Article Text
PDF
Abstract
Background In a lung cancer survey in 2000 we showed significantly less favourable stage distribution and lower resection rate in Teesside (UK) than in the comparable industrialised area of Varese (Italy). Lung cancer services in Teesside were subsequently reorganised according to National Cancer Plan recommendations.
Methods For all new lung cancer cases diagnosed in Teesside (n=324) and Varese (n=260) during the 12 months October 2010 to September 2011 (hereafter ‘the 2010 cohort’), demographic, clinico-pathological and disease management data were prospectively recorded using the same database and protocol as the 2000 survey. Findings were analysed focusing on resection rate.
Results In the 2010 cohort compared with 2000, both in Teesside and Varese emergency referral decreased (p<0.001), performance status improved (p<0.001), but cancer stage shift was not seen; resection rate improved in Teesside, from 7% to 11% (p=0.054), and was unchanged in Varese (24%). Moreover, in Teesside compared with Varese the stage distribution remained less favourable, stage I–II non-small cell lung cancer (NSCLC) proportion being respectively 12% and 19% (p=0.040), and resection rate in all lung cancers remained lower (11% and 24%; p<0.001). On multivariate analysis, resection predictors in Teesside were as follows: stage I–II NSCLC (OR 86.14; 95% CI 31.80 to 233.37), performance status 0–1 (OR 5.02; 95% CI 1.48 to 17.07), belonging to 2010 cohort (OR 2.85; 95% CI 1.06 to 7.64).
Conclusions In Teesside the main independent predictor of resection was disease stage; in 2010–2011 compared with 2000, lung cancer service improved but stage shift did not occur, and resection rate increased but remained significantly lower than in Varese.
- Lung Cancer
- Thoracic Surgery
- Clinical Epidemiology
Statistics from Altmetric.com
Key messages
What is the key question?
▸ A decade after the 2000 National Cancer Plan implementation, were lung cancer stage distribution and resection rate still worse in Teesside (UK) than in the comparable industrialised area of Varese (Italy)?
What is the bottom line?
▸ In Teesside, lung cancer service improved in 2010–2011 relative to 2000, but early-stage diagnosis was still infrequent. Cancer stage did not shift and resection rate remained significantly lower than in Varese.
Why read on?
▸ Findings suggest that in Teesside further efforts should focus on diagnosing lung cancer at an earlier stage, as this is the main predictor of resection.
Introduction
Observational evidence supports the concept that surgery is the main intervention that may improve long-term survival of patients with lung cancer (LC),1–4 and surgical resection rate is commonly used as one measure of a region's effectiveness in treating this disease.5 In 2013 the IQR of resection rates for all LCs (including histologically confirmed and unconfirmed) was a low 11.5–17.4% in the UK6 compared with 20–25% resection rates in other high-income European countries and the US.7–9 This disparity is related to differences in demography, tumour biology and comorbidities; suboptimal management of the disease in the UK, including delayed diagnosis and variable access to thoracic surgeons, may also play a role.7 ,10–12 Previous comparisons of international LC resection rates with those in the UK, based on cancer registries, were limited by lack of information on patient clinico-pathological characteristics and stage-specific treatment; thus the impact of these key factors in determining the generally lower resection rate in the UK remains unclear.13 ,14
Aiming to minimise problems in direct comparison of LC management, in 2000 we carried out a 12-month prospective study of LC presentation and treatment modalities in Teesside (UK) and in the comparable industrialised area of the Province of Varese (Italy), using an identical database and data collection protocol for the two populations. We found that patients with LC in Teesside had more comorbidities, were diagnosed at a later stage, had more aggressive subtypes, less frequently underwent active oncological treatment and had a markedly lower resection rate in all LCs, relative to Varese (7% vs 24%).7 After 2000 the Teesside LC service was enhanced with waiting time guidelines and dedicated thoracic surgeons’ attendance at multidisciplinary team (MDT) meetings, according to National Health Service Cancer Plan recommendations.15 In Varese the LC service organisation has remained essentially unchanged. A decade after the 2000 survey, we repeated the comparison of LC presentation and management in these two geographical areas, and in this paper we focus on the evaluation of changes in resection rate.
