Benchmarking ChatGPT-4 on ACR Radiation Oncology In-Training Exam (TXIT): Potentials and Challenges for AI-Assisted Medical Education and Decision Making in Radiation Oncology

The potential of large language models in medicine for education and decision making purposes has been demonstrated as they achieve decent scores on medical exams such as the United States Medical Licensing Exam (USMLE) and the MedQA exam. In this work, we evaluate the performance of ChatGPT-3.5 and ChatGPT-4 in the specialized field of radiation oncology using the 38th American College of Radiology (ACR) radiation oncology in-training exam (TXIT). ChatGPT-3.5 and ChatGPT-4 have achieved the scores of 63.65% and 74.57%, respectively, highlighting the advantage of the latest ChatGPT-4 model. Based on the TXIT exam, ChatGPT-4's strong and weak areas in radiation oncology are identified to some extent. Specifically, ChatGPT-4 demonstrates good knowledge of statistics, CNS & eye, pediatrics, biology, and physics but has limitations in bone & soft tissue and gynecology, as per the ACR knowledge domain. Regarding clinical care paths, ChatGPT-4 performs well in diagnosis, prognosis, and toxicity but lacks proficiency in topics related to brachytherapy and dosimetry, as well as in-depth questions from clinical trials. While ChatGPT-4 is not yet suitable for clinical decision making in radiation oncology, it has the potential to assist in medical education for the general public and cancer patients. With further fine-tuning, it could assist radiation oncologists in recommending treatment decisions for challenging clinical cases based on the latest guidelines and the existing gray zone database.

The Segment Anything foundation model achieves favorable brain tumor autosegmentation accuracy on MRI to support radiotherapy treatment planning

Background: Tumor segmentation in MRI is crucial in radiotherapy (RT) treatment planning for brain tumor patients. Segment anything (SA), a novel promptable foundation model for autosegmentation, has shown high accuracy for multiple segmentation tasks but was not evaluated on medical datasets yet. Methods: SA was evaluated in a point-to-mask task for glioma brain tumor autosegmentation on 16744 transversal slices from 369 MRI datasets (BraTS 2020). Up to 9 point prompts were placed per slice. Tumor core (enhancing tumor + necrotic core) was segmented on contrast-enhanced T1w sequences. Out of the 3 masks predicted by SA, accuracy was evaluated for the mask with the highest calculated IoU (oracle mask) and with highest model predicted IoU (suggested mask). In addition to assessing SA on whole MRI slices, SA was also evaluated on images cropped to the tumor (max. 3D extent + 2 cm). Results: Mean best IoU (mbIoU) using oracle mask on full MRI slices was 0.762 (IQR 0.713-0.917). Best 2D mask was achieved after a mean of 6.6 point prompts (IQR 5-9). Segmentation accuracy was significantly better for high- compared to low-grade glioma cases (mbIoU 0.789 vs. 0.668). Accuracy was worse using MRI slices cropped to the tumor (mbIoU 0.759) and was much worse using suggested mask (full slices 0.572). For all experiments, accuracy was low on peripheral slices with few tumor voxels (mbIoU, <300: 0.537 vs. >=300: 0.841). Stacking best oracle segmentations from full axial MRI slices, mean 3D DSC for tumor core was 0.872, which was improved to 0.919 by combining axial, sagittal and coronal masks. Conclusions: The Segment Anything foundation model, while trained on photos, can achieve high zero-shot accuracy for glioma brain tumor segmentation on MRI slices. The results suggest that Segment Anything can accelerate and facilitate RT treatment planning, when properly integrated in a clinical application.

Deep Learning for automatic head and neck lymph node level delineation

Background: Deep learning-based head and neck lymph node level (HN_LNL) autodelineation is of high relevance to radiotherapy research and clinical treatment planning but still understudied in academic literature. Methods: An expert-delineated cohort of 35 planning CTs was used for training of an nnU-net 3D-fullres/2D-ensemble model for autosegmentation of 20 different HN_LNL. Validation was performed in an independent test set (n=20). In a completely blinded evaluation, 3 clinical experts rated the quality of deep learning autosegmentations in a head-to-head comparison with expert-created contours. For a subgroup of 10 cases, intraobserver variability was compared to deep learning autosegmentation performance. The effect of autocontour consistency with CT slice plane orientation on geometric accuracy and expert rating was investigated. Results: Mean blinded expert rating per level was significantly better for deep learning segmentations with CT slice plane adjustment than for expert-created contours (81.0 vs. 79.6, p<0.001), but deep learning segmentations without slice plane adjustment were rated significantly worse than expert-created contours (77.2 vs. 79.6, p<0.001). Geometric accuracy of deep learning segmentations was non-different from intraobserver variability (mean Dice per level, 0.78 vs. 0.77, p=0.064) with variance in accuracy between levels being improved (p<0.001). Clinical significance of contour consistency with CT slice plane orientation was not represented by geometric accuracy metrics (Dice, 0.78 vs. 0.78, p=0.572) Conclusions: We show that a nnU-net 3D-fullres/2D-ensemble model can be used for highly accurate autodelineation of HN_LNL using only a limited training dataset that is ideally suited for large-scale standardized autodelineation of HN_LNL in the research setting. Geometric accuracy metrics are only an imperfect surrogate for blinded expert rating.

Deep learning for brain metastasis detection and segmentation in longitudinal MRI data

Brain metastases occur frequently in patients with metastatic cancer. Early and accurate detection of brain metastases is very essential for treatment planning and prognosis in radiation therapy. To improve brain metastasis detection performance with deep learning, a custom detection loss called volume-level sensitivity-specificity (VSS) is proposed, which rates individual metastasis detection sensitivity and specificity in (sub-)volume levels. As sensitivity and precision are always a trade-off in a metastasis level, either a high sensitivity or a high precision can be achieved by adjusting the weights in the VSS loss without decline in dice score coefficient for segmented metastases. To reduce metastasis-like structures being detected as false positive metastases, a temporal prior volume is proposed as an additional input of the neural network. Our proposed VSS loss improves the sensitivity of brain metastasis detection, increasing the sensitivity from 86.7% to 95.5%. Alternatively, it improves the precision from 68.8% to 97.8%. With the additional temporal prior volume, about 45% of the false positive metastases are reduced in the high sensitivity model and the precision reaches 99.6% for the high specificity model. The mean dice coefficient for all metastases is about 0.81. With the ensemble of the high sensitivity and high specificity models, on average only 1.5 false positive metastases per patient needs further check, while the majority of true positive metastases are confirmed. The ensemble learning is able to distinguish high confidence true positive metastases from metastases candidates that require special expert review or further follow-up, being particularly well-fit to the requirements of expert support in real clinical practice.