Factors Influencing Costs of Cancer Care for Patients with Neuroendocrine Neoplasms

Neuroendocrine neoplasms (NENs) are comparatively rare tumours arising from the enterochromaffin cells. NENs can origin from various organs. In Europe, most common NENs are well-differentiated neoplasms of the gastroentero-pancreatic tract (GEP-NENs), followed by lung primary carcinoids (typical [TC] or atypical lung carcinoids) [1, 2]. Neuroendocrine carcinomata (NECs) may represent some 10–20% of all NENs [2]. However, the percentage of NEC is likely overestimated based on previously common, inaccurate labelling of high grade (grade 3, Ki-67 index >20%), but well-differentiated NENs as NECs [3]. NECs differ in morphology, behaviour, and prognosis and require different strategies. Recent updates of European Neuroendocrine Tumour Society (ENETS) guidance papers for the management of people with GEP-NEN [4‒10] and lung carcinoids [11, 12] have been published. The 20-year limited-duration prevalence of NENs is now deemed to be ∼48/100,000 [2, 13]. However, this is almost certainly still an underestimate considering that not all people harbouring a NEN are diagnosed, and even if so, inclusion of all people who were diagnosed with a NEN in the respective databases cannot be guaranteed, related to limitations of available coding options and applying inaccurate codes when using standardised systems such as the International Classification of Diseases (ICD-10-CM) tool. Due to prolonged survival in many cases of metastatic NENs, combined with the rising incidence of NENs, and earlier detection, prevalence is increasing relevantly, with implications on healthcare resources.

For the purpose of this commentary/narrative review, a literature search in PubMed was conducted, using different combinations of the following keywords: “neuroendocrine neoplasms,” “neuroendocrine tumour,” “neuroendocrine carcinoma,” “cost,” and “economic.” Only articles written in English were included. Articles and review papers not directly relating to NEN cancer care were excluded. Focus was on the most impactful and clinically relevant reports and included published review articles. Considering constant changes in healthcare costs over time, focus was on articles published from 2020, but only few studies were identified. All costs given relate to National Health System (NHS) prices, unless otherwise indicated. Conversion rates GBP to EUR and USD were calculated as per January 26, 2025.

Rare diseases are defined in the European Union as diseases with a prevalence of <5 cases out of a population of 10,000 [14]. However, the Surveillance of Rare Cancers in Europe (RARECARE) project proposes a definition of rare cancers based on incidence [14]. According to this definition, rare cancers are identified as those with an incidence of <6/100,000 persons per year. Based on this, at least in terms of prevalence, NENs do now approach the threshold of not qualifying to be classified as a “rare cancer,” with increasing implications on healthcare resources.

Economic implications of NEN cancer diagnostics and treatment have been addressed in some recent studies [15‒20]. Most studies were performed in the USA, and NEN-related costs in other countries and health systems may not be comparable. In one of the few studies performed outside the USA, a population-based register-linkage study from Sweden [19] including 478 grade 1 and grade 2 metastatic GEP-NEN patients with mainly small bowel primaries (80% small intestinal NEN; 10% pancreatic NEN; 41% with symptoms of carcinoid syndrome (CS) found that main contributor to total costs were costs of direct medical care (77%), of those 42% related to costs of drugs (mainly due to use of somatostatin analogues [SSAs]); and 35% related to healthcare resource use. The reminder of costs (22%) was related to production loss, including sickness absences and healthcare visits/hospital admissions [19]. The total annual cost (in 2013) was reported as EUR 37,300/GBP 31,332/USD 39,165 in pancreatic NEN, EUR 24,800/GBP 20,832/USD 26,040 in small bowel NEN, and EUR 18,600/GBP 15,624/USD 19,530 in metastatic GEP-NEN with other primaries. Although treatment with SSA was more frequent in patients with small bowel (77%) versus pancreatic (31%) NEN, total costs were higher in pancreatic NEN related to increased production loss, and higher cost of molecular targeted treatment. In a study performed in Canada, Hallet and colleagues compared drivers of costs in 3,827 NEN patients with 9,320 colon cancer patients [16]. Authors identified higher costs in NEN patients leading to a diagnosis, and a more than 3-fold higher cost long-term post-diagnostic in NEN, driven by drug costs. Comorbidities and higher age predicted increased NEN costs also, as well as socioeconomic status; and the site of the primary NEN [16].

Grande and colleagues [18] performed a systematic review in 2018, identifying 14 studies that included cost-of-illness analyses (n = 4), economic evaluations (n = 7), and budget impact analyses (n = 3). Results indicated that reducing adverse events such as diarrhoea caused by CS could be an area where cost savings could be achieved, but the exact impact of adverse events on costs was not available [18]. A later study performed in Spain also identified uncontrolled CS resulting in higher health resource use, including emergency department attendances, and more frequent and prolonged hospitalisation; with mean annual total healthcare costs 60% higher in patients with uncontrolled versus controlled CS [21]. In a retrospective economic study on dose escalation of SSA including patients between 2000 and 2012, Huynh and colleagues found that patients with CS symptom improvement had significantly lower annual healthcare costs (USD 14,766/GBP 11,813/EUR 14,028 per patient) than patients with continued symptoms of flushing or diarrhoea. This study included an analysis of indirect costs, e.g., hours spent at clinic visits and corresponding loss of earnings [22]. Ramage and co-workers concluded that health economic evaluations (HEEs) in NEN were mainly industry funded and most HEEs did not meet published health economic criteria used to assess quality [20]. This is relevant considering that, e.g., in the UK, HEEs are increasingly used by the government to assess treatments; and economic evaluation appears to have a major influence on the decisions of the National Institute for Health and Care Excellence (NICE) whether a newly available treatment will be recommended [20, 23].

Incidences of NENs in 2025 can only be estimated but based on previous publications, the incidence of GEP-NENs has increased more than 6-fold between 1997 and 2012 [13] and appears to be further increasing since [24‒26]. The incidence of GEP-NENs in the USA based on an update of the Surveillance, Epidemiology and End Results (SEER) database is estimated to be 3.56/100,000/year, and was reported, on average, slightly lower in European populations, but in some countries, e.g., Norway, has been reported as high as 5.83/100,000 [27]. Further, whereas in most of Europe small intestinal and pancreatic NEN appear to be most common, more frequently observed primary sites of GEP-NENs reported in North America are of small intestinal and colorectal origin, and again different in Asia, where rectal, gastric, and pancreatic NENs tend to be predominant [3]. In addition, there appear to be relevant and unexplained differences even between European countries, e.g., in Iceland incidence of GEP-NENs appears to be unchanged; and a prospective multi-centre tumour registry in Greece reported gastric NENs as the most common primary. The reason for these observed differences is unknown but could be related to biological differences between populations of different ethnicities; and environmental factors including dietary habits [3] that may not substantially change even when moving to another geographical location.

Heterogeneity in reported results of different studies is substantial, e.g., some studies reported GEP-NEN incidence rates based on primary tumour sites, others by disease stage, or by ethnicity; or age adjusted; or, in cross-sectional studies, prevalence only can be reported [3]. Therefore, findings of epidemiological studies commenting on NEN incidence are challenging to compare. To assess the true healthcare burden of NENs, uniformly conducted population-based epidemiologic studies in each country would be required [3]. However, even if this could be implemented, now more widely available high-sensitivity diagnostic tools such as 68Ga PET-CT scans and constantly further improving cross-sectional imaging in more recent studies lead to higher detection rates per se, so any 1: 1 comparison with older epidemiological reports will remain to be challenging.

