The increased use of cells in biomanufacturing and as therapeutic products over the last ten years has urged the development and publication of two ISO Cell Counting Standards, ISO 20391 – 1:2018 and ISO 20391 – 2:2019.
The aim is to offer better guidance on general principles relating to cell counting and to create an approach to sufficiently assess the quality of cell counting methods.
This work demonstrates the practical implementation of the experimental protocol detailed in ISO Cell Counting Standard Part 2 and a BlandAltman comparative analysis to assess the performance and comparison of cell counting methods.
Two cell types, two image cytometry instruments, and two fluorescent stains are compared to evaluate cell counting method performance and thus calculate the precision, coefficient of determination (R^{2}), and a proportionality index (PI) parameter.
In addition, the cell counting results are subjected to direct comparison to evaluate bias between two cell counting methods. The protocol is appropriate for evaluating and comparing the performance of multiple cell counting methods to select for downstream assays.
In the recent decade, cell and gene therapies have significantly improved their efficacy and have become key players in cancer treatment.^{1,2} With two chimeric antigen receptor (CAR) T cell therapies approved by the U.S. Food and Drug Administration (FDA) in 2017, the numbers of clinical studies and tests on new and novel cell therapy products have also increased.^{3–5}
Generally speaking, cellular therapies necessitate genetic modification of the immune cells (i.e., T cells, NK cells) acquired from patients, culture expansion, and reintroduction of the final products back into the patients.
Therefore, it is of key importance to offer precise cell counting for the administration of proper dosages, which may otherwise result in inefficacy or provoke unwanted autoimmune responses in patients receiving therapeutic treatments.^{6–8}
In the 21st Century Cures Act, the United States Congress has also acknowledged the importance of standardization for simplifying development, quality assurance, and facilitating regulatory approval of cell and gene therapy products.^{9}
In the “Synergizing Efforts in Standards Development for Cellular Therapies and Regenerative Medicine Products” workshop led by the FDA on March 31st, 2014, cell counting and viability measurement assurance was recognized as opportunities for standards development.^{10,11}
ISO has since published two cell counting standards, “ISO 203911:2018 Biotechnology – Cell Counting – Part 1: General Guidance on Cell Counting Methods” and “ISO 203912:2019 Biotechnology – Cell Counting – Part 2: Experimental Design and Statistical Analysis to Quantify Counting Method Performance”.
These standards can act as guidance for researchers working in the field of immunotherapy and adoptive cell therapy, where both necessitate high quality and robust cell counting measurements for biologics and cell products.^{12,13}
Derived from standard concepts described in ISO Cell Counting Standard Part 1, it is proposed that 6 key factors can offer guidance on the selection of cell counting methods and enhance the cell counting measurement quality:
 Establish the intended purpose of the cell counting result (e.g., cell count for normalization of bioassays, posttumor digestion for single cellbased transcriptome analysis, cell therapy dosing, mouse tissue processing for cytotoxicity assays or isolation of human PBMCs for immunophenotyping analysis, etc.)
 Evaluate to understand cell sample composition (e.g., chemical impurities, particle debris, various cell types, and suspension medium), as well as the morphological appearances of the cells when subjected to microscopy.
 Comprehend the assay principles and select the cell counting assay most suitable for the application, such as total, live and dead cell count, viability, or cell population analysis
 Study the capabilities and choose the appropriate cell counting systems, where the system is made up of reagents, consumables, instrument and software algorithms, as well as assay performance criteria (i.e., precision, range, linearity, etc.)
 Treat each cell counting method as a complete process, including sampling, diluting, and staining, which are crucial for carrying out appropriate sample preparation
 Deliver continuous operator training to guarantee consistent cell counting results
It is also acknowledged that the cell counting needs for cell and gene therapies are broad due to an extensive range of biological sample types with different formulations and bioprocessing steps, which are intricate, dynamic, and heterogeneous.
Since there are no current reference materials for live mammalian cells that are certified for cell concentration, the accuracy parameter detailed in the ICH Harmonized Tripartite Guideline – Validation of Analytical Procedures: Text and Methodology Q2 (R1) cannot be freely applied, thus increasing the challenge and difficulty of validating the cell counting accuracy.^{14–16}
Therefore, the ISO cell counting standards can act as an invaluable tool to assess and select cell counting methods that are fitforpurpose to boost confidence in the cell counting results.
Development of the protocol
Therefore, employing the guidance from ISO 20391 – 2:2019 Biotechnology – Cell Counting – Part 2 and utilizing information from the ICH Q2 (R1) enabled the development of an appropriate protocol to assess the performance of selected cell counting methods.
The ICH Q2 (R1) guidance document puts forward multiple parameters for the verification of analytical methods, including linearity, detection range, robustness, limits of detection (LOD), limits of quantitation (LOQ) and accuracy (repeatability, intermediate precision, reproducibility).
It is crucial to note that since no reference material exists to offer a reference value for cell concentration, assessment of the accuracy parameter needs to be indirectly evaluated by orthogonal comparative methods.
The ISO Cell Counting Standard Part 2 outlines a detailed protocol to simultaneously evaluate precision (repeatability), coefficient of determination (R^{2} ) and proportionality. Using the ISO Cell Counting Standard Part 2 document, several key parameters have been identified that can sufficiently assess the performance of cell counting methods in quick succession.^{17,18}
This article focuses on an experimental protocol taken from the ISO Cell Counting Standard Part 2, which assesses the coefficient of determination (R^{2} value), precision (repeatability – coefficient of variation [CV]) and proportionality index (PI) of a cell counting method.
The proportionality index is a metric presented in the ISO Cell Counting Standard Part 2 that quantifies the degree to which a cell counting method is in accordance with the principle of proportionality, where it is anticipated that cell counts will proportionally scale with dilution.
The principle of proportionality is a basic underlying property of any cell counting method, and any deviation from proportionality would signal a systematic or nonsystematic error generating a loss of measurement accuracy.
To better evaluate systematic deviation from proportionality, which is a sign of loss of accuracy, the PI is determined by fitting a proportional model to the dilution series data, then summarizing residuals predicated on smoothed data, thus limiting the influence of random variation on the evaluation of proportionality.^{17}
There are numerous approaches to determine PI, where some PI metrics may be more appropriate based on the fitforpurpose need of the cell counting method. Some metrics penalize more for outliers, while others evenly weigh errors across the dilutions or enable more contribution by higher cell concentrations.
In this work, the PI model was used, as previously published by the National Institute of Standards and Technology (NIST).^{17} Note that sources of systematic error that are proportional to sample dilution cannot be determined with this approach.
For instance, if debris is mixed up with the cell suspension and erroneously identified as cells, the concentration of both cells and debris would be reduced proportionally with dilution, and the false counts would not influence the proportionality.
To demonstrate the suitable usage of the proposed experiments, these protocols were tested employing various image cytometry systems from Nexcelom Bioscience LLC. (Lawrence, MA).
BlandAltman comparative analysis method
Comparative analysis methods can be utilized for the performance comparison of different cell counting methods. While the lack of reference material prevents the direct measurement of cell counting accuracy, comparison of orthogonal methods may act as a possible alternative.
