Population-based assessment of the national comprehensive cancer network recommendations for baseline imaging of hepatocellular carcinoma.

This analysis aims to evaluate the performance characteristics of alternative baseline imaging thresholds in a cohort of hepatocellular carcinoma (HCC) patients from the Surveillance, Epidemiology, and End Results (SEER) database. HCC patients within the SEER database (2010-2015) who had complete information on clinical T and N stages as well as complete information on metastatic sites were eligible for the current study. Various performance characteristics associated with baseline imaging were investigated, including specificity, sensitivity, positive likelihood ratio (LR), negative LR, number needed to investigate (NNI), negative predictive value (NPV), positive predictive value (PPV), and accuracy. A total of 27,201 HCC patients were included. Based on current recommendations that advocate for the use of cross-sectional chest imaging in all newly diagnosed cases of HCC, these recommendations would yield a PPV of 5.0% for the detection of lung metastases. This would translate to an NNI of 20.0. When T1N0 patients were excluded from routine chest or bone imaging, this resulted in a PPV of 6.8% for the identification of lung metastases and an NNI of 14.7. Likewise, this translated to a PPV of 4.6% for the identification of bone metastases and an NNI of 21.7. Similarly, when patients with T1N0 disease and normal alpha-fetoprotein (AFP) were excluded from routine imaging, this resulted in a PPV of 5.6% for the identification of lung metastases and an NNI of 17.8. Also, this translated to a PPV of 3.8% for the identification of bone metastases and an NNI of 26.3. The current study suggests that the omission of routine baseline chest imaging may be considered in selected patients with asymptomatic early-stage HCC and normal AFP.