Thorax 1993;48:70-74

Pulmonary gangrene and the air crescent sign

Jerome M Reich

AbstractBackground A study was carried out toincrease familiarity with the aetiology,pathogenesis, and radiographic featuresthat characterise pulmonary gangrene.Patients Four patients with one of thedisorders vasoinvasive aspergillosis,infarcted tuberculous cavity, chronicnecrotising aspergillosis, and gangrenedue to Pseudomonas aeruginosa wereselected because they showed the varia-tions of the typical radiographic patternand illustrated the pathogenesis. A fifthcase is also presented, in which pulmon-ary gangrene was simulated by the in-vagination of a loculated pleural effusioninto the wall ofa contiguous lung abscess.Conclusions Evolution of a crescent orrim of air within a homogeneous shadowis the feature that both heralds thedevelopment and facilitates the recog-nition of pulmonary gangrene. It is mostoften the result of vascular thrombosisinduced by the infecting organism. Theoutcome of treatment is often unfavour-able, principally because of the severityof the predisposing systemic or localunderlying disorder, although a delay indiagnosis, possibly due to unfamiliaritywith the radiographic pattern, may havecontributed to the adverse outcome insome instances.

(Thorax 1993;48:70-74)

Division of PulmonaryMedicine, Bess KaiserMedical Center,Portland, Oregon,USAJ M ReichRequests for reprints to:Jerome M Reich,3414 North Kaiser CenterDrive, Portland, Oregon97227, USADate received2 October 1991Returned to Authors28 November 1991Revised received28 January 1992Accepted 9 June 1992

Pulmonary gangrene comprises a subset of theradiographic pattern of a mass within a cavityin which both features are the consequence oftissue necrosis induced by infection. The dis-tinctive pathogenesis, prognosis, and thera-peutic implications set it apart from othercauses of this pattern. The term "sphacelus,"derived from the Greek sphakelos' (gangrene),was first used in 1575 by Banister.2 It wasemployed by Laennec' to denote a necroticmass within the cavity wall and more recentlyby Khan et al,4 who defined it as "a mass ofsloughing, gangrenous, or necrotic matter."The purpose of this communication is to

illustrate the distinguishing clinical andradiographic features of pulmonary gangrene,and to attempt to provide an explanation for itsdevelopment in each instance.

Case historiesCASE 1: ASPERGILLOMA COMPLICATING INVASIVEASPERGILLOSISA 68 year old woman with erythroleukaemia

developed widespread pulmonary infiltratesduring a one week period of profound neutro-penia (absolute neutrophil count <100/mm3).The diffuse infiltrate resolved as theneutropeniaresponded to chemotherapy for-her leukaemia,leaving a homogeneous opacity in the rightupper lobe, which subsequently cavitated, re-vealing a mass surmounted by an air crescent(fig 1). Fibreoptic bronchoscopy and trans-bronchial biopsy were unhelpful. No diagnosiswas made; the lesion continued to expand andthe patient subsequently died. At necropsy amass was identified that occupied much of theright upper lobe. It was composed of necroticlung supplied by a large thrombosed vessel.The lung tissue, arterial wall, and thrombuswere infiltrated by fungal hyphae. Similarelements, later identified as Aspergillusfumigatus, were found in the left upper lobe,liver, and spleen.The rapid onset ofdiffuse pulmonary infiltra-

tion, the absence of the organism in the bron-chial washings, and the necropsy evidence ofextrapulmonary spread suggest that the initiat-ing event was haematogenous dissemination.Theresolutionofthediffuseinfiltratecoincidingwith the appearance ofthe pulmonary gangrenecan most easily be accounted for by therecovery of the neutrophil count, enabling theeradication of aspergilli in areas in which theblood supply remained intact. Release ofneutrophilic proteolytic enzymes resulted incavitation in the right upper lobe, where infarc-tion due to vasoinvasive aspergillosis hadalready taken place.56Pulmonary gangrene and the air crescent

sign that accompanies it is most often caused by

Figure 1 Case 1: The dense shadow to the right of thehilum, which was previously homogeneous, has undergonecavitation, revealing a mass surmounted by an aircrescent.

