In this phase 1 study, azacitidine (AZA) was given before high-dose cytarabine (HiDAC) and mitoxantrone (mito) based on the hypothesis that epigenetic priming with a hypomethylating agent before cytotoxic chemotherapy would improve response rates in patients with high-risk acute myeloid leukemia (AML), including relapsed/refractory disease. The primary objective was to establish the recommended phase 2 dose of AZA given before standard HiDAC/mito. In a dose escalation scheme, 46 patients (median age, 66 years) received AZA at 37.5, 50, or 75 mg/m2 subcutaneously or IV once daily on days 1 to 5 followed by HiDAC (3000 mg/m2) and mitoxantrone (30 mg/m2) once each on days 6 and 10 (the HiDAC/mito dose was reduced 33% in elderly subjects). Two dose-limiting toxicities occurred (both in the same patient): acute liver failure and kidney injury at the 50 mg/m2 dose. The 30-day induction death rate was 2.2% (1 of 46). The overall response rate, including complete remission and complete remission with incomplete count recovery, was 61% (28 of 46). Previously untreated patients aged ≥60 years with therapy-related AML and de novo AML were more likely to respond than untreated patients with AML progressing from an antecedent hematologic disorder (myelodysplastic syndrome and chronic myelomonocytic leukemia). Patients with favorable European Leukemia Network risk (P = .008), NPM1 mutations (P = .007), or IDH2 mutations (P = .03) were more likely to respond, and those with TP53 mutations (P = .03) were less likely to respond. The recommended phase 2 dose of AZA is 75 mg/m2 per day on days 1 to 5 followed by HiDAC (3000 mg/m2) and mitoxantrone (30 mg/m2) once each on days 6 and 10. This trial was registered at www.clinicaltrials.gov as #NCT01839240.

PURPOSE Large-scale analysis of real-world evidence is often limited to structured data fields that do not contain reliable information on recurrence status and disease sites. In this report, we describe a natural language processing (NLP) framework that uses data from free-text, unstructured reports to classify recurrence status and sites of recurrence for patients with breast and hepatocellular carcinomas (HCC). METHODS Using two cohorts of breast cancer and HCC cases, we validated the ability of a previously developed NLP model to distinguish between no recurrence, local recurrence, and distant recurrence, based on clinician notes, radiology reports, and pathology reports compared with manual curation. A second NLP model was trained and validated to identify sites of recurrence. We compared the ability of each NLP model to identify the presence, timing, and site of recurrence, when compared against manual chart review and International Classification of Diseases coding. RESULTS A total of 1,273 patients were included in the development and validation of the two models. The NLP model for recurrence detects distant recurrence with an area under the curve of 0.98 (95% CI, 0.96 to 0.99) and 0.95 (95% CI, 0.88 to 0.98) in breast and HCC cohorts, respectively. The mean accuracy of the NLP model for detecting any site of distant recurrence was 0.9 for breast cancer and 0.83 for HCC. The NLP model for recurrence identified a larger proportion of patients with distant recurrence in a breast cancer database (11.1%) compared with International Classification of Diseases coding (2.31%). CONCLUSION We developed two NLP models to identify distant cancer recurrence, timing of recurrence, and sites of recurrence based on unstructured electronic health record data. These models can be used to perform large-scale retrospective studies in oncology.

