Background and Aims

Constitutional mismatch repair deficiency syndrome (CMMRD) is a severe pediatric cancer predisposition syndrome resulting in early-onset brain, gastrointestinal (GI), hematological and other cancers. International guidelines for surveillance exist; however, no study has evaluated their efficacy.

Methods

Data were collected from 105 patients with confirmed CMMRD registered to the International Replication Repair Deficiency Consortium. Tumor spectrum, efficacy of the recommended surveillance modalities and malignant transformation of low-grade lesions were analyzed in this cohort. Seventy-seven patients were followed prospectively to assess survival outcome.

Results

A total of 193 malignant tumors were identified. Brain tumors were most common (44%) followed by GI (27%), hematological (19%) and other (10%) malignancies. Surveillance uncovered 46 asymptomatic malignant tumors including 15 brain, 24 GI, 2 hematological and 5 other tumors. Additionally, 5 brain and 10 hematological malignancies were detected symptomatically despite patients being under surveillance. Longer than recommended time between scans or difficulties in patient access were reported as causative factors for symptomatic brain tumors. In contrast, for hematological malignancies, CBC and CT scans were largely ineffective at detecting these cancers. GI screenings and whole-body MRI identified 100% of GI cancers and other tumors, respectively.

To test the role of early detection, patient outcome was measured per tumor. Five-year overall survival was 90% and 50% for patients when cancer was detected asymptomatically or symptomatically, respectively (p=0.001). Patient outcome was subsequently measured based on adherence to surveillance protocol. Four-year overall survival was 74% for patients undergoing full surveillance (n=25), 44% for partial surveillance (n=17) and 12% for those not under surveillance (n=35) (p<0.0001).

Lastly, 72 low-grade tumors were detected. Cumulative chance of transformation from low-grade to high-grade over time showed ~80% chance for GI within 8 years and 100% for brain within 6 years.

Conclusions

These data demonstrate significant survival benefit of routine surveillance practices for patients with CMMRD.

Background and Aims

Pediatric replication repair deficient (RRD) cancers are hypermutant and may respond to immune checkpoint inhibition (ICI). To determine response, survival and biomarker for therapy, we performed a consortium registry study of ICI for recurrent RRD cancers.

Methods

Clinical, genomic and immune data were collected longitudinally on all patients from May 2015 to October 2019. Biomarkers included tumor mutational burden (TMB), neoantigens and genetic signatures obtained from whole genome and exome sequencing. Immune inference was obtained by RNAseq and T cell rearrangement (TCR) was collected in the tumor and in blood throughout treatment. Immunohistochemistry (IHC) was performed on tumors to probe for PD-L1, CD3, CD8, CD4 and CD68 markers.

Results

Thirty-nine patients were treated for 46 tumors. These included 32 brain tumors with glioblastoma in 96%, 11 solid tumors, and 3 hematologic cancers. Rapid, objective responses (>50%) were observed in 50% of all tumors. Three year overall survival (OS) for the whole cohort was 48+/-8% which compares favorably with historical controls. Brain tumors fared worse than the other solid tumors, with OS of 39+/-10% and late recurrences were observed even after 2 years of therapy (p=0.02). Tumor burden and acute “flare” constitute poor outcome throughout all cancers. Molecular biomarkers reveal that while all tumors were hypermutant, TMB and predicted neoantigens correlated with response to ICI (p=0.003). Importantly, for tumors with lower TMB, microsatellite instability predicted outcome better than other biomarkers. Immune markers such as PD-L1 and CD8 count on pre-treated tumors predicted response and survival independently. Acute tumor “flare” was associated with increased immune response by RNA inference and TCR in biopsied tumors suggesting acute nonspecific immune response and not true progression.

Conclusions

RRD tumors exhibit favorable outcome and responses to ICI. Genomic and immune signatures of these cancers may provide significant clinical insight on response and treatment decisions.

Background and Aims

This phase 1 study evaluated regorafenib plus vincristine/irinotecan in pediatric patients with rhabdomyosarcoma (RMS) and other solid tumors.