Although higher grade NENs and particularly NECs typically behave more aggressively, many lower grade, well-differentiated NENs tend to be slow growing, resulting in prolonged survival even in the presence of metastatic (stage IV) disease. The largest available retrospective population-based study using nationally representative data from the SEER program in the USA identified 64,971 NEN patients between 1973 and 2012 [13]. Analyses using associated population data included assessment 5-year overall survival (OS) rates. On multivariable analyses, the median 5-year OS rate varied significantly by primary tumour site, stage, grade, age at diagnosis, and time period of diagnosis. The OS rate for all NENs significantly improved from the 2000–2004 period to the 2009–2012 period (hazard ratio 0.79; 95% CI, 0.73–0.85), with even more pronounced improvements noted in stage IV gastrointestinal (OS improved by 29%) and pancreatic NENs (OS improved by 44%) [3, 13]. In a study performed in Germany identifying 2821 patients with GEP-NENs between 1976 and 2006, the 5-year OS rate increased between the 1976–1988 period and the 1998–2006 period for all GEP-NEN from 50% to 79% [28]. In another population-based study population-based study from the SEER 18 registry (2000–2016), 8,944 patients with pancreatic NENs were identified [29]. Annual incidence rates increased from 0.27 to 1.00 per 100,000, which was largely explained by an increasing number of patients diagnosed with localised disease in more recent years (2012–2016). Median OS was 68 months and 5-year OS rates in localised, regional, and metastatic disease were 83%, 67%, and 28%, respectively. Again, OS significantly improved across all disease stages for patients diagnosed between 2009 and 2016 (median OS 85 months), compared with patients diagnosed between 2000 and 2008 (median OS 46 months) [29]. Reported survival rates are expected to improve further related to earlier detection of NENs at lower disease stages, due to increased awareness and use of more sensitive diagnostic tools; and advances in treatment; ultimately – if not curative – adding to cost pressures.

Earlier detection of a NEN has higher chances of lower disease stages at the time of diagnosis, with potentially curative, surgical resection, which also from an economical point of view is the preferred option. However, unless cured by surgical resection, most patients eventually have disease progression.

Treatment of advanced NENs is guided by various features including the primary tumour location, tumour stage and grade, presence of symptoms related to hormone secretion, and somatostatin-receptor expression, with goals of systemic therapy being slowing down of tumour growth and controlling clinical symptoms [30]. Liver debulking surgery, if technically feasible, may improve survival in metastatic well-differentiated NENs [31‒34] and can improve functioning symptoms. Liver transplant in highly selected patients with metastatic NENs has been proposed, but this treatment option comes with risks including the need for indefinite immunosuppression, and efficacy is discussed controversially [35‒38], also considering selection bias when prioritising younger and fitter patients for transplant procedures [39, 40].

First landmark progresses related to NEN drug treatments were made in the 1980s, with availabilities of SSA and a-interferon. The latter is now rarely used related to side effects and limited availability, but treatment with SSA remains the first-line treatment in people with well-differentiated NENs, either in preparation for surgical, potentially curative resection, or in an attempt slowing down tumour growth, and to tackle functioning symptoms in patients with CS. Improvement of flushing and diarrhoea is achieved in 70–80% of patients by using long-acting formulations [2]. Costs of treatment with SSA are significant (e.g., the NHS drug tariff price of long-acting lanreotide is given as GBP 937/EUR 1,115/USD 1,171 for 1 standard dose of 120 mg 4 weekly; NHS tariff price for long-acting octreotide is given as GBP 998/EUR 1,188/USD 1,248 for 1 standard dose of 30 mg 4 weekly) and were identified as the largest contributor (91% of the total drug cost) to the direct medical costs in a Swedish GEP-NEN cohort [19]. In case of progression or worsening of CS, it is common practice to increase the SSA dose to greater than the standard dose by reducing the injection interval (e.g., from 4 weekly to 3 weekly, or 2 weekly), although there are only limited prospective data to support this approach [2]. Advanced SSA preparation is in development, including new, advanced self-injectables, and oral capsules.

Since the early 2010s, the treatment landscape in NENs has further evolved, including availability of molecular targeted agents; and treatment using alkylating agent chemotherapy with temozolomide-based regimens for patients with advanced pancreatic NENs [41]. Although temozolomide-based chemotherapy is still evolving, it has partially replaced streptozocin-based regimens in pancreatic NENs related to better tolerability, and could have benefits in, e.g., high-grade, non-functioning well-differentiated small bowel NENs and lung carcinoids also [42]. Mutation targeting treatment is not promising in low-grade NEN due to low mutational burden but could be promising in NEC and high-grade (grade 3) NEN. However, molecular testing, e.g., of O(6)-methylguanine-DNA methyltransferase (MGMT; a DNA repair enzyme that functions against the DNA methylation caused by alkylating agents such as temozolomide) status may be useful to predict objective tumour response to treatment with temozolomide and prognosis [43, 44].

Further treatment options now include mammalian target of rapamycin (mTOR) inhibitors [45]; approved for the treatment of progressive, advanced, low, or intermediate-grade pancreatic NENs; with potential additional applications in non-functioning lung and small bowel NENs being investigated. Tyrosine kinase inhibitors (TKIs) such as sunitinib [46] add to the therapeutic arsenal and are currently approved for the treatment of advanced pancreatic NENs. New TKI targeting multiple tyrosine kinases including MET and AXL (both MET and AXL may facilitate resistance against vascular endothelial growth factor receptor-targeted tyrosine kinase therapy at least in clear cell renal cell carcinoma [47]), in addition to targeting vascular endothelial growth factor receptors are in development, showing significantly improved progression-free survival in patients with previously treated, progressive, advanced extra-pancreatic or pancreatic NENs, as compared with placebo [41].

Telotristat ethyl (TE) is an oral inhibitor of tryptophan hydroxylase, a rate-limiting enzyme in the synthesis of serotonin, which showed significant improvement in the number of bowel movements in a placebo-controlled phase III trial (TELESTAR) in 135 patients with refractory CS-related diarrhoea. Durable response (defined as 30% improvement in bowel movements for >50% of the 12-week core study period) occurred in 44% and 42% of the patients treated with 250 mg or 500 mg 3 times daily, respectively [48]. A second placebo-controlled trial (TELECAST) including 76 patients with less frequent bowel movements supported the efficacy and good tolerability of TE [48]. We previously reported significant weight gain and improved nutritional status in patients with CS who were treated with TE [49, 50]. In TELSTAR, 32.5% of patients treated with TE experienced significant, dose-dependent weight gain, associated with reduced diarrhoea severity and improved biochemical and metabolic parameters [49], and these findings were confirmed in a combined analysis using pooled data from the 36-week open-label extensions of both trials, showing that improvements in weight and nutritional parameters were sustained, or further improved in patients with CS through to week 48 of treatment with TE [50]. These studies suggest that TE, alongside routine clinical management, may provide long-term benefits in nutritional intake and weight evolution in patients with CS [49, 50].

Treatment with now advanced peptide-receptor radionuclide therapy (PRRT) options, e.g., PRRT using 177-Lutetium (177-Lu)-DOTATATE was approved by the European Medicines Agency only in 2017 and by the US Food and Drug Administration (FDA) in 2018, for the treatment of somatostatin receptor (SSR) positive, advanced GEP-NEN [51]. Use of Lutetium PRRT (Lu-PRRT) is established in progressive, non-resectable, metastatic somatostatin-receptor expressing GEP-NEN. In 2024, the FDA approved Lu-PRRT for paediatric patients from age 12 years with the same indications [51]. Use as first-line treatment in high grade 2 and grade 3 NENs [52] and in lung carcinoids [53] is under development.

Further drugs in clinical trials or early development include combination therapies with alpha emitters; advances in cancer immunotherapy, such as use of bispecific T-cell engagers; or insulin receptor antibodies in insulinoma. Ongoing emergence of novel treatment options can be anticipated, which is, of course, crucially important considering the still limited long-term survival in people with advanced NENs; however, unless resulting in cure, available treatments will also further contribute increasing the costs of NEN cancer care.

Based on a web-based survey performed in England [54] (n = 229 with complete responses), median time from first symptom to diagnosis of a NEN was 53.8 months, and 80% of respondents visited their GP a median of 11 times prior to diagnosis. Authors concluded that there is a long time from onset of symptoms to diagnosis in all types of NENs, despite many respondents having had alarming symptoms such as pain (33%), diarrhoea (22%), or flushing (17%). Further, many people are asymptomatic before being diagnosed with a NEN, and where symptoms do occur, the most frequent are non-specific, e.g., tiredness, bloating, or loose stools. Therefore, “Simply getting to the point of suspicion is a significant milestone, given the often non-specific symptoms of NENs” [55]. Ongoing education and awareness of NENs in the primary care setting may support earlier diagnosis. Apart from relieving long-standing and sometimes debilitating NEN symptoms earlier, this will also optimise resource use, related to reduced costs when treating NENs at earlier disease stages, in addition to driving down production losses, and releasing GP workforce.

