Repeat needle biopsies combined with clinical observation are safe and accurate in the management of a solitary pulmonary nodule

Abstract

BACKGROUND

Sixty percent of removed solitary pulmonary nodules (SPNs) are benign. An approach that reduces the unnecessary excision of benign nodules is consistent with the oncologic objective of organ preservation.

METHODS

A prospective observational study was performed at a lung cancer referral center in which consecutive patients were evaluated who presented with SPNs measuring < 4 cm on computed tomography (CT) scans. Patients underwent transbronchial biopsy (TBB), percutaneous needle aspiration (PCNA), clinical observation, repeat CT scans, and repeat biopsies. Patients were observed clinically and underwent repeat biopsies in an effort to reduce unnecessary surgical intervention.

RESULTS

One hundred eighteen patients underwent 194 biopsy sessions, including 137 TBB sessions and 57 PCNA sessions. The mean follow-up was 4 years. The shortest follow-up of a benign lesion was 3 years. The incidence of malignancy was 61%. The positive predictive value, negative predictive value, sensitivity, specificity, and accuracy all were 100%. Five patients had a delayed change in diagnosis from benign to malignant. This delay in diagnosis neither resulted in a change in tumor stage nor had an impact on patient management or survival.

CONCLUSIONS

Repeat needle biopsies combined with clinical observation and repeat CT scans can classify an SPN as benign versus malignant with 100% accuracy (95% confidence interval, 96.1–100.0%). An SPN diagnostic approach that includes a TBB, then PCNA, clinical observation, repeat CT scans, and repeat biopsies for continued suspicion of malignancy appears to reduce the unnecessary surgical excision of benign nodules from the current rate of 60% to 5% of SPN resections without affecting the survival of patients who have malignant SPNs. Cancer 2005. © 2005 American Cancer Society.

Recent advances have improved the methods available for the evaluation of the solitary pulmonary nodule (SPN). The presence of benign lesions in 60% of removed nodules indicates a need for an approach that better enables lung preservation.1 It has been determined that both computed tomographic (CT) scans and needle biopsies are cost effective in the management of the SPN and should be performed routinely.2 Other authors are not convinced of the value of the needle biopsy in the evaluation of SPNs and demand additional evidence prior to supporting its use.3

When screening for lung carcinoma, the ability of chest CT scans to identify surgically resectable tumors in 90% of newly identified malignancies is considerably better than the 22% rate of chest X-rays and has generated considerable enthusiasm for this screening technique.4 Although screening CT is awaiting evidence identifying its ability to reduce mortality, medical economic authors are embracing this practice as cost effective.5, 6 Although the National Lung Screening Trial will not provide mortality data until the year 2013, this support from economists, along with the enhanced ability of screening CT scans to identify malignancies while they are surgically resectable, is creating a movement toward the ubiquitous offering of screening chest CT scans to tobacco smokers.

Pulmonary nodules are identified in 23% of patients age ≥ 60 years who undergo a screening chest CT scan.4 However, only 2.7% of screened patients age ≥ 60 years4 and 1.1% of patients age > 40 years have a malignancy.7 In addition, subsequent annual screening CT scans identify pulmonary nodules in 2.5% of patients and identify malignancy in 0.6% of patients.8 If, as screening becomes more common, a strategy is followed that considers needle biopsy techniques inadequate, then the rate of benign nodule resection would increase from the current rate of 60% to closer to 98%. Our screening program evaluated 123 patients last year. Without needle biopsy, this would result in 120 unnecessary surgeries per year at our institution alone and demands evaluation of less invasive diagnostic techniques.

The two commonly used biopsy techniques include the transbronchial biopsy (TBB) and percutaneous needle aspiration biopsy (PCNA). TBB is performed through fiberoptic bronchoscopy with or without fluoroscopic guidance. PCNA is performed under fluoroscopic or CT guidance. The combination of these techniques reportedly has a sensitivity of 95.2% and a specificity of 99.5%, with a complication rate of 4.8% for TBB and 14.4% for PCNA. Severe complications, defined as complications that require treatment, were limited to 3.5%, and no patients suffered permanent morbidity or death.9 In a similar study, the diagnostic yield of this combination of techniques was 90.3%.10 Another study evaluating the role of CT-guided PCNA alone in evaluation of the SPN reported similar sensitivity, specificity, and severe complication rates.11

