Abstract: Fluorescence in situ hybridization (FISH) is widely used for diagnosing cancer and genetic disorders due to its high specificity and accuracy. However, traditional methods face challenges such as suboptimal focus adjustments, subjective signal counting errors, and inefficiencies in imaging, limiting their use in high-throughput screening. To address these issues, we introduced the integrated FISH imaging and analysis system (FAST), an innovative solution that combines rapid filter switching, automated focusing, multilayer fluorescence signal fusion, and the improved ResNet152 deep learning framework. Compared with clinical manual counts and analysis of case reports of 10 patients with chronic lymphocytic leukemia (CLL), the FAST achieved an average nuclei segmentation accuracy of 98.28%. For abnormal gene detection, the model achieved an accuracy of 97.86%. Additionally, its intuitive interface allows the operator to complete the entire workflow—from scanning to report generation—within 45 min. FAST represents a significant advancement in cancer and genetic disorder diagnostics, offering a powerful tool for early detection.

基于深度学习的荧光原位杂交自动基因异常检测

[1]AbrolV, DhallaS, GuptaS, et al., 2023. An automated segmentation of leukocytes using modified watershed algorithm on peripheral blood smear images. Wireless Personal Communications, 131(1):197-215.

[2]BattledayRM, PetersonJC, GriffithsTL, 2020. Capturing human categorization of natural images by combining deep networks and cognitive models. Nature Communications, 11(1):5418.

[3]BiancalaniT, ScaliaG, BuffoniL, et al., 2021. Deep learning and alignment of spatially resolved single-cell transcriptomes with tangram. Nature Methods, 18(11):1352-1362.

[4]CaicedoJC, CooperS, HeigwerF, et al., 2017. Data-analysis strategies for image-based cell profiling. Nature Methods, 14(9):849-863.

[5]ChoudharyG, SethiD, 2023. Mathematical modeling and simulation of multi-focus image fusion techniques using the effect of image enhancement criteria: a systematic review and performance evaluation. Artificial Intelligence Review, 56(11):13787-13839.

[6]ChyadM, ZaidanBB, ZaidanAA, et al., 2025. Exploring adversarial deep learning for fusion in multi-color channel skin detection applications. Information Fusion, 114:102632.

[7]CuiJ, LiuYT, LvR, et al., 2024. Fluorescence in situ hybridization reveals the evolutionary biology of minor clone of gain/amp(1q) in multiple myeloma. Leukemia, 38(6):1299-1306.

[8]DaiY, IdornM, SerreroMC, et al., 2024. TMEFF1 is a neuron-specific restriction factor for herpes simplex virus. Nature, 632(8024):383-389.

[9]Durak ArasB, IsikS, Uskudar TekeH, et al., 2021. Which prognostic marker is responsible for the clinical heterogeneity in CLL with 13q deletion? Molecular Cytogenetics, 14(1):2.

[10]ElkriefA, MontesionM, SivakumarS, et al., 2024. Intratumoral escherichia is associated with improved survival to single-agent immune checkpoint inhibition in patients with advanced non-small-cell lung cancer. Journal of Clinical Oncology, 42(28):3339-3349.

[11]HuCS, LiH, MaT, et al., 2024. An improved image enhancement algorithm: radial contrast-limited adaptive histogram equalization. Multimedia Tools and Applications, 83(36):83695-83707.

[12]HuY, JonesD, ZhaoWQ, et al., 2023. Incidence, clinicopathologic features, HER2 fluorescence in situ hybridization profile, and oncotype DX results of human epidermal growth factor receptor 2-low breast cancers: experience from a single academic center. Modern Pathology, 36(7):100164.

[13]HuberD, Voith von VoithenbergL, KaigalaGV, 2018. Fluorescence in situ hybridization (FISH): history, limitations and what to expect from micro-scale FISH? Micro and Nano Engineering, 1:15-24.

[14]IsmaelAM, ŞengürA, 2021. Deep learning approaches for COVID-19 detection based on chest X-ray images. Expert Systems with Applications, 164:114054.

[15]JardimS, AntónioJ, MoraC, 2022. Graphical image region extraction with K-means clustering and watershed. Journal of Imaging, 8(6):163.

[16]JonkmanJ, BrownCM, WrightGD, et al., 2020. Tutorial: guidance for quantitative confocal microscopy. Nature Protocols, 15(5):1585-1611.

[17]KnöllM, DheenadayalanRP, YosifovDY, et al., 2024. Biallelic deletion 13q in chronic lymphocytic leukemia treated with targeted agents. Blood, 144(S1):4608.

