Yee, Andrew J.

As machine learning (ML)-based decision support tools proliferate in clinical practice, understanding how clinicians integrate personalized ML predictions alongside randomized controlled trial (RCT) evidence is critical. We designed a web-based clinical decision support system (CDSS) presenting survival and adverse event data from a simulated RCT and ML model across 12 synthetic multiple myeloma scenarios. In a within-subjects study with 32 physicians, we evaluated how clinicians synthesize competing evidence sources to make treatment decisions. When ML and RCT outputs were concordant, physicians reported greater confidence than with RCT data alone. When results were discordant, most physicians shifted toward the ML-supported treatment, often before reviewing any information about model training or validation, suggesting a tendency toward automation bias rather than algorithm avoidance. Despite reporting higher perceived reliability after viewing model quality disclosures, physicians were largely unable to describe the validation procedures they had reviewed. Taken together, these findings reveal that clinicians may over-rely on ML recommendations even when equipped with tools designed to support critical appraisal. We discuss implications for CDSS design, clinician training, and the institutional safeguards needed before ML-based systems are deployed in high-stakes oncology settings.

Purpose Selinexor, a first-in-class, oral, selective exportin 1 (XPO1) inhibitor, induces apoptosis in cancer cells through nuclear retention of tumor suppressor proteins and the glucocorticoid receptor, along with inhibition of translation of oncoprotein mRNAs. We studied selinexor in combination with low-dose dexamethasone in patients with multiple myeloma refractory to the most active available agents. Patients and Methods This phase II trial evaluated selinexor 80 mg and dexamethasone 20 mg, both orally and twice weekly, in patients with myeloma refractory to bortezomib, carfilzomib, lenalidomide, and pomalidomide (quad-refractory disease), with a subset also refractory to an anti-CD38 antibody (penta-refractory disease). The primary end point was overall response rate (ORR). Results Of 79 patients, 48 had quad-refractory and 31 had penta-refractory myeloma. Patients had received a median of seven prior regimens. The ORR was 21% and was similar for patients with quad-refractory (21%) and penta-refractory (20%) disease. Among patients with high-risk cytogenetics, including t(4;14), t(14;16), and del(17p), the ORR was 35% (six of 17 patients). The median duration of response was 5 months, and 65% of responding patients were alive at 12 months. The most common grade ≥ 3 adverse events were thrombocytopenia (59%), anemia (28%), neutropenia (23%), hyponatremia (22%), leukopenia (15%), and fatigue (15%). Dose interruptions for adverse events occurred in 41 patients (52%), dose reductions occurred in 29 patients (37%), and treatment discontinuation occurred in 14 patients (18%). Conclusion The combination of selinexor and dexamethasone has an ORR of 21% in patients with heavily pretreated, refractory myeloma with limited therapeutic options.