Second Primary Malignancies in Patients with a Neuroendocrine Neoplasm in England Open Access

Neuroendocrine neoplasms (NENs) are a heterogeneous group of malignant tumours that originate from neuroendocrine cells. These tumours can occur throughout the body with the most commonly occurring NENs originating in the gastrointestinal system (e.g., stomach, small and large bowel, pancreas, and rectum) and the lungs [1, 2]. NENs may be further classified as well differentiated neuroendocrine tumours (NETs) or poorly differentiated carcinomas (NECs) which include small cell and large cell NECs [3]. The incidence of NENs has been increasing significantly during the last 40 years at varying rates across the world [2, 4, 8]. A recent cohort study reported over 15,000 cases of NEN in the UK between 2013 and 2015 with an age-standardized incidence rate of 8.1 per 100,000 in females and 9.1 in males (2013–2015 combined) [9].

Patients are often diagnosed at an advanced stage of disease due to the wide variety of signs and symptoms experienced by patients. Furthermore, patients with NENs frequently go on to develop a second primary malignancy (SPM) [10]. Exact reasons for this are unknown but may include genetic susceptibility, shared risk factors, and increased surveillance [11]. Whilst studies exist which have investigated these SPMs, many are only case studies or single-centre studies and only look at crude incidence. There are, however, a few studies such as those by Tichansky et al. (USA) [12] and Tsai et al. (Taiwan) [13] which have looked at comparative data to calculate standarized incidence ratios (SIR). The study by Tichansky et al. [12] was a national study which utilized data from the National Cancer Institute Surveillance, Epidemiology, and End Result (SEER) database. This study, however, only included colorectal NETs and concluded that these patients had an increased rate of cancer of the colon, rectum, small bowel, oesophagus, stomach, lung, urinary tract, and prostate, when compared to a background population. In their nationwide study, Tsai et al. [13] also concluded that the risk of a second cancer after NET is increased, though they did not observe a predominance of second cancer type according to primary site of NET.

We therefore sought to investigate the incidence of SPMs following NEN occurring at different primary sites in a nationwide European cohort of patients which to the best of our knowledge has not been published before. Therefore, the aim of this study was to calculate the SIR of SPMs which occur after diagnosis of NEN, when compared to the background incidence of such tumours in England. By identifying which patients are at a higher risk of these tumours, we hope to inform on the screening of these patients to ensure early and efficient diagnosis is achieved, and to better inform patients themselves about their condition.

These data incorporate over 99% of all cancers diagnosed in England and are obtained from several sources including patient medical records, histopathology, and haematology services as well as screening services, general practitioner practices, and other UK cancer registries. This project involves data derived from patient-level information collected by the NHS, as part of the care and support of cancer patients. The data are collated, maintained, and quality assured by the National Cancer Registration and Analysis Service, which is part of NHS Digital. Access to the data was facilitated by the Office for Data Release [14].

Data were extracted on all patients aged 16 and over who were diagnosed with a NEN between 2012 and 2018. The WHO International Classification of Disease Edition-10 (ICD-10) was used to identify NENs occurring at all anatomical sites (C00 – C80). Morphology codes included 8013 (excluding lung [C34 and C78]), 8041–8045 (excluding lung), 8150–8158, 8240–8247, 8249, and 9091 according to the WHO International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) (online suppl. Table 1; for all online suppl. material, see www.karger.com/doi/10.1159/000530238).

As goblet cell carcinomas ICD-O-3: 8243 have been reclassified as non-NEN by the WHO classification of tumours [15], these were excluded from analysis [15]. Patients were classified into one of eight NEN site groups including: appendix, caecum, colon, lung, pancreas, rectum, small intestine, and stomach. Other sites, which were non-descript and sites likely to be metastasis such as liver were excluded. We estimate the proportion of secondary NENs excluded to be around 3%.

Data on patients’ demographics such as age, sex, ethnicity, and indices of multiple deprivations were available. Clinical information such as date of diagnosis as well diagnosis of a SPM was extracted. ICD-10 codes were used to identify patients who had been diagnosed with second primary malignant neoplasms of the lung (C34), oesophagus (C15), stomach (C16), small intestine (C17), colon (C18), rectum (C19-20), pancreas (C25), prostate (C61), bladder (C67), renal (C64-65), breast (C50), ovary (C56), uterus/corpus uteri (C54-55), cervix (C53), and thyroid (C73).