[18]KrayemB, Moustafa-HawashN, FrischA, et al., 2024. Optical genome mapping reclassifies patients with intermediate risk acute myeloid leukemia. Blood, 144(S1):1559.

[19]LankfordA, SwaminathanM, GonuguntaAS, et al., 2024. Predictors for time-to-first-treatment in patients with del(17p) and/or TP53-mutated chronic lymphocytic leukemia. Blood, 144(S1):3246.

[20]Le GoallecA, DiaiS, CollinS, et al., 2022. Using deep learning to predict abdominal age from liver and pancreas magnetic resonance images. Nature Communications, 13(1):1979.

[21]LewisSM, Asselin-LabatML, NguyenQ, et al., 2021. Spatial omics and multiplexed imaging to explore cancer biology. Nature Methods, 18(9):997-1012.

[22]LiDG, BledsoeJR, ZengY, et al., 2020. A deep learning diagnostic platform for diffuse large B-cell lymphoma with high accuracy across multiple hospitals. Nature Communications, 11(1):6004.

[23]LiFD, YangYQ, WeiYN, et al., 2022. Predicting neoadjuvant chemotherapy benefit using deep learning from stromal histology in breast cancer. npj Breast Cancer, 8(1):124.

[24]LiPY, TianJD, TangYD, et al., 2021. Deep retinex network for single image dehazing. IEEE Transactions on Image Processing, 30:1100-1115.

[25]LiXJ, HouGJ, LiKQ, et al., 2022. Enhancing underwater image via adaptive color and contrast enhancement, and denoising. Engineering Applications of Artificial Intelligence, 111:104759.

[26]LiYZ, PillarN, LiJX, et al., 2024. Virtual histological staining of unlabeled autopsy tissue. Nature Communications, 15(1):1684.

[27]MalikH, NaeemA, Sadeghi-NiarakiA, et al., 2024. Multi-classification deep learning models for detection of ulcerative colitis, polyps, and dyed-lifted polyps using wireless capsule endoscopy images. Complex & Intelligent Systems, 10(2):2477-2497.

[28]MinaeeS, BoykovY, PorikliF, et al., 2022. Image segmentation using deep learning: a survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(7):3523-3542.

[29]NasirES, ParvaizA, FrazMM, 2023. Nuclei and glands instance segmentation in histology images: a narrative review. Artificial Intelligence Review, 56(8):7909-7964.

[30]OrdoñezR, ZhangWM, EllisG, et al., 2024. Genomic context sensitizes regulatory elements to genetic disruption. Molecular Cell, 84(10):1842-1854.e7.

[31]OtomoK, OmuraT, NozawaY, et al., 2024. descSPIM: an affordable and easy-to-build light-sheet microscope optimized for tissue clearing techniques. Nature Communications, 15(1):4941.

[32]PatinoT, 2024. Imaging DNA origami by fluorescence in situ hybridization. Nature Nanotechnology, 19(1):1-2.

[33]Perez-LopezR, Ghaffari LalehN, MahmoodF, et al., 2024. A guide to artificial intelligence for cancer researchers. Nature Reviews Cancer, 24(6):427-441.

[34]PlassM, KarglM, KiehlTR, et al., 2023. Explainability and causability in digital pathology. The Journal of Pathology: Clinical Research, 9(4):251-260.

[35]PunnNS, AgarwalS, 2022. Modality specific U-Net variants for biomedical image segmentation: a survey. Artificial Intelligence Review, 55(7):5845-5889.

[36]Ramos-SotoO, Rodríguez-EsparzaE, Balderas-MataSE, et al., 2021. An efficient retinal blood vessel segmentation in eye fundus images by using optimized top-hat and homomorphic filtering. Computer Methods and Programs in Biomedicine, 201:105949.

[37]SaeedN, KhanZF, JehanzebH, et al., 2024. Frequency of deletion 13q14.3 and its impact on outcome in patients of chronic lymphocytic leukemia: a single-centered institutional study from pakistan. Journal of Haematology and Stem Cell Research, 4(1):120-124.

[38]SchermellehL, FerrandA, HuserT, et al., 2019. Super-resolution microscopy demystified. Nature Cell Biology, 21(1):72-84.

[39]ScottDW, KingRL, StaigerAM, et al., 2018. High-grade B-cell lymphoma with MYC and BCL2 and/or BCL6 rearrangements with diffuse large B-cell lymphoma morphology. Blood, 131(18):2060-2064.

[40]SharmaAK, NandalA, DhakaA, et al., 2022. Enhanced watershed segmentation algorithm-based modified ResNet50 model for brain tumor detection. BioMed Research International, 2022:7348344.

