Adjuvant Surgical Treatment of NontuberculousMycobacterial Lung DiseaseYuji Shiraishi, MD, Naoya Katsuragi, MD, Hidefumi Kita, MD, Akira Hyogotani, MD,Miyako H. Saito, MD, and Kiyomi Shimoda, MDSection of Chest Surgery, Fukujuji Hospital, Kiyose, Tokyo, Japan

GENERALTHORACIC

Background. According to the 2007 American Thoracic

Society/Infectious Diseases Society of America statementon nontuberculous mycobacterial diseases, more evidencefor the benefits of adjuvant nontuberculous mycobacteriallung disease surgical intervention is needed before itswide application can be recommended.

Methods. A retrospective review was conducted of 60consecutive patients who met American Thoracic Society/Infectious Diseases Society of America diagnostic criteriaand underwent pulmonary resection for localized non-tuberculous mycobacterial lung disease between January2007 and December 2011. All patients were receivingchemotherapy before resection.

Results. Included were 41 women (68%) and 19 men(32%), with a median age of 50 years (range, 20 to 72 years).Of these, 55 patients (92%) had Mycobacterium aviumcomplex disease. Bronchiectatic disease was noted in 29patients, cavitarydisease in25,both in4, andnodulardisease

Accepted for publication March 1, 2013.

Presented at the Poster Session of the Forty-ninth Annual Meeting of TheSociety of Thoracic Surgeons, Los Angeles, CA, Jan 26-30, 2013.

Address correspondence to Dr Shiraishi, Section of Chest Surgery,Fukujuji Hospital, 3-1-24 Matsuyama, Kiyose, Tokyo 204–8522, Japan;e-mail: yujishi@mvb.biglobe.ne.jp.

� 2013 by The Society of Thoracic SurgeonsPublished by Elsevier Inc

in 2. The indications for resection were a poor response todrug therapy in 52patients, hemoptysis in 6, anda secondaryinfection in 2. Sixty-five pulmonary resections were per-formed: 1pneumonectomy, 3bilobectomies, 39 lobectomies,17 segmentectomies, 3 lobectomies plus segmentectomies,and 2wedge resections. Therewerenooperative deaths, andall patients attained sputum-negative statuspostoperatively.Eleven postoperative complications occurred in 8 patients(12%); relapse was observed in only 2 (3%).Conclusions. Pulmonary resection combined with

chemotherapy is safe, with favorable treatment outcomes,for patients with localized nontuberculous mycobacteriallung disease. Our results support the liberal use ofoperations for nontuberculous mycobacterial lungdisease whenever indicated.

(Ann Thorac Surg 2013;96:287–92)� 2013 by The Society of Thoracic Surgeons

ontuberculous mycobacterial lung disease has

Nbecome an increasingly significant public healthproblem [1, 2]. In 2007, the American Thoracic Society(ATS) and the Infectious Diseases Society of America(IDSA) published an official statement, dedicated entirelyto nontuberculous mycobacterial disease, to help clini-cians diagnose, treat, and prevent this condition [3].Although the primary treatment remains chemotherapyusing multidrug regimens, the 2007 ATS/IDSAguidelines state that combining this treatment withresection of the involved portion of lung may bebeneficial to patients with limited but refractorynontuberculous mycobacterial lung disease; this wouldinclude disease caused by Mycobacterium avium complexspecies and M abscessus.

Favorable treatment outcomes have been reported forresectional operations [4–10], supporting its use asadjunctive treatment. However, the guidelines also warnthat patient selection criteria have not been established,severe postoperative complications may occur, and

experience with operations for this indication is limited[3]. More evidence is therefore needed before thewide application of resections for nontuberculousmycobacterial lung disease can be recommended. Wereport our experience with adjuvant lung resection,performed according to the 2007 ATS/IDSA guidelines,in patients with nontuberculous mycobacterial disease.

Patients and Methods

This study was approved by the Fukujuji HospitalInstitutional Review Board for Human Research. Patientconsent for participation was waived.

PatientsThe records of 60 consecutive patients who underwentpulmonary resection for localized nontuberculousmycobacterial lung disease at Fukujuji Hospitalbetween January 2007 and December 2011 were retro-spectively reviewed. No patients tested positive forhuman immunodeficiency virus. All patients werediagnosed with nontuberculous mycobacterial lungdisease at their regional hospital or at our institutionand met the ATS/IDSA diagnostic criteria. Bronchos-copy was performed in 15 patients to confirm thediagnosis or to rule out contralateral disease or coex-isting malignancies, or both.

