- Email: [email protected]
Percutaneous Lung Biopsy in Immunocompromised Pediatric Patients
CLINICAL STUDY
Percutaneous Lung Biopsy in Immunocompromised Pediatric Patients Heather Cleveland, BSRS, Alex Chau, MD, Zachary Jeng, MD, Gregory Gardner, MD, Raphael Yoo, MD, Wei Zhang, PhD, and Jose Hernandez, MD
ABSTRACT Purpose: To determine the diagnostic yield and safety of image-guided lung biopsies in immunocompromised pediatric patients. Materials and Methods: This was a retrospective pediatric cohort study conducted from June 2000 to April 2017. Subjects were 0–17 years of age (median, 10 years of age). There were 46 males (48%). A total of 73 consecutive image-guided lung biopsies were performed in 68 patients (weight range, 4.9–97.3 kg [median, 25.3 kg]). The indication for biopsy was to isolate an organism to tailor medical therapy. All patients were immunocompromised with an underlying history of bone marrow transplantation (n ¼ 50), primary immunodeficiency (n ¼ 14), and solid organ transplantation (n ¼ 4). Patient and technical factors were analyzed for rates of complication. Results: Overall diagnostic yield was 43 of 73 patients (60%). There were 14 minor (19%) and 8 major (11%) complications. Major complications included pneumothorax or hemoptysis requiring intervention (n ¼ 6), and death (n ¼ 2). The histological diagnosis was an infectious cause in 5 of 8 major complications (63%). There were statistically significant differences between the rates of complications with the imaging modality used (P ¼ .02) and the use of fine needle aspiration (P ¼ .02). Conclusions: Image-guided percutaneous lung biopsy can be helpful in isolating an organism to tailor therapy. Biopsies performed in immunosuppressed patients result in an elevated complication risk of up to 30% and demonstrate lower diagnostic yield and increased mortality, which should warrant detailed discussion with the primary team and family.
Lung lesions in children can be caused by a variety of pathological processes including infection and inflammation, primary or secondary malignancy, and post-transplantation lymphoproliferative disease in immunosuppressed patients (1). With continued refinement and advances in dose reduction for cross-sectional imaging in children, the request for direct tissue sampling of these increasingly commonly detected lesions has grown (2). Although percutaneous image-guided biopsy offers many advantages over conventional open, and more recently, video- assisted surgical biopsy or endoscopic bronchoalveolar lavage, historical referral patterns, and competition have resulted in more extensive
From the Department of Interventional Radiology (H.C., A.C., G.G., R.Y., W.Z., J.H.) Texas Children’s Hospital, 6621 Fannin Street, Houston, TX 77030-2608; and the Department of Interventional Radiology (A.C., Z.J., G.G., R.Y., J.H.), Baylor College of Medicine, Houston, Texas. Received August 25, 2018; final revision received July 17, 2019; accepted July 18, 2019. Address correspondence to J.H.; E-mail: [email protected]
discussion of surgical rather than percutaneous options in the pediatric reports (3). Consequently, although image-guided percutaneous lung biopsy is well established in adult patients, comparatively less evidence exists for this procedure in children (4,5). In the limited pediatric reports, percutaneous lung biopsies are generally considered safe, with a high diagnostic accuracy (3,6). Although 1 study has documented the risk of bleeding as high as 46% for percutaneous biopsies in children with aspergillosis, which is especially prevalent among immunosuppressed children, the largest series to date has reported a low rate of complications (only 1 major complication in 64 patients), with nearly all studies documenting a high diagnostic accuracy of 85% to 100% (3,7). However, no research exists assessing differences in safety and diagnostic accuracy for percutaneous lung biopsies performed in children with immunosuppression. Thus, the purpose of this study was to review a single-center experience with image-guided percutaneous lung biopsy in immunosuppressed children to determine procedural safety and efficacy.
None of the authors have identified a conflict of interest. From the SIR 2018 Annual Scientific Meeting.
MATERIALS AND METHODS
© SIR, 2019
Demographics
J Vasc Interv Radiol 2020; 31:93–98
This was a retrospective Institutional Review Boardapproved study. The Epic electronic medical records (Epic
https://doi.org/10.1016/j.jvir.2019.07.016
94 ▪ Percutaneous Lung Biopsy in Immunocompromised Pediatrics
Systems, Verona, Wisconsin) for a single tertiary pediatric referral center were reviewed to identify all percutaneous image-guided lung biopsies performed from June 2000 to April 2017 in patients between the ages of 0 and 17 years (median, 10 years). Patient weight ranged from 4.9 to 97.3 kg (median, 25.3 kg) (Table 1). Demographic information, comorbidities, diagnostic imaging, technical details of the procedure, pathology reports, laboratory data, surgical records, procedural complications, and the postprocedure course until patient discharge were obtained for all patients through chart review. The initial percutaneous biopsy diagnostic adequacy and final diagnosis were determined based on pathology reports. Adequacy was determined by confirmation of histological sample and ability to diagnose the infection by the pathologist. A total of 81 consecutive biopsy procedures were performed in 76 immunocompromised patients (50 males and 26 females) referred for biopsy of indeterminate lesions or masses identified on prior imaging. Excluded from this study were patients in whom there was a suspicion of malignancy based on clinical presentation (n ¼ 8). The remaining 73 procedures in 68 patients were performed after preprocedure noncontrast chest computed tomography (CT) (n ¼ 73). Lesions were either pleural (n ¼ 61) or parenchymal (n ¼ 12) based. Pleural-based lesions were defined as having contact with the pleural surface. Biopsy specimens were sent for microbiological analysis to isolate an organism to guide treatment (n ¼ 73) and pathologic analysis (n ¼ 18) in which a fellowship-trained pediatric cytopathologist was present to evaluate the specimen. All biopsies were performed in immunocompromised patients with an absolute neutrophil count (ANC) <1.5 103 UL of cells receiving chemotherapy or those receiving immunosuppressive medications (most commonly tacrolimus or high-dose prednisone).
Procedure Preprocedural laboratory values were obtained in all patients, with a required minimum platelet count 50,000/μL and an international normalized ratio (INR) 1.5 with fresh frozen plasma and platelet transfusions administered as necessary. There was no cutoff value for preprocedural ANC levels (based on the normal pediatric range of 1.5–8 [1,000/ mm3]). A fellowship-trained pediatric anesthesiologist was present throughout each procedure to administer monitored anesthesia care or general anesthesia. Complications were graded based on the SIR classification system (8). The preprocedural cross-sectional imaging for each patient was reviewed to determine the optimal positioning and approach for percutaneous biopsy. The choice of needle, gauge (20 G), throw length, number of passes (standard of 4 passes, 2 FNAs, and 2 core samples), and technique (coaxial vs. tandem needle) was determined at the time of the procedure by the pediatric interventional radiologist. Lesion sizes (ranging from 10–40 mm, based on the longest
Cleveland et al ▪ JVIR
Table 1. Patient, Technical, and Lesion Characteristics (n ¼ 73) Characteristic % of females Age (y)* Weight (kg)* Height (cm)* Body mass index*
Total 34 10 (0-17) 25.3 (4.9-97.3) 127.5 (59-172) 17.3 (13.1-33.6)
Laboratory values Absolute neutrophil count (1,000/mm3)*
3.3 (0-19)
International normalized ratio*
1.1 (0.9-1.5)
Hemoglobin (g/dL)* Technical factors
9.2 (3.6-13.1)
CT-guided
27
US guided
41
Combined CT and US guidance Interventionalist experience (y)
5 16 (7-30)
Lesion characteristics Size (mm) % Centrally located
25 (5-55) 16
*Values are median.
