Efficacy of FOLFOX in Patients with Aggressive Pancreatic Neuroendocrine Tumors After Prior Capecitabine/Temozolomide

Abstract

Background

5‐fluorouracil, leucovorin, and oxaliplatin (FOLFOX) has activity in pancreatic neuroendocrine tumors (pNETs), but its use is limited, partly because of toxicities. pNETs can often become aggressive over time. We evaluated the efficacy of FOLFOX in patients with aggressive pNETs who had progressed after capecitabine plus temozolomide (cap/tem) among other treatments.

Materials and Methods

This was a retrospective study of all patients with well‐differentiated metastatic pNETs, treated at an academic cancer center between January 2008 and June 2019, who received FOLFOX and had received cap/tem in the past. The primary endpoint was objective response rate.

Results

Thirty‐one patients met eligibility criteria. Twenty‐five received FOLFOX, and six received FOLFOX with bevacizumab. Patients were heavily pretreated, having received a median of three prior lines of systemic therapy prior to FOLFOX (range, 1–8). Three (9.7%) patients had grade [G]1 tumors, 16 (51.6%) had G2, and 6 (19.4%) had G3, and grade was unspecified in 6 (19.4%) patients. Fourteen (45.2%) exhibited a best response of partial radiographic response per RECIST 1.1 criteria, 15 (48.4%) stable disease, and 2 (6.4%) progressive disease; overall response rate was 45.2% and disease control rate was 93.5%. Median progression‐free survival was 6 months (95% confidence interval [CI], 5.0–7.0), and median overall survival was 16 months from onset of study treatment (95% CI, 11.3–20.7) and 67 months from date of diagnosis (95% CI, 49.8–84.2). Median duration of treatment was 3 months, and median duration of response was 2 months. Toxicity profile was consistent with known adverse events associated with this regimen.

Conclusion

FOLFOX is active in aggressive, heavily pretreated pNETs that have progressed on prior cap/tem chemotherapy; response durations are relatively short.

Implications for Practice

FOLFOX chemotherapy has robust activity in patients with rapidly progressive, heavily pretreated pancreatic neuroendocrine tumors (NETs), a setting in which few, if any, other options are likely to be effective. Durations of response, however, are relatively short, and new treatments are urgently needed for patients with aggressive transformation of pancreatic NETs.

Short abstract

There is increasing evidence that well‐differentiated neuroendocrine tumors can become highly aggressive over time. This article assesses FOLFOX activity in a population of heavily pretreated patients with rapidly progressive pancreatic neuroendocrine tumors.

Introduction

Multiple systemic therapies are available for patients with advanced, progressive, well‐differentiated pancreatic neuroendocrine tumors (NETs), including everolimus, sunitinib, capecitabine plus temozolomide, and ¹⁷⁷Lu‐DOTATATE [1, 2, 3, 4, 5]. These treatments have primarily been evaluated early in the course of disease, often after progression on first‐line somatostatin analog. There is increasing recognition that well‐differentiated NETs can become highly aggressive over time. Significant increases in tumor proliferative activity (measured by the mitotic rate of the ki‐67 index) have been observed in patients with rapidly progressive tumors who undergo repeat biopsies [6, 7]. These transformations appear to occur primarily in pancreatic NETs (compared with midgut NETs) and are likely caused by the clonal accumulation of mutations, possibly induced with by DNA‐damaging therapies.

Platinum‐based regimens such as cisplatin plus etoposide are known to be active in aggressive neuroendocrine neoplasms, particularly poorly differentiated neuroendocrine carcinomas (NECs), in which they represent the standard first‐line therapy [8, 9, 10]. Other platinum regimens consisting of oxaliplatin and a fluoropyrimidine (either 5‐fluorouracil [5‐FU] or capecitabine) have been studied in a variety of neuroendocrine neoplasms, both well and poorly differentiated, showing heterogeneous activity depending on tumor type and line of therapy [11, 12, 13]. Relatively high response rates have been observed with 5‐FU plus oxaliplatin (FOLFOX) or capecitabine plus oxaliplatin (CAPEOX) and bevacizumab in well‐differentiated pancreatic NETs. However, the overall use of these regimens remains infrequent, primarily because toxicities such as chronic neuropathy.

Our institutional practice pattern has been to prescribe FOLFOX in select patients with aggressively progressing well‐differentiated pancreatic NETs, typically in patients who received prior capecitabine plus temozolomide chemotherapy among other treatments. This choice of regimen is based on the activity of platinum‐based chemotherapy in aggressive NETs and the fact that FOLFOX is often more tolerable than cisplatin or carboplatin with etoposide, as long as patients are monitored carefully for the development of neuropathy.

