolid component and GGO, regardless ofsize, represents malignancy with sufficientlikelihood to warrant further evaluation.The evaluation should include the timelyperformance of PET or preferably PET/CT,since there is a greater likelihood that theselesions represent invasive tumors for whichpreoperative staging and assessment ofprognosis is warranted. Less clear is therole for transbronchial and/or transtho-racic biopsy in these cases, given the limitedvalue of accurate differentiation betweenBAC and invasive adenocarcinomas, espe-cially for lesions with less than 50% solidcomponents, and the likelihood that theselesions will be resected, regardless.

Multiple Subsolid NodulesIn cases of multiple lesions smaller than 5mm with pure GGO, at least 1-year fol-low-up surveillance CT study should beperformed on the premise that these pa-tients may be at greater risk than thegeneral population for developing cancer.However, continued long-term follow-upshould not be considered necessary.

In general, follow-up CT surveillanceis to be preferred in cases in which mul-tiple small (5–10-mm in size) lesions are

identified, as these most likely representeither multifocal AAH or in smokers, re-spiratory bronchiolitis.

In distinction, (a) surgical resectionshould be considered, especially in cases inwhich there are dominant lesions, definedas GGOs greater than 10 mm in size orlesions with mixed solid component andGGO; (b) PET or preferably PET/CTshould be performed following a similarlogic as outlined above for solitary lesionswith mixed solid component and GGO; and(c) limited lung-sparing resections may beconsidered as an option to routine lobec-tomy given the likelihood that at least someof the remaining lesions will continue togrow.

In conclusion, new appreciation of theimportance of subsolid nodules has led tothe need for a reappraisal of the naturalhistory of such lesions. While numerouscontroversial aspects remain, the main pur-pose of this report has been to set out in-terim guidelines based on best-guess esti-mates, the authors’ extensive albeit anec-dotal experience, and especially currentlyavailable published data. It cannot be over-emphasized that the guidelines proposed inthis review pertain to subsolid nodules onlyand are not intended to supplant guidelinesregarding the management of solid nodulesthat have already been published both as aconsensus statement of the Fleischner So-ciety (85) and more recently by the Amer-ican College of Chest Physicians (87).

It is anticipated that future develop-ments based on multidisciplinary effortswill result in greater consensus regardingoptimal CT classification of subsolid le-sions and ultimately more definitive, evi-dence-based guidelines leading to morerigorous standardization and ultimatelyimproved clinical treatment of patientswith subsolid lung nodules.

Acknowledgments: The authors acknowledgethe assistance provided by Martha Helmers, BA,for her contribution with the images and DaizukeNonaka, MD, for providing histopathologic images.

References1. Shiau MC, Bonavita J, Naidich DP. Adeno-

carcinoma of the lung: current concepts inradiologic diagnosis and management. CurrOpin Pulm Med 2007;13:261–266.

2. Henschke CI, Yankelevitz DF, Mirtcheva R,McGuinness G, McCauley D, Miettinen OS.

CT screening for lung cancer: frequency andsignificance of part-solid and nonsolid nodules.AJR Am J Roentgenol 2002;178:1053–1057.

3. Travis WD, Garg K, Franklin WA, et al.Evolving concepts in the pathology and com-puted tomography imaging of lung adenocar-cinoma and bronchioloalveolar carcinoma.J Clin Oncol 2005;23:3279–3287.

4. Jemal A, Siegel R, Ward E, et al. Cancer statis-tics, 2008. CA Cancer J Clin 2008;58:71–96.

5. Barsky SH, Cameron R, Osann KE, Tomita D,Holmes EC. Rising incidence of bronchioloalveo-lar lung carcinoma and its unique clinicopatho-logic features. Cancer 1994;73:1163–1170.

6. Auerbach O, Garfinkel L. The changing patternof lung carcinoma. Cancer 1991;68:1973–1977.

7. Beasley MB, Brambilla E, Travis WD. The2004 World Health Organization classifica-tion of lung tumors. Semin Roentgenol 2005;40:90–97.

8. Christiani DC, Pao W, DeMartini JC, et al.BAC consensus conference, November 4–6,2004: epidemiology, pathogenesis, and pre-clinical models. J Thorac Oncol 2006;1(9suppl):S2–S7.

9. Gandara DR, Aberle D, Lau D, et al. Ra-diographic imaging of bronchioloalveolarcarcinoma: screening, patterns of presenta-tion and response assessment. J Thorac On-col 2006;1(9 suppl):S20–S26.