The evaluation of an SPN requires a multidisciplinary approach. The clinician, radiologist, and pathologist all need to communicate to reach the correct diagnosis. An approach that utilizes close clinical observation along with diagnostic studies to minimize the unnecessary excision of benign nodules has been recommended by others.12 This approach is consistent with the current oncologic management objective of organ preservation. The importance of this minimally invasive approach will increase as additional SPNs are identified by increasingly available and affordable CT scanners and new reading techniques.13 The objective of this study was to evaluate the efficacy of a diagnostic strategy that utilizes repeat needle biopsies and clinical observation with routine CT scans in an attempt to determine the etiology of the SPN and to reduce unnecessary surgical resections.

MATERIALS AND METHODS

Consecutive patients who presented with an SPN that measured ≤ 4 cm on a CT scan were included in the study. Patients with characteristic benign patterns of calcification on a CT scan or with stability of nodular size for > 2 years were excluded from further investigation based on the high likelihood that these were benign nodules. Patients who had an active, previously diagnosed malignancy were excluded, whereas patients who had a previously diagnosed malignancy that was believed to be in remission were included. Patients who had mediastinal adenopathy that measured > 2 cm in greatest dimension were excluded. In addition to the radiologist, a member of the Lung Center analyzed the CT and radiographic studies. All masses were measured in three dimensions on radiographs (cephalad-caudad, ventral-dorsal, and medial-lateral) and in two dimensions on CT scans (ventral-dorsal and medial-lateral). In the event of a disagreement between the radiologist and the Lung Center specialist, a second member of the Lung Center reviewed the CT or radiograph. The agreement of the two specialists was accepted as the official reading. The location of the mass was defined by the segment from which sampling was performed during bronchoscopy.

From 1995 to 2000, all patients age > 40 years with an SPN, as defined above, were investigated consecutively using the same protocol. After an initial interview, a physical examination, and obtaining informed consent, each patient underwent CT of the chest. Next, a single operator performed fiberscopic bronchoscopy with transbronchial needle aspiration (TBNA) of both the parenchymal nodule and the mediastinal lymph nodes. If this initial biopsy failed to reveal a malignant diagnosis, then a PCNA was performed under CT guidance or fluoroscopy. If that technique also failed to reveal a malignant diagnosis, then the patient was reviewed. If it was found that there was potential to improve the adequacy of the sample obtained, then the patient underwent another biopsy using the method that the investigators believed offered the most information. This pursuit was stopped either when a malignant diagnosis was obtained or when the samples were the best obtainable. At that point, patients with a benign diagnosis were followed clinically, and chest CT scans were obtained at 3 months, 6 months, 12 months, and 24 months.

Transbronchial biopsy is performed in a dedicated bronchoscopy suite by a single pulmonologist. The patient lies in the supine position and undergoes conscious sedation, typically using fentanyl (Sublimaze) along with midazolam (Versed). Local anesthesia with 2% lidocaine spray in the throat and lidocaine jelly in the nares provides additional patient comfort. Patients who have features that create concerns for the possible need for conversion to general anesthesia (i.e., tracheal mass, laryngeal spasm, or severe cough) are studied in the presence of an anesthesiologist. A variety of bronchoscopes are used (models BF P40 and BF P20D; Olympus; Tokyo, Japan) along with the typical biopsy needles, forceps, brushes, and brush needles. After a complete inspection of the bronchial tree, including the subsegmental bronchi, the mass is visualized in multiple planes using a C-arm fluoroscope. Once precise localization is obtained, and multiple biopsies of the mass are performed with all of the noted instruments. All samples are prepared and delivered to the laboratory, as noted below. All patients undergo a postprocedure chest X-ray prior to discharge from the Lung Center.

