The following six clinical studies led to the development of a palatable pediatric formulation of crizotinib to treat pediatric patients with cancer. The decisions made were based on the goal of developing a pediatric formulation that was palatable, tolerated, dose-flexible, easy to administer, and bioequivalent to the commercial adult formulation.

2.1 Study 1: ADVL0912 (NCT00939770)—Development of Crizotinib Pediatric Oral Solution

Crizotinib is commercially available as hard gelatin capsules of 200 mg or 250 mg for adults and children who are able to swallow intact capsules [8]. Crizotinib was initially evaluated in pediatric patients in a Children’s Oncology Group (COG) phase I/II study (Study ADVL0912 [ClinicalTrials.gov identifier: NCT00939770]; Fig. 1a) as an oral solution for patients unable to swallow capsules [16]. The first clinical dosage form developed for the pediatric population was a crizotinib 25-mg/mL oral solution [17]. The development of this oral solution focused on inclusion of excipients needed for taste masking to improve palatability, including a screening of sweeteners, flavors, and mouthfeel modifiers selected specifically to mask the bitter taste and burning sensation associated with the drug. The excipient screening formulation development for oral liquid taste-masking excipient selection involved a combination of research based on published articles, in vitro compatibility studies (physical and chemical compatibility of the excipient with the active ingredient in solution), and healthy adult taste panel evaluations prior to the introduction, and further palatability assessments in pediatric patients. In total, across the research-based considerations, physical and chemical compatibility assessments, and taste panel evaluations, six sweeteners, 16 flavors, and 68 potential mouthfeel modifiers were evaluated. This work enabled minimization of the perception of bitterness and decreased the burning mouthfeel sensation to the extent possible for a challenging taste profile compound delivered as an oral liquid.

Fig. 1

Clinical study designs of different crizotinib formulations. a This COG phase I/II study in pediatric patients evaluated crizotinib initially as an oral solution in patients unable to swallow capsules. b This COG phase I PK study was conducted to determine safety, tolerability, and the recommended phase II dose of crizotinib in combination with cytotoxic chemotherapy for children and young adults with ALK-positive refractory solid tumors and ALK-positive ALCL. c This study in healthy adults in the fasted state evaluated three coated microsphere formulations, two pH-dependent systems (Smartseal-1 and Smartseal-2) and one pH-independent system (Opadry II), to determine the optimal balance between PK and acceptable palatability. d This phase I study in healthy adults evaluated the potential effect of food and PPI (esomeprazole) on PK of crizotinib given as a microsphere formulation. e Using six different treatment sequences in healthy adults, this study evaluated the two redesigned clinical prototype formulations against the oral solution (for palatability), against the FC (for relative bioavailability), and for the potential effect of food or PPI on the PK of crizotinib. f This phase I study in healthy adults was conducted to establish bioequivalence of the Opadry SGR–coated microsphere formulation to the FC using direct-to-mouth sprinkle administration. Each period in the treatment sequence consisted of a single dose of crizotinib 250 mg in the specified form, given on Day 1 after an overnight fast of at least 10 hours, with a washout period of at least 14 days between doses. ALK anaplastic lymphoma kinase, BMI body mass index, BOR best overall response, cMS crizotinib microsphere formulation, CNS central nervous system, COG Children’s Oncology Group, eMS encapsulated microsphere formulation, FC formulated capsule, MRD minimal residual disease, MTD maximum tolerated dose, PK pharmacokinetics, PPI proton pump inhibitor, PS performance status, RP2D recommended phase II dose

2.2 Study 2: ADVL1212 (NCT01606878)—Clinical Evaluation of Oral Crizotinib and Oral Solution Palatability

Study ADVL1212, a COG phase I PK study (NCT01606878; Fig. 1b), was conducted to determine the safety, tolerability, and recommended phase II dose of crizotinib in combination with cytotoxic chemotherapy for children and young adults (aged 1–21 years) with ALK-positive refractory solid tumors and ALK-positive ALCL [18]. Three crizotinib formulations, the formulated oral solution, formulated capsules, and an early prototype microsphere formulation, were initially evaluated.

To overcome the palatability issues with the oral solution, focus was shifted toward developing an oral solid multiparticulate platform that could be used to develop a patient-centric pediatric formulation that was palatable for patients, tolerated, easy to administer at the required doses, and bioequivalent to the commercial capsule formulation.

