Streptozotocin, 1982–2022: Forty Years from the FDA’s Approval to Treat Pancreatic Neuroendocrine Tumors

On May 7, 1982, the US Food and Drug Administration (FDA) approved the use of streptozotocin (STZ) for the treatment of pancreatic neuroendocrine tumors (panNETs). Thus, this year marks 40 years since that landmark date. This review of STZ to treat panNETs is intended to commemorate this anniversary. Inasmuch as it is a tribute to this milestone, this text has been broached from a historical perspective, punctuating it with a series of events, recalling the dates and the leading figures of the discoveries. Figure 1 depicts a timeline of the most pivotal developments. Nevertheless, this perspective will not keep us from addressing in the last section the latest aspects and taking note of the prospects that lie ahead on the future horizon.

Neuroendocrine tumors (NETs) have traditionally been considered rare tumors and, although they still are, their incidence is on the rise in the USA [1], Europe [2], and Japan [3]. Due to the relatively long survival of patients with metastatic NETs, the prevalence is substantially higher than the incidence. They can develop in different organs. In the digestive system, they originate in the diffuse endocrine system or the endocrine cells of the pancreatic islets. This hallmark differentially defines these tumors with the expression of specific markers, such as chromogranin or synaptophysin, that can be detected in tumor specimens by immunohistochemistry. Furthermore, some of these tumors can secrete biogenic amines or peptide hormones. The release of these bioactive products that cause specific endocrine symptoms associated with the so-called functioning NETs enables their plasma concentrations to be quantified and used to diagnose and monitor response to treatment. The functioning status of NETs is critical for clinical management, as symptom relief is always mandatory in addition to antitumor treatment. A historical classification defines NETs on the basis of their origin in the embryological gut as foregut (up to the upper duodenum, including pancreas), midgut (up to the middle of the transverse colon), and hindgut (distal colorectal section) tumors. The most common advanced gastrointestinal NETs include panNETs originating in the foregut and ileal NETs (formerly denominated ileal carcinoids) arising from the midgut. Although highly variable between studies, panNETs may account for 10–35% of digestive NETs [3].

Already in 1888, Lubarsch [4] reported the existence of digestive endocrine tumors other than adenocarcinomas, while Ransom [5] published in 1890 one of the earliest reports of metastatic small bowel NET, likely associated with carcinoid syndrome. Oberndorfer [6] first coined the term “carcinoid” in 1907, and while the term no longer exists as a diagnostic entity outside of the lungs, it is still in common use and lives on as the carcinoid syndrome.

NETs are of epithelial origin and, subsequently, classified as carcinomas. Although the term “carcinoma” is used for all neoplasms in other locations regardless of aggressiveness, in NETs, with the evolution of classifications, it is reserved for poorly differentiated carcinoma. According to the WHO classification, last revision in 2019, the term neuroendocrine neoplasm is the broader term encompassing all tumors, including grade 1, 2, and 3 well-differentiated NETs, and the more aggressive poorly differentiated grade 3 large-cell or small-cell neuroendocrine carcinomas [7]. Current terminology using morphology next to grade-associated NET terminology has replaced older terms such as “carcinoid tumor” or even more obsolete terms, such as “APUDomas” [8].

Given their distinct biology and low incidence, the treatment of these neoplasms poses a permanent challenge. Obviously, the entire NET management, from diagnosis to treatment and follow-up, must be multidisciplinary, involving an experienced pathologist, endocrinologist, surgeon, medical oncologist, radiologist, and nuclear medicine specialist.

