Citations

Abstract Background Ripretinib 150 mg once daily (QD) is indicated for advanced gastrointestinal stromal tumors (GISTs) as at least fourth-line therapy. In INVICTUS, ripretinib intrapatient dose escalation (IPDE) to 150 mg b.i.d. was allowed after progressive disease (PD) on 150 mg QD by blinded independent central review using modified RECIST 1.1. We report the efficacy and safety of ripretinib IPDE to 150 mg b.i.d. after PD among patients randomized to ripretinib 150 mg QD in the INVICTUS study. Materials and Methods Tumor imaging was performed every 28-day cycle for the first four cycles in the ripretinib 150 mg QD period and then every other cycle, including the 150 mg b.i.d. period. Among the ripretinib IPDE patients, progression-free survival (PFS)1 was the time from randomization until PD; PFS2 was the time from the first dose of ripretinib 150 mg b.i.d. to PD or death. Results Among 43 ripretinib IPDE patients, median PFS1 was 4.6 months (95% confidence interval [CI], 2.7–6.4) and median PFS2 was 3.7 months (95% CI, 3.1–5.3). Median overall survival was 18.4 months (95% CI, 14.5–not estimable). Ripretinib 150 mg b.i.d. (median duration of treatment 3.7 months) was well tolerated with new or worsening grade 3–4 treatment-emergent adverse events (TEAEs) of anemia in six (14%) and abdominal pain in three (7%) patients. Ripretinib 150 mg b.i.d. was discontinued because of TEAEs in seven (16%) patients. Conclusion Ripretinib 150 mg b.i.d. after PD on 150 mg QD may provide additional clinically meaningful benefit with an acceptable safety profile in patients with at least fourth-line GISTs. Implications for Practice Of the 85 patients with advanced gastrointestinal stromal tumor having received at least three prior anticancer therapies randomized to ripretinib 150 mg once daily (QD) in the phase III INVICTUS study, 43 underwent ripretinib intrapatient dose escalation (IPDE) to 150 mg b.i.d. after progressive disease (PD). Median progression-free survival was 4.6 months before and 3.7 months after ripretinib IPDE. The safety profile of ripretinib 150 mg b.i.d. was acceptable. These findings indicate ripretinib IPDE to 150 mg b.i.d. may provide additional clinical benefit in patients with PD on ripretinib 150 mg QD, for whom limited treatment options exist.

Abstract Aims We sought to investigate whether artificial intelligence (AI) and specifically deep neural networks (NNs) for electrocardiogram (ECG) signal analysis can be explained using human-selected features. We also sought to quantify such explainability and test if the AI model learns features that are similar to a human expert. Methods and results We used a set of 100 000 ECGs that were annotated by human explainable features. We applied both linear and non-linear models to predict published ECG AI models output for the detection of patients’ age and sex. We further used canonical correlation analysis to quantify the amount of shared information between the NN features and human-selected features. We reconstructed single human-selected ECG features from the unexplained NN features using a simple linear model. We noticed a strong correlation between the simple models and the AI output (R2 of 0.49–0.57 for the linear models and R2 of 0.69–0.70 for the non-linear models). We found that the correlation of the human explainable features with either 13 of the strongest age AI features or 15 of the strongest sex AI features was above 0.85 (for comparison, the first 14 principal components explain 90% of the human feature variance). We linearly reconstructed single human-selected ECG features from the AI features with R2 up to 0.86. Conclusion This work shows that NNs for ECG signals extract features in a similar manner to human experts and that they also generate additional novel features that help achieve superior performance.

