The cytogenetic dicentric chromosome (DC) assay quantifies exposure to ionizing radiation. The Automated Dicentric Chromosome Identifier and Dose Estimator software accurately and rapidly estimates biological dose from DCs in metaphase cells. It distinguishes monocentric chromosomes and other objects from DCs, and estimates biological radiation dose from the frequency of DCs.
Biological radiation dose can be estimated from dicentric chromosome frequencies in metaphase cells. Performing these cytogenetic dicentric chromosome assays is traditionally a manual, labor-intensive process not well suited to handle the volume of samples which may require examination in the wake of a mass casualty event. Automated Dicentric Chromosome Identifier and Dose Estimator (ADCI) software automates this process by examining sets of metaphase images using machine learning-based image processing techniques. The software selects appropriate images for analysis by removing unsuitable images, classifies each object as either a centromere-containing chromosome or non-chromosome, further distinguishes chromosomes as monocentric chromosomes (MCs) or dicentric chromosomes (DCs), determines DC frequency within a sample, and estimates biological radiation dose by comparing sample DC frequency with calibration curves computed using calibration samples. This protocol describes the usage of ADCI software. Typically, both calibration (known dose) and test (unknown dose) sets of metaphase images are imported to perform accurate dose estimation. Optimal images for analysis can be found automatically using preset image filters or can also be filtered through manual inspection. The software processes images within each sample and DC frequencies are computed at different levels of stringency for calling DCs, using a machine learning approach. Linear-quadratic calibration curves are generated based on DC frequencies in calibration samples exposed to known physical doses. Doses of test samples exposed to uncertain radiation levels are estimated from their DC frequencies using these calibration curves. Reports can be generated upon request and provide summary of results of one or more samples, of one or more calibration curves, or of dose estimation.
Radiation biodosimetry uses biological markers, mostly chromosomal aberrations such as dicentric chromosomes (DCs) and chromosome translocations to measure radiation doses that individuals are exposed to. A biologically absorbed dose may be different from the physical dose measured by instruments due to variability between individuals. Similarly, radiation of a certain physical dose can produce different biological exposures due to underlying physiological or environmental conditions. Knowledge of the biological dose is of particular importance for both diagnosis and treatment.
The DC assay is the gold standard of the World Health Organization (WHO) and International Atomic Energy Agency (IAEA) for assessing biological radiation exposure in people. It was the first assay recommended by the IAEA and WHO for radiation dose assessment. DC frequency is relatively stable for approximately 4 weeks after radiation exposure1 and their quantitative correlation with emitted radiation dose is accurate, which make DCs the ideal biomarker. The relationship between radiation dose (referenced in Gray [Gy] units), and DC frequency (referenced as number of DCs per cell) can be expressed as a linear-quadratic function.
The cytogenetic DC assay has been the industry standard for about 55 years2. It has been performed manually, requiring 1 – 2 days to analyze microscope data from a single blood sample. Several hundred to several thousand images are needed to accurately estimate radiation exposure depending on the dose3. At doses exceeding 1 Gy, IAEA recommends a minimum of 100 DCs be detected. Examination of 250 – 500 metaphase images is common practice in biodosimetry cytogenetic laboratories. For samples with exposures <1 Gy, 3,000 – 5,000 images are suggested due to the lower probabilities of DC formation. In either case, it is a labor-intense task.
Cytogenetic biodosimetry laboratories create their own in vitro radiation biodosimetry calibration curves before assessing biological doses in test samples. Blood samples from normal, control individuals are exposed to radiation and lymphocytes are then cultured and prepared for metaphase chromosome analysis. Using these samples, biological doses received are calibrated to the known physical doses emitted by a standard radiation source. After metaphase cell images are recorded, experts examine images, count DCs and calculate DC frequencies for each sample. A calibration curve is built by fitting a linear-quadratic curve to the DC frequencies at all doses. Then, exposures in test sample from individuals can be inferred by matching the DC frequencies to the calibrated doses on the curve or by specifying them in the corresponding linear quadratic formula.
We have automated both the detection of DCs and dose determination to expedite this procedure using software. Automated Dicentric Chromosome Identifier and Dose Estimator (ADCI) uses machine learning-based image processing techniques to detect and discriminate dicentric chromosomes (DCs) from monocentric chromosomes (MCs) and other objects and automates radiation dose estimation. The software aims to significantly reduce or eliminate the necessity for manual verification of DC counts and to accelerate dose estimation through automation. It has been developed with the involvement of reference biodosimetry laboratories at Health Canada (HC) and Canadian Nuclear Laboratories (CNL). Their feedback will ensure that performance will continue to meet IAEA criteria for this assay.
The software performs the following functions: 1) filtering DCs and selecting optimal metaphase cell images for analysis, 2) chromosome recognition, DC detection, and DC frequency determination, and 3) estimating radiation dose from dose-calibrated, cytogenetic radiation data. This software processes groups of metaphase images from the same individual (termed a sample), counts the number of DCs in each using image processing techniques, and returns the estimated radiation dose received by each sample in units of Grays (Gy).
The software has been designed to handle a range of chromosome structures, counts, and densities. However, the algorithm performs optimally in metaphase images containing a near complete complement of well-separated, linear chromosomes4. Images containing highly overlapped sets of chromosomes, multiple cells, incomplete metaphase cells, sister chromatid separation, nuclei, non-chromosomal objects, and other defects can reduce the accuracy of the algorithm. Dedicated image selection models and other object segmentation thresholds can filter out the majority of sub-optimal images and false positive DCs.
