Applying a Text-Search Algorithm to Radiology Reports Can Find More Patients With Pulmonary Nodules Than Radiology Coding Alone
Introduction: Chest imaging often incidentally finds indeterminate nodules that need to be monitored to ensure early detection of lung cancers. Health care systems need effective approaches for identifying these lung nodules. We compared the diagnostic performance of 2 approaches for identifying patients with lung nodules on imaging studies (chest/abdomen): (1) relying on radiologists to code imaging studies with lung nodules; and (2) applying a text search algorithm to identify references to lung nodules in radiology reports.
Methods: We assessed all radiology studies performed between January 1, 2016 and November 30, 2016 in a single Veterans Health Administration hospital. We first identified imaging reports with a diagnostic code for a pulmonary nodule. We then applied a text search algorithm to identify imaging reports with key words associated with lung nodules. We reviewed medical records for all patients with a suspicious radiology report based on either search strategy to confirm the presence of a lung nodule. We calculated the yield and the positive predictive value (PPV) of each search strategy for finding pulmonary nodules.
Results: We identified 12,983 imaging studies with a potential lung nodule. Chart review confirmed 8,516 imaging studies with lung nodules, representing 2,912 unique patients. The text search algorithm identified all the patients with lung nodules identified by the radiology coding (n = 1,251) as well as an additional 1,661 patients. The PPV of the text search was 72% (2,912/4,071) and the PPV of the radiology code was 92% (1,251/1,363). Among the patients with nodules missed by radiology coding but identified by the text search algorithm, 130 had lung nodules > 8 mm in diameter.
Conclusions: The text search algorithm can identify additional patients with lung nodules compared to the radiology coding; however, this strategy requires substantial clinical review time to confirm nodules. Health care systems adopting nodule-tracking approaches should recognize that relying only on radiology coding might miss clinically important nodules.
Rapid advances in imaging technology have led to better spatial resolution with lower radiation doses to patients. These advances have helped to increase the use of diagnostic chest imaging, particularly in emergency departments and oncology centers, and in screening for coronary artery disease. As a result, there has been an explosion of incidental findings on chest imaging—including indeterminate lung nodules.1,2
Lung nodules are rounded and well-circumscribed lung opacities (≤ 3 cm in diameter) that may present as solitary or multiple lesions in usually asymptomatic patients. Most lung nodules are benign, the result of an infectious or inflammatory process. Nodules that are ≤ 8 mm in diameter, unless they show increase in size over time, often can be safely followed with imaging surveillance. In contrast, lung nodules > 8 mm could represent an early-stage lung cancer, especially among patients with high-risk for developing lung cancer (ie, those with advanced age, heavy tobacco abuse, or emphysema) and should be further assessed with close imaging surveillance, either chest computed tomography (CT) alone or positron-emission tomography (PET)/CT, or tissue biopsy, based on the underlying likelihood of malignancy.
Patients who receive an early-stage lung cancer diagnosis can be offered curative treatments leading to improved 5-year survival rates.3,4 Consequently, health care systems need to be able to identify these nodules accurately, in order to categorize and manage them accordingly to the Fleischner radiographic and American College of Chest Physicians clinical guidelines.5,6 Unfortunately, many hospitals struggle to identify patients with incidental lung nodules found during diagnostic chest and abdominal imaging, due in part to poor adherence to Fleischner guidelines among radiologists for categorizing pulmonary nodules.7,8
The Veterans Health Administration (VHA) system is interested in effectively detecting patients with incidental lung nodules. Veterans have a higher risk of developing lung cancer when compared with the entire US population, mainly due to a higher incidence of tobacco use.6 The prevalence of lung nodules among veterans with significant risk factors for lung cancer is about 60% nationwide, and up to 85% in the Midwest, due to the high prevalence of histoplasmosis.7 However, only a small percentage of these nodules represent an early stage primary lung cancer.
Several Veterans Integrated Service Networks (VISNs) in the VHA use a radiology diagnostic code to systematically identify imaging studies with presence of lung nodules. In VISN 23, which includes Minnesota, North Dakota, South Dakota, Iowa, and portions of neighboring states, the code used to identify these radiology studies is 44. However, there is high variability in the reporting and coding of imaging studies among radiologists, which could lead to misclassifying patients with lung nodules.8
Some studies suggest that using an automated text search algorithm within radiology reports can be a highly effective strategy to identify patients with lung nodules.9,10 In this study, we compared the diagnostic performance of a newly developed text search algorithm applied to radiology reports with the current standard practice of using a radiology diagnostic code for identifying patients with lung nodules at the Iowa City US Department of Veterans Affairs (VA) Health Care System (ICVAHCS) hospital in Iowa.