Data on the background incidence of malignant neoplasms were extracted from the Office for National Statistics (OND) website [16]. Cancer incidence was extracted by year (2012–2018), age, and sex at diagnosis.

Descriptive tables were produced to characterize the study cohort. Continuous variables were given as mean with standard deviation. The frequency and proportions of second cancers which occurred before, on the same date and after a NEN diagnosis date were summarized. Standardized incidence ratios (SIRs) were then produced for the total NEN population for each SPM (lung, oesophagus, stomach, small intestine, colon, rectal, pancreatic, prostate, bladder, renal, breast, ovarian, uterine or corpus uteri, cervical, and thyroid) which occurred after NEN diagnosis (metachronous tumours). We chose to include these patients only, as incidental findings in the investigation of the second cancer occurring simultaneously with the NEN diagnosis may be due to other reasons and represent a different patient cohort. The SIRs were produced by comparing the observed incidence of each metachronous tumour type to the expected incidence according to the data from the background population in England. SIRs were then calculated based on NEN type (appendix, caecum, colon, lung, pancreas, rectum, small intestine, and stomach).

A total of 20,579 patients diagnosed with a NEN between 2012 and 2018 were identified for inclusion in the study cohort (Table 1). There was a mean age of 61 (standard deviation: 16.70) with a relatively even split between males (48%) and females (52%). There was a gradual increase in incidence of NENs over the 7 years with 2,455 diagnosed in 2012 and 3,316 in 2018.

Overall, 3,127 patients were diagnosed with a non-NEN cancer at any time point (15%), of which 19% were diagnosed after NEN, 70% were diagnosed prior NEN and 11% on the same date. The frequencies and proportions of each second cancer are outlined in Figure 1 and online supplementary Table 2. When looking at non-NEN cancers which occurred prior to NEN diagnosis, the most common were the breast (22%), prostate (22%), and colon (21%). The lowest proportions were seen in the cervical (1%), pancreatic (1%), and small intestine (<1%) malignancies. When looking at the non-NEN cancers which were diagnosed after the patient’s NEN diagnosis, the most commonly occurring cancers were the prostate (20%), lung (20%), and breast (15%). The ovarian (2%), small intestine (<1%), and cervical (<1%) malignancies had the lowest proportions diagnosed after NEN.

When including all NEN types (Fig. 1; online suppl. Table 3), statistically significant SIRs were observed for non-NEN cancers of the lung (SIR = 1.85, 95% CI: 1.55–2.22), colon (SIR = 1.78, 95% CI: 1.40–2.27), prostate (SIR = 1.56, 95% CI: 1.31–1.86), kidney (SIR = 3.53, 95% CI: 2.72–4.59), and thyroid (SIR = 6.31, 95% CI: 4.26–9.33). When stratified by sex, statistically significant SIRs remained for the lung, renal, colon, and thyroid tumours (Fig. 2). Only females had a statistically significant SIR for stomach cancer (SIR = 2.65, 95% CI: 1.26–5.57) and bladder cancer (SIR = 2.61, 95% CI: 1.36–5.02) (Fig. 3).

