Background

Lung cancer is the most frequently diagnosed cancer and continues to be the leading cause of cancer mortality among both males and females [1]. The 5-year survival rate of lung cancer is only 18%, largely due to late-stage diagnosis [1]. Thus, early diagnosis is especially critical to improve long-term survival. Biopsy is important for identification and confirmation of lung cancer. In clinical practice, conventional flexible bronchoscopy is supposed to be difficult for biopsy of small lesions in lung parenchyma or mediastinum. Therefore, we focused on the following approaches for biopsy according to a variety of lesion location in lung parenchyma, i.e., CT-guided PTNB(Percutaneous Transthoracic Needle Biopsy), ENB (Electromagnetic Navigation Bronchoscopy), EBUS-TBNA (Endobroncheal Ultrasonography-Transbronchial Needle Aspitation) and mediastinoscopy. Furthermore, the studies regarding liquid biopsies, e.g., CTC (Circulating Tumor Cell) detection are timely and hot, and warrant to be systematically reviewed.

Therefore, we evaluated the published studies in the last 20 years which focused on CT-guided PTNB, ENB, EBUS-TBNA, mediastinoscopy and CTC, aiming to reveal the clinical updates, development trends, detection rates and complications.

Methods

We used systematic review to identify relevant studies, assess study eligibility, evaluate methodological quality, and summarize findings regarding postoperative clinical outcomes. Medline and manual searches were performed by investigators CJD and FQD independently and jointly to identify all published articles in English journals from January 1, 2000 to November 9, 2017 that addressed the issues regarding detection of lung cancers by using CT-guided PTNB, ENB, EBUS-TBNA, mediastinoscopy and CTCs, respectively. The Medline search was done on PubMed (http://www.ncbi.nlm.nih.gov). The search strategies and yielded citations were shown in Tables 1 and 2, respectively. Investigators CJD and FQD performed the actual search and data abstraction.

Table 1 Data sources and searches regarding Clinical updates of approaches for biopsy
Full size table
Table 2 Information of yielded citations regarding approaches for biopsy
Full size table

Data abstraction

From the eligible articles, investigators CJD and FQD reviewed the following information, i.e., PMID, year of publication, study design, number of patients, average age of patients, nodules size and location, operation time, biomarkers for detection, diagnostic sensitivity, relative complication morbidity, treatment of complications, outcome and follow-up period.

Statistical analysis

The association between detection rate of ENB and nodule size, number of cases, operation time, average age of patients, sex, and mean distance of the lesions from the pleura was performed using Pearson’s correlation analysis. The impact of nodule location on detection rate of ENB was analyzed by using ANOVA analysis. The association between morbidity of pneumothorax following ENB and nodule size was performed using Pearson’s correlation analysis. The analyses were performed using SPSS Version 11.0 software for Windows (SPSS, Inc., Chicago, IL, USA). P < 0.05 (two-sided) was considered to indicate a statistically significant difference.

Results

CT-guided PTNB: Biopsy of lesion in lung parenchyma mapped on CT images

In last 20 years, the overall sensitivity, specificity, and accuracy of CT-guided PTNB were 92.52 ± 3.14%, 97.98 ± 3.28%, and 92.28% ± 5.40%, respectively. The top two complications of CT-guided PTNB were pneumothorax (1111/4822:23.04%) and hemorrhage (287/3503:8.19%), respectively. Two cases with severe complications were reported [2, 3]. Bronchial artery embolization was performed in one patient due to massive hemoptysis [3]. The other one suffered from cardiopulmonary arrest leading to death [2].

Diagnostic accuracy and incidence of complications seemed to be decreased [3,4,5] and increased [2,3,4,5,6,7,8,9], respectively, by smaller lesion size or longer needle path length (P < 0.05).

ENB: Biopsy of lesion in lung parenchyma and mediastinal area

The detection rate of lung cancer by ENB increased gradually (Fig. 1a) and was recently reported to be 96.8% [10]. There seemed to be no significant correlation between detection rate and number of cases, average age of patients, sex, nodule size, lobar location of nodule, mean distance from pleura to nodule and operation time. As shown in Fig. 1b, pneumothorax was the top one complication following ENB (86/1648:5.2%). In 86 pneumothorax cases, 34 cases (34/86) were administrated with closed drainage [10,11,12,13,14,15,16,17,18,19,20,21], and one case (1/86) was managed with manual aspiration and observation [19]. The other 51 cases with mild pneumothorax were discharged for rehabilitation. Intriguingly, the incidence of pneumothorax was significantly negatively correlated with nodule size (R = − 0.512, P = 0.018, Fig. 1c). The three hemorrhage cases were observed carefully without further intervention and were discharged for rehabilitation [16, 22]. Three cases of respiratory failure were reported without detailed depiction [16]. There were no ENB related death [10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30]. ENB plus EBUS imaging seemed to yield a higher detection rate as compared with sole use of ENB (59% vs. 88% [20] and 71.42% vs. 73.07% [11]). Surprisingly, studies combining fluoroscopy with ENB to confirm navigation success reported lower diagnostic yields (56.3 vs. 69.2% without fluoroscopy, p = 0.006) [31].