It is also often best to identify how closely the results of one method will be in agreement with another, such as when an instrument is upgraded after several years in the lab. One useful method is the production of a Tukey meandifference plot, also known as a BlandAltman plot.^{20–22}
The BlandAltman analysis gives rise to the calculation of a bias (with corresponding confidence interval) between two methods, signaling which method counts higher or lower on average and by how much.
The analysis also offers an estimate of what to expect with the correlation between the two methods for a single sample.
Through modifying the dilution series experimental design as described in ISO Cell Counting Standard Part 2 document, it was possible to acquire data appropriate for a BlandAltman analysis while also falling in line with the standards requirements for calculating CV, R^{2} and PI.
Typically, BlandAltman plots are comprised of absolute differences between two measurements plotted against their average. In the case of cell counting, variance is not constant for different concentrations but is usually proportional to the number of cells counted.^{17}
For BlandAltman analysis to be suitable for such an application, the data can be transformed to accomplish an approximate constant variance across a range of concentrations. In this protocol, percent differences were used rather than absolute differences to acquire a more uniform variance.
The BlandAltman analysis method generates three metrics of comparison:
 The bias between two methods, which is the mean of the differences.
 The limits of agreement (LoA) are a multiple of the standard deviation of the differences.
 The confidence interval (CI) of the bias, which is a multiple of the standard error of the mean of the differences.
The bias details the mean difference between measurement results acquired via the two methods. As a result of biological variation in the samples and variability in the measurement process for both methods, it is not possible to precisely predict how much the measurement of any single sample will differ between the two methods.
However, when measurements of several samples are averaged, a bias – even a small one – may emerge. The bias can be interpreted as one method measuring higher or lower than another on average. However, when measuring a single sample, the difference between methods may vastly differ.
The limits of agreement describe how vastly these differences may vary. When introduced to and subtracted from the bias, the LoA determines a range within which the difference between the measurements from two methods of a single sample is expected to be identified.
In this protocol, the limits of agreement that estimate a 95% confidence interval (1.96 × standard deviations) for normal distribution were used. This is an appropriate estimation for such purposes and a 72measurement sample size.
If fewer measurements are obtained, confidence intervals determined from the appropriate t distribution (rather than the Normal distribution) are recommended. If the percent differences between the results from the two methods trail a normal distribution, it is to be expected that 95% of the differences will fall within one LoA from the value of the bias.
The reality is that the values will not be normal, but the approximation is practical when evaluating subsequent measurements.^{23} If more statistical rigor is necessary, tests for normality can be employed, and the confidence intervals can be calculated more exactly.^{24}
Depending on the variation seen between samples relative to the variation between replicate measurements from each sample, it may be useful to involve random effects terms usually included in the analysis of hierarchical experiments.
This study was not concerned with the sampletosample variation in the proposed experiments.
The confidence interval of the bias delivers approximate uncertainty for the calculated bias value and indicates a range within which the true value of the bias between the two cell counting methods can be potentially found.
Contrary to the LoA, this confidence interval is constrained with an increased number of samples measured. If the 95% confidence interval is greater than the total value of the bias (i.e. the CI brackets the value 0), the method comparison has not shown a statistically significant bias between the two methods (at α = 0.05 significance level).
With enough samples, even the slightest bias may be measured confidently. A slight bias is often nominal in comparison to sample variation.
Researchers should consider how a cell count is being utilized to establish acceptable levels of bias in their case. Before moving on with BlandAltman comparative analysis for cell counting, researchers should remember the following:
 Establish the range of cell concentration values for which comparison between the two methods is preferable;
 Determine what values of the bias and LoA are appropriate for their application,
 Select cell samples that are representative of the population for which the comparison is required and the range determined in step 1; and
 Measure each sample using the various cell counting methods, taking care that the sample does not alter between measurements (minimal delay between measurements, proper mixing, etc.).^{25}
A greater number of paired measurements can limit the uncertainties of the bias and LoA, e.g. the confidence interval of the bias lessens with further measurements. Researchers should calculate the precision they require and increase the number of paired measurements as a result – a minimum of 20 paired measurements should be used as a recommended starting point.
Finally, it is feasible to consider that either the bias or the variation will vary with cell concentration. In such an instance, the bias and LoA captured for the entire group of data may not be representative of how the two methods compare over a smaller range of concentrations. It may be worthwhile performing BlandAltman analysis on smaller subsets of data.
Applications of the method
The cell counting method performance comparison and evaluation protocols can be utilized for research, analytical method development, process development and preclinical or clinical trials.
Additionally, the method can be applied to a wide range of research fields requiring the usage of cells, such as cellular and gene therapy, cell line development and biologics production, immunooncology and immunotherapy, virology and infectious disease, toxicology, regenerative medicine, food science and even renewable energy.
The quality of cell counting results is crucial for a broad range of cell types used in the research fields detailed above.
These cell types can be comprised of primary cells, such as human or mouse whole blood, bone marrow aspirate, adipose tissue, cord blood, hepatocytes, PBMCs, platelets, leukapheresis sample, tumor or tissue digests. Moreover, bacteria and yeast cells are typically used to generate biologics or used for beverage production.
Experimental design
The cell counting method performance assessment proposed here is comprised of a dilution series experiment and comparative analysis for numerous methods. The experimental design is shown using CHOS and Jurkat cell lines fluorescently stained with acridine orange and a green nuclear dye.
A comparison of two cell counting systems – the Cellaca MX HighThroughput Cell Counter (Cellaca MX) and the Celigo Image Cytometer (Celigo) – is performed.
It is crucial to note that ISO Cell Counting Part 2 requires users to evaluate pipetting error contributions to dilution integrity to determine confidence in dilution and sampling.
Here, a preevaluation was conducted, which took into account pipetting error, which is not described in this protocol. It is also crucial to evaluate the stability of the target cell sample prior to performing the experiment to prevent drift in concentration and viability during the assay time frame.
Figure 1. Cell sample preparation and measurement process diagram. Procedure sequence from left to right: (1) Collect your target cell sample and prepare different concentrations with specific dilution fractions using cell media. (2) Repeat this process to generate 3 replicates for each dilution fraction. (3) Label each tube in random order from 1–18. (4) Prepare and measure each tube 4 times with each selected cell counting method. (5) Analyze the images with each cell counting method to generate cell counting results. (6) Utilize the cell counting results to generate proportionality index (PI), coefficient of variation (CV), and coefficient of determination (R^{2}). Image Credit: Nexcelom Bioscience LLC