Methods
Data collection
The LC cases diagnosed during the 12 months from October 2010 to September 2011 in the referral hospitals of Teesside and Varese (hereafter ‘the 2010 cohort’) were prospectively collected using an identical data collection protocol as in the year 2000. The characteristics of the referral hospitals in Teesside (University Hospital of North Tees, Stockton-on-Tees, and University Hospital of Hartlepool) and in Varese (University Hospital Azienda Ospedale di Circolo di Varese and Hospital S. Antonio Abate di Gallarate) have not substantially changed since 2000.7 At both sites, during the study period all new clinical or pathological diagnoses of LC (ICD-10 C33–C34) were included; we excluded cases of mesothelioma, carcinoid, adenocystic carcinoma, and adenosquamous carcinoma of the lung. Patients presenting with suspected LC in Teesside and Varese were investigated using similar protocols, as specified below. During the 2000 and 2010–2011 surveys a voluntary programme of chest radiography screening for LC at the population level was active in the Varese area, covering about 10% of high-risk smokers.7
Collected data were analysed and compared with those obtained in the year 2000 survey. Research Ethics Committee approval was obtained in Varese; in Teesside the study was approved by the Trust's Research and Development Department as part of a service improvement programme.
Diagnosis and management
Conventional diagnostic and staging procedures were performed as indicated by British Thoracic Society (BTS) guidelines,16 using chest X-rays (CXRs), CT, positron emission tomography (PET), sputum cytology, fibreoptic bronchoscopy, fine needle aspiration cytology, endobronchial ultrasound-transbronchial needle aspiration and video-assisted thoracoscopic surgery. Treatment was carried out according to BTS recommendations,16 namely, surgical resection for eligible patients with early stage (stages I and II) non-small cell LC (NSCLC) and for a selected group with low-risk NSCLC stage IIIA disease; radical radiotherapy (≥50 Gy) for patients with early stage NSCLC not suitable for surgery; chemotherapy and/or radiotherapy for patients with good performance status and inoperable NSCLC (stages III and IV); chemotherapy with/without radiotherapy for patients with small cell LC (SCLC). Palliation in advanced or inoperable cases was effected by chemotherapy, radiotherapy, pain control and nutritional support. In the two UK hospitals chemotherapy was administered, while radiotherapy and surgery were performed after referral to the Middlesbrough subregional centre, where the volume of lung resections was approximately 80 procedures/year. In Varese, chemotherapy was administered at both hospitals, while radiotherapy and surgery for all patients were carried out at the University Hospital Azienda Ospedale di Circolo di Varese (lung resections volume, approximately 70 procedures/year). In Teesside and Varese, management decisions were based on discussions at MDT meetings. Notably, after 2000 in Teesside the lung cancer service organisation was implemented with at least one thoracic surgeon attending MDT meetings and with cancer waiting time guidelines and targets introduced by the Cancer Plan in England; specifically, 31 days from diagnosis to treatment, and 62 days from urgent referral to treatment.15 In Varese these basic features of LC service organisation, with minor differences of waiting times, were already present in 2000 and remained unchanged in 2010–2011.
Data recorded
The following data were gathered from LC case records: age at diagnosis; gender; LC risk factors (smoking habit, occupational risk); comorbidities and Adult Comorbidity Evaluation-27 (ACE-27) score;17 percentage of predicted FEV1; WHO Eastern Cooperative Oncology Group performance status; source of patient referral to LC specialist; mode of presentation (symptomatic/asymptomatic); clinical diagnosis or histological subtype; clinical stage according to Mountain;18 management (radiotherapy, chemotherapy, surgical resection, no active cancer treatment). Causes of non-operability (advanced stage; comorbidity; SCLC histology) were recorded.
