ERS statement on harmonised standards for lung cancer registration and lung cancer services in Europe

Drivers, development and implementation of two national lung cancer audit programmes: Denmark and England

Value of a pan-European dataset

PAG members were in agreement that the implementation of a lung cancer dataset across Europe would be particularly useful in:

• developing and monitoring diagnostic standards;

• developing and monitoring standards of care in lung cancer;

• assisting evidence-based analysis of data across countries; and

• establishing what treatments work and for whom.

Patient access to data would be of interest and value to patients as a way to understand more about their condition and what could be viewed as usual or unusual in comparison with others. This would give individuals a useful comparator to discuss their condition with clinicians. In light of this, patient access to the data should be considered as part of any dataset development.

Gathering data

It was noted that patients could provide valuable input in defining the relevant importance of different quality of life (QoL) data and the considerations to be aware of when collecting these data. Patients felt it was important for QoL data to be collected verbally, directly from patients, to ensure consistency, and to identify patterns across the pathway which could then lead to the identification of relevant support where appropriate. A crucial factor for the successful gathering of information from the patient is the level of trust that exists between clinician and patient. The task force identified the QoL questionnaire from the European Organization for Research and Treatment of Cancer (EORTC QLQ-C30) as a possible resource because this is a standardised tool in current use [27]. The PAG members thought that some of the questions were relevant, but that a subset of these questions might be more effective.

PAG members suggested that there could be a beneficial role for caregivers, nurses and hospice staff in helping to gather QoL information and assisting patients, especially at times of high stress and anxiety. They identified these times as often being at the point of diagnosis; when patients may not have much energy, e.g. during chemotherapy; or when receiving palliative care. This will vary with each individual and further discussions would be beneficial to ensure data are both sensitively and effectively gathered. Patient-reported outcomes have demonstrated positive impact on treatment outcomes and their use is expected to increase in the future [28]. Specific recommendations on patient-reported outcomes are not dealt with in this recommendation.

The PAG thought QoL data collection would be most valuable at diagnosis, post primary treatment (3 months) and at the end of primary treatment (6 months). Several PAG members felt that it would be helpful to gather data after 12 months. It was also suggested that collecting QoL from patients at the end of a 5-year recurrence-free follow-up could be valuable in sharing hope among patients.

PAG members recommended that the gathering of comorbidity data should be patient led and clinician reported. These data should emerge from discussion and agreement between the clinician and patient. This would have the additional advantage of patients being better informed about potential comorbidities and provide opportunities for pre-agreement with their clinician about what to do should symptoms appear, potentially leading to lower patient anxiety in the long term.

Implementation topics of importance to patients

Implementation topics identified as important to patients included informed patient consent, data protection and data security, data use and patient knowledge of how it is used, and information about clinical trial involvement. Providing personalised data summaries with pan-European comparisons would also be a valued option.

ELF would recommend that patient representatives are fully involved in future discussions about dataset development to ensure that all patient issues have been considered and any potential challenges addressed before any future roll-out across Europe.

Basic patient features

Table 1 illustrates data items for basic patient features. A record of ethnicity is important for several reasons. There is evidence of significant variations in the prevalence of somatic mutations in adenocarcinoma of the lung based on ethnicity [29, 30]. There is also evidence of variation in the route to accessing healthcare services based on patient ethnicity [31–33]. However, it is difficult to find one coding system for ethnicity that would capture the needs of every country in Europe. The ICHOM dataset definition [12] states that individual countries should determine the definition, and therefore this data item is not suitable for cross-country comparison. The task force therefore concluded that it was not possible to propose one list of ethnicity codes that would be relevant for every country in Europe (an example of a coding system is shown in supplement 1.3). The educational level of an individual was chosen as a surrogate for socioeconomic status. Some countries have well-established linkage between registries or independent lung cancer audit programmes and census data which allow them to stratify an individual's socioeconomic status. However, these are in the minority, and although socioeconomic status is a very important indicator of access to healthcare generally, as well as key clinical outcomes in thoracic oncology [34–36], a compromise was agreed. The task force adopted a simple outline of educational level achieved based on ICHOM (primary, secondary or tertiary) [12]. There is wide variation in the level of educational status achieved in different countries, and it is not an ideal surrogate for socioeconomic status, but despite this limitation it was thought that educational level would be a data item that could be captured.

