Safety and efficacy of avapritinib in advanced systemic mastocytosis: the phase 1 EXPLORER trial

Advanced systemic mastocytosis (AdvSM) is a rare hematologic neoplasm driven by the KIT D816V mutation and associated with poor survival. This phase 1 study (NCT02561988) evaluated avapritinib (BLU-285), a selective KIT D816V inhibitor, in patients with AdvSM. The primary endpoints were the maximum tolerated dose, recommended phase 2 dose and safety of avapritinib. Secondary endpoints included overall response rate and changes in measures of mast cell burden. Avapritinib was evaluated at doses of 30–400 mg once daily in 86 patients, 69 with centrally confirmed AdvSM. Maximum tolerated dose was not reached, and 200 mg and 300 mg daily were studied in dose-expansion cohorts. The most frequent adverse events observed were periorbital edema (69%), anemia (55%), diarrhea (45%), thrombocytopenia (44%) and nausea (44%). Intracranial bleeding occurred in 13% overall, but in only 1% of patients without severe thrombocytopenia (platelets <50 × 109/l). In 53 response-evaluable patients, the overall response rate was 75%. The complete remission rate was 36%. Avapritinib elicited ≥50% reductions in marrow mast cells and serum tryptase in 92% and 99% of patients, respectively. Avapritinib induced deep and durable responses, including molecular remission of KIT D816V in patients with AdvSM, and was well tolerated at the recommended phase 2 dose of 200 mg daily.

Avapritinib (BLU-285, Blueprint Medicines Corporation) is a new, selective type 1 KIT inhibitor with high potency for KIT D816V 12,[24][25][26] . The US Food and Drug Administration (FDA) approved avapritinib in June 2021 for adult patients with AdvSM, including patients with ASM, SM-AHN and MCL. It was also recently approved in the United States for the treatment of adults with unresectable or metastatic gastrointestinal stromal tumor (GIST) harboring platelet-derived growth factor receptor alpha (PDGFRA) exon 18 mutations at a recommended starting dose of 300 mg once daily (QD), based on dose-escalation and dose-expansion trials that evaluated efficacy and safety in this setting 27,28 . We describe the results of a phase 1 international, multicenter, open-label study (EXPLORER, NCT02561988) evaluating the safety, pharmacokinetics (PK), efficacy and patient-reported outcomes (PROs) of avapritinib in adult patients with AdvSM and other myeloid malignancies.

Results
Patients. Between 10 March 2016 and 18 March 2020, 86 patients (dose escalation, n = 32; dose expansion, n = 54) were enrolled at ten sites in the United States and the United Kingdom (Fig. 1).
Diagnoses were centrally confirmed retrospectively by review of pathology and adjudication of C-findings using World Health Organization (WHO) criteria (Methods). Of note, previously undiagnosed AHN was identified by central pathology in 41% of patients with a local diagnosis of ASM, resulting in reclassification as SM-AHN for analysis. Overall, 69 of the 86 enrolled patients had confirmed AdvSM, while the remaining patients were adjudicated to have diagnoses of indolent SM (n = 14), smoldering SM (n = 2) or chronic myelomonocytic leukemia (n = 1). Modified International Working Group-Myeloproliferative Neoplasm Research and Treatment-European Competence Network on Mastocytosis (mIWG-MRT-ECNM) response outcomes are reported here for patients with a centrally adjudicated diagnosis of AdvSM and at least one evaluable mIWG-MRT-ECNM C-finding Discontinued treatment (n = 13) • Adverse events (n = 7; 5 related) • Withdrew consent (n = 4) • Investigator decision/administrative (n = 2) Discontinued treatment (n = 28) • Disease progression (n = 14) • Adverse events (n = 8; 4 related) • Withdrew consent (n = 2) • Investigator decision/administrative (n = 4) AdvSM safety population (n = 69) Patients with centrally confirmed AdvSM diagnosis who received ≥1 dose of study drug. Included in safety analyses. Response not assessed (n = 16) • Absence of evaluable C-findings per mIWG-MRT-ECNM criteria (unless MCL; n = 12) • (the response-evaluable population), while safety is reported in all treated patients and for those with AdvSM. At baseline, measures of mast cell disease burden (BM mast cell burden, serum tryptase level) were similar in the safety, AdvSM safety and response-evaluable populations ( Table 1). The most common evaluable C-findings per mIWG-MRT-ECNM criteria at baseline were marked splenomegaly (53%), thrombocytopenia (36%), elevated alkaline phosphatase (34%), anemia (32%) and ascites (19%; Supplementary Table 1).
