Eur Radiol DOI 10.1007/s00330-015-3861-6

INTERVENTIONAL

Comparison of cone-beam CT-guided and CT fluoroscopy-guided transthoracic needle biopsy of lung nodules Nicola Rotolo 1 & Chiara Floridi 2 & Andrea Imperatori 1 & Federico Fontana 2 & Anna Maria Ierardi 2 & Monica Mangini 2 & Veronica Arlant 1 & Giuseppe De Marchi 2 & Raffaele Novario 3 & Lorenzo Dominioni 1 & Carlo Fugazzola 2 & Gianpaolo Carrafiello 2

Received: 30 September 2014 / Revised: 12 May 2015 / Accepted: 21 May 2015 # European Society of Radiology 2015

Abstract Purpose To compare the diagnostic performance of conebeam CT (CBCT)-guided and CT fluoroscopy (fluoro-CT)guided technique for transthoracic needle biopsy (TNB) of lung nodules. Methods The hospital records of 319 consecutive patients undergoing 324 TNBs of lung nodules in a single radiology unit in 2009–2013 were retrospectively evaluated. The newly introduced CBCT technology was used to biopsy 123 nodules; 201 nodules were biopsied by conventional fluoro-CT-guided technique. We assessed the performance of the two biopsy systems for diagnosis of malignancy and the radiation exposure. Results Nodules biopsied by CBCT-guided and by fluoro-CTguided technique had similar characteristics: size, 20±6.5 mm (mean±standard deviation) vs. 20±6.8 mm (p=0.845); depth from pleura, 15±15 mm vs. 15±16 mm (p=0.595); malignant, 60 % vs. 66 % (p=0.378). After a learning period, the newly introduced CBCT-guided biopsy system and the conventional fluoro-CT-guided system showed similar sensitivity (95 % and 92 %), specificity (100 % and 100 %), accuracy for diagnosis of malignancy (96 % and 94 %), and delivered nonsignificantly different median effective doses [11.1 mSv

* Gianpaolo Carrafiello [email protected]

(95 % CI 8.9–16.0) vs. 14.5 mSv (95 % CI 9.5–18.1); p=0.330]. Conclusion The CBCT-guided and fluoro-CT-guided systems for lung nodule biopsy are similar in terms of diagnostic performance and effective dose, and may be alternatively used to optimize the available technological resources. Key Points • CBCT-guided and fluoro-CT-guided lung nodule biopsy provided high and similar diagnostic accuracy. • Effective dose from CBCT-guided and fluoro-CT-guided lung nodule biopsy was similar. • To optimize resources, CBCT-guided lung nodule biopsy may be an alternative to fluoro-CT-guided. Keywords Lung nodule biopsy . Transthoracic needle biopsy . C-arm cone-beam CT . CT fluoroscopy . Radiation dose

Abbreviations CBCT Cone-beam computed tomography CT Computed tomography Fluoro-CT CT fluoroscopy NPV Negative predictive value PPV Positive predictive value

1

Center for Thoracic Surgery, Insubria University, Viale Borri 57, 21100 Varese, Italy

Introduction

2

Radiology Department, Insubria University, Viale Borri 57, 21100 Varese, Italy

3

Medical Physics Department, Insubria University, Viale Borri 57, 21100 Varese, Italy

Management of pulmonary nodules, many of which are incidentally diagnosed as a result of extensive use of computed tomography (CT) examinations, is generally based on the estimated probability of cancer. After evaluation of clinical risk

Eur Radiol

factors and of nodule imaging characteristics, biopsy is recommended for lesions with intermediate or high probability of cancer, and transthoracic needle biopsy (TNB) is frequently used to achieve definitive pathological diagnosis [1–3]. Generally, TNB is performed in the outpatient setting, with few complications. The procedure of choice (ultrasound or CT or fluoroscopy guidance) depends on nodule location and depth in the parenchyma, on accessibility of the technical devices, and on the radiologist’s experience or preference. Currently, TNB with CT fluoroscopy (hereafter fluoro-CT) guidance is the commonest and state-of-the-art technique, achieving greater than 90 % diagnostic accuracy for malignant lung lesions, with reasonably low morbidity and minimal mortality [3–5]. In recent years a new, real-time needle guidance technique using C-arm cone-beam CT (hereafter CBCT) was developed that can be used to visualize under fluoroscopic guidance the needle trajectory from skin to target pulmonary lesions during interventional radiology procedures [6–10]. It has been suggested that the CBCT system, compared to fluoro-CT, facilitates the needle biopsy procedure and likely reduces radiation exposure of patient and operator [11–15]; however a direct comparison with evaluation of the advantages of CBCT over conventional fluoro-CT biopsy is not currently available [15]. In this study we aimed to compare the performance of CBCT-guided and fluoro-CT-guided TNB for diagnosis of malignancy of indeterminate lung nodules, and to assess the radiation exposure with the two procedures.