Definitions used in the study
For ‘date of diagnosis’, ‘occupational risk’, ‘comorbidity’, ‘source of referral’, ‘urgent/emergency admission’, ‘lung cancer specialist’, ‘active treatment’ we used the same definitions as in the year 2000 study.7 To facilitate comparison of tumour stage distribution between the 2010–2011 and 2000 survey, only the clinical stage assessed according to Mountain18 was reported. Resection rate, expressed as the proportion of cases for which an operation was performed to eradicate the cancer, was calculated against the three main denominators: all LCs, NSCLCs, stage I–II NSCLCs.
Analysis of results and statistical methods
Continuous variables were reported as mean with SD, or median with range. Categorical data were presented as numbers and percentages, and were compared by χ2 test or Fischer's exact test. Comparison between groups was made using Student's t test or the Mann–Whitney U test for continuous variables. p Values <0.05 were considered statistically significant.
Extracted database variables were tabulated using Microsoft Excel (Microsoft Corp, Redmond, Washington, USA). Factors affecting the probability of use of surgery in LC treatment were examined using logistic regression models. The multivariable model included all factors with a p value <15% from the univariate model that considered all recorded data for LC cases. We excluded histology from the multivariable model because of collinearity with NSCLC stage. Sensitivity analysis was performed by backwards selection methods, to test potentially significant predictors of resection among those available. Statistical analysis was made with MedCalc Statistical Software V.13.3.3 (MedCalc Software bvba, Ostend, Belgium).
Results
From 1 October 2010 to 30 September 2011, 324 patients in Teesside and 265 in Varese presented with LC (annual crude incidence respectively of 98 and 71 per 100 000 population). After case record review, five patients in Varese were excluded from analysis due to non-adherence to histological subtype criteria. The final totals were 324 patients with LC in Teesside and 260 in Varese. The data of cases diagnosed in 2010–2011 were compared with the 2000 survey data,7 focusing on resection rate.
Demographics and clinical presentation
Table 1 summarises the characteristics of patients with LC. The crude number of newly diagnosed cases increased in the 2010 cohort relative to 2000, by 21% in Teesside and by 7% in Varese. Also the patient mean age at diagnosis significantly increased in Teesside (p=0.018) and in Varese (p=0.003). The smoking history of patients with LC fell to 40.0 median pack-years in both geographical locations. The prevalence of patients with occupational risk significantly decreased in Teesside (p<0.001) and approached that of Varese, where it did not significantly vary. The prevalence of patients with comorbidities and the pulmonary function measured as percent of predicted FEV1 were unchanged at both sites. Recording of FEV1 remained poor (65% in Teesside, 54% in Varese). The ACE-27 score distribution was similar in the 2010 cohorts (p=0.184) and patient performance status at diagnosis improved in Teesside (p<0.001) and in Varese (p<0.001), remaining significantly better in Varese. In Teesside the urgent/emergency referrals decreased significantly (p<0.001), and general practitioners were still the main source of referral to LC specialists; in Varese, urgent/emergency referral of patients with LC was also less frequent than in the year 2000 (p<0.001). In the 2010 cohorts the vast majority of patients with LC still were diagnosed with symptoms, significantly more frequently so in Teesside than in Varese (93% and 75%; p<0.001); analysis of symptomatic patients only still showed that, comparing Teesside to Varese, a greater proportion of patients had performance status 2–4 (46% vs 19%; p<0.001) and poorer lung function (predicted FEV1%: 66±24 vs 80±17; p<0.001).
