Research in context Evidence before this study We performed a comprehensive PubMed search for articles published in English up to August 1, 2024, using the search terms "radiotherapy" or "radiation therapy" in combination with "survival prediction" or "mortality prediction." This search yielded a total of 345 studies. The majority of these studies focused on survival prediction for specific cancer types, with relatively few addressing survival prediction following radiotherapy more broadly. Most of the identified studies employed statistical models requiring manually structured variables that are not easily extractable from electronic health records (EHRs). Only four studies utilized variables that could be easily extracted from EHRs for survival prediction, but these studies lacked critical information about disease status and overall patient condition, which are typically captured in unstructured EHR data. Instead, they relied on traditional, structured data such as blood test results or national registry information, or small datasets that were manually structured. Notably, no studies employed advanced flexible models, such as large language models (LLMs), to automate the structuring of unstructured data and incorporate it into survival prediction. Added value of this study Our findings suggest the potential of LLMs to process extensive unstructured data, which would be impractical for manual structuring. LLMs demonstrated high accuracy in structuring unstructured data, even without extensive tuning, using a single-shot example approach. Our study is the first to demonstrate that the appropriate application of LLMs can improve the prognosis of patients and the quality of healthcare delivery. Implications of all the available evidence The RT-Surv framework developed in this study has broad applications beyond radiation oncology. As unstructured clinical records form the basis of EHR data across all medical specialties, this framework can be adapted to reduce overall hospital mortality rates, predict length of stay, and assess complication risks. Its ability to automatically structure large volumes of unstructured data enables more accurate and efficient use of clinical data across various domains.

Predicting postoperative risk can inform effective care management & planning. We explored large language models (LLMs) in predicting postoperative risk through clinical texts using various tuning strategies. Records spanning 84,875 patients from Barnes Jewish Hospital (BJH) between 2018 & 2021, with a mean duration of follow-up based on the length of postoperative ICU stay less than 7 days, were utilized. Methods were replicated on the MIMIC-III dataset. Outcomes included 30-day mortality, pulmonary embolism (PE) & pneumonia. Three domain adaptation & finetuning strategies were implemented for three LLMs (BioGPT, ClinicalBERT & BioClinicalBERT): self-supervised objectives; incorporating labels with semi-supervised fine-tuning; & foundational modelling through multi-task learning. Model performance was compared using the AUROC & AUPRC for classification tasks & MSE & R2 for regression tasks. Cohort had a mean age of 56.9 (sd: 16.8) years; 50.3% male; 74% White. Pre-trained LLMs outperformed traditional word embeddings, with absolute maximal gains of 38.3% for AUROC & 14% for AUPRC. Adapting models through self-supervised finetuning further improved performance by 3.2% for AUROC & 1.5% for AUPRC Incorporating labels into the finetuning procedure further boosted performances, with semi-supervised finetuning improving by 1.8% for AUROC & 2% for AUPRC & foundational modelling improving by 3.6% for AUROC & 2.6% for AUPRC compared to self-supervised finetuning. Pre-trained clinical LLMs offer opportunities for postoperative risk predictions with unseen data, & further improvements from finetuning suggests benefits in adapting pre-trained models to note-specific perioperative use cases. Incorporating labels can further boost performance. The superior performance of foundational models suggests the potential of task-agnostic learning towards the generalizable LLMs in perioperative care.

Congenital heart defects (CHD) are the most prevalent birth defects in the United States and surgical outcomes vary considerably across the country. The outcomes of treatment for CHD differ for specific patient subgroups, with non-Hispanic Black and Hispanic populations experiencing higher rates of mortality and morbidity. A valid comparison of outcomes within racial/ethnic subgroups is difficult given large differences in case-mix and small subgroup sizes. We propose a causal inference framework for outcome assessment and leverage advances in transfer learning to incorporate data from both target and source populations to help estimate causal effects while accounting for different sources of risk factor and outcome differences across populations. Using the Society of Thoracic Surgeons' Congenital Heart Surgery Database (STS-CHSD), we focus on a national cohort of patients undergoing the Norwood operation from 2016-2022 to assess operative mortality and morbidity outcomes across U.S. geographic regions by race/ethnicity. We find racial and ethnic outcome differences after controlling for potential confounding factors. While geography does not have a causal effect on outcomes for non-Hispanic Caucasian patients, non-Hispanic Black patients experience wide variability in outcomes with estimated 30-day mortality ranging from 5.9% (standard error 2.2%) to 21.6% (4.4%) across U.S. regions.

Question Can a machine learning algorithm applied to electronic health record data predict patients' short-term risk of death at the time that they begin chemotherapy? Findings In this cohort study of 26 946 patients with cancer starting 51 774 discrete chemotherapy regimens, those at high risk of 30-day mortality were accurately identified across palliative and curative chemotherapy regimens and many types and stages of cancer. The algorithm was more accurate than predictions based on randomized clinical trials or population-based registry data. Meaning A machine learning algorithm accurately identified individuals at high risk of short-term mortality and may help to guide patient and physician decisions about chemotherapy initiation and advance care planning. Importance Patients with cancer who die soon after starting chemotherapy incur costs of treatment without the benefits. Accurately predicting mortality risk before administering chemotherapy is important, but few patient data–driven tools exist. Objective To create and validate a machine learning model that predicts mortality in a general oncology cohort starting new chemotherapy, using only data available before the first day of treatment. Design, Setting, and Participants This retrospective cohort study of patients at a large academic cancer center from January 1, 2004, through December 31, 2014, determined date of death by linkage to Social Security data.