10. Arenberg D. Bronchioloalveolar lung cancer:ACCP evidence-based clinical practice guide-lines (2nd edition). Chest 2007;132(3 suppl):306S–313S.

11. Dutta PR, Maity A. Cellular responses toEGFR inhibitors and their relevance to can-cer therapy. Cancer Lett 2007;254:165–177.

12. Noguchi M, Morikawa A, Kawasaki M, et al.Small adenocarcinoma of the lung: histologiccharacteristics and prognosis. Cancer 1995;75:2844–2852.

13. Ishikawa H, Koizumi N, Morita T, et al. Ultras-mall pulmonary opacities on multidetector-rowhigh-resolution computed tomography: a pro-spective radiologic-pathologic examination.J Comput Assist Tomogr 2005;29:621–625.

14. Koga T, Hashimoto S, Sugio K, et al. Lungadenocarcinoma with bronchioloalveolarcarcinoma component is frequently associ-ated with foci of high-grade atypical adeno-matous hyperplasia. Am J Clin Pathol 2002;117:464–470.

15. Carter D, Vazquez M, Flieder DB, et al. Com-parison of pathologic findings of baseline andannual repeat cancers diagnosed on CTscreening. Lung Cancer 2007;56:193–199.

16. Travis WD, Garg K, Franklin WA, et al.Bronchioloalveolar carcinoma and lung ad-



enocarcinoma: the clinical importance andresearch relevance of the 2004 World HealthOrganization pathologic criteria. J ThoracOncol 2006;1(9 suppl):S13–S19.

17. Aoki T, Nakata H, Watanabe H, et al. Evolu-tion of peripheral lung adenocarcinomas: CTfindings correlated with histology and tumordoubling time. AJR Am J Roentgenol 2000;174:763–768.

18. Yang ZG, Sone S, Takashima S, et al. High-resolution CT analysis of small peripherallung adenocarcinomas revealed on screeninghelical CT. AJR Am J Roentgenol 2001;176:1399–1407.

19. Nakata M, Saeki H, Takata I, et al. Focalground-glass opacity detected by low-dosehelical CT. Chest 2002;121:1464–1467.

20. Suzuki K, Kusumoto M, Watanabe S,Tsuchiya R, Asamura H. Radiologic classifi-cation of small adenocarcinoma of the lung:radiologic-pathologic correlation and itsprognostic impact. Ann Thorac Surg 2006;81:413–419.

21. Saito H, Yamada K, Hamanaka N, et al. Initialfindings and progression of lung adenocarci-noma on serial computed tomography scans.J Comput Assist Tomogr 2009;33:42–48.

22. Takashima S, Maruyama Y, Hasegawa M,et al. CT findings and progression of smallperipheral lung neoplasms having a replace-ment growth pattern. AJR Am J Roentgenol2003;180:817–826.

23. Naidich DP, Webb WR, Muller NL, Vlahos I,Krinsky GA. Computed tomography andmagnetic resonance of the thorax. 4th ed.Philadelphia, Pa: Lippincott, Williams &Wilkins, 2007.

24. Kodama K, Higashiyama M, Yokouchi H,et al. Prognostic value of ground-glass opac-ity found in small lung adenocarcinoma onhigh-resolution CT scanning. Lung Cancer2001;33:17–25.

25. Yokose T, Suzuki K, Nagai K, Nishiwaki Y,Sasaki S, Ochiai A. Favorable and unfavorablemorphological prognostic factors in peripheraladenocarcinoma of the lung 3 cm or less indiameter. Lung Cancer 2000;29:179–188.

26. Oda S, Awai K, Liu D, et al. Ground-glassopacities on thin-section helical CT: differenti-ation between bronchioloalveolar carcinomaand atypical adenomatous hyperplasia. AJRAm J Roentgenol 2008;190:1363–1368.

27. Kim TJ, Goo JM, Lee KW, Park CM, LeeHJ. Clinical, pathological and thin-sectionCT features of persistent multiple ground-glass opacity nodules: comparison with soli-tary ground-glass opacity nodule. Lung Can-cer 2009;64:171–178.

28. Ikeda K, Awai K, Mori T, Kawanaka K, Ya-

mashita Y, Nomori H. Differential diagnosis ofground-glass opacity nodules: CT numberanalysis by three-dimensional computerizedquantification. Chest 2007;132:984–990.

29. Nomori H, Ohtsuka T, Naruke T, SuemasuK. Differentiating between atypical adeno-matous hyperplasia and bronchioloalveolarcarcinoma using the computed tomographynumber histogram. Ann Thorac Surg 2003;76:867–871.