We, therefore, conducted a retrospective analysis of all pancreatic NET patients treated at our institution with FOLFOX (with or without bevacizumab). The purpose of this study was to assess whether FOLFOX was particularly active in a select population of heavily pretreated, aggressively progressive pancreatic NETs.

Materials and Methods

This was a retrospective analysis of adult patients with progressive pancreatic NETs treated at the Moffitt Cancer Center with FOLFOX between January 2008 and June 2019. There were no limitations on the number of prior lines of therapy. A key eligibility requirement was that one of the prior lines of therapy consist of capecitabine plus temozolomide. Patients with a history of well‐differentiated NET of any grade (low, intermediate, or high) were eligible; poorly differentiated carcinomas (NECs) were excluded.

All baseline and follow‐up scans were evaluated by RECIST v1.1 by a radiologist (B.M.) blinded to outcomes. Target lesions were selected per RECIST v1.1 criteria, selecting those lesions which were progressing on scans prior to initiation of FOLFOX. The primary endpoint was objective response rate (ORR). Secondary endpoints were progression‐free survival (PFS), overall survival (OS), and toxicity. Other data points included prior lines of therapy, the median time to progression prior to initiation of FOLFOX, number of treatment cycles, doses, changes in symptoms, and reasons for discontinuation.

We conducted a subset analysis to assess outcomes stratified by patients who received bevacizumab along with FOLFOX. Institutional review board approval was obtained and waiver of consent was obtained because of the retrospective nature of this study. Data was analyzed using IBM SPSS version 25. Survival curves were estimated using the Kaplan‐Meier method, and categorical variables were analyzed using logistic regression or categorical response models. A p value set at .05 was used for Pearson correlations and χ² analyses.

Results

Table 1 presents patient demographics and tumor characteristics. We identified 31 patients with metastatic pancreatic NETs who were treated with FOLFOX and who met eligibility criteria. Among these patients, 6 had received bevacizumab along with their chemo and 25 received FOLFOX alone. Median time to progression prior to initiating treatment was 3 months (range, 1–8). Median number of prior therapies was three (range, 1–8). Per eligibility criteria, all patients had received prior capecitabine plus temozolomide. Seventy‐one percent (n = 22) of patients had received prior somatostatin analog, 58% (n = 18) had received prior everolimus, 45% (n = 14) had received prior sunitinib, and 19% (n = 6) had received prior peptide receptor radiotherapy with ¹⁷⁷Lu‐DOTATATE. Four patients underwent biopsy of their aggressively progressing disease, demonstrating Ki‐67% increases in three patients (two with increase from grade 2 to 3) and no change in one patient. One patient underwent a biopsy after completion of FOLFOX, revealing a Ki‐67% increase from 1% to 34%. Three patients were given FOLFOX with the goal of shrinking disease and rendering patient eligible for surgical debulking.

Table 1.

Patient demographics and tumor characteristics

Gender	n (%)
Male	23 (74.2)
Female	8 (25.8)
Age, yr
30–45	6 (19.4)
46–60	10 (32.2)
61–75	14 (45.2)
75+	1 (3.2)
Race
White	25 (80.6)
Black	5 (16.1)
Asian	1 (3.2)
Tumor grade
1	3 (9.7)
2	16 (51.6)
3	6 (19.4)
Unknown	6 (19.4)
kI‐67%
0–2	1 (3.2)
3–10	11 (35.5)
11–20	3 (9.7)
20+	7 (22.6)
Not reported	9 (29.0)

Fourteen patients achieved a partial radiographic response (PR) corresponding to an ORR of 45%. Median duration of response was 2 months. Median PFS was 6 months (95% confidence interval [CI], 5.0–7.0; Fig. 1). Median OS was 16 months from onset of study treatment (95% CI, 11.3–20.7) and 67 months from date of diagnosis (95% CI, 49.8–84.2; Figs. 2 and 3, respectively). Fifteen patients (48%) had stable disease as best response, corresponding to disease‐control rate (PR + stable disease) of 93%. Sixty‐four percent of patients discontinued treatment for progression and 26% for toxicity. One patient completed her treatment course (planned for debulking surgery), and two were lost to follow‐up after initiating treatment. Fourty‐two percent of patients (n = 13) had symptomatic improvement from their treatment.

Figure 1.

Progression‐free survival.

Figure 2.

