Deep learning enables rapid and robust analysis of fluorescence lifetime imaging in photon-starved conditions

Chen, Yuan‐I; Chang, Yin-Jui; Liao, Shih-Chu Jeff; Nguyễn, Tùng Lâm; Yang, Jianchen; Kuo, Yu‐An; Hong, Suk-In; Liu, Yen-Liang; Rylander, H. G.; Santacruz, Samantha R.; Yankeelov, Thomas E.; Yeh, Hsin‐Chih

doi:10.1101/2020.12.02.408195

Cited by 4 publications

(2 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As NLSF was usually adopted in previous studies [ 25 , 27 , 35 ], we compared the inference performances of ELM and deconvolution-based NLSF (implemented with lsqcurvefit (·) function in MATLAB using iterative Levenberg–Marquardt algorithm) in Figure 2 . As such, 2000 simulated testing datasets were generated for recovery for single and double lifetimes.…”

Section: Synthetic Data Analysismentioning

confidence: 99%

Fast Analysis of Time-Domain Fluorescence Lifetime Imaging via Extreme Learning Machine

Zang

Xiao

Wang

et al. 2022

Sensors

View full text Add to dashboard Cite

We present a fast and accurate analytical method for fluorescence lifetime imaging microscopy (FLIM), using the extreme learning machine (ELM). We used extensive metrics to evaluate ELM and existing algorithms. First, we compared these algorithms using synthetic datasets. The results indicate that ELM can obtain higher fidelity, even in low-photon conditions. Afterwards, we used ELM to retrieve lifetime components from human prostate cancer cells loaded with gold nanosensors, showing that ELM also outperforms the iterative fitting and non-fitting algorithms. By comparing ELM with a computational efficient neural network, ELM achieves comparable accuracy with less training and inference time. As there is no back-propagation process for ELM during the training phase, the training speed is much higher than existing neural network approaches. The proposed strategy is promising for edge computing with online training.

show abstract

Section: Synthetic Data Analysismentioning

confidence: 99%

Fast Analysis of Time-Domain Fluorescence Lifetime Imaging via Extreme Learning Machine

Zang

Xiao

Wang

et al. 2022

Sensors

View full text Add to dashboard Cite

show abstract

“…3-D [8] and 1-D CNNs [9] were proposed to process entire 3-D tensors and pixel-wise 1-D histograms, respectively. A generative adversarial network [10] was utilized in photon-starved conditions. Thanks to high hardware integration technologies, modern time-correlated singlephoton counting (TCSPC) systems [11] have been integrated on a single board.…”

Section: Introductionmentioning

confidence: 99%

Hardware Inspired Neural Network for Efficient Time-Resolved Biomedical Imaging

Zang

Xiao

Wang

et al. 2022

2022 44th Annual International Conference of the IEEE Engineering in Medicine &Amp; Biology Society (EMBC)

View full text Add to dashboard Cite

Longitudinal FRET Imaging of Glucose and Lactate Dynamics and Response to Therapy in Breast Cancer Cells

et al. 2021

Self Cite

View full text Add to dashboard Cite

The reprogramming of cellular metabolism is a hallmark of cancer. The ability to noninvasively assay glucose and lactate concentrations in cancer cells would improve our understanding of the dynamic changes in metabolic activity accompanying tumor initiation, progression, and response to therapy. Unfortunately, common approaches for measuring these nutrient levels are invasive or interrupt cell growth. This study transfected FRET reporters quantifying glucose and lactate concentration into breast cancer cell lines to study nutrient dynamics and response to therapy. Procedures: Two FRET reporters, one assaying glucose concentration and one assaying lactate concentration, were stably transfected into the MDA-MB-231 breast cancer cell line. Correlation between FRET measurements and ligand concentration were measured using a confocal microscope and a cell imaging plate reader. Longitudinal changes in glucose and lactate concentration were measured in response to treatment with CoCl 2 , cytochalasin B, and phloretin which, respectively, induce hypoxia, block glucose uptake, and block glucose and lactate transport. Results: The FRET ratio from the glucose and lactate reporters increased with increasing concentration of the corresponding ligand (p < 0.005 and p < 0.05, respectively). The FRET ratio from both reporters was found to decrease over time for high initial concentrations of the ligand (p < 0.01). Significant differences in the FRET ratio corresponding to metabolic inhibition were found when cells were treated with glucose/lactate transporter inhibitors. Conclusions: FRET reporters can track intracellular glucose and lactate dynamics in cancer cells, providing insight into tumor metabolism and response to therapy over time.

show abstract

Deep learning enables rapid and robust analysis of fluorescence lifetime imaging in photon-starved conditions

Cited by 4 publications

References 60 publications

Fast Analysis of Time-Domain Fluorescence Lifetime Imaging via Extreme Learning Machine

Fast Analysis of Time-Domain Fluorescence Lifetime Imaging via Extreme Learning Machine

Hardware Inspired Neural Network for Efficient Time-Resolved Biomedical Imaging

Longitudinal FRET Imaging of Glucose and Lactate Dynamics and Response to Therapy in Breast Cancer Cells

Contact Info

Product

Resources

About