axis of the lesion), location (central, nonpleural-based vs. peripheral, pleural-based), and presence of consolidation (ill-defined homogeneous opacities with obscuration of vessels and alveolar air bronchograms vs. lesions with some ground glass components) were determined based on preprocedural or intraoperative imaging. Procedures were performed by 7 interventional radiologists with experience ranging from 7 to 30 years (median, 16 years). All but 1 operator completed training in a dedicated pediatric interventional radiology fellowship because this training was not part of routine practice when he began practicing in the 1990s. Beginning in 2013, a cytopathologist was present during all lung biopsy cases to evaluate the FNA samples prior to core biopsy. Preliminary diagnosis was often obtained from the FNA. Core biopsies were invariably requested to provide adequate tissue for further immunologic, histologic, cytologic, or pathologic evaluation. All biopsies were sent for microbiologic and/or pathological evaluation. Ultrasonography-guided biopsies were performed using free-hand technique, with placement of the ultrasonography probe along the longest axis of the lesion at the most appropriate intercostal space. All procedures were performed during the respiratory phase allowing for the best visualization of the lesion, most often at end-expiration (n ¼ 39). All lesions were biopsied through the pleural surface. In CT-guided biopsies, patients were positioned according to the optimal approach based on preprocedural imaging. Patient positioning included prone (n ¼ 18), supine (n ¼ 6), and right lateral decubitus (n ¼ 3). All biopsies were performed using coaxial needle technique (median passes, 5; range, 1–20). Tract embolization was not performed.
Volume 31 ▪ Number 1 ▪ January ▪ 2020
95
Table 2. Lung Biopsy Tissue Diagnosis and Complications (n ¼ 73) Diagnosis
Biopsy with Complication (n ¼ 22)
Biopsy without Complication (n ¼ 51)
Total (N ¼ 73)
Fungal
4 minor 5 major
10
19
Bacterial
1 minor
11
12
Viral Malignancy
2 minor 2 minor 1 major
2 4
4 7
Sequela of drug reaction
1 minor
Inconclusive or Inadequate
4 minor 2 major
1 24
30
A chest radiograph of all patients was taken 2 hours after the procedure to assess for complications. If patients showed signs and symptoms of complications such as pneumothorax or bleeding, the imaging interval was shortened. When biopsies were preformed using CT guidance, a postprocedural chest CT was obtained immediately after needle removal to assess for complications.
Statistical Analysis Statistical analysis was performed to compare the effects of various factors on the complication rate. Continuous variables were compared using the Wilcoxon rank test, and categorical variables were compared using the Fisher exact test.
RESULTS Four patients (1 female and 3 males) required repeat biopsies. Two of these children underwent 2 biopsies, both of different lung lesions. Two remaining patients underwent 3 biopsies, both of the same lesion subsequently and of a different lung lesion on the third biopsy. Of the 4 patients who had repeated biopsies, 3 underwent surgical biopsy for diagnosis, and 1 had a conclusive percutaneous biopsy. Procedures were performed with the patient under monitored anesthesia care (n ¼ 25, 34%) or general anesthesia (n ¼ 48, 66%). Core biopsy samples were obtained with
16-G to 20-G Cook (n ¼ 38; Cook Medical, Bloomington, Indiana) and BioPince (n ¼ 35; Argon Medical Devices, Frisco, Texas) devices. FNA was performed using 20-G to 25-G Chiba needles (n ¼ 18; Cook Medical). When 5 FNA passes were performed, the median was 6 passes (range, 5–20 passes). When 4 FNA passes were performed, the median was 4 passes (range, 1–4). Core biopsies alone were obtained in a majority of the cases (n ¼ 55, 75%). FNA was performed in conjunction with core needle biopsy (n ¼ 18, 22%) beginning in 2013. Pleural-based lesions (n ¼ 61, 84%) were preferentially biopsied by using ultrasonography guidance if adequately visualized (n ¼ 41, 56%). Adjunctive CT guidance (n ¼ 5, 7%) was used if pleural-based lesions were unable to be fully visualized with ultrasonography. The remaining pleural-based and parenchymal lesions were biopsied under CT guidance (n ¼ 27, 37%) (Table 1). All biopsies were performed in immunocompromised patients with underlying conditions including post-bone marrow transplantation (n ¼ 52, 71%), primary immunodeficiency (n ¼ 16, 22%), or post-solid organ transplantation (n ¼ 5, 7%). Diagnostic yield, where histological or microbiological diagnosis was successfully obtained, was 43 of 73 (60%). The final diagnoses for biopsies performed for each indication are listed in Table 2. Of the inconclusive or inadequate biopsies, patients were empirically treated with antifungal agents and broad spectrum antibiotics (n ¼ 8); underwent surgical biopsy (n ¼ 5) or bronchoalveolar lavage (n ¼ 1); had a repeat biopsy that was conclusive (n ¼ 1); received tailored therapy based on a fungal biopsy sample of a liver lesion (n ¼ 1) or died during biopsy (n ¼ 1). There were no progress notes available for the remaining patients (n ¼ 13). Overall, complications occurred in 22 procedures (30%). A total of 14 minor (SIR categories A and B) complications (19%) occurred that required no further intervention or escalation in care. Minor complications included hemoptysis and bleeding (n ¼ 5); small, asymptomatic pneumothorax (n ¼ 7); hemothorax (n ¼ 1); and hemopneumothorax (n ¼ 1). A total of 8 major complications (10%) (Table 3) occurred, including pneumothorax requiring chest tube placement (n ¼ 3) (SIR categories C and D); hemoptysis requiring extended intubation, and intensive care unit care
Table 3. Major Complications Age
Tissue Diagnosis
Imaging Modality
Complication
1
Patient
6
Fungal
CT
Pneumothorax, chest tube
C
2 3
8 17
Malignancy Fungal
US CT
Pneumothorax, chest tube Pneumothorax, PICU
C D
4
2
Fungal
US
Hemoptysis, PICU
D
5
3
Fungal
US
Hemoptysis, PICU
D
6
1
Inconclusive
CT
Hemorrhage, PICU
D
7
2
Fungal
US/CT
Death
F
8
3
Inconclusive
US
Death
F
BAL ¼ bronchoalveolar lavage; PICU ¼ pediatric intensive care unit.
