Sunitinib shrinks NET-G3 pancreatic neuroendocrine neoplasms

Abstract

Purpose

The 2017 revised World Health Organization classification of pancreatic neuroendocrine neoplasms classified conventional G3 tumors into well-differentiated (NET-G3) and poorly differentiated (NEC-G3) tumors. However, guidelines for selection of drug therapy were not established in the 2017 revision. This study aimed to elucidate the rates of maximum tumor reduction of sunitinib, progression-free survival, and overall survival in the new classification.

Methods

We investigated the reduction rate over time using computed tomography for 60 patients with unresectable or distant metastatic pancreatic neuroendocrine neoplasms who received 37.5 mg of sunitinib in our department from April 2013 to November 2017.

Results

Of the 60 cases, 42, 10, and 5 were NET-G1/G2, NET-G3, and NEC-G3, respectively. The prognostic factors were analyzed according to clinicopathological factors using the Cox hazard model. The median observation period was 19 months, and the median duration of sunitinib administration was 7 months. The median maximum reduction rate of sunitinib was 18.3%. Tumor response was classified according to the Response Evaluation Criteria in Solid Tumors: 20 cases (33.3%) showed partial response, 29 cases (48.3%) showed stable disease, and 11 cases (18.3%) showed progressive disease. In a multivariate analysis of factors contributing to progression-free survival from the start of sunitinib administration, only histologically poor differentiation was a significant factor (p = 0.010). Progression-free survival and overall survival were significantly better in patients with NET-G3 than that in patients with NEC-G3 (p = 0.005, p = 0.012), while it was not different between those with NET-G3 and those with NET-G1/2.

Conclusion

Our results indicate that sunitinib is as effective for NET-G3 as for NET-G1/2.

Electronic supplementary material

The online version of this article (10.1007/s00432-018-2636-2) contains supplementary material, which is available to authorized users.

Introduction

Although pancreatic neuroendocrine neoplasms (PNENs) are rare, their rate of diagnosis has been increasing primarily owing to advances in medical technology (Yao et al. 2008; Ito et al. 2015). In the 2010 World Health Organization (WHO) classification, they are classified as NET-G1, G2, and NEC according to their malignant potential. Immunohistochemically, NET-G1/G2 and NEC (G3) has less than 20% Ki-67 index and a mitotic index of less than 20/10 high-power field (HPF), while NEC has ≥ 20% Ki-67 index and a mitotic index of ≥ 20 per 10 HPF. Some studies reported that G3 has a subgroup with a profile close to NET-G1/G2 based on immunostaining and gene analysis results (Yachida et al. 2012; Hijioka et al. 2017; Konukiewitz et al. 2017). Thus, G3 PNEN was divided into well-differentiated (NET-G3) and poorly differentiated (NEC-G3) tumor in the 2017 WHO classification (Lloyd et al. 2017).

The 5-year survival rate of patients with PNENs is below 43% (Pape et al. 2004). Systemic drug therapy is the standard treatment for patients with unresectable disease or metastatic tumors, while surgical resection is the optimal treatment for resectable disease. Molecular targeted drugs such as sunitinib and everolimus are recommended for patients with NET-G1/G2 (Pavel et al. 2016). Platinum-based chemotherapies are the first-line treatment for all PNENs that are not NET-G1/G2. However, the NORDIC study reported that the susceptibility to platinum-based chemotherapy varies among categories even in the same G3 grade (Sorbye et al. 2013). In addition, several reports have questioned the usefulness of platinum-based chemotherapy for NET-G3 tumors because it is less susceptible to chemotherapy than NEC-G3 tumors (Hijioka et al. 2015, 2017; Velayoudom-Cephise et al. 2013; Raj et al. 2017). Moreover, it is unclear whether the therapy can improve the overall survival (OS) rate of G3 tumors with classified as NEC.