The Jurkat and CHO cells’ stability used in this work has been tested previously and showed no discernable trends (Supplementary Figure 1).
The dilution experiment is made up of a 6point concentration series of the target cell types, where each concentration is produced independently from the original stock (rather than the other dilutions) to limit propagation of dilution error that can influence proportionality.
The dilution series should span the conventional concentration range of the target cell samples to assess the performance of the cell counting method in the range as specified.
Three replicate samples are produced per concentration, and each replicate sample is measured 4 times per cell counting method so that each method offers a total of 12 measurements per concentration and a total of 72 measurements in a 6point concentration series.
The measurements are then used to establish the coefficient of determination (R^{2}), precision (repeatability – Coefficient of Variation, CV) and proportionality index (PI) parameters for each cell counting method.
It is crucial to note that the tested Jurkat and CHO cells were stained with acridine orange and a Nuclear Green dye to measure only the overall cell concentration in this study. The BlandAltman method should be applied for performance comparison between two cell counting methods.
Like the proportionality measurement, the comparison results are only valid for the intended use of the particular methods (cell type, assay type, exact instruments, etc.) and only for the range of cell concentrations incorporated into the test.
Therefore, it is crucial to first define the precise methods, test conditions and range of cell concentrations over which comparison is preferred.
For the most accurate results, the BlandAltman analysis should include as many measurements as possible, encompassing the sources of variation that are to be anticipated for the normal operation of the cell counting method, such as multiple operators, reagent lots and cell culture flasks.
Each point on the BlandAltman plot is acquired using both cell counting methods to measure a single sample. The sample should be mixed carefully to guarantee homogeneity before portions are extracted for measurement with each cell counting method.
Measurements should be made simultaneously with minimal lag time between them if possible. If the experiment described above is conducted using the same tubes of cells used in both methods at the same time, BlandAltman analysis may be performed with the resulting data.
If preferred, tighter confidence intervals on the calculated bias or less uncertainty on the Limits of Agreement can be acquired by supplementing the data with additional samples. Concentrations covering the selected concentration range should be equally represented in the samples used.
Expertise needed to implement the protocol
Generally speaking, the expertise necessary to apply the cell counting method performance evaluation is proper training by an expert user in the operation of the cell counting systems. Additionally, the users should be trained on sample preparation to guarantee the consistent performance of the dilution, sampling and staining steps of the cell counting process.
Limitations
Accuracy is one of the leading parameters for the validation of an analytical method; however, it cannot be applied directly to a majority of cell counting methods. Since there are limited live cell reference standards, attempting to assess the accuracy of a cell counting method can be challenging.
Therefore, proportionality is a viable alternative parameter when evaluating accuracy relative to dilution fraction, which acts as the internal control, as well as utilizing orthogonal methods for comparison. It should be acknowledged that R^{2} values calculated over a range of concentrations are strongly dependent on the range chosen.
A greater range of linear data results in an R^{2} value closer to 1. If an attempt is made to contrast R^{2} values, it is of utmost importance that the range for the two calculations remains the same.
Additionally, it should be reported that the proportionality index, as detailed here, is not normalized to the number of dilution fractions and the number of biological replicates per dilution fraction.
It is necessary that the equivalent experimental design be used if PI is to be compared meaningfully between methods.
Materials
Documentation materials
 ISO 203911:2018 Biotechnology – Cell Counting – Part 1: General Guidance on Cell Counting Methods
 ISO 203912:2019 Biotechnology – Cell Counting – Part 2: Experimental Design and Statistical Analysis to Quantify Counting Method Performance
Biological materials
 Chinese Hamster Ovary (CHOS) cell line (Gibco, #11619012)
 Jurkat, Clone E61 cell line (ATCC, TIB152™)
Growth medium & supplements
 Antibiotic Antimycotic Solution (100X) (SigmaAldrich, #A5955100ML)
 CD CHO Medium (1X) (Gibco, #10743011)
 Fetal Bovine Serum (FBS) (Access, #A19023)
 GlutaMAX1 (100X) (Gibco, #35050061)
 HT Supplement (100X) (Gibco, #11067030)
 RPMI Medium 1640 (1X) (Gibco, #11875093)
Fluorescent staining reagents
 ViaStain™ AOPI Staining Solution (AOPI, Nexcelom Bioscience, CS201065mL)
 ViaStain™ AO Staining Solution (AO, Nexcelom Bioscience, CS201085mL)
 ViaStain™ Total Cell Nuclear Green (Nuclear Green, Nexcelom Bioscience, CS1V00081)
Other reagents & chemicals
 HyClone™ Water, Cell Culture Grade (EndotoxinFree) (GE Health, #SH3052903)
 Phosphate Buffered Saline (PBS) powder (SigmaAldrich, #P38135)
Equipment
 Automatic pipettor (Fisherbrand™ Pipet Controller, #FB14955202)
 Cell culture incubator (Thermo, Forma 370)
 Plate rocker (Boekel, RockerII 260350)
 Centrifuge (Eppendorf, 2702)
 Cellometer Spectrum and operating laptop computer (Spectrum, Nexcelom Bioscience)
 Cellaca MX HighThroughput Automated Cell Counter and operating laptop computer, concentration range of 1 x 10^{5} – 1 x 10^{7} cells/mL (Nexcelom Bioscience)
 Celigo Image Cytometer and operating desktop computer (Nexcelom Bioscience)
 Manual pipettors (P10, P100, P1000) (VWR, 110UL, 10100UL, 1001000UL)
 Tissue culture hood (Forma Scientific, ClassII A/B3 BSC)
Disposable instruments
 15mL centrifuge tube (Greiner Bio, 188271)
 Cell counting slides (Nexcelom Bioscience, CHT4SD100002)
 Cellaca MX Highthroughput Automated Cell Counter Plates (Cellaca MX plates, Nexcelom Bioscience, CHM24A100001)
 Microtubes 1.5 mL (VWR, 89000028)
 Microtubes 0.5 mL (CellTreat, 229440)
 Pipette tips (P10 and P1000) (VWR, 7320561, 83007380)
 Pipette tips (P200) (USA Scientific, 11111210)
 T75 cm^{2} flask (USA Scientific, CC768248)
 Serological Pipets 5 mL, 10 mL, 25 mL (USA Scientific, #10750110, #10710810, #10725410)
Reagent setup
PhosphateBuffered Saline (PBS) solution
Formulate the PBS solution by mixing 5 L of H_{2}O with 1 packet of PBS powder to produce a solution of 0.01M PBS at pH 7.4 with NaCl at 0.138 M and KCl at 0.0027 M.
CHOS cell culture medium
Prepare CHOS medium (500 mL) using the CD CHO Medium (1X) and an addition of 5 mL of the GlutaMAX1 (100X) and 5 mL of the HT Supplement (100X).
Jurkat cell culture medium
Prepare Jurkat medium (500 mL) with the RPMI Medium 1640 (1X) and supplement with 10% FBS (50 mL) and 5 mL of the Antibiotic Antimycotic Solution (100X).
ViaStain™ AOPI staining solution
The acridine orange (AO) and propidium iodide (PI) staining solution is readily prepared to the correct concentration prior to staining at 1:1 with the cells.
ViaStain™ AO staining solution
The acridine orange (AO) staining solution is readily prepared to the correct concentration prior to staining at 1:1 with the cells.
ViaStain™ Total Cell Nuclear Green staining solution
Preparation of a 2X staining solution (10 µM) by mixing PBS and the Nuclear Green stock solution at 5 mM. Pipette 10 mL of PBS into a 15mL centrifuge tube and add 20 µL of the Total Cell Nuclear Green stock solution. Close the 15mL centrifuge tube and invert 10X to mix the staining solution prior to use.
Equipment setup
Cellometer Spectrum
Using a USB cable, connect the Cellometer Spectrum to the main laptop computer and plug in the power cord. Turn the power on from the back side and then find and open the Cellometer Spectrum analysis software. Once in the Cellometer Spectrum software system, select the “AOPI Viability Assay_S5” default assay type for cell counting.
Cellaca™ MX HighThroughput Automated Cell Counter
Pair the Cellaca MX with the operating laptop computer using the USB cable and plug in the power cord. Turn the power on the computer from the back side and then start the Cellaca MX analysis software. In the Cellaca MX software (v1.2), select the “MX04.0_ AOPI_LiveDead” default assay type for cell counting.
Celigo^{®} Image Cytometer