30. Suzuki K, Yokose T, Yoshida J, et al. Prog-nostic significance of the size of central fibro-sis in peripheral adenocarcinoma of the lung.Ann Thorac Surg 2000;69:893–897.

31. Ninomiya H, Hiramatsu M, Inamura K, et al.Correlation between morphology and EGFR mu-tations in lung adenocarcinomas: significance ofthe micropapillary pattern and the hobnail celltype. Lung Cancer 2009;63:235–240.

32. Kobayashi N, Toyooka S, Ichimura K, et al.Non-BAC component but not epidermalgrowth factor receptor gene mutation is as-sociated with poor outcomes in small adeno-carcinoma of the lung. J Thorac Oncol 2008;3:704–710.

33. Seki N, Sawada S, Nakata M, et al. Lung cancerwith localized ground-glass attenuation repre-sents early-stage adenocarcinoma in nonsmok-ers. J Thorac Oncol 2008;3:483–490.

34. Takashima S, Maruyama Y, Hasegawa M,et al. Prognostic significance of high-resolutionCT findings in small peripheral adenocarci-noma of the lung: a retrospective study on 64patients. Lung Cancer 2002;36:289–295.

35. Vazquez M, Carter D, Brambilla E, et al.Solitary and multiple resected adenocarcino-mas after CT screening for lung cancer: his-topathologic features and their prognostic im-plications. Lung Cancer 2009;64:148–154.

36. Tanoue LT. Staging of non-small cell lungcancer. Semin Respir Crit Care Med 2008;29:248–260.

37. Lee HJ, Goo JM, Lee CH, et al. Predictive CTfindings of malignancy in ground-glass nod-ules on thin-section chest CT: the effects onradiologist performance. Eur Radiol 2009;19:552–560.

38. Kim HY, Shim YM, Lee KS, Han J, Yi CA,Kim YK. Persistent pulmonary nodularground-glass opacity at thin-section CT: his-topathologic comparisons. Radiology 2007;245:267–275.

39. Takashima S, Sone S, Li F, Maruyama Y,Hasegawa M, Kadoya M. Indeterminate soli-tary pulmonary nodules revealed at popula-tion-based CT screening of the lung: usingfirst follow-up diagnostic CT to differentiatebenign and malignant lesions. AJR Am JRoentgenol 2003;180:1255–1263.

40. Park CM, Goo JM, Kim TJ, et al. Pulmonarynodular ground-glass opacities in patientswith extrapulmonary cancers: what is theirclinical significance and how can we deter-mine whether they are malignant or benignlesions? Chest 2008;133:1402–1409.

41. Usuda K, Saito Y, Sagawa M, et al. Tumordoubling time and prognostic assessment ofpatients with primary lung cancer. Cancer1994;74:2239–2244.

42. Hasegawa M, Sone S, Takashima S, et al.Growth rate of small lung cancers detectedon mass CT screening. Br J Radiol 2000;73:1252–1259.

43. Revel MP, Bissery A, Bienvenu M, Aycard L,Lefort C, Frija G. Are two-dimensional CTmeasurements of small noncalcified pulmo-nary nodules reliable? Radiology 2004;231:453–458.

44. Jennings SG, Winer-Muram HT, Tarver RD,Farber MO. Lung tumor growth: assessmentwith CT—comparison of diameter andcross-sectional area with volume measure-ments. Radiology 2004;231:866–871.

45. Wormanns D, Kohl G, Klotz E, et al. Volu-metric measurements of pulmonary nodulesat multi-row detector CT: in vivo reproduc-ibility. Eur Radiol 2004;14:86–92.

46. Ko JP, Rusinek H, Jacobs EL, et al. Smallpulmonary nodules: volume measurement atchest CT—phantom study. Radiology 2003;228:864–870.

47. Revel MP, Lefort C, Bissery A, et al. Pulmo-nary nodules: preliminary experience withthree-dimensional evaluation. Radiology 2004;231:459–466.

48. Yankelevitz DF, Reeves AP, Kostis WJ, ZhaoB, Henschke CI. Small pulmonary nodules:volumetrically determined growth ratesbased on CT evaluation. Radiology 2000;217:251–256.

49. Tran LN, Brown MS, Goldin JG, et al. Com-parison of treatment response classificationsbetween unidimensional, bidimensional, andvolumetric measurements of metastatic lunglesions on chest computed tomography.Acad Radiol 2004;11:1355–1360.

