nab-Sirolimus for Patients With Malignant Perivascular Epithelioid Cell Tumors

PURPOSE Malignant perivascular epithelioid cell tumor (PEComa) is a rare aggressive sarcoma, with no approved treatment. To our knowledge, this phase II, single-arm, registration trial is the first prospective clinical trial in this disease, investigating the safety and efficacy of the mammalian target of rapamycin inhibitor nab-sirolimus (AMPECT, NCT02494570). PATIENTS AND METHODS Patients with malignant PEComa were treated with nab-sirolimus 100 mg/m2 intravenously once weekly for 2 weeks in 3-week cycles. The primary end point was objective response rate evaluated by independent radiology review. Key secondary end points included duration of response, progression-free survival, and safety. A key exploratory end point was tumor biomarker analysis. RESULTS Thirty-four patients were treated (safety evaluable), and 31 were evaluable for efficacy. The overall response rate was 39% (12 of 31; 95% CI, 22 to 58) with one complete and 11 partial responses, 52% (16 of 31) of patients had stable disease, and 10% (3 of 31) had progressive disease. Responses were of rapid onset (67% by week 6) and durable. Median duration of response was not reached after a median follow-up for response of 2.5 years, with 7 of 12 responders with treatment ongoing (range, 5.6-47.2+ months). Twenty-five of 31 patients had tumor mutation profiling: 8 of 9 (89%) patients with a TSC2 mutation achieved a confirmed response versus 2 of 16 (13%) without TSC2 mutation (P < .001). The median progression-free survival was 10.6 months (95% CI, 5.5 months to not reached), and the median overall survival was 40.8 months (95% CI, 22.2 months to not reached). Most treatment-related adverse events were grade 1 or 2 and were manageable for long-term treatment. No grade ≥ 4 treatment-related events occurred. CONCLUSION nab-Sirolimus is active in patients with malignant PEComa. The response rate, durability of response, disease control rate, and safety profile support that nab-sirolimus represents an important new treatment option for this disease.


INTRODUCTION
Perivascular epithelioid cell tumors (PEComas) are mesenchymal neoplasms, composed of histologically and immunohistochemically distinctive epithelioid cells. 1,2 Most PEComas are clinically benign and do not metastasize, but malignant PEComas demonstrate local invasion and/or metastatic spread. Malignant PEComas are classified as an ultrarare soft tissue sarcoma (STS) with an estimated annual incidence of # 1/1,000,000 population, 3 arise most commonly at visceral sites (especially renal, uterine, and gastrointestinal), and have a female predominance.
Malignant PEComa has no approved treatment. Although often treated with cytotoxic chemotherapy regimens, these have shown modest benefit. 4 Some patients with PEComas benefited from treatment with mTORC1 inhibitors (including sirolimus, everolimus, and temsirolimus), as described in case reports and retrospective analyses. [4][5][6][7][8][9] PEComas commonly have loss-of-function mutations in or deletions of TSC1 or TSC2. 10 In addition, PEComas often show evidence of mTORC1 activation with phosphorylation of p70S6K and ribosomal protein S6 by immunohistochemistry (IHC). 11 Aberrant mTORC1 signaling is a key driver of cell proliferation and tumor formation, 12 suggesting that mTORC1 inhibition may be a promising therapeutic approach for PEComas.
The orally available mammalian target of rapamycin (mTOR) inhibitors sirolimus and everolimus have variable absorption, often require therapeutic drug monitoring, and have incomplete target suppression. 13-15 nab-Sirolimus (nanoparticle albumin-bound sirolimus; ABI-009, formerly known as nab-rapamycin) is a novel intravenous (IV) mTOR inhibitor with significantly higher tumor growth inhibition, higher intratumoral drug accumulation, and greater mTOR target [phospho-S6 (pS6)] suppression compared with oral inhibitors, as demonstrated in preclinical models. 16 To our knowledge, this trial (AMPECT) is the first prospective clinical trial in advanced malignant PEComa. Herein, we describe the safety and efficacy of nab-sirolimus in patients with this disease.