Figure 4 and online supplementary Table 4 display the SIRs for each non-NEN cancers occurring after NEN diagnosis which were found to be significant in the overall analysis, when stratified by NEN site. There was no clear pattern of non-NEN tumour incidence by NEN site with all eight NEN sites being associated with an increased incidence of at least one non-NEN tumour site. Incidence of lung cancer was increased in several NEN sites including the appendix (SIR = 2.04, 95% CI: 1.13–3.68), caecum (SIR = 2.71, 95% CI: 1.22–6.03), lung (SIR = 1.95, 95% CI: 1.41–2.70), pancreas (SIR = 2.59, 95% CI: 1.72–3.89), and stomach (SIR = 2.27, 95% CI: 1.26–4.10). In females, stomach cancer incidence was increased significantly in NEN of the pancreas (SIR = 6.69, 95% CI: 1.67–26.73), rectum (SIR = 16.96, 95% CI: 4.02–64.22), and stomach (SIR = 11.74, 95% CI: 2.94–46.92). Significant SIRs were reported for colon cancer in patients with NEN of the appendix (SIR = 3.42, 95% CI: 1.89–6.17), caecum (SIR = 4.03, 95% CI: 1.68–9.69), colon (SIR = 4.82, 95% CI: 2.17–10.73), and stomach (SIR = 298, 95% CI: 1.49–5.96). Prostate cancer incidence was higher in patients with a NEN of the caecum (SIR = 2.88, 95% CI: 1.37–6.04), lung (SIR = 1.68, 95% CI: 1.15–2.45), and rectum (SIR = 3.46, 95% CI: 2.15–5.56). Renal cancer incidence was significantly higher in patients with a NEN of the lung (SIR = 3.85, 95% CI: 2.36–6.29), pancreas (SIR = 3.51, 95% CI: 1.75–7.01), rectum (SIR = 6.03, 95% CI: 2.71–13.43), small intestine (SIR = 3.73, 95% CI: 2.35–5.91), and stomach (SIR = 3.49, 95% CI: 1.31–9.30). Thyroid cancer incidence was significantly higher in patients with a NEN of the caecum (SIR = 8.29, 95% CI: 1.17–58.88), lung (SIR = 13.32, 95% CI: 8.03–22.09) and small intestine (SIR = 6.68, 95% CI: 3.00–14.86).

Online supplementary Table 5 shows the SIRs when stratified by NEN site, for each second primary tumour occurring after NEN diagnosis which was not found to be significant in the overall analysis. Most had very few expected or observed cases, and hence SIRs were not able to be calculated. Incidence of rectal cancer was significantly increased in patients with several NEN sites including the lung (SIR = 3.85, 95% CI: 2.36–6.29), pancreas (SIR = 3.51, 95% CI: 1.75–7.01), rectum (SIR = 6.03, 95% CI: 2.71–13.43), small intestine (SIR = 3.73, 95% CI: 2.35–5.91), and stomach (SIR = 3.49, 95% CI: 1.31–9.30).

The time between diagnosis of NEN and non-NEN tumours varied greatly across the non-NEN tumour types (online suppl. Table 6). The median times ranged from just over 2 months for small intestine tumours to 40 months for stomach tumours. For the non-NEN tumours found to have significant SIRs, the longest median time between diagnoses was for prostate cancer at just under 17 months (IQR: 6.80–30.10). The shortest median time between diagnoses was for renal tumours at just under 4 months (IQR: 1.80–14.90). Both colon and thyroid non-NEN tumours were diagnosed at a median time of 7 months after NEN diagnosis, whilst the median time between NEN and lung tumours was 14 months (IQR: 1.90–35.40).

Results from this study have shown that patients with NENs are at an increased risk of lung, colon, prostate, renal, and thyroid cancer compared to the background population in England. In particular, renal and thyroid cancers displayed markedly increased SIRs. There was no clear pattern of metachronous tumour incidence when stratified by NEN site. All eight NEN sites investigated in this study were associated with an increased incidence of at least one metachronous tumour site.

In their 2013 study, Krausch et al. [17] summarized studies looking at second malignancies in NEN patients. They reported rates ranging from 7.1 to 31.3% [18, 25]. Whilst overall, 15% of patients in this study had a SPM, only 3% of these were diagnosed after NEN which is lower than previously reported.

Results from this study somewhat echo the results from the study by Tichansky et al. [12] which utilized the SEER database. In their study, the authors concluded that patients with colorectal carcinoids had an increased rate of cancer in the colon, rectum, small bowel, oesophagus, stomach, lung, urinary tract, and prostate. Our study was broader, including patients with eight different NEN sites: appendix, caecum, colon, lung, pancreas, rectum, small intestine, and stomach. Like Tichansky et al. [12], our study also found an increased rate of second cancer of the colon, lung, and prostate in NEN patients. When our results are stratified by NEN site, those with a NEN in the colon had an increased rate of non-NEN cancer of the colon, and those with a NEN in the rectum had an increased rate of non-NEN cancer of the stomach (females only), prostate and kidney, echoing results from Tichansky et al. [12] We speculate that the association of the colon NET with colon cancer may be partly due to surveillance imaging as part of cancer follow-up.

