Bayesian-Optimized Convolutional Neural Networks for Classifying Primary Tumor Origin of Brain Metastases from MRI

Nawabi, Jawed; Eminovic, Semil; Hartenstein, Alexander; Baumgaertner, Georg Lukas; Schnurbusch, Nils; Rudolph, Madhuri; Wasilewski, David; Onken, Julia; Siebert, Eberhard; Wiener, Edzard; Bohner, Georg; Dell'Orco, Andrea; Wattjes, Mike P.; Hamm, Bernd; Fehrenbach, Uli; Penzkofer, Tobias

doi:10.3390/brainsci15050450

Open AccessArticle

Bayesian-Optimized Convolutional Neural Networks for Classifying Primary Tumor Origin of Brain Metastases from MRI

by

Jawed Nawabi

^1,†

,

Semil Eminovic

^2,*,†

,

Alexander Hartenstein

³,

Georg Lukas Baumgaertner

²

,

Nils Schnurbusch

²,

Madhuri Rudolph

²,

David Wasilewski

⁴,

Julia Onken

⁵,

Eberhard Siebert

¹

,

Edzard Wiener

¹,

Georg Bohner

¹

,

Andrea Dell'Orco

¹

,

Mike P. Wattjes

¹,

Bernd Hamm

²,

Uli Fehrenbach

^2,‡

and

Tobias Penzkofer

^2,6,‡

¹

Department of Neuroradiology, Charité—Universitätsmedizin, 10117 Berlin, Germany

²

Department of Radiology, Charité—Universitätsmedizin, 10117 Berlin, Germany

³

Medical Affairs and Pharmacovigilance, Bayer AG, 13353 Berlin, Germany

⁴

Department of Neurosurgery, Medical Faculty & University Hospital Düsseldorf, Heinrich Heine University Dusseldorf, 40225 Dusseldorf, Germany

⁵

Department of Neurosurgery, Charité—Universitätsmedizin, 10117 Berlin, Germany

⁶

Berlin Institute of Health (BIH), 10117 Berlin, Germany

^*

Author to whom correspondence should be addressed.

^†

Jawed Nawabi and Semil Eminovic are shared first authors.

^‡

Uli Fehrenbach and Tobias Penzkofer are shared last authors.

Brain Sci. 2025, 15(5), 450; https://doi.org/10.3390/brainsci15050450

Submission received: 25 March 2025 / Revised: 13 April 2025 / Accepted: 16 April 2025 / Published: 25 April 2025

(This article belongs to the Section Neuro-oncology)

Download

Browse Figures

Versions Notes

Abstract

Background/Objectives: This study evaluates whether convolutional neural networks (CNNs) can be trained to determine the primary tumor origin from MRI images alone in patients with metastatic brain lesions. Methods: This retrospective, monocentric study involved the segmentation of 1175 brain lesions from MRI scans of 436 patients with histologically confirmed primary tumor origins. The four most common tumor types—lung adenocarcinoma, small cell lung cancer, breast cancer, and melanoma—were selected, and a class-balanced dataset was created through under-sampling. This resulted in 276 training datasets and 88 hold-out test datasets. Bayesian optimization was employed to determine the optimal CNN architecture, the most relevant imaging sequences, and whether the masking of images was necessary. We compared the performance of the CNN with that of two expert radiologists specializing in neuro-oncological imaging. Results: The best-performing CNN from the Bayesian optimization process used masked images across all available MRI sequences. It achieved Area-Under-the-Curve (AUC) values of 0.75 for melanoma, 0.65 for small cell lung cancer, 0.64 for breast cancer, and 0.57 for lung adenocarcinoma. Masked images likely improved performance by focusing the CNN on relevant regions and reducing noise from surrounding tissues. In comparison, Radiologist 1 achieved AUCs of 0.55, 0.52, 0.45, and 0.51, and Radiologist 2 achieved AUCs of 0.68, 0.55, 0.64, and 0.43 for the same tumor types, respectively. The CNN consistently showed higher accuracy, particularly for melanoma and breast cancer. Conclusions: Bayesian optimization enabled the creation of a CNN that outperformed expert radiologists in classifying the primary tumor origin of brain metastases from MRI.

Keywords: brain metastases; neural networks; Bayesian optimization

Share and Cite

MDPI and ACS Style

Nawabi, J.; Eminovic, S.; Hartenstein, A.; Baumgaertner, G.L.; Schnurbusch, N.; Rudolph, M.; Wasilewski, D.; Onken, J.; Siebert, E.; Wiener, E.; et al. Bayesian-Optimized Convolutional Neural Networks for Classifying Primary Tumor Origin of Brain Metastases from MRI. Brain Sci. 2025, 15, 450. https://doi.org/10.3390/brainsci15050450

AMA Style

Nawabi J, Eminovic S, Hartenstein A, Baumgaertner GL, Schnurbusch N, Rudolph M, Wasilewski D, Onken J, Siebert E, Wiener E, et al. Bayesian-Optimized Convolutional Neural Networks for Classifying Primary Tumor Origin of Brain Metastases from MRI. Brain Sciences. 2025; 15(5):450. https://doi.org/10.3390/brainsci15050450

Chicago/Turabian Style

Nawabi, Jawed, Semil Eminovic, Alexander Hartenstein, Georg Lukas Baumgaertner, Nils Schnurbusch, Madhuri Rudolph, David Wasilewski, Julia Onken, Eberhard Siebert, Edzard Wiener, and et al. 2025. "Bayesian-Optimized Convolutional Neural Networks for Classifying Primary Tumor Origin of Brain Metastases from MRI" Brain Sciences 15, no. 5: 450. https://doi.org/10.3390/brainsci15050450

APA Style

Nawabi, J., Eminovic, S., Hartenstein, A., Baumgaertner, G. L., Schnurbusch, N., Rudolph, M., Wasilewski, D., Onken, J., Siebert, E., Wiener, E., Bohner, G., Dell'Orco, A., Wattjes, M. P., Hamm, B., Fehrenbach, U., & Penzkofer, T. (2025). Bayesian-Optimized Convolutional Neural Networks for Classifying Primary Tumor Origin of Brain Metastases from MRI. Brain Sciences, 15(5), 450. https://doi.org/10.3390/brainsci15050450

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

Bayesian-Optimized Convolutional Neural Networks for Classifying Primary Tumor Origin of Brain Metastases from MRI

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI