Elsevier

Lung Cancer

Volume 54, Issue 2, November 2006, Pages 177-184
Lung Cancer

Nodule management protocol of the NELSON randomised lung cancer screening trial

https://doi.org/10.1016/j.lungcan.2006.08.006Get rights and content

Summary

In December 2003, the Dutch–Belgian NELSON trial, a Dutch acronym for “Nederlands-Leuvens Longkanker Screenings ONderzoek”, has been launched. Primary objective of the NELSON trial is to investigate whether screening for lung cancer by 16-detector multi-slice CT with 16 mm × 0.75 mm collimation and 15 mm table feed per rotation (pitch = 1.5) in year 1, 2 and 4 will lead to a decrease in lung cancer mortality in high risk subjects of at least 25% compared to a control group which receives no screening. In this paper, the screening regimen and the classification and management of the screen-detected nodules at baseline and incidence screening is presented. This is the first large lung cancer screening trial in which the nodule management protocol is based on volumetric nodule assessment and the presence or absence of growth. Furthermore, the quality assurance measures and the NELSON management system (NMS) are presented.

Introduction

Lung cancer is currently a serious public health problem. In Europe alone, an estimated 375,000 people die from lung cancer every year and worldwide 1.4 million per year [1]. At the time of diagnosis, over 75% of persons with lung cancer have loco-regional spread or distant metastases, substantially reducing the chances of survival [2]. Theoretically, primary prevention, quitting smoking or more importantly, measures to reduce starting smoking may totally eliminate the disease, but although several such measures have been successful, the number of lung cancer deaths each year is still unacceptably high. One of the most promising recent preventive measures is early detection using multi-detector low-dose computed tomography (MDCT) screening. Cohort studies have shown that lung cancer can be detected in a much earlier disease stage, but it is yet unknown if earlier detection will eventually reduce lung cancer mortality [2]. To address this question, in the US the National Lung Screening trial (NLST) has been launched in 2002. It is a very large multi-center trial with 53,476 participants in 46 institutes across the US, comparing CT screening with Chest X-ray screening in the control arm [3]. In Europe, the only large randomised trial is the Dutch–Belgian NELSON trial with 15,523 males and to a lesser extent females participants in four institutes, which have been launched in December 2003 [4]. Primary objective of the NELSON trial is to investigate whether screening for lung cancer by MDCTs in year 1, 2 and 4 will lead to a decrease in lung cancer mortality in high risk subjects of at least 25% compared to a control group which receives no screening, and to estimate the cost-effectiveness of this screening programme. In collaboration with a single institute in Copenhagen Denmark where 4104 participants have been enrolled in a randomised MDCT screening trial with almost the same design as NELSON, the target of 20,000 participants has almost been reached.

Screening is not merely a radiological technique, but instead a complex process of identification and selection of the target population, call and recall of participants, work-up and evaluation of positive screenees and adequate communication of the test results to the participants and all involved health care professionals. The recruitment procedure, the selection criteria and the power calculation used in the NELSON trial have been described elsewhere [4]. This paper deals with the screening algorithm and the classification and management of screen-detected nodules at baseline and incidence screening.

The management of persons with pulmonary nodules detected in a screening context differs markedly from usual clinical practice. Screening deals with asymptomatic ‘healthy’ individuals, approached by a letter of invitation and health care professionals participating in a screening programme carry thus extra responsibility for the information and safety of the individuals included in such a programme. Therefore, in this paper, also attention will be paid to quality assurance aspects and the role of the NELSON management system (NMS) in it. Given the fact that more and more advanced multi-detector CT scanner with smaller collimations are being used, also outside screening programmes, clinicians are more and more faced with the problem of small non-calcified pulmonary nodules. Prevalence rates up to 50% have been reported [5]. New software tools to assess volumes and volume doubling times become rapidly and widely available. Therefore, this management protocol could also be useful for the non-screening setting and provide new tools on how to deal with pulmonary nodules by using volumetric software.

Section snippets

NELSON management system (NMS)

To conduct this logistically complex multi-center study, the NELSON management system (NMS) has been developed. It is a web-based interactive database application used for data collection and management of all study related processes such as the selection and randomisation of participants, electronic storage of questionnaires and informed consent forms, completely trackable data collection, study monitoring, reporting of scan results and scheduling of appointments for follow-up scans. Because

Screens

The participants randomised to the screen arm were invited by an invitation letter to one of the four screening sites (University Hospital Groningen, University Hospital Utrecht and Kennemer Gasthuis Haarlem in the Netherlands, and University Hospital Gasthuisberg Leuven in Belgium. The CT scans used were all 16 detector MSCT scanners (M×8000 IDT or Brilliance 16P, Philips Medical Systems, Cleveland, OH, USA, or Sensation-16, Siemens Medical Solutions, Forchheim, Germany). All scans were