Maximum tolerated dose and recommended phase 2 dose. Part 1 studied doses of 30, 60, 100, 130, 200, 300 and 400 mg QD, with doses of 200 mg and higher achieving exposures consistently above, and doses of 100 mg and 130 mg partly above, the xenograft 90% inhibitory concentration (IC 90 ) for KIT D816V inhibition at steady state (Results and Extended Data Fig. 1). No maximum tolerated dose (MTD) was determined. During part 1, one patient experienced dose-limiting toxicity (DLT) of grade 4 vomiting at 400 mg daily. A recommended phase 2 dose (RP2D) of 300 mg QD was initially selected for use in part 2; however, dose reduction to 200 mg QD was typical, most commonly for cytopenias. Importantly, 200 mg QD had similar exposure and time to response, so a second expansion cohort with a 200 mg QD starting dose was added by protocol amendment. In total, 15 patients initiated treatment at <200 mg QD, 21 patients initiated at 200 mg QD, 43 patients initiated at 300 mg QD and 7 patients initiated at 400 mg QD.
Safety. At data cutoff, 35 of 69 patients with AdvSM (51%) had discontinued treatment: 14 (20%) due to disease progression, 7 (10%) due to related AEs, 6 (9%) due to unrelated AEs, 3 (4%) due to consent withdrawal and 5 (7%) due to investigator decision or administrative causes. The most frequent AEs are presented in Table 2, for the safety and AdvSM safety populations, and treatment-related AEs as determined by the investigator are presented in Extended Data Fig. 2.
In patients with AdvSM, the most frequent hematologic AEs (any grade and grade 3 or 4) were anemia (58% and 36%), thrombocytopenia (51% and 41%) and neutropenia (22% and 19%). Patients with baseline cytopenias were more likely to have cytopenia AEs on study. The risk of ≥grade 3 thrombocytopenia was 20% in the absence of baseline thrombocytopenia, but 70% where thrombocytopenia was present at baseline. Neutropenia of ≥grade 3 occurred in 25% without, and 63% with baseline neutropenia. For ≥grade 3 leukopenia, the risk ratios were 67% versus 19%, with or without any baseline leukopenia, respectively.
Hair color changes, typically grade 1, occurred in 13 patients (19%), and grade 1 skin hypopigmentation was reported in two patients (3%), possibly consistent with inhibition of wild-type KIT. There was one case of anaphylactic reaction (not considered treatment related).
Cognitive effects (memory impairment, cognitive disorder, confusional state and encephalopathy) occurred in 21 patients (30%). These events were mostly grade 1 (20%) or grade 2 (7%; Extended Data Fig. 3). Grade 3 cognitive effects occurred in two patients (3%), each in the setting of concomitant benzodiazepine and/or opioid treatment. Cognitive effects were less frequent with starting doses of ≤200 mg and rarely led to treatment discontinuation (two patients; 3%).
Intracranial bleeding (ICB) was reported in nine patients with AdvSM (13%; Extended Data Fig. 3). Five of the events were asymptomatic (grade 1) and identified by prespecified imaging per protocol, two were grade 2, one was grade 3 and one was grade 5 (which was associated with head trauma). Notably, seven of the events were associated with antecedent severe thrombocytopenia (platelet count <50 × 10 9 /l), and one additional ICB event occurred in the context of rapidly progressive MCL associated with severe thrombocytopenia. Only one (1%) patient experienced ICB in the absence of prior severe thrombocytopenia. Avapritinib dose, platelet count, coagulation studies and platelet transfusions for patients who had ICB are reported in Extended Data Fig. 4. Due to the increased incidence of ICB in association with antecedent severe thrombocytopenia, the protocol was amended to exclude the enrollment of new patients with severe thrombocytopenia (<50 × 10 9 /l), to increase monitoring of the platelet count, and to provide guidance for pausing avapritinib and providing platelet transfusion support for severe thrombocytopenia.