Materials and methods We retrospectively evaluated the hospital records of the 391 consecutive patients 18–90 years of age, referred to the interventional radiology unit of the Varese University Hospital for TNB of lung nodule(s) ≤3 cm in diameter between January 2009 and December 2013. After excluding biopsy candidates with (a) inability to comply with breath-holding instructions, (b) impaired coagulation, (c) high index of suspicion for benign lung lesion, (d) endobronchial lesion, (e) pure ground glass opacity (GGO), 319 subjects [215 male, 104 female; mean age, 68±10 years (range, 27–89 years)] were included in this study. Our institutional review board approved this study. All included patients gave written informed consent for imaging-guided TNB the day before the procedure. The TNB procedures were performed by the team of three radiologists with experience in CT-guided interventional procedures (GC, FF, AMI) who introduced the new C-arm CBCT technology in our hospital in 2010 [9, 10]. This study was not randomized. Our interventional radiologists rotated daily between the fluoro-CT suite and the angiography suite equipped with CBCT technology and performed the programmed TNBs using the available imaging system they were assigned for the

day. We recorded the location, size and depth of each nodule in the lung. Lesion size was defined as the nodule’s longest diameter on lung window setting CT images; depth was measured as the nodule’s distance from the parietal pleural surface. Transthoracic needle biopsy: technical procedure Lung biopsies included in the study were performed either in the CT room under fluoro-CT guidance (Toshiba, Aquilion 64, Japan) or in the angiography suite equipped with C-arm CBCT system (Philips Allura Xper FD20 system, Philips, Best NL) and with dedicated BXperGuide^ software (Philips Healthcare, Best, NL). The patient was positioned prone or supine, depending on the lesion location and on the access considered optimal. Before and during the procedure, the nursing staff monitored heart rate, blood oxygen saturation, respiratory rate and blood pressure. Fluoro-CT guided lung biopsy A preliminary chest CT scan was obtained to establish the cutaneous access point that was marked on the patient skin. After skin disinfection, local anaesthetic (2 % mepivacaine) was injected at the chest access site. Then, under fluoro-CT guidance the operator advanced the 20-gauge cutting needle with a coaxial system (Biopsy Bell, Medical Device, Modena, Italy) into the target lung lesion. The needle path was followed by repeated fluoro-CT acquisitions until the needle was correctly placed; the biopsy was then taken with the coaxial system. Finally, one last CT was obtained to assess complications of the procedure (perilesional haemorrhage, pneumothorax). CBCT-guided lung biopsy The first step was a CBCT scan for adequate planning of the biopsy. Briefly, using the XperGuide software, the operator or the training assistant at the workstation monitor established the target in the context of the lesion (Btarget point^) and decided the needle entry point at the skin surface. In all cases we used a 20-gauge cutting needle with a coaxial system (Biopsy Bell, Medical Device, Modena, Italy). The virtual segment connecting the entry and target points marked the needle path (Fig. 1). On the basis of this planning, the XperGuide software automatically calculated the correct C-arm position to display the needle Bentry point^ on the patient’s skin. This phase of the procedure, followed by C-arm positioning, was called Bentry point positioning^. After skin disinfection at the access area, and under local anaesthesia (2 % mepivacaine), via fluoroscopy the operator positioned the needle tip at the cutaneous Bentry point^, based on the CBCT image appearing on the monitor as a fusion of the 3D volume previously acquired using CBCT and the real-time fluoroscopy bidimensional plane. At this time, the C-arm was rotated in the

Eur Radiol

Fig. 1 Preprocedural CBCT shows needle pathway planned with XperGuide software (a, b) and visualization of entry point (c)

Bprogression view^ position, perpendicular to the previous one. Then the needle was advanced into the chest to reach the target point, following the virtual path previously determined and displayed on the monitor in real-time fluoroscopy. Thereafter, a second CBCT scan was performed to verify the needle correct placement (Fig. 2), and the biopsy was taken. A third and final scan was acquired to assess any biopsy-related complications (Fig. 3). The biopsy material was macroscopically assessed by the interventional radiologist, and another satisfactory specimen was obtained if necessary. Rapid on-site cytopathology examination [2] was not available at our unit. After all biopsies the patient was monitored for at least 4 h in the recovery room, and before discharge a chest X-ray was performed in postero-anterior projection and with maximum expiration, to highlight possible pneumothorax or haemorrhage. Biopsy tissue samples were formalin-fixed and paraffinembedded for pathology examination. The biopsy material was considered inadequate if sampled tissue was insufficient

Fig. 2 Intraprocedural CBCT shows the needle within the lesion (a), enlarged view shows the needle exiting sleeve (empty arrow) to collect biopsy material, positioned correctly in the middle of the lesion (white arrow) (b)

for pathologic diagnosis of the pulmonary lesion (no specimen, necrotic tissue, normal lung/chest wall tissue).