SIR Classification
96 ▪ Percutaneous Lung Biopsy in Immunocompromised Pediatrics
Table 4. Statistical Analysis of Categorical Factors’ Effect on the Complication Rate Categorical Factors
Number of Patients
Complication Count (% of N)
5 y
24
11 (45.8)
>5 y
49
11 (22.4)
CT
27
13 (48.1)
US
41
7 (17.1)
Age
.06
Modality
US and CT Gauge
.02
5
1.0 33
10 (30.3)
18
38
12 (31.6)
16
2
0
Consolidation
21
9 (42.9)
Lesion
52
Lesion or consolidation
.16 13 (25) .50 9
2 (22.2)
11-19
12
6 (50)
20-29
28
9 (32.1)
30-39
12
3 (25)
40
12
2 (16.7)
4
26
8 (30.8)
5 Multiple
39 8
10 (25.6) 4 (50.0)
With
18
10 (55.6)
Without
55
12 (21.8)
Central
12
5 (41.7)
Peripheral
61
17 (27.9)
61
19 (31.1)
12
3 (25)
<1.5
18
2 (11.1)
1.5 – 8
38
15 (39.5)
>8
14
5 (35.7)
29 43
11 (37.9) 11 (25.6)
50
13
4 (30.8)
51-99
18
100
41
16 (39)
10
43
10 (23.3)
10.1
29
12 (41.4)
Number of passes
.38
FNA
.02
Location of lesion
Pleural vs parenchymal Pleural Parenchymal
.49
1.0
ANC (3 missing)
.09
INR (1 missing) 1 1.1-1.5
.30
Platelets (1 missing)
.09 2 (11.1)
Hemoglobin (1 missing)
Table 4. Statistical Analysis of Categorical Factors’ Effect on the Complication Rate (continued) Categorical Factors Diagnostic Yield Adequate Inadequate Inconclusive
Number of Patients
Complication Count (% of N)
43 7 23
14 (32.6) 3 (42.9) 5 (21.7)
P Value .44
ANC ¼ absolute neutrophil count; FNA ¼ fine needle aspiration; INR ¼ international normalized ratio.
2 (40)
20
Size of lesion (mm) 10
P Value
Cleveland et al ▪ JVIR
.12
continued
(n ¼ 3) (SIR categories C and D); and death (n ¼ 2) (SIR category F) from respiratory arrest presumably secondary to asphyxiation from pulmonary hemorrhage. The histological diagnosis was determined to be caused by infection in 5 of 8 patients (63%) with major complications. Six of the 8 patients (75%) with major complications had biopsies performed with FNA, of which 4 were found to be inconclusive. In the 3 patients requiring chest tube placement, all chest tubes were removed within 3 days after the procedure, and 1 patient was able to have the chest tube removed within hours of admission to the postanesthesia care unit. In the 3 patients with hemoptysis requiring extended intubation, all required admission to the pediatric intensive care unit. All 3 patients were successfully extubated within 2 days after the procedure and had uneventful return to the regular inpatient floor. Several technical factors were assessed to determine the effect on the complication rate. There was a statistically significant correlation with the complication rate with the use of ultrasonography versus CT (P ¼ .02) (Table 4). There was a lower rate of complications for ultrasonography-guided procedures (17%). All 41 of the procedures performed using ultrasonography only were pleural-based lesions. There were statistically significant differences in complications with the use of FNA during biopsy (P ¼ .02). Of the patients who had conclusive diagnoses, fungal infection had a 47% (9 of 19) complication rate, whereas the other diagnoses had a 29% (7 of 24) complication rate (P ¼ .34). Several patient demographics and laboratory values were also assessed to determine their effect on the complication rate. Neither patient age, weight, height, nor body mass index were statistically significant (P ¼ .13, P ¼ .17, P ¼ .06, P ¼ .1, respectively)(Table 5). The preprocedure laboratory values were found to be insignificant (Table 4). Two deaths occurred due to complications experienced during biopsy. The first patient, a 2-year-old girl with graftversus-host disease underwent biopsy to determine possible fungal infection (the initial FNA samples were inconclusive). The patient died at the end of the procedure and was found to have pneumocystis fungal hyphae with angioinvasion on autopsy. The second patient, a 12-year-old boy with a history of relapsed acute lymphoblastic leukemia underwent biopsy to evaluate suspected multifocal
Volume 31 ▪ Number 1 ▪ January ▪ 2020
97
Table 5. Statistical Analysis of Continuous Factors’ Effect on the Complication Rate Continuous Factor Age
Overall (N ¼ 73)
Without Complication (n ¼ 51)
With Complication (n ¼ 22)
P Value
8.7 ± 5.6
9.3 ± 5.1
7.1 ± 6.5
.13
Weight (kg)
32.8 ± 21.1
Height (cm)
126.1 ± 32.2
33 ± 17.4 131.4 ± 29
32.3 ± 28.1
.17
114.2 ± 36.4
.06
Body mass index
18.6 ± 4.5
17.8 ± 3.3
20.6 ± 6.2
.10
Years of experience
19.5 ± 8.1
20 ± 7.6
18.5 ± 9.3
.14
pneumonia that had worsened despite prolonged antibacterial and antifungal therapy. The patient died and was found to have nonspecific acute lung injury, with a diagnosis that was never able to be precisely determined on autopsy.
DISCUSSION Open biopsy is used for establishing a histopathological diagnosis of lung lesions in children with case series reporting up to >95% diagnostic yield in critically ill children (9,10). However, its numerous associated comorbidities (thoracostomy drainage for invariable pneumothoraces, respiratory deterioration, and intrapulmonary hemorrhage, to name a few) have allowed percutaneous image-guided methods (both CT and ultrasonography) to gain favor as a minimally invasive alternative and thus increasing the use of these image-guided methods (11–13). The complication rates of percutaneous image-guided biopsy have also been shown to be lower than those with surgical biopsy; for example, although pneumothoraces do occur in 10%–15% of percutaneous procedures, they require chest tube placement in only 2% of cases with the overall complication rate lower than in open surgical biopsy (14). In the present study, there was a statistically significant lower rate of complications in patients undergoing lung biopsy with ultrasonography guidance (P ¼ .02), which could potentially be due to all procedures being done with ultrasonography only that were pleural-based lesions (n ¼ 41), thus resulting in little to no lung parenchyma being traversed by the biopsy needle. In an effort to improve the diagnostic yield of percutaneous lung biopsies performed at the authors’ institution, a cytopathologist has been present during these procedures since 2013. The role of the cytopathologist is to assess the FNA samples prior to core needle biopsy. For staining and immediate cytologic assessment to be possible, the FNA samples must be of adequate cellularity. If the initial samples demonstrate poor cellularity, the biopsy needle can be redirected into a more cellular component of the lesion and thereby improve the ability of any subsequent FNA and core biopsies to yield a diagnostic sample. Although a full evaluation of the effects of cytopathology assistance during lung biopsies is outside the scope of this study, diagnostic yields likely have improved with this immediate evaluation of tissue adequacy. However, in the present study, the statistically significant differences in complication rates of patients undergoing biopsy with FNA (P ¼ .02) were apparent. In the study by Cahill et al. (3), FNA was performed in 19 patients