Fifty (72%) patients with AdvSM experienced at least one dose reduction due to AEs (most commonly for cytopenias), with a median time to first dose reduction of 8 weeks (range, 0-109 weeks). The median daily dose was 164 mg QD (range, 30-317 mg) for all patients, which was similar for patients who started at 200 mg QD. Treatment-related AEs associated with treatment discontinuation are shown in Extended Data Fig. 5. There were six deaths due to AEs (acute myeloid leukemia, gastric hemorrhage, ICB (considered treatment related; occurring in the setting of antecedent severe thrombocytopenia and a recent fall with head trauma), cardiac arrest, staphylococcal sepsis and septic shock).
The depth of clinical response was generally correlated with the elimination of measurable KIT D816V allele burden in the BM, with a limit of detection of 0.17% by droplet digital polymerase chain reaction (ddPCR; Supplementary Fig. 2). A post hoc analysis according to best mIWG-MRT-ECNM response showed that complete elimination of measurable KIT D816V VAF occurred in 63% patients with a CR (n = 8), in 50% patients with a CRh (n = 8), in 23% patients with a PR (n = 13) and in only 15% of patients with a clinical improvement (n = 2) or stable disease (n = 11).

Patient-reported outcomes and symptom improvement. The
AdvSM-symptom assessment form (AdvSM-SAF) patient diary evaluated symptoms across two domains-gastrointestinal and skin. Patients provided information about the severity of abdominal pain, nausea, vomiting, diarrhea, spots, itching, flushing and fatigue, and the frequency of vomiting and diarrhea. Reductions from baseline were seen in both the gastrointestinal and skin domains. Treatment with avapritinib also yielded consistent reductions in the patient-reported AdvSM-SAF total symptom score (TSS), which encompassed gastrointestinal and skin symptoms and fatigue. The mean TSS at baseline was 19.1 points (n = 40). Statistically significant improvements in symptoms occurred rapidly and were sustained through C11 (mean change from baseline of −10.9 points, P = 0.002, n = 20; Fig. 2e).

Discussion
Patients with AdvSM have life-threatening organ damage and poor survival, caused by the increased proliferation of neoplastic mast cells driven by the KIT D816V mutation in ~95% of patients 6 . Midostaurin, until recently the only approved therapy for all subtypes of AdvSM, does not selectively target the KIT D816V mutant and only rarely do patients achieve a complete remission 7 . In a post hoc exploratory analysis of the pivotal midostaurin phase 2 study using IWG-MRT-ECNM criteria, responses occurred in 28% of patients with a 2% CR rate 32,33 .
Based on data from this phase 1 study and the ongoing phase 2 PATHFINDER (NCT03580655) study, avapritinib received United States FDA approval in June 2021 for adult patients with all subtypes of AdvSM, at a recommended dose of 200 mg orally QD. Due to the risks of ICB associated with severe thrombocytopenia, avapritinib is not recommended for the treatment of patients with AdvSM with a platelet count of <50 × 10 9 /l. The results of this phase 1 dose-escalation/expansion study demonstrated that potent and selective inhibition of KIT D816V by avapritinib elicited rapid, profound reductions of measures of mast cell disease burden, resolved organ damage, and improved patients' symptoms across the spectrum of AdvSM subtypes.
Responses (CR/CRh/PR/CI) occurred in 75% of patients, including 36% achieving CR/CRh. These responses were adjudicated using stringent mIWG-MRT-ECNM response criteria, which require a sustained response with 12-week confirmation and full resolution of one or more C-findings. Responses were observed irrespective of disease subtype, prior therapy or presence of high-risk S/A/R mutations.
Notably, KIT D816V became undetectable in the BM of 30% of patients with AdvSM, using a highly sensitive ddPCR assay. This demonstrated that avapritinib induced deep molecular responses, an outcome that represents a new response benchmark in AdvSM. In an earlier multivariate analysis of 38 midostaurin-treated patients with AdvSM, a mere 25% reduction in KIT D816V RNA-expressed allele burden was statistically associated with prolonged survival 29 .
