# Machine learning predicting acute pain and opioid dose in radiation treated oropharyngeal cancer patients

## TL;DR

This study uses machine learning to predict pain and opioid needs in cancer patients undergoing radiation therapy, helping improve personalized pain management.

## Contribution

The study introduces ML models to predict acute pain and opioid doses in oropharyngeal cancer patients during radiation therapy.

## Key findings

- Gradient Boosting Machine (GBM) showed best performance for predicting pain intensity (AUROC 0.71).
- Logistic Regression (LR) outperformed other models in predicting total opioid doses (AUROC 0.67).
- Baseline pain scores and vital signs were identified as the most important predictive features.

## Abstract

Acute pain is common among oral cavity/oropharyngeal cancer (OCC/OPC) patients undergoing radiation therapy (RT). This study aimed to predict acute pain severity and opioid doses during RT using machine learning (ML), facilitating risk-stratification models for clinical trials.

A retrospective study examined 900 OCC/OPC patients treated with RT during 2017–2023. Pain intensity was assessed using NRS (0-none, 10-worst) and total opioid doses were calculated using morphine equivalent daily dose (MEDD) conversion factors. Analgesics efficacy was assessed using combined pain intensity and total MEDD. ML predictive models were developed and validated, including Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), and Gradient Boosting Machine (GBM). Model performance was evaluated using discrimination and calibration metrics, while feature importance was investigated using bootstrapping.

For predicting pain intensity, the GBM demonstrated superior discrimination performance (AUROC 0.71, recall 0.39, and F1 score 0.48). For predicting the total MEDD, LR model outperformed other models (AUROC 0.67). For predicting analgesics efficacy, the SVM achieved the highest specificity (0.97), while the RF and GBM models achieved the highest AUROC (0.68). RF model emerged as the best calibrated model with an ECE of 0.02 and 0.05 for pain intensity and MEDD prediction, respectively. Baseline pain scores and vital signs demonstrated the most contributing features.

ML models showed promise in predicting end-of-treatment pain intensity, opioid requirements and analgesics efficacy in OCC/OPC patients. Baseline pain score and vital signs are crucial predictors. Their implementation in clinical practice could facilitate early risk stratification and personalized pain management.