Five data items relate to the diagnosis of lung cancer, i.e. how it was made, with the inclusion of pertinent dates; these will be powerful points of reference when interpreting the lung cancer pathway and processes within each country or between countries. Delays in referral to a lung cancer specialist have been proposed as a reason for differences in outcomes, so the date of referral to a lung cancer specialist is included. The route to a lung cancer specialist varies across Europe, and it often does not involve a primary care physician [10]. There is a hierarchical basis to the date of diagnosis, which is taken from the ENCR minimum dataset for cancer registries [9]. The date of the final pathology report reflects the need to identify delays in obtaining a complete pathological diagnosis that are consequent upon increasingly complex processing. The mode of presentation is an essential data item because it is known to influence prognosis.

The basis of diagnosis (clinical/radiological or pathological) is crucial because of the association with prognosis: a more precise identification of the denominator for the whole cohort allows international comparisons to reduce selection bias. Comparisons must use the same denominator because cohorts that only include patients with pathological confirmation do not include those patients with an often worse prognosis, who are diagnosed purely on the basis of a high level of clinical suspicion; such patients are often too unwell or too frail to undergo further tests. There is evidence that the likelihood of obtaining pathological confirmation in individuals believed to have lung cancer is affected by several factors. These include age [37], socioeconomic status [38] and performance status (PS) [39]. Equally, factors relating to the lung cancer service could account for variation in pathological confirmation rates, and hence the recommendation for agreed standards among lung cancer services in Europe (see “Part 2. Manual of standards for lung cancer services in Europe”). Internationally there is no agreed pathological confirmation rate, but the NLCA of England found that higher pathological confirmation rates were most strongly associated with survival in patients with stage I/II disease who had a PS of 0–1 [40]. Thus, a stratified approach to pathological confirmation based on clinical features was suggested, rather than a single benchmark figure for pathological confirmation rate. The basis of diagnosis is the same as that defined by ICHOM [12].

Tumour details

Data that specify details of the tumour (table 2) are essential for international comparisons because of the strong influence on prognosis, type of treatment offered and prediction of treatment response. The pathological subtype is vital, and we know that different countries within Europe use different systems. The majority use the International Classification of Diseases for Oncology, 3rd edition, which incorporates all subtypes according to the current 2015 World Health Organization classification of lung tumors [41], including the new lung adenocarcinoma classification originally proposed by the International Association for the Study of Lung Cancer (IASLC), American Thoracic Society and ERS [42]. However, Denmark uses Systematized Nomenclature of Medicine (SNOMED). The task force recommends that data are entered into the system using whichever classification is standard practice within each country. Retrieval of specific pathological subtypes could then be reconstructed with automated algorithms. For those countries without a specific pathological classification system, we have created a small but clinically relevant list of pathological subtypes. The field of molecular analysis is expected to expand in the future and so the data collection system needs to be able to include new definitions as clinical practice changes. These changes could be incorporated during a revision programme every 2 years, in order to balance clinical development with practical utility.

Stage of disease at diagnosis is compliant with the IASLC staging system [43]. The basis of the stage reflects access to certain procedures as well as national guidelines for diagnosing lung cancer, so there are data fields to record which investigations have been performed prior to the formation of a “final pretreatment clinical stage”. The version of the staging system (7th or 8th edition) is selected first, and then the individual T, N and M stage is entered. Further details about tumour size and the number and location of nodes and metastases then follow. Sub-classification of the extent of N2 disease is included as part of the desirable dataset, which could then be used to categorise the patient cohort based on either the Robinson classification [44] or the IASLC staging project (assuming N1 disease is also sub-classified; table 2) [45]. This level of detail from a pan-European cohort of individuals with lung cancer will allow for a comprehensive and very detailed analysis of the prognostic value of the current IASLC staging system.

Extended patient features

Table 3 captures the extended patient features. The main data item in this section is the PS of the individual at the time of diagnosis with lung cancer. The Eastern Cooperative Oncology Group (ECOG) system [46] (also known as the World Health Organization PS) is the most widely used method for recording this feature, although there is evidence that PS is only routinely collected in less than a third of European countries [11]. It is paramount that collecting this patient feature becomes routine in all registries or audit programmes given the important role PS plays in predicting outcome [39, 47–50].