Consistent with these findings, early outcomes with avapritinib were excellent. While midostaurin and avapritinib have not been directly compared in a prospective, randomized trial, the survival of patients with AdvSM in the current study is noteworthy. The median OS was not reached in patients with AdvSM on avapritinib, with an estimated 24-month OS rate of 76% (95% CI, 64-87%). The published literature confirms the particularly poor prognosis of patients with MCL, with only a 24% probability of 24-month OS in this subgroup 34 . In the current study of avapritinib, the 24-month estimated median OS in patients with MCL was 92% (95% CI, 76-100%), while in the previous trial of midostaurin, 24-month OS in patients with MCL was 26% (95% CI, 6-54%) 7 . Importantly, avapritinib demonstrated high ORR rates and similar OS in patients with and without prior midostaurin exposure, suggesting that cross-resistance to avapritinib is rare.
Although patients with and without S/A/R mutations at baseline had similar response rates, the presence of S/A/R mutations or a MARS score of ≥2 was associated with shorter OS. These findings suggest that although avapritinib reduced mast cell burden and C-findings regardless of co-mutation status, patients with adverse myeloid co-mutations such as S/A/R were more likely to experience disease progression. In contrast, patients with AdvSM and a MARS score of 0-1 had excellent outcomes on avapritinib monotherapy in this study, as demonstrated by a 100% 24-month OS rate.
Avapritinib at ≥200 mg QD starting doses induced rapid responses, with a median time to CR/CRh/PR of 2 months, as compared with over 9 months for <200 mg QD starting doses. Given the poor overall survival characteristic of AdvSM, the rapid responses observed with a starting dose of at least 200 mg are clinically important.
Avapritinib was generally well tolerated with few discontinuations due to AEs. The most common AEs, including fluid retention, gastrointestinal effects and cytopenias, were consistent with those reported with other inhibitors of KIT/PDGFR, and usually responded to dose modification. Cytopenias were the most common ≥grade 3 AEs, and they occurred at a higher rate in patients who had baseline cytopenias. Cognitive events and ICB have previously been reported in patients with GIST treated with avapritinib 27 . Although symptoms such as brain fog and memory impairment are known disease features of SM 35 , cognitive effect AEs were less frequent in this AdvSM population than those previously observed in patients with GIST 27 , which may be related to the lower starting doses in this study.
ICB was more frequent in the AdvSM setting compared with GIST and was strongly associated with antecedent severe thrombocytopenia (platelets <50 × 10 9 /l). The incidence of ICB in AdvSM in the absence of severe thrombocytopenia (1%) was similar to the incidence in GIST. The observation of an association between severe thrombocytopenia and ICB led to the implementation of several patient selection and risk mitigation strategies in both the phase 1 EXPLORER and ongoing phase 2 PATHFINDER (NCT03580655) studies: exclusion of severely thrombocytopenic patients, increased platelet count monitoring, dose interruption in patients developing severe thrombocytopenia, and permanent treatment discontinuation after ICB of any grade. Antiplatelet agents and anticoagulants were avoided, and any coagulation abnormalities were corrected, if possible. The risk of ICB was mitigated in the aforementioned phase 2 study following the implementation of these measures. A proactive patient management strategy to minimize ICB risk will require platelet monitoring with dose interruption, reduction and platelet support in the event of severe thrombocytopenia.
The primary objectives of this dose-escalation/expansion study were to determine the MTD, safety and the RP2D of avapritinib, an investigational potent and selective KIT D816V inhibitor. Considering efficacy, tolerability and PK, the RP2D was identified as 200 mg QD based on consistent exposure above the KIT D816V in vivo IC 90 and rapid time to response. At this dose, avapritinib induced deep and durable clinical, pathologic and molecular responses in patients with AdvSM, reduced disease-related symptoms, and was generally well tolerated.
The ongoing phase 2 PATHFINDER study has been designed to further characterize the safety and efficacy of avapritinib in patients with AdvSM at the 200 mg QD starting dose.

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/ s41591-021-01538-9.