Turn on the Celigo power on from the front and open the Celigo analysis software. Navigate to the top right and select ‘Administration’ and then click ‘Manage Plate Profiles’. After the “Plate Profile Management” window opens, select the ‘Import’ button and click the plate profile for Cellaca 12 × 2 plate. Return to the home screen for image acquisition and analysis.
Procedure
Maintenance of CHOS cells
Timing: 20 – 30 minutes for passaging the cells and measuring their concentration and viability.
 Transport the CHOS cells when they reach between 2 to 4 × 10^{6} cells/mL. Allowing the cells to grow above that concentration may diminish cell division as well as decrease viability due to poor nutrients in the media.
 Heat the CHOS cell culture medium at 37 °C for 15 minutes in the incubator or in a water bath at 37 °C for 5 minutes before transporting.
 Under the biosafety cabinet, using a 10 mL pipette, pipette up and down at least 10 times to separate the cell clumps and generate a homogenous cell suspension in the T75 flask.
 Remove 200 µL of cells from the T75 flask and move them into a 1.5 mL microtube before the cells have a chance to settle.
 Acquire a CHT4SD100 cell counting slide and removed the protective plastic film on the top and bottom, and place the slide on a KimWipe.
 Mix 20 µL of CHOS cell sample and 20 µL of AOPI within a 0.5 mL microtube.
 Pipette 20 µL of stained cell sample into one chamber on the cell counting slide.
 Insert the cell counting slide into the Spectrum and select the “AOPI Viability Assay_S5”.
 Measure the cell concentration and viability.
 Based on the measured concentration, determine the ratio of cells to new media that is necessary to achieve a concentration of 2 × 10^{5} cells/mL.
 Remove the calculated cell volume from the flask and substitute for an appropriate amount of warmed CHOS cell culture medium.
 Place the passaged flask back onto the plate rocker inside the 8% CO_{2} incubator at 37 °C.
 Monitor the growth of cells daily, and continue to passage as and when required (usually 3 times a week).
Maintenance of Jurkat cells
Timing: 20–30 minutes for passaging the cells and measuring their concentration and viability.
 Passage the Jurkat cells when they are between 1 to 2 × 10^{6} cells/mL. Enabling the cells to grow above that concentration may limit cell division as well as decrease viability due to poor nutrients in the media.
 Warm the Jurkat cell culture medium at 37 °C for 15 minutes in the incubator or in a water bath at 37 °C for 5 minutes before passaging.
 Under the biosafety cabinet, use a 10 mL pipette, pipette up and down at least 10 times to break up the cell clumps, and create a homogenous cell suspension in the T75 flask.
 Remove 200 µL of cells from the T75 flask and transfer them into a 1.5 mL microtube before the cells have had a chance to settle.
 Acquire a CHT4SD100 cell counting slide, remove the protective plastic film on the top and bottom and place the slide on a KimWipe.
 Mix 20 µL of Jurkat cell sample and 20 µL of AOPI within a 0.5 mL microtube.
 Pipette 20 µL of stained cell sample into one chamber on the cell counting slide.
 Insert the cell counting slide into the Spectrum and select “AOPI Viability Assay_S5”.
 Measure the cell concentration and viability.
 Based on the measured concentration, determine the ratio of cells to new media that is needed to achieve a concentration of 2 × 10^{5} cells/mL.
 Remove the calculated cell volume from the flask and substitute for an appropriate amount of warmed Jurkat cell culture medium.
 Place the passaged flask back inside the 5% CO_{2} incubator at 37 °C.
 Monitor the growth of cells daily, and continue to passage as needed (usually 3 times a week).
Stock cell sample preparation from cell culture
Timing: 15 minutes for collecting the cells from cell culture flasks, 5 minutes for cell counting and viability analysis and 10 minutes for adjusting cell sample concentration if necessary.
 Collect a stock of CHOS and Jurkat cell sample separately into a 15mL tube from cell culture following aseptic techniques.
 Obtain a CHT4SD100 cell counting slide, peel off the protective plastic film on the top and bottom and place the slide on a KimWipe.
 Pipette 20 µL of the cell sample using a P100 pipettor into a 0.5 mL microtube.
 Pipette 20 µL of the AOPI and add to the 0.5 mL microtube.
 Aspirate the mixture of cells and AOPI up and down at least 5 times.
 Pipette 20 µL of the stained cells into one chamber on the cell counting slide.
 Insert the cell counting slide into the Spectrum and count the stained cells to generate cell count and viability.
 Amend the stock cell sample concentration to ~5 × 10^{6} cells/mL for both CHOS and Jurkat cells.
 Decrease the concentration by dilution in cell media.
 Increase the concentration by centrifugation and resuspend in cell media.
 Repeat steps 28–33 to make sure the concentration is adjusted to ~5 × 10^{6} cells/ mL.
Sample preparation & cell counting preparation for cell counting methods performance evaluation & comparison
Timing: 15–30 minutes with a single, manual pipette for sample preparation. 15–20 minutes for incubation of cell samples when mixing with Nuclear Green (Figure 1). ~10 minutes per Cellaca MX plate with a single, manual pipette for preparation of cell counting.
It is critical to note that under the guidance of ISO Cell Counting Part 2, an introductory accuracy validation experiment of pipetting volume using the experimental pipettors is needed to boost sampling confidence.
This kind of validation can be conducted using a sensitive and wellcalibrated laboratory balance and a fluid of known density, but the procedure will not be highlighted in this protocol.
Directly dilute the cells to produce independent dilution samples rather than serial dilution to reduce the propagation of pipetting error that can influence proportionality.
 Acquire the prepared stock CHOS cell and Jurkat cell samples at the highest concentration for the intended use and range (~5 × 10^{6} cells/mL).
 Prepare other samples from the stock of CHOS and Jurkat cell samples at 0.1, 0.3, 0.5, 0.7, 0.9 and 1.0 dilution fractions (DFs) respectfully (Table 1).
 Prepare replicate samples with cell culture media or PBS.
 Pipette 120 µL of CHOS or Jurkat stock cell sample into the 1st microtube for the 1.0 DF sample.
 Pipette 108 µL of CHOS or Jurkat stock cell sample into the 2nd microtube and add 12 µL of PBS for the 0.9 DF sample.
 Pipette 84 µL of CHOS or Jurkat stock cell sample into the 3rd microtube and add 36 µL of PBS for the 0.7 DF sample.
 Pipette 60 µL of CHOS or Jurkat stock cell sample into the 4th microtube and add 60 µL of PBS for the 0.5 DF sample.
 Pipette 36 µL of CHOS or Jurkat stock cell sample into the 5th microtube and add 84 µL of PBS for the 0.3 DF sample.
 Pipette 12 µL of CHOS or Jurkat stock cell sample into the 6th microtube and add 108 µL of PBS for the 0.1 DF sample.
 Repeat Steps 38–43 two more times to produce a total of 3 replicate samples at each DF, where a total of 18 tubes of cell samples are generated.
 Pipette 120 µL of AO staining solution into the 1st microtube of each DF sample to create a 1:1 mixed sample. After this step, a total of 240 µL cell sample is prepared in the 1st microtube at each DF.
 Invert the 0.1 DF microtube 10 times to make sure the mixture is uniform.
 Transfer 50 µL from the mixed 0.1 DF microtube into the A1 loading well on the 1st Cellaca MX plate. Repeat the transfer 3 more times into the A2 – A4 loading wells of the 1st Cellaca MX plate.
 Repeat Step 46–47 for the 1st microtubes of the residual DFs (0.3, 0.5, 0.7, 0.9, and 1.0) samples into the remaining loading wells on the 1st Cellaca MX plate, following the plate map as displayed below. Subsequently, the 1st Cellaca MX plate should be prepared (Table 2).
 Randomize Step 46–48 if applicable. This is recommended by ISO Cell Counting Standard Part 2 to reduce the systematic timedependence effects on the proportionality index and other metrics of the cell counting measurement process quality.
 Repeat Step 45–48 for the 2nd and 3rd replicate samples at various DFs to prepare the 2nd and 3rd Cellaca MX plates.
 Prepare each Cellaca MX plate immediately before the image capture, instead of preparing all Cellaca MX plates at the beginning, to reduce the time gap between sample preparation and image acquisition.
 Repeat Steps 36–49 and stain with 120 µL of Nuclear Green.
 Incubate the Nuclear, Greenstained cell samples for 15–20 minutes at room temperature. Incubation time can be reduced at 37 °C.
Table 1. Dilution fractions and the corresponding volumes preparation for cell sample and PBS. Source: Nexcelom Bioscience LLC
DF 
Cell volume (μL) 
PBS volume (μL) 
1.0 
120 
0 
0.9 
108 
12 
0.7 
84 
36 
0.5 
60 
60 
0.3 
36 
84 
0.1 
12 
108 
Table 2. Cellaca plate map for cell samples at different DFs. Source: Nexcelom Bioscience LLC
Plate 1 