The antibiotic STZ (manufacturer’s drug code NSC-85998 and trade name Zanosar®) was discovered by researchers who first detected it in the fermentation broth of Streptomyces achromogenes at the Upjohn Laboratory in Kalamazoo, MI in 1959 [9-11]. In 1967, these same researchers were able to make a full description of the new drug [12]. The international nonproprietary name of the drug, which belongs to the nitrosoureas family, has two options: streptozocin and streptozotocin. In this article, we use the latter, with its acronym STZ, given that it is the most widely used in publications about the drug. STZ is a natural, monofunctional D-glucopyranose derivative of N-Methyl-N Nitrosourea, with the molecular formula C8H15N307 and a molecular weight of 266 [12]. Its chemical structure has two parts; a glucopyranosyl group and a nitrosourea group (Fig. 2). The glucopyranosyl group is linked to its affinity for pancreatic tissue since this part of the molecule fosters uptake by the pancreatic beta insular cells by means of a process involving STZ incorporation into the beta cell through the glucose transporter GLUT2 transporter [13-16]. This attraction by beta cells increases specificity against panNETs, thereby achieving its antitumor effect with less bone marrow toxicity [13]. The nitrosourea group is responsible for its alkylating activity and, in turn, for its chemotherapeutic mechanism of action [17]. Thus, in the laboratories of the Upjohn Company in Kalamazoo, a new drug had been synthesized whose chemical structure had two parts: one that conferred it with affinity for pancreatic beta cells and the other that gave it antitumor activity. Fermentation as an industrial manufacturing process entailed very high costs, so the company developed the chemical synthesis of STZ in 1969 [18]. STZ is an alkylating agent [19] as it can add alkyl groups to the double helix structure of DNA. An alkyl group is a chain of CH2 units, the simplest expression of which has a single carbon formulated as CH3, called the methyl group. When STZ decomposes, it produces reactive methylcarbonium ions that alkylate DNA [20]. Alkylation harms the cell by causing DNA structural damage in the form of strand breaks [21], creating interchain cross-bridges [22], provoking DNA mismatches [23], blocking DNA repair [24] as well as DNA and RNA synthesis, and eventually resulting in cell death by necrosis or apoptosis [25]. In addition, cytogenetic damage by STZ can result in chromosomal aberrations and micronuclei or sister chromatid exchanges [17], all with a remarkable specificity to target pancreatic beta cells [26]. It also induces diabetes in animal models [15]. STZ was finally developed as an antineoplastic drug in vitro and in vivo in animal models [13, 19]. In contrast, it was deemed too toxic for use as an antibiotic; consequently, in 1976, Upjohn filed a Marketing Authorization for the treatment of NETs that was granted by the FDA on May 7, 1982.

The pharmacokinetics of STZ is linear and predictable. It was first described in 1975 when preclinical studies in animals [27-29] and a clinical trial conducted in 15 cancer patients [30] were published. The drug is detectable in plasma for only 3 h; some of its metabolites remain quantifiable for 24 h, all materialized in a three-phase concentration/time curve with an initial half-life of 6 min, 3.5 h in a second phase, and 40 h in the final phase [30]. Excretion was shown to be largely urinary, especially in the first hour, after which only traces appeared in bile [27, 29]. Seventy percent of the STZ dose administered in humans can be detected in urine in the first 4 h [30]. Early studies have already revealed that, following injection of radioactive isotope-labeled forms, STZ accumulates first in the liver and is then cleared by the kidneys [28]. In vitro data did not indicate the involvement of microsomal CYP enzymes in the degradation of STZ, and the drug was not found to inhibit CYP450 enzymes in vitro [31]. These metabolic characteristics reduce the risk of drug-drug interactions.

STZ was found to have diabetogenic effects in some animal studies, given that it selectively destroys insulin-producing beta-cells in pancreatic islets [32]. The effect has subsequently been described in detail in preclinical laboratory models in rats [33], and it has prompted the use of STZ to induce diabetes in laboratory animals for research purposes. However, human pancreatic beta cells are more resistant, and only rarely does STZ cause clinically relevant diabetes in patients with a predisposition to hyperglycemia, although the reasons for this different behavior between the preclinical and clinical settings are unknown. Though uncommon, as sporadic clinical cases of patients with hyperglycemia have been reported, glycemic monitoring is recommended for individuals receiving treatment with STZ [34].

The literature includes multiple case reports and retrospective case series, prospective single-arm clinical trials, and three randomized clinical trials.

