Keywords CT-guided biopsy - lung metastasis - small pulmonary nodules
Introduction
Transthoracic biopsy plays an important role in the diagnosis and staging of primary
lung cancer.[1 ] Indeterminate lung nodules on imaging in patients with known malignancy need sampling
to rule out any underlying metastatic foci, especially when curative options are considered.[2 ]
[3 ] However, biopsy from lung lesions smaller than 2 cm in size is a challenging task.
In spite of availability of endobronchial biopsies (EBBs) and video-assisted thoracoscopic
(VATS) biopsies, computed tomography (CT)-guided transthoracic biopsy remains the
preferred method of choice for such smaller lesions.[3 ] CT biopsy has advantages over VATS and lobectomy since it is less invasive. Compared
to the EBB, the CT-guided biopsy samples are bigger. Core needle lung biopsy is preferred
over fine-needle aspiration cytology because of its higher diagnostic accuracy in
case of nonmalignant etiology, better tissue characterization, and option of additional
staining with immunohistochemistry markers for confirmation of malignant lesions.[4 ]
[5 ]
[6 ] The reported diagnostic accuracy of CT-guided core biopsy is 62 to 97% in lesions
larger than 2 cm.[7 ] Successful core biopsy of smaller lesions remains challenging, especially when the
lesions are situated in the lower lobe. There is concern for increased rate of complications
as the size of the lesion decreases. The primary aim of this study was to report the
diagnostic accuracy and rate of complications of CT-guided transthoracic core biopsy
of pulmonary nodules measuring less than 2 cm in long-axis diameter. The secondary
aim was to identify risk factors associated with clinically significant postprocedure
complications, including pneumothorax and pulmonary hemorrhage.
Materials and Methods
This study was a retrospective chart review and the necessity for an informed consent
was waived by the Institutional Review Board. Between January 2015 and December 2018,
we collected data on consecutive CT-guided core lung biopsies on 169 patients who
had a solitary lung lesion smaller than or equal to the size of 2 cm. Repeat biopsies
were not included. The biopsies were performed by two interventional radiologists
with a minimum of 2 years of experience with CT-guided biopsy. The clinical decision
to perform a biopsy was made by the referring physicians and the feasibility and approach
of the biopsy were determined by the interventional radiologists. All biopsies were
performed as a daycare procedure. Informed procedural consent was obtained from the
patient and one first-degree relative prior to the biopsy procedure. The prothrombin
time-to-international normalized ratio (PT-INR) and platelet count were checked before
the biopsy procedure. Necessary corrections were made before the procedure in patients
with PT-INR > 1.5 and platelet count < 50,000/mm3 .
Biopsies were performed using two different 64-slice scanners (Discovery, General
Electric and Biograph, Siemens Medical Systems). The prebiopsy CT images of the patients
were reviewed to determine the patient position during the biopsy. The patient was
transferred to the CT table and placed in a predetermined position. The safest and
most accessible path for the biopsy needle was selected to avoid bones, large vessels,
and blebs/bullae. Fissures were avoided as much as possible. After taking all aseptic
precautions, approximately 10 mL of 2% lignocaine was injected into the skin and subcutaneous
tissue for local anesthesia. All the biopsies were performed using the coaxial method
(17-gauge introducer needle with 18G automatic biopsy needle and 19G coaxial introducer
with 20G automatic biopsy needle). Patients were instructed not to take deep breaths.
Under free breathing, the coaxial needle was introduced. Once the position of the
tip of the coaxial needle was deemed satisfactory, the inner stylet was removed, and
multiple tissue cores (minimum 2) were taken using an automatic biopsy gun. CT fluoroscopy
was not available in our institute. The number of samples varied and that depended
on the quality of the sample, bleeding, and the patient's condition. All tissue samples
were preserved in a 10% formaldehyde solution. Additional tissue cores were preserved
in normal saline where infective etiology was a differential diagnosis for the pulmonary
nodule under evaluation. A check CT scan was performed at the end of the procedure
to rule out pneumothorax or pulmonary hemorrhage. All patients were observed for at
least 2 hours in the recovery ward after the biopsy.
Individualized treatment decisions were made for any patient who developed a pneumothorax
depending upon the volume of air and the general condition of the patient. Patients
with small pneumothorax were monitored with a follow-up CT after 1 hour to check for
progression of the pneumothorax. Asymptomatic, small, and stable cases were not treated.