1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
A 
0.1 
0.1 
0.1 
0.1 
0.3 
0.3 
0.3 
0.3 
0.5 
0.5 
0.5 
0.5 
B 
0.7 
0.7 
0.7 
0.7 
0.9 
0.9 
0.9 
0.9 
1.0 
1.0 
1.0 
1.0 
Plate 2 

1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
A 
0.1 
0.1 
0.1 
0.1 
0.3 
0.3 
0.3 
0.3 
0.5 
0.5 
0.5 
0.5 
B 
0.7 
0.7 
0.7 
0.7 
0.9 
0.9 
0.9 
0.9 
1.0 
1.0 
1.0 
1.0 
Plate 3 

1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
A 
0.1 
0.1 
0.1 
0.1 
0.3 
0.3 
0.3 
0.3 
0.5 
0.5 
0.5 
0.5 
B 
0.7 
0.7 
0.7 
0.7 
0.9 
0.9 
0.9 
0.9 
1.0 
1.0 
1.0 
1.0 
Image acquisition and analysis for each cell counting method
Timing: Scanning and analysis are 6 minutes per plate for the Cellaca MX and 5–10 minutes per plate for the Celigo.
 After preparation, load the 1st Cellaca MX plate into the Cellaca MX.
 Select the “MX04.0_AOPI_LiveDead” default assay type for cell counting in the Cellaca MX software for cell samples stained with AO staining solution.
For cell samples stained with Nuclear Green, increase the FL1 exposure time by 50–100%. Check the fluorescent intensity of the nuclear green stained cells in the preview images before image acquisition.
 Use the default analysis parameters for enumerating cells in the captured Cellaca MX bright field and FL1 fluorescent images. Export the concentration data.
 Transfer the 1st Cellaca MX plate to the Celigo. Select the plate profile for Cellaca MX plates. Utilize the default experiment setting for image capture and analysis. Export total cell counts from the fluorescence images captured.
 Repeat 51–54 for the 2nd and 3rd Cellaca MX plates.
Cell counting method performance evaluation
Timing: ~30 minutes to determine and analyze the parameters for performance evaluation for one cell line and one stain.
 Calculate the cell concentration utilizing the overall cell counts from the data exported from the Celigo and multiply by a factor of 1383.979, which is the conversion ratio based on the counted volume and dilution factor from staining with AO and Nuclear Green.
 Calculate the average concentration M_{Ai} acquired with method A (Cellaca MX) from a total number of n_{Ai }replicate measurements for sample i using Equation 1.