Percutaneous needle biopsy typically is performed with CT guidance. Occasionally, pleural-based masses are sampled easily using fluoroscopic guidance. Regardless of the imaging technique used, the patient undergoes local anesthesia with 1% lidocaine. The biopsy needle (model PCN-1915; Mill Rose Laboratories, Inc., Mentor, OH) is a double needle with an inner 21-gauge needle and an outer 19-gauge needle and is 15 cm in length. This needle is positioned into the mass using radiology guidance. The inner needle is retracted and, with suctioned applied, the outer needle is advanced into three separate areas of the mass and is removed. This typically is performed once and occasionally is performed twice. All samples are prepared and delivered to the laboratory, as noted below. Immediately, CT images at two levels in the region of the biopsy are obtained in search of a pneumothorax. All patients undergo a postprocedure chest X-ray prior to discharge from the Lung Center. Further details of the techniques used for TBNA and PCNA at our institution were previously published.14, 15

A CT scan of the chest was obtained with a 120-KV, 220-mA tomoscan (model Somatom Plus 4; Seimens; Munich, Germany), which administers a total of 16.5 mGy of radiation. Our standard protocol includes a tomogram and images taken from the lung apices to the top of the kidneys at a speed of 1.0 seconds per rotation in helical motion. The slice thickness is 5 mm through the mediastinum (from the top of the aorta caudal to 2 cm below the bifurcation of the main stem bronchus) and 8 mm elsewhere. One hundred milliliters of Optiray 240 (St. Louis, MO) is administered as the intravenous contrast agent at 1.5 mL per second. The images are reconstructed using the Kernal protocols of AB50 for the mediastinum and AB82 Lo for the lung windows.

Pathologic specimens are sent as both tissue biopsies and cytology smears. The tissue specimens are placed in formalin and are transferred to the pathology laboratory, where they are examined macroscopically and placed in an automated processor (Shandon-Hypercenter XP; Cheshire, England). The tissue processor performs a standard automated sequence of fixation, dehydration, and paraffin infusion. The resultant blocks are sectioned on the microtome, heated, and stained with hematoxylin and eosin stain prior to review by the pathologist. Special stains are performed subsequently as needed.

All cytology specimens are obtained through needle aspiration, brushings, or washings but have two methods of delivery to the pathology laboratory. One set of specimens is smeared on slides and placed in 95% ethanol in the bronchoscopy suite. These specimens are delivered to the pathology laboratory, where they are stained in Papanicolou (Pap) stain and reviewed by the pathologist immediately. A second set of specimens is delivered in saline and centrifuged. If they are highly cellular, then conventional smears are fixed in 95% ethanol. If there is a paucity of cells, then the specimen is cytocentrifuged prior preparing smears and fixing in 95% ethanol. These slides also are stained with Pap stain prior to pathologist examination.

All patients undergo a postprocedure chest X-ray and are monitored for procedure-associated complications. Major complications were defined as any morbidity created by a procedure that required a treatment. Minor complications were defined as any morbidity created by a procedure that did not affect the management of the patient.

Statistics

Results from patients who underwent lung resection and had malignant preresection pathology findings with benign postresection pathology findings were categorized as false-positive results. True-positive results were defined as a needle biopsy-identified malignancy that was confirmed by the postresection pathology findings. True-negative results were defined as masses that were identified as benign on needle biopsy that did not progress during 2 years of follow-up. We believed that this was an acceptable substitution for the gold standard of comparison with open biopsy based on the morbidity of such a procedure and the rapidly progressive nature of lung carcinoma. False-negative results were identified as benign findings after needle biopsy but malignant findings after resection. Using these predefined terms, sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and diagnostics yield of malignancy were determined.

Using the Centers for Disease Control EPIINFO 2000 program (version 1.0; Atlanta GA), the chi-square test was used to identify significant differences in the patients' baseline characteristics. EPIINFO 6 (version 6.04) was used to determine confidence limits for the screening test results and overall accuracy using the Epitable component of the program, which uses the Fleiss quadratic approximation for calculating confidence intervals that are valid, even when the event proportion is near or at 100%.

RESULTS

One hundred patients underwent 194 biopsy sessions. A biopsy session was defined as the patient receiving an anesthesia, during which multiple passes of the needle, forceps, or brush may be used to obtain tissue. If the patient was allowed to recover and a second anesthesia was required for further sampling (occasionally on the same day), then this was considered another biopsy session. One hundred thirty-seven of these sessions utilized transbronchial techniques, and 57 sessions utilized PCNA. The mean patient age was 67 years (range, 40–86 years). Accurate clinical data were available for all patients with benign diagnoses as of February, 2004. The mean follow-up was 4 years (range, 1.5–7.0 years). The shortest follow-up of a patient with a benign lesion was 3 years. The right upper lobe was the most common location for nodules compared with the other lobes. There were 72 malignant lesions and 46 benign lesions identified, resulting in a 61% incidence of malignancy. Nonsmall cell lung carcinoma was the most common etiology of the SPN, with an incidence of 44% (Fig. 1). Adenocarcinoma was the most common malignant cell type, occurring in 35% of malignancies.