Insulinoma was one of the first panNETs treated with STZ [35-42]. By way of a historical report, we will comment on the case reported by Arnould et al. [36] in 1969 in The Lancet. A 55-year-old male patient with hypoglycemic episodes associated with elevated plasma insulin underwent a liver scan with ¹⁹⁸Au and a liver phlebography that indicated perfusion defects. A laparoscopy showed liver metastases, and a biopsy revealed a pancreatic islet cell tumor. Treatment with diazoxide was initiated that only temporarily controlled the hypoglycemic crises; later, when his condition worsened, treatment was started with 1.5 g of intravenous STZ, followed by a dose of 3 g after 3 weeks. The patient’s symptoms improved; he resumed light activities at home; his blood glucose normalized; and his liver size decreased on physical examination. At 4 months, his general status continued to improve. Laparoscopy was performed with liver biopsies that failed to demonstrate any malignant tumor cells.

In several cases of Verner-Morrison syndrome, formerly denominated “pancreatic cholera” or “endocrine cholera” caused by hypersecretion of vasoactive intestinal peptide (VIP) by a pancreatic NET that is currently known as “VIPoma” have also been successfully treated with STZ [43-46], in one instance, also with intra-arterial application [47]. Overall, since the 1970s, there have been reports of all types of functioning panNETs treated with STZ, not only in insulinomas or VIPomas but also in glucagonomas [48] or gastrinomas (Zollinger-Ellison Syndrome) [49-53]. In these cases, STZ treatment successfully controlled the functioning syndrome associated with hormonal hypersecretion and resulted in clinical improvement.

The first retrospective series on 52 patients reported a 64% overall response rate in 1973 with STZ alone [54]. However, most of the retrospective studies published (Table 1) report the results of STZ in combination with either 5 fluorouracil (5FU) or doxorubicin in patients with panNETs. These kinds of studies are useful, as they afford a faithful reflection of daily clinical practice, although the retrospective, non-randomized design implies a low level of evidence and highly variable outcomes.

Two different retrospective studies recently reported objective response rates of 43% [55] and 28% [56], for STZ-5FU, disease control rate about 90%, median overall survival longer than 50 months [55, 56], and long-term survival rates at 5- and 10-year of 38% and 17%, respectively [56]. A retrospective review published in Japan reported a 24% overall response rate for STZ-based combination regimens in patients with pan NETs and highlighted the need to make STZ available to patients, overcoming STZ shortage issues experienced in Japan and other countries [57]. Two small retrospective studies of STZ in combination with doxorubicin failed to support the efficacy of the combination previously reported by Moertel et al. [58]. Indeed, Cheng and Saltz [59] reported the results of 16 patients with panNETs treated with STZ-doxorubicin and found only 1 patient who had a partial response by CT criteria. Similarly, McCollum et al. [60], in a retrospective study of 16 patients, reported only 1 patient with a partial response (Table 1). After these conflicting results were reported, two retrospective studies by Kouvaraki et al. [61] and Delaunoit et al. [62] were performed to evaluate the efficacy of STZ-5FU and STZ-doxorubicin with more modern methods than those used in the randomized trials. Kouvaraki et al. [61] reported the results of the combination of STZ-5FU-doxorubicin in 84 patients with a 39% overall response rate, an 89% disease control rate, and a median overall survival of 37 months. The combination STZ-doxorubicin resulted in a 36% overall response rate, a 60% disease control rate, and a median overall survival of 22.4 months when used as a front-line treatment [62]. A retrospective study published in 2015 reported a 34% overall response rate and a 75% disease control rate in 56 patients treated with either STZ-5FU or STZ-doxorubicin [63].

Table 2 lists prospective, non-randomized, clinical trials of STZ treatment in NETs. The first phase II clinical trial of STZ in panNETs was published by Moertel et al. [58] in 1971. As a historical fact, the title of the article mentioned the international nonproprietary name and the manufacturer’s code: Phase II study of streptozotocin NSC-85998 in the treatment of advanced gastrointestinal cancer [58].

A non-randomized, comparative study was published in 1990 with two cohorts of patients with panNETs: 25 received STZ-doxorubicin and 19 received STZ-5FU [64]. This study is worthy of note because, despite the small sample size and the outdated methods to assess tumor response, it provides a direct, albeit non-randomized, comparison of the two main STZ-containing regimens in combination. The response rate was 58% versus 36%, and the median duration of response was 36 versus 22 months in the STZ-5FU and the STZ-doxorubicin arms, respectively. Tumor response assessment was mainly based on decreased serum tumor markers or, less frequently, objective radiological response. However, an unbalanced mix of baseline patient characteristics, such as 32 with subsequent interferon treatment, or 22 with a diagnosis of type I multiple endocrine neoplasia, may have impacted the outcomes and made interpretation of the results somewhat more complicated.