Another study by Verrico et al. [26] examined data from a single institution in Italy over a 10 year period. In their study, the authors found that the incidence of a SPM in patients with a gastroenteropancreatic NET was 11.45%. They reported that 30 patients had a SPM tumour, 20 of which were metachronous tumours. The most common locations for these metachronous tumours were the colon and pancreas. In the current study, prostate and lung were the two most common non-NEN sites to be diagnosed after NEN diagnosis, however, colon was the only non-NEN cancer occurring frequently as a second tumour diagnosed at any time around NEN diagnosis.

Tsai et al. [13] conducted a similar study to ours in the Taiwan Cancer Registry. They found no significant differences in the cohort demographics for the patients with and without a second cancer. Whilst they found that patients with a NEN were at an increased overall risk of developing a second cancer compared to the general population, they only found that this association remained for kidney and bladder cancer. However, like the current study, the authors concluded that there appeared to be no preference over second cancer type according to the primary site of NETs.

While exact reasons for our finding of the increased incidence of second tumours are yet to be elucidated, there are several possible explanations which we suggest warrant further investigation. There may be a genetic link between the NENs and the other tumour types meaning that patients who have a NEN are more susceptible to developing a second malignancy. This link was also postulated by Tsai et al. [13] in their study with mutations in the MEN-1 and RET genes being suggested. Another mechanism might be a defect in mismatch repair genes, implicated in the oncogenesis of many cancers. Further reasons may include shared risk factors including lifestyle (e.g., smoking and alcohol), age, and sex. A small proportion of patients may also have had radiotherapy to treat their NEN. This treatment may have led to the development of another cancer, though there is no strong evidence for this link thus far. Repeated computerized tomography (CT) scanning may also accumulate radiation doses, with some NEN patients undergoing tens of CT scans in their treatment and surveillance course. Prommegger et al. [25] additionally postulated an association between neuropeptides which are produced and secreted from the NENs, and second malignancies. Examples of these neuropeptides include gastrin and cholecystokinin which can stimulate gastric mucosal and pancreatic cell growth [25]. Furthermore, other non-neuropeptide growth factors such as PDGF, FGF, TGF, IGF-I, and IGF-II, also produced by NENs, are known to contribute to the development of second tumours and may provide another explanation [17].

Another possible reason for the increased rate of metachronous tumours is incidental findings or “incidentalomas” in patients who are already undergoing scans for surveillance of their NEN. This may be especially true for lung cancers in patients undergoing chest CT scans. Thyroid abnormalities detected through scans, which subsequently spark further investigation, may also in part explain the high SIR observed in this study for thyroid cancer. Having said this, the median time lag between diagnosis of NEN and thyroid tumours was 7 months implying some second tumours are diagnosed at a slightly later date to the corresponding NEN. It is difficult to estimate a background rate of incidental tumours due to the absence of any true control population; however, several studies have indicated this to be high within their study populations [27, 29].

Results from this study suggest that clinicians looking after patients with NEN in England should be vigilant for second cancers, in particular cancers of the lung, prostate, kidney, colon, and thyroid. We would agree with the proposal by Prommegger et al. [25] that a NEN should be regarded as an “index tumour” and that risk-adapted follow-up with rigorous but appropriate investigations should be employed for these patients. However, we must also be mindful that screening for colorectal cancer is not always deemed beneficial as discussed by Bretthauer et al. [30]. To this end, we have formulated a table with some suggested surveillance regimens for different types of NETs (Table 2). This is for information purposes only and should not be treated as a guideline.

It is important to advise clinicians and patients accordingly on these findings. With regards to screening for the cancers with a higher risk compared to the general population in our study (lung, prostate, kidney, colon and thyroid), there are several investigations available. In England, men aged over 50 can request a prostate-specific antigen (PSA) test through their general practitioner as a measure of patients who may be at a higher risk of prostate cancer [31]. Whilst this is not yet part of a national screening programme, it may be beneficial for some patients. The bowel cancer screening programme is open for patients aged between 60 and 74 and is currently being expanded to patients aged 50–59, thereby fitting in with the mean age of patients included in our study.

