Overdiagnosis Risk Tempers Mortality Reduction Benefit In CT Lung Cancer Screening

More indolent lung tumours are detected using low-dose computed tomography than chest radiography

medwireNews: The risk of overdiagnosis is an important drawback of a potential lung cancer screening programme based on low-dose computed tomography (LDCT), researchers warn.

The team used data from the National Lung Screening Trial (NLST), which found that LDCT versus chest radiography screening was associated with a 20% reduction in mortality, but also an increase in the number of cancers detected, some of which may be clinically insignificant.

This latest analysis, by Edward Patz of the Duke University Medical Center, Durham, North Carolina, USA and colleagues, suggests that any tumour detected by LDCT has an 18.5% chance of being overdiagnosed – detected even though it would not have become clinically apparent.

The figure is based on 1089 tumours detected by LDCT among 26,722 participants and 969 detected by chest radiography among 26,730 participants, giving an excess of 120 cases. It is based on three screens and around 7 years of follow-up, and therefore represents an upper limit on the rate of overdiagnosis, say the researchers. Longer follow-up should result in a lower rate, as some tumours not initially detected by chest radiography later become clinically apparent.

However, the team notes that chest radiography itself has been associated with overdiagnosis.

The original NLST findings showed that it would be necessary to screen 320 patients with LDCT to prevent one lung cancer death. This current analysis indicates that, across the NLST follow-up period, 1.38 of the tumours detected among 320 patients would be overdiagnosed.

For specific subtypes, the likelihood of a tumour being an overdiagnosis was 22.5% for non-small-cell lung cancer (NSCLC), and was particularly high for bronchioloalveolar cell carcinoma (BAC), at 78.9%.

"These data raise the question as to the necessity and type of therapy required if a diagnosis of minimally invasive adenocarcinoma is established and challenge the diagnostic community to develop a classification scheme that could accurately Phenotype all lung tumors," the researchers write in JAMA Internal Medicine.

The high risk of BAC overdiagnosis with LDCT persisted even with extended follow-up. In a convolution model incorporating an average sojourn time of 32.1 years, the rate of overdiagnosis was 71% with 7 years of follow-up and 41% with lifetime follow-up. For NSCLC, with a sojourn time of 3.6 years, the corresponding rates were 19% and 9%.

The researchers stress that these estimates of true overdiagnosis derived from the model "must be treated cautiously".

They conclude: "[T]he limitations of the screening process, including the magnitude of overdiagnosis, should be considered when guidelines for mass screening programs are constructed."


Patz E, Pinsky P, Gatsonis C, et al. Overdiagnosis in Low-Dose Computed Tomography Screening for Lung Cancer. JAMA Intern Med; Advance online publication, December 09, 2013. doi:10.1001/jamainternmed.2013.12738

medwireNews (www.medwireNews.com) is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2013