Data collection For each needle biopsy procedure we recorded any complications and length of hospital stay. After biopsy, lesions temporarily labelled Bbenign^ and unresected were followed up for at least 2 years to ascertain their benign status; if follow-up was less than 2 years, lesions were defined as Bindeterminate^. All surgically excised lesions underwent definitive pathologic confirmation that was used as a reference to assess the specificity, sensitivity and accuracy of diagnosis of CBCT-guided and fluoro-CT-guided biopsies. To this effect we classified the result of lung nodule biopsy as true positive, true negative, false positive or false negative, according to the criteria described in detail by Priola et al. [3]. For the newly introduced CBCT-guided biopsy technique, we separately evaluated the diagnostic performance in the learning period (arbitrarily the

Eur Radiol Fig. 3 Postprocedural CBCT shows minimal paralesional bleeding (a, empty arrow) and the pneumothorax stratum positioned caudally at the base of the lung (b, white arrow)

first 15 CBCT-guided adequate biopsies for each of the three interventional radiologists), and in the subsequent procedures. The effective doses were retrospectively evaluated for the 70 biopsies (45 CBCT and 25 fluoro-CT procedures) performed in the last year of this study (2013), as for these cases all the relevant data were available to estimate radiation exposure. For CBCT-guided biopsies, the effective dose (expressed in mSv) was estimated from dose–area product (expressed in mGy cm2), using the previously described conversion factor for the Allura Xper FD20 system (0.31 mSv Gy−1 cm−2) that was obtained from a separate phantom study [11]. For fluoro-CT procedures the effective dose was obtained from the value of dose–length product (expressed in mGy cm) according to the method of Christner et al. [16]. For these 70 cases the body mass index (BMI; body mass divided by the square of the height) of corresponding patients was retrospectively assessed to evaluate the eventual influence of this factor (Table 1). All data were recorded in a dedicated database using Excel (Microsoft, Seattle, USA).

A p value less than 0.05 was considered statistically significant.

Results In 319 patients we performed a total of 324 TNBs of lung nodules, comprising 123 with CBCT guidance and 201 with fluoro-CT guidance. The five repeat biopsies in this study were evaluated as new cases (Table 1). Between the CBCT-guided and the fluoro-CT guided biopsy group there was no significant difference of gender distribution (p = 0.926), age (p = 0.709) and BMI (p = 0.77) (Table 1); lung nodule characteristics of the two groups were similar: diameter (p=0.845), depth in the lung (p=0.595) and lung lobe of location (p=0.895). The proportion of malignant lesions was also similar (60 % vs. 66 %; p=0.378) (Table 2). Lung nodule biopsy was completed in all cases without major complications. Of the 319 patients, 294 (92 %) were discharged on the same day of the procedure, while 25 (8 %) developed pneumothorax requiring chest tube drainage (see details below) and remained hospitalized overnight.

Statistical analysis

Diagnostic performance

Sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) were recorded for both CBCT-guided and fluoro-CT guided biopsies. Data were expressed as mean value and standard deviation, or median and 95 % confidence interval (CI), or range. Statistical analysis was performed using MedCalc 13.2.2 (MedCalc Software, Ostend, Belgium). Differences between proportions were analysed with the chi-square test or Fisher exact test. Continuous normally distributed variables were compared by Student t test. Normal distributions of sample values were tested by D’Agostino–Pearson test [17]. Non-normally distributed variables were compared by Mann–Whitney U test.