The administration of antitumor agents in galenic formulations other than the original one is a line of research that has also been used in the field of NET treatment. This strategy aims to achieve the same efficacy results with a more favorable safety profile. This strategy has been proposed with the use of a liposomal formulation of doxorubicin. Thus, Fjällskog et al. [65] reported the results of a study of STZ-liposomal doxorubicin in 30 patients with panNETs in 2008. The overall response rate was 40%, which is remarkable since, in this study, tumor response was assessed using WHO criteria versus clinical criteria used in older trials. The median progression-free survival (PFS) was 13 months, and the median overall survival was 52 months. Two patients suffered from grade 4 toxicity – one stroke and one gastric bleeding. No cardiac events were observed.

Another prospective trial published in 2010 reported the results of treatment with STZ-5FU-cisplatin in 79 patients with advanced NETs, of whom 47 were panNETs. The overall response rate was 38%, and overall 1- and 2-year survival was 77% and 66%, respectively [66]. Results were not analyzed by primary tumor site, so the survival data for panNETs are unknown. However, a non-pancreatic primary tumor was identified as one of the factors associated with worse overall survival, suggesting that overall survival was longer in the panNET subgroup. A small prospective study with 12 patients reported in 1998 a 54% response rate with STZ-5FU-doxorubicin [67].

The single-arm, open-label, French multicenter phase II trial (BETTER-1 study) evaluated the efficacy and safety of the addition of bevacizumab to STZ-5FU in well-differentiated panNETs and was published in 2014 [68]. Thus, being one of the most recent studies, its relevance derives from the design and methods being more current. Therefore, the primary endpoint was PFS assessed by RECIST v1.0 criteria; tumors had to express Ki67 <15%, and a double evaluation of response was planned by the study investigator and by an external review done by experienced independent radiologists. The treatment regimen was STZ-5FU at standard doses plus bevacizumab 7.5 mg/kg intravenous Q3W. Thirty-four patients were included, of whom 79% had nonfunctioning tumors. The overall response rate was 56% as determined by the local investigators and 52% as determined by the external review, while the median PFS was 23.7 and 26.3 months, respectively. Overall survival at 12 and 24 months was 94% and 88%, respectively. These very encouraging results prompted the design and performance of the randomized BETTER-2 study, a currently ongoing trial comparing two common chemotherapy regimens, STZ-5FU and capecitabine-temozolomide, administered with or without bevacizumab in patients with panNETs.

Other old prospective studies with a small sample size enrolled a mix of patients in front-line and previously treated with chemotherapy [69, 70]. These studies are shown in Table 2.