Diarrhoea in people with GEP-NENs can be multifactorial, with misdiagnosis leading to delayed patient recovery and inefficient resource use [56]. Diarrhoea in CS is resulting from excess serotonin secretion, usually but not exclusively in presence of metastatic liver involvement, and is always secretory, persists with fasting, may occur during the night, and is associated with abdominal discomfort and faecal urgency [57]. CS caused diarrhoea is burdensome, and, according to a retrospective study (2002–2012) performed in the USA, generates adjusted mean annual costs of USD 81,610/GBP 65,288/EUR 77,530 [58].

However, based on a review of ∼10,000 incident cases of NENs in patients 65 years of age and older diagnosed over more than a decade, the aggregate proportion of patients manifesting CS was 18.8% only, ranging from 7.6% in lung carcinoids to 32.4% in small bowel NENs, respectively [59]. Apart from CS, other causes of diarrhoea in patients with NENs must be considered and may include pancreatic masses, previous surgical treatment of pancreatic NENs, pancreatic exocrine deficiency for various reasons, complications following treatment of small bowel NENs including small intestinal bacterial overgrowth, bile acid malabsorption, short-bowel syndrome, and/or mesenteric ischaemia. NEN-unrelated causes such as celiac’s, thyrotoxicosis, colitis, or gastroenteritis should be excluded as well. Further, possible side effects of certain treatments leading to diarrhoea should be considered, such as sunitinib-induced pneumatosis cystoides intestinalis. In a recent systematic review covering the differential diagnosis of diarrhoea in patients with NENs, authors highlighted studies where presence of bile acid malabsorption was as high as 80% of NEN patients who had been referred to a gastroenterologist service and were tested using the 75-selenium homocholic acid taurine scan, and presence of small intestinal bacterial overgrowth, diagnosed by breath tests, was reported in 23.6–62% of NEN patients who had been tested for the condition [56]. Another study reported that 20% of patients with small bowel or pancreatic NENs met the criteria for irritable bowel syndrome (IBS) with diarrhoea, demonstrating that patients with GEP-NENs can be initially misdiagnosed with IBS, or may have synchronous IBS [60]. Considering other than CS-induced potential causes of diarrhoea in people with NEN is important to optimise treatment, reduce chances of inappropriate treatment [61, 62], improve the quality of life of the patient, and reduce costs [56].

Malnutrition is common in people with cancer, influencing quality of life, treatment options, and survival [63] with additional implications on costs. Patients with GEP-NEN who are on treatment with SSA appear to be at particular risk of malnutrition [64‒68], possibly in part related to SSA treatment-induced pancreatic exocrine insufficiency (PEI) [62, 69]. After acute, short-term administration, rapid onset suppression of pancreatic exocrine secretion by SSA with associated steatorrhea has been reported (discussed in [64]). In a prospective, observational study in 50 patients with advanced, well-differentiated NENs, 24% of the patients developed SSA-related PEI after a median of 2.9 months from SSA initiation [69]. Measurement of faecal elastase was a reliable screening tool for PEI, especially if symptomatic (abdominal bloating, flatulence and/or diarrhoea, steatorrhea) [69].

In 2019, the Global Leadership Into Malnutrition (GLIM) criteria for the diagnosis of malnutrition were published, now including the presence of sarcopenia as one of the criteria, in addition to body weight, weight loss, and BMI [70]. Sarcopenia in cancer patients is common, e.g., prevalence of sarcopenia in patients with a recent diagnosis of gastrointestinal adenocarcinoma is reported as high as 30–60% [71], and associated with poorer outcomes of surgery, increased risk of chemotherapy toxicity, physical impairment, and shorter survival [71]. Few studies investigated presence of sarcopenia in patients with metastatic GEP-NENs [71‒73]. In the largest study published today (n = 183; of those 74% with metastatic small bowel NENs and 24% with metastatic pancreatic NENs), sarcopenia was present in 128 patients (69%) and unrelated to BMI. In patients with pancreatic NENs, the presence of sarcopenia was independently associated with shorter OS. In another study including 104 patients (of those, n = 52 patients had metastatic disease) with a recent diagnosis of GEP-NENs, 87% of patients had sarcopenia [72]. Based on these findings, the prevalence of sarcopenia at the time of diagnosis of a metastatic GEP-NEN could be even higher than in patients with metastatic adenocarcinoma.

Only one study in n = 118 has assessed GLIM criteria in patients with GEP-NENs on SSA (40% of patients had functioning symptoms), including assessment of associations with OS [73]. Overall, malnutrition was present in 88 patients (75%), based on low BMI in 26 (22%) patients, based on weight loss in 35 (30%) patients, and based on sarcopenia in 83 (70%) patients. Vitamin D deficiencies were present in 64 patients (54%), and vitamin A deficiencies in 29 patients (25%). Combined, presence of malnutrition appears to be excessively high in GEP-NEN patients who are on treatment with SSA. When malnutrition is related to weight loss, there are associations with poor OS [73]. Patients could benefit from regular weight monitoring and early nutritional intervention, including consideration of pancreatic enzyme replacement therapy if appropriate [62, 74].

Appropriate nutritional support is associated with cost savings in cancer patients [75]. Further, recognising potential SSA-induced side effects causing, e.g., bloating, diarrhoea, and possibly weight loss related to suppression of pancreatic digestive enzymes is important to guide appropriate treatment, and can reduce cost of under these circumstances inappropriate up-titration of SSA treatment [61, 62].

Even if replacement with pancreatic enzymes is being considered, this is frequently under-dosed; and, therefore, not efficient. In this context, is has not helped that higher strengths (containing, e.g., 40,000 units of lipase per capsule) preparations of pancreatin replacement are not available anymore, at least in the NHS. Currently available pancreatic capsules contain 5,000–25,000 units of lipase. Under normal physiologic circumstances, postprandial lipase secretion has been estimated at 9,000–18,000 units/min, totalling between 120,000 and ∼ 2,200,000 units in the 3-h postprandial period [76]; of those, some 50% being absorbed. Although based on intubation studies in humans, only 5–10% of normal pancreatic enzyme output may be required for sufficient fat absorption [76], some patients may need more than 2 or 3 capsules of pancreatin (25,000 units of lipase per capsule) to relief symptoms, especially when consuming meals with higher fat contents. However, increasing, the replacement dose to more than 5–6 capsules per meal often does not result in further improvement of steatorrhea, and may even cause issues related to the highly acidic enteric coating, with the possible side effect of developing strictures due to fibrosing colonopathy [76]. Prices of pancreatin 25,000 units capsules vary but on average are given as some GBP 28/EUR 33/USD 35 per 100 capsules (NHS indicative price, as per British National Formula [BNF] accessed 11/2024). Sufficient dosing is important both for symptom relief and cost-efficient use of this treatment.

Diabetes treatment and management are a considerable cost burden, accounting for more than six per cent, e.g., of the UK health budget in 2021/22 [77].

Patients with NENs may have impaired glucose tolerance or diabetes for various reasons. In case of increased BMI and family history of type 2 diabetes, the risk of developing the condition in people with NEN is expected to be similar to the general population. However, many patients with NENs are on treatment with SSA, which can interfere with the physiological secretion of various hormones including insulin and glucagon [78]. The exact effects of SSA on hormones affecting glucose tolerance in the body is not predictable, e.g., in some people glucose lowering insulin secretion is more relevantly affected than secretion of, e.g., counter-regulatory (and as such glucose increasing) glucagon. Thus, in people with pre-diabetes related to increased insulin resistance caused by high BMI and body fat mass, treatment with SSA may worsen (or possibly improve) glucose control and closer monitoring of the glucose levels is recommended.

Other factors known to affect glucose control include now less frequently used treatment with streptozocin (with the known “side effect” of directly targeting the insulin producing beta-cells in the pancreas, and, therefore, widely used to induce “type 1” diabetes in experimental models [79]) in patients with pancreatic NENs; and treatment with mTOR inhibitors such as everolimus, with mTOR also known to be a key step in the insulin signalling cascade [80‒82].

