Circulating tumour cells and tumour biomarkers in functional midgut neuroendocrine tumours

Abstract CALM‐NET was a phase IV exploratory study in the UK that aimed to evaluate if the presence of circulating tumour cells (CTCs) at baseline predicted symptomatic response in patients with midgut neuroendocrine tumours (NETs) treated with lanreotide autogel (LAN). Adults with functional, well/moderately differentiated (Ki‐67 <20%) midgut NETs received LAN 120 mg/28 days for 1 year. CTCs were present in blood if enumeration was >0. Primary endpoint was the clinical value of baseline CTCs to predict symptomatic response (decrease in diarrhoea or flushing of ≥50% frequency, or ≥1 severity level). Other endpoints included progression‐free survival (PFS) and correlations between plasma and urinary biomarkers (including 5‐hydroxyindoleacetic acid [5‐HIAA]). Fifty patients were enrolled; 40 completed the study. Baseline CTCs were present in 22 (45.8%) patients (missing baseline CTC status n = 2). Overall, 87.5% (95% confidence interval [CI]: 73.9; 94.5) of patients had a symptomatic response; a 5.9‐fold higher odds of symptomatic response in patients without CTC versus patients with CTC at baseline was observed, although this was not statistically significant (odds ratio: 0.17 [95% CI: 0.02; 1.65], p = .126). One‐year PFS rate was 66.4% (95% CI: 48.8; 79.2). Biomarker concentrations did not correlate to baseline CTC status. However, there was a strong correlation between plasma and urinary 5‐HIAA (Spearman correlation coefficients ≥0.87 [p < .001], all time points). In conclusion, patients without CTC at baseline may be more likely to achieve a symptomatic response following LAN treatment than patients with CTC. Plasma 5‐HIAA correlated with urinary 5‐HIAA during LAN treatment. ClinicalTrials.gov identifier: NCT02075606.


| INTRODUC TI ON
Neuroendocrine tumours (NETs) are a rare group of tumours originating from neuroendocrine cells occurring most commonly in the gastrointestinal tract (11%-41%), 1 pancreas, pulmonary bronchi and thyroid gland. In England between 2013 and 2014, the estimated annual incidence of neuroendocrine neoplasms (neuroendocrine carcinoma and NETs) was approximately 8 per 100,000. 2 NETs are characterized by the overexpression of somatostatin receptors (SSTRs), of which SSTR2 and SSTR5 are highly expressed in midgut NETs. 3 The management of NETs can be challenging due to the typically late diagnosis and lack of biomarkers to guide treatment, though advances have been made in expanding treatment options. 4,5 Tests for the classic single-analyte biomarkers, such as urinary 5-hydroxyindoleacetic acid (5-HIAA), chromogranin A (CgA) and neurokinin A (NKA), have inadequate specificities and sensitivities for diagnosis and a lack of predictive value, necessitating the development of novel predictive biomarkers. 4,6 The presence of circulating tumour cells (CTCs) has been shown to predict survival outcomes in a number of solid tumour malignancies following treatment, including metastatic breast, colorectal, prostate, and lung cancer. [7][8][9][10][11][12] CTCs have been shown to be detectable in approximately 50% of patients with metastatic NETs, and in these patients, progression-free survival (PFS) and overall survival were shorter than in those without CTCs; CTC number was also found to correlate with tumour burden, grade and CgA. 13,14 Reductions in CTC levels post-treatment have been associated with slower disease progression and improved survival in patients with NETs. 15 There is increasing interest in the use of CTCs to inform treatment decisions and monitor tumours during treatment, as analysis is relatively noninvasive compared to biopsy, and results can be returned quickly. 7 Lanreotide autogel (LAN) is a somatostatin analogue (SSA) that has antiproliferative effects through direct activation of SSTRs on tumour cells. LAN has been shown to improve symptoms and PFS in patients with functional NETs when administered at doses of up to 120 mg every 28 days. [16][17][18][19] LAN prolongs PFS in those with nonfunctional, locally advanced or metastatic enteropancreatic NETs. 20 The aim of the CALM-NET exploratory study was to determine whether pretreatment CTC status can help predict the clinical symptomatic response to LAN in patients with functional midgut NETs; CTCs were also enumerated during the study.