STZ has been evaluated in 3 randomized clinical trials [71-73] (Table 3). The first randomized clinical trial was the trial that Dr. Moertel published in The New England Journal of Medicine in 1980. It randomized 103 patients with locally advanced, unresectable or metastatic panNETs. The 84 evaluable patients were randomized to receive STZ single agent (n = 42) versus STZ-5FU (n = 42). Treatment regimens were STZ 500 mg/m²/day, days 1–5 Q6W in the STZ alone arm, and STZ 500 mg/m²/day, days 1–5 plus 5FU 400 mg/m²/day, days 1–5 Q6W. STZ single agent was considered the standard arm. Of the functioning tumors, the most frequent ones were insulinoma, gastrinoma, and VIPoma; the associated disease was still called “pancreatic cholera” at that time. Overall response rate was significantly higher in the STZ-5FU combination arm versus STZ alone (63 vs. 36%; p < 0.01), as was the rate of complete responses (33 vs. 12%; p = 0.01). This study evaluated response rates by means of a combination of clinical, biochemical, and radiological data. In some patients, tumor response was evaluated by measurement of liver size. Nowadays, using RECIST criteria to assess tumor response, we may find this perplexing and of limited value, but we must take into account that, at that time, between the 1970s and the 1980s, these were high-standard procedures that posed no problem for the publication of the study in The New England Journal of Medicine. The overall survival results showed a trend towards improved median overall survival in the combination arm (26 vs. 16.5 months) and the percentage of patients still alive at 12 months (31 vs. 17%). These overall survival results were not statistically significant, probably due to the small sample size that provided scant statistical power. The most significant adverse effects were vomiting in both arms, 60% moderate to severe in the STZ alone arm, 54% in the combination arm, nephrotoxicity in one-third of patients with no differences between arms, and no patients requiring dialysis. Hematologic toxicity was more frequent in the combination arm, both leukopenia (73 vs. 5%) and thrombopenia (27 vs. 5%). Some patients discontinued treatment because of adverse effects (4 due to vomiting and 3 to nephrotoxicity). Again, the historical perspective should help us understand the differences in how toxicity was reported with qualitative (moderate, severe) rather than the formality of the numerically-scored classifications we use today. Furthermore, today’s co-adjuvant antiemetic drugs and improved hydration regimens have dramatically decreased the most limiting non-hematologic toxicities. At the time, the scientific community deemed it to be a positive trial, establishing the combination of STZ-5FU as the new standard first-line treatment for panNETs that has lasted over 30 years and continues to do so.

These results have been consolidated by a second trial published in 1992, also in The New England Journal of Medicine [72]. The study design called for randomization to three arms: the standard STZ-5FU arm at the same doses as in the 1980 trial, an investigational arm with STZ plus doxorubicin, and another investigational arm with chlorozotocin. Twelve years had passed since the first randomized clinical trial published by Moertel, and tumor response was appraised biochemically and by CT imaging, although physical examination was still regarded as valid if hepatomegaly was >5 cm. A total of 105 evaluable patients were included, and the overall response rate was significantly different (p = 0.005) across the three arms: STZ-doxorubicin (69%, with 14% complete responses), STZ-5FU (45%, with 4% complete responses), and chlorozotocin (30% with 6% complete responses). Median overall survival was significantly longer in the STZ-doxorubicin arm compared to the STZ-5FU arm (26 vs. 17 months; p < 0.004) or the chlorozotocin arm (26 vs. 18 months; p < 0.03). In this trial, STZ-5FU and STZ-doxorubicin were associated with severe vomiting in 40% and 20% of the cases, chronic renal insufficiency in 7% and 4% of the cases, and leukopenia less than 2,000 × 10⁹ cells/L in 25% and 5% of the cases, respectively.

Several conclusions can be drawn from the two randomized studies. First, combining STZ with either doxorubicin or 5-FU is superior to STZ alone. Secondly, the very high response rates observed in these studies, above 60%, have not been reproduced after using the most current and stricter response evaluation criteria such as WHO or RECIST [61, 62]. Finally, the STZ-doxorubicin combination appeared more active than STZ-5FU. However, the latter became more popular and is now widely used. This may be due to the different toxicity profiles (alopecia and total lifetime dose-limiting cardiac toxicity). Furthermore, there was a strong influence of the Nordic school of thought, in particular, the Uppsala group that had previously published the non-randomized comparative study showing improved outcomes with STZ-FU versus STZ-doxorubicin [64] and whose STZ-FU regimen differed from Moertel’s that has been commonly referred to as the “Nordic Scheme” (Table 4).