2.3 Development of the Oral Solid Multiparticulate Pediatric Platform

Palatability challenges in early testing with the oral solution triggered the quest for a pediatric formulation that could effectively provide taste masking without impacting the absorption of the product. It was hypothesized that if an acceptable taste-masked product could be achieved for crizotinib, a drug with challenging palatability issues, while maintaining bioequivalence to the commercial capsule formulation, these learnings could be leveraged more broadly across a wide range of compounds. The three areas of focus for the platform were designing the multiparticulate core, coating the core with a barrier membrane, and commercial manufacturing of the final drug product [19]. Pfizer has adopted melt-spray-congeal (MSC) technology as one of the flexible oral solid multiparticulate platforms of particular interest in our drug product toolbox for challenging active pharmaceutical ingredients requiring an effective taste-masked coating.

2.3.1 Multiparticulate Core

Multiple technologies, such as minitablets, extrusion-spheronization, drug-layered beads, ion exchange resin, rotary and fluid bed granulation, and lipid encapsulation by MSC spinning disk atomization, were considered to make the multiparticulate core. During the designing of the core as a pediatric platform approach, critical quality attributes included (1) size: tunable and monodisperse particle-size distribution; (2) shape: spherical and symmetrical; and (3) strength: low friability. The spherical and symmetrical core was to provide a better substrate for downstream processing (i.e., coating, encapsulation) and improved flow properties of the multiparticulates. A core with low friability was to ensure that the substrate could withstand downstream processing.

Ultimately, development focused on the MSC spinning disk atomization technology. MSC was selected because it is a continuous process, can be developed with a small amount of drug substance (gram scale), and can be scaled up with little to no additional development work. The MSC technology produces highly spherical particles with a tight particle-size distribution [20] in which the crystalline drug substance is encapsulated in a lipid carrier, providing the first level of taste-masking performance. The particles produced also perform well with respect to downstream processing.

The MSC spinning disk atomization technique is used to produce a multiparticulate pharmaceutical dosage form. The final drug product consists of a lipid (stearyl alcohol) microsphere with the crystalline active pharmaceutical ingredient dispersed throughout. The drug is released from the microsphere core via porous diffusion through channels formed as the active pharmaceutical ingredient dissolves or through pores created by various soluble excipients (e.g., poloxamer 407) that can be added to the formulation.

The MSC process involves preparing a suspension of the crystalline drug in the lipid, feeding the suspension onto a spinning disc for atomization, and collecting particulates formed by congealing of droplets exiting from the edge of the spinning disc. The key process parameters that influence the microsphere size and distribution in the MSC process are the feed rate of suspension onto the spinning disk, rotational speed of the disk, and operating temperature of the system [21]. The primary material attribute of the suspension that influences the microsphere size is its viscosity.

The target particle size for uncoated microspheres, 150–250 µm, was selected to avoid the gritty mouthfeel reported with larger particles [22]. This target particle size was also desirable for downstream processing, including fluid bed coating and encapsulation [23].

2.3.2 Multiparticulate Coating

When developing a barrier coating for taste masking, there were two main approaches to consider: (1) pH-dependent membranes, in which the solubility of the coating is triggered based on the physiological pH in the body, and (2) pH-independent membranes, in which the solubility of the coating is triggered based on exposure time and not the physiological pH.

Taste masking using pH-dependent membranes can be achieved by using reverse enteric polymers that are typically soluble below a certain pH. These coatings prevent the release of the drug in the mouth (pH ≈6.2) [24,25,26] where the membrane is insoluble, while allowing its release in the stomach (fasted state, pH ≈1.2) [26, 27] where the membrane is soluble. The PK performance of pH-dependent membranes is highly dependent on the gastrointestinal physiological condition. For example, these reverse enteric coatings may not dissolve completely, or at all, depending on the pH conditions in the stomach or the duration of gastric exposure. Incomplete dissolution of the membrane causes slow or incomplete release of drug and may result in reduced drug absorption and systemic exposures.

pH-independent membranes typically use pH-neutral polymers that result in a dissolution time delay. Drug release is less dependent on the gastrointestinal physiological condition, but the taste masking for some molecules may not be as effective as a pH-dependent coating.

2.4 Study 3: A8081041 (NCT02006277)—Taste and Relative Bioavailability of Three Crizotinib Taste-Masked Microsphere Formulations

The search for the optimal barrier-coated multiparticulate system that would achieve the required balance between PK and acceptable palatability started with Study A8081041 (NCT02006277; Fig. 1c) in healthy adults in the fasted state (N = 20). Three coated microsphere formulations, two pH-dependent (reverse enteric) systems (Smartseal-1 and Smartseal-2) and one pH-independent system (Opadry II), were evaluated in this study for both PK and palatability. Participants were split into two cohorts and received either a 75-mg dose of crizotinib (Cohort 1) or a 250-mg dose of crizotinib (Cohort 2). In Cohort 1, crizotinib 75 mg was given as crizotinib microsphere formulation (cMS) 0.529, 0.470, or 0.420 mg/mg (Treatments A, B, and C, respectively) versus the oral solution (Treatment D), using a “swirl and spit” technique for taste evaluation. In Cohort 2, crizotinib 250 mg was administered orally as cMS 0.529, 0.470, or 0.420 mg/mg (Treatments F, G, and H, respectively) versus the formulated capsule (Treatment E) for relative bioavailability assessment, and Treatments F, G, and H versus the oral solution (Treatment I) for taste evaluation. There was a washout period of at least 14 days between Treatments E, F, G, and H, followed by a washout period of at least 5 days before Treatment I.