(25%) with no major complications after using this technique. In the present study, FNA was performed in 18 patients (25%), accounting for 6 major complications. There was an increased mean number of passes (mean, 6; range, 1–20) in this study versus that by Cahill et al. (3) (mean, 5; range, 1– 8). The larger number of major complications in the present study could possibly be attributed to the increased number of passes required to obtain an adequate cellular sample. Despite these efforts, the overall diagnostic yield in the present study remains low. A previous study of 75 pediatric patients reported a 91% diagnostic yield of lung lesions biopsied (3). The study by De Bazelaire et al. (14) demonstrated diagnostic yields as high as 90% in immunocompetent patients and as low as 40% in immunosuppressed patients with aspergillus or mucormycosis (15). In the study by Cahill et al. (3), 48% of the patient population was immunosuppressed. Both the study by Cahill et al. (3) and the present study have lower numbers of patients biopsied with a diagnosis of infectious cause (39% and 48%, respectively) and comparable biopsies performed with fungal infection (55% and 54%, respectively) (3). Lesions diagnosed with infectious cause (n ¼ 35) accounted for 63% of the major complications and 50% of the minor complications in this study. These findings contrast markedly with the results of Cahill et al. (3), again due to the differences in number of patients biopsied for infection assessment. Although there are few differences in the overall rates of minor complications in the present study compared to those in the study by Cahill et al. (3) (19% vs. 28%, respectively), the disparity in major complication rate (11% vs. 1.5%, respectively) is due almost entirely to the 6 major complications that occurred in the much larger subset of patients whose infections were biopsied. Because neither the technical components of the biopsies (other than biopsy modality) nor patient factors (age, weight, hemoglobin, INR, ANC, and so forth) have a statistically significant effects on the complication rates, the principal reason for the higher number of complications among patients biopsied might have been due to the aforementioned higher rate of infection. Although lung biopsy cultures for bacterial infection are close to 100%, microbiological results for fungal infection range from 30% to 70% (15). Pulmonary fungal lesions are highly angio-invasive, with vascular invasion leading to central thrombosis and inflammatory hypervascularity of the periphery, thereby greatly increasing the risk of a bleeding complication during percutaneous biopsy (16). In patients with invasive
98 ▪ Percutaneous Lung Biopsy in Immunocompromised Pediatrics
pulmonary aspergillosis after bone marrow transplantation, the risks of hemorrhage and death are as high as 82%, with multicenter case fatality rates ranging between 69% and 85% for invasive pulmonary fungal disease (9,17,18). Due to the high morbidity of pulmonary fungal infection, children who are unable to undergo subsequent surgical resection often do not survive (18). In both of the present patients’ deaths, massive periprocedural pulmonary hemorrhage leading to cardiopulmonary arrest was the principal cause of death. Although no diagnosis was ever obtained in the 12-year-old patient, an autopsy performed on the 2-yearold child affirmed angio-invasive pneumocystis fungi. Although fungal infections specifically remain highly suspicious for increased complication rates, in the present study this could not be concluded with statistical significance. The present study is limited by its retrospective nature and the variances in operator technique, experience, and equipment between the different pediatric interventional radiologists over the long study period. Incomplete patient data (primarily laboratory data) in the electronic medical records prior to 2006 (the year the authors’ institution converted from paper records to an electronic medical system) also limits the study. Additionally, the introduction of cytopathology present during the procedure, thus increasing the use of FNA, in 2013 limits the comparative analysis on the rate of complication. Due to the large number of inconclusive diagnoses of lung lesions in this study, the association between diagnoses and complication may not be generalizable to the whole cohort. This was a small-cohort study. Because it was limited by a small sample size, a multivariate logistic regression analysis was not performed to evaluate the independent contribution of potential risk factors to the complication adjusting for other variables. In conclusion, image-guided percutaneous lung biopsy can be helpful in isolating an organism to tailor therapy. Biopsies performed in immunosuppressed patients have an elevated complication risk of up to 30%, demonstrate lower diagnostic yield and increased mortality, which should warrant detailed discussion with the primary team and family.
Cleveland et al ▪ JVIR
REFERENCES 1. Lim GY, Newman B, Kurland G, et al. Post-transplantation lymphoproliferative disorder: manifestations in pediatric thoracic organ recipients. Radiology 2002; 222:699–708. 2. Larson DB, Molvin LZ, Wang J, Chan FP, Newman BFD. Pediatric CT quality management and improvement program. Pediatr Radiol 2014; 44(Suppl 3):519–524. 3. Cahill AM, Baskin KM, Kaye RD, Fitz CRTR. CT-guided percutaneous lung biopsy in children. J Vasc Interv Radiol 2004; 15:955–960. 4. Wilkinson AG, Paton JYGN. CT-guided 14-G cutting needle lung biopsy in children: safe and effective. Pediatr Radiol 1999; 29:514–516. 5. Kropshofer G, Kneer A, Edlinger M, et al. Computed tomography guided percutaneous lung biopsies and suspected fungal infections in pediatric cancer patients. Pediatr Blood Cancer 61:1620–1624. 6. Hoffer FA, Gow KFP. Accuracy of percutaneous lung biopsy for invasive pulmonary aspergillosis. Pediatr Radiol 2001; 31:144–152. 7. Sacks D, McClenny TE, Cardella JF LC. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 14(9 Pt 2):S199– 202. 8. Saugier-Veber P, Devergie ASA. Epidemiology and diagnosis of invasive pulmonary aspergillosis in bone marrow transplant patients: results of a 5 year retrospective study. Bone Marrow Transpl 1993; 12: 121–124. 9. Steinberg R, Freud E, Ben-Ari J, et al. Open lung biopsy: successful diagnostic tool with therapeutic implication in the critically ill paediatric population. Acta Paediatr 1998; 87:945–948. 10. Stefanutti D, Morais L, Fournet J, et al. Value of open lung biopsy in immunocompromised children. J Pediatr 2000; 137:165–171. 11. Hayes-Jordan A, Benaim ERS. Open lung biopsy in pediatric bone marrow transplant patients. J Pediatr Surg 2002; 37:446–452. 12. Hayes-Jordan A, Connolly B, Temple M, et al. Image-guided percutaneous approach is superior to the thoracoscopic approach in the diagnosis of pulmonary nodules in children. J Pediatr Surg 2003; 38: 745–748. 13. Lucidarme O, Howarth N, Finet JF. Intrapulmonary lesions: percutaneous automated biopsy with a detachable, 18-gauge, coaxial cutting needle. Radiology 1998; 207:759–765. 14. De Bazelaire C, Coffin A, Cohen-Zarade S, et al. CT-guided biopsies in lung infections in patients with haematological malignancies. Diagn Interv Imaging 2013; 94:202–215. 15. Segal BH. Current approaches to diagnosis and treatment of invasive aspergillosis. Am J Respir Crit Care Med 2006; 183:707–717. 16. Abbasi S, Shenep JL, Hughes WTFP. Aspergillosis in children with cancer: a 34-year experience. Clin Infect Dis 1999; 9:1210–1219. 17. Groll AH, Kurz M, Schneider W, Witt V, Schmidt H, Schneider MSD. Fiveyear-survey of invasive aspergillosis in a paediatric cancer centre. Epidemiology, management and long-term survival. Mycoses 1999; 42(7–8): 431–442. 18. Bond SJ, Lee DJ, Stewart DL BJ. Open lung biopsy in pediatric patients on extracorporeal membrane oxygenation. J Pediatr Surg 1996; 31: 1376–1378.
