EMCTO News: Evaluation of performance of lung cancer screening strategy in the first three rounds of the NELSON trial

Volumetric computer tomography screening for lung cancer

The Dutch-Belgian Randomized Lung Cancer Screening Trial (Dutch acronym: NELSON study) was designed to investigate whether screening for lung cancer by low-dose multidetector computed tomography (CT) in high-risk subjects will lead to a decrease in 10-year lung cancer mortality of at least 25% compared with a control group without screening. The study started in 2003 and Dr Nanda Horeweg of the Department of Public Health and Pulmonology at Erasmus Medical Centre in Rotterdam, Netherlands reported at the European Multidisciplinary Conference in Thoracic Oncology (EMCTO), held 9-11 May, 2013 in Lugano, Switzerland, findings from the evaluation of performance of the screening strategy in the first three rounds of the study and risk calculations made for a follow-up period of 5.5 years.

Lung cancer is most often detected in patients who present with already advanced disease, resulting in a poor chance for successful treatment. Therefore it is necessary to explain the risk of developing lung cancer to smokers, those family members who have been regularly exposed to second hand smoke and other high risk populations and to counsel on the potential harms and benefits of the screening procedure and how it could affect lung cancer detection at earlier stage. Low-dose CT scanning has been proposed as one sensitive screening modality; proponents estimate that it could detect approximately three times as many small lung nodules as chest X-ray. The NELSON screening strategy considers lung nodules with a volume >500mm³ or a volume-doubling time <400days positive; volumes 50-500mm³ or volume-doubling times 400-600 days indeterminate; and all other nodules as negative.

Dr Horeweg reported that one or more positive screen results were seen in 6% of the subjects and 200 participants were subsequently diagnosed with lung cancer, yielding a positive predictive value of 40.6% for the scans. False positive results were seen in just 1.2% of all scans. Over the 5.5 year evaluation it was found that the risk of lung cancer detected by screening was influenced by the results of the first scan; the risk was 1.0% after a negative primary scan at baseline, 5.7% after an indeterminate baseline and rose to 48.3% in participants with a positive baseline scan.

The authors concluded that the positive predictive value and low false positive rate arising from these results of an analysis of 5.5 years of the program support the use of low-dose CT for lung cancer screening and provide an additional tool for counselling potential candidates for screening.

Is low-dose CT lung cancer screening ready for large-scale population-based implementation?

The current guideline of the European Society for Medical Oncology (ESMO) for diagnosis, treatment and follow-up of early-stage and locally-advanced (non-metastatic) non-small cell lung cancer does not recommend it for screening, but the guideline text summarizes the evidence available until 2010. The ESMO Guidelines Working Group is currently evaluating newer data for the guideline’s update. However, it seems that in 2013 low-dose CT screening is not ready yet for large-scale population-based implementation, because of remaining questions on definition of the at-risk population, timing, interval and method of CT (especially 2D versus 3D nodule interpretation), how to handle false positive findings, and especially cost-effectiveness in relation to other prevention strategies, mainly smoking cessation. Further analyses of several ongoing European trials are eagerly awaited. Other potential methods of screening, such as sputum, exhaled breath or blood biomarkers lack validation and are not recommended in clinical practice, as well.

The study discussant, Dr Giulia Veronesi said that diagnostic algorithms aim to strike a balance between a too invasive work-up that exposes screened persons to useless invasive procedures and overtreatment, and insufficient surveillance that increases the risks of delayed diagnosis and false negatives. Many screening programmes have adopted 5 mm as the cut-off between positive and negative nodules on CT. The Fleischner's guidelines suggested 4 mm as threshold below which no follow-up was needed. This cut-off was used by the NLST trial and resulted in a very high rate of baseline positivity. Other investigators consider nodule volumes instead of nodule diameter to obtain more reliable assessments of nodule growth.

Dr Veronesi said that NELSON is the largest randomized controlled trial with LDCT screening in Europe and the first study to incorporate software calculated volume doubling time of nodules into a management algorithm to distinguish between positive and negative. According to Dr Veronesi remaining questions for the authors to address are the recall rate after baseline and after consecutive screening rounds. When we compare the NELSON recall rate with that of NLST study, the NELSON one is lower even when considering positive and indeterminate nodules together. She asked if the higher rate of recalls in the NLST was due only to the lower size cut off (4 mm instead of 50 mm3 equal to 5 mm) or there may be other causes. Her feeling is that when the multidisciplinary staff decides how to consider nodules, the rate of recalls is reduced.

In the Cosmos 1 study a total of 16 cases were missed at one CT scan and picked up a year later. Most of these were central lesions for which the LDCT scan has a recognised limitation; others had a very high growth rate (usually SCLC) or were below the threshold size in the previous year. Some delayed diagnoses were due to human failure as they were not recognized by the radiologists. Ten percent of diagnoses were reported as delayed by NY-ELCAP, a proportion closely similar to that of COSMOS. Dr Veronesi asked which is the rate in the NELSON study? T

he next question Dr Veronesi had for the authors of the NELSON study is about the standard treatment of positive cases. Did they use CT/PET or routine FNAB in every positive case? The next point she stressed is the treatment of slow growing nodules. She asked how do the NELSON investigators manage slow growing nodules suspicious for malignancy, do they consider wait and see strategy or surgical resection. And in case of surgery, would a limited resection can be an adequate treatment instead of lobectomy? It is expected that the results from the current and other ongoing screening studies answer to these questions.

EMCTO is organised in partnership between the European Society for Medical Oncology (ESMO), The European Society for Radiotherapy and Oncology (ESTRO), the European Society of Thoracic Surgeons (ESTS) and the European Respiratory Society (ERS), and the European Thoracic Oncology Platform (ETOP). The five partners have created a programme that integrated perspectives from the different disciplines and demonstrated how the multidisciplinary team can combine knowledge for personalised treatment of the whole range of thoracic oncology tumours.

All authors have declared no conflicts of interest.