Methods
Patients. Patients aged ≥18 years with locally diagnosed AdvSM or relapsed or refractory myeloid malignancies with evidence of aberrant KIT or PDGFR signaling were enrolled in the dose-escalation phase (part 1), and only patients with locally diagnosed AdvSM according to WHO criteria 36 were enrolled in the dose-expansion phase (part 2). Patients with ECOG performance status of 0-3 were eligible. Patients with platelets <25 × 10 9 /l were excluded initially; the threshold was subsequently raised to <50 × 10 9 /l to mitigate the risk of ICB. The complete inclusion and exclusion criteria are listed in Supplementary  Table 3. Diagnoses were retrospectively adjudicated by central pathology and by an Adjudication Committee using WHO criteria. In patients with a centrally confirmed diagnosis of AdvSM, evaluable baseline C-findings and responses were defined using mIWG-MRT-ECNM 13 consensus criteria (Supplementary  Tables 4 and 5).
Study objectives. The primary endpoints were MTD, RP2D and assessment of safety. Secondary endpoints included the ORR based on measurable C-finding responses per mIWG-MRT-ECNM criteria, which require 12-week confirmation of response. Responses in C-findings were centrally adjudicated by a Response Adjudication Committee (RAC) comprising a subset of study investigators (see below for further details and a list of RAC members). ORR included: CR or CRh (including complete resolution of all C-findings, elimination of marrow mast cell aggregates and serum tryptase <20 ng ml −1 ); PR (resolution of ≥1 C-finding and ≥50% reduction in marrow mast cells and serum tryptase); and clinical improvement (resolution of ≥1 C-finding).
Other secondary endpoints included PK, DOR and changes in measures of mast cell burden (percentage of BM mast cells, serum tryptase concentration, KIT D816V VAF by ddPCR (detection limit 0.17%), and spleen and liver volumes). PROs were measured by daily completion of AdvSM-SAF (described below) for patients in part 2 as a secondary outcome [37][38][39] . Exploratory endpoints included time to response, PFS and OS.
Study treatment and design. This phase 1, open-label study comprised a dose-escalation phase (part 1) and a dose-expansion phase (part 2; Fig. 1). During dose escalation, patients received oral avapritinib at starting doses ranging from 30 to 400 mg QD in continuous 28-d treatment cycles. Dose escalation followed a 3 + 3 design (described below) until the MTD or RP2D was determined. During part 1, intra-patient dose escalation to assess dose levels not exceeding the MTD was allowed. Dose expansion was initially conducted at a starting dose of 300 mg QD; a second expansion cohort at 200 mg QD was subsequently introduced via protocol amendment. Dosing continued until patients experienced unacceptable toxicity, disease progression, death or withdrew consent.
3 + 3 dose-escalation study design. The 3 + 3 dose-escalation design used cohorts of three to six patients. The first cohort received avapritinib at a starting dose of 30 mg daily. Dose escalation then proceeded at increments up to 100% until one or more patients treated at a given dose level had a ≥grade 2 non-hematologic AE possibly related to avapritinib or a grade 4 hematologic AE possibly related to avapritinib, or if the given dose exceeded the highest dose determined to be safe in the first-in-human study of avapritinib in patients with GIST (400 mg). Three patients were enrolled initially in each cohort with an additional accrual of three patients if the cohort required expansion due to DLT. Once the escalation cohort was full, up to three additional patients (diagnosis of AdvSM only) could be enrolled into an enrichment cohort at a lower dose. Data from these patients were intended for further exploration of PK, pharmacodynamics and safety in patients with AdvSM.
On-study corticosteroid treatment. Corticosteroid treatment, not exceeding doses administered during screening, was permitted; doses higher than 20 mg daily prednisone equivalent were avoided, and prolonged treatment (>14 d) with high-dose corticosteroids rendered patients ineligible for response assessment. All patients provided written informed consent. Participants were not compensated, except for reimbursement of reasonable travel expenses. All authors had access to all data, reviewed and provided critical input to the manuscript and made the decision to submit for publication. All authors vouch for the validity of the study results and adherence to the protocol.