PURPOSE CMB305 is a heterologous prime-boost vaccination regimen created to prime NY-ESO-1–specific CD8 T-cell populations and then activate the immune response with a potent TLR-4 agonist. This open-label randomized phase II trial was designed to investigate the efficacy and safety of adding the CMB305 regimen to atezolizumab (anti–programmed death ligand-1 therapy) in comparison with atezolizumab alone in patients with synovial sarcoma or myxoid liposarcoma. PATIENTS AND METHODS Patients with locally advanced, relapsed, or metastatic synovial sarcoma or myxoid liposarcoma (any grade) were randomly assigned to receive CMB305 with atezolizumab (experimental arm) or atezolizumab alone (control arm). The primary end points were progression-free survival (PFS) and overall survival (OS) analyzed using the Kaplan-Meier method. Safety and immune responses were assessed. RESULTS A total of 89 patients were enrolled; 55.1% had received ≥ 2 prior lines of chemotherapy. Median PFS was 2.6 months and 1.6 months in the combination and control arms, respectively (hazard ratio, 0.9; 95% CI, 0.6 to 1.3). Median OS was 18 months in both treatment arms. Patients treated with combination therapy had a significantly higher rate of treatment-induced NY-ESO-1–specific T cells ( P = .01) and NY-ESO-1–specific antibody responses ( P < .0001). In a post hoc analysis of all dosed patients, OS was longer (36 months) in the subset who developed anti-NY-ESO-1 T-cell immune response (hazard ratio, 0.3; P = .02). CONCLUSION Although the combination of CMB305 and atezolizumab did not result in significant increases in PFS or OS compared with atezolizumab alone, some patients demonstrated evidence of an anti-NY-ESO-1 immune response and appeared to fare better by imaging than those without such an immune response. Combining prime-boost vaccines such as CMB305 with anti–programmed death ligand-1 therapies merits further evaluation in other clinical contexts.

Abstract Aims Spectrum bias can arise when a diagnostic test is derived from study populations with different disease spectra than the target population, resulting in poor generalizability. We used a real-world artificial intelligence (AI)-derived algorithm to detect severe aortic stenosis (AS) to experimentally assess the effect of spectrum bias on test performance. Methods and results All adult patients at the Mayo Clinic between 1 January 1989 and 30 September 2019 with transthoracic echocardiograms within 180 days after electrocardiogram (ECG) were identified. Two models were developed from two distinct patient cohorts: a whole-spectrum cohort comparing severe AS to any non-severe AS and an extreme-spectrum cohort comparing severe AS to no AS at all. Model performance was assessed. Overall, 258 607 patients had valid ECG and echocardiograms pairs. The area under the receiver operator curve was 0.87 and 0.91 for the whole-spectrum and extreme-spectrum models, respectively. Sensitivity and specificity for the whole-spectrum model was 80% and 81%, respectively, while for the extreme-spectrum model it was 84% and 84%, respectively. When applying the AI-ECG derived from the extreme-spectrum cohort to patients in the whole-spectrum cohort, the sensitivity, specificity, and area under the curve dropped to 83%, 73%, and 0.86, respectively. Conclusion While the algorithm performed robustly in identifying severe AS, this study shows that limiting datasets to clearly positive or negative labels leads to overestimation of test performance when testing an AI algorithm in the setting of classifying severe AS using ECG data. While the effect of the bias may be modest in this example, clinicians should be aware of the existence of such a bias in AI-derived algorithms.