50. Jennings SG, Winer-Muram HT, Tann M,Ying J, Dowdeswell I. Distribution of stage Ilung cancer growth rates determined withserial volumetric CT measurements. Radiol-ogy 2006;241:554–563.

51. Kakinuma R, Ohmatsu H, Kaneko M, et al.Progression of focal pure ground-glass opac-ity detected by low-dose helical computed to-mography screening for lung cancer.J Comput Assist Tomogr 2004;28:17–23.

52. Lindell RM, Hartman TE, Swensen SJ, et al.Five-year lung cancer screening experience:



CT appearance, growth rate, location, andhistologic features of 61 lung cancers. Radi-ology 2007;242:555–562.

53. Kakinuma R, Kodama K, Yamada K, et al.Performance evaluation of 4 measuringmethods of ground-glass opacities for pre-dicting the 5-year relapse-free survival of pa-tients with peripheral nonsmall cell lungcancer: a multicenter study. J Comput AssistTomogr 2008;32:792–798.

54. Lee HY, Goo JM, Lee HJ, et al. Usefulness ofconcurrent reading using thin-section andthick-section CT images in subcentimetresolitary pulmonary nodules. Clin Radiol2009;64:127–132.

55. Funama Y, Awai K, Liu D, et al. Detection ofnodules showing ground-glass opacity in thelungs at low-dose multidetector computedtomography: phantom and clinical study.J Comput Assist Tomogr 2009;33:49–53.

56. Liu D, Awai K, Funama Y, et al. Identifica-tion and characterization of focal ground-glass opacity in the lungs by high-resolutionCT using thin-section multidetector helicalCT: experimental study using a chest CTphantom. Radiat Med 2008;26:21–27.

57. Ohno Y, Koyama H, Kono A, et al. Influenceof detector collimation and beam pitch foridentification and image quality of ground-glass attenuation and nodules on 16- and64-detector row CT systems: experimentalstudy using chest phantom. Eur J Radiol2007;64:406–413.

58. Marom EM, Erasmus JJ, Patz EF. Lung can-cer and positron emission tomography withfluorodeoxyglucose. Lung Cancer 2000;28:187–202.

59. Higashi K, Ueda Y, Seki H, et al. Fluorine-18-FDG PET imaging is negative in bronchi-oloalveolar lung carcinoma. J Nucl Med1998;39:1016–1020.

60. Kim BT, Kim Y, Lee KS, et al. Localized formof bronchioloalveolar carcinoma: FDG PETfindings. AJR Am J Roentgenol 1998;170:935–939.

61. Nomori H, Watanabe K, Ohtsuka T, NarukeT, Suemasu K, Uno K. Evaluation of F-18fluorodeoxyglucose (FDG) PET scanning forpulmonary nodules less than 3 cm in diame-ter, with special reference to the CT images.Lung Cancer 2004;45:19–27.

62. Tsunezuka Y, Shimizu Y, Tanaka N,Takayanagi T, Kawano M. Positron emissiontomography in relation to Noguchi’s classifi-cation for diagnosis of peripheral non-small-cell lung cancer 2 cm or less in size. WorldJ Surg 2007;31:314–317.

63. Veronesi G, Bellomi M, Veronesi U, et al. Roleof positron emission tomography scanning in the

management of lungnodules detectedat baselinecomputed tomography screening. Ann ThoracSurg 2007;84:959–965.

64. Yap CS, Schiepers C, Fishbein MC, PhelpsME, Czernin J. FDG-PET imaging in lungcancer: how sensitive is it for bronchioloal-veolar carcinoma? Eur J Nucl Med Mol Imag-ing 2002;29:1166–1173.

65. Goudarzi B, Jacene HA, Wahl RL. Diagno-sis and differentiation of bronchioloalveo-lar carcinoma from adenocarcinoma withbronchioloalveolar components with meta-bolic and anatomic characteristics usingPET/CT. J Nucl Med 2008;49:1585–1592.

66. Higashi K, Ito K, Hiramatsu Y, et al. 18F-FDG uptake by primary tumor as a predictorof intratumoral lymphatic vessel invasionand lymph node involvement in non-smallcell lung cancer: analysis of a multicenterstudy. J Nucl Med 2005;46:267–273.

67. Higashi K, Ueda Y, Arisaka Y, et al. 18F-FDG uptake as a biologic prognostic factorfor recurrence in patients with surgically re-sected non-small cell lung cancer. J NuclMed 2002;43:39–45.

68. Shimizu K, Ikeda N, Tsuboi M, Hirano T,Kato H. Percutaneous CT-guided fine needleaspiration for lung cancer smaller than 2 cmand revealed by ground-glass opacity at CT.Lung Cancer 2006;51:173–179.

69. Kim TJ, Lee JH, Lee CT, et al. Diagnosticaccuracy of CT-guided core biopsy ofground-glass opacity pulmonary lesions. AJRAm J Roentgenol 2008;190:234–239.

70. Krishna G, Gould MK. Minimally invasivetechniques for the diagnosis of peripheralpulmonary nodules. Curr Opin Pulm Med2008;14:282–286.

71. Yamato Y, Tsuchida M, Watanabe T, et al. Earlyresults of a prospective study of limited resectionfor bronchioloalveolar adenocarcinoma of thelung. Ann Thorac Surg 2001;71:971–974.

72. Nakata M, Sawada S, Saeki H, et al. Pro-spective study of thoracoscopic limited re-section for ground-glass opacity selected bycomputed tomography. Ann Thorac Surg2003;75:1601–1605.

73. Mun M, Kohno T. Efficacy of thoracoscopicresection for multifocal bronchioloalveolarcarcinoma showing pure ground-glass opaci-ties of 20 mm or less in diameter. J ThoracCardiovasc Surg 2007;134:877–882.

74. Park JH, Lee KS, Kim JH, et al. Malignantpure pulmonary ground-glass opacity nod-ules: prognostic implications. Korean J Ra-diol 2009;10:12–20.

75. Ginsberg RJ, Rubinstein LV. Randomizedtrial of lobectomy versus limited resection

for T1 N0 non-small cell lung cancer. LungCancer Study Group. Ann Thorac Surg 1995;60:615–622.

76. Asamura H, Suzuki K, Watanabe S, MatsunoY, Maeshima A, Tsuchiya R. A clinicopatho-logical study of resected subcentimeter lungcancers: a favorable prognosis for groundglass opacity lesions. Ann Thorac Surg 2003;76:1016–1022.

77. Carretta A, Ciriaco P, Melloni G, et al. Sur-gical treatment of multiple primary adeno-carcinomas of the lung. Thorac CardiovascSurg 2009;57:30–34.

78. Watanabe S, Watanabe T, Arai K, Kasai T,Haratake J, Urayama H. Results of wedgeresection for focal bronchioloalveolar carci-noma showing pure ground-glass attenuationon computed tomography. Ann Thorac Surg2002;73:1071–1075.

79. Yoshida J, Shirota T, Tanimura A, Akao M.Clinical pathway for impalpable or small lunglesions treated with coil marking and thora-coscopy. Jpn J Thorac Cardiovasc Surg2001;49:108–112.

80. Henschke CI, McCauley DI, Yankelevitz DF,et al. Early Lung Cancer Action Project:overall design and findings from baselinescreening. Lancet 1999;354:99–105.

81. Gohagan J, Marcus P, Fagerstrom R, Pinsky P,Kramer B, Prorok P. Baseline findings of a ran-domized feasibility trial of lung cancer screeningwith spiral CT scan vs chest radiograph: theLung Screening Study of the National CancerInstitute. Chest 2004;126:114–121.

82. Reich JM. A critical appraisal of overdiagnosis:estimates of its magnitude and implications forlung cancer screening. Thorax 2008;63:377–383.

83. Yankelevitz DF, Kostis WJ, Henschke CI,et al. Overdiagnosis in chest radiographicscreening for lung carcinoma: frequency.Cancer 2003;97:1271–1275.

84. Yankelevitz DF. Quantifying “overdiagnosis”in lung cancer screening [letter]. Radiology2008;246:332.

85. MacMahon H, Austin JH, Gamsu G, et al.Guidelines for management of small pulmo-nary nodules detected on CT scans: a state-ment from the Fleischner Society. Radiology2005;237:395–400.

86. Goodney PP, Lucas FL, Stukel TA,Birkmeyer JD. Surgeon specialty and opera-tive mortality with lung resection. Ann Surg2005;241:179–184.

87. Gould MK, Fletcher J, Iannettoni MD, et al.Evaluation of patients with pulmonary nod-ules: when is it lung cancer? ACCP evidence-based clinical practice guidelines (2nd edition).Chest 2007;132(3 suppl):108S–130S.



Documents

Subsolid Pulmonary Nodules and the Spectrum of Peripheral