In 28 biopsies the sampled tissue was inadequate for diagnosis (8 % of CBCT-guided; 9 % of fluoro-CT-guided; Table 2); five of these TNBs were repeated (four CBCT-guided and one fluoro-CT-guided). Of the 23 cases that remained indeterminate, nine were ascertained by lung wedge resection and 14 were followed up clinically. In the 113 adequate CBCT-guided biopsies, we found 11 false-negative diagnoses (10 %). Overall in the 103 CBCTguided TNBs for which final pathologic diagnosis was available (Table 2), we recorded 87 % (95 % CI 78–93 %) sensitivity, 100 % (95 % CI 78–100 %) specificity, 89 % (95 % CI 83–95 %) accuracy, 100 % (95 % CI 94–100 %) PPV and

Eur Radiol Table 1 Demographic data of patients and characteristics of lung nodules undergoing biopsy Man, n (%) Mean age, years (range) BMIa Lesion size, mm 4–10, n (%) 11–20, n (%) 21–30, n (%) Mean diameter±SD, mm (range) Depth in the lung, mm ≤20, n (%) >20, n (%) Mean depth±SD, mm (range) Location Upper and middle lobe, n (%) Lower lobe, n (%)

CBCT-guided biopsy n=123

Fluoro-CT-guided biopsy n=201

p value

82 (66.7) 68±11 (27–88) 22.2±1.3 (17–26)

133 (66.2) 68±10 (38–89) 22.3±1.2 (19–27)

0.926 0.709 0.77

8 (6.5) 64 (52.0) 51 (41.5) 19.8±6.5 (7–30)

18 (8.9) 90 (44.8) 93 (46.3) 19.6±6.8 (4–30)

0.845

86 (69.9)

146 (72.6)

37 (30.0) 15.4±15.1 (0–67)

55 (27.4) 15.2±16.1 (0–65)

78 (63.4) 45 (36.6)

126 (62.7) 75 (37.3)

0.595 0.895

CBCT cone-beam computed tomography, CT computed tomography a

Table 2 Histological diagnosis of CBCT-guided and fluoro CTguided biopsies

The BMI was calculated only for the 70 cases (45 CBCT, 25 CT) used for the dose evaluation

Diagnosis

CBCT-guided biopsy n=123

Fluoro CT-guided biopsy n=201

Malignant lesions Primary lung cancer

74 (60 %)a

132 (66 %)a

Adenocarcinoma Squamous cell carcinoma Neuroendocrine tumor Atypical adenomatous hyperplasia Non-small cell lung cancer Adenocarcinoma with ground glass Adenoid cystic carcinoma Small cell carcinoma Metastases Cancer not specified Benign lesions

41 11 2 1 1 – – – 15 3 39 (32 %)b

65 25 7 5 3 2 2 1 20 2

Anthracosis Inflammatory infiltrates No neoplasia Sclerosis Hamartoma Epithelioid granulomas Sarcoidosis Interstitial lung disease Tuberculosis Inadequate sample

8 14 2 9 1 2 2

51 (25 %)c 12 18 7 4 3 2 0

1 0 10 (8 %)

3 2 18 (9 %)

CBCT cone-beam computed tomography, CT computed tomography a

Proportion of malignant nodules in the two groups was not statistically different (chi-square test)

b

10 of 39 are currently indeterminate/benign (follow-up continues)

c

4 of 51 are currently indeterminate/benign (follow-up continues)

p=0.378

p=0.268

p=0.957

Eur Radiol

62 % (95 % CI 42–79 %) NPV for diagnosis of malignancy (Table 3). Because proficiency in performing TNB improves with time, with a suggested minimum of 10 procedures to establish basic competency [18], for the new CBCT-guided biopsy technique we evaluated the diagnostic performance in the learning period and separately in subsequent procedures. In the learning period the diagnostic performance was suboptimal; however, in the subsequent experience of 58 CBCT-guided procedures, we observed 95 % sensitivity, 100 % specificity and 96 % accuracy for diagnosis of malignancy (Table 3). Overall the fluoro-CT guided adequate biopsies (n=179) yielded six false-negative results (3 %) and showed 92 % (95 % CI 86–96 %) sensitivity, 100 % (95 % CI 88–100 %) specificity, 94 % (95 % CI 90–98 %) accuracy, 100 % (95 % CI 96–100 %) PPVand 77 % (95 % CI 62–87 %) NPV. Notably, after the learning period, the newly introduced CBCT-guided and the conventional fluoro-CT-guided technique had essentially the same high sensitivity, specificity and accuracy for diagnosis of nodule malignancy (Table 3). To analyse the influence of nodule size on the accuracy of biopsy, we classified nodules into three subgroups, based on the lesion longest diameter on CT scan: small (≤10 mm), medium (11–20 mm), large (21–30 mm) (Table 1). Subgroup analysis based on nodule size and depth in the parenchyma showed similar accuracy for the CBCT-guided and the fluoroCT- guided systems (Table 4).

