CT-Guided Percutaneous Lung Biopsies in Patients With Suspicion for Infection May Yield Clinically Useful Information

Transcription

1 Vascular and Interventional Radiology Original Research Haas et al. CT-Guided Percutaneous Lung Biopsies Vascular and Interventional Radiology Original Research Brian M. Haas 1 Joshua D. Clayton Brett M. Elicker Karen G. Ordovas David M. Naeger Haas BM, Clayton JD, Elicker BM, Ordovas KG, Naeger DM Keywords: immunocompromise, lung biopsy, pneumonia, pulmonary consolidations DOI: /AJR Received February 5, 2016; accepted after revision August 2, All authors: Department of Radiology and Biomedical Imaging, University of California, San Francisco, 505 Parnassus Ave, M-391, Box 0628, San Francisco, CA Address correspondence to D. M. Naeger (david.naeger@ucsf.edu). AJR 2017; 208: X/17/ American Roentgen Ray Society CT-Guided Percutaneous Lung Biopsies in Patients With Suspicion for Infection May Yield Clinically Useful Information OBJECTIVE. The objective of our study was to assess the frequency and time frame with which CT-guided lung biopsies for suspected infection yield information that can affect patient management. MATERIALS AND METHODS. All CT-guided lung biopsies over a 68-month period performed for the purpose of diagnosing a suspected infection were reviewed to determine the proportion that yielded information affecting patient management. Patients were included if infection was the only consideration causing the pulmonary lesion in question. RESULTS. Twenty-one biopsies were performed to identify a specific organism causing infection in patients with suspected infection; all patients were receiving antibiotics, 20 (95%) were immunocompromised, and 15 (71%) had undergone a prior bronchoscopy. Material collected from the biopsy provided a diagnosis in nine (43%) patients, whereas the biopsy results were nondiagnostic in the remaining 12 (57%). Of the nine patients for whom the biopsy yielded a diagnosis, eight biopsies revealed the species causing an infection (38%) and one biopsy (5%) detected posttransplant lymphoproliferative disease. Of the nine diagnoses, management was changed as a result of the biopsy in six patients (29% of all patients). The organisms identified by CT-guided lung biopsy in eight patients were fungi of the order Mucorales (i.e., mucormycosis) (n = 3), Aspergillus (n = 3), Pseudomonas (n = 1), and Nocardia (n = 1). The mean elapsed time between biopsy and pathologic diagnosis was 4 days (median, 3 days). There was no association between prior bronchoscopy and nondiagnostic biopsy results. CONCLUSION. CT-guided lung biopsies in patients with a high pretest suspicion for infection result in information sufficient to change patient management in 29% of patients. Organisms identified in these patients were most frequently fungi. P ercutaneous CT-guided lung biopsy is a reliable method for establishing a diagnosis of cancer [1 5]. However, performance of CT-guided lung biopsy in diagnosing infection is less well studied [6 9]. In our clinical practice, we most commonly perform CT-guided lung biopsies in patients in whom there is a high probability of cancer. Occasionally, we perform CT-guided lung biopsies to determine if a lesion is an infection or a malignancy, particularly in cases in which the clinical presentation is acute and followup imaging alone would be inappropriate. Uncommonly, we biopsy lesions highly suspected of being an infection when identification of the species causing the infection is urgently needed to guide therapy. We set out to study this last group of patients to better understand this patient population and how CT-guided lung biopsy performs in these patients. Our first objective was to rigorously describe the patient population who undergo CT-guided lung biopsies because of a high pretest suspicion for infection. Our second objective was to determine the proportion of patients with biopsies that yielded diagnostic information, including to assess the time frame in which this information became available. Finally, we aimed to identify the infections ultimately diagnosed in this patient population and the percentage of patients for whom biopsy results changed management. We hypothesized that patients undergoing biopsy to identify the species of organism causing an infection would be exclusively immunocompromised hospitalized patients actively undergoing some antibiotic therapy who had already undergone bronchoscopy. We hypothesized that few biopsies would result in a diagnostic result and that even fewer biopsies would yield results that would AJR:208, February