Equation 1 
Where M_{Ai} is the concentration acquired with method A for sample i during replicate measurement r.
 Calculate the mean concentration M_{Ak} acquired with method A (Cellaca MX) for dilution fraction k (DF_{k} ) using Equation 2.

Equation 2 
 Calculate the variance of concentration var_{Ai} acquired with method A (Cellaca MX) from a total number of n_{Ai} replicate measurements for sample i using Equation 3.

Equation 3 
 Calculate the pooled variance of concentration var_{AK} acquired with method A (Cellaca MX) for DF_{k} using Equation 4.

Equation 4 
 Calculate the pooled standard deviation of concentration σ_{Ak} acquired with method A (Cellaca MX) for DF_{k} using Equation 5.

Equation 5 
 Calculate the pooled CV, CV_{Ak} acquired with method A (Cellaca MX) for DF_{k} using equation Equation 6.

Equation 6 
 Repeat 57–62 to calculate the mean concentration M_{Bk}, the pooled standard deviation of concentration σ_{Bk} and the pooled CV acquired with method B (Celigo) for DF_{k}.
 Use mean concentrations (M_{Ai}, M_{Bi}) from all samples at 6 different DFs to generate a concentration series for both Cellaca MX and Celigo. Perform a proportional fit with the concentration series for each method using the iteratively reweighted least squares (IRLS) model. Set the weights of the least squares proportional to the reciprocal of the variances, which can be estimated by mean concentrations under the assumption of a quasiPoisson distribution that the variances of cell concentrations are proportional to their respective mean concentrations (var_{Ai} = φM_{Ai}), where φ is a scalar estimated from the experimental data that cancels out when used in the weighting of every least squares term.^{17} Rerun the model fitting by updating weights using predicted values of the mean concentrations until the proportional fit is optimized. Generate a list of predicted values of mean concentrations from the IRLS model.
 Determine the coefficient of determination (R^{2} value) from the IRLS model for method A (Cellaca MX) using Equation 7.^{26,27} Use the same method to determine the R^{2} value for method B (Celigo).