70

on Decem

ber 19, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.1.70 on 1 January 1993. Dow

nloaded from

Pulmonary gangrene and the air crescent sign

Figure 2 Case 2: (Top) View of S cm cavity of theright upper lobe with a small amount ofpericavitaryinfiltration and a convoluted "cavity within a cavity"(arrows) thought to represent a slough of the interioraspect of the outer cavity. (Bottom) Lateral view of thecavity showing the concentric relation of the cavity andits slough.

angioinvasive pulmonary aspergillosis.78Failure to recognise the diagnostic and thera-peutic importance of this distinctive radio-graphic pattern may have dire consequences.

CASE 2: INFARCTED TUBERCULOUS CAVITYA 39 year old black male patient, known to betuberculin positive since childhood, developedfever and a cough productive of copioussputum two months before admission. Threeyears previously he underwent partial pan-createctomy to reduce pain caused by alcoholinduced pancreatitis. Poorly controlled insulindependent diabetes mellitus ensued. A chestradiograph (fig 2) showed an inner cavity wallsurrounded successively by a rim of air, anouter concentric cavity wall, and pericavitaryinfiltration. Three sputum smears werenegative for acid fast bacilli. Bronchoscopyshowed erythema and small quantities ofpus ina narrowed right upper lobe bronchus. Trans-

bronchial biopsy of the inner cavity wallshowed fragments of necrotic tissue that con-tained numerous acid fast bacilli. Bronchialwashing revealed few acid fast bacilli and wasculture positive for Mycobacterium tuberculosis.The patient responded favourably to treatmentfor tuberculosis.A clinical diagnosis of tuberculosis in this

patient with multiple risk factors, a stronglysuggestive history, and a compatible chestradiograph was nearly discarded because of therepeated absence of stainable acid fast bacilli insputum specimens. Two features support theinterpretation that infarction of the tuber-culous cavity, a rarely reported event,4 is likelyto have taken place-namely, the presence ofconcentric cavitary walls that suggest a tissueslough and the necrotic tissue recovered fromthe inner cavity wall at transbronchial biopsy.Its occurrence may have reduced the number oforganisms present in the sputum. Smearnegative cavitary tuberculosis has notpreviously been associated with or attributed tocavitary infarction. Among the four patientswith tuberculous pulmonary gangrene repor-ted by Khan et al,4 three had a postmortemexamination and were found to have extensivetuberculous pulmonary arteritis with secon-dary thrombosis. By contrast with case 2, allfour had strongly positive sputum smears. Allhad very advanced disease and it is possible thatthe organisms identified on the smears may havearisen from regions ofthe lung unaffected by theinfarction.

CASE 3: CHRONIC NECROTISING PULMONARYASPERGILLOSIS COMPLICATING RADIATIONINDUCED FIBROSISA 66 year old diabetic woman was admitted tohospital in 1986 for treatment of a broncho-pleural fistula complicated by empyema. In1978, she underwent radiotherapy of a largebut localised breast cancer. Profound radiationfibrosis of both the breast and the underlyinglung ensued. In 1982, pulmonary gangrenedeveloped in the affected area (fig 3); andtomography showed very dilated airways.Fibreoptic bronchoscopy, undertaken toevaluate haemoptysis, showed pale, constric-

Figure 3 Case 3: Linear tomography of the left upperlobe showing a mass within a cavity.

71

on Decem

ber 19, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.1.70 on 1 January 1993. Dow

nloaded from

72 Reich

Figure 4 Case 4: Notethe narrow rim of airsurrounding the mass anddemarcating itfrom thecavity wall.

ted, and foreshortened bronchi in the left upperlobe. Aspergillus fumigatus was recovered on

bronchial washing, and a radioactive immu-noassay for specific Aspergillus antibodyshowed it to be substantially increased. Theextent of pulmonary disease increased withtime. Four years later a pneumothoraxdeveloped, complicated by a bronchopleuralfistula that led to a fatal, non-group D, astreptococcal empyema. No necropsy was per-formed.