The subsequent data items would allow a detailed evaluation of the clinical outcomes from lung cancer within and between countries. The majority of European countries do not collect many of these items, and it would take significant investment and political support to achieve this. Assessing comorbidities is a fundamental part of patient evaluation prior to making a treatment plan, and there is good evidence for the influence of comorbidity on outcome [51–54]. The Charlson index was developed in the late 1970s and validated on a cohort of patients with breast cancer [55]. It has subsequently been used in numerous studies, but it has limited functionality given the complexity of the score, the lack of clarity regarding severity of comorbid disease and the out-of-date weighting given to HIV/AIDS. The Adult Comorbidity Evaluation-27 (ACE-27) score is an alternative model used by some to quantify comorbid disease [56], and some countries record specific comorbid diseases, but the list is variable [11]. Therefore, the ICHOM list of comorbid diseases (based on Sangha et al. [57]) is recommended, but should be derived from the medical notes after consultation between the clinician and the patient. It is hoped this will ensure accurate recording of all known comorbidities. The EORTC QLQ-C30 patient-completed questionnaire is recommended for QoL [27]. This is based on the fact it is a validated research tool [58], although our patient group felt only a subset of the questions were relevant. This questionnaire should be completed at diagnosis, after completion of first-line treatment and at 6 months post-diagnosis. This may be difficult to achieve but QoL for patients is a fundamental outcome measure, often neglected. Members of our PAG felt that ideally we would also collect QoL data at 12 months and after 5-year survival, where applicable. A revised QoL questionnaire is in development, which incorporates elements of the QLQ-C30 with specific reference to side effects from medical treatments including chemotherapy and targeted therapies [59].

Lung cancer pathway/outcomes

The final section of the European recommended minimum dataset relates to aspects of the lung cancer pathway, specifically the outcomes in terms of treatment and survival (table 4). Patients in some countries have identified how important contact with a LCNS is because they provide significant support to patients and their families throughout the lung cancer pathway. Although there is no accepted international definition for a LCNS, the task force suggests the following: a LCNS is one whose primary role is to meet individuals with lung cancer at diagnosis, sometimes before, and then to provide support to the patient and their family in terms of education, access to benefits, liaising with primary care physicians and emotional support. The role may include other duties, such as administering chemotherapy, although this does not on its own meet the essential elements of holistic care described above.

Treatment data items are shown in table 4 and provide a comprehensive list of treatment options and associated secondary questions, which would not apply to all cases. In order for meaningful analysis of lung cancer outcomes to take place, and the influence of treatment modalities on survival to be assessed, every effort must be made to capture all relevant information.

UK

In 1995, the report by Calman and Hine, “A Policy Framework for Commissioning Cancer Services” [65], set the basic standards for cancer services in England and Wales including MDTs working as a core element of cancer services. Since then, the NHS has further developed and regularly updated standards of cancer centre-based care in the UK, and standards have been monitored through a national peer review process and the NLCA [19, 20, 22]. Furthermore, a national optimal clinical pathway for suspected and confirmed lung cancer: from referral to treatment” has been published [66]. In 2014, Cancer Research UK named the “Lung Cancer Centres of Excellence”, jointly based in London and Manchester, whose aim is to develop and promote high-level lung cancer research [67].

Denmark

In Denmark, as mentioned above, the DLCG, through the DLCR and the national lung cancer guideline programme and in collaboration with national healthcare authorities, catalysed a process of continuous improvement of lung cancer care which has, among other things, instigated a re-organisation with centralisation of Danish lung cancer services [17]. Supplement 1.4 depicts the lung cancer service in the region of Southern Denmark.

France

France has a national task force against cancer that has developed three national “cancer plans” [68]. The first cancer plan, launched by President Chirac (2003–2007), set the basis of a national strategy for multidisciplinary management of cancer. It legalised the compulsory multidisciplinary discussion of each individual cancer patient. MDTs are organised according to organ or system; within thoracic oncology, pleural mesothelioma and thymic epithelial tumours fall within the remit of rare tumour boards (national), rather than the lung MDT. The MDT discussion must lead to a consensual personalised treatment plan, which is a written document given to the patient during a structured consultation, and a nurse coordinator is also present to offer psychological or social support if required. The plan is also sent to the patient's general practitioner and all corresponding doctors.

The first cancer plan also elaborated on the accreditation of units caring for patients with cancer, and in particular of surgical units. A surgical unit should host at least two surgeons, have access to an intensive care unit and an endoscopy suite, and have a frozen section analysis available on site. Minimum thresholds have been set per organ, which result in a minimal caseload of 20 major resections per surgeon (respectively 30 cases per unit, given some surgeons work on more than just cancer).

The Ministry of Health created a national institute of cancer (Institut National du Cancer (INCa)) in 2005, which coordinates research and treatment in oncology. INCa publishes an annual report and collaborates with 25 regional oncology networks, which coordinate screening and treatment at the regional level. INCa is also connected to the Higher Authority of Health (Haute Autorité de la Santé, HAS), which is in charge of editing guidelines and quality control. Finally, INCa has accredited and coordinates eight inter-groups for clinical research, including the French Intergroup of Thoracic Oncology (IFCT).

Two subsequent cancer plans have been launched by presidents Sarkozy (2009–2013) and Hollande (2014–2019). The third cancer plan is an ambitious document [69] that not only aims to improve treatment, but also to act before diagnosis (prevention, screening) and after treatment (follow-up, social re-integration).

