This is a phase I/II study to evaluate the safety of combining intravenous (IV) atezolizumab and bevacizumab every three weeks, with daily oral cyclophosphamide and pharmacokinetic (PK)-guided sorafenib in children and adolescent and young adults (AYA) with relapsed or refractory solid malignancies (Part 1), and then evaluate the response rate of this combination in children, AYA with relapsed or refractory fibrolamellar carcinoma (FLC) and other rare solid malignancies (Part 2).

Primary Objectives Part 1

• To establish the safety associated with the administration of the combination of cyclophosphamide, PK-guided sorafenib, bevacizumab and atezolizumab in children and AYA with relapsed or refractory solid tumors
• To determine if sorafenib systemic exposure can be successfully targeted to an AUC between 20 and 55 hr·µg/mL by Day 21 of cycle 1 in 60% of evaluable patients, when given in combination with cyclophosphamide, bevacizumab, and atezolizumab in children and AYA with relapsed or refractory solid tumors

Part 2

• To evaluate the response rate (CR+PR) of the combination of cyclophosphamide, PK-guided sorafenib, bevacizumab and atezolizumab in children and AYA with relapsed or refractory FLC following two cycles of therapy
• To determine if the use of PK-guided sorafenib dosing to maintain a systemic exposure between 20 and 55 reduces the interpatient pharmacokinetic variability of sorafenib and the incidence of sorafenib- induced skin toxicities in children and AYA with relapsed or refractory FLC and other rare solid tumors

Parts 1 & 2

• To determine if the combination of cyclophosphamide, PK-guided sorafenib and atezolizumab will result in increased intratumoral T-cell infiltration of CD8+C45RO+ cells between baseline and following two courses of therapy in pediatric children and AYA with relapsed or refractory solid tumors following two cycles of therapy
• To characterize the pharmacokinetics of atezolizumab in combination with cyclophosphamide, PK-guided sorafenib and bevacizumab in children and AYA with relapsed or refractory solid tumors
• To assess the feasibility of performing contrast enhanced ultrasound and explore the correlation between quantitative CEUS parameters and clinical response.

Secondary Objectives

Part 1

• To describe the response rate (CR+PR) of the combination of cyclophosphamide, PK-guided sorafenib, bevacizumab and atezolizumab in children and AYA with relapsed or refractory solid tumors following two cycles of therapy

Part 2

• To describe the response rate (CR+PR) of the combination of cyclophosphamide, PK-guided sorafenib, bevacizumab and atezolizumab in children and AYA with relapsed or refractory FLC, HCC, desmoplastic small round cell tumor, malignant rhabdoid tumor, and other rare solid tumors following two cycles of therapy

Parts 1&2

• To describe the number of children with liver tumors, initially judged unresectable at diagnosis, that can have their primary tumor resected after treatment with oral cyclophosphamide and sorafenib with intravenous bevacizumab and atezolizumab
• To describe changes in immune cells in the peripheral blood at periodic times before and after treatment with this combination chemoimmunotherapy
• To describe the PFS, EFS, and OS in patients treated with the combination of cyclophosphamide, PK-guided sorafenib, bevacizumab, and atezolizumab in patients with relapsed or refractory FLC, DSRCT, MRT, HCC and other rare solid tumors

Exploratory Objectives (Parts 1 & 2)

• To determine the number of HCC and FLC xenografts that can be successfully established from children with relapsed or refractory disease (enrollment on MAST with biopsy for fresh tissue required)
• To evaluate the immune cell infiltrates in children with relapsed or refractory solid tumors before and after two cycles of cyclophosphamide, sorafenib, bevacizumab and atezolizumab
• To evaluate the number, type, specificity, repertoire, and activity of intratumoral T cells after two cycles of therapy, compared to baseline
• To characterize changes in tumor associated macrophages (polarization) between baseline and after two cycles of cyclophosphamide, sorafenib, bevacizumab and atezolizumab
• To measure changes in tumor mutational burden and mutation heterogeneity
• To explore associations between T stem cell epigenetic signatures and response to treatment
• To explore the association between baseline CD45RO+/CD8+ T-cell infiltration and PD-L1 expression and response to treatment
• To characterize tumor heterogeneity and microenvironment using single-cell/nuclear RNA sequencing pre- and post-treatment
• To longitudinally assess and quantify numerous metrics of quality of life (QoL), family distress and functional impairment for patients enrolled on ANGIOA and their primary caretakers
• To qualitatively assess patient and family physical, emotional and psychosocial experiences prior to and after receipt of therapy on ANGIOA through semi-structured interviews
• To assess the acceptability and feasibility of capturing patient/caregiver interview data at the time of enrollment and discontinuation on ANGIOA
• To assess the feasibility of performing quantitative multiparametric mapping sequences in patients already undergoing MRI for staging purposes and explore the correlation between quantitative enhancement and other parameters and clinical response

