Nondiagnostic computed tomography-guided percutaneous lung biopsies are more likely when infection is suspected

O RIGINAL A RTICLE Nondiagnostic Computed Tomography–guided Percutaneous Lung Biopsies Are More Likely When Infection Is Suspected Brian M. Haas, MD,* Brett M. Elicker, MD,* Janet Nguyen, MD,* Karen G. Ordovas, MD,* Kirk D. Jones, MD, w Travis S. Henry, MD,* and David M. Naeger, MD* P carry certain risks: most commonly, pneumothorax, and rarely death. 2–4 An additional “risk” is that the biopsy procedure may not yield diagnostic material, either due to a technical failure, insufficient material being obtained, sam- pling error, or inability of the pathologist to confidently identify the true abnormality. 2,5–7 Such an outcome is var- iably called “negative,” “nondiagnostic,” or “nonspecific” in the literature. The rate of nondiagnostic CTLB reported in the lit- erature ranges from 15% to 22%. 2,8 Variability in biopsy success has been attributed to technique, including number of passes, 8 the use of core versus fine-needle aspiration sampling, 9–11 needle trajectory, 12 and having a pathologist present to evaluate the sample. 13 Other groups have pub- lished that nodule characteristics affect the nondiagnostic rate of CTLB, with small size and presence of necrosis leading to lower rates of success. 2,5,8,14 To our knowledge, no studies have investigated the association between the pretest likelihood for cancer versus infection and the non- diagnostic rate. Overall, understanding the nondiagnostic rate on the basis of a set of predictive factors is important to accurately consent patients and to inform referring clinicians who must decide between various diagnostic options to obtain a diagnosis. In this study, we had 2 primary aims. First, we sought to assess the rates of nondiagnostic results on the basis of the pretest suspicion for cancer, infection, or for lesions considered clinically uncertain. Second, we aimed to determine the ultimate clinical diagnosis in patients with nondiagnostic CTLBs, overall and stratified by the pre- biopsy likelihood of the lesion being cancer versus infection. We hypothesized that our overall nondiagnostic rate would be comparable to other institutions. We hypothe- sized that a high pretest likelihood of infection would be correlated with a higher nondiagnostic biopsy rate. We also hypothesized that nondiagnostic biopsies would most commonly represent infections, particularly in cases in which infection was initially suspected. From the Departments of *Radiology and Biomedical Imaging; and wPathology and Laboratory Medicine, University of California, San Francisco, San Francisco, CA. The authors declare no conflicts of interest. Correspondence to: David M. Naeger, MD, Department of Radiology and Biomedical Imaging, University of California, San Francisco, 505 Parnassus Ave., M-391, Box 0628, San Francisco, CA 94143- 0628 (e-mail: david.naeger@ucsf.edu). Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/RTI.0000000000000207 The institutional review board approved this study and waived requirement of informed consent for this HIPAA-compliant study protocol. All patients presenting for CTLB over a 5-year period between October 1, 2009 and September 30, 2014 at 2 clinical sites were identified for review. The 2 clinical sites included an academic tertiary care hospital and an affiliated Veterans Affairs medical center. All biopsy requests were Purpose: The purpose of this study was to assess the incidence of nondiagnostic computed tomography–guided lung biopsy results, stratified by biopsy indication, and determine the final diagnosis in such cases. Materials and Methods: Following institutional review board approval, pathology results from CT-guided lung biopsies over a 5- year period at 2 institutions were categorized as diagnostic or nondiagnostic. Each biopsy’s indication was categorized as being for a lesion considered likely to be cancer, infection, or uncertain. For all nondiagnostic biopsies, the medical chart was reviewed to determine the final clinical diagnosis. Results: A total of 660 biopsies were evaluated, 139 (21%) of which were nondiagnostic. Of these 139 patients, the final clinical diag- nosis was infection in 37%, cancer in 30%, and a benign non- infectious diagnosis in 10%; 23% remained undiagnosed at last available follow-up. Among the patients in whom there was a high pretest suspicion for cancer, 13% were nondiagnostic, 45% of which were cancer and 27% were infection. Among biopsies of lesions with pretest probability for both cancer and infection, 51% were nondiagnostic; on clinical follow-up these were determined to be infection in 34% and cancer in 14%. When there was high pretest suspicion for infection, 73% were nondiagnostic, of which 13% were cancer on clinical follow-up, and 88% were infection. The rate of nondiagnostic biopsies was statistically significantly different (P < 0.001) among the 3 groups. Conclusions: Nondiagnostic biopsies are common and occur most frequently when there is a moderate or high pretest suspicion for infection. Among all nondiagnostic biopsies, regardless of indica- tion, cancer and infection were diagnosed on follow-up in similar proportions. Key Words: percutaneous lung biopsies, chest biopsies, non- diagnostic, pneumonia, lung cancer (J Thorac Imaging 2016;31:151–155) ercutaneous computed tomography (CT)–guided lung biopsy (CTLB) is an important way to establish a diagnosis of cancer and, in some cases, infection. 1 It is well known that these types of biopsies, although generally safe, MATERIALS AND METHODS J Thorac Imaging Volume 31, Number 3, May 2016 www.thoracicimaging.com | Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.