In people with impaired (e.g., due to a large, NEN-related mass in the pancreatic head, disturbing the insulin producing beta-cells; or non-NEN-related causes, e.g., previous episodes of pancreatitis related to gallstones or other causes, including endoscopic ultrasound-guided fine-needle aspiration (EUS FNA) induced pancreatitis [83]), or absent (e.g., following Whipple’s surgery for a pancreatic NEN located in the pancreatic head) insulin secretion, the resulting secondary (type 3c diabetes) can be a relevant issue if insulin replacement is not optimal. In contrast to type 1 diabetes, with selective loss of insulin producing beta-cells, people with type 3c diabetes may also lose counter-regulatory physiological hormone responses from the endocrine pancreatic cells involved in the regulation of glucose control, most importantly glucagon and somatostatin, in addition to loss of insulin producing cells. Therefore, in people with type 3c diabetes, failure to ensure a reliable 24-h cover with basal insulin [84] and even relatively minor mis-dosing of fast acting bolus insulin can result in rapid swings including severe hypoglycaemic events, with delayed counter-regulation by (other than glucagon and somatostatin) hormones in the body, such as cortisol, growth hormone, and adrenalin, resulting in a delayed, prolonged hyperglycaemic state, and potentially life-threatening ketoacidosis [85].

A recent study investigating annual diabetes healthcare resource utilisation in adults with diabetes in the USA found that people with type 1 diabetes had significantly higher mean total costs than the type 2 diabetes cohort (USD 18,817/GBP 15,054/EUR 17,876 vs. USD 14,148/GBP 11,318/EUR 13,441 per patient per year) [86]. It is anticipated that costs of diabetes care could be higher in type 3c diabetes, given the frequently observed particular difficulties achieving stable glucose control in this type of diabetes. In people with type 3c diabetes, long-term specialist input from a diabetologist team is strongly recommended.

Cost of measuring NEN-related biomarkers are relatively modest today. However, when using these markers to support diagnosis and monitoring of people with NEN, it is important to measure these markers under standardised conditions, to minimise the risk of false-high and as such misleading results, which can influence diagnostic and therapeutic decisions and, as such, unnecessarily add to cost pressure. Costs vary between countries and institutions; in our facilities, a blood test for measuring chromogranin A (CgA) comes at GBP 57/EUR 68/USD 71, a full “gut hormone profile” (pancreatic neuroendocrine markers, gastrin) costs GBP 97/EUR 115/USD 121, and analysis of a 24-h urine for measurement of urine 5-hydroxyindolacetic acid (5-HIAA) costs GBP 27/EUR 32/USD 34. Measurement of plasma metanephrines and normetanephrines costs GBP 22/EUR 26/USD 28.

More sensitive methods such as analysis of multiple neuroendocrine tumour transcripts or measurement of circulating tumour cells are promising future clinical tools [87], but to date these methods are not yet widely available in the routine clinical setting, also related to constraints on available funding [88]. Therefore, measurement of “gut hormone profiles,” CgA, and 5-HIAA remain key NEN biomarkers in current clinical practice [89]. Although most NEN Centres sample full gut hormone profiles in the overnight fasted state, in a survey amongst ENETS Centres of Excellence, we found that in the responding Centres in the UK, only 50% of the centres by default invite patients after an >10-h overnight fast for an isolated measurement of CgA [88]. Notably, although sampling of CgA under fasted conditions had been recommended in some of the previous guidelines [90], in other guidelines of the leading Neuroendocrine Tumour Societies either no clear recommendation is made how CgA should be sampled [91, 92], or it is explicitly stated that a fasted specimen for CgA is not required [93]. In the largest reported study assessing effects of food intake on CgA levels, including 28 patients with GEP-NENs in a randomised double-crossover intervention [88], intake of a 5-item English breakfast moderately raised plasma CgA levels in all participant groups but was more pronounced (up to 34%) in controls and in patients with GEP-NENs who were not on treatment with long-acting SSA. Therefore, use of non-fasted plasma CgA measurements may result in false positive findings especially in screening and in follow-up of patients with assumed completely resected disease, and should be discouraged [88]. A reliable >10 h fasted state can be challenging in some patient groups, including individuals with secondary diabetes. Further, in NEN centres with large numbers of patients, it can be logistically challenging restricting the sampling window to the earlier part of the morning for an overnight fasted blood sample. Repeating a CgA measurement in the >10 h overnight fasted state in case of borderline pathological non-fasted CgA results was suggested [88]. Further, in people taking proton-pump inhibitors, CgA levels can be relevantly above normal, potentially causing misleading results in both screening and follow-up [88]. Therefore, if clinically tolerated, proton-pump inhibitor should ideally be paused for up to 2 weeks prior to a measurement of fasted CgA [88].

Measurement of urine 5-HIAA remains another crucial biomarker particularly in NEN patients with small bowel or lung primaries, who are at relevant risk of developing CS and cardiac complications including right heart valvular fibrosis (Hedinger syndrome; “carcinoid heart disease” [CHD]). A urine 5-HIAA level >50 μmol/24 h is considered compatible with the diagnosis of CS [57]; and urine 5-HIAA >300 μmol is associated with a 2- to 3-fold increase in risk of development of CHD and progression [57]. Importantly, although CgA has been suggested as a marker for CS and CHD, CgA levels do not appear to differentiate between NEN patients with or without CS [57]. To reduce the risk of “false” high 5-HIAA results, it is commonly accepted that sampling of urine 5-HIAA should be done following avoidance of certain foods known to contain, or trigger the release of serotonin (e.g., bananas, pineapples, tomatoes, walnuts) for at least 24-h prior to starting the collection period. Various drugs can cause false high 5-HIAA results. Treatment with oral aminosalicylates, e.g., sulfasalazine is known to particularly affect 5-HIAA measurements, related to possible assay interference dependent on the measurement method used [94]). We have repeatedly observed excessively raised 5-HIAA (>10 × upper limit of normal) levels in patients on treatment for inflammatory bowel disease, which normalised after pausing the respective medication (if tolerated) for a few days only. Presence of inflammatory bowel disease such as Crohn’s (ileitis terminalis) can be particularly misleading given these patients may present with skin rushes, which are quite different from flushing as observed in NEN patients with CS but may be confused with “flushing” by patients and healthcare professional less familiar with NENs. Vice versa, some of the patients with assumed inflammatory bowel disease may indeed have a NEN [95]. Keeping in mind that urine 5-HIAA measurements can be strongly influenced by various factors is important when interpreting test results that may lead to a series of potentially unnecessary and costly additional investigations.

Other rather frequently observed issues include sampling of plasma metanephrines and normetanephrine (MTN) under non-standardised conditions (e.g., not sampled in the overnight fasted state following 15 min in the recumbent position) in screening or follow-up for patients with pheochromocytoma or paraganglioma, which can lead to “false”-high, stress related increased MTN levels, thus triggering anxiety, and in many cases resulting in unnecessary additional investigations including scans involving exposure to radiation.

Finally, omitting regular measurements of plasma N-terminal pro-brain natriuretic peptide (NT-proBNP) in routine follow-up in patients with CS and all patients with increased urine 5-HIAA [57] can lead to a delayed diagnosis of carcinoid heart disease (CHD). Although CHD is rare in terms of total numbers, it can develop in some 20–50% of patients with CS [57], with major, independent, negative prognostic implications [96]. If untreated, CHD results in significant morbidity and mortality caused by right heart failure. Importantly, progression of valve involvement and clinical decompensation may happen rapidly, in some cases within months, in >50% of patients [96, 97]. Early suspicion and aggressive lowering of 5-HIAA levels may prevent or delay progression [96]. Interdisciplinary management of CHD is crucial to improve the prognosis of patients with CS, and is expected to drive down costs (NHS states cost of complex repair of multiple heart valves, surgical procedure GBP 13,488–18,092/EUR 16,051–21,529/USD 16,860–22,615; US studies show annual median total medical costs for heart failure care ∼ USD 24,383/GBP 19,506/EUR 23,244/USD 23,164 per patient, with heart failure-specific hospitalisations identified as man driver of costs (median USD 15,879/GBP 12,703/EUR 15,085 per patient) [98].

Patients with suspected or confirmed NENs may present with complex features. Abdominal discomfort is the most frequent initial symptom in patients with small bowel NENs [99]. Combination of lack of physician awareness, often vague presentation in terms of symptoms, and typically slow growth of low-grade small bowel NENs can lead to long delays in diagnosis. Metastases at presentation are seen in ∼30% of patients with small bowel NENs in population-based studies and as frequent as >60% of patients at large NEN referral centres [100]. Especially prior to proceeding with major surgical interventions, staging including the use of 68Ga PET-CT is recommended to identify patients who (unless in case of acute small bowel obstruction) would not be eligible for surgical treatment based on widespread, small volume metastatic involvement, e.g., in the bones that may not be detectable using other scan modalities.