(p=0.570). Also the rate of pneumothorax requiring chest tube drainage was not significantly different (p=0.134) (Table 5). Minor intrapulmonary haemorrhage occurred in 21.9 % of CBCT-guided and in 18.9 % of fluoro-CT-guided biopsies (Table 5). Only one episode of haemoptysis was observed (after fluoro-CT procedure), which resolved spontaneously. Effective dose The mean effective dose through CBCT-guided biopsies and fluoro-CT-guided biopsies was 14.3 ± 10.0 and 16.2 ± 10.0 mSv, respectively (Fig. 4). We ascertained that effective dose values were not normally distributed in the sample of both biopsy techniques (D’Agostino–Pearson test, p=0.0002 and p=0.0241, respectively); therefore, the median values were compared. The median effective dose was 11.1 mSv (95 % CI 8.9–16.0 mSv) for CBCT-guided biopsies and 14.5 mSv (95 % CI 9.5–18.1 mSv) for fluoro-CT-guided biopsies. Thus, the median effective dose from CBCT procedures was 23 % lower compared to that of fluoro-CT procedures; however, this difference was not statistically significant (Mann–Whitney U test, p=0.330). We did not evaluate the influence of BMI because there was no significant difference between the BMI of the two groups.

Complications

Discussion

Pneumothorax (including all grades of severity of pneumothorax) occurred at a similar rate in the two types of procedure

Initial experiences suggested that CBCT-guided TNB could be a useful, safe and accurate procedure for diagnosis of

Table 3 Diagnostic performance of CBCT-guided biopsy and fluoro CT-guided biopsy

CBTC-guided procedures

Sensitivity Specificity Accuracy PPV NPV Malignant lesions Benign lesions Mean size±SD, mm Mean depth±SD, mm

Cumulative experience (n=103 biopsies)a

Initial experience (n=45 biopsies)

Later experience (n=58 biopsies)

87 % 100 % 89 % 100 % 62 % 74 (72 %) 29 (28 %) 19.8±6.5 15.4±15.1

83 % 100 % 87 % 100 % 63 % 30 (67 %) 15 (33 %) 19.9±7.3 14.2±14.9

95 % 100 % 96 % 100 % 78 % 44 (76 %) 14 (24 %) 20.0±6.4 16.3±15.2

Fluoro CT-guided procedures n=179 biopsiesb

92 % 100 % 94 % 100 % 77 % 132 (74 %) 47 (26 %) 19.6±6.8 15.2±16.1

CBCT cone-beam computed tomography, CT computed tomography, PPV positive predictive value, NPV negative predictive value, SD standard deviation a

Analysis made in the 103 biopsies for which final pathologic diagnosis was available (of 20 missing, 10 were inadequate samples and 10 were indeterminate/benign lesions currently on follow-up)

b Analysis made in the 179 biopsies for which final pathologic diagnosis was available (of 22 missing, 18 were inadequate samples and four were indeterminate/benign lesions currently on follow-up)

Eur Radiol Table 4 Diagnostic performance of CBCT-guided and fluoro-CTguided biopsy, as related to lung nodule size and depth in the parenchyma

Sensitivitya CBCT

Specificitya

Accuracya

Fluoro-CT

CBCT

Fluoro-CT

CBCT

Fluoro-CT

100 % (15) 87 % (78) 91 % (86)

100 % (4) 100 % (26) 100 % (28)

100 % (15) 100 % (78) 100 % (86)

100 % (4) 88 % (26) 96 % (28)

100 % (15) 95 % (78) 93 % (86)

100 % (40) 100 % (18)

100 % (131) 100 % (48)

95 % (40) 88 % (18)

Nodule size (mm) 4–10 100 % (4) 11–20 86 % (26) 21–30 96 % (28) Depth (mm) ≤20 93 % (40) >20 88 % (18)

90 % (131) 90 % (48)

92 % (131) 91 % (48)

Data within parenthesis are numbers of lesions CBCT cone-beam computed tomography, CT computed tomography a