Fig. 1
figure 1

Analysis of clinical points regarding ENB. a Correlation between detection rate and publication time showing the detection rate increased gradually. b Pneumothorax was the top one complication following ENB (86/1648:5.2%). c The morbidity of pneumothorax was significantly negatively correlated with nodule size (R = −0.512, P = 0.018)

Full size image

EBUS-TBNA: Biopsy of lesion in subcarinal and bilateral hilar area

The detection rate of lung node by EBUS-TBNA remained to be 86.06 ± 9.70%. The diagnostic sensitivity, specificity, accuracy, positive predictive value and negative predictive value of EBUS-TBNA for the mediastinal staging of lung cancer were 85.48% ± 12.89%, 99.09% ± 3.15%, 92.88% ± 4.99%, 98.70% ± 3.03%, 83.03% ± 15.46%, respectively. As shown in Fig. 2a, the top three complications following EBUS-TBNA were hemorrhage (53/8662:0.61%), pneumothorax (46/12432:0.37%) and infection (34/11250:0.30%), respectively. Four hemorrhage cases were administrated with further intervention with one perioperative death. The other 49 cases with mild hemorrhage were discharged for rehabilitation [32, 33]. In 46 pneumothorax cases, nine cases (9/46) and 37 cases (37/46) were administrated with closed drainage and conservative treatment, respectively [32,33,34,35]. Perioperative mortality was relatively low (4/11189:0.04%). Besides the above mentioned one case died of severe hemorrhage, there was one case died of cerebral infarction and two unexplained deaths [32, 33, 36].

Fig. 2
figure 2

Analysis of clinical points regarding EBUS-TBNA and mediastinoscopy. a The top three complications following EBUS-TBNA were hemorrhage (53/8662:0.61%), pneumothorax (46/12432:0.37%) and infection (34/11250:0.30%), respectively. b The detection rate by using mediastinoscopy increased slightly. c The positive rate of 4thR (91.5% ± 9.35%) and 7th (80.56% ± 19.47%) lymph node by using mediastinoscopy were significantly higher than others (P < 0.05). d Hoarseness (67/4387:1.53%) was the top one complication following mediastinoscopy

Full size image

Mediastinoscopy: Biopsy of the lesion or lymph node in the vicinity of the trachea, the subcarinal and the bronchi area

The detection rate of lung cancer by mediastinoscopy increased slightly (Fig. 2b) which was reported to be 96% in recent years [37]. The diagnostic sensitivity, specificity, accuracy, positive predictive value and negative predictive value of mediastinoscopy for the mediastinal staging of lung cancer were 82.83% ± 10.63%, 100%, 93.98% ± 4.68%, 100%, 87.64% ± 13.00%, respectively. Intriguingly, the positive rates of 4thR (91.5% ± 9.35%) and 7th (80.56% ± 19.47%) lymph node were significantly higher than others (P = 0.03) (Fig. 2c). As shown in Fig. 2d, hoarseness (67/4387:1.53%) was the top one complication following mediastinoscopy. Among the abovementioned 67 cases with hoarseness, nine cases (9/67) suffered from permanent hoarseness, two cases (2/67) recovered partially by vocal cord medialization and six cases (6/67) recovered within a few months [37,38,39,40,41,42,43,44,45]. Perioperative mortality was relatively low (4/2137: 0.19%). The death causes among three cases were aortic laceration, stroke, and cardiac arrest, respectively, and one case die of unexplained cause [46].

CTC: Biopsies of tumor cells shed from solid tumor lesion into peripheral blood

The mean sensitivities of a variety of methods to detect CTC remained to be 63.05%. As shown in Fig. 3a, sensitivity of PCR seemed to be highest (78.81 ± 14.72%). Sensitivity of Density-gradient, ISET and Magnetic bead seemed to be higher than 60% (71.32% ± 2.8%, 67.75% ± 21.22% and 67.85% ± 25.24%, respectively). Specificity of ISET, PCR and Cell search was relatively high (100%, 90.88 ± 0.53% and 94.33% ± 9.82%, respectively). There was no published data regarding specificity of Magnetic bead and density-gradient.