Percutaneous Lung Biopsy in Immunocompromised Pediatric Patients Heather Cleveland, BSRS, Alex Chau, MD, Zachary Jeng, MD, Gregory Gardner, MD, Raphael Yoo, MD, Wei Zhang, PhD, and Jose Hernandez, MD
ABSTRACT Purpose: To determine the diagnostic yield and safety of image-guided lung biopsies in immunocompromised pediatric patients. Materials and Methods: This was a retrospective pediatric cohort study conducted from June 2000 to April 2017. Subjects were 0–17 years of age (median, 10 years of age). There were 46 males (48%). A total of 73 consecutive image-guided lung biopsies were performed in 68 patients (weight range, 4.9–97.3 kg [median, 25.3 kg]). The indication for biopsy was to isolate an organism to tailor medical therapy. All patients were immunocompromised with an underlying history of bone marrow transplantation (n ¼ 50), primary immunodeficiency (n ¼ 14), and solid organ transplantation (n ¼ 4). Patient and technical factors were analyzed for rates of complication. Results: Overall diagnostic yield was 43 of 73 patients (60%). There were 14 minor (19%) and 8 major (11%) complications. Major complications included pneumothorax or hemoptysis requiring intervention (n ¼ 6), and death (n ¼ 2). The histological diagnosis was an infectious cause in 5 of 8 major complications (63%). There were statistically significant differences between the rates of complications with the imaging modality used (P ¼ .02) and the use of fine needle aspiration (P ¼ .02). Conclusions: Image-guided percutaneous lung biopsy can be helpful in isolating an organism to tailor therapy. Biopsies performed in immunosuppressed patients result in an elevated complication risk of up to 30% and demonstrate lower diagnostic yield and increased mortality, which should warrant detailed discussion with the primary team and family.
Lung lesions in children can be caused by a variety of pathological processes including infection and inflammation, primary or secondary malignancy, and post-transplantation lymphoproliferative disease in immunosuppressed patients (1). With continued refinement and advances in dose reduction for cross-sectional imaging in children, the request for direct tissue sampling of these increasingly commonly detected lesions has grown (2). Although percutaneous image-guided biopsy offers many advantages over conventional open, and more recently, video- assisted surgical biopsy or endoscopic bronchoalveolar lavage, historical referral patterns, and competition have resulted in more extensive
From the Department of Interventional Radiology (H.C., A.C., G.G., R.Y., W.Z., J.H.) Texas Children’s Hospital, 6621 Fannin Street, Houston, TX 77030-2608; and the Department of Interventional Radiology (A.C., Z.J., G.G., R.Y., J.H.), Baylor College of Medicine, Houston, Texas. Received August 25, 2018; final revision received July 17, 2019; accepted July 18, 2019. Address correspondence to J.H.; E-mail: [email protected]
discussion of surgical rather than percutaneous options in the pediatric reports (3). Consequently, although image-guided percutaneous lung biopsy is well established in adult patients, comparatively less evidence exists for this procedure in children (4,5). In the limited pediatric reports, percutaneous lung biopsies are generally considered safe, with a high diagnostic accuracy (3,6). Although 1 study has documented the risk of bleeding as high as 46% for percutaneous biopsies in children with aspergillosis, which is especially prevalent among immunosuppressed children, the largest series to date has reported a low rate of complications (only 1 major complication in 64 patients), with nearly all studies documenting a high diagnostic accuracy of 85% to 100% (3,7). However, no research exists assessing differences in safety and diagnostic accuracy for percutaneous lung biopsies performed in children with immunosuppression. Thus, the purpose of this study was to review a single-center experience with image-guided percutaneous lung biopsy in immunosuppressed children to determine procedural safety and efficacy.
None of the authors have identified a conflict of interest. From the SIR 2018 Annual Scientific Meeting.
MATERIALS AND METHODS
© SIR, 2019
Demographics
J Vasc Interv Radiol 2020; 31:93–98
This was a retrospective Institutional Review Boardapproved study. The Epic electronic medical records (Epic
https://doi.org/10.1016/j.jvir.2019.07.016
94 ▪ Percutaneous Lung Biopsy in Immunocompromised Pediatrics
Systems, Verona, Wisconsin) for a single tertiary pediatric referral center were reviewed to identify all percutaneous image-guided lung biopsies performed from June 2000 to April 2017 in patients between the ages of 0 and 17 years (median, 10 years). Patient weight ranged from 4.9 to 97.3 kg (median, 25.3 kg) (Table 1). Demographic information, comorbidities, diagnostic imaging, technical details of the procedure, pathology reports, laboratory data, surgical records, procedural complications, and the postprocedure course until patient discharge were obtained for all patients through chart review. The initial percutaneous biopsy diagnostic adequacy and final diagnosis were determined based on pathology reports. Adequacy was determined by confirmation of histological sample and ability to diagnose the infection by the pathologist. A total of 81 consecutive biopsy procedures were performed in 76 immunocompromised patients (50 males and 26 females) referred for biopsy of indeterminate lesions or masses identified on prior imaging. Excluded from this study were patients in whom there was a suspicion of malignancy based on clinical presentation (n ¼ 8). The remaining 73 procedures in 68 patients were performed after preprocedure noncontrast chest computed tomography (CT) (n ¼ 73). Lesions were either pleural (n ¼ 61) or parenchymal (n ¼ 12) based. Pleural-based lesions were defined as having contact with the pleural surface. Biopsy specimens were sent for microbiological analysis to isolate an organism to guide treatment (n ¼ 73) and pathologic analysis (n ¼ 18) in which a fellowship-trained pediatric cytopathologist was present to evaluate the specimen. All biopsies were performed in immunocompromised patients with an absolute neutrophil count (ANC) <1.5 103 UL of cells receiving chemotherapy or those receiving immunosuppressive medications (most commonly tacrolimus or high-dose prednisone).
Procedure Preprocedural laboratory values were obtained in all patients, with a required minimum platelet count 50,000/μL and an international normalized ratio (INR) 1.5 with fresh frozen plasma and platelet transfusions administered as necessary. There was no cutoff value for preprocedural ANC levels (based on the normal pediatric range of 1.5–8 [1,000/ mm3]). A fellowship-trained pediatric anesthesiologist was present throughout each procedure to administer monitored anesthesia care or general anesthesia. Complications were graded based on the SIR classification system (8). The preprocedural cross-sectional imaging for each patient was reviewed to determine the optimal positioning and approach for percutaneous biopsy. The choice of needle, gauge (20 G), throw length, number of passes (standard of 4 passes, 2 FNAs, and 2 core samples), and technique (coaxial vs. tandem needle) was determined at the time of the procedure by the pediatric interventional radiologist. Lesion sizes (ranging from 10–40 mm, based on the longest
Cleveland et al ▪ JVIR
Table 1. Patient, Technical, and Lesion Characteristics (n ¼ 73) Characteristic % of females Age (y)* Weight (kg)* Height (cm)* Body mass index*
Total 34 10 (0-17) 25.3 (4.9-97.3) 127.5 (59-172) 17.3 (13.1-33.6)
Laboratory values Absolute neutrophil count (1,000/mm3)*
3.3 (0-19)
International normalized ratio*
1.1 (0.9-1.5)
Hemoglobin (g/dL)* Technical factors
9.2 (3.6-13.1)
CT-guided
27
US guided
41
Combined CT and US guidance Interventionalist experience (y)
5 16 (7-30)
Lesion characteristics Size (mm) % Centrally located
25 (5-55) 16
*Values are median.
