Normal tissue complication probability (NTCP) models can be used to guide radiation therapy (RT) decisions by estimating side-effect risks pretreatment to minimize (late) side-effects. Recently, a comprehensive individual toxicity risk (CITOR) profile of NTCP models addressing common side-effects in head and neck cancer (HNC) patients was developed. This study investigates the generalizability of these models in an international setting, with different treatment approaches and side-effect assessments, promoting their integration into more widespread clinical practice. From a prospective registry study, 407 HNC patients were included who were treated with definitive RT with or without systemic therapy between 2015 and 2022. NTCP models predicting dysphagia, aspiration, xerostomia, sticky saliva, taste loss, speech problems, oral pain, and fatigue at 6 and 12 months after RT were evaluated. All side-effects were patient-rated using the MDASI-HN, except dysphagia which was reported by clinicians using the PSS-HN diet normalcy score. Model performance was appraised by discrimination (area under the curve [AUC]) and calibration. CITOR models showed moderate-to-high performance in this cohort (mean AUC = 0.67[range = 0.55-0.80], moderate-to-good calibration). NTCP models for dysphagia, xerostomia, sticky saliva, and fatigue were the top performing models. Models for aspiration, taste loss and speech problems performed moderately well, which was partly explained by lower incidences. Despite differences between the CITOR development and this evaluation cohort, including use of different side-effect scoring systems, most models exhibited moderate-to-high performance. This demonstrated that the dose-effect relations were generalizable. Therefore, this study supports further integration of these NTCP models in clinical practice.
Late radiation-associated dysphagia after head and neck cancer (HNC) significantly impacts patient's health and quality of life. Conventional normal tissue complication probability (NTCP) models use discrete dose parameters to predict toxicity risk but fail to fully capture the complexity of this side effect. Deep learning (DL) offers potential improvements by incorporating 3D dose data for all anatomical structures involved in swallowing. This study aims to enhance dysphagia prediction with 3D DL NTCP models compared to conventional NTCP models. A multi-institutional cohort of 1484 HNC patients was used to train and validate a 3D DL model (Residual Network) incorporating 3D dose distributions, organ-at-risk segmentations, and CT scans, with or without patient- or treatment-related data. Predictions of grade ≥ 2 dysphagia (CTCAEv4) at six months post-treatment were evaluated using area under the curve (AUC) and calibration curves. Results were compared to a conventional NTCP model based on pre-treatment dysphagia, tumour location, and mean dose to swallowing organs. Attention maps highlighting regions of interest for individual patients were assessed. DL models outperformed the conventional NTCP model in both the independent test set (AUC = 0.80-0.84 versus 0.76) and external test set (AUC = 0.73-0.74 versus 0.63) in AUC and calibration. Attention maps showed a focus on the oral cavity and superior pharyngeal constrictor muscle. DL NTCP models performed significantly better than the conventional NTCP model, suggesting the benefit of using 3D-input over the conventional discrete dose parameters. Attention maps highlighted relevant regions linked to dysphagia, supporting the utility of DL for improved predictions.
Purpose: To develop and internally validate prediction models with machine learning for future potentially preventable healthcare utilization in patients with multiple long term conditions (MLTC). This study is the first step in investigating whether prediction models can help identify patients with MLTC that are most in need of integrated care. Methods: A retrospective cohort study was performed with electronic health record data from adults with MLTC from an academic medical center in the Netherlands. Based on demographic and healthcare utilization characteristics in 2017, we predicted ≥ 12 outpatient visits, ≥ 1 emergency department (ED) visits, and ≥ 1 acute hospitalizations in 2018. Four machine learning models (elastic net regression, extreme gradient boosting (XGB), logistic regression, and random forest) were developed, optimized, and evaluated in a hold-out dataset for each outcome. Results: A total of 14,486 patients with MLTC were included. Based on the area under the curve (AUC) and calibration curves, the XGB model was selected as final model for all three outcomes. The AUC was 0.82 for ≥ 12 outpatient visits, 0.76 for ≥ 1 ED visits and 0.73 for ≥ 1 acute hospitalizations. Despite adequate AUC and calibration, precision-recall curves showed suboptimal performance. Conclusions: The final selected models per outcome can identify patients with future potentially preventable high healthcare utilization. However, identifying high-risk patients with MLTC and substantiating if they are most in need of integrated care remains challenging. Further research is warranted investigating whether patients with high healthcare utilization are indeed the most in need of integrated care and whether quantitively identified patients match the identification based on clinicians’ experience and judgment.
