---

About

Model Facts

Summary

This algorithm segments the whole prostate gland in biparametric MRI (bpMRI). Development of this model was geared toward robust prostate segmentation, at the expense of fine-grained zonal segmentation.

If you are using the AI-generated prostate segmentations, please cite the following article:

A. Saha, J. J. Twilt, J. S. Bosma, B. van Ginneken, D. Yakar, M. Elschot, J. Veltman, J. J. Fütterer, M. de Rooij, H. Huisman, "Artificial Intelligence and Radiologists at Prostate Cancer Detection in MRI: The PI-CAI Challenge (Study Protocol)", DOI: 10.5281/zenodo.6667655

@article{PICAI_Study_design, author={Anindo Saha AND Jasper J. Twilt AND Joeran S. Bosma AND Bram van Ginneken AND Derya Yakar AND Mattijs Elschot AND Jeroen Veltman AND Jurgen Fütterer AND Maarten de Rooij AND Henkjan Huisman}, title={{Artificial Intelligence and Radiologists at Prostate Cancer Detection in MRI: The PI-CAI Challenge (Study Protocol)}}, year={2022}, doi={10.5281/zenodo.6667655} }

Mechanism

This algorithm is a deep learning-based model, which ensembles five independent nnU-Net models (using 5-fold cross-validation). To prioritize robust segmentation, we trained these models with Cross-Entropy + Focal loss. We trained these models with a total of 438 prostate biparametric MRI (bpMRI) scans paired with a manual prostate segmentation. These scans were sourced from two independent hospitals: 299 cases from Radboudumc (of which 248 part of ProstateX) and 139 cases from Prostate158. Annotations for ProstateX cases were retrieved from https://github.com/rcuocolo/PROSTATEx_masks/.

We ensured there is no patient overlap between this algorithm's training dataset and the PI-CAI Hidden Validation and Tuning Cohort or Hidden Testing Cohort.

Source code and trained models are publicly vailable.

Validation and Performance

This algorithm is evaluated using 5-fold cross-validation using the dataset described in Mechanism. Segmentation performance and its standard deviation across cases are provided below.

Dice similarity coefficient	Jaccard index
0.8968 ± 0.0547	0.8169 ± 0.0820

Uses and Directions

• For research use only. This algorithm is intended to be used only on biparametric prostate MRI examinations. This algorithm should not be used in different patient demographics.

• Target population: This algorithm was trained on patients without prior treatment (e.g. radiotherapy, transurethral resection of the prostate (TURP), transurethral ultrasound ablation (TULSA), cryoablation, etc.), without prior positive biopsies, without artefacts and with reasonably-well aligned sequences.

• MRI scanner: This algorithm was trained and evaluated exclusively on prostate bpMRI scans derived from Siemens Healthineers (Skyra/Prisma/Trio/Avanto) MRI scanners with surface coils. It does not account for vendor-neutral properties or domain adaptation, and in turn, is not compatible with scans derived using any other MRI scanner or those using endorectal coils.

• Sequence alignment and position of the prostate: While the input images (T2W, HBV, ADC) can be of different spatial resolutions, the algorithm assumes that they are co-registered or aligned reasonably well.

• General use: This model is intended to be used by radiologists for predicting prostate volume in biparametric MRI examinations. The model is not meant to guide or drive clinical care. This model is intended to complement other pieces of patient information in order to determine the appropriate follow-up recommendation.

• Before using this model: Test the model retrospectively and prospectively on a cohort that reflects the target population that the model will be used upon to confirm the validity of the model within a local setting.

• Safety and efficacy evaluation: To be determined in a clinical validation study.

Warnings

• Risks: Even if used appropriately, clinicians using this model can estimate prostate volume incorrectly.

• Inappropriate Settings: This model was not trained on MRI examinations of patients with prior treatment (e.g. radiotherapy, transurethral resection of the prostate (TURP), transurethral ultrasound ablation (TULSA), cryoablation, etc.), prior positive biopsies, artefacts or misalignment between sequences. Hence it is susceptible to faulty predictions and unintended behaviour when presented with such cases. Do not use the model in the clinic without further evaluation.

• Clinical rationale: The model is not interpretable and does not provide a rationale. Clinical end users are expected to place the model output in context with other clinical information.

• Inappropriate decision support: This model may not be accurate outside of the target population. This model is not designed to guide clinical diagnosis and treatment for prostate cancer.

• Generalizability: This model was primarily developed with prostate MRI examinations from Radboud University Medical Centre and the Andros Kliniek. Do not use this model in an external setting without further evaluation.

• Discontinue use if: Clinical staff raise concerns about the utility of the model for the intended use case or large, systematic changes occur at the data level that necessitates re-training of the model.

Common Error Messages

Information on this algorithm has been provided by the Algorithm Editors, following the Model Facts labels guidelines from Sendak, M.P., Gao, M., Brajer, N. et al. Presenting machine learning model information to clinical end users with model facts labels. npj Digit. Med. 3, 41 (2020). 10.1038/s41746-020-0253-3