Enlarging or large pneumothoraces were treated with aspiration alone, if asymptomatic,
and with pigtail catheter insertion, if symptomatic. All symptomatic pneumothoraces
were treated with an 8F pigtail catheter insertion and drainage to a water-seal bag.
All patients with pneumothorax were reviewed after 24 hours with a chest radiograph.
In patients with pigtail drainage, after the radiograph showed complete lung expansion,
the pigtail tube was removed after clamping for 24 hours.
The severity of lung hemorrhage was divided into four grades. Grade 0 was defined
as no pulmonary hemorrhage, grade 1 as needle tract or perfocal hemorrhage < 2 cm
in width, grade 2 as hemorrhage > 2 cm in width but sublobar, and grade 3 as lobar
hemorrhage or greater, and grade 4 as hemothorax[8 ] ([Fig. 1 ]). Patients who had minor hemoptysis received conservative care. Patients with moderate
to large hemoptysis were admitted and given treatment based on the severity of their
symptoms. Intravenous tranexamic acid was administered to patients with mild to moderate
desaturation along with oxygen support. On the basis of merit, patients with gross
desaturation were intubated.
Fig. 1 Grades of hemorrhage. (A , C , and E ) Axial computed tomography (CT) sections showing lesions (arrows) before biopsy.
(B , D , and F ) Postbiopsy images showing grade 1, 2, and 3 hemorrhages, respectively (curved arrow).
The technical success of core needle biopsies (CNBs) was defined by availability of
a conclusive pathological diagnosis, either benign or malignant. CNB was deemed unsuccessful
if the pathology report was inconclusive due to a suboptimal or inadequate sample.
Data was collected regarding the final diagnosis (benign vs. malignant) of the successful
biopsies. Final diagnosis of the lesion was determined by either final pathology (in
malignant) or follow-up imaging (in biopsies that were negative for malignancy but
showed nonspecific benign findings). CNB findings of malignancy were considered true
positives if the histopathological findings were consistent with the known primary
malignancy or the subsequent clinical course was consistent with malignancy. Malignant
CNB findings were considered false positives when the subsequent clinical course was
inconsistent with malignancy or the surgical specimen contradicted the initial biopsy
specimen's histopathological finding. Benign CNB findings were considered true positives
when the lesion disappeared or decreased in size with or without antibiotics or the
lesion remained stable for at least 1 year after biopsy or there was surgical confirmation.
Benign CNB findings were considered false positives when the subsequent clinical course
was inconsistent with a benign diagnosis or surgical findings showed malignancy. The
inadequate or paucicellular CNB specimens were considered false negatives and none
of the lesions were classified as true negatives. Lesion size, location (relation to
pleura, lower lobe vs. nonlower lobe, central vs. peripheral), presence of emphysema
or fibrosis in background lung, patient position (prone vs. supine), needle size (18G
vs. 20G), and needle path (90 degrees to pleura vs. oblique) were recorded. Number
and degree of pulmonary hemorrhage and pneumothorax were recorded. Factors predicting
clinically significant pneumothorax and hemorrhage were examined by statistical analysis.
Results
A total of 169 nodules were biopsied (32–80 years) ([Table 1 ]). Size of the nodules ranged from 0.5 to 2.0 cm (average 1.4 ± 0.4 cm). The pathologically
reportable quality biopsy samples (success rate) were obtained in 92.9% (157/169)
cases and pathological inadequate or nondiagnostic results were obtained in 7.1% (12/169)
cases. The success rate for lesions larger than > 1 cm was 97.1% (101/104) and for
subcentimeter lesions it was found to be 86.2% (56/65).