In case of an emergency surgery scenario related to acute complete obstruction, meticulous planning of the optimal surgical strategy is difficult. However, prior to obstruction, most patients with small bowel NENs may present with episodic, crampy abdominal pains, and recurrent partial bowel obstruction, often over many years. In these cases, especially if no other explanation for the symptoms can be identified, and a mass is discovered on cross-sectional imaging, a possible NEN must be considered and referral to a centre with NEN MDT access is recommended. This would help planning for appropriate surgical treatment. Further, considering a small bowel NEN as a differential diagnosis would also usually trigger peri-operative cover with SSA, which may reduce risk of carcinoid crisis due to induction of anaesthesia or the surgical procedure [101]. Guidance related to the optimal management of NENs of different primaries is constantly evolving and without access to a multi-speciality centre with expertise in the management of NENs, there is increased risk of inappropriate treatment, potential harm to the patient, and increased costs.

Progression of “desmoplasia” may cause vascular encasement. Clinically, this may cause abdominal discomfort, diarrhoea, malnutrition, weight loss, and possibly life-threatening situations with acute small bowel obstruction or intestinal gangrene [102]. Notably, a key issue in appropriate surgical treatment of patients with small bowel NEN’s is not necessarily the resection of the primary tumour itself [102]. Surgery is to focus on preserving bowel function whilst selectively resecting mesenteric lymph node involvement, via cautious dissection around the superior mesenteric vessels whilst preserving vascular supply, which can be technically challenging [102]. Further, multifocal primary small bowel tumours are common, reported in 25–44% of cases [101]. Many of these multifocal NENs are tiny (<1 cm) and can only be identified intraoperatively. Therefore, careful palpation of the entire small bowel is deemed crucial to identify small NENs and multifocal disease.

Resection of the small bowel primary(ies) together with dissection of superior mesenteric root lymph nodes and peri-mesenteric involvement significantly improve outcomes for the patient [103], with reported 5- and 10-year survival rates of 100% in stage I and stage II disease, and > 95% and > 80% in stage III disease [102]. In a recent multi-centre study including 229 patients with stage I-III small intestinal NENs, we reported recurrence free survival rates after 5 and 10 years following locoregional resective surgery with curative intent of 66.6% and 49.3%, respectively [104]. However, prophylactic locoregional resection in asymptomatic stage IV small bowel NEN is associated with more re-operations and longer hospital stay with no clear survival benefit [105].

Peritoneal metastases in small bowel NENs are common, noted in some 20% of the patients [106]. Synchronous resection of mesenteric/peritoneal lymph node metastases should be performed as completely as technically safely possible to reduce the risk of future ischaemic bowel complications related to intra-abdominal fibrosis [101, 102, 107]. Leaving behind principally resectable peritoneal metastatic tissue can have grave consequences for the patient at later stages, given that desmoplastic mesenteric lesions do not respond (radiographically) to treatment with PRRT [108], which may need to be considered as a treatment option in case of disease progression. In NETTER-1, the estimated rate of progression-free survival at month 20 was 65.2% in the Lu-PRRT group [107], but at some point, most patients show progressive disease. Thus, starting from some 2–3 years following treatment with PRRT, these patients can present with debilitating symptoms related to vascular encasement and/or bowel obstruction, with often no promising further surgical options, and limited other options, such as stenting, and oncological treatment in case of progressive well-differentiated small bowel NENs (as compared to pancreatic NENs) being moderately effective at best [42], and not recommended by ENETS guidance in non-pancreatic NENs unless in NEN G2 with Ki-67 > 15%, or in case of aggressive biological behaviour (RECIST progression in 3–6 months) [42].

Once liver metastatic involvement is present, allowing serotonin bypassing of hepatic clearance from the portal circulation, patients may experience functioning symptoms of CS [99]. Rarely (∼5% of cases), CS symptoms can happen without metastatic liver involvement, e.g., in patients with retroperitoneal or ovarian metastases [99]. It is commonly accepted that liver debulking, if technically achievable based on the pattern of liver involvement, benefits patients with CS. However, even in the absence of CS, prognostic benefits of liver debulking have been reported [31‒34], which may not be offered to patients not discussed in a MDT with NEN expertise.

Lack of access to a MDT with expertise in the management of NEN may not only increase the risk of delayed diagnosis and under-treatment of a NEN, but also may increase the risk of over-treatment. In our Centre looking after currently ∼1,000 active patients with NEN, a relevant number of patients with low risk (e.g., well-differentiated, low grade, size <10 mm, clear resection margins (R0), absence of serosal perforation), resected appendix NEN were identified retrospectively, following completion hemicolectomy performed elsewhere, whereas the appropriate advice would have been that the patient was already cured following the appendectomy. Based on two large multi-centre studies [109, 110] and additional evidence, the guidance of the ENETS covering management of appendix NEN was recently updated [5], suggesting that the most robust criterion favouring right hemicolectomy to avoid metastatic potential appears to be tumour size >2 cm. Additional criteria include incomplete tumour resection and high grade appendix NEN, although without a specific Ki-67 cut-off being defined, likely related to the rarity of high-grade appendix NEN in the published studies [5]. Cost of right hemicolectomy comes at GBP 15,580/EUR 18,540/USD 19,475, with postoperative mortality rates in a recent Dutch retrospective population-based study in 29,274 patients with a diagnosis of colon cancer 2012–2020 still being reported as 3.5% in the most recent subgroup (2018–2020) [111], and postoperative including non-surgical complications noted in as many as 28.8% and 17.6% of the cases, respectively [111]. Although complication rates of prophylactic hemicolectomy in often young and otherwise healthy people with the incidental finding of an appendix NEN might be lower, exposing patients to any unnecessary major procedures must be avoided. People with incidental findings of a low risk, completely resected, small, low-grade appendix NEN can be discharged based on guidelines [5], and limited healthcare resources allocated differently.

Another example of potential over-treatment (due to lack of access to NEN centres) includes resection of incidentally detected, well-differentiated, non-functioning, asymptomatic, small (<20 mm) pancreatic NENs, in the absence of higher risk factors such as higher grade, or dilatation of the pancreatic duct (>3 mm). If located in the pancreatic head, people with a small, low-risk pancreatic NEN may be inappropriately exposed to Whipple’s surgery (pancreaticoduodenectomy), which is one of the most complex and technically challenging surgical procedures [112]. The overall morbidity and mortality in patients undergoing pancreaticoduodenectomy remain high, although it has improved significantly over the decades. Current mortality rates vary from 2% to 10%, with morbidity rates reported as high as 60% [112]. Cost of Whipple’s’ surgery in the UK is currently GBP 38,853/EUR 46,235/USD 48,566. Consequences include a high chance of losing the insulin secretion capacity, resulting in secondary (type 3c diabetes), which can be challenging to control, further affecting the quality of life of the patient and driving up cost. In the interim analysis (n = 500) of ASPEN, the largest prospective investigation to date, with a median follow-up of 25 months [113], a non-operative strategy seems safe as only a negligible fraction of patients had an increase in tumour size and no patient developed distant metastases during follow-up [114]. Still, follow-up was relatively short-term and there is a risk of later distant metastases. Further, histological features of aggressiveness were noted in almost 20% of operated tumours (n = 94 patients had surgical resection; of those 13% had severe complications [defined as those with a Clavien-Dindo grade >III] [114]). Publication of the final analyses in n = 1,000 patients is anticipated in the next future and will help defining personalised management strategies for patients with pancreatic NENs >1 cm, including possibly different strategies in young patients, also weighing the risks and inconveniences of long-term, close monitoring with exposure to scans; and in the presence of measurable growth of the NEN.