Abstract Background: Letetresgene autoleucel (lete-cel; GSK3377794) is an autologous T-cell therapy expressing a genetically modified T-cell receptor (TCR) to improve recognition of cancer cells expressing NY-ESO-1 and/or LAGE-1a. Next generation NY-ESO-1 TCR T-cell therapies, including GSK3901961 and GSK3845097, incorporate further genetic modifications to enhance anticancer activity. GSK3901961 co-expresses the CD8α chain to stabilize TCR-human leukocyte A (HLA) class I interactions on CD4+ T cells, enhancing T-cell persistence and increasing helper functions such as Type 1 T-helper anti-tumor responses. GSK3845097 co-expresses a dominant negative transforming growth factor-β (TGF-β) type II receptor to reduce TGF-β pathway activation and maintain T-cell proliferation, cytokine production, and cytotoxicity in the tumor microenvironment. A first-time-in-human master protocol (NCT04526509) will evaluate the safety, tolerability, and recommended Phase 2 dose (RP2D) of these two therapies and possible subsequent ones. Substudy 1 will assess GSK3901961 in patients with advanced non-small cell lung cancer (NSCLC) or synovial sarcoma (SS). Substudy 2 will assess GSK3845097 in patients with advanced SS. Methods: Each substudy includes a dose confirmation stage to assess RP2D and a dose expansion stage. Table 1 lists eligibility criteria. Primary endpoints are safety (adverse events) and tolerability (dose-limiting toxicities). Secondary endpoints include investigator-assessed overall response rate, duration of response, and maximum expansion/persistence and phenotype of infiltrating transduced T cells. Exploratory endpoints include laboratory parameters, overall survival, and anti-GSK3901961 and anti-GSK3845097 titers for the respective substudies. The substudies are open and recruiting. Funding: GSK (209012; NCT04526509) Initial screening criteriaInclusion criteriaExclusion criteriaSubstudies 1 and 2 (SS and NSCLC)≥18 years of agePrior malignancy that is not in complete remission or clinically significant systemic illnessMeasurable disease per RECIST v1.1 criteriaPrevious treatment with genetically modified NY-ESO-1-specific T cells, NY-ESO-1 vaccine, or NY-ESO-1 targeting antibodyExpression of HLA-A*02:01, A*02:05, or A*02:06Prior gene therapy using an integrating vectorExpression of NY-ESO-1/LAGE-1a in tumor archival or fresh biopsyPrevious allogeneic hematopoietic stem cell transplant within the last 5 years or solid organ transplantSubstudies 1 and 2 (SS only)Histologically confirmed advanced (metastatic or unresectable) SS diagnosisCentral nervous system metastasesPresence of t(X;18) translocation(Allowed for NSCLC participants on a case-by-case basis)Received, completed, or intolerant to treatment with anthracycline or anthracycline with ifosfamide for advanced (metastatic or unresectable) disease and has progressedSubstudy 1 (NSCLC only)Histologically or cytologically confirmed Stage IV NSCLCReceived or failed ≥3 lines of systemic therapyReceiving or previously received ≥1 prior line(s) of standard of care treatment including programmed death receptor-1/programmed death ligand-1 checkpoint blockade therapy, and received or be intolerant to doublet taxane and platinum chemotherapyPresence of actionable genetic aberration (activation epithelial growth factor receptor, anaplastic lymphoma kinase/c-ros oncogene 1) per NCCN guidelines Citation Format: Adam J. Schoenfeld, Mehmet Altan, Taofeek K. Owonikoko, Sandra D'Angelo, Brian H. Ladle, Jonathan Noujaim, Kai He, David Liebner, Adrian G. Sacher, John B.A.G. Haanen, Jeffrey Yachnin, Chao Huang, Brian A. Van Tine, Aisha Hasan, Thomas Faitg, Emily Butler, Aiman Shalabi, Steven Attia, Dejka M. Araujo. Master protocol to assess the safety and recommended Phase 2 dose of next generation NY-ESO-1-specific TCR T-cells in HLA-A*02 patients with synovial sarcoma and non-small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr CT219.

11536 Background: Ripretinib is a switch-control tyrosine kinase inhibitor that broadly inhibits KIT and PDGFRA kinase signaling. In the INVICTUS study (NCT03353753), patients with advanced GIST (≥4th-line) receiving ripretinib had a median progression-free survival (mPFS) of 6.3 months vs 1.0 month for patients receiving placebo (HR = 0.15, p <0.0001). In an earlier phase 1 dose escalation study, the maximum tolerated dose was not reached with doses up to 200 mg twice daily (BID). Here, we report efficacy, safety, and pharmacokinetic data for IPDE patients initially randomized to ripretinib from the phase 3 INVICTUS study based on data as of 10 Aug 2020. Methods: A total of 129 patients were randomized 2:1 to ripretinib 150 mg once daily (QD; n = 85) or placebo (n = 44). Patients receiving ripretinib 150 mg QD who had progressive disease (PD) as assessed by blinded independent central review (BICR) using mRECIST were given the option for IPDE to 150 mg BID. Tumor response assessments were performed every 28 days for 4 cycles and every 56 days thereafter (including after IPDE). The primary endpoint was PFS. For this exploratory analysis, PFS1 for IPDE patients was defined as the time from randomization to PD; PFS2 for IPDE patients was the time from the first dose of ripretinib 150 mg BID to PD or death. PFS1 and PFS2 were based on BICR. Results: Of 85 patients treated with ripretinib 150 mg QD, 43 patients with BICR PD dose escalated to 150 mg BID. Baseline characteristics of IPDE patients at time of study entry were similar to those observed in the original ripretinib QD arm and similar to the 22 patients with BICR PD who either remained on 150 mg QD or discontinued treatment. IPDE patients had a mPFS1 of 4.6 months (95% CI, 2.7–6.4) and a mPFS2 of 3.7 months (95% CI, 3.1–5.3); mPFS2/mPFS1=80%. The IPDE dosing period was well tolerated without the emergence of new safety concerns. The most common new or worsening (unchanged or improving grades not included) treatment-emergent adverse event (TEAE) for IPDE patients during the BID period was abdominal pain (30.2% all grades and 7% Grade 3–4 vs 41.9% and 4.7% in QD period, respectively). The most common Grade 3–4 TEAE in the BID period was anemia (14% vs 2.3% in QD period). IPDE from QD to BID resulted in an approximately 2-fold increase in the steady state trough concentration. Conclusions: Similar to the phase 1 study wherein IPDE to 150 mg BID following PD provided clinical benefit with a mPFS1 of 5.5 months and mPFS2 of 4.6 months (mPFS2/mPFS1=84%) for patients with ≥4th-line GIST, these analyses from INVICTUS indicate that IPDE to ripretinib 150 mg BID can also provide additional meaningful clinical benefit and a similar tolerability profile to the 150 mg QD regimen for patients with ≥4th-line GIST that progressed following treatment with ripretinib 150 mg QD. Clinical trial information: NCT03353753.

