Citations

Objectives: To develop and verify a digital twin model of critically ill patient using the causal artificial intelligence approach to predict the response to specific treatment during the first 24 hours of sepsis. Design: Directed acyclic graphs were used to define explicitly the causal relationship among organ systems and specific treatments used. A hybrid approach of agent-based modeling, discrete-event simulation, and Bayesian network was used to simulate treatment effect across multiple stages and interactions of major organ systems (cardiovascular, neurologic, renal, respiratory, gastrointestinal, inflammatory, and hematology). Organ systems were visualized using relevant clinical markers. The application was iteratively revised and debugged by clinical experts and engineers. Agreement statistics was used to test the performance of the model by comparing the observed patient response versus the expected response (primary and secondary) predicted by digital twin. Setting: Medical ICU of a large quaternary- care academic medical center in the United States. Patients or Subjects: Adult (> 18 year yr old), medical ICU patients were included in the study. Interventions: No additional interventions were made beyond the standard of care for this study. Measurements and Main Results: During the verification phase, model performance was prospectively tested on 145 observations in a convenience sample of 29 patients. Median age was 60 years (54–66 d) with a median Sequential Organ Failure Assessment score of 9.5 (interquartile range, 5.0–14.0). The most common source of sepsis was pneumonia, followed by hepatobiliary. The observations were made during the first 24 hours of the ICU admission with one-step interventions, comparing the output in the digital twin with the real patient response. The agreement between the observed versus and the expected response ranged from fair (kappa coefficient of 0.41) for primary response to good (kappa coefficient of 0.65) for secondary response to the intervention. The most common error detected was coding error in 50 observations (35%), followed by expert rule error in 29 observations (20%) and timing error in seven observations (5%). Conclusions: We confirmed the feasibility of development and prospective testing of causal artificial intelligence model to predict the response to treatment in early stages of critical illness. The availability of qualitative and quantitative data and a relatively short turnaround time makes the ICU an ideal environment for development and testing of digital twin patient models. An accurate digital twin model will allow the effect of an intervention to be tested in a virtual environment prior to use on real patients.

Purpose Endoxifen is a tamoxifen metabolite with potent antiestrogenic activity. Patients and Methods We performed a phase I study of oral Z-endoxifen to determine its toxicities, maximum tolerated dose (MTD), pharmacokinetics, and clinical activity. Eligibility included endocrine-refractory, estrogen receptor–positive metastatic breast cancer. An accelerated titration schedule was applied until moderate or dose-limiting toxicity occurred, followed by a 3+3 design and expansion at 40, 80, and 100 mg per day. Tumor DNA from serum (circulating cell free [cf); all patients] and biopsies [160 mg/day and expansion]) was sequenced. Results Of 41 enrolled patients, 38 were evaluable for MTD determination. Prior endocrine regimens during which progression occurred included aromatase inhibitor (n = 36), fulvestrant (n = 21), and tamoxifen (n = 15). Patients received endoxifen once daily at seven dose levels (20 to 160 mg). Dose escalation ceased at 160 mg per day given lack of MTD and endoxifen concentrations > 1,900 ng/mL. Endoxifen clearance was unaffected by CYP2D6 genotype. One patient (60 mg) had cycle 1 dose-limiting toxicity (pulmonary embolus). Overall clinical benefit rate (stable > 6 months [n = 7] or partial response by RECIST criteria [n = 3]) was 26.3% (95% CI, 13.4% to 43.1%) including prior tamoxifen progression (n = 3). cfDNA mutations were observed in 13 patients ( PIK3CA [n = 8], ESR1 [n = 5], TP53 [n = 4], and AKT [n = 1]) with shorter progression-free survival ( v those without cfDNA mutations; median, 61 v 132 days; log-rank P = .046). Clinical benefit was observed in those with ESR1 amplification (tumor; 80 mg/day) and ESR1 mutation (cfDNA; 160 mg/day). Comparing tumor biopsies and cfDNA, some mutations ( PIK3CA, TP53, and AKT) were undetected by cfDNA, whereas cfDNA mutations ( ESR1, TP53, and AKT) were undetected by biopsy. Conclusion In endocrine-refractory metastatic breast cancer, Z-endoxifen provides substantial drug exposure unaffected by CYP2D6 metabolism, acceptable toxicity, and promising antitumor activity.