Image reading

Images were read on Siemens workstations using the Syngo Lungcare software package (Version Somaris/5 VB 10A-W) for multi-dimensional image processing and computer viewing. Lung windows were assessed at a width of 1500 and a level of −650 Hounsfield Units. After a first reading, the data were stored locally on the PACS system, and sent overnight via a protected Internet connection to Groningen for second reading and central storage in the radiological database. The second readers were unaware

Baseline screen protocol

NCNs were classified in four nodule categories (NODCAT) based on size, either 3-D (solid and partial solid lesions) or 2-D (solid pleural lesions and non-solid lesions) or based on growth (GROWCAT) according to formula (1) (Table 1). NODCAT 1 was defined as benign, NODCAT 2 as non-significantly small, NODCAT 3 as indeterminate and NODCAT 4 as potentially malignant. Based on the highest nodule category found, participants with NODCATs 1 and 2 received a negative test result, and were invited for

Incidence screen protocol

At annual repeat screening (incidence screening), there are two possibilities: either an NCN is existing, and comparison with baseline screening is possible, or the NCN is new. For the new nodules, the same classification according to size was made as for the baseline screening round. Follow-up was different, however, because at incidence screen new nodules are supposed to have a relatively higher growth rate (Table 4).

For all existing nodules, except for NODCAT 1, always a comparison with the

Management of NODCAT 4 and GROWCAT C nodules

Before describing the work-up and staging procedures for the different nodule categories, it is important to realise that especially in a screening setting unnecessary surgery for benign nodules should be avoided as much as possible. This imposes special problems for the diagnostic strategy. In general, non-invasive diagnostic procedures should be applied before invasive ones if possible, so that the latter can be reserved for lesions with a high probability of malignancy and resources can be

Baseline: NODCAT 4

If the highest category was a NODCAT 4, the participant was referred to the chest physician of choice via the general practitioner, usually the chest physician associated with the screening center. Primary objective was to confirm the presence of malignancy by performing routine physical examinations, routine laboratory tests and a bronchoscopy (bronchial washing for cytology and culture, and transbronchial biopsy or brushing on indication). A percutaneous CT-guided fine needle aspirate (FNA)

Baseline or incidfence: GROWCAT C

The work-up for participants with growing lesions (GROWCAT C) was essentially the same as for NODCAT 4, except that for these nodules a final histological diagnosis had to be obtained either by FNA, video-assisted thoracoscopic surgery (VATS), or wedge resection and examination on frozen section, and that further observation by follow-up CT scans was no longer allowed. If malignant, the nodule had to be surgically removed after appropriate staging. If the outcome of the investigation(s) was

Staging

Staging included a standard CT with intravenous contrast of the chest and upper abdomen including the liver and adrenal glands. A bone scintigraphy and MRI brain were only made on clinical indication. If the nodule was a NODCAT 4 or GROWCAT C larger than 500 mm3, a mediastinoscopy was only performed if the PET scan showed positive mediastinal lymph nodes, if there were enlarged lymph nodes on CT (short axis > 1 cm), and in the presence of a peripheral adenocarcinoma or a centrally located tumor.

Surgical resection

The treatment of small malignant lesions (T1) found at screening is according to standard practice [9], i.e. if possible at least a lobectomy should be performed due to a high frequency of local recurrence after more limited resections. Only in patients with poor pulmonary function who are judged by the surgeon not to tolerate a lobectomy, a segmentectomy or wedge resection could be performed. This may in some cases be performed as a minimal invasive VATS procedure. Because the small ground

Quality assurance

In order to promote the expertise of the investigators and to ensure the lung cancer screening trial's compliance with the quality demands of the National Health Council, several measures were taken. All radiological images are interpreted as well locally as centrally in Groningen for second reading, with the intention to promote the quality and to optimise the sensitivity of the screening. To this end, two full-time dedicated radiologists were appointed in Groningen, and a third one joined

Discussion

With the advent of high resolution CT screening, physicians are faced now with the very early stages of lung cancer among large numbers of insignificant, benign nodules. What the optimal management protocol is to discriminate between malignant and benign lesions is yet unknown. Several differences exist between the various nodule management protocols used world-wide and also the definitions used vary or are undefined, as for example, what should be regarded as an indeterminate nodule and what

Acknowledgements

We would like to acknowledge Prof. Dr. Claudia Henschke, Prof. Dr. David Yankelevitz and Prof. Anthony Reeves, Department of Radiology, Cornell Medical Center, New York, for sharing their expertise and knowledge in the development of our NELSON radiology protocol and for their advice in developing our management system. We also acknowledge Ton de Jongh, Artex VOF, Capelle a/d IJssel, the Netherlands, for developing and maintaining the NELSON management system (NMS).

References (20)

There are more references available in the full text version of this article.

Cited by (0)

View full text