Patients
Eligible adults (age $ 18 years) had an Eastern Cooperative Oncology Group performance status score # 1, had not previously received an mTOR inhibitor, and had a histologically confirmed diagnosis of either metastatic or locally advanced (ineligible for surgery) malignant PEComa and measurable disease according to the RECIST, v1.1. Histology was assessed locally in each institution at enrollment and subsequently confirmed by central review at the Dana-Farber/Harvard Cancer Center (J.L.H.). Pathologic confirmation of PEComa was based on characteristic histologic features and evidence of melanocytic (HMB-45 and/or melan A) and smooth muscle (smooth muscle actin and/ or desmin) differentiation by IHC; PEComas with either marked nuclear atypia or pleomorphism combined with mitotic activity were considered malignant. 1 Eligible patients had adequate hematologic, hepatic, and renal function, including an absolute neutrophil count of $ 1.5 3 10 9 /L, platelets of $ 100 3 10 9 /L, a hemoglobin level of $ 9 g/dL, a bilirubin level of # 1.5 3 upper limit of normal, and a serum creatinine level of # 1.5 3 upper limit of normal.

Trial Design and Treatment
In this multicenter, open-label, phase II registration study, patients received nab-sirolimus 100 mg/m 2 IV over 30 minutes once on days 1 and 8 of a 21-day cycle. A maximum of two dose reductions to 75 and 56 mg/m 2 were permitted for toxicity. Treatment continued until disease progression, unacceptable toxicities, or patient preference.

End Points and Statistical Analysis
The primary efficacy end point was overall objective response rate (ORR) by 6 months, evaluated by independent radiology review. The primary analysis was preplanned to occur when the last enrolled patient had been treated for 6 months. The sample size estimation assumed an observed 30% ORR and a sample size of 30 patients, which would exclude values , 15% for the lower bound of the 95% CI. Secondary end points included duration of response (DOR), progression-free survival (PFS), PFS at 6 months (PFS6), overall survival (OS), and safety. A key exploratory end point evaluated the association of tumor mutational and biomarker analyses with clinical response.
The DOR, PFS, and OS reported here are based on an additional 1.5-year follow-up after the primary analysis date.

Assessments
All patients receiving at least one dose of nab-sirolimus were evaluable for toxicity. All adverse events (AEs) were collected from the time the patient signed informed consent until 28 days after the last dose of nab-sirolimus. AEs were graded by National Cancer Institute Common Terminology Criteria for Adverse Events v4.03 and were coded using the Medical Dictionary for Regulatory Activities.
Patients were evaluated by contrast-enhanced computed tomography or magnetic resonance imaging if computed tomography was contraindicated every 6 weeks for the first year and then every 12 weeks until disease progression. Patients evaluable for efficacy had $ 1 dose of nab-sirolimus and centrally confirmed PEComa.
Tumor response was evaluated by RECIST v1.1 by investigators and independent review by two radiologists and an adjudicator, who were unaware of the investigators' assessment.
Patients were followed for survival every 12 weeks until death, loss to follow-up, or study closure.

Biomarker Study Methodology
Targeted exome next-generation sequencing using the OncoPanel test (Center for Advanced Molecular Diagnostics, Brigham, and Women's Hospital, Boston, MA) 17 was performed to assess mutations, copy number changes, and translocation events in approximately 500 genes. In addition, pS6, phospho-4EBP1 (p4EBP1), SPARC, %Ki67, and percent of cleaved caspase 3 were assessed by IHC, and TFE3 translocation by fluorescence in situ hybridization.
Mutational and biomarker analyses were blinded with respect to clinical outcome.
A multivariable analysis was conducted using Pearson correlation to correlate clinical response with altered genes and the following biomarkers: TFE3, phospho-AKT, pS6, p4EBP, SPARC, Ki67, and cleaved caspase 3.

Trial Oversight
The study was approved by the institutional review board of each participating site and was conducted in accordance with the International Conference on Harmonization requirements for Good Clinical Practice and with the ethical principles outlined in the Declaration of Helsinki. All patients provided written informed consent before the initiation of the study. Independent Data Monitoring Committee meetings were convened when 14 and 26 patients had completed $ 1 cycle of therapy and reported no concerns regarding the safety of nab-sirolimus requiring study modification or intervention.