Equation 7 
 Perform a fit with the concentration series for each method (Cellaca MX, Celigo) using a higherorder polynomial model as a flexible model. Set the order of the polynomial to be the number of DFs minus 1. Generate a list of predicted values of mean concentrations from the polynomial model.
 Determine the proportionality index (PI) based on the smoothed sum of absolute scaled residuals (PI_{A}^{SAbsSR}, PI_{A}^{SAbsSR}) for both Cellaca MX and Celigo using Equation 8 following previous publication.^{17,18}

Equation 8 
 Apply the BlandAltman method to compare the performance between two cell counting methods.
 An internally developed software application was used, which derived from the ISO Cell Counting Standard Part 2 and BlandAltman comparative method to automatically determine the coefficient of determination, precision, proportionality index parameters, as well as the Bland Altman analysis parameters (bias, LoA, the CI of the bias).
BlandAltman comparative method: Data calculation
Timing: ~30 minutes to analyze and plot the BlandAltman comparison data for one cell line and one stain.
 Calculate the percent difference Y_{i} between the measurement M_{Ai} acquired with method A and the measurement M_{Bi} acquired with method B for each sample i using the Equation 9, only if the samples are paired between method A and B.

Equation 9 
where X_{i} is the sample mean given by
 If measurements M_{Air} and M_{Bir} from replicate r are paired, calculate the percent difference Y_{ir} between the measurement M_{Air} acquired with method A and the measurement M_{Bir} acquired with method B for each replicate r of sample i using the Equation 10.

Equation 10 
where X_{ir} is the sample mean given by
 Calculate the bias from method A to method B (Bias_{AB}) by averaging the Y_{i} values using Equation 11 or by averaging the Y_{ir} values using Equation 12.

Equation 11 
where N is the number of samples (for paired samples, unpaired replicates, i.e., for each sample, different replicates are measured with each method).

Equation 12 
where N is the total number of replicate measurements (for paired samples with paired replicates, i.e., for each sample, the same replicates are measured using both methods).
 Calculate the LoA by multiplying 1.96 to the mean for percent differences determined in step 70 using Equation 13 or 14. LoA are defined as the onesided 95% confidence interval for a single sample.

Equation 13 
where N is the number of samples (paired samples, unpaired replicates).

Equation 14 
where N is the total number of replicate measurements (paired samples, unpaired replicates).
 Calculate the CI of the bias using Equation 15.

Equation 15 
where N is the number of samples (for paired samples without paired replicates) or the total number of replicate measurements (if both samples and replicates are paired).
BlandAltman comparative method: graphical representation
 Plot a single point on the BlandAltman diagram for each sample, with X_{i} (sample mean) on the horizontal axis and Y_{i} (percent difference) on the vertical axis.
 Plot a horizontal line that crosses the vertical axis at the value of Bias_{AB} calculated in step 71.
 Plot two additional horizontal lines that cross the vertical axis at the values of Bias_{AB} + LoA and Bias_{AB} – LoA, where the LoA is calculated as highlighted in step 72. These lines determine a range of values for the percent difference expected between the two methods for a single sample.
 Plot two additional horizontal lines at the values Bias_{AB} + CI_{Bias} and Bias_{AB} – CI_{Bias}. This range offers a sense of the uncertainty on the bias value itself.
 Evaluate the plot and note any concentration dependence in either the bias or variation.
Troubleshooting
Follow the troubleshooting Table 3 to optimize the experiments and output.
Table 3. Troubleshooting table. Source: Nexcelom Bioscience LLC
Step 
Problem 
Possible reason 
Solution 
62, 63 
CV is too large at one or a few DFs 
Sampling or pipetting error 
Properly mix and pipette samples following ISO cell counting standard 


Counting errors
due to clumps 
Adjust the counting parameters
Remove the outliers if severe counting errors are observed 
67 
Poor
Proportionality 
Propagation of
pipetting error 
Directly dilute to generate independent dilution samples instead of serial dilution to eliminate the propagation of pipetting error 
68, 70–78 
A large bias
between two
cell counting
methods 
Sample variation (i.e. different stocks of samples) 
Use the same stock of cell samples for both cell counting methods
If possible, use the same cell sample in the same piece of consumable to conduct cell counting comparison
Test the stability of the cell sample for concentration and viability for the duration of the assay. If a trend is observed, then the results may be invalid 


Sample condition change (e.g. photobleaching,
sample dryout) 
Practice cell counting performance evaluation and comparison experiments
Use presets in the software
Finish image acquisitions in a short time duration 


Cell counting analysis variation (e.g. declumping) 
Adjust the imaging and counting parameters in the software to ensure that cells are counted properly 


Instrument comparison 
Ensure the exact instruments are compared in repeated experiments 


Instrument
calibration 
Ensure both instruments are well calibrated and data acquisition and analysis parameters are optimized before use. 
Timing
 Step 1–26, maintenance of CHOS and Jurkat cells: 20–30 minutes for passaging the cells and measuring their concentration and viability per cell line.
 Steps 27–35, stock cell sample preparation from cell culture: 15 minutes for collecting the cells from cell culture flasks, 5 minutes for cell counting and viability analysis, and 10 minutes for modifying cell sample concentration if necessary.
 Steps 36–50, sample preparation and cell counting preparation for cell counting methods performance evaluation and comparison: 15–30 minutes with a single, manual pipette for sample preparation, 15–20 minutes for incubation of cell samples mixed with Nuclear Green, and ~10 minutes per Cellaca MX plate for cell counting preparation.
 Step 51–55, image acquisition and analysis: 6 minutes per plate for the Cellaca MX and 5–10 minutes per plate for the Celigo.
 Step 56–78, performance evaluation and BlandAltman comparison analysis: 1 hour per cell line per staining solution.
Anticipated results & discussion
Two cell lines (CHOS, Jurkat), two dyes (AO, Nuclear Green), and 2 cell counting methods were assessed to show the application of cell counting method performance evaluation and BlandAltman comparative analysis.
Figure 2. Cell counting results for (a) mean concentrations and (b) pooled CV of 6point concentration series of Jurkat Cells Stained with AO. The mean concentrations and CV values measured by Cellaca MX and Celigo were highly comparable at each dilution fraction. It is clear that as concentration decreased, the pooled CV increased, likely due to the Poisson Noise (Random Error) at lower concentrations. Image Credit: Nexcelom Bioscience LLC
Figure 2 exhibits the average and pooled CV, respectively, of the 6point concentration series of Jurkat cells stained with AO utilizing both cell counting systems. Table 4 exhibits the numerical results for the average and pooled CV.
Figure 3. Linear regression fitting of the 6point concentration series as a function of dilution fractions (DFs), which shows the distribution of concentration measurements for Cellaca MX and Celigo at each dilution fraction. Image Credit: Nexcelom Bioscience LLC
The concentration range determined in the experiment was ~5 × 10^{5} to ~6 × 10^{6} cells/mL. Both Cellaca MX and Celigo have consolidated CVs ranging from 1.8– 7.6% for all replicates per concentration.
Table 4. Calculated mean concentration and pooled CV for each dilution fraction using ISO Cell Counting Standard Part 2. Source: Nexcelom Bioscience LLC
Cell counting method 
DF 
n 
Mean (cells/mL) 
Pooled CV (%) 
Cellaca MX 
0.1 
12 
5.37E+05 
7.5 
0.3 
12 
1.78E+06 
4.2 
0.5 
12 
2.88E+06 
3.3 
0.7 
12 
3.93E+06 
2.2 
0.9 
12 
5.05E+06 
3.8 
1.0 
12 
5.74E+06 
2.3 
Celigo 
0.1 
12 
5.21E+05 
7.6 
0.3 
12 
1.71E+06 
3.9 
0.5 
12 
2.80E+06 
3.6 
0.7 
12 
3.91E+06 
2.7 
0.9 
12 
5.06E+06 
3.6 
1.0 
12 
5.80E+06 
1.8 
According to the measurements in Table 4, the coefficient of determination and proportionality index can be determined via regression analysis. Figure 3 demonstrates the proportional fits of the 6point concentration series as a functionofdilution fraction for both Cellaca MX and Celigo.
Figure 4. Comparison of the R^{2} and PI values between Cellaca MX and Celigo. The R^{2} and PI values are both highly comparable. Image Credit: Nexcelom Bioscience LLC
Table 5. Calculated R^{2} and PI values for performance evaluation. Source: Nexcelom Bioscience LLC