axis of the lesion), location (central, nonpleural-based vs. peripheral, pleural-based), and presence of consolidation (ill-defined homogeneous opacities with obscuration of vessels and alveolar air bronchograms vs. lesions with some ground glass components) were determined based on preprocedural or intraoperative imaging. Procedures were performed by 7 interventional radiologists with experience ranging from 7 to 30 years (median, 16 years). All but 1 operator completed training in a dedicated pediatric interventional radiology fellowship because this training was not part of routine practice when he began practicing in the 1990s. Beginning in 2013, a cytopathologist was present during all lung biopsy cases to evaluate the FNA samples prior to core biopsy. Preliminary diagnosis was often obtained from the FNA. Core biopsies were invariably requested to provide adequate tissue for further immunologic, histologic, cytologic, or pathologic evaluation. All biopsies were sent for microbiologic and/or pathological evaluation. Ultrasonography-guided biopsies were performed using free-hand technique, with placement of the ultrasonography probe along the longest axis of the lesion at the most appropriate intercostal space. All procedures were performed during the respiratory phase allowing for the best visualization of the lesion, most often at end-expiration (n ¼ 39). All lesions were biopsied through the pleural surface. In CT-guided biopsies, patients were positioned according to the optimal approach based on preprocedural imaging. Patient positioning included prone (n ¼ 18), supine (n ¼ 6), and right lateral decubitus (n ¼ 3). All biopsies were performed using coaxial needle technique (median passes, 5; range, 1–20). Tract embolization was not performed.
Volume 31 ▪ Number 1 ▪ January ▪ 2020
95
Table 2. Lung Biopsy Tissue Diagnosis and Complications (n ¼ 73) Diagnosis
Biopsy with Complication (n ¼ 22)
Biopsy without Complication (n ¼ 51)
Total (N ¼ 73)
Fungal
4 minor 5 major
10
19
Bacterial
1 minor
11
12
Viral Malignancy
2 minor 2 minor 1 major
2 4
4 7
Sequela of drug reaction
1 minor
Inconclusive or Inadequate
4 minor 2 major
1 24
30
A chest radiograph of all patients was taken 2 hours after the procedure to assess for complications. If patients showed signs and symptoms of complications such as pneumothorax or bleeding, the imaging interval was shortened. When biopsies were preformed using CT guidance, a postprocedural chest CT was obtained immediately after needle removal to assess for complications.
Statistical Analysis Statistical analysis was performed to compare the effects of various factors on the complication rate. Continuous variables were compared using the Wilcoxon rank test, and categorical variables were compared using the Fisher exact test.
RESULTS Four patients (1 female and 3 males) required repeat biopsies. Two of these children underwent 2 biopsies, both of different lung lesions. Two remaining patients underwent 3 biopsies, both of the same lesion subsequently and of a different lung lesion on the third biopsy. Of the 4 patients who had repeated biopsies, 3 underwent surgical biopsy for diagnosis, and 1 had a conclusive percutaneous biopsy. Procedures were performed with the patient under monitored anesthesia care (n ¼ 25, 34%) or general anesthesia (n ¼ 48, 66%). Core biopsy samples were obtained with
16-G to 20-G Cook (n ¼ 38; Cook Medical, Bloomington, Indiana) and BioPince (n ¼ 35; Argon Medical Devices, Frisco, Texas) devices. FNA was performed using 20-G to 25-G Chiba needles (n ¼ 18; Cook Medical). When 5 FNA passes were performed, the median was 6 passes (range, 5–20 passes). When 4 FNA passes were performed, the median was 4 passes (range, 1–4). Core biopsies alone were obtained in a majority of the cases (n ¼ 55, 75%). FNA was performed in conjunction with core needle biopsy (n ¼ 18, 22%) beginning in 2013. Pleural-based lesions (n ¼ 61, 84%) were preferentially biopsied by using ultrasonography guidance if adequately visualized (n ¼ 41, 56%). Adjunctive CT guidance (n ¼ 5, 7%) was used if pleural-based lesions were unable to be fully visualized with ultrasonography. The remaining pleural-based and parenchymal lesions were biopsied under CT guidance (n ¼ 27, 37%) (Table 1). All biopsies were performed in immunocompromised patients with underlying conditions including post-bone marrow transplantation (n ¼ 52, 71%), primary immunodeficiency (n ¼ 16, 22%), or post-solid organ transplantation (n ¼ 5, 7%). Diagnostic yield, where histological or microbiological diagnosis was successfully obtained, was 43 of 73 (60%). The final diagnoses for biopsies performed for each indication are listed in Table 2. Of the inconclusive or inadequate biopsies, patients were empirically treated with antifungal agents and broad spectrum antibiotics (n ¼ 8); underwent surgical biopsy (n ¼ 5) or bronchoalveolar lavage (n ¼ 1); had a repeat biopsy that was conclusive (n ¼ 1); received tailored therapy based on a fungal biopsy sample of a liver lesion (n ¼ 1) or died during biopsy (n ¼ 1). There were no progress notes available for the remaining patients (n ¼ 13). Overall, complications occurred in 22 procedures (30%). A total of 14 minor (SIR categories A and B) complications (19%) occurred that required no further intervention or escalation in care. Minor complications included hemoptysis and bleeding (n ¼ 5); small, asymptomatic pneumothorax (n ¼ 7); hemothorax (n ¼ 1); and hemopneumothorax (n ¼ 1). A total of 8 major complications (10%) (Table 3) occurred, including pneumothorax requiring chest tube placement (n ¼ 3) (SIR categories C and D); hemoptysis requiring extended intubation, and intensive care unit care
Table 3. Major Complications Age
Tissue Diagnosis
Imaging Modality
Complication
1
Patient
6
Fungal
CT
Pneumothorax, chest tube
C
2 3
8 17
Malignancy Fungal
US CT
Pneumothorax, chest tube Pneumothorax, PICU
C D
4
2
Fungal
US
Hemoptysis, PICU
D
5
3
Fungal
US
Hemoptysis, PICU
D
6
1
Inconclusive
CT
Hemorrhage, PICU
D
7
2
Fungal
US/CT
Death
F
8
3
Inconclusive
US
Death
F
BAL ¼ bronchoalveolar lavage; PICU ¼ pediatric intensive care unit.