11517 Background: Doxorubicin (doxo) remains standard first-line therapy for advanced STS. Doxo in combination with olaratumab (O) demonstrated superior clinical activity compared to doxo alone in a Ph 2 trial (NCT01185964), although this was not confirmed in the subsequent Ph 3 trial (NCT02451943). Gemcitabine (G) plus docetaxel (D) is a second line therapy for advanced STS. Here, we report a concurrent Ph 2 study that explored a second-line addition of O to G and D for advanced STS (ANNOUNCE 2 NCT02659020). Methods: Adult patients (pts) with unresectable locally advanced or metastatic STS, ≤ 2 prior lines of systemic therapy, and ECOG PS 0-1 were eligible. Pts were enrolled from 2 cohorts: O-naïve and O-pretreated. In both cohorts, pts were randomized 1:1 to either O, G plus D or placebo (PBO), G plus D. Pts received 21-day cycles of O (20 mg/ kg cycle 1 and 15 mg/kg other cycles, day (d) 1 and d8), G (900 mg/m2, d1 and d8) and D (75 mg/m2, d8). Pts continued treatment until progression, toxicity, or withdrawal. Randomization was stratified by histology (leiomyosarcoma [LMS] vs non-LMS), prior systemic therapy, ECOG PS, and prior pelvic radiation. The primary objective was overall survival (OS) in the O-naïve population using an alpha level of 0.20. Secondary endpoints included OS (O-pretreated) and other efficacy parameters, as well as safety and pharmacokinetics (PK). Results: 167 pts were enrolled in the O-naïve cohort and 89 pts in the O-pretreated cohort. Baseline patient characteristics were well balanced. OS for O-naïve pts was 16.8 vs 18.0 months (m) (hazard ratio [HR] = 0.95, 95% CI: 0.64-1.40; p = 0.78) for the investigational vs control arm, respectively. Other efficacy outcomes are presented in the table. Safety was manageable across treatment arms. PK parameter estimates for O were consistent with previous studies. Conclusions: There was no statistically significant difference in OS between the two arms in the O-naïve population. However, while not statistically significant, the combination of O, G and D demonstrated favorable OS in the O-pretreated cohort, and PFS and objective response rate (ORR) in both cohorts. For O-naïve pts, a clinically meaningful progression-free survival (PFS) improvement was observed. Further investigations in specific histological subtypes are ongoing. Clinical trial information: NCT02659020. [Table: see text]

