US-guided transthoracic biopsy of peripheral lung lesions: pleural contact length influences diagnostic yield

Abstract

Background

Transthoracic biopsy of peripheral lung lesions under ultrasonography (US) guidance is a useful diagnostic technique. However, factors affecting diagnostic yield of US-guided transthoracic biopsy of peripheral lung lesions are not well established.

Purpose

To determine the factors that influence diagnostic yield of US-guided transthoracic biopsy in peripheral lung lesions.

Material and Methods

A total of 100 consecutive patients underwent US-guided percutaneous cutting biopsy of peripheral lung lesions from October 2007 to March 2009. After seven unconfirmed cases were excluded, 97 procedures in 93 consecutive patients were included in this study. The accuracy of the lung biopsies was assessed by comparing the biopsy results with the final diagnoses. We divided the cases into a correct group (true-positive and true-negative) and an incorrect group (false-positive, false-negative, and non-diagnostic results) and analyzed the differences in the lesions, procedures, and patient variables between the two groups.

Results

According to the final diagnoses, 56 cases (57.7%) were malignant and 41 cases (42.3%) were benign. An overall diagnostic accuracy of 91.8% was obtained. The median size of the lesions was 46.0 mm (interquartile range [IQR], 30.0–69.5 mm), and the median lesion-pleura contact arc length (LPCAL) was 31.0 mm (IQR, 18.0–51.0 mm). Multivariate logistic regression analysis showed that only LPCAL (odds ratio, 1.16; 95% CI, 1.04–1.30) was a significant predictor of a correct diagnosis. When we divided the lesions into those with LPCAL values >30 mm and LPCAL values ≤30 mm, the sensitivity (96.6% vs. 74.1%; P = 0.02) and the accuracy (98% vs. 85.4%; P = 0.03) were significantly higher in the group with larger LPCAL.

Conclusion

In US-guided transthoracic biopsy of peripheral lung lesions, the LPCAL of the lesions is an important factor for a correct diagnosis.

Keywords

Ultrasound transthoracic biopsy lung lung biopsy diagnostic yield

Introduction

Ultrasound (US) examination can be used to visualize thoracic lesions in the chest wall, the pleura, the diaphragm, and the peripheral lungs when lesions are in contact with the pleura. Additionally, peripheral lung lesions with an accessible sonic window created by the pleural contact area of the lesion or consolidation interposed between the lesion and the pleura can undergo US-guided percutaneous biopsy.

Transthoracic needle biopsy under US guidance is a useful diagnostic technique for thoracic lesions with several advantages, including dynamic properties, a short examination time, and the lack of radiation exposure (1 –5). Most studies of US-guided transthoracic needle biopsy included thoracic lesions in the peripheral lungs, the pleura, the mediastinum, and the chest wall. These studies focused mainly on the feasibility, the safety, and the diagnostic yield of US-guided transthoracic needle biopsy or compared the diagnostic rates of different types of needles (2,4,6,7).

To the best of our knowledge, no published studies have systematically analyzed the factors determining diagnostic yield of US-guided biopsy in a large number of peripheral lung lesions. The purpose of this study was to determine the factors affecting diagnostic yield of US-guided transthoracic cutting biopsy in peripheral lung lesions.

Material and Methods

From October 2007 until March 2009, 100 consecutive patients underwent US-guided percutaneous transthoracic cutting biopsy of peripheral lung lesions. Seven patients were excluded due to unconfirmed final clinical outcomes, resulting in 97 procedures in 93 patients being included in this study. Four patients underwent two repeated biopsies for the same lesions, and the repeated biopsies were considered new procedures in the statistical analysis. In all the patients, the platelet count exceeded 100,000/µL and the prothrombin times and activated prothrombin time were within normal limits.

Before US-guided biopsy, routine chest computed tomography (CT) scans were obtained using a 64- or 4-slice multidetector computed tomography (CT) scanner (Brilliance 64 CT scanner, Philips Medical Systems, Cleveland, OH, USA, and LightSpeed Plus, GE Healthcare, Milwaukee, WI, USA).

