John-Jose Nunez | Nunez Lab for Computational Mental Health & Cancer Care

Investigating fine-tuning versus zero-shot learning for general large language models when predicting cancer survival from initial oncology consultation documents

Thu, 01 Jan 2026 00:00:00 +0000

This study investigates whether fine-tuning open-weight large language models (LLMs) improves cancer survival prediction from oncology consultation notes compared to zero-shot approaches. Using Meta’s Llama models on 59,800 patient records from BC Cancer, fine-tuning consistently improved performance over zero-shot inference, but did not outperform smaller task-specific NLP models. The findings suggest both approaches merit continued investigation, with the best choice depending on clinical context and practical constraints such as hardware, privacy, and deployment feasibility.

Patients' Attitudes toward Artificial Intelligence (AI) in Cancer Care: A Scoping Review Protocol

Wed, 15 Jan 2025 00:00:00 +0000

This scoping review protocol explores how patients with cancer perceive and accept artificial intelligence in their medical care, addressing an important gap in understanding patient perspectives on AI implementation in oncology.

A Randomized Evaluation of MoodFX, a Patient-Centred e-Health Tool to Support Outcome Measurement for Depression

Mon, 01 Jul 2024 00:00:00 +0000

MoodFX is a patient-centered e-health tool designed to support outcome measurement for depression. This randomized evaluation demonstrates its effectiveness in helping patients track symptoms and improve treatment outcomes.

Predicting Which Patients with Cancer Will See a Psychiatrist or Counsellor from Their Initial Oncology Consultation Document Using Natural Language Processing

Tue, 09 Apr 2024 00:00:00 +0000

This groundbreaking study demonstrates that artificial intelligence can predict psychosocial needs of cancer patients by analyzing initial oncology consultation documents. Different linguistic patterns predict psychiatrist versus counselor referrals, providing clinically relevant insights for patient care.

Machine Learning Prediction of Quality of Life Improvement During Antidepressant Treatment of Patients With Major Depressive Disorder: A STAR*D and CAN-BIND-1 Report

Wed, 01 Nov 2023 00:00:00 +0000

This study demonstrates that machine learning can predict quality of life improvements in depression treatment, supporting the use of early clinical indicators to identify patients likely to benefit from antidepressant therapy in terms of functional outcomes.

Response Trajectories during Escitalopram Treatment of Patients with Major Depressive Disorder

Fri, 01 Sep 2023 00:00:00 +0000

This study reveals distinct response trajectories to escitalopram treatment in depression by applying unsupervised machine learning to clinical data. The findings highlight that subjective mood and anhedonia are central to treatment response, while other symptom domains show more variable patterns.

Predicting the Survival of Patients With Cancer From Their Initial Oncology Consultation Document Using Natural Language Processing

Mon, 27 Feb 2023 00:00:00 +0000

This study demonstrates that natural language processing applied to initial oncology consultation documents can effectively predict patient survival outcomes across multiple cancer types. The models perform comparably with or better than previous prediction models while using readily available clinical data.

Supportive Care and Health Care Service Utilization in Older Adults with a New Cancer Diagnosis: A Population Based Review in British Columbia, Canada

Sun, 01 Jan 2023 00:00:00 +0000

This publication examines supportive care and health care service utilization patterns in older adults with new cancer diagnoses in British Columbia.

Replication of Machine Learning Methods to Predict Treatment Outcome with Antidepressant Medications in Patients with Major Depressive Disorder from STAR*D and CAN-BIND-1

Tue, 01 Jun 2021 00:00:00 +0000

This replication and external validation study demonstrates that machine learning methods can successfully predict antidepressant treatment outcomes, with prediction of remission performing better than prediction of response. These findings support the clinical utility of machine learning approaches for personalized depression treatment planning.