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.
Text Search Adjustments
Refining the text search criteria algorithm and the chart review process may decrease the rate of false positives significantly without affecting detection of lung nodules. In subsequent simulations, we found that by adding an exclusion criteria to text search algorithm to remove reports with specific keywords we could substantially reduce the number of false positive reports without affecting the detection rate of the lung nodules. These exclusion criteria would exclude any reports that: (1) contain “nodul” within the next 8 words after mentioning “no”; (2) contain “clear” within the next 8 words after mentioning “lung” in the text (eg, “lungs appear to be clear”); (3) contain “clear” within the next 4 words after mentioning “otherwise” in the text (eg, “otherwise appear to be clear”). Based on our study results, we further refined the text search strategy by limiting the search to only chest imaging studies. When we applied the revised algorithm to a random sample of imaging reports, we found all the code 44 radiology reports were still captured, but we were able to reduce the number of radiology reports needing review by about 80%.
Although classification approaches are being refined to improve radiology performance in multiple categories of nodules, this study suggests that alternative approaches based on text algorithms can improve the capture of pulmonary nodules that require surveillance. These algorithms also can be used to augment radiologist reporting systems. This represents an investment in resources to build a team that should include a bioinformatics specialist, lung nodule registry personnel (review charts of the detected imaging studies with lung nodules, populating the lung nodule database, and determining and tracking the need of imaging follow up), a lung nodule clinic nurse coordinator, and a dedicated lung nodule clinic pulmonologist.
Radiology departments could employ this text search approach to identify missed nodules and use an audit and feedback system to train radiologists to code lung nodules consistently at the time of the initial reading to avoid delays in identifying patients with nodules. Alternatively, the more widespread use of a standardized CT chest radiology reports using Fleischner or the American College of Radiology Lung Imaging Reporting and Data System (Lung RADS) templates might improve the detection of patients with lung nodules.5,13,14