Less drastic related to potential severe negative outcomes to the patient, but likely even more relevant in terms of total cost related to the frequency of events are issues related to incomplete (R1) resection of rectum polyps containing, on histopathology, an incidental finding of a small NEN. Problems related to incomplete resection of incidental rectum NENs have been discussed in detail in a recent opinion statement of Frydman and Srirajaskanthan [115]. In brief, in contrast to series reported from Asian countries, data from European multi-centre studies demonstrate high rates of incomplete (R1) resections. A French multi-centre study in 329 patients with rectal NENs reported that in only 18% a NEN was suspected at index procedure, resulting in incomplete resection in 76% [116]. Incomplete resection is associated with poorer outcomes [115], and at a minimum will usually trigger a string of additional investigations that are inconvenient for the patient, and drive-up costs. An additional issue is than not all endoscopy reports comment on the estimated size of a polyp (e.g., stating “small polyp,” but no estimate in mm is given), causing further uncertainties. Apart from grade and resection status, lesion size is a key factor in rectal NENs, defining the risk of lymph node and systemic involvement [8]. Endoscopic resection is the standard approach for lesions <10 mm, and may be considered in lesions 10–20 mm depending on tumour characteristics, and if complete (R0) resection can be achieved [8]. For rectal NEN >20 mm, oncological resection is recommended [11, 13]. Further, although small rectal NENs <10 mm have a low risk of metastases, lymph node metastases in rectal NEN <10 mm have been reported [115, 117, 118]. A study of 132 patients with rectal NENs showed 0% lymph node metastases in tumours ≤6 mm, but 10.3% in tumours sized 7–10 mm; and lymphatic invasion was still noted in 9.7% of rectum NENs <6 mm [117]. Based on this, resection should be considered in all cases. Following receipt of a R1 resected rectum NEN (with additional uncertainties about the size if a biopsy only was performed), further investigations are required. These include an MRI scan (NHS England: GBP 193/EUR 230/USD 241); additional 68Ga PET-CT scan (GBP 525/EUR 625/USD 656 for scan plus GBP 1,700/EUR 2,023/USD 2,125 for isotope), and CT staging scan of the thorax, abdomen, and pelvis (GBP 112/EUR 133/USD 140) in case of concerning lymph node involvement on MRI; and repeat sigmoidoscopy (diagnostic sigmoidoscopy GBP 504/EUR 600/USD 630; therapeutic sigmoidoscopy/polypectomy GBP 953/EUR 1,134/USD 1,191), including evaluation of the scar using endoscopic ultrasound (EUS) and biopsy (additional cost GBP 1,200/EUR 1,428/USD 1,500) prior to second resection [115]. Another repeat sigmoidoscopy including EUS and biopsies may be required [116]. In addition, there will be costs of repeated Clinic appointments (follow-up appointment Gastroenterology Clinic GBP 94/EUR 114/USD 118; Endocrinology Clinic GBP 109/EUR 130/USD 136) and repeat measurements of biomarkers, e.g., CgA, all of this avoidable if R0 resection was achieved at the first encounter and the patient discharged based on a low grade, completely resected rectum NEN <10 mm, as per guidelines [8].

Based on studies from Asian countries [115], higher rates of R0 resections appear to be achievable. Training of endoscopists should result in higher rates of recognition of rectum NENs and change to appropriate resection techniques, with significantly higher rates of R0 resections [116, 119]. When initial resection of rectal NENs is incomplete (>R0), there is risk of disease recurrence, and systematic resection of the visible scar after incomplete endoscopic resection is important [120]. In a retrospective analysis of consecutive endoscopic revisions in n = 100 patients, Cheminel and colleagues reported residual NENs in 43% of cases [120]; and salvage endoscopic procedure using endoscopic submucosal dissection, or endoscopic full-thickness resection showed an R0 rate of near 100% [120]. Routine implementation of these measures is expected to significantly reduce costs related to the management of small, well-differentiated rectum NENs.

Long-term (in many cases life-long) follow-up after resection of a lung carcinoid is essential to monitor for recurrence, although the optimal follow-up protocol remains unclear [121]. Historically, therapeutic management and follow-up of lung carcinoid patients was led by thoracic surgeons or a general tumour board. In a recent report summarising results of a survey conducted among 34 experts in NEN-related disciplines, authors found that 70.6% of participants reported the absence of a dedicated MDT for lung carcinoid patients in their respective hospitals [121]. Only recently multidisciplinary management with input from NEN experts was advocated [12]. This led to a campaign of the ENETS to include lung NEN cases in NEN MDTs. Currently, out of 68 ENETS Centres of Excellence (CoE), 18 institutions have been certified as PULM-NET centres, in addition to the GEP-NEN centre certification [122], with the aim to achieve PULM-NET status in all ENETS CoE.

Apart from potentially suboptimal therapeutic decisions, e.g., in complex cases with metastatic disease and/or features of CS, lack of lung carcinoid MDT access can lead to commonly observed issues when providing follow-up in assumed low-risk, potentially cured patients with completely resected (R0), low-stage TCs. Whilst in some countries such as Germany doctors in training are allocated to units typically longer term, a more UK specific issue is the often short-term (3–4 months) rotation of specialist registrars in training, who may not always be familiar with the specifics related to lung carcinoids; and, thus, may apply standard lung cancer protocols when requesting follow-up scans. This can result in over-scanning, e.g., series “standard” contrast chest CT scans, whilst the abdomen may not be included in the CT request; or, if so, the scans are not always requested as dual/phase (arterial/portal venous) or triple-phase scans, limiting their use when commenting on possible lung NEN-related involvement of the liver. In a high-risk patient, this would then require either repeating the CT scan in the correct technique (GBP 103/EUR 123/USD 129), exposing the patient to the inconvenience of having to attend another appointment, and additional radiation; or, if not contraindicated for other reasons (e.g., metal implants, certain pacemakers; claustrophobia), lead to an additional request of a liver MRI scan (GBP 193/EUR 230/USD 241). Vice versa, in a low-risk patient with completely resected TC and absence of higher risk factors, annual CT scans may not be required [11, 123]. More than 40% of lung carcinoids at first diagnosis may be incidentally detectable on a standard chest X-ray [124]. In a previous ENETS consensus paper from 2015, recommendation in resected, low-risk TCs (R0 resected; pN0), after more intense scanning over the first 2 years, was proceeding with annual chest X-rays and a CT every 3 years, long-term [11]; however, this concept was not mentioned in the more recent ESMO guidance [12]. In terms of cost savings, such a concept would be negligible but in terms of reducing cumulative exposure to unnecessary additional radiation, the difference would be dramatic, given the ∼ 470-fold (chest X-ray 0.014 mSV vs. chest CT scan 6.6 mSV) [125] higher radiation dose when proceeding with routine CT scans, although radiation dose in the latest CT scanner models is increasingly lower; and the option exists combining low dose chest CTs with a liver contrast MRI scan in routine follow-up of higher risk cases.

Although, ideally, every patient with a suspected or confirmed diagnosis of a NEN should be managed with input from NEN MDTs, in real life this cannot always be guaranteed, either related to unawareness of such facilities, or related to lack of access due to geographical distance to the next NEN Centre. This can result in various issues related to choosing suboptimal imaging modalities both for the diagnosis and follow-up care for people with NENs, and, as such, may result in delaying the diagnosis of a NEN, or underestimating the disease stage with potentially severe implications when referring to inappropriate treatment pathways, and drive-up costs.

CT is the basic radiological method for primary NEN diagnosis, staging, and monitoring [126]. CT is widely available and provides fast and detailed contrast enhanced imaging of extended body areas, including the chest, abdomen, and pelvis. Because of inadequate morphological criteria (short axis measurements) characterisation of lymph nodes by CT is difficult and bone metastases are often missed [126]. Further, CT is inferior to MRI scan for the assessment of NEN-related masses in the pancreas and the liver. When assessing liver involvement using CT, a so-called triple-phase examination is considered optimal. This involves examination before (non-enhanced, native) and during i.v. contrast-enhancement in the late arterial (portal venous inflow) phase, and in the venous phase. Omitting one or several of these phases in routine follow-up increases the risk of missing newly developed, well-vascularised metastases. Further, certain treatments of NENs including 5-fluorouracil (5-FU), e.g., sometimes used in combination for systemic treatment of pancreatic NENs, or G3 NENs of other primaries [42], may cause fatty liver infiltration by impairing mitochondrial fatty acid oxidation and enhancing reactive oxygen species accumulation in hepatocytes [127], and thus can significantly change appearances on repeat imaging [126]. Liver metastases initially diagnosed during the venous phase may no longer be visible at follow-up, but show up in the late arterial phase, or non-enhanced [126]. Triple-phase CT also reduces the risk of falsely identifying normal tissue as metastatic involvement in a fatty infiltrated liver [126]. To decrease radiation doses, the option of using dual phase CT to produce virtual pre-contrast images can be considered [126]. Further, in centres having 68Ga PET-CT available, a non-contrast CT scan is already included in this study; and one may prioritise abdomen/liver MRI, or, e.g., alternate liver MRI and dual phase CT scans including the chest on routine follow-up, if inclusion of the lung is indeed required on all follow-up scans [128], considering that concomitant progressive disease in already established metastases in the abdomen is evident in the majority of the patients when metachronous lung metastases develop [128].