Part 1 (safety/tolerability): Children with relapsed or refractory solid tumors with biopsy accessible and evaluable disease will be treated with two courses of oral PK-guided sorafenib (starting area under the curve (AUC) target: 20-55 ug/ml/hr), oral cyclophosphamide (50 mg/m2/dose, daily (qd) x 21 days), IV bevacizumab (15 mg/kg/dose, every (q) 21 days) and IV atezolizumab (15 mg/kg [max dose 1200 mg] q 21 days). Tumor biopsies are required before starting treatment and after course two (Section 4.1). Biopsied tissue will be used for enrollment on MAST and to evaluate changes in T-cell infiltration. Sorafenib PK will be obtained and dose adjusted to target an AUC between 20 and 55 hr·µg/mL by Day 21 of C1. Tolerability will be defined after completion of Course 1. Part 2 will begin once the recommended phase 2 dose (RP2D) is determined.

Part 2 (efficacy): Children and AYA with relapsed or refractory FLC will be treated with two courses of oral cyclophosphamide and sorafenib with IV bevacizumab and atezolizumab based on the RP2D from Part 1. Tumor response will be assessed after two courses according to immunologic and imaging criteria (Section 4.1). Eligible patients with hepatocellular (HCC), desmoplastic small round cell tumor (DSRCT) or non-central nervous system (CNS) malignant rhabdoid tumors (MRT) will be enrolled on separate strata but target accrual will be determined by patients with FLC.

Inclusion Criteria:

• Age: Patients must be < 30 years at the time of enrollment on study.

• Willingness to enroll on the St. Jude Molecular Analysis of Solid Tumors (MAST) study.

• Diagnosis

• Part 1: Patients with refractory or recurrent (relapsed) solid tumors accessible by biopsy for which there is no standard therapy are eligible.

• Part 2: Patients with one of the following diagnoses:

• Biopsy accessible refractory or recurrent (relapsed) hepatocellular carcinoma

• Biopsy accessible refractory or recurrent (relapsed)or FL-HCC, DSRCT or non-CNS MRT.

• Performance level: Karnofsky > 50 for patients > 16 years of age and Lansky > 50 for patients < 16 years of age (See Appendix III). Note: Patients who are unable to walk because of paralysis, but who are up in a wheelchair, will be considered ambulatory for the purpose of assessing the performance score.

• Disease status: Patients must tumors that are unresectable and have either measurable or evaluable disease that is accessible by biopsy

• Organ function: Must have adequate organ and bone marrow function as defined by the following parameters:

• Patients with solid tumor not metastatic to bone marrow:

  • Peripheral absolute neutrophil count (ANC) >1,000/mm3
  • Platelet count > 75,000/mm3 (no transfusion within 7 days of enrollment)
  • Hemoglobin > 8 g/dL (with or without support)

• Patients with solid tumor metastatic to bone marrow will be eligible for study but not evaluable for hematologic toxicity. These patients must not be known to be refractory to red cell or platelet transfusions. At least 2 of every cohort of 3 patients must be evaluable for hematologic toxicity. If dose limiting hematologic toxicity is observed at any dose level, all subsequent patients enrolled must be evaluable for hematologic toxicity.

• Adequate renal function defined as serum creatinine based on age as shown in Table 1, or creatinine clearance or radioisotope GFR 50 ml/min/1.73m2 (GFR 40 ml/min/1.73m2 if < 2 years of age).