A formal independent data monitoring committee was not used for this study. The sponsor had access to the safety data on a regular basis, as this was an open-label study. The sponsor's clinical study team hosted investigator teleconferences on a regular basis during the study. During part 1 of the study, the clinical study team and the investigators spoke by teleconference at the end of each treatment cohort to discuss and evaluate all of the gathered safety data. At the dose-escalation teleconference, safety information, including DLTs and all grade 2 or worse AEs reported during C1, and all available PK data were described and reviewed for each patient in the current dose cohort. Updated safety, PK and other data for all other ongoing patients, including data from later cycles, were also discussed. In addition, emerging safety and efficacy data were reviewed by the sponsor at quarterly safety meetings that included the study medical monitor, site personnel and biostatisticians.
In the event a patient was withdrawn from study drug administration or the follow-up phase of the study, the medical monitor was informed. If there was a medical reason for withdrawal, the patient remained under the supervision of the investigator or designee until the condition returned to baseline or stabilized.
Central adjudication of diagnosis. To ensure a consistent diagnosis, retrospective central pathologic review of enrolled patients' BM aspiration and biopsy samples and peripheral blood smears were assessed by an independent central pathologist. In addition, documentation (including site-entered data, as well as requested de-identified radiology reports and clinic notes) of baseline and historical (if relevant) C-findings were reviewed by the RAC to adjudicate and obtain consensus on SM diagnosis and subtyping based on the sum of available data. Diagnosis and subtyping were performed according to WHO criteria, with the following clarifications: the C-finding of weight loss was defined as documented weight loss of >10% within a 6-month (±3 months) period, and large osteolytic lesions were defined as ≥2 cm. These clarifications and strict requirement for documentation led to locally diagnosed ASM patients being reclassified as having indolent SM or smoldering SM by central diagnosis (based on data available to the RAC) for the purposes of data analysis. The MCL category also included MCL cases with evidence of AHN. The resulting comparison of local and centrally adjudicated diagnoses is presented in Supplementary Table 6. six 28-d treatment cycles with ≥2 post-baseline BM assessments, or having an end-of-study assessment). The response-evaluable population was used to report adjudicated response rates and PFS. The PK population comprised all patients with sufficient plasma concentration-time data to reliably estimate the PK parameters of avapritinib.
The percentage change from baseline in spleen and liver volume as measured by central radiographic assessment using an MRI or a CT scan was summarized over time in the safety population. In addition, maximum percentage reduction was presented in waterfall plots.
Spleen response was determined for patients with baseline splenomegaly (spleen ≥ 5 cm by palpation) and was assessed by palpation or spleen volume by imaging (that is, ≥35% reduction from baseline in spleen volume based on central imaging); presence of either criterion was considered a response. DOR started at the first evidence of a response in any evaluable C-finding (per mIWG-MRT-ECNM criteria) and ended at the time a response was lost (that is, loss of response per mIWG-MRT-ECNM criteria). For patients who were still responding at the time of analysis, DOR was censored at the latest time point that the patient is assessed as having clinical improvement or better. OS was analyzed in the safety population and was defined as the time from first dose to the time of death due to any cause. PK data were summarized for patients with sufficient plasma concentration-time data. Descriptive statistics were used to summarize PK parameters for each dose level, as appropriate.
No imputation was performed for missing data elements. Where the date of onset of an AE was missing, event onset was assumed to be the date of treatment to conservatively report the event as treatment emergent.
Study assessments. Responses were assessed according to the protocol schedule. Spleen size was assessed by palpation and by imaging with calculation of volume centrally. KIT D816V VAF in the BM (or in blood if unavailable) was centrally measured using a validated ddPCR assay with a lower limit of detection of 0.17% (Bio-Rad). Other somatic mutations in BM were centrally assayed at baseline by next-generation sequencing.
ORR was defined as the proportion of patients with a confirmed best response of CR, CRh, PR or clinical improvement by mIWG-MRT-ECNM consensus criteria. Time to CR/CRh/PR was defined as the time from the start of treatment to the time a CR/CRh/PR by mIWG-MRT-ECNM was first met. DOR was defined as the time from first documented response to the date of first documented progressive disease/loss of response or death due to any cause, whichever occurred first. OS was defined as the time from the start of treatment to the date of death. PFS was defined as the time from the start of treatment to the date of first documented progressive disease or death due to any cause, whichever occurred first. Progressive disease was defined per mIWG-MRT-ECNM criteria as worsening of ≥1 baseline C-finding confirmed for at least 4 weeks, or progression to acute myeloid leukemia.
