Optimizing Channel Selection for U-Net Segmentation in MIBI-TOF of Triple-Negative Breast Cancer

• Victor Gabriel Pereira de Sousa UFPI
• Daniel de Sousa Luz UFPI
• Daniel Rodrigues de Sousa UFPI
• Flávio Henrique Duarte de Araújo UFPI

DOI: https://doi.org/10.5753/sbcas.2026.21491

Resumo

Multiplexed ion-beam imaging enables simultaneous quantification of dozens of protein markers at subcellular resolution. In this paper, we investigate how channel selection impacts U-Net–based segmentation performance on a public triple-negative breast cancer MIBI-TOF dataset with 41 images each with 48 channels. Our experiments show that compact, data-driven marker subsets achieve higher Dice than the recommended panel, suggested by the dataset paper, while using fewer channels. In summary, our experiments show that it is possible to reduce from seven channels to one without losing segmentation accuracy: our best single-channel model (dsDNA) achieves 89.36% Dice, practically matching the 89.41% Dice obtained by the full recommended panel.

Referências

Andrews, L. P., Marciscano, A. E., Drake, C. G., and Vignali, D. A. A. (2017). LAG3 (CD223) as a cancer immunotherapy target. Immunological Reviews, 276(1):80–96.

Angelo, M., Bendall, S. C., Finck, R., Hale, M. B., Hitzman, C., Borowsky, A. D., Levenson, R. M., Lowe, J. B., Liu, S. D., Zhao, S., Natkunam, Y., and Nolan, G. P. (2014). Multiplexed ion beam imaging of human breast tumors. Nature Medicine, 20(4):436–442.

Bray, F., Laversanne, M., Sung, H., Ferlay, J., Siegel, R. L., Soerjomataram, I., and Jemal, A. (2024). Global cancer statistics 2022: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: a cancer journal for clinicians, 74(3):229–263.

Chen, Z., Soifer, I., Hilton, H., Keren, L., and Jojic, V. (2020). Modeling multiplexed images with spatial-lda reveals novel tissue microenvironments. Journal of Computational Biology, 27(8):1204–1218.

Croft, M. (2010). Control of immunity by the TNFR-related molecule OX40 (CD134). Annual Review of Immunology, 28:57–78.

Etzerodt, A. and Moestrup, S. K. (2013). CD163 and inflammation: biological, diagnostic, and therapeutic aspects. Antioxidants & Redox Signaling, 18(17):2352–2363.

Farber, D. L., Yudanin, N. A., and Restifo, N. P. (2014). Human memory T cells: generation, compartmentalization and homeostasis. Nature Reviews Immunology, 14(1):24–35.

Geijtenbeek, T. B. H., Kwon, D.-S., Torensma, R., van Vliet, S. J., van Duijnhoven, G. C. F., Middel, J., Cornelissen, I. L. M. H. A., Nottet, H. S. L. M., KewalRamani, V. N., Littman, D. R., Figdor, C. G., and van Kooyk, Y. (2000). DC-SIGN, a dendritic cell-specific HIV-1-binding protein that enhances trans-infection of T cells. Cell, 100(5):587–597.

Hori, S., Nomura, T., and Sakaguchi, S. (2003). Control of regulatory T cell development by the transcription factor Foxp3. Science, 299(5609):1057–1061.

Jie, H., Ma, W., and Huang, C. (2025). Diagnosis, prognosis, and treatment of triple-negative breast cancer: A review. Breast Cancer: Targets and Therapy, pages 265–274.

Keren, L., Bosse, M., Marquez, D., Angoshtari, R., Jain, S., Varma, S., Yang, S.-R., Kurian, A., Van Valen, D., West, R., Bendall, S. C., and Angelo, M. (2018). A structured tumor-immune microenvironment in triple negative breast cancer revealed by multiplexed ion beam imaging. Cell, 174(6):1373–1387.e19.

Kouzarides, T. (2007). Chromatin modifications and their function. Cell, 128(4):693–705.

Liu, C. C., Greenwald, N. F., Kong, A., McCaffrey, E. F., Leow, K. X., Mrdjen, D., Cannon, B. J., Rumberger, J. L., Varra, S. R., and Angelo, M. (2023). Robust phenotyping of highly multiplexed tissue imaging data using pixel-level clustering. Nature communications, 14(1):4618.

Magness, A., Colliver, E., Enfield, K. S., Lee, C., Shimato, M., Daly, E., Moore, D. A., Sivakumar, M., Valand, K., Levi, D., et al. (2024). Deep cell phenotyping and spatial analysis of multiplexed imaging with tracerx-phlex. Nature Communications, 15(1):5135.

Neefjes, J., Jongsma, M. L. M. J., Paul, P., and Bakke, O. (2011). Towards a systems understanding of MHC class I and MHC class II antigen presentation. Nature Reviews Immunology, 11(12):823–836.

Nimmerjahn, F. and Ravetch, J. V. (2008). Fcγ receptors as regulators of immune responses. Nature Reviews Immunology, 8(1):34–47.

Pardoll, D. M. (2012). The blockade of immune checkpoints in cancer immunotherapy. Nature Reviews Cancer, 12(4):252–264.

Passarelli, M. K. and Winograd, N. (2011). Lipid imaging with time-of-flight secondary ion mass spectrometry (ToF-SIMS). Biochimica et Biophysica Acta, 1811(11):976–990.

Patwa, A., Yamashita, R., Long, J., Risom, T., Angelo, M., Keren, L., and Rubin, D. L. (2021). Multiplexed imaging analysis of the tumor-immune microenvironment reveals predictors of outcome in triple-negative breast cancer. Communications biology, 4(1):852.

Petinrin, O. O., Saeed, F., Toseef, M., Liu, Z., Basurra, S., Muyide, I. O., Li, X., and Wong, K.-C. (2023). Machine learning in metastatic cancer research: Potentials, possibilities, and prospects. Computational and Structural Biotechnology Journal, 21:2454–2470.

Picarda, E., Ohaegbulam, K. C., and Zang, X. (2016). Molecular pathways: Targeting B7-H3 (CD276) for human cancer immunotherapy. Clinical Cancer Research, 22(14):3425–3431.

Ronneberger, O., Fischer, P., and Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, volume 9351 of Lecture Notes in Computer Science, pages 234–241. Springer.

Saxton, R. A. and Sabatini, D. M. (2017). mTOR signaling in growth, metabolism, and disease. Cell, 168(6):960–976.