Aspergillus species may manifest theirinvasive potential in fibrotic areas of the lung,most often in the apices. Binder et al 9 suggestedthe term "chronic necrotising pulmonaryaspergillosis" for this unusual entity, which isintermediate between the angioinvasive pat-tern seen principally in patients with profoundand protracted neutropenia and the relativelybenign, non-invasive mycetoma that developsin patients with pre-existing cavities. Gefteret al described a similar appearance inpatients who developed radiation induced pul-monary fibrosis. A possible explanation for thislocalised occurrence (it is not associated withdiffuse pulmonary fibrosis) is that profound

Figure 5 Case 5: A mass of eccentric location and irregular shape lies close to the innerwall of the cavity. A left lower lobe infiltrate is present.

dilation of airways due to traction from theintense local scarring" favours colonisation byAspergillusfumigatus.CASE 4: PULMONARY GANGRENE DUE TOPSEUDOMONAS AERUGINOSA IN AIDSA 51 year old homosexual man, who hadexperienced many infections attributed toAIDS for three years, among them Pneumo-cystis carinii pneumonia, was admitted fortreatment of pneumonia. Three days earlierhe had developed chills, fever, and a non-productive cough. At the time of admission histreatment consisted of dapsone, azidothym-idine, and fluconazole, to which ethambutol,clofazimine and ciprofloxacin had been addedbecause of suspected systemic Mycobacteriumavium-intracellulare disease. His chest radio-graph showed a homogeneous lingular infil-trate. His white blood cell count was 2 6 x 109/1,of which 85% were granulocytes. Blood cul-tures were negative. Empirical treatment withcefotaxime and metronidazole was unhelpful.Bronchoscopy showed no abnormality. A30 ml bronchial lavage of the lingula showed afew neutrophils with no stainable organisms; afew colonies of Pseudomonas aeruginosa wererecovered. Treatment was changed to piper-acillin and tobramycin, but the infiltrateincreased in size and pulmonary gangrene wasindicated by the presence of a rim of air (fig 4).At thoracotomy, an 8 x 6 cm intracavitarynecrotic mass was found and a lobectomyperformed. Microscopical examination of thesurgical specimen showed mononuclear infil-tration and extensive necrosis with no thrombior vasculitic changes. Silver methenamine,Ziehl-Nielson, and Gram stains were allnegative; tissue culture was positive only forPseudomonas and the patient died with a post-operative empyema from which the sameorganism was recovered.Pseudomonas species do not figure promi-

nently among the pulmonary bacterial patho-gens afflicting patients with AIDS."2 It seemslikely that its role as a pulmonary pathogen wasconditioned by the niche created by theantibiotic treatment received at the time theinfection evolved. The suppressive effect ofthisregimen would also account for the absence ofabnormalities at bronchoscopy and the fewbacteria recovered in the bronchial washspecimen.

CASE 5: PSEUDOSPHACELUSAn 82 year old male cigarette smoker (40 pack-years) was admitted to hospital in May 1990 fortreatment of an unresolved right lower lobepneumonia. He had lifelong asthma withoccasional exacerbations requiring hospitalcare, but he was not dependent on oral steroidtreatment. He took cimetidine daily for a pepticulcer. Bronchoscopy showed mucoid secre-tions that grew normal flora. An organisingpneumonia was suspected, antibiotics werestopped, and he was asked to return for a followup chest radiograph. This showed that the areaof infiltration had undergone cavitation. Lungabscess due to aspiration of oral organisms wassuspected, and he was treated with clin-damycin. Sputum production became less and

on Decem

ber 19, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.1.70 on 1 January 1993. Dow

nloaded from

73Pulmonary gangrene and the air crescent sign

Figure 6 Case 5:Computed tomographylung window: the"intracavitary mass" is aninterlobar effusion thatinvaginated theinferolateral aspect of thecavity wall.

the cavity diminished in size. Subsequently, a

mass appeared within the cavity and an exten-sive infiltrate developed at the base of theopposite lung (fig 5). This led to the suspicionthat a sphacelus was intermittently obstructingthe bronchus draining the cavity, causingperiodic spillage. Repeat bronchoscopyshowed no new abnormalities. Lobectomy was

considered but the patient responded to a

change of antibiotic regimen to amoxacillin-clavulanate. Computed tomography (fig 6)showed that the appearance of an intracavitarymass had been simulated by invagination of theabscess cavity by an interlobar extension of thepleural reaction. The process resolved withcontinued treatment except for a residualparenchymal and interlobar pleural scar.