2 Haas et al. return in a time frame to allow meaningful changes to patient management. Materials and Methods The institutional review board approved this study and waived the requirement of informed consent for this HIPAA-compliant study protocol. All CT-guided lung biopsies performed over the 68-month period between October 1, 2009, and May 31, 2015, as a means of identifying a causative organism for a presumed pulmonary infection were reviewed. All biopsies were performed in a single tertiary care hospital in an academic medical center. All biopsy requests were prospectively reviewed by an attending thoracic radiologist; when it was clinically feasible, the radiologist requested that all the medical evaluations for pulmonary findings suspicious for infection be completed before biopsy. An attending thoracic radiologist or a closely supervised cardiothoracic radiology fellow performed all biopsies. Biopsies were performed on 64-MDCT scanners; fine-needle aspiration samples were obtained via a 19-gauge coaxial system using 20- to 23-gauge biopsy needles. Core specimens were occasionally, but not routinely, obtained at the discretion of the performing radiologist in consultation with an onsite pathologist. A pathologist was always present in the procedure room to evaluate sample adequacy. Core samples were typically obtained if the onsite pathologist thought the obtained aspirates were going to be insufficient for diagnosis. Patients with a previous CTguided lung biopsy that was already included in the study population were excluded. Clinical notes were reviewed to ascertain antibiotic treatment administered at the time of biopsy, a history of and results from a prior bronchoscopy performed to identify an organism, and the occurrence of any biopsy complications. CT images were reviewed to determine the size of the biopsied abnormality (mean of longest diameter and orthogonal diameter of lesion). Pathology and microbiology reports were reviewed to determine which biopsies yielded diagnostic information and to determine the time that had elapsed between the biopsy and a positive result. Clinical notes were reviewed to determine if the biopsy results affected patient management including initiating, changing, or discontinuing antibiotics. Biopsies were considered nondiagnostic if the pathology and microbiology studies performed on a biopsy specimen failed to identify a causative organism or family of organisms to enable targeted antibiotic therapy and also failed to identify any other specific pathologic entity, such as cancer or a specific benign entity. Specifically, biopsies that showed blood, necrosis, fibrous tissue, benign bronchial cells, macrophages, atypical cells, or nonspecific inflammation were considered nondiagnostic. A Fisher exact test was performed to assess for statistically significant differences in the diagnostic yield of CT-guided lung biopsy when there was and was not a prior bronchoscopy. Statistical significance was defined as a p value Statistical analysis was performed using statistics software (QuickCalcs, version 2016, GraphPad) and Excel (version Mac2011, Microsoft). Results During the study period, 556 CT-guided lung biopsies were performed. Twenty-one (3.8%) of these biopsies were performed in patients with a high pretest suspicion for infection, and the biopsy objective was to identify a causative organism. The study group was composed of 14 men (67%) and seven women (33%); the mean age was 54 years (range, years). Demographic information and patient history before biopsy are reported in Table 1. All patients (n = 21) were receiving antibiotic therapy at the time of biopsy. Twenty (95%) of the patients were immunocompromised at the time of the biopsy: 62% (n = 13) due to related to leukemia, 10% (n = 2) due to from nonneoplastic disease (myelodysplastic syndrome or hemophagocytic lymphohistiocytosis), and 24% (n = 5) due to pharmacologic immunosuppression after organ transplant. Only fine-needle aspirates were obtained in 71% (n = 15) of patients, both fine-needle aspirates and core needle samples in 24% (n = 5), and only core needle samples in 5% (n = 1) (Table 2). Of the 21 patients, 95% (n = 20) were hospitalized at the time of biopsy, and 71% (n = 15) had previously undergone bronchoscopy in an attempt to diagnose the infection. Of the 15 patients with a history of bronchoscopy, 12 had undergone at least three prior bronchoscopies during the workup of the pulmonary abnormality (Table 1). The results from the bronchoscopies were negative, misleading, or not helpful in all but three of