Comparison of demographic and clinical data at presentation in the Teesside and Varese 2000 cohort7 and 2010 cohort
Histology and staging
As summarised in table 2, in the 2010 cohort the rates of histologically confirmed LC in Teesside (72%) and in Varese (83%) were stable compared with a decade earlier. The frequency of histological subtypes varied: adenocarcinoma increased in Teesside (p=0.044) and in Varese (p<0.001); SCLC decreased in Teesside (p=0.021), while in Varese it did not change. The proportion of LCs that were diagnosed as stage I–II NSCLC did not significantly vary over time, thus remaining significantly lower in Teesside than in Varese (12% vs 19%; p=0.040). In both locations there was no shift towards a more favourable stage distribution (figures 1 and 2). The frequency of advanced (stage III–IV) NSCLC diagnosis was higher in the later survey than in 2000; this increase was more marked in Teesside.
Histology and staging of lung cancers diagnosed in the Teesside and Varese 2000 cohort7 and 2010 cohort
Stage and histology distribution in patients with lung cancer at diagnosis in Teesside in the 2000 cohort7 and in the 2010 cohort. NSCLC, non-small cell lung cancer.
Stage and histology distribution in patients with lung cancer at diagnosis in Varese in the 2000 cohort7 and in the 2010 cohort. NSCLC, non-small cell lung cancer.
Treatment and resection rate
In the 2010 cohort compared with 2000 the proportion of patients receiving active oncological treatment increased significantly in Teesside from 50% to 66%; it remained stable at 75% in Varese (table 3). The overall use of radiotherapy increased in Teesside, while it decreased in Varese; the proportion of cases treated with chemotherapy as the sole therapy, or in combination with other treatments, rose at both sites. Details of surgical resection rates in Teesside and in Varese are shown in table 3. The resection rates increased in Teesside, reaching 11% in all LCs (p=0.054), 18% in NSCLC (p=0.096), 67% in stage I–II NSCLC (p=0.056). In Varese the corresponding resection rates were stable over the considered decade (24% in all LCs, 33% in NSCLC, 88% in stage I–II NSCLC), and in the 2010 cohort they remained significantly higher than in Teesside. The 30-day postoperative mortality in the 2010 cohort was nil at both sites. Most resections at both locations were lobectomies. The pneumonectomy rate fell in Teesside from 41% of all resections in 2000 to 17% a decade later. Mean age of resected patients increased in both populations: from 64±9 to 68±7 years in Teesside (p=0.080) and from 65±9 to 69±8 years in Varese (p=0.022). Notably, if the patients with incidental findings were excluded from the analysis, resection rates in all LCs were not significantly different in the two cities’ 2010 cohorts (10% and 14%, p=0.205) (table 4).
Treatment modalities of lung cancers diagnosed in the Teesside and Varese 2000 cohort7 and 2010 cohort
Comparison of demographics, pathology and treatment of lung cancer in Teesside and Varese in the 2010 cohort
In both locations, over the decade no significant changes were seen in the distribution of causes of non-operability; however a trend of more frequent inoperable advanced stage NSCLC was recorded in Teesside (table 5).
Causes of non-operability in unresected cases in the Teesside and Varese 2000 cohort7 and 2010 cohort
Predictors of resection
To investigate the predictors of surgical treatment in Teesside, logistic regression analysis was performed considering age, gender, symptoms, referral, performance status, cancer stage and period of LC diagnosis. At univariate analysis the unadjusted OR of undergoing surgery significantly increased for NSCLC stage I–II (p<0.001), performance status 0–1 (p<0.001), non-emergency referral (p<0.001), asymptomatic at diagnosis (p<0.001), belonging to 2010 cohort (p=0.055) (table 6). As shown in table 6, independent predictors of LC resection were stage I–II NSCLC (OR 86.14), performance status 0–1 (OR 5.02), age+1 year (OR 0.95), belonging to 2010 cohort (OR 2.85).
Univariate and multivariate analysis by logistic regression of predictors of radical resection in 585 patients with lung cancer diagnosed in Teesside (UK) in the 2000 cohort7 and 2010 cohort
The analysis of predictors of resection in Varese showed that independent predictors of resection were (in order of importance): stage, age, performance status, asymptomatic diagnosis (table 7).