Germany

In 2008, the German Cancer Society in collaboration with the German Respiratory Society and the German Society of Thoracic Surgery initiated a certification programme for lung cancer centres as part of the German National Cancer Plan. In September 2016 there were 44 certified lung cancer centres in Germany and two in Switzerland. The certification process comprises:

• an annual updated parameter manual with mandatory and recommended elements of structure, process and outcome data which are used for self-assessment and subsequent external validation;

• annual visits to the respective lung cancer service by trained external specialists;

• an extensive evaluation of the results by an independent institute; and

• a final evaluation [70, 71].

The German parameter manual contains 10 chapters covering mainly medical aspects of the lung cancer pathway. Multi-professional/disciplinary working is encouraged and there are specific mandatory standards for centres. These include diagnosing and treating ≥200 new patients with pathologically confirmed lung cancer per year and ≥75 anatomical lung cancer resections per year, and recording performance indicators such as a ≤5% 30-day mortality after anatomical lung cancer resections, and ≥10% proportion of broncho-/angioplastic resections on all anatomical resections. Clinical lung cancer registration and follow-up data collection are mandatory in every certified lung cancer centre and their close linkage to the newly established clinical cancer registries of the 16 German Federal States is promoted [70, 71]. The process has seen improvements in multidisciplinary working.

Only 33% of all new cases of lung cancer in 2016 were covered by certified lung cancer centres. The main obstacles for broader implementation are the mandatory thresholds for new cases and surgical resections [70, 71]. Other medical societies in Germany have established independent certification programmes related to lung cancer care (i.e. the “Oncological Centres” of the German Society for Haematology and Medical Oncology and the “Thoracic Centres” of the German Society of Thoracic Surgery, the latter initiative appraising both benign and malignant disease) [72, 73].

Organisation of the lung cancer service

The manual of standards for a lung cancer service in Europe is provided as supplement 3. This first section addresses the relevant organisational aspects of the lung cancer service as a whole. Although a multidisciplinary network environment is an essential requirement, it was agreed that the specific membership should be determined according to the local and/or national setting. Advanced standards have been formulated to encourage lung cancer services that have a full range of diagnostic and/or treatment facilities to offer these to partner organisations. It is important to note that there is no one infrastructure that every service should adopt. Every aspect of the lung cancer pathway should be available to the individual patient, but the delivery of this may vary at the local level. A real-life example of a multi-site lung cancer service from Denmark is included in supplement 1.4.

The PAG also elaborated on the standards for patient- and carer-centred care. The evidence base for this is limited, although not strictly necessary when a patient expert group has commented and when there are several national lung cancer guidelines recommending shared decision-making on the basis of information that is easy to access and understand [75].

The task force identified further “essential standards” in a lung cancer service which relate to:

• adherence to evidence-based care, with use of regularly updated guidelines;

• access to specialised care;

• timeliness of care;

• documentation, accessibility and provision of patient and carer-related information;

• communication and environment for communication;

• education for healthcare professionals, patients and carers (e.g. the Thoracic Oncology HERMES syllabus and curriculum [76–80]);

• clinical cancer registration;

• quality assurance, quality improvement and risk management; and

• collaboration with external healthcare professionals and other external stakeholders by the lung cancer service.

The utilisation of the proposed pan-European dataset for lung cancer registration is recommended as an advanced standard. Advanced collaborative measures have been proposed by the task force group to facilitate local, regional, national and international networking.

Lung cancer pathway

The second section of the manual encompasses the entire lung cancer pathway within the lung cancer service, from diagnosis through treatment, follow-up, relapse and end of life or survivorship (supplement 3). The underlying international and national guidelines which provide recommendations related to most of the essential and advanced standards within this section are listed in supplement 1.2.

Cross-pathway care is included in this section. This is often important to ensure that the patient's experience is maximised when care is needed from services outside the lung cancer pathway; these may include emergency care, intensive care and services for specific symptom management. Palliative care is included here but it is noted that this should be provided throughout the entire pathway (figure 1) [81].

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Novel integrated lung cancer care concept with diagnostics, systemic therapy, radiotherapy, surgery, palliative care and follow-up as equitable pillars of lung cancer care. Adapted from [81].

Pre-existing statement papers and recommendations issued by other international medical societies have been reviewed and incorporated into the manual of standards where appropriate. These include imaging [82], fitness for diagnostics and radical therapy [83], thoracic surgery [84], radiotherapy [85–93] and palliative care [94]. Owing to limited evidence and heterogeneity among and within European countries, the task force group was unable to define standards for individual or institutional volumes of care and timeliness/waiting times.