these patients. In one patient, cytology results from the bronchial washings identified budding yeast, and a subsequent bone marrow biopsy confirmed an active infection with Candida organisms (patient 15 in Table 1). In the two other patients, an organism that presumably was the offending agent (rhinovirus in patient 18 and unidentified species of fungi in patient 20) was identified by bronchoscopy. In 43% of the patients (n = 9), material collected from the CT-guided lung biopsy yielded diagnostic information (95% CI, 24 63%) with a causative organism identified in 38% (n = 8) and posttransplant lymphoproliferative disease (PTLD) diagnosed in 5% (n = 1). The infectious agent was identified by cytology or Gram stain in five patients, surgical pathology in one patient, culture in four patients, and polymerase chain reaction (PCR) in one patient (Table 2). The organism was identified only by cytology, Gram stain, or both in three patients; by culture only in one patient; and by PCR assay only in one patient (Table 2). The organisms identified by CT-guided lung biopsy of eight patients were fungi of the order Mucorales (i.e., mucormycosis) in three patients, Aspergillus in three patients, Pseudomonas in one patient, and Nocardia in one patient (Table 3). The mean and median elapsed time between biopsy samples obtained and pathologic diagnoses were 4 and 3 days, respectively (range, 0 13 days). Of the nine patients with a diagnosis identified by the CT-guided lung biopsy, six patients had a substantial change in clinical management as a result (29% of the total patients; 95% CI, 14 50%). The three patients with mucormycosis diagnosed by CT-guided lung biopsy underwent lobectomies, and all three patients survived more than 2 months. Of the five other patients with infections diagnosed (three with Aspergillus infection, one with Nocardia infection, and one with Pseudomonas infection), antibiotic coverage was narrowed because of the biopsy results in one patient, a new antibiotic was started and other antibiotics were stopped in one patient, and there was no immediate change to the antibiotic regimen in three patients. The patient with PTLD received appropriate therapy. Detailed information about all study patients is presented in Table 1. In the remaining 12 patients with nondiagnostic biopsy results, the abnormality was eventually diagnosed as cancer in two patients (pulmonary adenocarcinoma and lymphoma), was diagnosed as mucormycosis in one patient, and was presumed to be infection without confirmation of a causative organism in the remaining nine patients. For patients who had not undergone bronchoscopy before CT-guided lung biopsy (n = 6), 50% (n = 3) of the CT-guided lung biopsies were diagnostic. For patients who had undergone prior bronchoscopy (n = 15), 40% (n = 6) of the CT-guided lung biopsies were diagnostic (p = 1.00). 460 AJR:208, February 2017

3 CT-Guided Percutaneous Lung Biopsies Discussion Of the 556 CT-guided lung biopsies performed for all indications at a single institution during the study period, 21 (3.8%) biopsies were performed in patients with a high pretest probability of pulmonary infection to identify a causative organism. As we had hypothesized, most of these biopsies were performed in immunocompromised inpatients who were being actively treated with antibiotics. Many of the biopsy requests in this patient population were for hospitalized patients. The percentage of biopsies that were successful in identifying a causative organism (38% [8/21]) was higher than we had hypothesized; additionally, the time elapsed between biopsy and positive results was shorter than we expected. We hypothesized that CT- TABLE 1: Prebiopsy Clinical Information for 21 Patients With High Suspicion for Infection Who Underwent CT-Guided Lung Biopsy to Identify Causative Organism Patient Prior Bronchoscopy Prior Blood Culture Age No. (y) Sex Clinical Information Antimicrobials Washings Biopsy Results No. Result AFB Special Tests Legionella Antibody 1 58 F AML and Antibacterial 0 0 NA 0 NA M HLH Antibacterial, 0 0 NA 2 Sputum ( ) F Prior renal transplant Antibacterial F ALL and Antifungal Sputum ( ) M ALL and Antibacterial, M AML and Antibacterial, 0 0 NA M Prior heart transplant Antibacterial M AML and Antibacterial, M Prior lung transplant Antifungal NA M AML and Antibacterial, M Aplastic anemia and Antibacterial, Sputum and M Immunocompromised; chronic cavitary consolidation Antibacterial + 0 FP a 0 NA Sputum and Cocci Titers CMV and Viral PCR M AML and Antibacterial Sputum and F AML and Antibacterial, 0 0 NA M AML and Antibacterial, + 