0003-4975/$36.00http://dx.doi.org/10.1016/j.athoracsur.2013.03.008

288 SHIRAISHI ET AL Ann Thorac SurgADJUVANT OPERATIONS FOR NTM 2013;96:287–92

GENERALTHORACIC

Preoperative ManagementMultidrug regimens were initiated for all patients at theirregional hospital or at our institution. Patients withM avium complex disease received daily clarithromycin,rifampicin, and ethambutol, with the dose of clari-thromycin ranging from 600 to 800 mg, depending on thepatient’s body mass index. If necessary, aminoglycosideswere added to the regimen early in treatment. Patientswith M abscessus, M gordonae, or M xenopi receivedmultidrug regimens specialized for these organisms [3].Of the 60 patients, 41 (68%) had begun treatment attheir regional hospital and were referred to our hospitalto seek specialized care for refractory nontuberculousmycobacterial disease. If their preexisting drugregimens were appropriate, these medications werecontinued at our institution. If the regimens were notideal, they were changed at our institution.

After at least 3 months of chemotherapy, we reviewedsputum smears, culture results, and radiographic studiesto assess whether a resectional procedure was indicated.Resection of the typical bronchiectatic and cavitary lunglesions of nontuberculous mycobacterial disease wasrecommended for those patients who had persistentlypositive sputum despite chemotherapy. Of the patientswho had achieved negative sputum conversion inresponse to chemotherapy, resection was recommendedfor those who had residual bronchiectatic and cavitarydisease on imaging. One of these categories was appli-cable in 52 patients and were classified as having a poorresponse to drug therapy. Resection was also recom-mended for 6 patients with hemoptysis and for 2 witha secondary infection, regardless of the sputum smearand culture results.

Preoperative studies included chest roentgenogram,high-resolution chest computed tomography, pulmonaryfunction testing, arterial blood gas analysis, and a quan-titative lung ventilation-perfusion scan. An emphasis wasplaced on ascertaining whether patients had sufficientpulmonary reserve to tolerate pulmonary resection andwhether the most affected lung areas were reasonablylocalized for safe resection. We accepted patients withscattered nodules in other ipsilateral lobes and in thecontralateral lung on the presumption that these lesionscould be managed with postoperative chemotherapy.Limited contralateral bronchiectatic and cavitary diseasewas managed with staged bilateral resections. Improve-ment of nutritional status of thin patients preoperativelyis indispensable. If necessary, a nutritional support teamwas involved in preoperative and postoperative care ofpatients.

Surgical TechniqueThe operations were performed under general anesthesiawith the use of a double-lumen endobronchial tube. Forpatients in whom dense pleural adhesions were antici-pated, such as those with apical fibrocavitary disease withthickened pleura, we performed a posterolateral thora-cotomy. For patients in whom dense pleural adhesionswere not anticipated, such as those with middle lobe andlingular bronchiectatic disease, we performed a 6- to 8-cm

anterolateral thoracotomy and used thoracoscopicassistance.In all patients, the bronchus was divided and closed

with staples; lymphadenectomy was not performed.Bronchial stumps were reinforced with muscle flaps inpatients at high risk for bronchopleural fistula, such asthose with positive sputum and those who underwentpneumonectomy. When the residual space was large, weattempted to reduce it using a latissimus dorsi muscletent [11]. All resected specimens were sent for pathologicand bacteriologic examination.

Postoperative ManagementThe postoperative plan was for all patients to be main-tained on their multidrug regimens, generally the sameas the preoperative therapy, for at least 12 months afterthe time of operation or the time of sputum conversion.Follow-up data were obtained from outpatient orhospital records, or by direct contact with patients ortheir relatives. All follow-up was completed onDecember 31, 2012.Operative mortality was defined as all deaths that were

clearly related to the operation, regardless of the post-operative interval. The occurrence of bronchial stumpcomplications and empyema postoperatively was definedas a postoperative complication, because these conditionshave the potential to occur as late as 1 year later. Becausenontuberculous mycobacteria are found in the environ-ment and patients will occasionally demonstrate positivecultures containing small numbers of mycobacteriawithout radiographic change, this finding was notconsidered to be a sign of relapse. For the purposes of thisstudy, relapse was defined as positive sputum withprogression of residual lesions or the appearance of newlesions on radiography, or both.