Univariate and multivariate analysis by logistic regression of predictors of radical resection in 503 patients with lung cancer diagnosed in Varese (Italy) in the 2000 cohort7 and 2010 cohort
Discussion
By repeating in 2010–2011 the comparison of LC management in Teesside and in Varese, we aimed to analyse the magnitude and possible causes of persisting disparity in LC resection rate between these two locations. Acknowledging concerns about data completeness and comparability, efforts were made to capture all new LC cases diagnosed at both sites during the study period, under the principal investigators’ supervision (AI and RNH), using the same database and protocol as in the 2000 survey.7 We found that LC crude annual incidence in the 2010 cohort relative to 2000 rose by 21% in Teesside and by 7% in Varese. Such an increase, almost exclusively due to more diagnoses of advanced stage NSCLC, has several possible explanations. These include higher quality of preoperative staging processes and stage migration, population growth (+3% in Teesside;19 +7% in Varese20) and ageing over the considered decade. The more robust LC incidence rise in the UK site likely reflects improved local LC service organisation and markedly increased case ascertainment rate in the UK over the study period, as documented by the National Lung Cancer Audit.21
Our study shows that important demographic differences persisted between patients diagnosed with LC in the two locations. In the 2010 cohort the male/female ratio in Varese was three times that in Teesside, however the ratio decreased in both cities over a decade; this decrease was more pronounced in Varese, reflecting the trend of more frequent LC diagnosis in female patients in recent years in Italy.22 Moreover, after a decade, significant differences between cities persisted in lung function, performance status, source of referral, and rates of incidental diagnosis, histological confirmation, adenocarcinoma, total NSCLC, stage I–II NSCLC, active treatment, and surgical resection (table 4).
In Teesside, the earlier referral of symptomatic patients and application of waiting time guidelines recommended by the National Cancer Plan did not lead to a more favourable LC stage distribution; the proportion of cancers diagnosed as stage I–II NSCLC did not significantly change in the 2010 cohort relative to 2000 and remained significantly lower in Teesside than in Varese (12% vs 19%). Moreover, in Teesside after a decade the rate of histologically unconfirmed LC remained stable above European standards, and significantly higher than in Varese. Our findings mirror the outcome of LC awareness and early referral campaigns conducted in 2008 and 2011 in other UK areas, which resulted in higher incidence of advanced LC diagnoses, without stage shift.23 ,24 A recent study, however, showed a 3.1% increase in proportion of NSCLC diagnosed in stage I and 2.3% increase in resections for patients seen during a national campaign in the UK to raise public awareness of persistent cough as a LC symptom.25
It should be underscored that the proportion of asymptomatic LC diagnoses following incidental imaging was nearly three times greater in Varese compared with Teesside, suggesting a more conservative use of radiologic imaging exams in general in the UK setting. Overall, the proportion of screening detected LCs in the Varese 2010 cohort was minimal, likely due to scarce adherence to the screening programme.26 Therefore, we hypothesise that more frequent early-stage disease diagnosis by incidental imaging observed in Varese may reflect a greater LC awareness at population level and among general practitioners,27 and a more liberal use of CXR and chest CT for follow up in a variety of diseases. If LC screening by low-dose chest CT were implemented in the UK, as proposed by an expert panel11 in the wake of the National Lung Screening Trial results showing 20% LC mortality reduction,28 the number of asymptomatic LC diagnoses could rise. The results of the NELSON study, the Dutch large trial of CT screening for LC, will be announced by the end of 201529 and will likely affect the decision by health policy makers in European countries to promote LC screening at the population level.
Focusing on LC resection rate in our study, we found that the use of surgery in Teesside grew to 11% in the 2010 cohort (with all registered LCs used as denominator), reflecting the UK trend in recent years.10 ,30 The resection rate in NSCLC and in stage I–II NSCLC, that have been proposed as benchmark audit indicators for LC services,21 ,31 also rose in the Teesside 2010 cohort, but remained significantly lower than in Varese. We hypothesise that this difference in resection rate may be contributed to by higher incidence of asymptomatic LC findings in Varese; notably, if patients with incidental diagnosis were excluded, resection rates in all LCs are not significantly different in the two cities’ 2010 cohorts (10% and 14%, p=0.205).