Purpose: Pulmonary arterial hypertension (PAH) significantly affects the pulmonary vasculature, requiring accurate estimation of mean pulmonary arterial pressure (mPAP) and pulmonary vascular resistance index (PVRi). Although cardiac catheterization is the gold standard for these measurements, it poses risks, especially in children. This pilot study explored how machine learning (ML) can predict pulmonary hemodynamics from non-invasive cardiac magnetic resonance (CMR) cine images in pediatric PAH patients. Methods: A retrospective analysis of 40 CMR studies from children with PAH using a four-fold stratified group cross-validation was conducted. The endpoints were severity profiles of mPAP and PVRi, categorised as ‘low’, ‘high’, and ‘extreme’. Deep learning (DL) and traditional ML models were optimized through hyperparameter tuning. Receiver operating characteristic curves and area under the curve (AUC) were used as the primary evaluation metrics. Results: DL models utilizing CMR cine imaging showed the best potential for predicting mPAP and PVRi severity profiles on test folds (AUCmPAP=0.82 and AUCPVRi=0.73). True positive rates (TPR) for predicting low, high, and extreme mPAP were 5/10, 11/16, and 11/14, respectively. TPR for predicting low, high, and extreme PVRi were 5/13, 14/15, and 7/12, respectively. Optimal DL models only used spatial patterns from consecutive CMR cine frames to maximize prediction performance. Conclusion: This exploratory pilot study demonstrates the potential of DL leveraging CMR imaging for non-invasive prediction of mPAP and PVRi in pediatric PAH. While preliminary, these findings may lay the groundwork for future advancements in CMR imaging in pediatric PAH, offering a pathway to safer disease monitoring and reduced reliance on invasive cardiac catheterization.
Background & purpose Late radiation-associated dysphagia after head and neck cancer (HNC) significantly impacts patient’s health and quality of life. Conventional normal tissue complication probability (NTCP) models use discrete dose parameters to predict toxicity risk but fail to fully capture the complexity of this side effect. Deep learning (DL) offers potential improvements by incorporating 3D dose data for all anatomical structures involved in swallowing. This study aims to enhance dysphagia prediction with 3D DL NTCP models compared to conventional NTCP models. Materials & methods A multi-institutional cohort of 1484 HNC patients was used to train and validate a 3D DL model (Residual Network) incorporating 3D dose distributions, organ-at-risk segmentations, and CT scans, with or without patient- or treatment-related data. Predictions of grade ≥2 dysphagia (CTCAEv4) at six months post-treatment were evaluated using area under the curve (AUC) and calibration curves. Results were compared to a conventional NTCP model based on pre-treatment dysphagia, tumour location, and mean dose to swallowing organs. Attention maps highlighting regions of interest for individual patients were assessed. Results DL models outperformed the conventional NTCP model in both the independent test set (AUC=0.80-0.84 versus 0.76) and external test set (AUC=0.73-0.74 versus 0.63) in AUC and calibration. Attention maps showed a focus on the oral cavity and superior pharyngeal constrictor muscle. Conclusion DL NTCP models performed better than the conventional NTCP model, suggesting the benefit of using 3D-input over the conventional discrete dose parameters. Attention maps highlighted relevant regions linked to dysphagia, supporting the utility of DL for improved predictions.