In this patient, misinterpretation of an

extracavitary mass as intracavitary almost ledto unnecessary surgical intervention in a highrisk patient. The repeated episodes ofpneumonia and the original lung abscess weremost likely due to episodes of aspiration.

DiscussionThe phrase "mass within a cavity" is bothcumbersome and lacking in specificity. Itencompasses several entities that includemycetoma, Rasmussen aneurysm, and haema-toma, in which a mass comes to occupy a pre or

coexisting cavity, and pulmonary gangrene,

first described by Laennec in 1826,' in whichnecrotic lung tissue is sloughed in a manner

analogous to an osteomyelitic sequestrum. Inthis entity both cavity and mass are the con-sequence of infection induced tissue necrosis.Nearly synonymous terms are slough,'3 mas-sive sequestration of the lung,'4 gangrene withspontaneous amputation,'5 spontaneous lobec-tomy," 7 infarction of a pulmonary lobe,'8massive necrosis of lung,'9 and massive pul-monary gangrene.2"2Pulmonary gangrene is most often encoun-

tered in invasive aspergillosis7-'0 and with sup-purative pneumonias due to Friedlander'sbacillus, polymicrobial anaerobic organ-isms,'320 Pneumococcus, probably in associationwith other organisms,20 and Haemophilusinfluenza.4 Mucormycosis24 and tuberculosis4are rarer causative agents. The commonpathological feature in most cases is vascularthrombosis. The appearance of an air crescentwas first recognised and applied to the descrip-tion of aspergilloma by Pesle and Monod in1954.25The common pathological feature shared by

the first three cases is the loss or reduction ofblood supply. In the angioinvasive form ofaspergillosis, pulmonary infarction is the con-sequence of mycotic invasion of the vasovasorathat results in infarction of the media andsecondary thrombus formation. Tuberculousthrombosis of a pulmonary artery branchpresumably accounted for the cavitary wallslough in case 2. The chronic invasive form ofaspergillosis seems to have a more complexmechanism: it seems likely that, owing toradiation induced impairment of blood supply,the invasive potential of Aspergilli colonisingthe ectatic bronchi is enhanced.

In contrast the presence of proteolyticexotoxins seems the most likely explanation forthe nearly identical radiographic pattern in thepatient with a Pseudomonas induced pulmonarygangrene, as no vascular involvement was seenin the lung specimen. Most strains of thisorganism produce two or three proteolyticenzymes believed to cause the mixed haemor-rhagic and infarction pattern of ecthyma gan-grenosum.2S28 Whereas it seems reasonable tosuppose that the pathological sequence in thelingula ofthis patient was analogous to ecthymagangrenosum, the absence of previous reportsof pulmonary gangrene attributable toPseudomonas pulmonary infections, as contras-ted with Friedlander's bacillus and poly-microbial anaerobic infections,1323 casts doubton the adequacy of this explanation.

Pseudosphacelus may be regarded as aradiographic curiosity as no previous reportsexist of an intracavitary mass simulated by thefortuitous juxtaposition of an interlobareffusion and a contiguous cavitary structure.Careful analysis of the radiographic sequenceenables it to be distinguished from pulmonarygangrene: the mass made its appearance afterthe abscess was fully developed and it did notconform in shape to the cavity it occupied.

I express my appreciation to Kyle Fuchs, MD and KendallBarker, MD, for their many helpful modifications of the manus-cript.

on Decem

ber 19, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.1.70 on 1 January 1993. Dow

nloaded from

Reich

1 Sphacelus. Oxford English Dictionary. 2nd ed. Oxford:Clarendon Press, 1989;XVI:203.

2 Banister J. A treatise of chyrurgerie. London: ThomasMarshe, 1575.

3 Laennec RTH. Traite de l'auscultation mediate et desmaladies des poumons et du coeur; Tome premier, 2nd ed.Paris: JS Chaude, 1826.

4 Khan FA, Rehman M, Marcus P, Azueta V. Pulmonarygangrene occurring as a complication ofpulmonary tuber-culosis. Chest 1980;77:76-80.