Figure 1. This chart illustrates the final pathologic diagnosis of solitary pulmonary nodules. A-V: arterial-venous; Ca: carcinoma.

Baseline characteristics were collected and assessed for differences among patients with benign and malignant disease. Age achieved a statistically significant difference, with patients age < 56 years having statistically significantly fewer malignancies compared with older patients (21% vs. 66%). Smoking history did not reach statistical significance; however, tobacco history was not available for 60 of 140 patients (42%). Patient gender and location of mass did not affect the incidence of malignancy.

There were 72 true-positive results, 46 true-negative results, and no false-positive or false-negative results (Table 1). This positive predictive value, negative predictive value, sensitivity, and specificity all were 100%. Ninety-eight patients (83%) had their mass identified correctly as either benign or malignant after the first biopsy session. Twenty-six patients (56%) with benign lesions underwent > 1 biopsy, but their diagnosis did not change. Fifty-two of 72 patients (72%) with true-positive results were identified in their first biopsy session. Overall, 74 of 118 patients (63%) required a single biopsy session to be diagnosed correctly. The accuracy began at 83% with the first biopsy session and increased to 100% by the third biopsy session (Table 2). The overall accuracy of this diagnostic process was 118 of 118 patients (100%; 95% confidence interval [95%CI], 96.1–100.0%).

Table 1. Performance of a Diagnostic Strategy Utilizing Repeat Needle Biopsies and Clinical Observation

TBB: transbronchial biopsy; PCNA: percutaneous needle aspiration; PPV: positive predictive value; NPV: negative predictive value.
No. of patients	118
No. of biopsy sessions	194
TBB (no.)	137
PCNA (no.)	57
No. with benign results	46
No. with malignant results	72
Incidence of malignancy (%)	61
No. of true-positive results	72
No. of true-negative results	46
No of false-positive results	0
No. of false-negative results	0
PPV (%)	100
NPV (%)	100
Sensitivity (%)	100
Specificity (%)	100

Table 2. Accuracy per Biopsy Session

Biopsy session	Accuracy (%)
1	98/118 (83)
2	33/39 (85)
3	16/16 (100)
4	5/5 (100)

Twenty patients had a change in diagnosis from an initially benign diagnosis to malignant on repeat biopsy sessions. Fifteen of those biopsies were performed within an acceptable time frame of 2 months (Table 3). The remaining 5 patients had a change in diagnosis after a range 5–22 months (Table 4). All five patients underwent repeat biopsies based on continued clinical suspicion and lack of resolution of CT findings. None of those five patients had a change in disease stage from the original CT findings. Patients 1 and 2 had Stage IA disease, underwent successful resection, and were disease free of their original tumor at 5 years. Patient 2 had the longest delay in diagnosis due to poor compliance with reporting to the office for appointments: It is noteworthy that this patient developed a second primary lung carcinoma of a different cell type in a new location 5 years after undergoing resection. Patient 3 was not a surgical candidate due to poor clinical status but, surprisingly, lived for 4.5 years after diagnosis. Patient 4 had extensive pulmonary fibrosis, was followed with repeat CT scans, and underwent a repeat biopsy due to the persistence and enlargement of an ipsilateral hilar lymph node from an initial measurement of 1 cm to 2 cm on repeat CT scan. This patient underwent a successful lung resection and was disease free 1.5 years later, when the patient was lost to follow-up due a move outside the United States. Patient 5 had a prolonged time to diagnosis due to poor compliance with scheduled office appointments. He was alive and disease free 1.5 years after undergoing surgical resection but subsequently was lost to follow-up. These delayed diagnoses did not have an impact on patient management, and all five patients had an excellent clinical course.