Overall survival from date of first FOLFOX dose. Abbreviation: FOLFOX, 5‐fluorouracil, leucovorin, and oxaliplatin.

Figure 3.

Overall survival from date of diagnosis.

Grade 3 treatment‐emergent toxicities included peripheral sensory neuropathy (n = 2), neutropenia (n = 1), fatigue (n = 1), hypoglycemia (n = 1), hyperglycemia (n = 1), hepatic encephalopathy (n = 2), and small bowel obstruction (n = 1). Eleven patients in total had mild toxicities, the most common of which were nausea, cold sensitivity, neuropathy, fatigue, and thrombocytopenia.

The ORR, median PFS, and median OS (from time of treatment) of patients who received bevacizumab along with their chemotherapy were 50%, 6 months (95% CI, 1.8–10.2), and 14 months (95% CI, 5.6–22.4), respectively. The corresponding results among patients who received FOLFOX alone were 44%, 6 months (95% CI, 4.9–7.1), and 17 months (12.1–21.9), respectively. Differences between both groups were not statistically significant.

Discussion

Aggressive transformation of pancreatic NETs is a relatively common occurrence, particularly in heavily pretreated patients. Once such a transformation occurs, prognosis is poor and therapies primarily studied in low or intermediate‐grade tumors such as SSAs, sunitinib, and everolimus, are unlikely to be effective in these circumstances. Platinum‐etoposide based regimens have been primarily studied in poorly differentiated NECs, and their efficacy in heavily pretreated pancreatic NETs is unknown.

Our data demonstrate that FOLFOX is quite active in this population, even though all patients studied had received prior cytotoxic chemotherapy with capecitabine plus temozolomide. An ORR of 45% is exceptionally favorable in a heavily pretreated population. The median PFS and OS were much less encouraging, indicating that resistance to treatment develops rapidly.

Patients who require rapid symptom palliation are likely to be chief beneficiaries of FOLFOX in this setting. Likewise, patients who have focal progression of metastases that can be resected may benefit from a short course of neoadjuvant FOLFOX prior to locoregional therapy such as surgical resection or ablation of metastases.

Unlike prior studies of fluoropyrimidines plus oxaliplatin, our analysis focused on a very specific population: patients with pancreatic NET who were experiencing rapid disease progression and received prior capecitabine plus temozolomide. Prior studies have evaluated much more heterogeneous populations. A prospective combined analysis of FOLFOX or CAPEOX plus bevacizumab permitted enrollment of advanced, progressive NETs or NECs regardless of line of treatment, primary site, or differentiation [8]. In the pancreatic NET subset of patients, objective responses were observed in 9 of 28 patients (32%); median PFS was 21 months with FOLFOX plus bevacizumab (Bev) and 15.7 months with CAPEOX plus Bev, potentially reflecting the fact that a relatively small percentage of patients had received prior chemotherapy. Likewise, a large retrospective series of 72 patients treated with FOLFOX at a single institution evaluated patients with gastroenteropancreatic NET and NEC of any primary site, grade, differentiation, and line of therapy [13]. Median PFS in the pancreatic cohort was 9 months; objective responses were not categorized by primary site.

A limitation of our study was the retrospective nature of the analysis and lack of pathological data on most patients at time of progression. Although some institutions routinely rebiopsy patients at time of rapid disease growth, we consider rapid progression to be an a priori sign of increase in proliferative activity and likely of grade transformation (although not necessarily dedifferentiation). Our analysis was not randomized or large enough to accurately determine whether bevacizumab contributes to the efficacy of this regimen.

In conclusion, FOLFOX chemotherapy has robust activity in patients with aggressively progressive, heavily pretreated pancreatic NETs, a setting in which few, if any, other options are likely to be effective. Durations of response, however, are relatively short, and new treatments are urgently needed for patients with aggressive transformation of pancreatic NETs.

Author Contributions

Conception/design: Taymeyah Al‐Toubah, Jonathan Strosberg

Provision of study material or patients: Jonathan Strosberg

Collection and/or assembly of data: Taymeyah Al‐Toubah, Brian Morse, Eleonora Pelle, Jonathan Strosberg

Data analysis and interpretation: Taymeyah Al‐Toubah, Jonathan Strosberg

Manuscript writing: Taymeyah Al‐Toubah, Jonathan Strosberg

Final approval of manuscript: Taymeyah Al‐Toubah, Brian Morse, Eleonora Pelle, Jonathan Strosberg

Disclosures

Jonathan Strosberg: Novartis (C/A), Ipsen, Lexicon (Other‐ speaker's bureau). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board