5 Albeda SM, Talbot GH, Gerson SL, Miller WT, CassilethPA. Pulmonary cavitation and massive hemoptysis ininvasive pulmonary aspergillosis; influence of bonemarrow recovery in patients with acute leukemia. Am RevRespir Dis 1985;131:1 15-20.

6 Janoff A, Sandhaus RA, Hospelhom VD, Rosenberg R.Digestion oflung proteins by human leukocyte granules invitro. Proc Soc Exp Biol Med 1972;140:516-9.

7 Curtis AM, Walker Smith GJ, Ravin CE. Air crescent signof invasive aspergillosis. Radiology 1979;133:17-21.

8 Slevin ML, Knowles GK, Phillips MJ, Stansfeld AG, ListerTA. The air crescent sign of invasive pulmonary asper-gillosis in acute leukaemia. Thorax 1982;37:554-5.

9 Binder RE, Faling LJ, Pugnatch RD, Mahasaen C, SniderGL. Chronic necrotizing pulmonary aspergillosis: a dis-crete clinical entity. Medicine 1982;61:109-24.

10 Gefter WB, Weingrad TR, Epstein DM, Ochs RH, MillerWT. "Semi-invasive" pulmonary aspergillosis. Radiology1981;140:313-21.

11 Westcott JL, Cole SR. Traction bronchiectasis in end-stagepulmonary fibrosis. Radiology 1986;161:665-9.

12 Murray J. State of the art: Pulmonary infectious complica-tions of human immunodeficiency virus infection; part 1.Am Rev Respir Dis 1990;141:1356-72.

13 Brock RC. Lung abscess. Oxford: Blackwell ScientificPublications, 1952.

14 Brock RC. Studies in lung abscess. Guy's Hospital Report1946;95:40-7.

15 Lenoble E, Jegat Y. Gangrene limitee a'une portion dupoumon droit avec amputation spontanee du lobesuperieur. Bulletins et memoires de la Societe Medicale desHospitaux de Paris 1921;45:1534-8.

16 Humphreys DR. Spontaneous lobectomy. BMJ 1945;2:185-6.

17 Taylor JW. Spontaneous lobectomy. BMJ 1954;2:500-1.18 Rawson AJ, Cocke JA. Infarction of an entire pulmonary

lobe with subsequent aseptic softening causing sterilehemopneumothorax. Am J Med Sci 1947;214:520-4.

19 Thomas JH, Brewster OM. Friedlander's pneumonia withmassive necrosis of lung. BMJ 1952;2:817-8.

20 Danner PK, McFarland DR, Felson B. Massive pulmonarygangrene. American Journal of Roentgenology, RadiumTherapy, and Nuclear Medicine 1968;103:548-54.

21 Gutman E, Pongdee 0, Park YS. Massive pulmonarygangrene. Radiology 1973;107:293-4.

22 Knight L, Fraser RG, Robson HG. Massive pulmonarygangrene: a severe complication of klebsiella pneumonia.Can Med Assoc J 1975;112:196-8.

23 Cameron DI, Stewart JD, Slack RC, Lewis CL. Massivepulmonary gangrene [letter to editor]. Can Med Assoc J1975;112:1290.

24 Reich JM, Renzetti AD. Pulmonary phycomycosis; report ofa case of bronchocutaneous fistula formation and pul-monary arterial mycothrombosis. Am Rev Respir Dis1970;102:959-63.

25 Pesle GD, Monod 0. Bronchiectasis due to aspergilloma.Dis Chest 1954;125:172-83.

26 Liu PV. Extracellular toxins of pseudomonas aeruginosa.J Infect Dis 1974;130(suppl O):S94-9.

27 Liu PV. The roles of various fractions ofpseudomonas in itspathogenesis. ii. Effects oflecithinase and protease. J InfectDis 1966;116:112-6.

28 Hitschman F, Kreilich K. Zur Pathogenese des BacillusPyocyaneus und zur Aetiologie des Ekthyma Gangraen-osum. Wiener Klinische Wochenschrift 1897;10:1093-101.

74

on Decem

ber 19, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.1.70 on 1 January 1993. Dow

nloaded from