R^{2} 
PI 
Cellaca MX 
0.997 
0.44 
Celigo 
0.997 
0.42 
Significance 
No 
No 
Results for each parameter are displayed in Figure 4 and Table 5. Both Cellaca MX and Celigo show similar values of coefficient of determination (R^{2} values) and proportionality indices (PIs) from the proportional fits in this cell counting method evaluation.
Figure 5. BlandAltman plot between Cellaca MX and Celigo cell counting methods. The calculated percent differences show an increasing trend as the concentration increases. Image Credit: Nexcelom Bioscience LLC
No major differences are seen between Cellaca MX and Celigo for both R^{2} and PI values.
Next, the BlandAltman method is employed to compare the performance between Cellaca MX and Celigo cell counting methods. Figure 5 demonstrates a representative BlandAltman plot between Cellaca MX and Celigo.
In this plot, a positive percentage signals a greater concentration for Celigo measurements. Each point in the plot stands for a pair of measurements identified by both cell counting methods.
Table 6. BlandAltman comparative analysis results between Cellaca MX and Celigo. Source: Nexcelom Bioscience LLC
Bias 
Limit of agreement 
CI of bias 
1.5% 
6.9% to 3.9% 
2.1% to 0.8% 
Results of concentration bias, 95% confidence interval, and a conventional deviation of the bias between Cellaca MX and Celigo are exhibited in Table 6. A bias of ~ 1.5% (n = 72 replicate measurements) signifies that the concentration measured by Celigo is ~1.5% less than Cellaca MX in this cell counting method comparison.
Due to the fact that the value of 0 remains beyond the confidence interval of the bias, it is concluded that the concentration difference seen between these two cell counting methods is of some significance (p < 0.05), despite being comparatively small. The bias demonstrates a slight dependence on cell concentration in this case.
An additional coupled measurement conducted using two methods under equivalent conditions would be expected to fall within the range identified by the LoA’s in approximately 95% of cases.
Table 7. Summary table of cell counting performance evaluation and comparison results. Source: Nexcelom Bioscience LLC
Cell line 
Staining solution 
Cellaca MX 
Celigo 
Bias 
LoA 
95% CI of bias 
Significance of bias 
R^{2} 
Pooled CV range (%) 
PI 
R^{2} 
Pooled CV range (%) 
PI 
Jurkat 
Nuclear green 
0.998 
3.8% to 6.1% 
0.30 
0.997 
3.7% to 7.4% 
0.37 
3.4% 
6.0% to 12.8% 
2.3% to 4.5% 
Y 
Jurkat 
AO 
0.997 
2.2% to 7.5% 
0.44 
0.997 
1.8% to 7.6% 
0.42 
1.5% 
6.9% to 3.9% 
2.1% to 0.8% 
N 
CHO 
Nuclear green 
0.998 
3.4% to 5.8% 
0.40 
0.999 
2.7% to 6.8% 
0.35 
5.6% 
3.4% to 14.6% 
4.5% to 6.7% 
Y 
CHO 
AO 
0.998 
2.7% to 7.0% 
0.44 
0.996 
2.7% to 6.4% 
0.35 
5.1% 
2.4% to 12.5% 
4.2% to 5.9% 
Y 
Table 7 shows a summary of the cell counting performance evaluation and comparison results between Cellaca MX and Celigo using various combinations of targeted cell lines and staining solutions.
Concentration biases are 2–6% for the paired cell counting methods in these four cell counting methods. Since each cellstain combination was independently treated, the results were not combined, and the False Discovery Rate (FDR) or FamilyWise Error Rate (FWER) were not calculated.
Overall, multiple practical experiments following ISO Cell Counting Standard Part 2 with Celigo and Cellaca MX we performed, and the results displayed here are in the expected range of the cell counting measurement quality.
The ISO Cell Counting Standard Part 2 facilitates cell and gene therapy research, enabling researchers to carry out experiments for the evaluation and comparison of the quality of cell counting measurement processes.
In this practical application of the standard, it was determined that the evaluation of the Cellaca MX and Celigo using Jurkat and CHO cells could be customized for the bioprocessing and cell therapy communities.
The users of the ISO Cell Counting Standard Part 2 may apply the protocol to assess one or more cell counting methods. For the evaluation of a single method, it is possible to conduct the experiments to determine a baseline for each of the quality parameters, where this baseline can be monitored further by various operators, instruments, processes, etc.
For a complete comparison of cell counting methods, the quality parameters can be contrasted against one another via bootstrap analysis or replicate studies, and the difference or bias between the methods can be identified using the BlandAltman comparative analysis.
It is also crucial to note that the quality parameters acquired using the ISO Cell Counting Standard Part 2 are unique to the measurement process evaluated (e.g., operators, instruments, cell sample properties, etc.). Thus the strength of the measurement process should also be assessed to extend the findings of the study to comparable measurement processes.
Frequently asked questions
1. Can the number of replicate samples and measurements be reduced?
 Yes, to an extent; the minimum recommended experimental design is made up of at least 4 target dilution fractions, 3 replicate samples, and 3 replicate measurements.
 Quality indicators from experimental designs that fall short of these recommendations should be interpreted with caution and may necessitate additional studies to evaluate proportionality and precision.
 For instance, if an experimental design has only 2 replicate measurements, a direct CV assessment from this experimental design and statistical analysis may not be relevant. However, evaluation of proportionality can still be carried out.
 In this case, a second experiment with increased replicate measurements of fewer samples and/or fewer dilution fractions may be carried out to directly address the precision of the method.
2. What is the optimal cell concentration range that should be used?
 It should be fitforpurpose for the standard range of cell concentrations intended for the evaluation of cell type.
3. Should the pipettors be checked?
 Always utilize professionally calibrated pipettors
 Performing a check on pipettors will boost confidence in the results
 ISO 203912 also recommends an approach for producing measured dilution fractions, where the mass of solution pipetted while generating the fractions is used to calculate a more accurate measured dilution fraction value for use in the analysis of R^{2} and PI. In this case, minor errors in pipetting can be accounted for in the proportional model fit.
4. What do the results indicate?
 The quality indicators offer a means to quantify and compare the quality of cell counting methods predicated on principles that are essential to counting: precision and proportionality
 The ISO Counting Standard Part 2 analysis offers no assumptions about the true cell count and can make conclusions in relation to method quality in lieu of reference material or reference method.
 The BlandAltman comparative analysis will signal the percent difference between 2 methods.
 These approaches do not represent or compare the accuracy of the cell counting methods.
 Cell counting method selection should be conducted based on the quality of the method and on what is fitforpurpose for your measurement needs.
References
 Boyiadzis MM, Dhodapkar MV, Brentjens RJ et al. Chimeric antigen receptor (CAR) T therapies for the treatment of hematologic malignancies: clinical perspective and significance. J. ImmunoTher. Cancer 2018; 6(137): 1–12.
 Gonçalves GAR, Paiva RdMA. Gene therapy: advances, challenges, and perspectives. Einstein 2017; 15(3): 369–75.
 Huang R, Li X, He Y et al. Recent advances in CART cell engineering. J. Hematol. Oncol. 2020; 13(86): 1–19.
 Li Y, Huo Y, Yu L, Wang J. Quality Control and Nonclinical Research on CART Cell Products: General Principles and Key Issues. Engineering 2019. 5: 122–31.
 Seimetz D, Heller K, Richter J. Approval of First CARTs: Have we Solved all Hurdles for ATMPs? Cell Med. 2019; 11: 1–16.
 Nagata S, Hanayama R, Kawane K. Autoimmunity, and the Clearance of Dead Cells. Cell 2010; 140(5): 619–30.
 Rock KL, Kono H. The inflammatory response to cell death. Ann. Rev. Pathol. 2008; 3: 99–126.
 Zhang Y, Chen X, Gueydan C, Han J. Plasma membrane changes during programmed cell deaths. Cell Res. 2018; 28: 9–21.
 21st Century Cures. United States of America; 2016.
 Arcidiacono JA, Bauer SR, Kaplan DS, Allocca CM, Sarkar S, LinGibson S. FDA and NIST collaboration on standards development activities supporting innovation and translation of regenerative medicine products. Cytotherapy 2018; 20: 779–84.
 LinGibson S, Sarkar S, Elliott JT. Summary of the National Institute of Standards and Technology and US Food And Drug Administration cell counting workshop: Sharing practices in cell counting measurements. Cytotherapy 2018; 20(6): 785–95.
 Biotechnology – Cell counting – Part 1: General guidance on cell counting methods. In: Standardization IOf, editor; 2018.
 Biotechnology – Cell counting – Part 2: Experimental design and statistical analysis to quantify counting method performance. In: Standardization IOf, editor; 2019.
 Validation of Analytical Procedures: Text and Methodology Q2(R1). In: Guideline IHT, editor; 2005.
 Guidance for Industry: Process Validation: General Principles and Practices. In: Services USDoHaH, Administration FaD, (CDER) CfDEaR, (CBER) CfBEaR, (CVM) CfVM, editors; 2011.
 Guidance for Industry: Analytical Procedures and Methods Validation for Drugs and Biologics. In: Services USDoHaH, Administration FaD, (CDER) CfDEaR, (CBER) CfBEaR, editors; 2015.
 Sarkar S, Lund SP, Vyzasatya R, Vanguri P, Elliott JT, Plant AL, LinGibson S. Evaluating the quality of a cell counting measurement process via a dilution series experimental design. Cytotherapy 2017; 19(12): 1509–21.
 Sarkar S, Pierce L, LinGibson S, Lund SP. Standards Landscape in Cell Counting: Implications for Cell & Gene Therapy. Cell Gene Ther. Ins. 2019; 5(1): 117–31.
 Cohen J. The earth is round (p<.05). Am. Psychol. 1994; 49(12): 997–1003.
 Altman DG, Bland JM. Measurement in Medicine: The Analysis of Method Comparison Studies. J. Royal Stat. Soc. Series D (The Statistician) 1983; 32(3): 307–17.
 Bland JM, Altman DG. Statistical Methods For Assessing Agreement Between Two Methods of Clinical Measurement. Lancet 1986; 327(8476): 307–10.
 Tholudur A, Giron L, Alam K et al. Comparing Automated and Manual Cell Counts for Cell Culture Applications. BioProcess Int. 2006.
 Carkeet A. A Review of the Use of Confidence Intervals for BlandAltman Limits of Agreement in Optometry and Vision Science. Optometry Vision Sci. 2020; 97(1): 3–8.
 Carkeet A. Exact Parametric Confidence Intervals for BlandAltman Limits of Agreement. Optometry Vision Sci. 2015; 92(3): e71–e80.
 AbuArafeh A, Jordan H, Drummond G. Reporting of method comparison studies: a review of advice, an assessment of current practice, and specific suggestions for future reports. Br. J. Anaesthesia 2016; 117(5): 569–75.
 Navidi W. Statistics for Engineers and Scientists. New York: McGrawHill; 2013.
 Sokal RR, Rohlf FJ. Biometry. New York: W. H. Freeman and Company; 1969.
About Nexcelom Bioscience
Nexcelom Bioscience is a developer and marketer of image cytometry products for cell analysis in life science and biomedical research. Products range from cell viability counters (Cellometer) to high throughput microwell image cytometry workstations (Celigo), used in thousands of research laboratories in academic institutes, and pharmaceutical and biotech companies. The company contributes to the life science industry through innovation and expertise in the science of cell counting.
The product family includes instruments, consumables, and reagents. Nexcelom customers engage in a wide variety of research, such as cancer research, immunology, stem cell research, and neuroscience. Nexcelom offers different Cellometer models to count and analyze cell lines and primary cells, through brightfield and fluorescence imaging modes. In addition, Celigo is a powerful high image quality, highthroughput image cytometry system for adherent and suspension cells in microwell plates.
Nexcelom Bioscience is a fastgrowing company in a huge market. With its headquarters and manufacturing facilities in the Boston area, the company currently has over 80 global employees, who are fastpaced, customercentric, helpful to colleagues and customers, and passionate about their impact in life science.
Sponsored Content Policy: NewsMedical.net publishes articles and related content that may be derived from sources where we have existing commercial relationships, provided such content adds value to the core editorial ethos of NewsMedical.Net which is to educate and inform site visitors interested in medical research, science, medical devices, and treatments.