7505 Background: MCL is a rare lymphoma without a standard of care but several regimens have demonstrated clinical activity, the majority based on traditional chemotherapy. We hypothesized that adding venetoclax (V) to R2 would be safe and effective in MCL pts irrespective of age, morphology or stage. Here we present safety and efficacy data from the on-going phase 1b study of R2 + V in pts with newly diagnosed MCL. Methods: This multi-center phase 1 study (NCT03523975) enrolled pts aged ≥18 yrs with untreated MCL. The primary objective was to characterize the safety and tolerability of R2 + V and determine the MTD. During induction (12 months (m)) pts received lenalidomide (L) 20 mg daily on day 1-21, Rituximab (R) was given weekly during c1 then on day 1 of every even cycle, V was escalated over 4 weeks to 400 mg beginning day 8. Each cycle is 28 days (d). The DLT period was 42 d beginning C1D8. In maintenance, R every 8 weeks for 36m, L at 10 mg or half of last dose during induction for 24 m and V for minimum 12 m. No pts have been transplanted. Pts with progression (PD) came off study. MRD was analyzed in parallel with scans during induction by clonoSEQ assay (Adaptive Biotechnologies). Results: As of Feb. 1st, 2021, we have enrolled all 28 planned pts on study. Pt characteristics/responses are summarized in Table. Among the 28 pts who have received at least one dose, the median treatment duration so far is 278d (IQR 170-560), with 24 pts still on treatment (Tx). 1 pt is off from a unrelated condition. All pts escalated to V 400 mg w/o any DLTs noted. Treatment-emergent adverse events (TEAEs) were reported in 100% of pts, and grade 3+ TEAEs were reported in 26 (93%) patients. The most common all-grade TEAEs (≥50% of pts), regardless of relationship to study Tx, were fatigue, neutropenia and diarrhea. Grade ≥3 TEAEs reported in ≥50% pts were neutropenia (68%) and thrombocytopenia (50%). No pts have withdrawn or d/c Tx due to AEs. There was one grade 5 event, in a non-evaluable pt, related to a PE that occurred prior to DLT period. In the 28 evaluable pts the ORR (CR/PR) was 96% (27/28 pts) with CR/CRu of 89%. Of the responding pts, two had PD, one w/ CR and one w/ PR. All pts with PD had baseline TP53 mutation. MRD testing was successful in all pts. At time of submission 20 of 28 (71%) were MRD - at 10-6. Conclusions: Interim results show that at the MTD the combination of V 400 mg daily, L 20 mg, with R is safe with a manageable toxicity profile and a high ORR and MRD - in pts with newly diagnosed MCL. Safety data is consistent with the AE profile noted for each drug without any unexpected or unique AEs. Updated results including BH3 profiling will be presented at the meeting. Clinical trial information: NCT03523975. [Table: see text]

7523 Background: Outcomes of pts with newly diagnosed DLBCL with high/poor risk according to the revised International Prognostic Index (IPI) treated with immunochemotherapy (IC) remain suboptimal, with a 4-year survival rate of 55% (Sehn et al, Blood 2007). SC epco is a well-tolerated bispecific antibody with single-agent activity in the relapsed/refractory (R/R) aggressive B-cell NHL setting. The mechanism of action and safety profile of epco are distinct from IC, and epco is well suited for use in combinations and in earlier lines of therapy. Addition of a novel agent to standard of care IC may overcome the adverse prognosis of high-risk pts. Presented here are updated results of epco + rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) in previously untreated DLBCL pts with IPI 3–5 (EPCORE NHL-2 arm 1; NCT04663347). Methods: Adults with previously untreated CD20+ DLBCL and IPI ≥3 received SC epco (every week, cycle [C] 1–4; every 3 weeks, C5–6) + R-CHOP for 6 cycles (21 d) followed by single-agent epco every 4 weeks up to 1 y (in cycles of 28 d). To mitigate CRS, epco step-up dosing and corticosteroid prophylaxis were required. Response was assessed by PET-CT with contrast per Lugano 2014 criteria. Results: As of Dec 1, 2021, 33 pts (median age, 66 y; range, 19–82) received epco + R-CHOP (epco 24 mg, n = 4; 48 mg, n = 29). Median time from diagnosis to first dose was 25 d (range, 5–70). All pts had IPI ≥3 and ≥24% had double- or triple-hit DLBCL. Median follow-up was 3 mo (range, 0–9.7); median number of total cycles initiated was 5 (1–13). Overall, 94% of pts (31/33) remained on Tx. Tx-emergent adverse events (TEAEs) in ≥30% of pts were neutropenia (48%; febrile neutropenia in 9% of all pts), CRS (45%), infections (42%), anemia (39%), injection-site reactions (36%), nausea (33%), constipation (30%), and pyrexia (30%). No TEAEs led to epco discontinuation. AEs of special interest included CRS (42% grade [G] 1/2, 3% G3) and ICANS (3% G2); 1 pt had tumor lysis syndrome (3% G3). Most CRS events occurred in C1 and resolved after a median of 2 days (1–11); 4 pts with CRS received tocilizumab. No fatal TEAEs were reported. In efficacy-evaluable pts, the overall response rate (ORR) was 96% (24/25); 68% (17/25) had complete metabolic response (CMR). For the 10 pts who received 6 cycles of R-CHOP and had a subsequent response assessment, the ORR and CMR rate were 100% and 90%, respectively. As of the data cutoff, all of these 10 pts remained in response, with the longest duration of response 7.1+ mo and ongoing. Updated data will be presented. Conclusions: Epco is the first SC bispecific antibody assessed in combination with standard of care in previously untreated DLBCL. The safety profile of epco + R-CHOP is manageable. CRS events were mostly of low grade and did not lead to Tx discontinuation. ORR and CMR rate were high with no relapses as of the data cutoff date. Clinical trial information: NCT04663347.