Safety assessments. Safety assessments included determination of ECOG performance status, clinical laboratory testing, vital signs, electrocardiograms, brain imaging (MRI or CT scan) and physical examinations. AEs were graded according to National Cancer Institute Common Terminology Criteria (v4.03) 40 . Treatment-emergent AEs were defined as any AE occurring between the first dose of avapritinib through 30 d after the last dose of avapritinib, any event that was considered related to the study drug regardless of the start date of the event, or any event that was present at baseline but worsened intensity or was subsequently considered related to the study drug by the investigator.
Patient-reported outcomes assessment. The AdvSM-SAF comprises a ten-item questionnaire completed by the patient daily, specifically designed to assess symptoms related to AdvSM and measure the impact of treatments on patient symptom improvement 39 . It has been validated and demonstrated to be reliable and sensitive to clinical changes. The AdvSM-SAF assesses the severity of eight symptoms of AdvSM, including abdominal pain, nausea, vomiting, diarrhea, spots, itching, flushing and fatigue over a 24-h period on a numerical rating scale, and also measures the frequency of vomiting and diarrhea by asking patients to enter a discrete numerical value. Data were analyzed at the item level, domain level and total score level. Domain level analyses are: • Gastrointestinal symptoms score (abdominal pain, nausea, vomiting and diarrhea severity (range 0-40)) • Skin symptom score (spots, itching and flushing severity (range 0-30)) The TSS was achieved by summing all eight severity items (range 0-80).
Pharmacokinetic assessments and analysis. On D1 and D15 of C1, blood samples for PK analysis were collected before the dose, and 0.5, 1, 2, 4, 8 and 24 h (C1D1 only) after the dose. C2D1 pre-dose concentration was used as the C1D15 24-h concentration for PK analysis. Additional pre-dose samples were collected on D1 of C3 and C4. Avapritinib PK parameters were obtained by non-compartmental analysis of the plasma concentration data versus nominal time using Phoenix WinNonlin© (version 8.0, Certara). Parameters estimated included C max , T max , AUC 0-24 ; AUC 0-τ , apparent oral clearance, apparent volume of distribution, terminal half-life, trough concentrations 24 h after dose and accumulation ratio. Plasma concentration-time profiles across dose levels and dose versus systemic exposure (C max and AUC) relationships were assessed graphically. Systemic exposure parameters of interest included the arithmetic and geometric mean, C max and AUC, and 90% CIs.
Additional statistical methods. The total number of patients enrolled in part 1 was dependent on the observed safety profile, which determined the number of patients per dose cohort, as well as the number of dose escalations required to identify the MTD and RP2D. It was expected that approximately 25 patients who met the criteria for the dose-determining population (described below) would be enrolled in part 1. In part 2, to adequately assess the safety of avapritinib, approximately 45 patients were planned for enrollment in cohort 1 at a starting dose of 300 mg QD, and approximately 10 patients were planned to be enrolled in cohort 2 at a starting dose of 200 mg QD. With 55 patients, there is approximately 94% probability of observing an AE that occurs at a frequency of ≥5%.
Primary ORR analysis was based on RAC-adjudicated responses in the response-evaluable population. ORR was estimated using frequency, percentage and two-sided 95% CIs based on the exact binomial distribution (Clopper-Pearson). Statistical testing on binomial proportion against a null of 28% was performed using one-sided α = 0.025. Wald test P value was presented. CR + CRh + PR rate was summarized similarly as ORR. Statistical testing on binomial proportion against a null of 17% was performed using one-sided α = 0.025.
OS, PFS and DOR were determined by Kaplan-Meier estimates, and 95% CIs were estimated using Greenwood's formula 41 . Comparisons between two Kaplan-Meier functions was performed using a log-rank test.
Data analysis was performed using SAS version 9.3 software.