AI Systems Advance Treatment and Care for Blood Cancer Patients

Artificial intelligence is transforming palliative oncology for patients with hematologic malignancies by improving symptom management, pain classification, and mortality prediction. Clinical trials by researchers like Wilson and Manz demonstrate that AI-driven decision support tools and machine learning nudges successfully increase palliative care consultations, reduce hospital readmissions, and facilitate timely end-of-life conversations.

AI in palliative oncology

• ▪The medical community remains cautious about using artificial intelligence tools in practice due to unfamiliarity, accuracy concerns, and ethical considerations.
• ▪Artificial intelligence enhances prognostication, symptom management, and personalized care for patients with hematologic malignancies in palliative oncology.

Symptom management tools

• ▪Artificial intelligence tools such as natural language processing can analyze electronic health records, text, and speech patterns to detect emotional distress.
• ▪Machine learning models developed by Masukawa and colleagues detected social distress with an AUROC of 0.98 and spiritual distress with an AUROC of 0.90 in terminally ill cancer patients.

Pain assessment models

• ▪Random forest models perform well in predicting pain intensity, opioid responsiveness, and the risk of chronic pain, outperforming traditional regression models.
• ▪Automated pain assessment using artificial intelligence utilizes facial expressions, language, posture, and neurophysiological signals to standardize and improve subjective self-reported pain.

Mortality prediction systems

• ▪A prospective study at Taipei Medical University Hospital monitored hand movements of 68 hospice patients using wearable devices to evaluate survival outcomes.
• ▪Liu and colleagues explored using artificial intelligence and wearable devices to predict 7-day mortality in terminally ill cancer patients.

Clinical trial evidence

• ▪A clinical trial by Wilson and colleagues showed that artificial intelligence-based decision support tools increased palliative care consultations and reduced hospital readmissions.
• ▪A clinical trial by Manz and colleagues demonstrated that machine learning-triggered behavioral nudges improved serious illness conversations and reduced systemic therapy use near the end of life.

2 sources