Table 3. Fifteen Patients Diagnosed with Malignant Masses < 2 Months After a Benign Initial Biopsy

Age (yrs)	Smoking (pack yrs)	Location	Tumor size (cm)	No. of biopsy attempts	Attempt interval (days)a	Malignancy
LUL: left upper lobe; NSCLC: nonsmall cell lung carcinoma; RUL: right upper lobe; RLL: right lower lobe; LLL: left lower lobe; RML: right middle lobe. a The attempt interval is the number of days between the first biopsy and the last biopsy.
67	40	LUL	3.0	2	30	NSCLC
69	50	RUL	2.3	2	3	NSCLC
57	68	LUL	3.0	2	9	NSCLC
60	80	LUL	4.0	2	2	NSCLC
60	80	RUL	2.5	3a	32	NSCLC
71	69	RUL	3.1	2	7	NSCLC
56	50	RUL	2.4	2	Same day	NSCLC
66	159	LUL	2.0	2	5	NSCLC
73	200	LUL	1.0	2	8	NSCLC
60	Unknown	RLL	2.6	3a	32	NSCLC
71	90	LLL	1.5	2	44	NSCLC
67	Unknown	LUL	3.0	2	4	Small cell
74	Unknown	RML	1.7	2	8	NSCLC
63	20	RLL	1.0	2	7	Small cell
63	105	RUL	3.7	2	4	NSCLC

Table 4. Five Patients Diagnosed with Malignant Masses > 2 Months After a Benign Initial Biopsy

Patient	Age (yrs)	Smoking (pack-yrs)	Location	Size (cm)	No. of biopsy attempts	Attempt interval (mos)	Malignancy	Stage	Classification	Treatment	Follow-up (yrs)
LUL: left upper lobe; CT: chemotherapy; RT: radiotherapy; RUL: right upper lobe; LLL: left lower lobe. a The patient developed small cell carcinoma in a different lung location. b Based on clinical status, the patient was not a good surgical candidate.
1	74	40	LUL	1.5	3	9	Squamous	IA	T1N0M0	Surgery	Alive (5.0)
2	68	132	LUL	1.0	3	22	Adenocarcinoma	IA	T1N0M0	Surgery	Alive (5.0)a
3	57	120	LUL	3.0	3	5	Squamous	IIB	T3N0M0	CT, RTb	Alive (4.5)
4	55	Unknown	RUL	2.3	2	17	Adenocarcinoma	IIA	T1N1M0	Surgery	Alive (1.5)
5	56	Nonsmoker	LLL	3.0	3	6	Bronchoalveolar	IIB	T2N1M0	Surgery	Alive (1.5)

Four patients underwent surgical resection for benign disease (Table 5). This represents 5% (4 of 76 patients) of our surgical resections. All four patients had benign preoperative pathology results but underwent surgery due to the specialist team's suspicion of malignant disease. Although needle biopsy failed to obviate the need for surgery, this was based on clinical suspicion and was not a biopsy failure. Therefore, the results from these four patients were analyzed as true-negative results. There were no surgical complications in these patients.

Table 5. Four Patients with Benign Masses who Underwent Surgical Resection

Age (yrs)	Smoking (packs per yr)	Location	Size (cm)	No. of attempts	Procedure diagnosis	Surgical diagnosis
RUL: right upper lobe; A-V: arterial-venous.
59	Unknown	RUL	2.3	1	Not diagnostic	A-V malformation
63	44	RUL	1.0	2	Not diagnostic	Granuloma
42	Non-Smoker	RUL	3.5	2	Pneumonia	Aspergilloma
70	Unknown	RUL	2.8	4	Not diagnostic	Pneumonia

The diagnostic yield is the percentage of biopsies that correctly identifies malignant disease. In the current study, the yield was 72% (18 of 26 biopsies) for PCNA, 70% (59 of 84 biopsies) for TBNA, and 100% (22 of 22 biopsies) for patients who underwent both PCNA and TBNA. The diagnostic yield was 72% (52 of 72 biopsies) for the first biopsy session, 75% (15 of 20 biopsies) for the second biopsy session, and 100% (5 of 5 biopsies) for the third biopsy session (Table 6). The diagnostic yield for PCNA and TBNA were raised from 72% to 100% by using both techniques and performing 3 biopsy sessions. Neither the size nor the location of the SPN affected the diagnostic yield.