The pneumonectomy rate fell in Teesside over a 10-year period. We think that better preoperative staging, closer adherence to surgical guidelines and the presence of a dedicated thoracic surgeon willing to take on difficult lobectomies played a role in reducing the pneumonectomy rate.
At both sites the causes of non-operability did not substantially change. However, among all inoperable cases the proportion of advanced stage NSCLC increased in 2010–2011, as the frequency of these advanced cancers grew. The increased use of surgery in Teesside is likely attributable in part to overall improvement of LC service, documented by earlier case referral, better patient performance status at diagnosis and >97% of new LC cases discussed at MDT meetings in 2010 and 2011.21 ,32 Routine participation of a specialist thoracic surgeon in lung MDT meetings in Teesside was likely a major contribution to LC resection rate rise, as suggested by others.12 ,33 Whether the improved LC resection rate in Teesside will translate into increased long-term survival needs to be evaluated by prolonged follow-up. Nevertheless, in the 2010 cohort the LC resection rate using all registered LCs as the denominator remained significantly lower in Teesside (11%) than in Varese (24%). Of note, comparison of four cancer registries in northern Italy in 2003–2005 showed that in the two Italian areas where >80% of LCs were histologically verified and a thoracic surgery unit was present, the resection rate in all LCs was 23.9% and 25.3% respectively, similar to that in Varese; it was 9.2% in the area lacking a thoracic surgery facility.3
The strengths of this study are the prospective design and use of the same database and protocol for data collection in 2000 and a decade later, both in Teesside and in Varese. The limitations are the lack of information on patient factors influencing the willingness to undergo surgery, and on surgeon factors such as degree of specialisation and propensity to take on high-risk surgical cases.
Multivariate analysis showed that in Teesside the most powerful predictor of surgical treatment, among the clinico-pathological factors considered, was stage I–II NSCLC (OR 86.14). Accordingly, in the 2010 cohort the low proportion of stage I–II NSCLC at diagnosis in Teesside was probably the dominant cause of the persistently low resection rate compared with Varese. However, the lower resection rates in Teesside remained heavily influenced by poorer performance status (OR 5.02). The increase in surgery and other anticancer treatments in Teesside after a decade reflect better performance status and the service change, with increased early referrals and fewer emergency presentations.
In conclusion, the main result of this study is that in Teesside the LC service improved in the 2010 cohort relative to 2000, but disease stage, the main predictor of resection, did not shift and the use of surgery remained significantly lower than in Varese.
Based on our findings it is unlikely that in Teesside the resection rate in LC will reach the 20% rate observed in other high-income countries unless LC is diagnosed at an earlier stage. Patients with LC need to be diagnosed before they become symptomatic.
References
Footnotes
Acknowledgements We gratefully acknowledge the assistance of Mr Timothy Sampson in editing the manuscript.
Contributors AI, RNH, LD, NR planned the study, analysed and interpreted the data. NL, VJ, JB, EN, EA, MCas, MCat collected, prepared, transmitted and considered the raw data. AI and RNH supervised data collection. NR, MCas, JB quality controlled the dataset. AI, RNH, LD, NR contributed to writing the report.
Funding This study was supported in part by University of Insubria research funds (FAR 2010–2014), by research grant No. 230/2011 from University of Insubria to Dr N Rotolo, by grants from Associazione PREDICA Onlus, Varese, Italy to Dr E. Altieri and Dr M. Castiglioni, and by research grant No. 3/2014 from Associazione Comunitaria del Varesotto—Onlus, Varese, Italy.
Competing interests None declared.
Patient consent Obtained.
Ethics approval Research Ethics Committee of Varese Hospital.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Airwaves