Patient Characteristics and Demographics
A total of 35 patients were enrolled between April 2016 and November 2018 at nine community and academic centers across the United States; 34 of 35 patients were treated with at least one dose of nab-sirolimus, and 31 were evaluable for efficacy (two did not have PEComa on central pathology review; one did not have sufficient tissue for review). The median age was 60 years (range, 27-78), 82% (25 of 31) were female, and 85% (26 of 31) had metastatic disease (Table 1). The most common primary sites of disease were the uterus (24%), pelvis and retroperitoneum (18% each), and lung and kidney (12% each; Table 1). Thirteen percent (4 of 31) of efficacyevaluable patients received prior chemotherapy for advanced disease, including gemcitabine-docetaxel, doxorubicin-ifosfamide, and doxorubicin-olaratumab.

Efficacy
Response evaluation at the primary analysis. The primary analysis was preplanned to occur when the last patient enrolled had been treated for 6 months (May 22, 2019). The confirmed ORR as assessed by independent radiologists was 39% (12 of 31; 95% CI, 22 to 58), all partial responses (PRs). One additional patient had an unconfirmed PR without subsequent confirmatory scans and was assessed as stable disease (SD) $ 12 weeks. SD occurred in 52% of patients (16 of 31, with 10 of 31 SD $ 12 weeks), and 10% of patients had progressive disease (3 of 31; Table 2). The disease control rate (defined as complete response [CR] 1 PR 1 SD $ 12 weeks) was 71% (22 of 31).
Responses were of rapid onset and durable. Sixty-seven percent (8 of 12) of PRs were seen at the first scan after baseline at week 6 (median 1.4 months; 95% CI, 1.3 to 2.8). The median DOR was not reached at the time of the primary analysis, with 9 of 12 responders still on treatment.
1.5-year follow-up after the primary analysis for DOR, PFS, and OS. Reponses and DOR At a 1.5-year follow-up after the primary analysis date (November 23, 2020; ie, 2 year after the last patient initiated treatment), 7 of 12 responders were still receiving treatment and the median DOR had not been reached after a median follow-up for response of 2.5 years (DOR range, 5.6 to 47.21 months, Table 2). Figures 1A-1C show the target tumor responses (waterfall plot) and changes over time (spider plot). Notably, one patient with a primary renal PEComa metastatic to the lungs and lymph nodes had a PR for 10 months that converted to a CR (Table 2), with response ongoing at 21.61 months. One additional patient had a CR in target lesion measurement; however, this patient still had an observed nontarget lesion and an overall assessment of PR.
Responses were independent of the primary site and were observed in tumors originating in the uterus (three), kidney (three), retroperitoneum (two), pelvis (two), liver (one), and small bowel (one; Fig 1D). Notably, 43% (3 of 7) of patients with uterine PEComa had a PR. Responses were also observed in 3 of 4 patients who had previously received chemotherapy, with ongoing DOR ranging from 31.51 to 47.21 months. PFS The median PFS was 10.6 months (95% CI, 5.5 months to not reached; Fig 2A). The PFS rates at 3, 6, 12, and 24 months were 79%, 69%, 47%, and 47%, respectively.  Two of five (40%) patients with locally advanced disease deemed not eligible for surgery at study entry by the investigators had resection of residual PEComa after treatment with nab-sirolimus. Before surgery, one patient had a 7.9% reduction in target lesions after one cycle and the other patient had a 22.3% reduction in target lesions after 10 cycles. Following surgery, both patients remained without disease recurrence at 3 and 3.5 years.
Twenty-four percent (8 of 34) of patients had treatmentrelated serious adverse events (TRSAEs), with 12 events in total. Most TRSAEs were in metabolism and nutrition disorders (33%, 4 of 12 events: four events of grade 3 dehydration occurring in two patients) and GI disorders (25%, 3 of 12: grade 2 abdominal pain, grade 2 diarrhea, and grade 3 enteritis, occurring in one patient each). Other TRSAEs were acute kidney injury, acute coronary syndrome, edema, and pancytopenia, all grade 3, occurring in one patient each.
Dose reductions occurred in 34% (13 of 34) of patients; 11 of 13 patients had one dose reduction, and two patients had two dose reductions. The most common reasons for dose reductions were mucositis and pneumonitis. In all cases, patients whose doses were reduced maintained their best response at the lower dose.  Fig 1D).
Additional confirmed PRs were seen in 20% (1 of 5) of patients with a TSC1 mutation, and in 9% (1 of 11) without a mutation in TSC1 or TSC2. Confirmed PRs were also observed in two of the six patients with tumors with unknown mutational status because of insufficient archival material for analysis.
TSC1 and TSC2 mutations were mutually exclusive. TSC2 mutations were not associated with specific anatomic sites: the primary sites of tumors for the nine patients with TSC2 mutations were retroperitoneum (three), kidney (two), uterus (two), liver (one), and small bowel (one). One of the seven patients with RB1 mutation responded to nab-sirolimus, whereas 9 of 18 patients without RB1 mutation responded (Fisher's exact P 5 .18).
At a 1.5-year follow-up after the primary analysis date, the median DOR had not been reached for TSC2 mutations Progression-Free Survival (%) after a median follow-up for response of 33.7 months (6 of 8 ongoing, range: 6.5 to 47.21 months). TSC2 mutational status was significantly associated with longer PFS and OS (medians not reached, Fig 2B). One patient with a TSC1 mutation and one patient with no TSC1 or TSC2 mutations had the DOR of 5.6 months and 33.41 months, respectively.