SIR Classification
96 ▪ Percutaneous Lung Biopsy in Immunocompromised Pediatrics
Table 4. Statistical Analysis of Categorical Factors’ Effect on the Complication Rate Categorical Factors
Number of Patients
Complication Count (% of N)
5 y
24
11 (45.8)
>5 y
49
11 (22.4)
CT
27
13 (48.1)
US
41
7 (17.1)
Age
.06
Modality
US and CT Gauge
.02
5
1.0 33
10 (30.3)
18
38
12 (31.6)
16
2
0
Consolidation
21
9 (42.9)
Lesion
52
Lesion or consolidation
.16 13 (25) .50 9
2 (22.2)
11-19
12
6 (50)
20-29
28
9 (32.1)
30-39
12
3 (25)
40
12
2 (16.7)
4
26
8 (30.8)
5 Multiple
39 8
10 (25.6) 4 (50.0)
With
18
10 (55.6)
Without
55
12 (21.8)
Central
12
5 (41.7)
Peripheral
61
17 (27.9)
61
19 (31.1)
12
3 (25)
<1.5
18
2 (11.1)
1.5 – 8
38
15 (39.5)
>8
14
5 (35.7)
29 43
11 (37.9) 11 (25.6)
50
13
4 (30.8)
51-99
18
100
41
16 (39)
10
43
10 (23.3)
10.1
29
12 (41.4)
Number of passes
.38
FNA
.02
Location of lesion
Pleural vs parenchymal Pleural Parenchymal
.49
1.0
ANC (3 missing)
.09
INR (1 missing) 1 1.1-1.5
.30
Platelets (1 missing)
.09 2 (11.1)
Hemoglobin (1 missing)
Table 4. Statistical Analysis of Categorical Factors’ Effect on the Complication Rate (continued) Categorical Factors Diagnostic Yield Adequate Inadequate Inconclusive
Number of Patients
Complication Count (% of N)
43 7 23
14 (32.6) 3 (42.9) 5 (21.7)
P Value .44
ANC ¼ absolute neutrophil count; FNA ¼ fine needle aspiration; INR ¼ international normalized ratio.
2 (40)
20
Size of lesion (mm) 10
P Value
Cleveland et al ▪ JVIR
.12
continued
(n ¼ 3) (SIR categories C and D); and death (n ¼ 2) (SIR category F) from respiratory arrest presumably secondary to asphyxiation from pulmonary hemorrhage. The histological diagnosis was determined to be caused by infection in 5 of 8 patients (63%) with major complications. Six of the 8 patients (75%) with major complications had biopsies performed with FNA, of which 4 were found to be inconclusive. In the 3 patients requiring chest tube placement, all chest tubes were removed within 3 days after the procedure, and 1 patient was able to have the chest tube removed within hours of admission to the postanesthesia care unit. In the 3 patients with hemoptysis requiring extended intubation, all required admission to the pediatric intensive care unit. All 3 patients were successfully extubated within 2 days after the procedure and had uneventful return to the regular inpatient floor. Several technical factors were assessed to determine the effect on the complication rate. There was a statistically significant correlation with the complication rate with the use of ultrasonography versus CT (P ¼ .02) (Table 4). There was a lower rate of complications for ultrasonography-guided procedures (17%). All 41 of the procedures performed using ultrasonography only were pleural-based lesions. There were statistically significant differences in complications with the use of FNA during biopsy (P ¼ .02). Of the patients who had conclusive diagnoses, fungal infection had a 47% (9 of 19) complication rate, whereas the other diagnoses had a 29% (7 of 24) complication rate (P ¼ .34). Several patient demographics and laboratory values were also assessed to determine their effect on the complication rate. Neither patient age, weight, height, nor body mass index were statistically significant (P ¼ .13, P ¼ .17, P ¼ .06, P ¼ .1, respectively)(Table 5). The preprocedure laboratory values were found to be insignificant (Table 4). Two deaths occurred due to complications experienced during biopsy. The first patient, a 2-year-old girl with graftversus-host disease underwent biopsy to determine possible fungal infection (the initial FNA samples were inconclusive). The patient died at the end of the procedure and was found to have pneumocystis fungal hyphae with angioinvasion on autopsy. The second patient, a 12-year-old boy with a history of relapsed acute lymphoblastic leukemia underwent biopsy to evaluate suspected multifocal
Volume 31 ▪ Number 1 ▪ January ▪ 2020
97
Table 5. Statistical Analysis of Continuous Factors’ Effect on the Complication Rate Continuous Factor Age
Overall (N ¼ 73)
Without Complication (n ¼ 51)
With Complication (n ¼ 22)
P Value
8.7 ± 5.6
9.3 ± 5.1
7.1 ± 6.5
.13
Weight (kg)
32.8 ± 21.1
Height (cm)
126.1 ± 32.2
33 ± 17.4 131.4 ± 29
32.3 ± 28.1
.17
114.2 ± 36.4
.06
Body mass index
18.6 ± 4.5
17.8 ± 3.3
20.6 ± 6.2
.10
Years of experience
19.5 ± 8.1
20 ± 7.6
18.5 ± 9.3
.14
pneumonia that had worsened despite prolonged antibacterial and antifungal therapy. The patient died and was found to have nonspecific acute lung injury, with a diagnosis that was never able to be precisely determined on autopsy.
DISCUSSION Open biopsy is used for establishing a histopathological diagnosis of lung lesions in children with case series reporting up to >95% diagnostic yield in critically ill children (9,10). However, its numerous associated comorbidities (thoracostomy drainage for invariable pneumothoraces, respiratory deterioration, and intrapulmonary hemorrhage, to name a few) have allowed percutaneous image-guided methods (both CT and ultrasonography) to gain favor as a minimally invasive alternative and thus increasing the use of these image-guided methods (11–13). The complication rates of percutaneous image-guided biopsy have also been shown to be lower than those with surgical biopsy; for example, although pneumothoraces do occur in 10%–15% of percutaneous procedures, they require chest tube placement in only 2% of cases with the overall complication rate lower than in open surgical biopsy (14). In the present study, there was a statistically significant lower rate of complications in patients undergoing lung biopsy with ultrasonography guidance (P ¼ .02), which could potentially be due to all procedures being done with ultrasonography only that were pleural-based lesions (n ¼ 41), thus resulting in little to no lung parenchyma being traversed by the biopsy needle. In an effort to improve the diagnostic yield of percutaneous lung biopsies performed at the authors’ institution, a cytopathologist has been present during these procedures since 2013. The role of the cytopathologist is to assess the FNA samples prior to core needle biopsy. For staining and immediate cytologic assessment to be possible, the FNA samples must be of adequate cellularity. If the initial samples demonstrate poor cellularity, the biopsy needle can be redirected into a more cellular component of the lesion and thereby improve the ability of any subsequent FNA and core biopsies to yield a diagnostic sample. Although a full evaluation of the effects of cytopathology assistance during lung biopsies is outside the scope of this study, diagnostic yields likely have improved with this immediate evaluation of tissue adequacy. However, in the present study, the statistically significant differences in complication rates of patients undergoing biopsy with FNA (P ¼ .02) were apparent. In the study by Cahill et al. (3), FNA was performed in 19 patients