11506 Background: Uterine leiomyosarcoma (uLMS) is an aggressive sarcoma subtype with frequent metastatic relapse. After failure of front-line chemotherapy, remaining options provide limited benefit (trabectedin: ORR 11%, mPFS 4.0 mos; pazopanib: ORR 11%, median PFS 2.9 mos; dacarbazine: ORR 9%, mPFS 1.5 mos). Recent molecular studies suggest uLMS harbors characteristic defects in the homologous recombination (HR) DNA repair pathway, including somatic biallelic BRCA2 deletion in 10%, implicating potential sensitivity to PARP-inhibitor based treatment approaches. In preclinical uLMS models in which temozolomide (T, an oral equivalent to dacarbazine) or the PARP inhibitor olaparib (O) showed limited single agent activity, the combination of T + O was highly effective in inhibiting uLMS tumor growth and promoting apoptosis. Methods: NCI protocol #10250 is a single-arm, open-label, multi-center phase II study evaluating T + O in advanced uLMS. Pts had progression on ≥ 1 prior line and ECOG PS ≤ 2. Pts received T 75 mg/m2 PO daily + O 200 mg PO BID days 1-7 in 21-day cycles. Primary endpoint was ORR. A one-stage binomial design was used. If ≥ 5/22 responded, the treatment was considered promising (95% power; α = 0.06). All pts underwent paired tumor biopsies. Correlative assays to evaluate for HR deficiency (whole exome sequencing/RNAseq, RAD51 foci formation) and for intrinsic PARP inhibitor resistance (SLFN11 expression) will be correlated with response. Results: 22 patients were evaluable (median age 55, median prior treatment lines 3). Median follow-up was 10.8 mos. Primary endpoint, ORR within 6 mos of initiating treatment, was 23% (5/22). Overall ORR was 27% (6/22). Median PFS was 6.9 mos (95% CI: 5.4 mos – not estimable (NE)). Median duration of response (DOR) was 12.0 mos (95% CI: 9.5 mos – NE). Hematologic toxicity was common (grade 3/4 neutropenia, 77%; thrombocytopenia 32%) but manageable with dose modification. Correlative assays are ongoing with plans to present at the meeting. An immunohistochemical assay for RAD51 foci has been adapted for uLMS samples and clearly distinguishes BRCA2- deleted and wild-type tumors. Conclusions: NCI 10250 met the prespecified primary efficacy endpoint of ORR in a population of patients with heavily pre-treated uLMS. Responses are durable (median DOR 12 mos). Correlative assays are being completed to evaluate whether uLMS tumors with HR deficiency or with preserved SLFN11 expression are most sensitive to T + O and may underlie durable responses. A randomized study of the combination is planned. Clinical trial information: NCT03880019.

11505 Background: AL3818 (Catequentinib, Anlotinib) is a novel, orally administered, small molecule tyrosine kinase inhibitor. The primary objective of this Phase 3 study is to evaluate the efficacy of AL3818 monotherapy in patients (pts) with synovial sarcoma (SS) comparing with dacarbazine in randomization setting. Methods: Patients with a diagnosis of synovial sarcoma requiring second line or further line treatment were eligible for enrollment. The regimen was a 21-day cycle with oral AL3818 administered on 14 days on and 7 days off. This phase 3 trial is randomized in 2:1 ratio of AL3818 comparing to dacarbazine with option of crossover after PD of dacarbazine treatment. Progression free survival (PFS) with Log Rank test is the primary endpoint and this trial for SS is currently completed enrolled in US and Italy. Results: Total 79 pts initiated treatment and are evaluable, 52 received AL3818 as treatment arm (T), and 27 received dacarbazine (D) as control arm (C). Arms T/C median ages were 40.5/42.0 years (range: 18-70+) and 20/16 (38.5%/59.3%) were male. Overall, PFS was 2.89 months (95% CI: 2.73 – 6.87) for AL3818 and 1.64 (95% CI: 1.45 – 2.70) for D. The PFS of study met the primary endpoint with a p-value of 0.0015 and a HR of 0.449 (95% CI: 0.270– 0.744). At the month 4, 6, and 12, the percentages of progression free patients for AL3818 were 48.1%, 42.3% and 26.9%; and for D were 14.85%, 11.1% and 3.7%. For grade 3 treatment-related adverse events, 12(23.1%) of pts experienced for AL3818 and 7(25.9%) of pts experienced for D. The most common AL3818 related grade 3 AEs were diarrhea (5.8%) and hypertension (3.8%). Conclusions: This phase III trial demonstrates improved disease control and superior progression free survival for AL3818 vs dacarbazine in advanced SS. In addition, the study further confirms the acceptable benefit-risk profile of AL3818 from the prior randomized Phase 2b soft tissue sarcoma study (NCT02449343). AL3818 is a meaningful treatment option for pts with advanced SS. Clinical trial information: NCT 03016819 Clinical trial information: NCT03016819.