2.5 Study 4: A8081066 (NCT03137134)—PK of Crizotinib Taste-Masked Microspheres Under Fasted, Fed, and Proton Pump Inhibitor (PPI) Conditions

The potential for reduced exposure with the reverse enteric–coated system was initially evaluated by in vitro dissolution testing across a range of physiological gastric pH conditions with dissolution media from pH 1.2–6.2. The mean gastric pH after repeated dosing with 40 mg of esomeprazole, a common PPI, was 4.4 and 4.8 in volunteers with gastroesophageal reflux disease and healthy volunteers, respectively [28]. Therefore, criteria for dissolution targets were initially set to ensure complete dissolution of the barrier coating up to pH 5.0 to guide formulation design that was anticipated to support co-dosing of crizotinib with esomeprazole.

The phase I study A8081066 (NCT03137134) was conducted to evaluate the potential effect of food and PPI on the PK of crizotinib given as a microsphere formulation (Fig. 1d). Participants received at least four single doses of 300-mg cMS; each mg of microspheres by weight contained 0.535 mg of active crizotinib. The dose given was 300 mg of active cMS. There was a washout period of at least 14 days between cMS doses. The coating tested in this study was Kollicoat Smartseal 30 D (Smartseal-3), a pH-dependent (reverse enteric) coating. The reference in this study was the Smartseal-3–coated formulation in the fasted state. To further evaluate the potential impact of gastric pH change on the release and absorption from the reverse enteric–coated system, select participants were given an acidic beverage (orange juice) or food (applesauce) when receiving crizotinib and PPI.

2.6 Redesigning the Barrier Membrane Coating and Further Clinical Evaluation

Based on the results of Study A8081066, a formulation redesign was initiated to explore a barrier membrane coating that could provide acceptable taste masking while still maintaining drug release under elevated pH conditions (e.g., coadministration with a PPI) similar to that of the fasted condition. To pursue this goal, development continued with two redesigned barrier membrane systems: (1) a different pH-independent coating material, Opadry SGR, marketed as a sugar-coating system, and (2) a further optimized reverse enteric–coated microsphere (prototype Smartseal-4).

2.7 Study 5: A8081069 (NCT03978143)—Palatability and Relative Bioavailability of Two New Crizotinib Taste-Masked Microsphere Formulations

The two redesigned clinical prototype formulations were evaluated against the oral solution for palatability and against the formulated capsule for relative bioavailability in Study A8081069 (NCT03978143; Fig. 1e). In addition, the potential effect of food or PPI (using esomeprazole as probe) on the PK of crizotinib, given as the two clinical prototype formulations, was also evaluated to assist in selecting a clinically favorable microsphere formulation.

A visual analog scale (VAS) scoring tool was used to evaluate the overall palatability of each formulation. The VAS is an accepted methodology used to correlate the palatability of pediatric medicines with acceptability. Results are scaled from 0–100 (with 0 being “the best” and 100 being “the worst”). VAS scores of > 65 have been correlated with pediatric medicines having both poor taste and poor acceptability [29]. To demonstrate an acceptable formulation, a VAS score of < 65 is preferred for pediatric medicines [30].

2.8 Study 6: A8081074 (NCT04856293)—Bioequivalence of Crizotinib Encapsulated Microsphere Formulation Versus Formulated Capsule

With a successful formulation identified, commercialization activities of the formulation were initiated. A pivotal bioequivalence phase I study, A8081074 (NCT04856293), was conducted to establish bioequivalence of the Opadry SGR–coated microsphere formulation to the commercial formulated capsule (Fig. 1f). This study evaluated direct-to-mouth sprinkle administration with the Opadry SGR–coated microsphere formulation compared with the reference commercial formulated capsule. For administration of the encapsulated microsphere formulation (“encapsulated” in this context describes the microspheres being inside a capsule), capsules were opened, the microspheres were emptied into a dry glass, they were poured into the participant’s mouth without the participant chewing, and then they were swallowed.

A subgroup analysis was conducted to evaluate swallowing the intact sprinkle capsule (Opadry SGR intact capsule) compared with swallowing the direct-to-mouth sprinkle administration of the microspheres followed by drinking water (Opadry SGR).