(25%) with no major complications after using this technique. In the present study, FNA was performed in 18 patients (25%), accounting for 6 major complications. There was an increased mean number of passes (mean, 6; range, 1–20) in this study versus that by Cahill et al. (3) (mean, 5; range, 1– 8). The larger number of major complications in the present study could possibly be attributed to the increased number of passes required to obtain an adequate cellular sample. Despite these efforts, the overall diagnostic yield in the present study remains low. A previous study of 75 pediatric patients reported a 91% diagnostic yield of lung lesions biopsied (3). The study by De Bazelaire et al. (14) demonstrated diagnostic yields as high as 90% in immunocompetent patients and as low as 40% in immunosuppressed patients with aspergillus or mucormycosis (15). In the study by Cahill et al. (3), 48% of the patient population was immunosuppressed. Both the study by Cahill et al. (3) and the present study have lower numbers of patients biopsied with a diagnosis of infectious cause (39% and 48%, respectively) and comparable biopsies performed with fungal infection (55% and 54%, respectively) (3). Lesions diagnosed with infectious cause (n ¼ 35) accounted for 63% of the major complications and 50% of the minor complications in this study. These findings contrast markedly with the results of Cahill et al. (3), again due to the differences in number of patients biopsied for infection assessment. Although there are few differences in the overall rates of minor complications in the present study compared to those in the study by Cahill et al. (3) (19% vs. 28%, respectively), the disparity in major complication rate (11% vs. 1.5%, respectively) is due almost entirely to the 6 major complications that occurred in the much larger subset of patients whose infections were biopsied. Because neither the technical components of the biopsies (other than biopsy modality) nor patient factors (age, weight, hemoglobin, INR, ANC, and so forth) have a statistically significant effects on the complication rates, the principal reason for the higher number of complications among patients biopsied might have been due to the aforementioned higher rate of infection. Although lung biopsy cultures for bacterial infection are close to 100%, microbiological results for fungal infection range from 30% to 70% (15). Pulmonary fungal lesions are highly angio-invasive, with vascular invasion leading to central thrombosis and inflammatory hypervascularity of the periphery, thereby greatly increasing the risk of a bleeding complication during percutaneous biopsy (16). In patients with invasive
98 ▪ Percutaneous Lung Biopsy in Immunocompromised Pediatrics
pulmonary aspergillosis after bone marrow transplantation, the risks of hemorrhage and death are as high as 82%, with multicenter case fatality rates ranging between 69% and 85% for invasive pulmonary fungal disease (9,17,18). Due to the high morbidity of pulmonary fungal infection, children who are unable to undergo subsequent surgical resection often do not survive (18). In both of the present patients’ deaths, massive periprocedural pulmonary hemorrhage leading to cardiopulmonary arrest was the principal cause of death. Although no diagnosis was ever obtained in the 12-year-old patient, an autopsy performed on the 2-yearold child affirmed angio-invasive pneumocystis fungi. Although fungal infections specifically remain highly suspicious for increased complication rates, in the present study this could not be concluded with statistical significance. The present study is limited by its retrospective nature and the variances in operator technique, experience, and equipment between the different pediatric interventional radiologists over the long study period. Incomplete patient data (primarily laboratory data) in the electronic medical records prior to 2006 (the year the authors’ institution converted from paper records to an electronic medical system) also limits the study. Additionally, the introduction of cytopathology present during the procedure, thus increasing the use of FNA, in 2013 limits the comparative analysis on the rate of complication. Due to the large number of inconclusive diagnoses of lung lesions in this study, the association between diagnoses and complication may not be generalizable to the whole cohort. This was a small-cohort study. Because it was limited by a small sample size, a multivariate logistic regression analysis was not performed to evaluate the independent contribution of potential risk factors to the complication adjusting for other variables. In conclusion, image-guided percutaneous lung biopsy can be helpful in isolating an organism to tailor therapy. Biopsies performed in immunosuppressed patients have an elevated complication risk of up to 30%, demonstrate lower diagnostic yield and increased mortality, which should warrant detailed discussion with the primary team and family.
Cleveland et al ▪ JVIR
REFERENCES 1. Lim GY, Newman B, Kurland G, et al. Post-transplantation lymphoproliferative disorder: manifestations in pediatric thoracic organ recipients. Radiology 2002; 222:699–708. 2. Larson DB, Molvin LZ, Wang J, Chan FP, Newman BFD. Pediatric CT quality management and improvement program. Pediatr Radiol 2014; 44(Suppl 3):519–524. 3. Cahill AM, Baskin KM, Kaye RD, Fitz CRTR. CT-guided percutaneous lung biopsy in children. J Vasc Interv Radiol 2004; 15:955–960. 4. Wilkinson AG, Paton JYGN. CT-guided 14-G cutting needle lung biopsy in children: safe and effective. Pediatr Radiol 1999; 29:514–516. 5. Kropshofer G, Kneer A, Edlinger M, et al. Computed tomography guided percutaneous lung biopsies and suspected fungal infections in pediatric cancer patients. Pediatr Blood Cancer 61:1620–1624. 6. Hoffer FA, Gow KFP. Accuracy of percutaneous lung biopsy for invasive pulmonary aspergillosis. Pediatr Radiol 2001; 31:144–152. 7. Sacks D, McClenny TE, Cardella JF LC. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 14(9 Pt 2):S199– 202. 8. Saugier-Veber P, Devergie ASA. Epidemiology and diagnosis of invasive pulmonary aspergillosis in bone marrow transplant patients: results of a 5 year retrospective study. Bone Marrow Transpl 1993; 12: 121–124. 9. Steinberg R, Freud E, Ben-Ari J, et al. Open lung biopsy: successful diagnostic tool with therapeutic implication in the critically ill paediatric population. Acta Paediatr 1998; 87:945–948. 10. Stefanutti D, Morais L, Fournet J, et al. Value of open lung biopsy in immunocompromised children. J Pediatr 2000; 137:165–171. 11. Hayes-Jordan A, Benaim ERS. Open lung biopsy in pediatric bone marrow transplant patients. J Pediatr Surg 2002; 37:446–452. 12. Hayes-Jordan A, Connolly B, Temple M, et al. Image-guided percutaneous approach is superior to the thoracoscopic approach in the diagnosis of pulmonary nodules in children. J Pediatr Surg 2003; 38: 745–748. 13. Lucidarme O, Howarth N, Finet JF. Intrapulmonary lesions: percutaneous automated biopsy with a detachable, 18-gauge, coaxial cutting needle. Radiology 1998; 207:759–765. 14. De Bazelaire C, Coffin A, Cohen-Zarade S, et al. CT-guided biopsies in lung infections in patients with haematological malignancies. Diagn Interv Imaging 2013; 94:202–215. 15. Segal BH. Current approaches to diagnosis and treatment of invasive aspergillosis. Am J Respir Crit Care Med 2006; 183:707–717. 16. Abbasi S, Shenep JL, Hughes WTFP. Aspergillosis in children with cancer: a 34-year experience. Clin Infect Dis 1999; 9:1210–1219. 17. Groll AH, Kurz M, Schneider W, Witt V, Schmidt H, Schneider MSD. Fiveyear-survey of invasive aspergillosis in a paediatric cancer centre. Epidemiology, management and long-term survival. Mycoses 1999; 42(7–8): 431–442. 18. Bond SJ, Lee DJ, Stewart DL BJ. Open lung biopsy in pediatric patients on extracorporeal membrane oxygenation. J Pediatr Surg 1996; 31: 1376–1378.