The third randomized trial was a multicenter UK trial led by Dr. Meyer et al. [73] from the Royal Free Hospital in London. The results were published in the European Journal of Cancer in 2014, more than 20 years after its predecessor, which denotes the magnitude of the challenge. The inclusion criteria differed from the two Moertel trials that exclusively included patients with panNETs. The trial by Meyer et al. [73] studied 86 patients with foregut gastrointestinal NETs (foregut, from the esophagus to the duodenum, including panNETs) or NETs of unknown origin. Patients were randomized to two treatment arms: the control arm (CS) was a modification of the classic Moertel (STZ-5FU) regimen substituting 5FU for capecitabine, and the experimental arm (CSC) consisted of the same as in the control arm with the addition of cisplatin. The drug doses were as follows: STZ 1,000 mg/m² Q3W + capecitabine 625 mg/m² bid for the 21 days of the cycle, without interruption, and the same regimen plus 70 mg/m² of cisplatin Q3W following STZ infusion. Randomized patients received CS (n = 44) or CSC (n = 42). The primary endpoint was the overall response rate as per RECIST v1.0 criteria. The pancreas was the most frequent primary tumor site (in 48% of cases), followed by GI NETs (20%). The overall response rate was 12% in patients in the CS arm and 16% in patients receiving CSC. Median PFS was 10.2 versus 9.7 months in the CS and CSC arms, respectively, and the rate of progression-free patients at 12 months was 47% versus 33%. Median overall survival was 26.7 versus 27.5 months, and the 12-month survival rate was 73% versus 72% in the CS and CSC arms, respectively, in the overall population. The overall response rate and disease control rates were 17% and 86% for patients with pancreatic primaries (n = 41), 6% and 83% for patients with other foregut primaries (n = 17), 15% and 73% for patients with an unknown primary site (n = 18), respectively. These differences did not reach statistical significance. Participants assigned to the CSC arm scored worse for quality of life as measured by the EORTC-QLQC30 questionnaire. Toxicity was significantly higher in the CSC arm. Indeed, toxicity grade 3 or higher was experienced by 44% versus 68% in the CS and CSC arms, respectively. In this study, grade 3–4 adverse events rates were: renal toxicity 2% versus 10% in the CS and CSC arms respectively, vomiting 4% versus 18% (of note CSC includes cisplatin), liver 2% versus 3% and neutropenia 0 versus 8%. Fewer subjects suffered from nephrotoxicity than in the Moertel studies since only 5 patients experienced grade ≥3 nephrotoxicity, four of them in the CSC arm that included cisplatin. Therefore, in summary, the results of the Meyer et al. [73] randomized trial demonstrate that adding cisplatin does not bring increased benefit to the STZ-capecitabine combination (or its STZ-5FU equivalent) and is more toxic. Consequently, the combination of STZ with an intravenous fluoropyrimidine remained the standard of care treatment.

STZ is currently administered in 4 combination regimens, never as a single agent. STZ is combined with 5FU in three of the regimens (Moertel and Uppsala regimens with 5FU bolus and the BETTER-2 regimen with 5FU in continuous infusion, administered as 5FULV instead of a bolus) and with doxorubicin in one of them (Table 4).

Adequate antiemetics and hydration must accompany these STZ regimens. Normal saline (0.9% sodium chloride) 1,000 mL in 2 h before STZ infusion is recommended. Contraindications for normal saline, such as heart failure, are evaluated on an individual basis. Current NCCN guidelines recommend a four-drug antiemetic regimen of olanzapine, NK1 receptor antagonist, 5HT3 receptor antagonist, and dexamethasone [74].

In fact, STZ, which is eliminated by the kidney, can provoke nephrotoxicity due to injuries to the renal tubule. Nephrotoxicity was frequent and sometimes severe in early studies where patients were treated with doses that were no longer recommended [75]. The rates and grades of renal toxicity decreased over time in clinical studies. Preclinical and clinical studies showed dose-dependent proximal tubule dysfunction leading to glycosuria, hypophosphatemia, tubular acidosis, and tubular proteinuria [76]. In the Moertel’s 1992 trial, STZ-5FU and STZ-doxorubicin were associated with chronic renal insufficiency in 7% and 4% of the patients, respectively [72]. Recently, the Streptotox-FFCD 0906 study, published in 2021 [77], investigated the incidence of GFR decrease during STZ treatment in individuals treated intravenously or with hepatic intra-arterial embolization using an ambispective design. Thirty patients were included in the prospective cohort and 108 in the retrospective cohort. The change in glomerular filtration rates prior to and following STZ treatment was analyzed, and the results demonstrated that STZ-induced nephrotoxicity can be prevented by proper prehydration. Indeed, renal toxicity was lower in the prospective cohort than the retrospective cohort: the rate of deterioration of GFR ≥15% was lower (22 vs. 37%), and none of the patients presented with grade 3–4 renal toxicity versus one in the retrospective cohort. In the most contemporary randomized trial with the safety reported more accurately, the Meyer’s trial [73], renal toxicity was 2% versus 10% in the CS and CSC arms, respectively.