Most well-differentiated NEN expresses SSRs on their cell membrane. Given the high sensitivity (92%; range 64–100%) and specificity (95%; range 83–100%) for imaging of most types of well-differentiated NENs, ENETS guidance recommends using 68Ga PET in combination with CT (PET-CT) as part of tumour staging, preoperative imaging, and re-staging [2]. In a large study in 728 patients with confirmed or suspected NENs, use of 68Ga PET changed the scheduled treatment plan in 40.9% of patients [129].

SSR scintigraphy (SRS) (“Ocretotide scans”; used in the EU since 1995; or, authorised in Poland since 2004 but approved in 18 other EU countries only since 2023, “Tekrotyd” scans, which come with lower radiation exposure for patients, lower costs, and, based on limited data, appear to show some higher sensitivity as compared to octreotide scans [130‒133]) is used as an alternative if PET-CT (and, more restricting, timely access to gallium-68) is not available. SRS scans are useful to inform about suitability of treatment with SSA or PRRT, but these methods are relevantly inferior compared to 68Ga PET-CT in terms of sensitivity, specificity, and diagnostic accuracy, with particular weaknesses in the detection of small bowel primaries, avid lymph nodes, bone, and peritoneal involvement [2, 129, 133, 134]. Further, octreotide scans come with relevantly higher exposure to radiation (∼12 mSv vs. 2.9 mSv in a 70 kg person), and are less convenient for the patient, requiring 2 patient visits 24-h apart versus one-stop-shop [126]. Relying on SRS rather than 68Ga PET-CT scans prior to major surgical interventions (especially if performed with curative intent) must be discouraged given the increased risk of missing potentially present widespread, small volume disease that would have resulted in opting against surgical treatment [129, 134], and offering different, non-surgical treatment options if 68Ga PET-CT scan was used preoperatively. In that situation, a patient could be severely harmed related to the morbidity and mortality risks inherent to the surgical procedure, without a clear, prognostic benefit of having the surgical procedure performed, e.g., in the presence of bone metastases, which now are assumed to be as common as 12% in overall NEN populations [129]. Regardless of the location of the primary, 68Ga PET-CT findings affect management in a large proportion of people with NENs, with the most frequent and important change being a switch from initially planned surgical treatment to treatment with PRRT, or chemotherapy [129], and additional treatment for bone involvement, if appropriate [135]. Therefore, SSR imaging using 68Ga PET-CT, if accessible, can be considered mandatory for NEN tumour staging.

Another promising option and possible alternative to the current gold standard (68Ga-labelled SSAs) for SRS imaging in patients with NENs could be the use of certain 18F-labelled SSAs [136, 137], with some showing aspects of superiority to 68Ga-labelled SSAs in recent non-inferiority studies [136, 138], combined with logistic advantages related to higher stability [137].

In many of the 27 EU member states, availability of PET-CT scans has relevantly improved in recent years, with a total number of 1,114 PET-CTs available across EU member states as per 2022 [139]. However, density of cover is widely varying, e.g., Italy, France, and Germany have the largest number (54%) of PET-CT facilities, with 222, 216, and 166 up and running scanners as per 2022, whereas 8 of the EU member states still have only 3 or less PET-CT facilities [139]. In the UK, there are currently 70 PET-CT facilities, of those 64 are located in England including 16 in the London area [140]. A modern “standard” static PET scanner comes at a cost of some GBP 2.2/EUR 2.62/USD 2.75 million, whereas most advanced models allowing real-time dynamic imaging of the whole-body within ∼ 2 min (thereby, also reducing exposure to radiation) are in the region of GBP 10/EUR 11.9/USD 12.5 million. PET scanners are typically replaced every 8–10 years. Use of mobile PET scanners, e.g., for use in the theatre to support assessing resection margins in real-time is being investigated [141].

However, for logistical reasons, the majority of PET-CT facilities offer [¹⁸F] Fluorodeoxyglucose (FDG) PET-CT only (e.g., in England, only ∼7 out of 70 PET-CT facilities are currently offering 68Ga PET as per 2024, most of them located in the London area), or use of other radionuclides that are unrelated to NEN diagnostics. FDG is typically produced centrally and then shipped to the respective PET scanning facilities. Although the labelled FDG compound has a relatively short half-life of 109.8 min, it is sufficiently long to allow shipping the compound even to remote PET scanning facilities. Combining 68Ga and FDG-PET-CT imaging has some use in the diagnostic workup of certain NENs, e.g., high grade 2 and grade 3 well-differentiated NENs, NECs, including prognostic information [2]), given that higher grade NENs tend to lose SSR expression, together with increased glucose turnover. However, in the majority of people with lower grade NENs, use of FDG-PET scans does not add relevantly to the diagnostic work-up and is not generally recommended.

A main barrier for offering more widespread diagnostic services using 68Ga PET-CT is related to the short half-life of gallium-68, with the implication that in contrast to FDG, gallium-68 must be produced on site. Previously, application of gallium-68 in nuclear medicine was largely restricted due to cyclotron-dependency. Commercial availability of 68Ga generators since the late 1990s and availability of pharmaceutical grade generators for clinical usage since 2014 have transformed NEN diagnostics. A gallium-68 generator extracts the positron-emitting isotope 68 of gallium from decaying germanium-68, which has a half-life of 271 days and, as such, is suitable for in-hospital production of gallium-68. However, the product of this process, 68-gallium, has a half-life of 68 min only, and can typically be used for only up to 3–4 h post generation, and perhaps slightly longer (∼5–6 h) in case latest model PET scanners are available, related to increased sensitivity in modern devices. Typical scan time using a “standard” modern PET scanner is some 18–20 min but could possibly be slightly extended to make best use of the radionuclide. Combined, depending on both the sensitivity of the PET scan device and scanning time, using gallium-68 might could still yield useful results up to some 6-h post generation (generation to use in 68Ga PET-CT time). However, controlled studies would be needed to investigate whether such an approach will yield comparably accurate results.

Members of our Nuclear Medicine team are currently exploring the use of drones to tackle the issues of short half-lives of radiopharmaceuticals and time-sensitive transport. Using drone services have been trialled for transporting blood samples and implants between two hospital sites ∼14 miles/22 km apart (∼1,200 miles/2,000 km of successful flights being completed within 4 months) and demonstrated relevantly reduced transport times (8 min by drone vs. 45 min by car between the hospital sites), in addition to substantial carbon savings (90.5% greater than electric vans). Current activities include discussions with regulatory bodies such as the Civil Aviation Authority of transporting radioactive material related to legislative requirements, testing packages to transport the radiation in safely and securely, and performing test flights using approved air corridors. Once approved and implemented, strategically located radiopharmacies may use drones for transport of tracers such as gallium-68 to supply nuclear medicine departments/PET scans units within a 50-mile/80-km radius, thereby increasing diagnostic efficacy, reducing waiting and travelling time for patients, and driving down costs. However, in various countries, e.g., in central Europe the nuclear medicine services are often still remote, sometimes over 100 km away, so different solutions are required in these cases. These challenges might be also tackled by using certain 18F-labelled alternatives for SSR imaging (e.g., [136, 138]), considering the high activity yield combined with a longer half-life, enabling centralissed production of 18F-labelled tracers for distribution to distant PET centres [142]. Notably, some of the recently developed novel, 18F-based antagonist radioligands for PET imaging of SSRs showed in vitro and in vivo stability for 2 h without degradation, as compared to a decay of 18.8% of 68Ga-DOTA ligands even after 1 h [137].

In a hospital aiming to set up 68Ga PET scan facilities on site, special production facilities and staff skills are required. The nuclear medicine department will develop a Ga-68 labelled pharmaceutical service utilising existing PET/CT imaging equipment; and radiopharmacy facilities, centred on a germanium-68/gallium-68 (Ge-68/Ga-68) generator, a radiopharmaceutical biosynthesis unit (both required to be set up within a grade “A” aseptic environment), and quality control devices. Licencing procedures, validation of the production system, and quality controls are mandatory regulatory requirements to ensure safe operation. The annual cost of a gallium generator has increased by more than 50% over the last few years and is currently ∼ GBP 100,000/EUR 119,000/USD 125,000. The yield is decreasing over time and in our facilities, the generator is replaced every 9–10 months. Each Ge-68/Ga-68 generator will provide 3 patient injections per day for up to month 6, and 2 injections per day for the remaining period. Kit costs are ∼ GBP 1,000/EUR 1,190/USD 1,250 per vial.

