PFS for subjects randomised in the lanreotide group, assessed by central review using Response Evaluation Criteria In Solid Tumours Version 1.1 (RECIST v1.1) criteria every 12 weeks, defined as the time from randomisation to disease progression or death from any causes during either the double-blind phase, or the open-label treatment phase. The distribution of PFS times were estimated using the Kaplan-Meier product limit method.

The ITT population included all randomised subjects. Subjects were analysed as randomised, regardless of the treatment received. One subject randomised to lanreotide should have been censored in the PFS analysis treatment for discontinuation for toxicity or other reasons, however the baseline central radiological assessment was performed prior to the randomisation date, therefore the subject was excluded from the analysis.

PFS was assessed by central review using RECIST v1.1 criteria every 12 weeks, defined as the time from randomisation to disease progression or death from any causes during the double-blind phase. The distribution of PFS times were estimated using the Kaplan-Meier product limit method.

The ITT population included all randomised subjects. Subjects were analysed as randomised, regardless of the treatment received. One subject in the double-blind phase: lanreotide arm should have been censored in the PFS analysis for treatment discontinuation for toxicity or other reasons, however the baseline central radiological assessment was performed prior to the randomisation date, therefore the subject was excluded from the analysis.

PFS was assessed by local review using RECIST v1.1 criteria every 12 weeks, defined as the time from randomisation to disease progression or death from any causes during the double-blind phase. The distribution of PFS times were estimated using the Kaplan-Meier product limit method.

The ITT population included all randomised subjects. Subjects were analysed as randomised, regardless of the treatment received. Two subjects should have been censored in the PFS analysis, however the baseline central radiological assessment was performed prior to the randomisation date, therefore the subjects were excluded from the analysis.

ORR was assessed by central review and local review using RECIST v1.1 criteria every 12 weeks, defined as the percentage of subjects who achieved a best overall response of complete response or partial response in the double-blind phase.

The ITT population included all randomised subjects. Subjects were analysed as randomised, regardless of the treatment received. Two subjects were excluded from the analysis as they had no best response recorded in the raw data.

TTF was defined as the time from randomisation to disease progression using RECIST v1.1, death, consent withdrawn, an adverse event, protocol deviations, lost to follow-up, the appearance of carcinoid syndrome or other hormone related syndrome necessitating the initiation of SSAs (rescue octreotide and/or long-acting release SSA), or initiation of anticancer treatment in the double-blind phase. The distribution of TTF times were estimated using the Kaplan-Meier product limit method.

The ITT population included all randomised subjects. Subjects were analysed as randomised, regardless of the treatment received.

Blood samples were collected to determine plasma CgA. Baseline was defined as the last non-missing measurement collected prior to the first dose of study treatment (lanreotide). The x of upper limit of normal (ULN) was calculated as raw value/ULN.

Double-blind phase: The ITT population included all randomised subjects with non-missing measurements. Subjects were analysed as randomised, regardless of the treatment received. Open-label phase: The open-label ITT population included all ITT subjects with non-missing measurements entering the open-label phase who received at least one injection of lanreotide autogel/depot in the open-label phase.

Measured in subjects with an elevated CgA at baseline (≥2 x ULN). Blood samples were collected to determine plasma CgA. Baseline was defined as the last non-missing measurement collected prior to the first dose of study treatment (lanreotide).

Double-blind phase: The ITT population included all randomised subjects with non-missing measurements and elevated CgA at Baseline. Subjects were analysed as randomised, regardless of the treatment received. Open-label phase: The open-label ITT population included all ITT subjects with non-missing measurements and elevated CgA at Baseline entering the open-label phase who received at least one injection of lanreotide autogel/depot in the open-label phase.

The EORTC QLQ-C30 (V3.0) consisted of 30 questions. The final 2 questions were related to global health status/QoL, with responses requested on a 7-point scale from 1 ('Very poor') to 7 ('Excellent'). The global health status/QoL scale ranges in score from 0 to 100. A high score for the global health status/QoL scale represents a high QoL, thus, an increase in score represents an increase in QoL. 95% Clopper-Pearson confidence intervals were estimated using the exact method for binomial distributions. Baseline was defined as the last non-missing measurement collected prior to the first dose of study treatment (lanreotide).

Double-blind phase: The ITT population included all randomised subjects with non-missing measurements. Subjects were analysed as randomised, regardless of the treatment received. Open-label phase: The open-label ITT population included all ITT subjects with non-missing measurements entering the open-label phase who received at least one injection of lanreotide autogel/depot in the open-label phase.

QoL deterioration was defined by a decrease from baseline in EORTC QLQ-C30 Global Health Status/QoL Score of at least 10 points. The EORTC QLQ-C30 (V3.0) consisted of 30 questions. The final 2 questions were related to global health status/QoL, with responses requested on a 7-point scale from 1 ('Very poor') to 7 ('Excellent'). The global health status/QoL scale ranges in score from 0 to 100. A high score for the global health status/QoL scale represents a high QoL, thus, an increase in score represents an increase in QoL. 95% Clopper-Pearson confidence intervals were estimated using the exact method for binomial distributions. Baseline was defined as the last non-missing measurement collected prior to the first dose of study treatment (lanreotide).

Double-blind phase: The ITT population included all randomised subjects with non-missing measurements. Subjects were analysed as randomised, regardless of the treatment received. Open-label phase: The open-label ITT population included all ITT subjects with non-missing measurements entering the open-label phase who received at least one injection of lanreotide autogel/depot in the open-label phase.

Measured in subjects with an elevated 5-HIAA at baseline (≥2 x ULN). The assessment of urinary 5-HIAA required subjects to collect their urine for the 24 hour period prior to the study visit. Baseline was defined as the last non-missing measurement collected prior to the first dose of study treatment (lanreotide). The x of ULN was calculated as raw value/ULN.

Double-blind phase: The ITT population included all randomised subjects with non-missing measurements and elevated 5-HIAA at baseline. Subjects were analysed as randomised, regardless of the treatment received. Open-label phase: The open-label ITT population included all ITT subjects with non-missing measurements and elevated 5-HIAA at baseline entering the open-label phase who received at least one injection of lanreotide autogel/depot in the open-label phase.