STZ has also been administered intra-arterially as part of chemoembolization treatment in patients with NETs of gastroenteropancreatic origin and liver involvement as the sole or main metastatic site [78]. Case series treated with STZ as a single agent [79-81], in combination with 5FU [82], with hepatic radiation [83], or with other embolization procedures such as drug-eluting-microspheres infusion [84], have been reported. Taken together, these studies yielded variable response rates of up to 60% and high efficacy in controlling symptoms of hormonal hypersecretion. Of note, trans-arterial chemoembolization using STZ requires general anesthesia due to painful administration. Due to the lack of well-designed comparative studies, several questions remain unanswered regarding the optimal drug to be used. A retrospective analysis of 67 patients (163 procedures) suggested that STZ use, compared with doxorubicin, was among the factors associated with higher tumor response in multivariate analysis [80]. In clinical practice, few sites use STZ to treat liver metastases in patients with panNETs, while other chemotherapy agents are used for tumors of other origins. TAE or trans-arterial chemoembolization is recommended in highly selected patients with exclusive or predominant hepatic metastatic disease that have previously progressed to systemic treatment or suffer from hormonal hypersecretion [85].

STZ-based chemotherapy is considered an option for the systemic treatment of advanced panNETs in the current international guidelines sharing therapeutic algorithms with somatostatin analogs, targeted agents (sunitinib, everolimus, surufatinib), or peptide receptor radionuclide therapy. Of note, significant differences exist between countries. STZ-based chemotherapy is considered a well-established systemic treatment in the European ESMO [85] and ENETS [86] guidelines. According to the ESMO guidelines, STZ-5FU is recommended for patients with advanced grade 1 or 2 panNET and can be considered upfront in patients with advanced tumors with high tumor burden or with bulky symptomatic tumors. In turn, NCCN guidelines include STZ-based chemotherapy as one of the available options with a level of recommendation lower than the European guidelines [87], pointing out that this option is less popular in the USA than in Europe. There are no published results from head-to-head comparative studies between the approved drugs, so the decision to use one or another is mostly based on indirect comparisons of the studies, respective safety/toxicity profiles, comorbidities, and individual tumor features.

The SEQTOR study is a randomized clinical trial comparing the two-sequence strategy; first everolimus, then STZ-5FU upon progression or the reverse sequence. The primary endpoint is the comparison of PFS duration until the first progression. This trial completed recruitment in October 2020, and its results are eagerly awaited, as it is the first head-to-head trial and will hopefully be able to answer this relevant question.

The BETTER2 study, a randomized phase II clinical trial of two chemotherapy regimens with or without bevacizumab in patients with panNETs, is currently underway. Patients are randomized to temozolomide + capecitabine or STZ-FU, with 5FU administered as a prolonged infusion combined with folinic acid (leucovorin), and then undergo a second randomization to add or not bevacizumab. The results of the study, which began recruiting in 2018, are also avidly anticipated, given that they will afford head-to-head information on possible benefits of the combination strategy of chemotherapy with antiangiogenic drugs. Figure 3 summarizes the design of the SEQTOR and BETTER2 studies.

Although chemotherapy has been used for decades in panNETs, several questions are still pending, mainly on overall strategy. We still do not know the optimal treatment duration and the possible benefits of strategies such as chemotherapy maintenance [88]. Novel therapeutic approaches for STZ, such as neoadjuvant STZ-based chemotherapy [89, 90], association of chemotherapy to peptide receptor radionuclide therapy [91, 92], adjuvant therapy (currently investigated phase II trial, SWOG S2104/NCT05040360), or the value of MGMT status [93] as a biomarker to predict tumor response in panNETs [93], as previously shown for Ki-67 expression [94], are still under clinical investigation. In the meantime, we can encourage all these novel approaches only under clinical trials to explore potentially more effective combinations with illustrative companion biomarkers. On the side of adverse events, the mechanisms related to the development of various toxicities (emesis, myelotoxicity, renal damage) are better known. Above all, current therapeutic resources to prevent adverse events are more refined, which has mitigated the perception of the seriousness of STZ-induced toxicity over time.