As such, the European Neuroendocrine Tumor Society guidelines for PNENs recommend other chemotherapy regimens such as streptozocin (STZ)/5-fluorouracil (5-FU) for NET-G3. However, there is no evidence on the efficacy of chemotherapies for NET tumors (Pavel et al. 2016). Moreover, no study has evaluated which chemotherapy regimen is effective for NET-G3. Somatostatin analogue and everolimus failed to improve the OS rate of pancreatic NET-G1/G2 (Rinke et al. 2009; Yao et al. 2016), and there is no report on whether they improve the OS of patients with NET-G3. To date, peptide receptor radionuclide therapy is only known to prolong the prognosis of somatostatin receptor-type 2 positive midgut NET; on the other hand, whether it prolongs the prognosis of advanced pancreatic NET-G3 is yet to be determined. Therefore, a treatment policy for NET-G3 needs to be established. Sunitinib is a multi-targeted tyrosine kinase inhibitor that as acts on the vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF), among other kinases (Abrams et al. 2003; Murray et al. 2003; Mendel et al. 2003). VEGF is a key molecule for angiogenesis in PNENs (Casanovas et al. 2005). Malignant PNENs express stem-cell factor (c-kit) receptor, VEGF receptors, and PDGF receptors (Fjallskog et al. 2003). In a phase III trial comparing sunitinib with placebo for PNENs, sunitinib improved progression-free survival (PFS), OS, and objective response rates (Raymond et al. 2011). Sunitinib is highly sensitive when the Ki-67 index is less than 5%. However, the number of patients with NET-G3 is extremely small and is thus inadequate to draw concrete conclusions on the susceptibility of NET-G3 to sunitinib. Despite limited evidence, NET-G3 is expected to be highly susceptible to sunitinib. In the present study, we aimed to determine whether sunitinib has sufficient tumor-shrinking capability and whether it can improve the prognosis of patients with PNENs, even those with NET-G3.

Methods

Patients and methods

Between 2002 and 2017, 235 patients with PNENs were treated at Tokyo Medical and Dental University. Of the patients who were administered sunitinib for ≥ 1 month, 60 patients who underwent computed tomography (CT) or magnetic resonance imaging (MRI) before and after sunitinib administration were included in this study (Fig. 1). All study procedures were approved by he Human Research Ethics Committee of Tokyo Medical and Dental University (Approval ID: 1080), and written informed consent was obtained from each participant.

Fig. 1.

Study design

In our hospital, the initial daily dosage of sunitinib is 18.75 mg (525 mg/28 days), and it can be increased to as high as 28 mg (787.5 mg/28 days) in the absence of grade 2 or higher toxicity. In cases of toxicity, the dose is reduced to 12.5 mg/day. If grade 2 side effects develop in increased dose, it is reverted to 18.75 mg/day. Patients have never been administered > 50 mg per day, with the maximum dose administered being 37.5 mg per day.

Background characteristics investigated included age, sex, genetic syndromes such as multiple endocrine neoplasia Type I, tumor functionality, tumor location, synchronous lymph node and liver metastases, and surgical procedure. All patients were examined at least every 3–6 months via laboratory tests and CT or MRI with a bolus injection of contrast medium. At least two radiologists diagnosed progression and relapse. PFS was measured from the start of sunitinib administration to the first day of progression or death due to any cause. We conducted a prognostic survey of all patients in December 2017.

Pathological findings analyzed included tumor size and immunohistochemical findings such as Ki-67 index and hormone production. Tumors were classified into four grades in terms of histological differentiation, mitotic index, and Ki-67 proliferative index according to the WHO 2017 Classification of Tumors of the Endocrine Organs (Lloyd et al. 2017). NET-G1 was defined as well differentiation, mitotic count of < 2/10 high-power field (HPF), and < 3% Ki-67 index. On the other hand, NET-G2 was defined as well differentiation, mitotic count of 2–20/10 HPF, and 3–20% Ki-67 index, while NET-G3 was well differentiation, mitotic count of > 20/10 HPF, and > 20% Ki-67 index. NEC-G3 has poor differentiation, mitotic count of > 20/10 HPF, and > 20% Ki-67 index. For tumors with a discrepancy between the Ki-67 index and mitotic count, the higher grade was assigned according to the WHO recommendation. We quantified the Ki-67 proliferative index and mitotic count by counting at least 500 cells in “hot spots.”

Statistical analysis

Statistical comparisons for significance of clinicopathological features were performed using Chi-square test or Fisher’s exact test with a single degree of freedom. Continuous data are expressed as the median (range), and continuous variables were compared among the groups using the Kruskal–Wallis test. Meanwhile, categorical variables were analyzed via Student’s t test. Survival curves were illustrated using the Kaplan–Meier method and compared via log-rank tests. Significant variables were subjected to univariate analysis using a Cox proportional hazards model. A p < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 23.0 (SPSS, Chicago, IL, USA).