Lung cancer screening trials have taken place in several countries including the National Lung Screening Trial in USA [32], the UK Lung Screening Trial (UKLST) [33], the NELSON trial (in Belgium and Holland) [34, 35], and the Danish Lung Screening Trial in Denmark [36]. However, despite a reduction in lung cancer mortality, these trials have failed to convince policymakers on the benefit of national lung screening programmes [37]. There does exist, however, an “NHS Lung Health Check” screening programme in certain parts of England involving a low-dose CT scan of smokers aged 55–75 [38]. Thyroid cancers are relatively rare and previous screening trials have increased incidence but failed to decrease mortality from the disease, hence no screening programme for this cancer exists [39, 41]. For kidney cancer, some patients who are at a higher risk, for example, those with von Hippel-Lindau disease may be referred for annual scans [42, 43]. However, there is no formal screening programme for kidney cancer in England.

Given there are currently no screening programmes for prostate, kidney, or lung cancer in England, we therefore propose that monitoring through scans, blood and urine testing and increased symptom awareness is recommended for the detection of these cancers in NEN patients. Furthermore, engagement in existing screening programmes is encouraged. Renal, lung and kidney cancers will often be picked up on CT scans of the chest, abdomen, and pelvis performed for at least 5 years after diagnosis of NET. However, prostate and colon cancers will only be picked up at a late stage by these scans and hence PSA, faecal immunochemical testing and colonoscopy should be considered during follow-up of NEN in high-risk groups. For thyroid cancers, since screening is not deemed beneficial, further studies are warranted to investigate possible reasons for the increased incidence and detection of thyroid cancers in NEN patients.

Our findings indicate a critical role in counselling NEN patients on the fact that other cancers may be found on the scans being done for their NEN. The outcome of detection of incidental malignancies can feel good and bad for patients, good because a potentially life-threatening cancer has been caught early and bad because non-serious findings can increase anxiety and impact quality of life or even have implications for healthcare in the form of insurance premia. As far as we know, no USA or EU guidelines exist for the handling of incidental finding of thyroid, colon, kidney or prostate cancer, therefore our study also indicates the need for these to exist [29].

To the best of our knowledge, this is the first nationwide European study to examine the rate of metachronous tumours in patients with NENs when compared to the incidence of a background population. The study benefitted from a large study cohort allowing for stratification by NEN type, however, despite this, some subgroups had small sample sizes due to the nature of NEN as a rare cancer. This study benefitted from the inclusion of data from the reference population to produce SIRs and not just crude results, as many previous studies have been limited to. A limitation to this study was the paucity of data on shared risk factors such as smoking status and alcohol intake as well as comorbidities such as diabetes. Other data which would have been beneficial to incorporate into the study include data on genetics and hormone secretion. Examining the possible genetic link between NENs and second cancers was not possible but would be a clear area for future research.

This study found that patients with a NEN experienced a metachronous tumour of the lung, prostate, kidney, colon, and thyroid at a higher rate than the general population of England. There was no clear pattern between the site of NEN and subsequent non-NEN tumour site. Based on our findings, we would suggest a need for increased vigilance by clinicians and the importance of counselling NEN patients on this risk. Surveillance and engagement in screening programmes are suggested to enable earlier diagnosis of second non-NEN tumours. Further studies are warranted to determine whether there is a genetic link between these tumours in order to effectively risk stratify patients.

The study was approved by the North West – Greater Manchester South Research Ethics Committee (IRAS project ID: 284875; REC reference: 20/NW/0342). The study was sponsored by the Hampshire Hospitals NHS Foundation Trust. Patient consent was not required as this study was based on publicly available data.

The authors have no conflicts of interest to declare.

This work was funded by Neuroendocrine Cancer UK.

Beth Russell: conceptualization, data curation, formal analysis, investigation, methodology, project administration, writing – original draft, and writing – review and editing.

Benjamin White: conceptualization, data curation, and writing – review and editing.

Brian Rous and Kwok Wong: data curation, validation, and writing – review and editing.

Catherine Bouvier Ellis: writing – review and editing.

Rajaventhan Srirajaskanthan: conceptualization, and writing – review and editing.

Mieke Van Hemelrijck: methodology, supervision, and writing – review and editing.

John Ramage: conceptualization, methodology, supervision, and writing – review and editing.

Data are available from the National Cancer Registration and Analysis Service (NCRAS), which is part of NHS Digital. Access to the data was facilitated by the Office for Data Release [8]. Further enquiries can be directed to the corresponding author.