Performance was analysed in 58 CBCT biopsies performed after completion of the learning period, and in 179 fluoro-CT biopsies (see note of Table 3)

indeterminate lung nodules, with possible advantages over fluoro-CT-guided biopsy [10, 12–14, 19]. It is well known that limited access due to a closed CT gantry increases the difficulty of TNB and increases radiation exposure for the operator, whereas the greater working space provided by the C-arm cone-beam system facilitates needle placement and speeds up the procedure [12, 20]. The CBCT system offers advanced needle planning under real-time needle guidance, using a combination of 3D images and fluoroscopy, with good angulation and rotation. Another advantage of CBCT is the improved resolution of images, which have superior contrast, without distortion, compared to fluoro-CT images [21]. Moreover, under CT guidance the needle path can be followed only on a single plane during needle insertion, a limitation in comparison with the 3D images offered by CBCT [12]. Furthermore, with fluoro-CT, greater patient cooperation is required to control breathing and to successfully perform the biopsy. Intuitively, these advantages of CBCT guidance should facilitate TNB of pulmonary lesions and improve the diagnostic performance of the procedure. Lee et al. reported their clinical experience with 1153 CBCT-guided TNBs of lung nodules, currently the largest published series, showing 95.7 % sensitivity, 100 % specificity and 97.0 % accuracy for the diagnosis of malignancy [15]. Table 6

summarizes the diagnostic performance of lung biopsy with the CBCT system in the series published from 2010 to date, showing that the diagnostic accuracy for malignancy ranges from 91.7 to 97 %. These data compare favourably with the 67–97 % accuracy reported by other authors for the conventional fluoro-CT-guided technique [3, 5, 22]. As there are no published reports comparing the two CT-guided techniques in the same institution, in the present study we directly compared the diagnostic performance of CBCT-guided and fluoro-CT-guided biopsy of lung nodules in the setting of our interventional radiology unit. The results of our interventional radiologists’ team indicate that after a short learning period the diagnostic accuracy of the newly introduced CBCT-guided biopsy procedure was high (96 %) and similar to that of conventional fluoro-CT-guided biopsy (94 %). Regarding complications of TNB of pulmonary nodules, the rate of pneumothorax reported in the literature varies widely, but is generally in the range of 15–30 % of CT-guided biopsies, while the proportion of TNBs requiring treatment by chest tube drainage ranges from 1 to 14 % [15, 23–29]. In our series the incidence of pneumothorax, the proportion of TNBs requiring chest

Table 5 Post-procedure complications Pneumothorax—all cases, n (%) Pneumothorax requiring drainage, n (%) Minor pulmonary haemorrhage, n (%)

CBCT guided biopsy n=123

Fluoro CT-guided biopsy n=201

P value

36 (29.3) 6 (4.9) 27 (21.9)

53 (26.7) 19 (9.4) 38 (18.9)

0.570 0.134 0.506

CBCT cone-beam computed tomography, CT computed tomography

Eur Radiol

Fig. 4 Mean effective dose±standard deviation (mSv) from CBTCguided and fluoro-CT-guided procedures of lung nodule biopsy

tube drainage and the incidence of minor pulmonary haemorrhage were in line with the literature data, and we found no significant differences between the CBTC group and the fluoro-CT group (Table 5). During CT-guided biopsy procedures there is potential for high radiation exposure. Although there are limited data in the literature on the radiation dose for CBCTguided lung biopsy, the mean effective dose with this procedure was reported to range from 4.6 to 12.3 mSv [7, 15, 19, 28, 30]. In our study the mean estimated effective dose through CBCT-guided biopsies was 14.3 ±10.0 mSv, slightly lower than that of fluoro-CT-guided biopsies (16.2±10.0 mSv) but the difference of effective dose between the two techniques was not statistically significant (Mann–Whitney U test, p=0.330). Unfortunately the dose estimates reported in different studies of CT-guided TNB are scarcely comparable because of biopsy procedure variables that may impact on radiation exposure and because of methodological differences in estimating the effective dose [28, 31]; therefore, further investigations of radiation exposure

Table 6 Sensitivity, specificity and accuracy of malignant diagnosis of CBCT-guided lung nodule biopsy in recently published series

comparing the CBCT and the fluoro-CT procedures are needed. This study has several limitations. First, it was retrospective and nonrandomized. However, our three interventional radiologists were assigned TNB sessions on rotation, either with the conventional fluoro-CT technique or with the CBCT system, without selecting nodules to biopsy with the new CBCT technique. Overall there were fewer TNB sessions assigned in the angiography suite equipped with CBCT technology; thus, the total number of CBCT-guided biopsies (123) was lower than that of fluoro-CT-guided ones (201). Notably, a wide variety of different nodules were biopsied over the years 2010–2013 in both groups, which turned out to be well matched and similar in terms of patient demographics, BMI, mean values of nodule characteristics and distribution of histological diagnoses (Tables 1, 2, and 3). Thus, comparison of the performance of fluoro-CT-guided and CBCTguided biopsy in our study seems reasonable, although the design was nonrandomized. Second, 8 % of CBCT-guided biopsies and 9 % of fluoro-CT guided proved to be inadequate for pathologic diagnosis, partly because of unavailability of on-site cytopathology examination at our unit. Another limitation is the relatively small sample size of our CBCTguided group (123 biopsies) compared to the 1153 biopsies reported by Lee et al. [15]; nevertheless, the accuracy of CBCT-guided biopsy was similar in our study and in Lee et al.’s [15] (96 % and 97 %). A strength of our study is the direct comparison between the CBCT system and the fluoro-CT system in a single centre with the same group of interventional radiologists, which likely reduced bias of differing procedure details and biopsy tools. In summary, the results of our study indicate that the accuracy, safety and effective dose for fluoro-CT-guided and CBCT-guided lung biopsy procedures are similar. Accordingly, in clinical practice these biopsy systems may be alternatively used to optimize the available technological resources.