Statistical AnalysisCategoric variables were analyzed using the c2 test orFisher exact test. Continuous variables were analyzedusing the Mann-Whitney U test. Statistical analyses wereperformed using StatView 5.0 software (SAS Institute Inc,Cary, NC). A p value of less than 0.05 was consideredstatistically significant.

Results

The patient group included 41 women (68%) and 19 men(32%), with a median age of 50 years (range, 20 to 72years). The median body mass index was 19.7 kg/m2

(range, 15.6 to 25.1 kg/m2). Eleven patients were smokersand 4 had diabetes mellitus (Table 1). Disease was causedby M avium complex in 55 patients (92%), M abscessus in 3,M gordonae in 1, and M xenopi in 1. Bronchiectatic diseasewas noted in 29 patients, cavitary in 25, bronchiectatic andcavitary disease in 4, and nodular disease in 2 (Table 1).Women were more likely than men to be older(p ¼ 0.0417), nonsmokers (p < 0.0001), and to havebronchiectatic disease (p < 0.0001). The median durationof preoperative chemotherapy was 14.2 months (range,3.3 to 75.2 months). Median preoperative vital capacity

Table 1. Patient Characteristics

VariableaOverall Male Female p Value(n ¼ 60) (n ¼ 19) (n ¼ 41) Male vs Female

Age, y 50 (20–72) 42 (20–68) 52 (25–72) 0.0417Body mass index, kg/m2 19.7 (15.6–25.1) 19.8 (15.6–25.1) 19.7 (16.2–23.4) 0.7147Smoker 11 10 1 <0.0001Diabetes mellitus 4 3 1 0.0894Type of disease <0.0001

Bronchiectatic 29 2 27Cavitary 25 16 9Both 4 0 4Nodular 2 1 1

a Continuous values are presented as median (range) and categoric values as numbers of patients.

289Ann Thorac Surg SHIRAISHI ET AL2013;96:287–92 ADJUVANT OPERATIONS FOR NTM

GENERALTHORACIC

was 2.83 L (range, 1.68 to 5.38 L), and forced expiratoryvolume in 1 second was 2.31 L (range, 1.25 to 4.46 L).

The 60 patients underwent 65 pulmonary resections(Table 2), and 5 patients underwent staged bilateralresections. Pneumonectomy was performed in 1 patient,bilobectomy in 3, lobectomy in 39, segmentectomy in 17,lobectomy plus segmentectomy in 3, and wedgeresection in 2. All bilateral resections consisted of rightmiddle lobectomy and subsequent left lingulectomy.The most common procedure was right upperlobectomy in men and right middle lobectomy, followedby left lingulectomy, in women. An anterolateral

Table 2. Types of Pulmonary Resection

Procedure

Overall(n ¼ 60)

No.

Male(n ¼ 19)

No.

Female(n ¼ 41)

No.

LobectomyRight upper and middle 2 (1a, 1b) 1 (1a) 1 (1b)Right middle and lower 1 (1b) 1 (1b)Right upper 12 (1a, 5b) 7 (1a, 3b) 5 (2b)Right middle 17 (1b) 17 (1b)Right lower 4 2 2

Lobectomy plus segmentectomyc

Right upper 2 (1a, 1b) 1 (1a) 1 (1b)Right middle 1 1

Segmentectomy of lower lobeRight superior 2 2Right basal 1 1

Left pneumonectomy 1 (1a) 1 (1a)Left upper lobectomy 3 (1a, 2b) 2 (1a, 1b) 1 (1b)Left lower lobectomy 3 1 2Left upper division

segmentectomy1 1

Left lingulectomy 12 1 11Left lingulectomy plus

segmentectomyc1 1

Left wedge resection 2 1 1Total 65 (5a, 11b) 19 (5a, 4b) 46 (7b)

a Resection with latissimus dorsi muscle flap. b Resection with inter-costal muscle flap. c Superior segmentectomy of the lower lobe.

thoracotomy with thoracoscopic assistance was used in55 procedures (85%). Bronchial stumps were reinforcedwith muscle flaps in 16 resections, latissimus dorsimuscle was used in 5 resections, and an intercostalmuscle in 11 resections. The median operative time was152 minutes (range, 37 to 560 minutes). The medianintraoperative blood loss was 50 mL (range, 5 to 665 mL).One patient required an intraoperative bloodtransfusion. Cultures of 31 of the 65 resectedspecimens (48%) were positive for nontuberculousmycobacteria.There were no operative deaths, and no bronchopleural

fistulae or empyema developed. A total of 11 post-operative complications occurred in 8 patients (12%;Table 3); all complications were managed withoutsurgical intervention. All patients attained sputum-negative status postoperatively. Relapse occurred in 2patients (3%). One relapse patient, who had undergonea right middle lobectomy for M avium complex disease,chose to discontinue postoperative chemotherapy 8months after the operation. Six months later, the bron-chiectatic disease in her left lingula progressed andscattered nodules appeared in the left lung; she wasrestarted on a multidrug regimen. The second patient,who had also undergone a right middle lobectomy forM avium complex disease, had positive sputum staining28 months postoperatively, while still taking chemo-therapy. Chest computed tomography 4 months laterdemonstrated new infiltrates in the right lower lobe. Shewas continued on her chemotherapeutic regimen.Themedian duration of postoperative chemotherapywas

24 months (range, 5.7 to 62 months), and 29 patients (48%)

Table 3. Postoperative Complications

Complication Patients, No.

Prolonged air leak 5Atelectasis 3Respiratory failure 1Atrial fibrillation 1Hemorrhage 1Total 11

290 SHIRAISHI ET AL Ann Thorac SurgADJUVANT OPERATIONS FOR NTM 2013;96:287–92

GENERALTHORACIC

remained on chemotherapy at the time of final follow-up.No late deaths occurred. Of the 60 patients, all of whomwere alive at the time of final follow-up, 58 (97%) wereconsidered free from relapse; these patients had a medianfollow-up of 34 months (range, 13.2 to 70.3 months).

Comment

The prevalence of nontuberculous mycobacterial lungdisease is increasing [1, 2], and the need to better controlthis disease is an emerging public health issue. Medicaltreatment with multidrug regimens remains the primarytreatment modality; however, a multidisciplinary treat-ment approach using surgical resection combined withmedical therapy may be beneficial for patients withintractable nontuberculous mycobacterial lung disease.The 2007 ATS/IDSA guidelines [3] state that surgicalresection of limited M avium complex lung disease can besuccessful, when combined with multidrug treatmentregimens, in patients with cardiopulmonary reserveadequate to withstand partial or complete lung resection.The guidelines also state that the only therapy thatreliably cures limited M abscessus lung disease is surgicalresection of the involved lung combined with multidrugchemotherapy. However, the following caveats are alsogiven: first, no criteria for patient selection have beenestablished; second, the operation is potentiallyassociated with severe perioperative complications; andthird, few institutions have gained extensive experiencewith mycobacterial operations [3]. More evidence of thebenefits of surgical resection is therefore needed beforeits wide application can be recommended.

The present study is the second-largest patient series inthe English literature to evaluate resection for non-tuberculous mycobacterial lung disease; the largestcohort was reported by the Denver group [7]. In thatstudy, 80% of patients had M avium complex disease [7],as did 92% of patients in our study. In our previousreport on resection for M avium complex disease, menand women were equally distributed, and cavitarydisease predominated [12]. In contrast, womenpredominated in the present study, and bronchiectaticand cavitary diseases were seen in equal proportions.

These changes in patient characteristics may reflectthat the proportion of women with bronchiectaticdisease has increased among the patients with non-tuberculous mycobacterial disease in Japan. Recentstudies from the United States also report a predomi-nance of women and of bronchiectatic disease. In theDenver group, 83% of patients were women, and bron-chiectasis was seen in 55% [7]. In another study from theUnited States, 83% of patients were women andbronchiectasis and nodular opacities were seen in 92%[10]. In women with nontuberculous mycobacteriallung disease, bronchiectasis is chiefly observed in themiddle lobe, the lingula, or both [13]; right middlelobectomy and left lingulectomy are therefore commonprocedures in women [7, 10].

The optimal duration to administer chemotherapybefore assessing patients for surgical indications remains

unclear. One prior report describes preoperative antibi-otic therapy lasting 2 to 6 months [7], whereas anotherlists treatment ranging from 2 to 37 months [6]. On thebasis of our experience with preoperative chemotherapyfor multidrug-resistant tuberculosis [14], we hypothesizethat several months of chemotherapy are probablynecessary for nontuberculous disease. This duration ofchemotherapy allows ample time for multidrugregimens to lessen the mycobacterial burden, helpingus evaluate which lesions actually require surgicalresection. The duration of preoperative chemotherapynoted in the present study was longer than weexpected. This is perhaps because of the indolent andnoncontagious nature of their disease, it often took timeto persuade patients to agree to an operation. It istherefore important to collect evidence of the benefits ofthese operations to present the argument for resectionmore clearly to patients.The surgical indications used at our institution are in

line with the ATS/IDSA guidelines [3]: a poor response todrug therapy, the development of macrolide-resistantdisease, or the presence of a significant disease-relatedcomplication such as hemoptysis. We did not routinelyperform clarithromycin-susceptibility testing on M aviumcomplex isolates during this study period, but we agreewith the guidelines that this test should be used morefrequently to create better treatment regimens. We arecurrently preparing to use this test routinely. The goal ofthe operation is to prevent disease progression byremoving the areas of lung most affected and harboringthe largest amounts of mycobacteria. The main surgicaltargets are therefore localized cavitary or bronchiectaticlesions that lend themselves to resection; the presence ofscattered nodules in ipsilateral and contralateral lobesdoes not preclude surgical resection. The present studyincluded 2 patients with nodular disease that began ascavitary disease, with the lesions converted by chemo-therapy. We offered surgical resection to these patientsbecause the nodules were considered to be at risk ofreturning to cavitary form.The postoperative mortality was 0% and morbidity was

12%. The most common complication was a prolonged airleak; no patients experienced bronchopleural fistula orempyema. The morbidity in the present study was lowerthan the 28.6% reported in our previous study [12] andlower than that reported by the Denver group, in whichthe major and minor morbidity rates were 11.7% and6.8%, respectively [7]. We previously reported thatpatients undergoing right pneumonectomy fornontuberculous mycobacterial lung disease are at riskfor bronchopleural fistula [15]; this finding was alsoobserved by the Denver group [7]. No patients in thisstudy underwent right pneumonectomy, and only 1patient underwent left pneumonectomy. Theinfrequency of this procedure may have contributed tothe fact that no study patients were affected bya bronchopleural fistula. The complication may also havebeen prevented by the use of bronchial stump coveragewith muscle flaps in high-risk patients. Mitchell andassociates [16] recently reported the safety of

291Ann Thorac Surg SHIRAISHI ET AL2013;96:287–92 ADJUVANT OPERATIONS FOR NTM

GENERALTHORACIC

thoracoscopic lobectomy and segmentectomy forpatients with infectious lung disease; they did notroutinely buttress the bronchial stump closure. Weagree that thoracoscopic resection without muscleflaps is safe for patients with middle lobe and lingularbronchiectasis because the risk of bronchopleuralfistula is low for these patients. We would emphasize,however, that patients at high risk of bronchopleuralfistula should undergo bronchial stump coverageusing muscle flaps.

Our relapse rate of 3% was acceptably low andcomparable to the relapse rates of 4% to 5% in previousreports [5, 6]. It is noteworthy that the median duration ofpostoperative chemotherapy in this study was longerthan that of previous reports. Prior studies report medianpostoperative chemotherapy durations of 6 months(range, 3 to 35 months) [5], and 8.5 months (range, 2 to 12months) [8]. The guidelines state that therapy should becontinued until sputum cultures have been negative for12 months [3]. This means that a 12-month duration ofpostoperative chemotherapy is appropriate if sputumconversion is achieved by the operation; however, theoptimal duration of postoperative chemotherapy isunknown.

One patient experienced progression of residualcontralateral lesions after intentionally discontinuingpostoperative chemotherapy. Our institutional policy is tokeep patients on chemotherapy for more than 12 monthsif their resected specimens yield a positive culture or ifthey have residual small cavitary lesions, limited bron-chiectasis, or scattered nodular disease. To minimize therisk of relapse, we think that the decision to discontinuepostoperative chemotherapy should be based on whetherpatients have residual disease. We would like toemphasize that surgical resection can achieve favorabletreatment outcomes only in the setting of multimodalitytherapy that includes rigorous postoperative chemo-therapy. Further investigation is needed to determinewhether the addition of an operation can shorten theoverall duration of chemotherapy.

This study has several limitations. First, the patientswere highly selected; in some patients the disease was tooextensive to be considered operable. Eligibility forsurgical resection can itself be a predisposing factor fora favorable outcome. Second, the duration of post-operative follow-up was relatively short, and half of thepatients remained on chemotherapy at the time of finalfollow-up. Lifelong follow-up is needed to assess the trueefficacy of adjunctive resection because relapse andreinfection may occur years after the completion oftherapy [17].