Table 1
Patient characteristics, lesion variables, and procedure variables
Age (mean)
59
Sex
Male
140
Female
29
Complications
Pneumothorax
37/169 (21.9%)
Hemorrhage
149/169 (88%)
Grade 1
122 (72.2%)
Grade 2
16 (9.5%)
Grade 3
11 (6.5%)
Minor hemoptysis
4
Major hemoptysis
1
Malignant disease was reported in 60.4% (102/169) and benign pathologies were noted
in 32.5% (55/169) cases ([Table 2 ]). Malignancy was reported in 82 lesions that were larger than 1 cm and 20 lesions
less than 1 cm. A size cutoff value of 1 cm for diagnosis of malignancy yielded a
sensitivity of 80%, specificity of 44.7%, accuracy of 66%, positive predictive value
of 68.9%, and negative predictive value of 60%. Average size of malignant lesions
was significantly larger (1.49 ± 0.4 cm, median 1.5 cm); however, there was significant
overlap between the groups. Average size comparison between benign, malignant, and
indeterminate lesions is given in [Table 3 ] Pneumothorax was seen in 21.9% cases (37/169) ([Tables 4 ] and [5 ]). Out of 37 patients who developed pneumothorax, only 5.3% (9/169) patients required
treatment (8 needle aspirations, 1 pigtail catheter drainage). Grade 1 or perifocal
hemorrhage was common (72.2%, 122/169), but clinically significant grade 3 pulmonary
lobar hemorrhage was seen only in 6.5% cases (11/169). One patient required intensive
care unit (ICU) admission due to desaturation from aspiration of blood. Note that
11.8% cases (20/169) had no hemorrhage and 9.5% cases (16/169) had grade 2 or sublobar
hemorrhage. No patient developed hemothorax ([Tables 6 ] and [7 ]). No significant association of hemorrhage was observed with location, size of the
needle used, or size of the lesion.
Table 2
Histopathology findings
Malignant
102
Benign
55
Fungal infection
1
Organizing pneumonia
5
Caseating granuloma
4
Noncaseating granuloma
32
Other nonspecific chronic inflammation
12
Hamartoma
1
Inconclusive
12
Total
169
Table 3
Relationship between the diagnostic accuracy and size
Indeterminate
(Group 1)
Benign
(Group 2)
Malignant
(Group 3)
Number [percentage]
12.0 [7.1]
55.0 [32.5]
102.0 [60.4]
Average size (cm)
[mean ± SD (SE)]
1.01 ± 0.4 (0.12)
1.31 ± 0.45 (0.06)
1.49 ± 0.4 (0.04)
Median [IQR]
0.95 (0.7–1.15)
1.2 (0.9–1.75)
1.5 (1.2–1.9)
Abbreviations: ANOVA, analysis of variance; IQR, interquartile range; SD, standard
deviation; SE, standard error.
Note: One-way ANOVA: F statistic = 8.93 [p = 0.0002].
Table 4
Incidence of pneumothorax and variables
Explanatory factor
Pneumothorax (n = 37)
%
No pneumothorax (n = 132)
%
Subcentimetric
15
40.5
50
37.9
Not subcentimetric
22
59.5
82
62.1
Lower lobe
20
54.1
50
37.9
Not lower lobe
17
45.9
82
62.1
Abnormal lung
8
21.6
15
11.4
Normal lung
29
78.4
117
88.6
18G needle
19
51.4
38
28.8
20G needle
18
48.6
94
71.2
Prone position
25
67.6
67
50.8
Supine position
12
32.4
65
49.2
> 90 angle of needle
16
43.2
65
49.2
< 90 angle of needle
21
56.8
67
50.8
Subpleural nodule
13
35.1
40
30.3
Not subpleural
24
64.9
92
69.7
Central nodule
2
5.4
21
15.9
Peripheral nodule
35
94.6
111
84.1
Fissure in needle path
3
8.1
18
13.6
No fissure
34
91.9
114
86.4
Table 5
Multivariate analysis of explanatory factors predicting postbiopsy pneumothorax
Odds ratio (95% CI)
p -Value
Subcentimetric nodule
3.62 [1.40–9.33]
0.00777**
Lower lobe location
2.97 [1.01–8.74]
0.0485*
Background lung disease (emphysema, ILD)
3.01 [1.00–9.04]
0.0493*
Use of 18G needle
3.52 [1.52–8.16]
0.00334**
Prone position
0.81 [0.28–2.36]
0.695
Oblique needle angle
1.64 [0.71–3.81]
0.251
Subpleural location
1.07 [0.45–2.55]
0.887
Fissure in needle path
0.98 [0.25–3.87]
0.971
Abbreviations: CI, confidence interval; ILD, interstitial lung disease.
Note: ** Statistically significant.
Table 6
Incidence of significant pulmonary hemorrhage with other variables
Explanatory factor
No hemorrhage (n = 20)
%
Hemorrhage (n = 149)
%
Subcentimetric
1
5.0
31
20.8
Not subcentimetric
19
95.0
118
79.2
Lower lobe
7
35.0
63
42.3
Not lower lobe
13
65.0
86
57.7
Abnormal lung
2
10.0
21
14.1
Normal lung
18
90.0
128
85.9
18G needle
3
15.0
54
36.2
20G needle
17
85.0
95
63.8
Prone position
11
55.0
81
54.4
Supine position
9
45.0
68
45.6
> 90 angle of needle
10
50.0
76
51.0
< 90 angle of needle
10
50.0
73
49.0
Subpleural nodule
4
20.0
49
32.9
Not subpleural
16
80.0
100
67.1
Central
0
0.0
23
15.4
Peripheral
20
100.0
126
84.6
Pneumothorax
2
10.0
35
23.5
No pneumothorax
18
90.0
114
76.5
Note: There was no statistically significant correlation with variables.
Table 7
Relationship of chance of clinically significant hemorrhage and different factors
OR
95 CI
z
p
Subcentimetric
2.01
0.59–6.89
1.12
0.26
Lower lobe
1.36
0.40–4.65
0.49
0.62
18G gun
0.72
0.18–2.83
0.47
0.64
Prone
0.29
0.07–1.14
1.78
0.08
Supine
3.44
0.88–13.45
1.78
0.08
Subpleural
1.27
0.36–4.54
0.37
0.71
Peripheral
1.62
0.20–13.27
0.45
0.65
< 90
0.75
0.22–2.57
0.45
0.65
Fissure in path
1.62
0.33–8.10
0.59
0.55
Abbreviations: CI, confidence interval; OR, odds ratio.
Note: There was no statistically significant correlation with location, size, needle
size, or patient position.
Pneumothorax was more common with the use of a 17G coaxial needle and an 18G biopsy
needle (odds ratio 2.61, 95% confidence interval 1.24–5.51, p = 0.01) ([Tables 4 ] and [5 ]). Pneumothorax was also more common when the patient was placed in a prone position;
however, this was not found to be statistically significant. No statistically significant
correlation of development of pneumothorax with location or size of the lesion was
noted.
Except for one nondiagnostic sample that turned out to be malignant, all pathologically
benign and nondiagnostic samples showed benign course during follow-up. No false positive
cases were detected in our series. If we assume nondiagnostic samples as negative
for malignancy (as happens in normal practice), the overall sensitivity of CT-guided
biopsy for small lung nodules was 99% and specificity was 100%.
Discussion
CT scan is the work horse in the diagnosis and staging of multiple neoplastic lesions
resulting in an increased detection of smaller lung nodules that are suspicious for
metastases. Biopsy even in suspected metastatic disease is important, since in the
presence of oligometastatic disease, treatment with curative intent can be considered.[9 ]
[10 ]
Accuracy of CT-guided biopsy is well documented in the literature; however, the same
is not established for smaller lesions, that is, lesions smaller than 2 cm.[11 ] These lesions are technically difficult to target, which may result in increased
risk of complications such as pneumothorax and pulmonary hemorrhage. In the present
study, attempts were made to address these issues.
The present study shows the overall success rate of CT-guided biopsy for obtaining
diagnostic quality tissues from < 2 cm lesion is 91.6%, which falls in the range of
accuracy reported for lung lesions in general.[11 ]
[12 ] This suggests that the accuracy of CT-guided biopsy is not greatly influenced by
smaller sized lesions. Interestingly, the diagnostic accuracy of CT-guided biopsy
for subcentimeter-sized lesions is not greatly different from lesions measuring 1
to 2 cm in size (86% vs. 97%, respectively).
In our series, the size of the smallest lesion that was biopsied was 5 mm and that
of the largest was 2 cm. Nondiagnostic biopsy was reported in 7.1% (12) of the cases.
Majority (75%, 9 out of 12) of nondiagnostic biopsies were reported from lesions that
measured less than a centimeter. Interestingly, the majority (83.3%, 10 out of 12)
of nondiagnostic biopsies turned out to be benign diseases on follow-up and only two
cases turned out to be malignant.
Only 40% (20/50) of lesions less than 1 cm turned out to be malignant in contrast
to 68.9% (82/119) of lesions > 1 cm that turned out to be malignant. This finding
may suggest the larger the size of the lesion, the higher the chances of malignancy
([Table 8 ]).
Table 8
Group comparison of diagnostic accuracy with size
Nondiagnostic
Benign
Malignant
Subcentimetric (n = 65)
9.0 (13.8%)
29.0 (44.6%)
27.0 (41.5%)
Not subcentimetric (n = 104)
3.0 (2.9%)
26.0 (25%)
75.0 (72.1%)
Note: Mann–Whitney U test [two-tailed]: U = 2273, z score = 3.57 (p = 0.00034). There is no statistically significant association between diagnostic
accuracy and lesion size.