Results

This study included 60 PNEN patients treated with sunitinib in a single high-volume center in Japan. The median observation period was 19 months. The patient’s baseline characteristics are shown in Table 1. The median patient age was 54.5 years. Five patients had genetic disease such as multiple endocrine neoplasms type 1 and von Hippel–Lindau disease. Nonfunctional tumors, insulinoma, gastrinoma, and VIPoma were observed in 49, 7, 2, and 2 patients, respectively. The median maximum tumor size was 49 mm. The number of tumors occurring in the pancreatic head and body/tail were even. The median Ki-67 index was 12%, and the median mitotic index was 2/10 HPF. Chromogranin A, synaptophysin, and CD-56 were positively stained in 49 (82%), 52 (87%), and 44 (73%) of the patients, respectively. A total of 52 and 5 patients had well differentiated and poorly differentiated tumors, respectively. Following the 2017 WHO classification, the 2, 40, 10, and 5 patients had NET-G1, NET-G2, NET-G3, and NEC-G3, respectively. The specimens of the three patients were insufficient for classification in the new WHO criteria.

Table 1.

Baseline patient characteristics

Characteristic	n = 60
Clinical
Age, years, median (range)	54.5 (18–83)
Sex, male/female, n	32/28
Prior surgery, n (%)	23 (38.3)
Posterior surgery, n (%)	17 (28.3)
Prior systemic chemotherapy, n (%)	18 (30.0)
Genetic syndrome
MEN type 1	3 (5.0)
VHL	2 (3.3)
Tumor
Maximum tumor size in mm, median (range)	49 (11–150)
Location
Head	30 (50.0)
Body/tail	30 (50.0)
Ki-67 index, %, median (range)	12 (1–90)
Mitosis, per 10 HPF, median (range)	2 (0–72)
Presence of liver metastasis, n (%)	46 (76.6)
Presence of lymph node metastasis, n (%)	27 (45.0)
Chromogranin A positive, n (%)	49 (81.6)
Synaptophysin positive, n (%)	52 (86.6)
CD-56 positive, n (%)	44 (73.3)
Tumor functionality, n (%)
Nonfunctioning	49 (81.7)
Insulinoma	7 (11.7)
Gastrinoma	2 (3.3)
VIPoma	2 (3.3)
Tumor grade (2017 WHO classification, n (%))
G1	2 (3.3)
G2	40 (66.7)
NET-G3	10 (16.7)
NEC-G3	5 (8.3%)
Unknown	3 (5.0)
Sunitinib treatment
Months since diagnosis, month, median (range)	10 (1–20)
Months on treatment, median (range)	7 (1–26)
Adverse event higher than Grade 3, n (%)	5 (8.3)
Dose reduction, n (%)	13 (21.6)

MEN multiple endocrine neoplasia, VHL von Hippel–Lindau, HPF high-power field, VIP vasoactive intestinal peptide, WHO World Health Organization, NET neuroendocrine tumor, NEC neuroendocrine carcinoma

The median duration from diagnosis until the beginning of sunitinib treatment was 10 months. A total of 23 and 17 patients underwent resection before and after sunitinib treatment, respectively. A total of 18 patients were administered other systemic chemotherapy such as platinum-based regimens before sunitinib administration. The median duration of sunitinib treatment was 7 months. Grade 3 or higher adverse events were observed in 5 patients (8.3%), and the sunitinib dose was reduced due to adverse events in 13 patients (21.6%).

The median maximum tumor reduction rate was 18.3% in the 60 patients. Figure 2a shows the best response during the entire administration period as evaluated according to the Response Evaluation Criteria in Solid Tumors. Of the 60 patients, 20 (33.3%), 29 (48.3%), and 11 (18.3%) had partial remission (PR), stable disease (SD), and progressive disease (PD), respectively. One after sunitinib administration, patients with worsening prognosis did not achieve PR (Fig. 2b). Meanwhile, patients who achieved PR status only once did not achieve PD. Except the 17 patients who underwent surgery, six of the 43 patients (14%) achieved PR for more than 1 year.

Fig. 2.