• Adequate hepatic function defined as total bilirubin < 5x upper limit of normal (ULN) and AST/ALT < 3 x ULN for age.

• Adequate cardiac function defined as shortening fraction > 28% OR ejection fraction of ≥ 47% by echocardiogram.

• Adequate blood clotting defined as PT/PTT < 1.2 x ULN without factor replacement products for 7 days

• Females of childbearing potential and males able to father a child must be willing to practice acceptable methods of birth control to prevent pregnancy during the study and for at least 5 months after last dose of therapy.

• Patients must have fully recovered from the acute toxic effects of chemotherapy, immunotherapy, surgery, or radiotherapy prior to entering this study:

• Myelosuppressive chemotherapy: Patient has not received myelosuppressive chemotherapy within 1 weeks of enrollment onto this study (within 2 weeks of estimated therapy start date) (4 weeks if prior nitrosourea).

• Hematopoietic growth factors: At least 7 days must have elapsed since the completion of therapy with a growth factor. At least 14 days must have elapsed after receiving pegfilgrastim.

• Biologic (anti-neoplastic agent): At least 7 days must have elapsed since completion of therapy with a biologic agent. For agents that have known adverse events occurring beyond 7 days after administration, this period prior to enrollment must be extended beyond the time during which adverse events are known to occur.

• Monoclonal antibodies: At least 14 days (at least 21 days from therapy start date) must have elapsed since the completion of therapy with a monoclonal antibody.

• Radiotherapy: At least 1 week (2 weeks from estimated therapy start date) must have elapsed since any irradiation; at least 5 weeks (at least 6 weeks from estimated therapy start date) must have elapsed since craniospinal RT or substantial bone marrow irradiation.

• Chemoembolization: at least 21 days (28 days from estimated therapy start date) must have elapsed since the completion of chemoembolization

• Radioembolization: at least 21 days (28 days from estimated therapy start date) must have elapsed since the completion of radioembolization

• Cardiac disease or hypertension: Patients must not have a history of myocardial - infarction, severe or unstable angina, or severe peripheral vascular disease. Hypertension must be well controlled on stable doses of medication for at least two weeks.

• Female participant who is post-monarchal must have a negative urine or serum pregnancy test.

• Life expectancy of at least 8 weeks

Exclusion Criteria:

• Pregnant or breastfeeding.
• Currently receiving other investigational drugs.
• Unwilling or unable to comply with the safety monitoring requirements of this protocol.
• Tumor not safely accessible by biopsy
• Inability or unwillingness of research participant or legal guardian / representative to give written informed consent.
• Surgical procedures and serious or non-healing wounds: patients with a documented, chronic non-healing wound, ulcer, or bone fracture or history of a major surgical procedure or significant traumatic injury within 28 days prior to beginning therapy are excluded due to preclinical evidence supporting the potential for delayed wound healing.
• Minor surgical procedures for minimally invasive biopsies will be allowed. For minor surgeries, the wound must be healed, and 7 days elapsed since surgery. For procedures such as the placement of an indwelling IV catheter, it is recommended that bevacizumab be postponed for at least 24 hours after the procedure.
• Thrombosis: Patients must not have a deep venous or arterial thrombosis (including pulmonary embolism) within the last three months prior to study entry and must not have a known thrombophilic condition (i.e., protein S, protein C or antithrombin III deficiency, Factor V Leiden, Factor II G20210A mutation, homocysteinemia or antiphospholipid antibody syndrome).

Individual participant de-identified datasets containing the variables analyzed in the published article will be made available (related to the study primary or secondary objectives contained in the publication). Supporting documents such as the protocol, statistical analyses plan, and informed consent are available through the CTG website for the specific study. Data used to generate the published article will be made available at the time of article publication. Investigators who seek access to individual level de-identified data will contact the computing team in the Department of Biostatistics (ClinTrialDataRequest@stjude.org) who will respond to the data request.

Data will be made available at the time of article publication.

Data will be provided to researchers following a formal request with the following information: full name of requestor, affiliation, data set requested, and timing of when data is needed. As an informational point, the lead statistician and study principal investigator will be informed that primary results datasets have been requested.