| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Commentary |
1 Department of Radiology, University Hospital, University of Cincinnati, 234 Goodman St., ML 0742, Cincinnati, OH 45267-0761.
Received July 27, 2006; accepted after revision August 1, 2006.
Address correspondence to C. A. Meyer.
Keywords: biopsy • chest
What makes the article by Khouri et al. [1], "Transthoracic Needle Aspiration Biopsy of Benign and Malignant Lung Lesions," one of the top 100 most-cited in the AJR from the large number of articles published supporting the virtues of percutaneous imaging-guided transthoracic needle aspiration (TTNA) of lung lesions? Was it the first? No, in fact, TTNA was first reported more than 100 years earlier by Leyden [2] to diagnose an inflammatory process. In 1886, Menetrier used the technique to diagnose a pulmonary neoplasm. The application of fluoroscopically guided TTNA for the diagnosis of lung cancer was later popularized by Dahlgren and Nordenstrom [3] in the 1960s. Haaga and Alfidi [4] first described a CT-guided lung biopsy in 1976. In its modern form, this procedure had been around for 19 years before the publication of the Khouri et al. article.
Was it the largest study? Although an impressive study with 650 patients reported over an 11-year period, the total number of patients described by Khouri et al. [1] falls well short of the largest series reported to date. That honor goes to the study by Lalli et al. [5] in 1978 describing TTNA results in a series of 1,204 patients followed closely by Sagel et al. [6] that same year with a series of 1,153 patients. By 1984, there were at least 22 studies reporting TTNA results on a combined total of 5,515 patients [7].
Did it have the best results for diagnosing malignancy? No, Khouri et al. [1] reported a sensitivity of 94.7% and a specificity of 96% for diagnosing lung cancer. By that time, most articles were reporting sensitivities and specificities for malignancy of greater than 90%. Sagel et al. [6], in the study of 1,153 patients, reported a sensitivity of 96% and a specificity of 99% for bronchogenic carcinoma. In the work preceding the Khouri et al. article, attention was focused on the performance of TTNA as a tool for diagnosing lung cancer and, as such, reported ranges of sensitivity from 72% to 98% and specificity from 93% to 100%, false-positive rates from 0% to 18%, and false-negative rates from 5% to 100% [7].
Is this the most important data? Not really; the diagnosis of malignancy in a surgical candidate with a resectable lesion did not alter the patient's management—surgical resection. With the treatment options available in 1984, the preoperative diagnosis of cancer served one of two purposes: to convince a reluctant patient to have surgery or to establish the diagnosis in patients with advanced-stage disease before chemotherapy and radiation therapy. The greatest clinical impact of a positive biopsy result from TTNA occurs when the pretest likelihood is low [8]. Currently, with the advent of PET, the diagnosis of advanced disease can often be made from more remote sites (up to 20% of clinical stage I lung cancers), thereby avoiding the necessity of lung biopsy.
The true impact of the Khouri et al. [1] article was the revelation of the strong performance of TTNA in nonmalignant diagnoses. The authors divided their nonmalignant biopsy results into specific and nonspecific benign lesions as previously described [9]. Khouri et al. reported a specific benign diagnosis in 67.8% of lesions and a nonspecific benign diagnosis in another 19.7% of lesions; this was a landmark achievement! Until this article, most TTNA studies reported rates for specific benign diagnoses between 12% and 44% [10]. A "no malignant cells identified" result by a cytologist has roughly the same value as an intermediate probability ventilation-perfusion scan.
An inconclusive biopsy presents a clinical dilemma. In one series of 397 TTNA biopsies, 132 failed to reveal malignancy; of these 132, only 12% yielded a specific benign diagnosis, whereas 33% were subsequently found to have malignancy [11]. In addition to a high yield of specific benign diagnoses, the authors suggested specific criteria by which a nonspecific benign diagnosis would justify watchful waiting. When on-site cytopathologic review fails to diagnose malignancy, the addition of biopsies with an automated cutting needle has been shown to increase the yield for specific benign diagnosis to 75-91%, somewhat better than Khouri et al. [1] achieved with TTNA [12-14].
The patients we really help are those whom we can keep out of the operating room by yielding a specific benign diagnosis. In this classic article from 1985, Khouri et al. [1] calculated that TTNA could reduce the number of diagnostic thoracotomies performed to less than 8%. This issue becomes even more important in the era of lung cancer screening because the overwhelming number of incidental lung nodules are benign. Given the very high false-positive rate for nodules detected at screening CT (in excess of 98%), needle biopsy is a critical diagnostic tool to reduce the number of unnecessary thoracotomies [15].
In the past, the number of thoracotomies performed for benign disease has been reported to be as high as 50% [16]. Think these data are of historical interest only? A recent review of the Mayo Clinic CT lung cancer screening data reported that of the 55 patients who underwent surgical procedures, 18.1% had benign disease [17]. The surgical excision of nodules is not without risk. Complications in the entire Mayo Clinic cohort were reported in 27% of patients, with an operative mortality of 1.7% [17].
Outcome statistics for thoracic surgical procedures are well described in several series. In 1983, the Lung Cancer Study Group published mortality results of surgical resections for lung cancer: 1.5% for wedge resection, 3% for lobectomy, and 6% for pneumonectomy [18]. A more recent series in published in 1999 from the National Veterans Affairs Surgical Quality Improvement Program reported a 30-day mortality rate of 4% for lobectomy and 11.5% for pneumonectomy in a total of 3,516 patients [19]. Although clearly safer, the reported mortality of video-assisted thoracoscopic lobectomy is 1% [20, 21]. The principle of primum non nocere (first do no harm) remains paramount. Death rates from TTNA are reported to be up to 0.15%, although the majority of deaths were due to bleeding during cutting biopsy of diffuse lung disease [22]. For biopsy of focal disease, TTNA is an exceedingly safe procedure, with mortality closer to 0.02% [23]. Our goal should be to prevent the postoperative discussion that begins with, "Good news! It's not cancer."
So what is the rationale for TTNA in 2007? In combination with core biopsy, TTNA can establish a specific benign diagnosis and thereby prevent unnecessary surgery. Westcott et al. [24] reported that 50% of patients with nodules biopsied that were less than 1.5 cm in size avoided thoracoscopy or thoracotomy. Although it has been suggested that TTNA should be reserved for lesions with a low pretest probability of malignancy, at our institution, the a priori knowledge that a patient has malignancy allows the surgeon to proceed directly to lobectomy, significantly shortening operating and anesthesia time by avoiding the potentially time-consuming steps of sampling a lesion intraoperatively and awaiting a frozen section before definitive surgery. This is especially useful in the face of minimally invasive surgery. The argument that a positive biopsy does not alter management is, therefore, no longer valid. Furthermore, because we diagnose smaller lung cancers, the likelihood of long-term survival increases, as does the likelihood of developing metachronous primary bronchogenic carcinoma.
The incidence of second primary bronchogenic carcinomas is roughly 2.5% per year. Among peripheral tumors of less than 1.5 cm, a preoperative cytodiagnosis contributes to decision making for intentionally limited resection [25]. Lung-preserving surgery will be essential to maintain surgical options for future malignancy. As radiologists detect an increasing number of pulmonary nodules at CT with a resultant decline in the pretest likelihood of malignancy, the importance of TTNA and core needle biopsy techniques for specific benign diagnoses cannot be overemphasized.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Related articles in AJR:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