Tumor shrinkage rates after sunitinib administration. a The maximum shrinkage rate of the target lesion in every patient as evaluated according to RECIST. PD progressive disease, SD stable disease, PR partial response. Maximum shrinkage rate (%) = [(Sum of tumor diameters at maximum reduction − baseline diameters)/baseline diameters] × 100 for SD or PR patients. Maximum shrinkage rate (%) = [(Sum of tumor diameters at maximum increase − baseline diameters)/ baseline diameters] × 100 for PD patients. b The time courses of the target tumor size in each patient. Note that the shrinkage rate was evaluated for the first 30 days after the initial administration to determine long-term time course

These results show which factors were correlated with the PFS of patients. Age > 55 years old, Ki-67 index > 20%, mitosis index > 20/10 HPF, negative stain of synaptophysin, and poor differentiation were risk factors for poor prognosis in univariate analysis (Table 2). Tumor functionality, tumor size, and no early enhancement as determined via CT were not risk factors of PFS. Meanwhile, poor differentiation was the only risk factor of PFS in multivariate analysis, with a hazard ratio of 5.3 (p = 0.001).

Table 2.

Univariate and multivariate analyses of progression-free survival of all patients

Clinicopathological factor	Univariate analysis		Multivariate analysis
	HR (95% CI)	P value	HR (95% CI)	P value
Clinical factors
Age, > 55 years	2.3 (1.0–5.3)	0.041		0.224
Sex, male	1.0 (0.7–1.4)	0.992
Lymph node metastasis (+)	1.2 (0.6–2.1)	0.655
Liver metastasis (+)	2.4 (0.7–7.9)	0.160
Functionality, nonfunctioning	2.5 (0.7–8.3)	0.142
Tumor factors
Tumor size, > 50 mm	1.3 (0.6–2.7)	0.589
Location, head	1.0 (0.7–1.5)	0.988
Ki-67 index, > 20%	2.5 (1.1–6.0)	0.033		0.528
Mitosis, > 20 per 10 HPF	8.4 (2.2–32.0)	0.002		0.778
No early enhancement	1.5 (0.7–3.4)	0.292
Chromogranin A, negative	2.0 (0.7–5.2)	0.183
Synaptophysin, negative	3.4 (1.3–9.4)	0.016		0.191
CD-56, negative	1.4 (0.5–4.1)	0.575
Poorly differentiation	7.1 (3.5–22.9)	0.001	5.3 (1.5–18.7)	0.001

HPF high-power field

Table 3 shows the factors correlated with the OS of the 60 patients. In univariate analysis, age > 55 years, Ki-67 index > 20%, mitotic index > 20/10 HPF, and poor differentiation were significant risk factors. Meanwhile, tumor functionality, tumor size, and no early enhancement were not risk factors of OS. Meanwhile, poor differentiation was the only risk factor of OS in multivariate analysis, with a hazard ratio of 24.1 (p < 0.001).

Table 3.

Univariate and multivariate analyses of overall survival of all patients

Clinicopathological factor	Univariate analysis		Multivariate analysis
	HR (95% CI)	P value	HR (95% CI)	P value
Clinical factors
Age, > 55 years	3.0 (1.0–9.0)	0.047		0.360
Sex, male	1.2 (0.7–2.0)	0.511
Positive lymph node metastasis	1.5 (0.5–4.3)	0.465
Positive liver metastasis	1.0 (0.3–3.3)	0.961
Functionality, nonfunctioning	1.3 (0.3–6.0)	0.705
Tumor factors
Tumor size, > 50 mm	1.1 (0.4–3.1)	0.807
Location, head	1.1 (0.6–1.7)	0.929
Ki-67 index, > 20%	4.7 (1.5–15.2)	0.009		0.512
Mitosis, > 20 per 10 HPF	13.8 (2.7–70.7)	0.002		0.352
No early enhancement	1.8 (0.6–5.0)	0.284
Chromogranin A, negative	3.2 (1.0–10.4)	0.057
Synaptophysin, negative	3.2 (0.7–15.4)	0.138
CD-56, negative	2.9 (0.7–11.8)	0.129
Poor differentiation	16.3 (3.5–75.7)	< 0.001	24.1 (4.2–138.7)	< 0.001