CBCT biopsies, n

Sensitivity

Specificity

Hwang et al., 2010 [19] Cheung et al., 2011 [13] Braak et al., 2012 [23] Choi et al., 2012 [14]

27 74 84 99

94 % 91 % 90 % 95.8 %

89 100 100 100

Lee et al., 2014 [15] Our series

1.153 58a

95.7 % 95 %

100 % 100 %

% % % %

Accuracy

NPV

92 % 93 % 91.7 % 97 %

89 % NA 66.7 % NA

97 % 96 %

NA 78 %

CBCT cone-beam computed tomography, CT computed tomography, NPV negative predictive value a

Statistical analysis made on the 58 biopsies performed after completion of learning period

Eur Radiol Acknowledgments The scientific guarantor of this publication is Gianpaolo Carrafiello. The authors of this manuscript declare no relationships with any companies whose products or services may be related to the subject matter of the article. The authors state that this work has not received any funding. One of the authors (Raffaele Novario) has significant statistical expertise. Institutional review board approval was obtained. Written informed consent was obtained from all subjects (patients) in this study. Some study subjects or cohorts have been previously reported in Floridi C, Muollo A, Fontana F et al (2014) C-arm cone-beam computer tomography needle path overlay for percutaneous biopsy of pulmonary nodules. Radiol Med. doi:10.1007/s11547-014-0406-z Methodology: retrospective, observational, performed at one institution.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

Smith-Bindman R, Miglioretti DL, Johnson E et al (2012) Use of diagnostic imaging studies and associated radiation exposure for patients enrolled in large integrated health care system, 1996– 2010. JAMA 307:2400–2409 Schmidt RL, Witt BL, Lopez-Calderon LE et al (2013) The influence of rapid onsite evaluation on the adequacy rate of fine-needle aspiration cytology. A systematic review and meta-analysis. Am J Clin Pathol 139:300–308 Priola AM, Priola SM, Cataldi A et al (2007) Accuracy of CTguided transthoracic needle biopsy of lung lesions: factors affecting diagnostic yield. Radiol Med 112:1142–1159 Tomiyana N, Yasuhara Y, Nakajima Y et al (2006) CT-guided needle biopsy of lung lesions: a survey of severe complication based on 9783 biopsies in Japan. Eur J Radiol 59:60–64 Hiraki T, Mimura H, Gobara H et al (2009) CT fluoroscopy-guided biopsy of 1,000 pulmonary lesions performed with 20-gauge coaxial cutting needles: diagnostic yield and risk factors for diagnostic failure. Chest 136:1612–1617 Tam AL, Mohamed A, Pfister M et al (2010) C-arm cone beam computed tomography needle path overlay for fluoroscopic guided vertebroplasty. Spine 35:1095–1099 Braak SJ, van Stijen MJL, van Es HW, Nievelstein RAJ, van Heesewijk JPM (2011) Effective dose during needle interventions: cone-beam CT guidance compared with conventional CT guidance. J Vasc Interv Radiol 22:455–461 Carrafiello G, Dizonno M, Colli V et al (2010) Comparative study of jaws with multislice computed tomography and cone-beam computed tomography. Radiol Med 115:600–611 Carrafiello G, Fontana F, Mangini M et al (2012) Initial experience with percutaneous biopsies of bone lesions using XperGuide conebeam CT (CBCT): technical note. Radiol Med 117:1386–1397 Floridi C, Muollo A, Fontana F et al (2014) C-arm cone-beam computer tomography needle path overlay for percutaneous biopsy of pulmonary nodules. Radiol Med. doi:10.1007/s11547-0140406-z Strocchi S, Colli V, Conte L et al (2012) Multidector CT fluoroscopy and cone-beam CT-guided percutaneous transthoracic biopsy: comparison based on patient doses. Radiat Prot Dosim 151:162– 165 Jin KN, Park CM, Goo JM et al (2010) Initial experience of percutaneous transthoracic needle biopsy of lung nodules using C-arm cone-beam CT system. Eur Radiol 20:2108–2115 Cheung JY, Kim Y, Shim SS et al (2011) Combined fluoroscopyand CT-guided transthoracic needle biopsy using a C-arm conebeam CT system: comparison with fluoroscopy-guided biopsy. Korean J Radiol 12:89–96

14.