[41]ShenX, QiYF, MaTF, et al., 2019. A dicentric chromosome identification method based on clustering and watershed algorithm. Scientific Reports, 9(1):2285.

[42]ShengG, HuG, WangX, et al., 2025. Cross-layer frequency-spatial domain feature interaction awareness fusion for fine-grained visual classification. Information Fusion, 103788.

[43]SiegelRL, MillerKD, WagleNS, et al., 2023. Cancer statistics, 2023. CA: A Cancer Journal for Clinicians, 73(1):17-48.

[44]SiegelRL, GiaquintoAN, JemalA, 2024. Cancer statistics, 2024. CA: A Cancer Journal for Clinicians, 74(1):12-49.

[45]SinghY, EatonJE, VenkateshSK, et al., 2025. Deep learning analysis of MRI accurately detects early-stage perihilar cholangiocarcinoma in patients with primary sclerosing cholangitis. Hepatology, 83(1):30-39.

[46]SkergetS, PenaherreraD, ChariA, et al., 2024. Comprehensive molecular profiling of multiple myeloma identifies refined copy number and expression subtypes. Nature Genetics, 56(9):1878-1889.

[47]Soltanian-ZadehS, SahingurK, BlauS, et al., 2019. Fast and robust active neuron segmentation in two-photon calcium imaging using spatiotemporal deep learning. Proceedings of the National Academy of Sciences of the United States of America, 116(17):8554-8563.

[48]SunLJ, KeXD, GuanA, et al., 2023. Intratumoural microbiome can predict the prognosis of hepatocellular carcinoma after surgery. Clinical and Translational Medicine, 13(7):e1331.

[49]SunY, ZhaoZC, JiangD, et al., 2022. Low-illumination image enhancement algorithm based on improved multi-scale retinex and ABC algorithm optimization. Frontiers in Bioengineering and Biotechnology, 10:865820.

[50]SyrykhC, van den BrandM, KatherJN, et al., 2025. Role of artificial intelligence in haematolymphoid diagnostics. Histopathology, 86(1):58-68.

[51]TaoYF, ZhouXL, SunLQ, et al., 2023. Highly efficient and robust π-FISH rainbow for multiplexed in situ detection of diverse biomolecules. Nature Communications, 14(1):443.

[52]TeierleSM, HuangY, KittaiAS, et al., 2023. Characteristics and outcomes of patients with CLL and CDKN2A/B deletion by fluorescence in situ hybridization. Blood Advances, 7(23):7239-7242.

[53]UedaHR, ErtürkA, ChungK, et al., 2020. Tissue clearing and its applications in neuroscience. Nature Reviews Neuroscience, 21(2):61-79.

[54]VincentL, DoughertyER, 1994. Morphological segmentation for textures and particles. In: Dougherty ER (Ed.), Digital Image Processing Methods. CRC Press, Boca Raton, USA, p.43-102.

[55]WangL, YouZH, LuW, et al., 2024. Attention-aware sobel graph convolutional network for remote sensing image change detection. IEEE Transactions on Geoscience and Remote Sensing, 62:4409912.

[56]XieXY, CuiQ, JiangTT, et al., 2024. A critical role of the endothelial S-phase kinase-associated protein 2/phosphatase and tensin homologue axis in angiogenesis and psoriasis. British Journal of Dermatology, 190(2):244-257.

[57]XuLS, ZhangHR, WangJL, et al., 2023. Information loss challenges in surgical navigation systems: from information fusion to AI-based approaches. Information Fusion, 92:13-36.

[58]ZhangB, QiuLY, XiaoW, et al., 2021. Reconstruction of the hypothalamo-neurohypophysial system and functional dissection of magnocellular oxytocin neurons in the brain. Neuron, 109(2):331-346.e7.

[59]ZhangRH, KhareP, BanerjeeP, et al., 2024. The DLEU2/miR-15a/miR-16-1 cluster shapes the immune microenvironment of chronic lymphocytic leukemia. Blood Cancer Journal, 14(1):168.

[60]ZhaoZJ, YangWB, KongR, et al., 2023. CircEIF3I facilitates the recruitment of SMAD3 to early endosomes to promote TGF-β signalling pathway-mediated activation of MMPs in pancreatic cancer. Molecular Cancer, 22(1):152.

[61]ZhengHD, SunYL, KongDW, et al., 2022. Deep learning-based high-accuracy quantitation for lumbar intervertebral disc degeneration from MRI. Nature Communications, 13(1):841.

[62]ZhongGL, ZhaoQ, ChenZL, et al., 2023. TGF-β signaling promotes cervical cancer metastasis via CDR1as. Molecular Cancer, 22(1):66.