HPF high-power field

These results indicate that histological poor differentiation is the most important factor for both PFS and OS in the 60 patients treated with sunitinib. Indeed, the framework of the WHO 2017 classification is extensively determined according to differentiation because it is the most important factor in distinguishing between NET and NEC before the Ki-67 and mitotic indices are calculated. Therefore, we drew a survival curve based on the WHO 2017 classification. Figure 3 shows both the PFS and OS calculated from the start of sunitinib administration for patients with NET-G1/2, NET-G3, and NEC-G3. PFS was not different between NET-G1/G2 and NET-G3 (p = 0.957). By contrast, NEC-G3 patients had significantly poorer PFS than NET-G1/2 (p < 0.001) and NET-G3 (p = 0.005) patients. The 1-year PFS rates for patients with NET-G1/G2, NET-G3, and NEC-G3 were 44, 40, and 0%, respectively. Moreover, NET-G3 patients had better OS than NEC-G3 patients under sunitinib administration (p = 0.012). OS was also not significantly different between NET-G1/2 and NET-G3 patients (p = 0.597). The 3-year OS rates for patients with NET-G1/G2, NET-G3, and NEC-G3 were 81, 60, and 0%, respectively.

Fig. 3.

Progression-free survival (a) and overall survival (b) rates from start of sunitinib treatment in patients with pancreatic neuroendocrine neoplasms classified according to the 2017 WHO criteria. Significant differences were found via log-rank test

We analyzed each case to determine the correlation of the tumor reduction rate by sunitinib with the WHO 2017 classification and PFS (Table 4). ten NET-G3 patients and five NEC-G3 patients were included in the analysis. Of the 15 G3 patients, 5 received platinum-based chemotherapy. However, this did not shrink the tumor by any rate. Among the ten NET-G3 patients who received sunitinib, six (60%), three (30%), and one (10%) achieved PR, SD, and PD, respectively. Meanwhile, of the five NEC-G3 patients who received sunitinib, one (20%) and four (80%) achieved SD and PD, respectively.

Table 4.

Sunitinib and platinum-based chemotherapy for each patient with pancreatic neuroendocrine tumor G3 (NET-G3)/NEC-G3

Age (years)	Sex	Ki-67 index (%)	Differentiation	WHO 2017 Grade	Sunitinib			Platinum-based chemotherapy
					Maximum change (%)	Best overall response	PFS	Regimen	Maximum change (%)	Best overall response	PFS
47	F	70	Well	NET-G3	37	PR	469	CDDP/VP-16	113	SD	51
45	F	26	Well	NET-G3	68	PR	174
45	F	26	Well	NET-G3	67	PR	555
52	F	37	Well	NET-G3	57	PR	147	CDDP/VP-16	108	SD	20
46	M	34	Well	NET-G3	63	PR	193
81	M	21	Well	NET-G3	130	PD	63
62	F	20	Well	NET-G3	61	PR	229
47	F	56	Well	NET-G3	97	SD	189
83	F	90	Well	NET-G3	90	SD	31
80	M	36	Well	NET-G3	118	SD	146	CDDP/VP-16	193	PD	35
28	F	30	Poor	NEC-G3	133	PD	105
71	M	42	Poor	NEC-G3	71	SD	133
60	M	75	Poor	NEC-G3	189	PD	40
61	M	80	Poor	NEC-G3	128	PD	47	CDDP/VP-16	151	PD	75
57	F	90	Poor	NEC-G3	163	PD	55	CDDP/CPT-11	158	PD	84

WHO World Health Organization, PFS progression-free survival, F female, M male, NET neuroendocrine tumor, NEC neuroendocrine carcinoma, PR partial response, SD stable disease, PD progressive disease, CDDP cisplatin, VP-16 etoposide, CPT-11 irinotecan

The maximum shrinkage rate of the five NEC-G3 patients was 108%. A total of two (40%) and three (60%) of the five patients achieved SD and PD, respectively. The PFS of all NET-G3 patients administered with chemotherapy was within 2 months, although it tended to exceed 2 months when sunitinib was administered. In NEC-G3 patients, chemotherapy or sunitinib did not substantially affect PFS, but it was rather long when sunitinib was administered. Moreover, the sunitinib tended to be more effective in NET-G3 patients than in those with NEC-G3 patients in terms of shrinkage rate and PFS time. These results suggest that sunitinib may be superior to platinum-based chemotherapy for NET-G3, although the generalizability of the result is still limited owing to the small cohort in the present study.

Adverse events due to sunitinib were observed in 47 patients (78.3%) (Supplementary Table 1). Adverse events higher than grade 3 occurred in 5 (8.3%) of the 60 patients, namely, thrombocytopenia (n = 2), tumor collapse syndrome (n = 2), nephrotic syndrome (n = 1), and neutropenia (n = 1).

Discussion

This study is the first to examine whether sunitinib is effective for PNET-G3. We showed evidence that sunitinib induces a similar shrinkage rate in NET-G3 tumor and NET-G1/G2 tumor. Poor differentiation was the only factor correlated with the PFS and OS of the patients who received sunitinib, while other factors such as the Ki-67 and mitotic indices did not affect prognoses. The survival curves of OS and PFS showed that patients with poor differentiation had poor survival. These results indicate that NET-G3 may be more responsive to sunitinib than NEC-G3. In particular, among patients with unknown drug response, well-differentiated tumors on biopsy will allow for the selection of the proper drug and thus improve prognosis. To this end, we hypothesized that well differentiation might be an indicator for estimating the maximum rate of tumor shrinkage and the prognoses of the patients administered sunitinib.

The 2017 WHO classification distinguished NEC-G 3 from NET-G 3, adding tumor differentiation to the WHO 2010 classification. The NORDIC study reported that patients with gastrointestinal neuroendocrine carcinomas (GI-NEC) with a Ki-67 of less than 55% had a lower sensitivity to platinum-based chemotherapy than those with Ki-67 over 55%; moreover, the treatment did not prolong OS (Sorbye et al. 2013). However, no phase III study on molecular targeted therapy for pancreatic NEC with Ki-67 of less than 55% has been performed because no such case has been reported by far, and the treatment guidelines are also ambiguous. Furthermore, the appropriate drug for the new NET-G3 classification should be determined in the future.

The present study involved ten patients with NET-G3 (18%) and five NEC-G3 (9%) categorized according to the WHO 2017 classification. The median Ki-67 mitotic indices were 12% and 2/10 HPF, respectively. The minimum and maximum values of Ki-67 index were 1 and 90%, respectively. This study included patients with aggressive tumors with high Ki-67 index. On the other hand, a previous study included relatively stable tumors with low Ki-67 index. In the subgroup analyses (except those with poorly differentiated tumors) in the phase III trial of sunitinib for pancreatic NETs, sunitinib was ineffective in 29 cases with Ki-67 exceeding 5%, while it was effective in 41 cases with Ki-67 less than 5% (Raymond et al. 2011). However, it is not clear which WHO classification was used in this paper; thus, the definitions of both well and poor differentiation were unknown. If the 2004 WHO classification was used, there is a possibility that the “poor differentiation” defined in the paper may be different from that in the 2010 or 2017 WHO classification because poor differentiation is defined only according to mitotic index, and not Ki-67 in the 2004 classification. In a phase II study of sunitinib in Japan that excluded poorly differentiated tumors classified according to the 2004 WHO classification, no data on Ki-67 index or mitotic index were shown. Moreover, only those with no disease progression in the last 12 months before starting sunitinib were targeted in the study. Given that only patients with very slow tumors were recruited, the cohort may have not included patients with Ki-67 index above 20%. The present study showed the maximum shrinkage rate of tumors. The objective response rate was 33.3%. Complete response (CR), PR, SD, and PD were observed in 0, 33.3, 48.3, and 18.3%, respectively. Despite comprising 27% of cases with Ki-67 of 20% or more, the PR rate of the current study was higher than that of the phase III trial, but lower than that of a phase II study in Japan (Ito et al. 2013). Our findings are consistent with those of previous reports that also included aggressive tumors. The CR, PR, SD, and PD rates in a phase III study were 2, 7, 63, and 14%, and the objective response rate was 9.3%. In the Japanese phase II study, CR, PR, SD, and PD were observed in 0, 50, 41.7, and 8.3%, respectively. The objective response rate was 50% because only 12 patients were involved in the study. Moreover, drug response was evaluated according to the tumor reduction rate 1 month after sunitinib treatment (Fig. 2b). This result was similar to that of the Japanese phase II study. In any case, PD can be approximately 10% if the Ki-67 index is less than 20%, although the objective response rate may vary. Considering the prognoses of NET-G1/G2 and NET-G3, the present study indicated that the effect of sunitinib on NET-G3 was equivalent to that in NET-G1/G2 classified according to the 2017 WHO criteria (Fig. 3). In the phase III study, sunitinib substantially improved the prognosis of patients with well-differentiated tumors categorized according to the 2005 WHO classification (Raymond et al. 2011), although it did not demonstrate whether it improved the prognosis of NET-G3 and NEC-G3 categorized according to the 2017 WHO classification. The Japanese phase II study also failed to demonstrate if sunitinib improved the prognoses of NET-G3/NEC-G3 patients.