Choi MJ, Kim Y, Hong YS et al (2012) Transthoracic needle biopsy using a C-arm cone-beam CT system: diagnostic accuracy and safety. Br J Radiol 85:182–187 15. Lee SM, Park CM, Lee KH et al (2014) C-arm cone-beam CTguided percutaneous transthoracic needle biopsy of lung nodules: clinical experience in 1108 patients. Radiology 27:291–300 16. Christner JA, Kofler JM, McCollough CH (2010) Estimating effective dose for CT using dose-length product compared with using organ doses: consequences of adopting International Commission on Radiological Protection publication 103 or dual–energy scanning. Am J Roentgenol 194:881–889 17. D’Agostino RB, Belanger A, D’Agostino RB Jr (1990) A suggestion for using powerful and informative tests of normality. Am Stat 44:316–321 18. Li-Han H, Chia-Chuan L, Jen-Sheng K (2004) Education and experience improve the performance of transbronchial needle aspiration. A learning curve at a cancer center. Chest 125:532–540 19. Hwang HS, Chung MJ, Lee JW et al (2010) C-arm cone-beam CTguided percutaneous transthoracic lung biopsy: usefulness in evaluation of small pulmonary nodules. Am J Roentgenol 195:400–407 20. Gupta R, Cheung AC, Bartling SH et al (2008) Flat-panel volume CT: fundamental principles, technology, and applications. Radiographics 28:2009–2022 21. El-Sheik M, Heverhagen JT, Alfke H et al (2001) Multiplanar reconstructions and three-dimensional imaging (computed rotational osteography) of complex fractures by using a C-arm system: initial results. Radiology 221:843–849 22. Kothary N, Lock L, Sze DY et al (2009) Computed tomographyguided percutaneous needle biopsy of pulmonary nodules: impact of nodule size on diagnostic accuracy. Clin Lung Cancer 10:360– 363 23. Braak SJ, Herder GJM, van Heesewijk JPM, van Strijen MJL (2012) Pulmonary masses: initial results of cone-beam CT guidance with needle planning software for percutaneous lung biopsy. Cardiovasc Intervent Radiol 35:1414–1421 24. Khan MF, Straub R, Moghaddam SR et al (2008) Variables affecting the risk of pneumothorax and intrapulmonal hemorrhage in CTguided transthoracic biopsy. Eur Radiol 18:1356–1363 25. Covey AM, Gandhi R, Brody LA et al (2004) Factors associated with pneumothorax and pneumothorax requiring treatment after percutaneous lung biopsy in 443 consecutive patients. J Vasc Interv Radiol 15:479–483 26. Yeow KM, Su IH, Pan KT et al (2004) Risk factors of pneumothorax and bleeding. Multivariate analysis of 660 CT-guided coaxial cutting needle lung biopsies. Chest 126:748–754 27. Saji H, Nakamura H, Tsuchida T et al (2002) The incidence and risk of pneumothorax and chest tube placement after percutaneous CTguided lung biopsy. Chest 121:1521–1526 28. Choi JW, Park CM, Goo JM et al (2012) C-arm cone-beam CTguided percutaneous transthoracic needle biopsy of small (≤20 mm) lung nodules: diagnostic accuracy and complications in 161 patients. Am J Roentgenol 199:322–330 29. Kim JI, Park CM, Lee SM, Goo JM (2015) Rapid needle-out patient-rollover approach after cone beam CT-guided lung biopsy: effect on pneumothorax rate in 1,191 consecutive patients. Eur Radiol. doi:10.1007/s00330-015-3601-y 30. Choo JY, Park CM, Lee NK, Lee SM, Lee HJ, Goo JM (2013) Percutaneous transthoracic needle biopsy of small (≤1 cm) lung nodules under C-arm cone-beam CT virtual navigation guidance. Eur Radiol 23:712–719 31. Kim GR, Hur J, Lee SM et al (2011) CT fluoroscopy-guided lung biopsy versus conventional CT-guided lung biopsy: a prospective controlled study to assess radiation doses and diagnostic performance. Eur Radiol 21:232–239