In summary, pulmonary resection combined withchemotherapy is safe, with favorable treatment outcomes,in patients with localized nontuberculous mycobacteriallung disease. Our results support the liberal use ofsurgical resection for nontuberculous mycobacterial lungdisease whenever indicated. We hope this may encourage

our medical colleagues to refer such patients for surgicalresection.

References

1. Adjemian J, Olivier KN, Seitz AE, Holland SM, Prevots DR.Prevalence of nontuberculous mycobacterial lung disease inU.S. Medicare beneficiaries. Am J Respir Crit Care Med2012;185:881–6.

2. Prevots DR, Shaw PA, Strickland D, et al. Nontuberculousmycobacterial lung disease prevalence at four integratedhealth care delivery systems. Am J Respir Crit Care Med2010;182:970–6.

3. Griffith DE, Aksamit T, Brown-Elliott BA, et al. An officialATS/IDSA statement: diagnosis, treatment, and preventionof nontuberculous mycobacterial diseases. Am J Respir CritCare Med 2007;175:367–416.

4. Pomerantz M, Madsen L, Goble M, Iseman M. Surgicalmanagement of resistant mycobacterial tuberculosis andother mycobacterial pulmonary infections. Ann Thorac Surg1991;52:1108–12.

5. Nelson KG, Griffith DE, Brown BA, Wallace RJ Jr. Results ofoperation inMycobacterium avium-intracellulare lung disease.Ann Thorac Surg 1998;66:325–30.

6. Watanabe M, Hasegawa N, Ishizaka A, et al. Early pulmo-nary resection for Mycobacterium avium complex lung diseasetreated with macrolides and quinolones. Ann Thorac Surg2006;81:2026–30.

7. Mitchell JD, Bishop A, Cafaro A, Weyant MJ, Pomerantz M.Anatomic lung resection for nontuberculous mycobacterialdisease. Ann Thorac Surg 2008;85:1887–93.

8. van Ingen J, Verhagen AF, Dekhuijzen PN, et al. Surgicaltreatment of non-tuberculous mycobacterial lung disease:strike in time. Int J Tuberc Lung Dis 2010;14:99–105.

9. Griffith DE, Girard WM, Wallace RJ Jr. Clinical features ofpulmonary disease caused by rapidly growing mycobacteria.An analysis of 154 patients. Am Rev Respir Dis 1993;147:1271–8.

10. Jarand J, Levin A, Zhang L, Huitt G, Mitchell JD, Daley CL.Clinical and microbiologic outcomes in patients receivingtreatment for Mycobacterium abscessus pulmonary disease.Clin Infect Dis 2011;52:565–71.

11. Rocco G. Pleural partition with intrathoracic muscle trans-position (muscle tent) to manage residual spaces aftersubtotal pulmonary resections. Ann Thorac Surg 2004;78:e74–6.

12. Shiraishi Y, Nakajima Y, Takasuna K, Hanaoka T,Katsuragi N, Konno H. Surgery for Mycobacterium aviumcomplex lung disease in the clarithromycin era. Eur JCardiothorac Surg 2002;21:314–8.

13. Reich JM, Johnson RE. Mycobacterium avium complexpulmonary disease presenting as an isolated lingular ormiddle lobe pattern. The LadyWindermere syndrome. Chest1992;101:1605–9.

14. Shiraishi Y, Nakajima Y, Katsuragi N, Kurai M, Takahashi N.Resectional surgery combined with chemotherapy remainsthe treatment of choice for multidrug-resistant tuberculosis.J Thorac Cardiovasc Surg 2004;128:523–8.

15. Shiraishi Y, Nakajima Y, Katsuragi N, Kurai M, Takahashi N.Pneumonectomy for nontuberculous mycobacterial infec-tions. Ann Thorac Surg 2004;78:399–403.

16. Mitchell JD, Yu JA, Bishop A, Weyant MJ, Pomerantz M.Thoracoscopic lobectomy and segmentectomy for infectiouslung disease. Ann Thorac Surg 2012;93:1033–40.

17. Wallace RJ Jr, Zhang Y, Brown-Elliott BA, et al. Repeatpositive cultures in Mycobacterium intracellulare lungdisease after macrolide therapy represent new infections inpatients with nodular bronchiectasis. J Infect Dis 2002;186:266–73.