DISCUSSION
The AMPECT study met its primary end point with an independently assessed ORR of 39% (95% CI, 22 to 58), exceeding the prespecified lower-bound ORR of 15% below which the regimen would be considered no more active than standard doxorubicin-based chemotherapy. The response rate, durability of response, disease control rate, and toxicity profile support that nab-sirolimus may represent an important new treatment option for patients with advanced malignant PEComa.
It is inherently difficult to perform randomized studies for ultrarare indications; thus, this study was a single-arm phase II study to estimate the response rate. There are no prior prospective trials in patients with malignant PEComa, but STS response rates are low and typically , 20%. [18][19][20] For example, in a recent study that included doxorubicin as control for an unselected patient population of STS, the response rate was 11.9%. 21 Although the small sample size in the present study provided a wide confidence interval for the 39% ORR, given the aggressive natural history of the disease not known to spontaneously regress, the responses are most likely due to antitumor activity of nab-sirolimus.
Therapeutics yielding PFS rates of $ 40% at 3 months and $ 14% at 6 months are considered to be potentially active in advanced STS. 22 A retrospective analysis of cytotoxic chemotherapy in malignant PEComa described a median PFS of 3.2-5.4 months. 4 In the current study, nabsirolimus significantly exceeded these benchmarks with PFS rates at 3 and 6 months of 79% and 70%, respectively. Although patients with tumors with TSC2 mutations had a longer PFS and OS compared with those without TSC2 mutations, we cannot differentiate between this being an effect of nab-sirolimus or potentially reflecting variations in clinical behavior of these genotypes. Overall, these outcomes for a targeted therapy are promising and demonstrate the importance of studying the molecular genetics of each type of sarcoma and other cancer types.
A subset of malignant PEComas are associated with mutations (inactivation or deletions) of TSC1 or TSC2, negative regulators of the mTOR signaling pathway. Retrospective analyses of patients with advanced PEComa treated with mTOR inhibitors sirolimus, temsirolimus, or everolimus showed evidence of antitumor activity, [4][5][6][7][8] suggesting that mTOR inhibitors may improve outcomes compared with chemotherapy and tyrosine kinase inhibitors. nab-Sirolimus is a novel albumin-bound mTOR inhibitor characterized by high tumor uptake and mTOR target suppression and may enhance tumor penetration and accumulation via albumin receptor-mediated (gp60) endothelial transcytosis. 23,24 Albumin has long plasma half-life and broad binding affinity and accumulates in tumors, areas of inflammation, and tissue remodeling, making it an ideal candidate for drug delivery. 23,25 nab-Paclitaxel was the first therapeutic agent using this technology and was shown to have greater drug delivery, safety, and efficacy compared with conventional solvent-based paclitaxel in different various solid tumors. 26,27 Similarly, nab-sirolimus has a distinct pharmacologic profile and pharmacokinetics compared with sirolimus and other mTOR inhibitors, 14,28 although a direct clinical comparison has not yet been performed.