Council for Harmonisation Consolidated Guideline on Good Clinical
Practice, and all patients provided written informed consent.
F I G U R E 1 CALM-NET study design. a For patients not undergoing a washout period. b LAN injection administered once every 28 days. c A washout period is required for patients previously receiving either s.c. octreotide (at least 2 weeks) or one injection of a long-acting SSA (at least 6 weeks) prior to study entry. For patients undergoing washout, information on symptoms at diagnosis and during prior SSA treatment was obtained from the medical notes of patients. CT

| Study population
Patients were eligible for inclusion if they were aged ≥18 years, had a documented diagnosis of functional well-or moderately differentiated (Ki-67 <20%) midgut NETs and had ongoing diarrhoea and/or flushing at enrolment. Eligible patients also had positive SSTR imaging and were clinically suitable for SSA treatment. Patients being treated with a short-acting SSA and those who had received a single injection of a long-acting SSA could be enrolled into the study, provided there was an adequate washout period (≥2 weeks for shortacting subcutaneous octreotide and ≥6 weeks for long-acting SSAs).
Those treated with interferon, chemotherapy, chemoembolization or radionuclide therapy in the previous 3 months were excluded.

| Study treatment and concomitant medication
Patients were treated with LAN every 28 days for up to 1 year (48 weeks, followed by 4 week follow-up period). The dose for the first three injections was 120 mg per injection; thereafter, the dose could be reduced to 90 mg or 60 mg per injection, depending on the clinician's decision, LAN tolerability and the review of the patient's symptoms, as per the summary of product characteristics. 21 Concomitant medications for acute symptomatic episodes (e.g., loperamide, nifuroxazide or octreotide infusion) were permitted during the study for ethical reasons. Information on their use was recorded in patients' case report forms and taken into account in the primary endpoint analysis.

| Study assessments
Patients self-reported their symptoms by telephone via an interactive voice response system (Figure 1), by answering pre-set questions to record symptom frequency ("how many times have you had diarrhoea in the last 24 h?"; "how many times have you had flushing in the last 24 h?") and severity of flushing ("overall, how would you rate your flushing events in the last 24 h? Please give an average rating if more than one flushing event occurred"). Severity of flushing was recorded using a three-point system: mild, moderate or severe. and end of study/early withdrawal ( Figure 1). Enumeration of CTCs was conducted using the CellSearch system (Veridex) as previously described. 22 Patients underwent computed tomography (CT) or magnetic resonance imaging (MRI) scans at screening, week 24, and end of study/early withdrawal ( Figure 1). Investigators reviewed the scan images locally to determine disease progression according to response evaluation criteria in solid tumours 1.1 criteria. NETs were staged using the European Neuroendocrine Tumour Society (ENETS) classification criteria, from stage I to IV.

| Study endpoints
The primary endpoint was the clinical value of enumeration of CTCs at baseline to predict symptomatic response to LAN. A symptomatic response was assessed as a qualitative variable (yes/no) defined as a reduction in the number of self-reported daily diarrhoea and/or flushing episodes by ≥50% and/or a decrease in the most commonly reported category of severity (mild, moderate or severe) of flushing by at least one level, between baseline (defined as the 7 days preceding the first study treatment injection) and last period (defined as

| Statistical analyses
A sample size of 50 patients was chosen for this exploratory study based on practical considerations relating to the number of patients who could be enrolled within the timeframe. It was estimated that this sample size would give 80% statistical power to detect a difference in clinical response rates of approximately 40%; for example, 55% in patients with no measurable CTCs (CTC−) versus 15% in patients with measurable CTCs (CTC+).
For the primary endpoint, clinical symptomatic response was described according to CTC presence (CTC+) at baseline and overall using descriptive qualitative statistics including mean 95% confidence intervals (CIs), and was analysed using univariate logistic regression. The relationship between CTC presence and clinical symptomatic response was also assessed adjusting for the presence of other potentially influential patient characteristics (i.e., age, sex, body mass index, time since NET diagnosis, location of primary tumour, ENETS disease staging, prior medication, and type of concomitant medications potentially impacting symptoms) using univariate logistic regression to identify parameters significant at the 20% level.
The potential impact of concomitant medication on the primary endpoint was evaluated according to whether the medication was likely to have a "full" or "partial" effect on symptoms. Medications defined as "antipropulsive", "somatostatin and analogues" or "other intestinal anti-infectives" were considered to have a potential effect on symptoms. If such medications were started before the start of the last period of symptom collection (Days 11 to 17 after the last injection) and ongoing during the last period of symptom collection, they were considered to have a potential 'full effect' on symptoms.
Medications that were not started before the start of the last period of symptom collection and were not ongoing during the last period of symptom collection were considered to potentially have a 'partial effect' on symptoms.
Progression-free survival was analysed using the Kaplan-Meier method; hazard ratios and corresponding 95% CIs were calculated using a Cox proportional hazard model. All other data were summarized using descriptive qualitative statistics. Correlations between plasma 5-HIAA and other biomarkers were assessed using the Spearman correlation coefficient, and correlation between biomarkers and CTC presence at each time point was assessed using the Wilcoxon test.

| Baseline characteristics
Overall, 50 patients were enrolled and 40 (80.0%) completed the study. Reasons for withdrawal were adverse events (AEs; n = 5; ongoing severe diarrhoea; liver failure; lump at injection site; chest pain, and hypoglycaemia), withdrawal of consent (n = 1) and other reasons (n = 4; investigator's decision; symptom management; increasing symptoms with LAN administered every 3 weeks; death).

| Primary endpoint
Circulating tumour cells data were available for 48 patients who were included in the primary endpoint analysis. Overall, 87.5% (95% CI: 73.9; 94.5) of patients achieved a clinical symptomatic response ( Figure 2). Patients who were CTC− at baseline had a 5.

| HRQoL
At the end of the study, there was no overall trend toward changes in global health status or functional HRQoL measured using

| Safety and tolerability
Overall, most treatment-emergent AEs (TEAEs) were mild or moderate in severity (Table 3), and only two of the TEAEs leading to discontinuation were considered treatment-related by the investigator (bowel obstruction in one patient and a lump at the injection site in another patient). The patient who experienced a bowel obstruction died from multi-organ failure following complications with surgery (ischemia) to treat the bowel obstruction. In addition, there were two further patients with TEAEs resulting in death during the study: one patient died of multi-organ failure following a severe myocardial infarction and pneumonia, and another patient, with a history of ischemic heart disease and carcinoid heart disease, died as a result of heart failure. No cause of death was deemed to be treatment-related by the investigator; however, the concomitant event of bowel obstruction in one patient was considered to be related. An additional death occurred prior to treatment administration as a result of disease progression. The most common treatment-related TEAEs were abdominal pain, diarrhoea and injection site mass (all 10%; Table 4).
Treatment-related serious TEAEs occurred in two patients; bowel obstruction in one, and abdominal pain, nausea and vomiting in another. LAN treatment was discontinued in the patient with bowel obstruction, but continued in the latter patient, whose gastrointestinal events resolved.

| DISCUSS ION
In NET transcripts (NETest) has shown promising results by identifying stable and progressive disease in response to SSA treatment, at an earlier time point than using CgA. 28,29 The tolerability profile observed with LAN was similar to that ob-  Severe and related 3 (6.0) [5] Missing and related 0 TEAEs leading to treatment discontinuation 5 (10.0) [8] TEAEs leading to study withdrawal 4 (8.0) [4] TEAEs leading to death 3 (6.0) a [3] Serious

| CON CLUS IONS
The results from the CALM-NET study demonstrate that a clinical symptomatic response was achieved for the majority of patients with midgut NETs treated with LAN. The safety profile of LAN was consistent with previous studies, with no new safety signals observed.
The symptomatic response did not statistically differ between those patients with baseline CTCs and those without. However, we acknowledge that the study was limited due to its small sample size and short follow up. In addition, a strong correlation between urinary 5-HIAA and plasma 5-HIAA concentrations was observed. This provides evidence of concordance between plasma 5-HIAA and 24-h urinary 5-HIAA measurements in patients with NETs.

ACK N OWLED G EM ENTS
The authors thank all patients involved in the study, as well as their caregivers, care team, investigators and research staff in are Ipsen employees.

DATA AVA I L A B I L I T Y S TAT E M E N T
Where patient data can be anonymized, Ipsen will share all individual participant data that underlie the results reported in this article with qualified researchers who provide a valid research question.
Study documents, such as the study protocol and clinical study report, are not always available. Proposals should be submitted to DataSharing@Ipsen.com and will be assessed by a scientific review board. Data are available beginning 6 months and ending 5 years after publication; after this time, only raw data may be available.