0 TP b M Myelodysplastic syndrome and Antibacterial, 0 0 NA M AML and Antifungal F AML and Antibacterial, + 0 TP c 3 Sputum and d F Prior liver transplant Antibacterial 0 0 NA F AML and Antibacterial, + 0 TP e M Prior kidney transplant Antibacterial, ND Sputum and 0 Note AFB = acid-fast bacillus, CMV = cytomegalovirus, PCR = polymerase chain reaction, AML = acute myelogenous leukemia, 0 = not obtained, NA = not applicable, HLH = hemophagocytic lymphohistiocytosis, minus sign ( ) = negative, plus sign (+) = positive, BAL = bronchoalveolar lavage, ALL = acute lymphocytic leukemia, FP = false-positive, TP = true-positive, ND = nondiagnostic because sample was unsatisfactory. a Few Klebsiella organisms. b Budding yeast forms. c PCR results positive for rhinovirus. d Rhinovirus. e Rare elongated fungi. AJR:208, February

4 Haas et al. guided lung biopsies in patients with a high suspicion for infection would rarely contribute to clinical management, which was incorrect. Our results showed that the management plan was altered in 29% (n = 6) of patients and that the management change involved a potentially life-saving lobectomy in 14% (n = 3). In clinical practice, we are not always informed of how biopsy results ultimately affect management, particularly when the results are delayed. This practice may have incorrectly informed our hypothesis that these biopsies are rarely useful. There are few studies in the literature about the performance of CT-guided lung biopsy to diagnose infection. One study examined the performance of CT-guided lung biopsy in diagnosing fungal infections in patients with hematologic malignancy. In that case series of 17 patients, the observed sensitivity of CTguided lung biopsy for diagnosing fungal infection was 71% [9]. However, that study included only patients with imaging findings suspicious for invasive fungal infection, whereas our study included all patients with imaging findings simply suspicious for an infection of any sort. Another study evaluated all CT-guided lung biopsies in patients with a history of solid-organ transplant [6]. That study included patients irrespective of pretest probability of infection versus cancer and observed a diagnostic yield of 53%; a substantial number of infections and malignancies were ultimately diagnosed (15 and 9, respectively) [6]. A third study, which evaluated patients with new diagnoses of hematologic malignancies and undiagnosed pulmonary lesions, also had large proportions of the biopsies yielding malignancy (63% of patients with diagnostic biopsy) [10]; therefore, the patient population in that study is substantially different from the population in our study. None of these studies examined the management implications of positive biopsies. Our study has a number of implications for patient care. Foremost, it is important to discuss with patients and referring physicians the risk of nondiagnostic biopsy results and the time course with which they can expect to receive the results. Our study highlights that biopsies in these patients are more commonly unhelpful than helpful, but they do occasionally provide useful information and can effect major changes in patient management. Further, our study shows that the ability of CT-guided lung biopsy to identify a causative organism in a proportion of ill immunocompromised patients remained despite the fact that many had negative results at prior bronchoscopy or were receiving ongoing treatment with an empirical antibiotic. Our study has several limitations. First, the sample size is small. We obtained cases TABLE 2: Data Pertaining to Biopsy Target and Outcome in 21 Patients With High Suspicion for Infection Who U nderwent CT-Guided Lung Biopsy No. Patient Age (y) Sex Clinical Information Inpatient Lesion Size (cm) Lesion Description No. of FNA Samples No. of Core Samples Complication 1 58 F AML and Yes 3.5 LUL cavitary consolidation 1 0 None 2 68 M HLH Yes 1.0 LUL nodule 1 0 None 3 51 F Prior renal transplant Yes 5.9 RUL consolidation 1 0 None 4 45 F ALL and Yes 6.3 LUL mass 3 1 None 5 54 M ALL and Yes 2.7 RLL mass 3 2 None 6 65 M AML and Yes 3.1 RLL mass 3 0 None 7 68 M Prior heart transplant Yes 1.4 RML nodule 2 0 None 8 66 M AML and Yes 7.5 RLL consolidation 6 3 None 9 61 M Prior lung transplant Yes 1.8 RLL nodule 4 0 None M AML and Yes 3.5 RLL mass 3 0 None M Aplastic anemia and Yes 1.9 LLL consolidation 3 0 None M Immunocompromised; chronic cavitary consolidation No 3.1 RML consolidation 2 0 None M AML and Yes 1.4 LLL consolidation 2 1 None F AML and Yes 4.2 RLL consolidation 1 0 None M AML and Yes 5.4 RLL infarct 6 0 None M Myelodysplastic syndrome and Yes 2.3 LUL nodule 3 0 None M AML and Yes 5.3 RUL consolidation 1 5 None F AML and Yes 5.5 LUL masslike consolidation 0 4 None F Prior liver transplant Yes 3.5 RML pleura-based nodule 3 0 Small PTX F AML and Yes 3.7 LLL mass 3 0 Small PTX M Prior kidney transplant Yes 7.6 LLL masslike consolidation 3 0 None Note FNA = fine-needle aspiration, AML = acute myelogenous leukemia, LUL = left upper lobe, HLH = hemophagocytic lymphohistiocytosis, RUL = right upper lobe, ALL = acute lymphocytic leukemia, RLL = right lower lobe, RML = right middle lobe, LLL = left lower lobe, PTX = pneumothorax. 462 AJR:208, February 2017

