Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Jun 7;7(2):026110.
doi: 10.1063/5.0153413. eCollection 2023 Jun.

AI-aided holographic flow cytometry for label-free identification of ovarian cancer cells in the presence of unbalanced datasets

F Borrelli, J Behal, A Cohen  1 , L Miccio  2 , P Memmolo  2 , I Kurelac, A Capozzoli  3 , C Curcio  3 , A Liseno  3 , V Bianco  2 , N T Shaked  1 , P Ferraro  2
Affiliations

AI-aided holographic flow cytometry for label-free identification of ovarian cancer cells in the presence of unbalanced datasets

F Borrelli et al. APL Bioeng. .

Abstract

Liquid biopsy is a valuable emerging alternative to tissue biopsy with great potential in the noninvasive early diagnostics of cancer. Liquid biopsy based on single cell analysis can be a powerful approach to identify circulating tumor cells (CTCs) in the bloodstream and could provide new opportunities to be implemented in routine screening programs. Since CTCs are very rare, the accurate classification based on high-throughput and highly informative microscopy methods should minimize the false negative rates. Here, we show that holographic flow cytometry is a valuable instrument to obtain quantitative phase-contrast maps as input data for artificial intelligence (AI)-based classifiers. We tackle the problem of discriminating between A2780 ovarian cancer cells and THP1 monocyte cells based on the phase-contrast images obtained in flow cytometry mode. We compare conventional machine learning analysis and deep learning architectures in the non-ideal case of having a dataset with unbalanced populations for the AI training step. The results show the capacity of AI-aided holographic flow cytometry to discriminate between the two cell lines and highlight the important role played by the phase-contrast signature of the cells to guarantee accurate classification.

PubMed Disclaimer

Conflict of interest statement

The authors have no conflicts to disclose.

Figures

FIG. 1.
FIG. 1.
Sketch of experimental arrangement. MO1—focusing microscope objective; P—pinhole; L— collimating lens; HWP—half-wave plate; PBS—polarizing beam splitter; MO—microscope objective; MC—microfluidic chip; Ms—mirrors; I—iris diaphragm; BS—beam splitter; CMOS—camera; SD—shearing device. The image in front of the MO provides a zoom into the MC with cells flowing inside the channel. Inserts in front of the CMOS illustrate replicas arising from the SD. The highlighted image portions (area1 and area 2) represent overlapping areas detected by the CMOS, which provide correct holographic signature.
FIG. 2.
FIG. 2.
PCA and T-SNE carried out over the data before feature selection to assess the complexity of the classification task by observing the natural data clustering.
FIG. 3.
FIG. 3.
Absolute value of Kendall correlation coefficient among the inspected features. The red square marks out the QPI features.
FIG. 4.
FIG. 4.
Results of Relief analysis performed over the features to assess their significance. The red arrows point out some of the QPI features, which are among the most informative.
FIG. 5.
FIG. 5.
Results of PCA and T-SNE analysis after the feature selection process.
FIG. 6.
FIG. 6.
Validation and test confusion matrices relative to the Cubic SVM classifier when all the features are employed.
FIG. 7.
FIG. 7.
Figures showing the (a) details of the MobileNet V2 architecture and (b) ResNet-18 architecture used.
FIG. 8.
FIG. 8.
Validation and test confusion matrices relative to MobileNet-V2 classifier.
FIG. 9.
FIG. 9.
Outline of the reconstruction pipeline: (a) acquired hologram (b), Fourier spectrum of the holograms with the three orders visible (c), in-focus amplitude of the complex field obtained after the refocusing procedure, and (d) and (e) QPMs of an OC cell and a monocyte.
See this image and copyright information in PMC

References

    1. Miccio L., Cimmino F., Kurelac I., Villone M. M., Bianco V., Memmolo P., Merola F., Mugnano M., Capasso M., Iolascon A., Maffettone P. L., and Ferraro P., VIEW 1, 20200034 (2020).10.1002/VIW.20200034 - DOI
    1. Kaldjian E. P., Ramirez A. B., Sun Y., Campton D. E., Werbin J. L., Varshavskaya P., Quarre S., George T., Madan A., Blau C. A., and Seubert R., Cytometry, Part A 93, 1220 (2018).10.1002/cyto.a.23619 - DOI - PMC - PubMed
    1. Zamay A. S., Zamay G. S., Kolovskaya O. S., Zamay T. N., and Berezovski M. V., “ Aptamer-based methods for detection of circulating tumor cells and their potential for personalized diagnostics,” in Isolation and Molecular Characterization of Circulating Tumor Cells, edited by Magbanua M. J. M. and Park J. W. ( Springer International Publishing, Cham, 2017), pp. 67–81. - PubMed
    1. Maheswaran S., Sequist L. V., Nagrath S., Ulkus L., Brannigan B., Collura C. V., Inserra E., Diederichs S., Iafrate A. J., Bell D. W., Digumarthy S., Muzikansky A., Irimia D., Settleman J., Tompkins R. G., Lynch T. J., Toner M., and Haber D. A., New Engl. J. Med. 359, 366 (2008).10.1056/NEJMoa0800668 - DOI - PMC - PubMed
    1. Denis J. A., Patroni A., Guillerm E., Pépin D., Benali-Furet N., Wechsler J., Manceau G., Bernard M., Coulet F., Larsen A. K., Karoui M., and Lacorte J.-M., Mol. Oncol. 10, 1221 (2016).10.1016/j.molonc.2016.05.009 - DOI - PMC - PubMed