On chest CT, all lung lesions were abutting the pleura and had variable lengths. The lobar location of each lesion was recorded. Lesion size and the lesion-pleura contact arc length (LPCAL) (Fig. 1) of each lesion were measured twice on different days by one radiologist using a picture-archiving and communication system (PACS) monitor and digital imaging and communications in medicine (DICOM) imaging software (PiView; Infinitt, Seoul, Korea). The average lesion size and LPCAL were recorded. The lesion size was measured along the long axis of the lesion in the mediastinal window setting. The LPCAL was measured along the maximal curvilinear length of the lesion that contacted the pleura in the mediastinal window setting. In addition, the presence of emphysema was recorded. Informed consent was obtained from all patients prior to performing the biopsy, and approval from the institutional review board was obtained.

Fig. 1.

A schematic cross-sectional view of a peripheral lung mass that is in contact with the overlying pleural surface. The mass lesions in figures (a) and (b) have the same diameter (D) but different lesion-pleura contact arc lengths (LPCALs, white dotted line between the arrows).

US examination was performed using the Acuson Sequoia 512 ultrasound system (Siemens, Erlangen, Germany) with a 4C1 convex transducer or a 15L8 linear array transducer. All the biopsies were performed by one experienced radiologist using an automated biopsy gun with an 18-gauge cutting needle (Bard Magnum, Bard Biopsy Systems, Tempe, AZ, USA) and a penetration depth of 15 mm. The same type of needle was used in all the patients. The patients were placed in the supine, prone, or lateral decubitus position to use the shortest and safest approach for each lesion. The skin was prepped, and local anesthesia was administered. Biopsies were performed using the free-hand technique without any puncture-guiding devices. While the needle was inserted, patient respiration was suspended at the most appropriate time. After sampling, the specimen was placed in 10% formalin for histological examination. The number of tissue samples for each patient was recorded. Postprocedure upright posteroanterior chest radiographs were obtained immediately after the biopsy and 4 h and 12 h after the biopsy to detect an immediate or delayed pneumothorax. The pathological reports of the biopsy specimens, the final diagnoses, and the clinical outcomes of the patients were recorded.

The biopsy results were divided into the following diagnostic categories: (i) malignancy; (ii) specific benign; (iii) non-specific benign without evidence of malignancy; and (iv) non-diagnostic. The non-diagnostic results included biopsies with inadequate tissue, necrosis, or the presence of skeletal muscle only and in these cases, the pathologists recommended repeated biopsies. Cases were classified as specific benign when a benign diagnosis with a precise etiology could be made based on the histological analysis, and the clinical course was compatible to the diagnosis. The specific benign, non-specific benign without evidence of malignancy, and non-diagnostic results were considered negative for malignancy in the statistical analysis.

The biopsy results were considered true-positive for malignancy based on the following criteria: (i) the lesion was surgically confirmed; (ii) the biopsy of the lung lesion and other sites revealed cancer with the same histological characteristics; (iii) the lesion size had increased; (iv) other metastases were identified; or (v) the lesion responded to specific chemotherapy. A benign result was considered a true-negative for malignancy when there was surgical confirmation, the lesion size had decreased, the lesion subsequently disappeared, or the lesion had remained stable for 2 years. A biopsy result that was positive for malignancy was considered a false-positive if surgical resection yielded a benign diagnosis, the lesion subsequently disappeared, or the lesion size decreased without specific treatment. A benign biopsy result was considered a false-negative when the histopathological analysis of the biopsy specimen did not reveal evidence of malignancy, but surgical resection or the subsequent clinical course indicated that the lesion was malignant.

We classified the biopsy results into a correct group (true-positive and true-negative) and an incorrect group (false-negative, false-positive, and non-diagnostic results), and we analyzed the factors that led to correct diagnostic results. To determine the factors that influenced correct results, the patient characteristics and the lesion and procedure variables were compared between the two groups. The patient variables included patient age, sex, and the presence of emphysema. The lesion variables included lesion size, LPCAL, and lesion location. The procedure variable was the number of obtained core samples. The variables were compared between the two groups in a univariate analysis using the Mann-Whitney U test for the numerical values and the chi-squared or Fisher’s exact test for the categorical values. Multivariate logistic regression analysis was also used to identify significant predictors of the correct diagnosis. The diagnostic accuracy was compared between the lesions that were >30 mm and ≤30 mm in size and between the lesions that were >30 mm and ≤30 mm in LPCAL. A P value of <0.05 was considered statistically significant. The statistical analysis was performed using the SPSS 18.0 software (SPSS Inc., Chicago, IL, USA).

Results

The study included 67 men and 26 women, with a mean age of 62.7 years (range, 28–88 years). According to the final diagnoses, 56 lesions were malignant, and 41 lesions were benign. Five (8.9%) of the 56 malignant lesions were confirmed through surgical resections. Seven lesions (12.5%) were confirmed based on the histopathological analysis of the lung biopsy specimens and the biopsies of other sites, which indicated the same histological results. Overall, 44 lesions (78.6%) were confirmed as malignant based on the histopathological diagnosis of the biopsy specimens and the subsequent clinical course. Thirty-eight of these patients were followed up for a mean period of 16 months. These patients had progressive disease or new metastatic lesions at other sites or died of cancer progression. A diagnosis of small cell lung cancer was made in six of these patients because the lesion responded to systemic chemotherapy.

Of the 41 lesions diagnosed as benign, three (7.3%) were confirmed through surgical resections. Overall, 37 lesions (90.2%) were confirmed as benign based on biopsy results and the subsequent diminution or disappearance of the lesion. One lesion (2.4%) was confirmed as benign because it did not change in size for more than 2 years.

Forty-eight of the 56 malignant lesions and 39 of the 41 benign lesions were diagnosed correctly. Two cases were diagnosed as benign based on the first biopsy but were confirmed as malignant after a repeated biopsy using CT and US. Histological examination resulted in non-diagnostic in eight lesions (8.2%) because of necrosis or inadequate specimens (Fig. 2). Six of these lesions were confirmed as non-small cell lung cancer through surgical resection or repeated biopsy, and two lesions were confirmed as benign. Specific benign diagnoses were made in 25 of 41 benign lesions (61%), these included eight cases of tuberculosis, five cases of organizing pneumonia, four cases of pneumonia, two abscesses, two cases of paragonimiasis, one case of aspergillosis, one case of cryptococcosis (Fig. 3), one case of cryptogenic organizing pneumonia, and one inflammatory pseudotumor.

Fig. 2.

A 67-year-old man presented with an incidentally found lung mass. (a) Chest CT shows a 43-mm mass in the periphery of the right lower lobe. The lesion-pleura contact arc length (LPCAL) (arrowheads) was 17 mm. (b) Biopsy was performed under US guidance. The result of the first biopsy was non-diagnostic, but an adenocarcinoma was diagnosed after a repeated US-guided biopsy. Note the relatively narrow US window (arrowheads) compared with the mass size. The inserted needle was visualized as an echogenic line (arrow).

Fig. 3.

A 65-year-old woman presented with incidentally detected lung lesions during chemotherapy for non-Hodgkin’s lymphoma. (a) Chest CT scan shows a small nodule (arrow) in the right lower lobe that was 9 mm in diameter with a lesion-pleura contact arc length (LPCAL) of 8 mm. (b) US-guided biopsy was successfully performed, and cryptococcosis was confirmed. The inserted needle was visualized as an echogenic line (arrow).

The median lesion size was 46.0 mm (interquartile range [IQR], 30.0–69.5 mm). The median LPCAL was 31.0 mm (IQR, 18.0–51.0 mm). The relationship between lesion size or LPCAL and diagnostic accuracy is shown in Table 1. Of the lesions in lobar locations, 16 (16.5%) were in the right upper lobe, seven (7.2%) in the right middle lobe, 38 (39.2%) in the right lower lobe, 15 (15.5%) in the left upper lobe, and 21 (21.6%) in the left lower lobe. A median of two core samples were obtained from each procedure (range, 1–4 samples). Emphysema was present in 21 cases.

Table 1.

Results of US-guided transthoracic biopsies according to lesion size and lesion-pleura contact arc length (LPCAL).

(mm)	Lesion size						LPCAL
	Malignant (n = 56)			Benign (n = 41)			Malignant (n = 56)			Benign (n = 41)
	TP	FN	ND	SB	NB	ND	TP	FN	ND	SB	NB	ND
≤10	2	1	1	3	2	4
11–30	9	2	6	3	1	17	1	4	9	6	2
31–50	9	3	12	6	1	10	1	9	5
≤51	28	2	6	5	18	3	3
Total	48	2	6	25	14	2	48	2	6	25	14	2

FN, false negative; LPCAL, lesion-pleura contact arc length; NB, non-specific benign; ND, non-diagnostic; SB, specific benign; TP, true-positive.

Of all of the lesions, 87 (true-positive and true-negative) were included in the correct group, and 10 (false-negative and non-diagnostic results) in the incorrect group. There were no significant differences in lesion size between the correct and the incorrect groups. However, the mean LPCAL in the correct group was significantly longer than in the incorrect group (P = 0.004) (Table 2). Other factors including age, sex, number of core samples, lesion location, and the presence of emphysema were not different between two groups. Multivariate logistic regression analysis showed that only LPCAL (odds ratio, 1.16; 95% CI, 1.04–1.30) was a significant predictor for correct diagnosis.

Table 2.

Characterization of the correct group (TP + TN) and the incorrect group (FP + FN + ND) in 97 US-guided transthoracic needle biopsies.

Characteristics	Correct group (n = 87)	Incorrect group (n = 10)	P
Age (years)	66.0	67.5	0.722*
Sex (male/female)	62/25	7/3	0.595^†
Size (mm)	46.0	34.0	0.069*
LPCAL (mm)	34.0	16.0	0.004*^‡
Core samples (n)	2.0	2.0	0.522*
Location (RUL/RML/RLL/LUL/LLL)	14/7/32/1/19	2/0/6/0/2	0.425^†
Emphysema (Yes/No)	18/69	3/7	0.447^†

The Mann-Whitney U test was used for the group comparisons.

†

Fisher’s exact test or a chi-squared test was used for the group comparisons.

‡

P < 0.05.

The numbers represent the medians or the numbers of patients.

FN, false-negative; FP, false-positive; LLL, left lower lobe; LPCAL, lesion-pleura contact arc length; LUL, left upper lobe; ND, non-diagnostic; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe; TN, true-negative; TP, true-positive.

The overall diagnostic accuracy was 91.8% (89/97). The sensitivity and the specificity for a diagnosis of malignancy were 85.7% (48/56) and 100% (41/41), respectively. There were significant decreases in sensitivity (96.6% vs. 74.1%; P = 0.02) and in accuracy (98% vs. 85.4%; P = 0.03) for lesions with LPCAL values ≤30 mm compared with lesions that had LPCAL values >30 mm. However, there were no differences in the diagnostic performance between the two groups according to lesion size (Table 3). Postbiopsy pneumothorax and hemoptysis occurred in two patients. Chest tube placement or other specific treatments were not necessary.

Table 3.

A comparison of the diagnostic accuracy for malignancy of US-guided transthoracic biopsy according to lesion size and lesion-pleura contact arc length (LPCAL).

	All (n = 97)	Size (mm)			LPCAL (mm)
	All (n = 97)	≤30 (n = 25)	>30 (n = 72)	P	≤30 (n = 48)	>30 (n = 49)	P
Sensitivity	85.7 (48/56)	78.6 (11/14)	88.1 (37/42)	0.40	74.1 (20/27)	96.6 (28/29)	0.02*
Specificity	100 (41/41)	100 (11/11)	100 (30/30)	1	100 (21/21)	100 (20/20)	1
Accuracy	91.8 (89/97)	88 (22/25)	93.1 (67/72)	0.42	85.4 (41/48)	98 (48/49)	0.03*
PPV	100 (48/48)	100 (11/11)	100 (37/37)	1	100 (20/20)	100 (28/28)	1
NPV	83.7 (41/49)	78.6 (11/14)	85.7 (30/35)	0.67	75 (21/28)	95.2 (20/21)	0.12

P < 0.05 (Fisher’s exact test).

The numbers of the patients that were used to calculate the percentages are in parentheses.

NPV, negative predictive value; PPV, positive predictive value.

Discussion

Ultrasonography is not a familiar tool for thoracic radiologists because of its infrequent use; however, this technique offers several advantages to both the operator and the patient. With an available sonic window, US can guide percutaneous biopsy of peripheral thoracic lesions in the chest wall, the pleura, the mediastinum, and the peripheral lungs (8,9).

Many studies have demonstrated that US is as effective as CT for guidance during transthoracic biopsy of peripheral lung lesions, and the accuracy of this approach has ranged from 84–95% in confirmative histological diagnoses (1,3,4,10 –13). In this study, the overall diagnostic accuracy was 91.8%.

Several factors can influence the diagnostic yield of transthoracic needle biopsy, including lesion size, needle type, needle size, and the expertise of the operator. Compared with fine-needle aspiration biopsy, tissue-core biopsy with a cutting needle provides comparatively larger specimens, enables specific benign diagnoses, and can be used to determine the subtypes of malignant tumors (4,10,12). There were no significant differences in the diagnostic performance of the different types of cutting biopsy needles (2). In this study, biopsies were performed by one radiologist using one type of cutting needle in all the patients; therefore, the lesion factors may have affected the diagnostic accuracy.

Several studies on the diagnostic yield of image-guided transthoracic needle biopsy focused on the relationship between lesion size and diagnostic accuracy, and have reported that diagnostic accuracy decreases with decrease in lesion size (4,14 –17). In most of these studies, CT was used as the guiding tool. In contrast, lesion size did not influence the diagnostic accuracy in studies that have used US as a guiding tool for transthoracic needle biopsy (2,4,18,19). Liao et al. and Yuan et al. reported diagnostic accuracies of 96% and 90% for US-guided transthoracic biopsy of peripheral thoracic lesions that were <3 cm, respectively (18,19). According to Yang et al., neither lesion location nor lesion size affected the results of US-guided needle biopsy of thoracic lesions (4). Our results were in agreement with their findings. Lesion size did not influence the diagnostic yield of US-guided biopsy of peripheral lung lesions. Additionally, there were no significant differences in the diagnostic accuracy for lesions that were >3 cm or <3 cm in diameter. In our study, the only factor that affected the diagnostic yield of US-guided transthoracic biopsy in univariate and multivariate analyses was the LPCAL. In addition, for lesions with LPCAL values ≤30 mm, the diagnostic accuracy decreased from 98% to 85.4%.

Under CT guidance, needle passage and tissue sampling must be performed blindly unless CT fluoroscopy is available, and the ribs and poor respiration holding hinder the accurate targeting of small lesions. Therefore, sampling errors could occur more frequently in small lesions using CT-guided biopsy. However, under US guidance, biopsy can be performed in the respiratory phase when the lesion is most accessible, and the needle tip can be visualized throughout the procedure to ensure proper placement of the needle (9). Therefore, lesion size dose not significantly affect the diagnostic yield of US-guided biopsy, especially when the biopsy is performed by an experienced radiologist. However, at times large lesions can have extensive central necrosis, which can lead to inadequate sampling for a histopathological diagnosis (20).

The contact area between the peripheral lung lesions and the pleura acts as an ultrasonic window, which is a prerequisite for US examinations. US-guided biopsy of lung lesions can be performed more easily when the lesions have wide sonic windows that permit a flexible approach route and enable the needle tip to be traced during the procedure. Therefore, the LPCAL is a more important factor than lesion size for successful targeting and sampling of peripheral lung lesions using US-guided biopsy.

Pneumothorax and other biopsy-associated complications were not significant even with an average of two needle passes because real-time monitoring under US guidance avoided passage through normal lung tissue and the laceration of the pleura based on the respiratory movement.

In conclusion, the LPCAL was the only significant predictor for correct diagnosis in US-guided transthoracic biopsy of peripheral lung lesions. Measuring LPCAL on CT images may help radiologists decide whether to use US as a guiding tool for biopsy of peripheral lung lesions.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Yang

Luh

Sheu

. Peripheral pulmonary lesions: ultrasonography and ultrasonically guided aspiration biopsy. Radiology 1985; 155: 451–456.

Pang

Tsang

Hom

. Ultrasound-guided tissue-core biopsy of thoracic lesions with Trucut and Surecut needles. Chest 1987; 91: 823–828.

Yang

Chang

. Ultrasound-guided core biopsy of thoracic tumors. Am Rev Respir Dis 1992; 146: 763–767.

Yang

Lee

. Ultrasonographically guided biopsy of thoracic tumors. A comparison of large-bore cutting biopsy with fine-needle aspiration. Cancer 1992; 69: 2553–2560.

Yang

. Ultrasound-guided transthoracic biopsy of peripheral lung, pleural, and chest-wall lesions. J Thorac Imaging 1997; 12: 272–284.

Ikezoe

Sone

Higashihara

. Sonographically guided needle biopsy for diagnosis of thoracic lesions. Am J Roentgenol 1984; 143: 229–234.

Pedersen

Aasen

Gulsvik

. Fine needle aspiration biopsy of mediastinal and peripheral pulmonary masses guided by real-time sonography. Chest 1986; 89: 504–508.

Wernecke

Vassallo

Peters

. Mediastinal tumors: biopsy under US guidance. Radiology 1989; 172: 473–476.

Cinti

Hawkins

. Aspiration biopsy of peripheral pulmonary masses using real-time sonographic guidance. Am J Roentgenol 1984; 142: 1115–1116.

10.

Hsu

Chiang

Hsu

. Ultrasound-guided fine-needle aspiration biopsy of lung cancers. J Clin Ultrasound 1996; 24: 225–233.

11.

Ikezoe

Morimoto

Arisawa

. Percutaneous biopsy of thoracic lesions: value of sonography for needle guidance. Am J Roentgenol 1990; 154: 1181–1185.

12.

Sagar

Gulati

Gupta

. Ultrasound-guided transthoracic co-axial biopsy of thoracic mass lesions. Acta Radiol 2000; 41: 529–532.

13.

Sconfienza

Mauri

Grossi

. Pleural and Peripheral Lung Lesions: Comparison of US- and CT-guided Biopsy. Radiology 2013; 266: 930–935.

14.

Boiselle

Shepard

. Diagnostic accuracy and safety of CT-guided percutaneous needle aspiration biopsy of the lung: comparison of small and large pulmonary nodules. Am J Roentgenol 1996; 167: 105–109.

15.

Ohno

Hatabu

Takenaka

. CT-guided transthoracic needle aspiration biopsy of small (< or = 20 mm) solitary pulmonary nodules. Am J Roentgenol 2003; 180: 1665–1669.

16.

Tsukada

Satou

Iwashima

. Diagnostic accuracy of CT-guided automated needle biopsy of lung nodules. Am J Roentgenol 2000; 175: 239–243.

17.

Montaudon

Latrabe

Pariente

. Factors influencing accuracy of CT-guided percutaneous biopsies of pulmonary lesions. Eur Radiol 2004; 14: 1234–1240.

18.

Liao

Chen

Chang

. US-guided transthoracic cutting biopsy for peripheral thoracic lesions less than 3 cm in diameter. Radiology 2000; 217: 685–691.

19.

Yuan

Yang

Chang

. Ultrasound-guided aspiration biopsy of small peripheral pulmonary nodules. Chest 1992; 101: 926–930.

20.

Yeow

Tsay

Cheung

. Factors affecting diagnostic accuracy of CT-guided coaxial cutting needle lung biopsy: retrospective analysis of 631 procedures. J Vasc Interv Radiol 2003; 14: 581–588.