Next, the susceptibility of NET-G3/NEC-G3 patients who received sunitinib to platinum-based chemotherapy was examined although the number of such patients is small. Platinum-based chemotherapies were administered in three NET-G3 patients. The maximum response to platinum-based chemotherapy was SD for one patient and PD for two, and no tumor shrinkage was observed (Table 4). Meanwhile, the maximum response to sunitinib was PR in two cases and SD in one case. The maximum tumor shrinkage rate from platinum-based chemotherapies was lower than that from sunitinib in all patients. These results raise the possibility that sunitinib has more capability to reduce tumor volume than platinum-based chemotherapies for NET-G3. Several studies reported that the tumor reduction capability of platinum-based chemotherapy is unknown (Sorbye et al. 2013; Velayoudom-Cephise et al. 2013; Hijioka et al. 2015). On the other hand, treatment response to sunitinib for NEC-G3 patients was SD for one patient and PD for four. The two patients who received platinum-based chemotherapies both achieved PD. Although no randomized trial on platinum-based chemotherapies for NEC-G3 has been conducted, the response rate has been reported to be 12–58% (Sorbye et al. 2013; Mani et al. 2008; Iwasa et al. 2010; Yamaguchi et al. 2014). The optimal chemotherapeutic regimen for NEC-G3 is an important problem.

This study illustrated that sunitinib has excellent antitumor effect and low rate of severe adverse events at only 8.3% (supplementary Table 1). The conservative protocol (i.e., low initial dosage, gradual dose increase, and rapid dose reduction before the development of any adverse event development) may reduce the severity of adverse event. Preventing adverse events through adequate dose adjustment and flexible dosing schedule allow for long-term sunitinib treatment (Lee et al. 2018). The lower incidence of adverse events in the present study than that in previous reports may be due to the administration method in our hospital. In phase III trials, adverse events were primarily managed through treatment discontinuation, dose reduction to 25 mg per day, and subsequent dose increase if grade 2 or higher toxicity did not recur. In patients without an objective tumor response and had few treatment-related adverse events during the first 8 weeks, the dose was increased to 50 mg per day (Raymond et al. 2011). In the phase II study, the dose was decreased to 25 mg depending on individual tolerance (Ito et al. 2013). The dose could also be increased to 50 mg/day if minimal clinical response is observed in the first 8 weeks and if the patient tolerates such dose.

In conclusion, sunitinib improved both PFS and OS by reducing the tumor volume in patients with NET-G3 classified according to the 2017 WHO classification. These results suggest that sunitinib may play a crucial role in the treatment of NET-G3 patients.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Funding

This work was supported by Grant-in-Aid for Scientific Research (C) Grant number 15K10046.

Compliance with ethical standards

Conflict of interest

The author Yuki Mizuno declares that he has no conflict of interest. The author Atsushi Kudo declares that he has no conflict of interest. The author Takumi Akashi declares that he has no conflict of interest. The author Keiichi Akahoshi declares that he has no conflict of interest. The author Toshiro Ogura declares that he has no conflict of interest. The author Kosuke Ogawa declares that he has no conflict of interest. The author Hiroaki Ono declares that he has no conflict of interest. The author Yusuke Mitsunori declares that he has no conflict of interest. The author Daisuke Ban declares that he has no conflict of interest. The author Shinji Tanaka declares that he has no conflict of interest. The author Ukihide Tateishi declares that he has no conflict of interest. The author Minoru Tanabe declares that he has no conflict of interest.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The procedures were approved by the Human Research Ethics Committee of Tokyo Medical and Dental University (Approval ID: 1080). This article does not contain any studies with animals performed by any of the authors.

Informed consent

Informed consent was obtained from all individual participants included in the study.