Table 6. Diagnostic Yields

Biopsy technique	Yield (%)	Biopsy session	Yield (%)
PCNA: percutaneous needle aspiration; TBNA: transbronchial needle aspiration.
PCNA	18/26 (69)	First	55/75 (73)
TBNA	62/87 (71)	Second	15/20 (75)
PCNA and TBNA	22/22 (100)	Third	5/5 (100)

The only complications were pneumothorax. These complications were divided equally between major and minor complications and benign and malignant diagnoses. A major complication was defined as a pneumothorax requiring tube thoracostomy for drainage. A minor complication was defined as a pneumothorax that resolved spontaneously. The overall complication rate was 7.2% (14 of 194 patients). All complications were observed after a PCNA, which had a total complication rate of 24% (14 of 57 patients), and the complication rate was 0% for transbronchial biopsies. Eleven of 14 incidents of pneumothorax occurred during the first PCNA procedure, suggesting that the risk associated with an additional PCNA procedure is equal to that of the first PCNA procedure. Seven of 194 patients (3.6%) required chest tubes, all of which were removed and the patients discharged home after 24 hours. All other patients with pneumothorax were observed for 4 hours and underwent repeat chest X-rays. If the repeat X-ray documented an unchanged or resolving (< 15%) pneumothorax, then the patient was allowed to go home on the same day as the procedure. None of these patients required delayed management of the pneumothorax, and they were seen only as scheduled for their next outpatient visit.

DISCUSSION

Combined repeat needle biopsy and clinical observation is a reliable method for classifying SPNs as benign or malignant and can reduce the surgical resection of benign SPNs. In the current study, it was found that the described approach to the evaluation of the SPN had 100% accuracy (95%CI, 96.1–100.0%). Using this combined technique, 83% of patients were diagnosed correctly after the first biopsy session, and 91% of patients who had benign disease avoided undergoing surgical resection. Only 4 of our 76 patients (5%) who underwent surgical resection had benign nodules. This is considerably lower than previous reports of 50–60%.1 Most important, all patients were identified correctly with benign or malignant SPN prior to undergoing surgical resection, and these outcomes are similar to those reported previously.9, 11

The negative predictive value with this approach is 100% at our institution. The higher incidence of lung carcinoma in Baltimore (61%) compared with the rest of the United States (40%)16, 17 suggests that the negative predictive value will not be < 100% in other communities. A similar negative predictive value can be expected in patients with an SPN and a previous primary extrapulmonary neoplasm, because it has been demonstrated that this group of patients has an incidence of malignancy of 40%.18 Because the negative predictive value increases with the incidence of benign disease, clinicians can be confident that the method described herein will not fail to identify patients with malignant SPN in the United States.

Twenty patients had a change in diagnosis from benign to malignant on repeat biopsy sessions. The clinicians' continued suspicions of malignancy lead to subsequent biopsies that demonstrated this change in diagnosis. Five patients had a delay in diagnosis of > 2 months. Two of those patients missed scheduled appointments and were pursued until they returned for an additional biopsy session. The one disturbing instance of a delay in diagnosis was a patient who had extensive pulmonary fibrosis. It appears that this patient's fibrotic CT findings lowered the clinician's suspicion of malignancy, resulting in a prolonged observation time, and a second biopsy was delayed for 17 months after the original discovery of the nodule. This was unique, in that patients with persistent nodules typically underwent a second biopsy within 1 month after the first biopsy. These delayed diagnoses neither resulted in a change in disease stage nor had an impact on patient management. This is consistent with the previous demonstration that a delay in surgical intervention of SPNs does not effect survival significantly.19 This supports the approach of continued patient monitoring, consideration of additional biopsies, and reservation of thoracotomy for only those patients with confirmed malignancy. By decreasing the number of patients with benign masses who undergo thoracotomy, this management approach to SPNs lowers patient morbidity and allows lung preservation while having no deleterious effect on survival.

No patient in the current study required more than three biopsy sessions to obtain an accurate diagnosis. Repeat biopsy sessions and the combination of transbronchial and percutaneous techniques resulted in an increased biopsy yield. The need for multiple biopsy sessions may be reduced by the availability of immediate cytologic analysis. It has been shown that immediate cytologic analysis increases the sensitivity and specificity of a single PCNA biopsy session from 90% and 96.2%, respectively, to 98.5% and 100%, respectively.11 Consistent with most institutions that lack pathology residents, our institution is not staffed adequately with pathologists to perform immediate cytologic analyses routinely. Although the improved sensitivity and specificity observed in association with immediate cytologic analysis may decrease the total number of patients who undergo multiple biopsy sessions, it cannot be known whether it would have increased the diagnostic yield to 100% after 1 or 2 biopsy sessions in our patient population. Others have been using transthoracic and transbronchial ultrasound for guidance of both PCNA and TBNA.20, 21 It is unknown whether that technique is capable of achieving the same accuracy and yield described herein. The ability of ultrasound to obtain biopsies in individuals who are unable to undergo CT scanning, its wider availability, and its lower expense make ultrasound a very attractive modality.