5 CT-Guided Percutaneous Lung Biopsies TABLE 3: Biopsy Outcome and Clinical Utility of CT-Guided Lung Biopsy in Patients With Positive Biopsy Results Patient Information No. Age (y) Sex Clinical Information Biopsy Outcome Time Between Biopsy and Organism Identification (d) Clinical Utility 1 58 F AML and Mucormycosis 0 Lobectomy 8 d after biopsy 2 68 M HLH Aspergillus infection 0 No change in management 3 51 F Prior renal transplant Nocardia infection 1 Narrowed antibiotic coverage 4 45 F ALL and Mucormycosis 1 Lobectomy 3 d after biopsy 5 54 M ALL and Aspergillus infection 5 No change in management 6 65 M AML and Aspergillus infection 6 Started new antibiotic, stopped other antibiotics 7 68 M Prior heart transplant Pseudomonas infection 6 No change in management 8 66 M AML and Mucormycosis 13 Lobectomy 14 d after biopsy 9 61 M Prior lung transplant PTLD NA Identified neoplasm Note AML = acute myelogenous leukemia, HLH = hemophagocytic lymphohistiocytosis, ALL = acute lymphocytic leukemia, PTLD = posttransplant lymphoproliferative disease, NA = not applicable. from a large time frame at a high-volume institution, but this indication remains uncommon. Second, we used a retrospective design; therefore, patient management and biopsy technique for individual patients were left to the discretion of the individual physicians. Third, this study is a single-institution study; endemic fungal infections vary across the world, and the case mix at other institutions is likely different from that at our institution. Finally, the level of support from our department of pathology is high, including providing on-site pathologists and performing advanced diagnostic evaluations of the material obtained. The level of pathology department support may differ between institutions. Conclusion CT-guided lung biopsies in patients with a high pretest suspicion for infection obtain clinically actionable information in 29% of patients. Organisms identified in these patients are most often fungi. References 1. Li H, Boiselle PM, Shepard JO, Trotman-Dickenson B, McLoud TC. Diagnostic accuracy and safety of CT-guided percutaneous needle aspiration biopsy of the lung: comparison of small and large pulmonary nodules. AJR 1996; 167: Tsukada H, Satou T, Iwashima A, Souma T. Diagnostic accuracy of CT-guided automated needle biopsy of lung nodules. AJR 2000; 175: Yeow KM, Tsay PK, Cheung YC, Lui KW, Pan KT, Chou AS. Factors affecting diagnostic accuracy of CT-guided coaxial cutting needle lung biopsy: retrospective analysis of 631 procedures. J Vasc Interv Radiol 2003; 14: Fontaine-Delaruelle C, Souquet PJ, Gamondes D, et al. Negative predictive value of transthoracic core needle biopsy: a multicenter study. Chest 2015; 148: Harter L, Moss A, Goldberg H, Gross B. CT-guided fine-needle aspirations for diagnosis of benign and malignant disease. AJR 1983; 140: Hsu JL, Kuschner WG, Paik J, Bower N, Vazquez Guillamet MC, Kothary N. The diagnostic yield of CT-guided percutaneous lung biopsy in solid organ transplant recipients. Clin Transplant 2012; 26: Lass-Flörl C, Resch G, Nachbaur D, et al. The value of computed tomography-guided percutaneous lung biopsy for diagnosis of invasive fungal infection in immunocompromised patients. Clin Infect Dis 2007; 45:e101 e Carrafiello G, Laganà D, Nosari AM, et al. Utility of computed tomography (CT) and of fine needle aspiration biopsy (FNAB) in early diagnosis of fungal pulmonary infections: study of infections from filamentous fungi in haematologically immunodeficient patients. Radiol Med 2006; 111: Nosari A, Anghilieri M, Carrafiello G, et al. Utility of percutaneous lung biopsy for diagnosing filamentous fungal infections in hematologic malignancies. Haematologica 2003; 88: Gupta S, Sultenfuss M, Romaguera JE, et al. CTguided percutaneous lung biopsies in patients with haematologic malignancies and undiagnosed pulmonary lesions. Hematol Oncol 2010; 28:75 81 AJR:208, February