Using 68Ga PET-CT scans is costly (currently in England the cost for a 68Ga PET-CT scan is ∼ GBP 2,225/EUR 2,648/USD 2,781; of those ∼ GBP 1,700/EUR 2,023/USD 2,125 for the isotope, and ∼ GBP 525/EUR 625/USD 656 for the scan), but apart from significantly higher diagnostic accuracy, cost-effectiveness of using 68Ga PET-CT scans has been demonstrated as well [143‒145]. Froehlich and colleagues evaluated the cost-effectiveness of 68Ga PET-CT-CT compared to CT and octreotide scans, using a decision model based on a decision-analytic software [145]. Authors demonstrated the cost-effectiveness of 68Ga PET-CT-CT for detecting NENs, related to more timely detection and reduced therapy costs. Authors concluded that in daily clinical practice, use of 68Ga PET-CT-CT can be considered the most economical approach for diagnostic workup of people with suspected NENs [145]. An Australian study found that the average cost of using 68Ga PET-CT-CT was 4 times less than using octreotide scans [144]. Economic efficiency of using 68Ga PET-CT scan can be related to more timely detection of NENs, facilitating the choice of tailored, optimised treatment options, reduced therapy costs, and better outcomes for the patients [145].

PRRT using 111-Indium-pentetreotide was first given to a NEN patient at the Erasmus Medical Centre in 1994. However, disease remission was rare, in part related to low SSR-2 affinity. To increase efficacy, higher energy beta (β)-particles were introduced (90-Yttrium [90-Y] and 177-Lutetium [177-Lu]). Although direct prospective comparisons between 177-Lu and 90-Y PRRT are not available, median progression-free survival and OS in single arm studies appear higher for 177-Lu vs. 90-Y-based PRRT treatment forms [51], and nephrotoxicity in Lu-based PRRT schemes appears to be relevantly lower [146]. Therefore, most NEN centres and currently running prospective trials are using Lu-based PRRT regimen. For hospitals/NEN centres considering to set up treatment with Lu-PRRT on site, logistical requirements are less complex as compared to setting up diagnostic 68Ga PET scan services. This is related to the much longer half-life of 177-Lu DOTATATE of 6.73 days, which means it will be stable for at least 2–3 days and can be shipped. Providing treatment with Lu-PRRT requires meeting regulatory standards, specialised handling of high energy radionuclides, involvement of radiopharmacy, dedicated clinical facilities, and trained staff [51].

To set up PRRT, services must be centred around NEN multidisciplinary teams. Centres/hospitals providing PRRT should be a recognised molecular radiotherapy centre and be undertaking the broad range of molecular radiotherapy treatments using other radioactive substances. Appropriately trained staff, facilities, and governance arrangements must be in place to ensure the safe delivery of PRRT. Further, trained staff such as the medical physics expert and in some NHS Trusts the Administration of Radioactive Substances Advisory Committee (ARSAC) license holder are required to be on site at the time of administration of this treatment, for clinical advice and to ensure safety management (e.g., in case of spillages) is being handled appropriately. Whilst 177-Lutetium may be delivered under the supervision of nuclear medicine or oncology services, there must be clear integration between departments, e.g., oncology, endocrinology, gastroenterology, and nuclear medicine, to ensure appropriate pathway management.

Exact tariffs hospitals are charging for providing PRRT services vary, but (in the NHS) are somewhere in the region of GBP 2,025/EUR 2,410/USD 2,531 per cycle. Staff costs including 2.5 h Oncology consultant time per cycle, 7.5 h NEN nurse time per cycle, costs to fund the nuclear medicine technologist, reporting time per cycle, and 1 × SPECT-CT per cycle add up to an annual expenditure of currently some GBP 80,000/EUR 95,200/USD 100,000. The cost of the drug is currently some GBP 17,000/EUR 20,230/USD 21/250 per cycle, but usually pass through, so the hospital is not required to pay for the drug. Based on this, for a hospital providing PRRT services, 10 NEN patients (40 administrations of Lu-177; 4 cycles per patient) annually would need to be treated to cover the costs. Therefore, providing PRRT services are unlikely financially sustainable in smaller centres.

Various factors contribute to cost pressures in NEN cancer care (Table 1). Some of the aspects including “true” increasing incidence are difficult to tackle, unless causal factors increasing this trend (e.g., diet? environmental causes?) can be identified; with preventive measures even if implemented unlikely to result in any short- or medium-term changes. Some aspects contributing to increasing prevalence can be seen as a positive, reflecting earlier diagnosis and more potent treatment options, resulting in increased survival, but ultimately increase cost as well. Other aspects adding to cost pressures could be avoided or improved, e.g., using suboptimal diagnostic tools, and some of the aspects must be avoided, e.g., under- or over-treatment of people with NENs, which could at least partially be prevented by increased awareness of NENs being a possible differential diagnosis and seeking NEN MDT discussion as early as possible.

There is only limited evidence whether intense versus less intense follow-up in cancer patients is cost efficient [147], and cost-efficacy of intense follow-up appears to be dependent on the type of cancer [148]. Further, many UK cancer protocols suggest discharge from the specialist cancer services 5 years after treatment, with ongoing follow-up provided, e.g., via the GP surgery. However, this practice is not applicable in most people with a previous diagnosis of a NEN. Except for very low risk findings such as small, low grade, completely resected appendix or rectum NENs in the absence of adverse histological features, life-long follow-up care is recommended [12, 149] and should ideally take place in specialist NEN centres, or at least in hospitals with close collaboration with NEN centres including NEN MDT access. Virtual NEN MDTs might be considered as well [150]. If this is difficult to achieve, some follow-up may happen in local hospitals, with referral back to the NEN centre if conditions are changing. These recommendations are based on observed high recurrence rates often many years following initial treatment [12], and the generally progressive nature of all NENs if not surgically cured. The ENETS recommends follow-up intervals for most well-differentiated NENs between 3 and 6–12 months, mainly based on the location of the primary NEN, and the tumour grade [149]. Recommended follow-up intervals are generally shorter in patients with residual, metastatic, or aggressive disease. Importantly, even in well-differentiated, often slow-growing NENs, disease evolution appears to be common. In a recent multi-centre study, we found ∼ 43% of patients undergoing re-biopsy (mean time to re-biopsy 48.8 months) having increased Ki-67 indices [151]; with even moderate changes of the Ki-67 (an increment in Ki-67 >/ = 1%, not necessarily changing the grade based on the currently proposed tumour grade brackets) being associated with a dramatic, statistically significant drop in OS (32.9 months, vs. 80.5 months in patients with unchanged Ki-67) [151]. A modified histopathological grading system applying Ki-67 cut-offs of 5% and 10% could be superior to predict differences in small bowel NEN patient survival outcomes [152]; similar to previous observations made in people with pancreatic NENs (e.g., [153‒156]). Implementation of modified grading systems may assist changing the treatment strategies and possibly identify patients who would benefit from more intense follow-up, thereby also potentially influencing the optimal allocation of limited healthcare resources. Finally, any increased costs incurred related to inappropriate diagnostics and treatment of people with NENs come with inconvenience, stress, and in the worst cases major negative outcomes. Any preventable issues causing harm to the patient must be avoided at all cost.

MOW would like to thank Dr. Lisa Rowley, Dr. Chen Sheng Low, and Dr. Olu Adesanya, Nuclear Medicine Department in University Hospitals Coventry and Warwickshire (UHCW) NHS Trust, Dr Chander Shekhar, Gastroenterology Department UHCW, and Dr Sharmila Sothi, Oncology Department UHCW, for valuable discussions; and Bal Marva, Head of Costing UHCW, and Phillip Coton, Commercial Analytics Specialist UHCW, for valuable support researching cost related specifics.

M.O.W. has no financial disclosures or conflicts of interests to declare related to this manuscript.

No funding was received for this work.

M.O.W. has researched the related literature and written this manuscript.