The female prevalence in this study is consistent with known epidemiology of this disease. The relatively high ORR in patients with primary uterine PEComas is consistent with that of the overall study population and contrasts with retrospective reports of lower sensitivity of uterine PEComas to mTOR inhibitors. 4 No new safety signals were observed despite relatively high doses of nab-sirolimus versus those reported with other mTOR inhibitors. 13,14, 29 The high degree of AE resolution and the ability of patients to continue on therapy for . 2 years suggest that nab-sirolimus is manageable for longterm treatment. On the basis of the pharmacokinetic and safety profile of nab-sirolimus in a phase I study, 28 as well as the safety and efficacy in the present study, therapeutic drug monitoring is not required with IV nab-sirolimus.
To our knowledge, AMPECT is the first trial to prospectively evaluate treatment outcome and exploratory correlation with mutational status and biomarkers in malignant PEComa. TSC1-or TSC2-inactivating mutations were seen in 56% (14 of 25) of patients. TSC2 mutations were a strong GLI2  GNA11  GNAS  HELQ  HNF1A  JAK1  JAZF1  KAT6B  KDM5A  KDM6A  KIT  KLF4  KMT2A  KMT2D  MAFB  MAP3K1  MAX  MCL1  MCM8  MECOM  MED12  MEN1  MGA  MSH2  MUTYH  MYCN  NF1  NOTCH3  NPM1  NR0B1  NRG1  NSD1  NTRK1  PARK2  PDGFRB  PIK3C2B  PMS2  PNKP  POLQ  PRKDC  PTCH1  PVRL4  RAD21  RAD50  RBBP8  RECQL4  RET  RIF1  RIT1  RNF43  ROS1  RPTOR  RSPO2  SDHA  SETBP1  SETD2  SF3B1  SLX4  SMARCA4  SMO  SOCS1  SUFU  TCEB1  TET1  TET2  TFE3  TLX3  TMPRSS2  TP53BP1  TRIM37  UIMC1  VHL  WHSC1L1  WRAP53  WRN  XPA   7  12  4  14  22  25  31  3  2  6  29  21  27  16  18  23  13  10  30  1  11  26  24  5  17   7  12  4  14  22  25  31  3  2  6  29  21  27  16  18  23  13  10  30  1  11  26  24  5   positive predictive factor for response to nab-sirolimus in these patients, which is consistent with the role of the TSC protein complex in mTORC1 regulation and warrants further studies for the role of nab-sirolimus in other tumors with TSC2-inactivating mutations. Extensive previous studies have shown that inactivation or loss of either TSC1 or TSC2 has similar effects on the activation of mTORC1. 30 Thus, it remains unclear as to why TSC2 mutations and not TSC1 mutations were associated with response in this study. This analysis is limited by the small number of patients (n 5 5) with TSC1-mutant tumors. Of the four patients with TSC1-mutant tumors that did not develop RECIST responses, three had SD, two for at least 12 weeks, suggesting that nab-sirolimus might have provided some, but incomplete, antitumor activity in this setting. Other possible explanations could be that the nature of the TSC1 mutation led to incomplete inactivation of the TSC1/TSC2 complex or other coincident mutations in these tumors such as mutations in TP53 may contribute to different tumor behavior and outcome (Fig 1D). No significant differences were identified in pharmacokinetic parameters of Cmax or area under the curve in these patients that could account for differences in outcome.
The absence of pS6 staining, which reflects lack of mTORC1 activation, was a strong negative predictor of response to nab-sirolimus. RB1 mutation was also uncommon in responders. Since the total number of patients with mutation and biomarker analysis was relatively small (n 5 25), further study of these biomarkers is warranted.
Although other mTOR inhibitors have been used off-label for treatment of advanced malignant PEComa, to our knowledge, the AMPECT study is the first prospective study in this disease and provides evidence that nab-sirolimus may offer an important benefit in a rare and aggressive sarcoma for which there are no approved therapies.