Fig. 3
figure 3

Analysis of clinical points regarding CTCs. a Sensitivity of PCR seemed to be highest (78.81 ± 14.72%). Specificity of ISET, PCR and Cell search was relatively high (100%, 90.88 ± 0.53% and 94.33 ± 9.82%). b Sensitivity of Multimarker assay seemed to be highest(90%) including 17 target genes: AGR2, CEACAM5, MMP11, STRN3, CEACAM6, COL5A2, AMPH, CEACAM7, ABCC3, THY1, COL6A3, ENO1, PNN, SCFD1, KDELR3, KIAA0391, TACSTD1

Full size image

Intriguingly, there are a variety of biomarker combination for CTCs identification by using PCR yielding different sensitivities. As shown in Fig. 3b, the sensitivity of Multimarker assay seemed to be highest (90%). Besides, the sensitivity of the combination of TSA-9, KRT-19, Pre-proGRP was satisfactory (84.3%).

Discussion

Considering the exquisite anatomy of the mediastinum, hilar and lung parenchyma, the equipment and technique, e.g., percutaneous lung biopsy, ENB, EBUS-TBNA, and Mediastinoscopy developed quickly. Furthermore, liquid biopsy, e.g., CTC detection has been introduced and a few pilot studies regarding early diagnosis of lung cancer have been published [47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65]. According to application in specific location and situation, we systemic reviewed clinical updates of these approaches focusing on development trends, detection rate and complications .

CT-guided PTNB is regarded as an effective and feasible procedure to detect a difficult nodule with advantage of accurate positioning and high detection accuracy. Nevertheless, once the lesion diameter is less than 2 cm or the needle path length is more than 8 cm, the detection rate will drop dramatically [4]. In addition, the lesions in the vicinity of mediastinum vessels are challengers to clinicians with regards to safety. Currently, ENB is developed for biopsy of the lesions in deep lung parenchyma or mediastinum.

ENB is recommended in patients with lesions in lung parenchyma difficult to reach with conventional bronchoscopy or CT-guided PTNB. The detection rate of ENB increased gradually probably due to improvement of software and hardware. Eberhardt et al. [20] found nodule location has been noted to be an important factor in diagnostic yield, e.g., the yields from the lower lobes were significantly lower (29%; p = 0.01). However, Jensen et al. [22] found lobar location of nodule did not affect the diagnostic yield (p = 0.59). Therefore, we systematically analyzed the results of six studies mentioning detection rate and nodule location [14, 20, 22, 27, 29, 66], and found that there seemed to be no association between them (p = 0.433). The highest incidence of complication is pneumothorax (5.2%). However, pneumothorax following ENB was reported to be unrelated with age or sex [16], accordant with our results. Intriguingly, the incidence of pneumothorax seemed to be significantly negatively correlated with nodule size, probably due to difficulties varying with the size. Additionally, there was no reported ENB associated death, proving that ENB is relatively safe.

Empirically, EBUS-TBNA is suitable for biopsy of lesion in subcarinal and bilateral hilar area. EBUS-TBNA is also well utilized in the peripheral area with radial probe EBUS and in conjunction with ENB. As EBUS-TBNA has relatively high false negative rates, especially at station 4R or 7 lymph node, mediastinoscopy is still required for patients with suspicious nodal disease in these stations [67]. Cytological samples are usually taken by EBUS-TBNA, however, larger histological tissue samples are possible to obtain by mediastinoscopy.

Mediastinoscopy is always recognized as the gold standard for surgical staging of lung cancer which is suitable for biopsy in the vicinity of the trachea, the subcarinal and the bronchi area. Especially, the positive rate of station 4Rth (91.5 ± 9.35%) and 7th (80.56 ± 19.47%) lymph node were significantly higher than other stations (Fig. 2c). Nevertheless, as mediastinoscopy is an invasive approach, the incidences of complications are relatively remarkable.

CTC is a kind of liquid biopsies of tumor cells shed from solid tumor lesions (primary foci and metastases) into peripheral blood. Although the mean sensitivities of CTC detection were not satisfactory, the convenience of this non-invasive method seems to be incomparable. Sensitivity of PCR remained to be highest (78.81% ± 14.72%) as compared with other methods. Intriguingly, the sensitivities of PCR varies with combined biomarkers. Expectedly, the sensitivity of combination of multimarkers assay is highest (90%). Furthermore, the specificity of the three methods, i.e., ISET, PCR and Cell search, was relatively high (100%, 90.88% ± 0.53% and 94.33% ± 9.82%, respectively). Currently, CTC can be used as an auxiliary diagnostic method to provide a higher detection rate.

Conclusions

The biopsy approaches should be chosen according to a variety of location and situation of lesions. CT-guided PTNB is regarded as an effective and feasible procedure for biopsy in lung parenchyma, however, diagnostic accuracy and incidence of complications may be impacted by lesion size or needle path length. ENB has an advantage for biopsy of smaller and deeper lesions in lung parenchyma. ENB plus EBUS imaging can further improve the detection rate. EBUS-TBNA and mediastinoscopy can be recommended for the biopsy in lower and upper mediastinum, respectively. The former is relatively safer and the latter provides more tissue acquisition and better diagnostic yield of 4R and 7th lymph node. CTC detection can be considered for adjuvant diagnosis.