Methods

Patients with relapsed/refractory tumors received intravenous vincristine (1.5 mg/m², Days 1,8) and irinotecan (50 mg/m²/day, Days1–5) plus once-daily oral regorafenib (6–<24 months: 60 mg/m² escalating to 65 mg/m²; 2–<18 years: 72 mg/m² escalating to 82 mg/m²) on either Days 1–14 (concomitant dosing) or Days 8–21 (sequential dosing) during each 21-day cycle. Per protocol, ≥50% of patients had to have RMS.

Results

Of 21 treated patients (RMS, n=12; Ewing sarcoma, n=5; neuroblastoma, n=3; Wilms tumor, n=1), 2 received concomitant (72 mg/m²) and 19 sequential (72 mg/m², n=6; 82 mg/m², n=13) dosing. Median age was 10 years (1.5–17.0). Median number of cycles was 3 (1–17); irinotecan dose reductions occurred in 62%. Grade 3 dose-limiting toxicities were reported with concomitant (peripheral neuropathy and liver injury [1]; pain, vomiting, febrile aplasia [1]) and sequential (rash and elevated aspartate aminotransferase [1]; thrombocytopenia [1]) dosing. Concomitant dosing was discontinued. Maximum tolerated dose and recommended phase 2 dose (RP2D) of regorafenib in sequential combination was 82 mg/m². Most common grade ≥3 toxicities were neutropenia (71%), thrombocytopenia (33%), leukopenia (29%), anemia (24%), and alanine aminotransferase increased (24%). Response rate was 38%, including 1 complete (RMS) and 7 partial responses (5 RMS, 2 Ewing sarcoma); 3 had prior irinotecan. Of 12 patients with RMS, 6 (4 with alveolar subtype) had a response. Nine patients (43%) had stable disease (maximum 17 cycles). After the cut-off, 2 additional patients (1 RMS, 1 Ewing sarcoma) had a partial response.

Conclusions

Regorafenib can be combined at its single-agent RP2D of 82 mg/m² with standard-dose vincristine/irinotecan (with appropriate dose modifications) in pediatric patients with refractory/relapsed solid tumors in a sequential dosing schedule. Activity was observed in patients with sarcoma.

Background and Aims

Proangiogenic signaling pathways cooperate with mTOR-mediated regulation to drive development of many pediatric cancers. We report results of the phase 1 dose escalation for LEN+EVE in pediatric patients with recurrent solid and CNS tumors conducted by Children’s Oncology Group.

Methods

Dose escalation was conducted using a rolling-6 design. Patients received LEN+EVE orally once daily in continuous 28-day cycles. Dose determination was based on toxicity during cycle 1. Pharmacokinetics of LEN and EVE were monitored.

Results

Seventeen patients (median [range] age: 10 [3–21] years; 8 had CNS tumors) were enrolled and evaluable for dose-limiting toxicity (DLT). Enrollment started at dose level 1 (DL1; LEN 11mg/m² + EVE 3mg/m²); after treatment of 3 patients, dose was de-escalated to DL–1 (LEN 8mg/m² + EVE 3mg/m²) due to DLT of proteinuria in 1 patient and self-resolving headache in another which, on review, was not DLT. No patients enrolled at DL–1 (n=5) experienced DLT. DLTs occurred in 2/6 patients enrolled at DL1: the initial patient (proteinuria) and 1 additional patient (hypertriglyceridemia, hypercholesteremia). Because the 2 DLTs were different categories and unrelated to pharmacokinetics, the DL1 cohort was expanded by 6 patients; none experienced DLT. 47% of patients had ≥1 treatment-related adverse events grade ≥3; proteinuria (n=2) was most frequent. On cycle 1 day 15, mean (SD) C_max (ng/mL) for LEN at DL–1 and DL1, respectively, was 314 (150) and 359 (270); mean (SD) AUC_0-8h (hr•ng/mL) for LEN was 1,570 (935) and 1,780 (1,100). No further dose escalation was recommended.

Conclusions

Recommended phase 2 dose of LEN+EVE in pediatric solid and CNS tumors was LEN 11mg/m² + EVE 3mg/m². Toxicity was no more than additive to single-agent therapy. Pharmacokinetic exposure was comparable to pediatric single-agent LEN and adults receiving LEN+EVE. Enrollment to the phase 2 portion is ongoing.