We did not identify a variation in yield based on size or location of the SPN, consistent with previous results using CT-guided PCNA for SPNs that measure ≤ 1 cm.22, 23 It was found previously that the size and location of the SPN affects the diagnostic yield of TBNA when it is performed alone.24 The accuracy described herein is greater than that described for either PCNA alone or TBNA alone.22, 24 Both the increased accuracy and yield is attributable to the improved access to lesions of smaller size and more peripheral location created by the combination of percutaneous and transbronchial techniques.

This technique did not eliminate surgery in 100% of patients who had benign disease. The decision to proceed to surgical resection is made by clinicians who consider the biopsy results along with other factors. Four of 46 patients (9%) in the current study who had benign disease underwent an unnecessary surgical resection. However, because only 4 of 76 patients (5%) who underwent surgical resection had benign nodules, our technique reduced the patient morbidity rate compared with the 50–60% benign nodule resection rate reported elsewhere.1 The ultimate objective is to eliminate unnecessary surgical resections, which may be achieved better through the utilization of additional tests, such as positron emission tomography (PET) scanning, dynamic magnetic resonance imaging, and enhanced CT scanning.25–27 In this setting, PET scan reportedly has a sensitivity of 95% and a specificity of 70–85%28, 29; however, PET scanning was not part of our protocol, and the value it adds to the approach described herein needs to be determined.

The only complication seen with needle biopsy in this study was pneumothorax. There were no complications associated with transbronchial biopsy due to the ability of any postbiopsy air leak to escape though the bronchial airway. This preferential association of complications with transcutaneous needle biopsy is consistent with the literature, which has shown that, in addition to a pneumothorax rate of 8–20% and a chest tube insertion rate of 3–12%, additional reported complications include hemoptysis (2–5%) and pulmonary hemorrhage (1%). In addition, air embolism and tumor seeding of the biopsy tract have been reported rarely and probably occur at a rate of 1 in 5000 biopsies. Death rarely has been reported and typically is associated with air embolism or hemorrhage.30–33 The 24% complication rate (14 of 57 patients) of percutaneous needle biopsy and the 7 of 57 patients (12%) who required chest tubes in the current study indicates that PCNA must be performed only after a nondiagnostic transbronchial biopsy for all patients with nonpleural-based SPNs. This can be performed effectively by a skilled bronchoscopist, even in the presence of peripheral nodules, through fluoroscopic guidance. Utilization of the transbronchial approach first may enable avoidance of the PCNA procedure, thereby reducing the overall complication rate for the patient. It has been shown that using a 25-gauge needle reduces the occurrence of PCNA-associated pneumothorax.34 Whether this smaller needle gauge can achieve accuracy and yield similar to our data has yet to be determined.

In conclusion, repeat needle biopsies combined with clinical observation and repeat CT scans can be used by a specialist team to classify an SPN as benign versus malignant with 100% accuracy (95%CI, 96.1–10.00%). The diagnostic approach toward patients who present with an SPN should include a transbronchial needle biopsy, then percutaneous needle biopsy, clinical observation, and repeat CT scans along with repeat biopsies when a suspicion of malignancy persists. This approach appears to reduce the unnecessary surgical excision of lung tissue containing benign nodules from the current rate of 60% to 5% of SPN resections. In addition, it may lead to decreased morbidity in patients who have benign SPNs without affecting the survival of patients who have malignant SPNs. To achieve this level of success, the medical team must include a specialist with extensive experience in transbronchial and percutaneous needle biopsy techniques.

Acknowledgements

The authors offer to a special note of appreciation to Dr. Ko-Pen Wang. In preparation of this trial, they believed that the results could be trusted uniquely if the data acquisition, data analysis, and article preparation were performed by individuals who did not perform bronchoscopy and thus had no special interest in the success of the techniques utilized. Dr. Wang was sufficiently noble to perform the procedures without expecting sufficient involvement in preparation of the publication to be listed as an author. The authors also thank Norman Dubin, Ph.D., for performing statistical analysis.