Joeran S. Bosma

Luc Builtjes, Joeran Bosma, Mathias Prokop et al.

Medical reports contain rich clinical information but are often unstructured and written in domain-specific language, posing challenges for information extraction. While proprietary large language models (LLMs) have shown promise in clinical natural language processing, their lack of transparency and data privacy concerns limit their utility in healthcare. This study therefore evaluates nine open-source generative LLMs on the DRAGON benchmark, which includes 28 clinical information extraction tasks in Dutch. We developed \texttt{llm\_extractinator}, a publicly available framework for information extraction using open-source generative LLMs, and used it to assess model performance in a zero-shot setting. Several 14 billion parameter models, Phi-4-14B, Qwen-2.5-14B, and DeepSeek-R1-14B, achieved competitive results, while the bigger Llama-3.3-70B model achieved slightly higher performance at greater computational cost. Translation to English prior to inference consistently degraded performance, highlighting the need of native-language processing. These findings demonstrate that open-source LLMs, when used with our framework, offer effective, scalable, and privacy-conscious solutions for clinical information extraction in low-resource settings.

Matin Hosseinzadeh, Anindo Saha, Joeran Bosma et al.

Quality of deep convolutional neural network predictions strongly depends on the size of the training dataset and the quality of the annotations. Creating annotations, especially for 3D medical image segmentation, is time-consuming and requires expert knowledge. We propose a novel semi-supervised learning (SSL) approach that requires only a relatively small number of annotations while being able to use the remaining unlabeled data to improve model performance. Our method uses a pseudo-labeling technique that employs recent deep learning uncertainty estimation models. By using the estimated uncertainty, we were able to rank pseudo-labels and automatically select the best pseudo-annotations generated by the supervised model. We applied this to prostate zonal segmentation in T2-weighted MRI scans. Our proposed model outperformed the semi-supervised model in experiments with the ProstateX dataset and an external test set, by leveraging only a subset of unlabeled data rather than the full collection of 4953 cases, our proposed model demonstrated improved performance. The segmentation dice similarity coefficient in the transition zone and peripheral zone increased from 0.835 and 0.727 to 0.852 and 0.751, respectively, for fully supervised model and the uncertainty-aware semi-supervised learning model (USSL). Our USSL model demonstrates the potential to allow deep learning models to be trained on large datasets without requiring full annotation. Our code is available at https://github.com/DIAGNijmegen/prostateMR-USSL.

Joeran S. Bosma, Anindo Saha, Matin Hosseinzadeh et al.

Deep learning-based diagnostic performance increases with more annotated data, but large-scale manual annotations are expensive and labour-intensive. Experts evaluate diagnostic images during clinical routine, and write their findings in reports. Leveraging unlabelled exams paired with clinical reports could overcome the manual labelling bottleneck. We hypothesise that detection models can be trained semi-supervised with automatic annotations generated using model predictions, guided by sparse information from clinical reports. To demonstrate efficacy, we train clinically significant prostate cancer (csPCa) segmentation models, where automatic annotations are guided by the number of clinically significant findings in the radiology reports. We included 7,756 prostate MRI examinations, of which 3,050 were manually annotated. We evaluated prostate cancer detection performance on 300 exams from an external centre with histopathology-confirmed ground truth. Semi-supervised training improved patient-based diagnostic area under the receiver operating characteristic curve from $87.2 \pm 0.8\%$ to $89.4 \pm 1.0\%$ ($P<10^{-4}$) and improved lesion-based sensitivity at one false positive per case from $76.4 \pm 3.8\%$ to $83.6 \pm 2.3\%$ ($P<10^{-4}$). Semi-supervised training was 14$\times$ more annotation-efficient for case-based performance and 6$\times$ more annotation-efficient for lesion-based performance. This improved performance demonstrates the feasibility of our training procedure. Source code is publicly available at github.com/DIAGNijmegen/Report-Guided-Annotation. Best csPCa detection algorithm is available at grand-challenge.org/algorithms/bpmri-cspca-detection-report-guided-annotations/.

Alessa Hering, Sarah de Boer, Anindo Saha et al.

The PI-CAI (Prostate Imaging: Cancer AI) challenge led to expert-level diagnostic algorithms for clinically significant prostate cancer detection. The algorithms receive biparametric MRI scans as input, which consist of T2-weighted and diffusion-weighted scans. These scans can be misaligned due to multiple factors in the scanning process. Image registration can alleviate this issue by predicting the deformation between the sequences. We investigate the effect of image registration on the diagnostic performance of AI-based prostate cancer diagnosis. First, the image registration algorithm, developed in MeVisLab, is analyzed using a dataset with paired lesion annotations. Second, the effect on diagnosis is evaluated by comparing case-level cancer diagnosis performance between using the original dataset, rigidly aligned diffusion-weighted scans, or deformably aligned diffusion-weighted scans. Rigid registration showed no improvement. Deformable registration demonstrated a substantial improvement in lesion overlap (+10% median Dice score) and a positive yet non-significant improvement in diagnostic performance (+0.3% AUROC, p=0.18). Our investigation shows that a substantial improvement in lesion alignment does not directly lead to a significant improvement in diagnostic performance. Qualitative analysis indicated that jointly developing image registration methods and diagnostic AI algorithms could enhance diagnostic accuracy and patient outcomes.

Natália Alves, Megan Schuurmans, Geke Litjens et al.

Early detection improves prognosis in pancreatic ductal adenocarcinoma (PDAC) but is challenging as lesions are often small and poorly defined on contrast-enhanced computed tomography scans (CE-CT). Deep learning can facilitate PDAC diagnosis, however current models still fail to identify small (<2cm) lesions. In this study, state-of-the-art deep learning models were used to develop an automatic framework for PDAC detection, focusing on small lesions. Additionally, the impact of integrating surrounding anatomy was investigated. CE-CT scans from a cohort of 119 pathology-proven PDAC patients and a cohort of 123 patients without PDAC were used to train a nnUnet for automatic lesion detection and segmentation (nnUnet_T). Two additional nnUnets were trained to investigate the impact of anatomy integration: (1) segmenting the pancreas and tumor (nnUnet_TP), (2) segmenting the pancreas, tumor, and multiple surrounding anatomical structures (nnUnet_MS). An external, publicly available test set was used to compare the performance of the three networks. The nnUnet_MS achieved the best performance, with an area under the receiver operating characteristic curve of 0.91 for the whole test set and 0.88 for tumors <2cm, showing that state-of-the-art deep learning can detect small PDAC and benefits from anatomy information.

Anindo Saha, Joeran Bosma, Jasper Linmans et al.

We hypothesize that probabilistic voxel-level classification of anatomy and malignancy in prostate MRI, although typically posed as near-identical segmentation tasks via U-Nets, require different loss functions for optimal performance due to inherent differences in their clinical objectives. We investigate distribution, region and boundary-based loss functions for both tasks across 200 patient exams from the publicly-available ProstateX dataset. For evaluation, we conduct a thorough comparative analysis of model predictions and calibration, measured with respect to multi-class volume segmentation of the prostate anatomy (whole-gland, transitional zone, peripheral zone), as well as, patient-level diagnosis and lesion-level detection of clinically significant prostate cancer. Notably, we find that distribution-based loss functions (in particular, focal loss) are well-suited for diagnostic or panoptic segmentation tasks such as lesion detection, primarily due to their implicit property of inducing better calibration. Meanwhile, (with the exception of focal loss) both distribution and region/boundary-based loss functions perform equally well for anatomical or semantic segmentation tasks, such as quantification of organ shape, size and boundaries.