In 2022, 40 years after its approval by the FDA, STZ remains one of the well-established standards-of-care options for the first-line treatment of patients with panNETs since this drug has accumulated the highest amount of data in the chemotherapy field. In 2018, European marketing authorization was granted in 11 countries based on a robust bibliographical dossier. However, there are still many questions to be answered to optimize the management of patients with panNETs, and we expect that clinical research will provide valuable data in the near future.

Dr. Jaume Capdevila: Speakers’ bureau, consulting or advisory role: Amgen, Bayer, AAA, Astrazeneca, Eisai, Exelixis, Ipsen, ITM, Lilly, Sanofi, Merck, Novartis, Pfizer. Dr. Michel Ducreux: Consulting or advisory role: Amgen, AstraZeneca, Basilea, Bayer, Daiichi Sankyo, Glaxo Smith Kline, HalioDX, Ipsen, Lilly, Merck Serono, MSD, Pierre Fabre, Rafael, Roche, Servier, Sotio, Zymeworks. Invited speaker: Amgen, AstraZeneca, Bayer, Lilly, Merck Kga, Pierre Fabre, Pfizer, Roche, Servier, Terumo. His wife is the head of the oncology business unit in the French Affiliate of Sandoz. Dr. Enrique Grande: Honoraria for speaker engagements, advisory roles, or funding of continuous medical education from Adacap, AMGEN, Angelini, Astellas, Astra Zeneca, Bayer, Blueprint, Bristol Myers Squibb, Caris Life Sciences, Celgene, Clovis-Oncology, Eisai, Eusa Pharma, Genetracer, Guardant Health, HRA-Pharma, IPSEN, ITM-Radiopharma, Janssen, Lexicon, Lilly, Merck KGaA, MSD, Nanostring Technologies, Natera, Novartis, ONCODNA (Biosequence), Palex, Pharmamar, Pierre Fabre, Pfizer, Roche, Sanofi-Genzyme, Servier, Taiho, Thermo Fisher Scientific, and Zodiac. E.G. has received research grants from Pfizer, Merck, Astra Zeneca, Astellas, and Lexicon Pharmaceuticals. Dr. Thorvardur Halfdanarson: Research Support: Thermo Fisher Scientific; Advanced Accelerator Applications (a Novartis company); Turnstone Biologics; Agios; Consultancy/Advisory Board; Ipsen; TerSera; ScioScientific; Curium; Advanced Accelerator Applications (a Novartis company); Terumo; ITM; Isotopen Technologien Muenchen; Crinetics; Viewpoint Molecular Targeting. Dr. Marianne Pavel: Honoraria for advisory boards from Riemser, AAA, IPSEN, Amgen, Crinetics, Hutchmed, and for presentations from AAA, IPSEN, Novartis, Boehringer-Ingelheim, Lilly, MSD. Dr. Vicente Valentí: No known competing financial interests or personal relationships related to this paper. Dr. Staffan Welin: No known competing financial interests or personal relationships related to this paper. Dr. Ramón Salazar: Employment: Sace Medhealth, S.L. (Owner of scientific consultancy Ltd. (& Onco Knowledge website); Consultant or Advisory Role: VCN-BCN, Agendia, Laboratorios Esteve, Guardant Health, Roche Diagnostics, WntResearch AB, Saga Diagnostics, Ferrer, Pfizer, Novartis, Ipsen, Amgen, Merck, Roche Farma, Lilly, Merck Sharp and Dohme (MSD), and Advanced Accelerator Applications (AAA). Stock Ownership: Sace Medhealth, S.L. Speaking: Pfizer, Amgen, Novartis, Merck, MSD, AstraZeneca, Celgene, and AAA.

ESTEVE SA provided an unrestricted grant to develop the project of writing this review.

Dr. Jaume Capdevila conceived the idea of writing this review; Dr. Jame Capdevila and Dr. Ramón Salazar coordinated the project; Dr. Vicente Valentí wrote the first draft; all the authors discussed and agreed on the structure of the review manuscript, discussed the content of the text in a number of drafting meetings, reviewed and corrected the first draft, approved the final draft, and agreed to send it to publish.