Except for one case, where a nondiagnostic report on pathology turned out to be malignant
on subsequent follow-up, the rest of the biopsy reports were consistent with clinical
course of the disease. Unlike larger lesions, while performing biopsy from a lesion
less than 2 cm in size, the entire lesion is likely to be sampled thus leaving very
little scope for sampling errors. This can explain the high accuracy of the pathological
biopsy reports.
Though the rate of occurrence of pneumothorax was higher (21.9%), patients who actually
required any active intervention were only 5.3% of the cases. Majority of pneumothoraces
were minor and stable, which were successfully managed conservatively. Higher rate
of pneumothorax most likely reflects the difficulty in targeting the lesion.[13 ] Larger gauge needles were associated with statistically significant pneumothorax
compared to smaller needles. No statistical significance could be demonstrated with
respect to location of the lesion (central or peripheral and upper or lower lobe,
or position of the patient).[14 ] Even though in our series the presence or absence of emphysema did not influence
the occurrence of pneumothorax, it is a safe practice to avoid the emphysematous areas
along the pathway of the needle.
Pulmonary hemorrhage was found to be very common (88%); however, most of them were
either tract hemorrhages or focal pulmonary hemorrhages not requiring any significant
intervention or management ([Fig. 1 ]). Grade 3 hemorrhage was seen in 6.5% cases (11/169). Two percent (4/169) had self-limiting
mild hemoptysis, which required no treatment. One patient had significant hemoptysis
who required ICU admission for 2 days. Unlike previous studies, no statistical significance
could be demonstrated between the occurrence of hemorrhage with the site of the lesion,
position of the patient, and gauge of the needle in our study.[14 ]
Targeting small pulmonary lesions can more often than not be very challenging. Lesions
located in the upper lobe are relatively easily targeted as compared to lower lobe
lesions, which tend to move with respiration.[15 ] Meticulous preprocedure planning plays a very important role for a successful biopsy.
Performing all manipulations in the biopsy gun in terms of site of entry and angulation
before piercing the pleura may help in significantly reducing complications.[15 ] The moving target may require extremely slow advancement of needle and manipulation
of the needle at each step toward the direction of the lesion. Appropriate biopsy
throw length needs to be selected keeping in mind not to unnecessarily sample normal
lung parenchyma beyond the lesion. Extra care needs to be taken about throw length
when the lesion is in close proximity to pulmonary vessels and bronchus as bleeding
and aspiration into the bronchus can be catastrophic.
Subpleural lesions in particular pose a challenge since these lesions tend to be highly
mobile with respiration and coaxial needle stability cannot be achieved within them.
It is thus preferred to approach such lesions close to the pleura tangentially for
better coaxial needle stability rather than taking direct short distance approach,
which may result in increasing chances of shearing of the pleura ([Fig. 2 ]).[15 ]
Although positron emission tomography (PET)/CT is better than conventional CT in detecting
bony metastasis, the sensitivity and specificity of PET/CT for pulmonary lesions in
excluding metastasis are roughly 79 and 81.8%, respectively. The significant prevalence
of well-known low 18F-fludeoxyglucose-avid tumors and the small nodule sizes were
the main causes of false negative PET/CT results.[16 ]
Limitation of our study includes small sample size and its retrospective nature. Negative
sample may not always imply nonmalignant lesions. Sampling error can happen especially
when these lesion are too small to target due to technical reasons. Though we included
patients with known malignancy, these findings can be generalized to other populations.
Conclusion
CT-guided biopsy of small pulmonary nodules is safe and feasible with good accuracy.
Adequate samples can be obtained in most cases. Majority of the complications are
minor, which do not require any intervention. In a patient with a known malignancy,
a lung nodule may indicate nonmetastatic disease in significant number of patients,
necessitating biopsy confirmation.
Fig. 2 (A ) To reduce the risk of pneumothorax, a longer approach is used to biopsy the subpleural
lesion rather than the shortest one (dotted line). (B ) To reduce the risk of pneumothorax and avoid ribs, a longer path is selected. (C ) Oblique biopsy path chosen to avoid fissure (dotted line) and the rib. (D ) Lesion biopsied in decubitus position to shorten the skin to the lesion's distance.
The posterior approach was not suitable because of a fissure (dotted line), and the
anterior entry was very long (arrow).