Dicentric chromosome detection is performed when an image is processed. The algorithm attempts to determine which objects in an image are chromosomes and then locates the two regions most likely to be centromeres on each chromosome. Then, a series of different Support Vector Machine (SVM) learning models distinguish chromosomes as either DCs or normal, monocentric chromosomes. The SVM models differ in sensitivity and specificity of DC detection (see Step 3.1.4 below), which can affect the DC frequencies that are determined in a sample.
ADCI processes sets of Giemsa- (or DAPI-) stained metaphase digital images (in TIFF or JPG format) for one or more samples. The software analyzes DCs in both calibration samples and test samples. The physical doses (in Gy) of calibration samples are known and are used in the generation of a calibration curve. The physical and biological doses of individuals with unknown exposures are inferred by the software from the machine-generated calibration curve. Although laboratories use comparable techniques, the calibration curves from different laboratories often vary3. Both calibration curve and test samples from the same laboratory should be processed for accurate dose estimation in test samples.
This software offers speed, accuracy and scalability which addresses the productivity required to handle an event in which many individuals must simultaneously be tested. It was developed from 2008 – 20174,5,6,7,8,9,10,11,12,13. Using recent computer hardware, this desktop PC software can process and estimate radiation dose in a patient sample of 500 metaphase genome equivalents in 10 – 20 min 4. The code is based on a set of proprietary image segmentation and machine learning algorithms for chromosome analysis. Expert analysis of each chromosome exposed to 3 Gy radiation gave comparable accuracies to ADCI. In a set of 6 samples of unknown exposures (previously used in an international proficiency exercise), the software estimated doses within 0.5 Gy of the values obtained by manual review of the same data by HC and CNL, fulfilling the IAEA's requirements for triage biodosimetry. Furthermore, inter-laboratory standardization and ultimately reproducibility of dose estimates benefit from having a common, automated DC scoring algorithm. Nevertheless, the software permits customization of image filtering and selection criteria, enabling differences in chromosome preparation methods and radiation calibration sources to be taken into account.
This software is a graphical user interface (GUI)-based system which analyzes sets of chromosome images containing Giemsa (or DAPI)-stained metaphase cells for abnormalities that result from exposure to ionizing radiation. The image sets are digitally photographed with a light (or epifluorescent) microscope system and each set corresponds to a different sample. The software utilizes image processing techniques to detect and discriminate DCs from MCs and other objects. Empirically-derived segmentation filters then automatically eliminate false positive DCs without affecting true DCs. Finally, the software automatically filters out undesirable images based on various image properties found poor quality metaphase images with precomputed (or user-specified) image selection models. These images include those containing excessive or insufficient numbers of "noisy" objects, multiple overlapping chromosomes, images lacking metaphase chromosomes, excessive numbers of sister chromatids4. The automatically curated image data are used to generate dose calibration curve from samples of known radiation dose and are used to estimate exposures of test samples exposed to unknown doses.
Output of the software can be viewed and saved as: 1) text-based output viewed in the console, 2) plots which can be saved as images, and 3) reports in HTML format. Many aspects of the software are customizable to suit the specific needs of different laboratories. Individual laboratories usually provide both calibration and test samples prepared and collected based on the cytogenetic protocol validated in that laboratory. This maintains uniformity of sample preparation and allows calibration curves generated from calibration samples to be meaningfully applied to test samples derived using the same protocol. Calibration curves may also be created from either curve coefficients or DC frequencies at defined doses. The most accurate dose estimates are obtained by filtering out lower quality images and false positive DCs (FPs). Selection of optimal image subsets within each sample is accomplished using 'Image selection models' that eliminate subpar images which tend to introduce FPs. A series of pre-validated models are included with the software, however additional models with customized thresholds and filters can be created and saved, by the user.
Once the software successfully loads, the main graphical user interface (GUI) is presented (see Figure 1). From this interface, samples, each consisting of a folder of metaphase cell image files, may be selected and processed to identify DCs, calibration curves may be created and compared, and radiation exposure dose of samples may be determined.
Figure 1: The Major Sectors of the Graphical User Interface Include: a list of samples (1), a list of calibration curves (2), the process queue (3), which monitors the status of DC detection in each set of images of each sample, a plot display (4), which summarizes statistical or other quantitative properties of a set of images in samples or calibration curves, and a console (5) which contains descriptive text as outputs of each operation performed by the program. Please click here to view a larger version of this figure.
Capabilities and limitations of the software
The protocol described in this paper introduces the typical step-wise procedure used in ADCI to import and process cytogenetic metaphase images, create radiation calibration curves, and estimate biological dose in individuals or samples exposed to unknown radiation levels. However, it is not necessary to carry out these instructions sequentially. For example, many test samples of unknown dose can be processed and analyzed using the same precomputed calibration curve. Furthermore, once processing is complete, the image selection and DC filtering models can be iterated by the user. Application of an appropriate image selection model depends on the characteristics and quality of the metaphase image data, which in turn relies both on the laboratory protocol used to prepare cells and the stringency criteria used to select cells with automated metaphase capture systems. Chromosome morphologies will differ among biodosimetry and cytogenetic laboratories, and thus, the image selection models should be evaluated by the user to determine whether the predefined image selection models supplied with the software will be adequate to produce accurate dose estimates, or whether custom models with user-defined thresholds need to be created. Based on our experience, the effectiveness of image selection models is influenced by the source and quality of the cell images. Users can design their own image selection criteria using different combinations of filters to eliminate false positive DCs and image selection models, and the corresponding threshold values to select desired images. There is flexibility in input of calibration curves and dose estimation, as coefficients of the linear-quadratic curve and DC frequencies can be modified or manually inputted.
Although the software is fully automated, images can be manually reviewed and selected. This capability is available to include or remove individually processed images through the Microscope Viewer function in the main GUI. Nevertheless, due to automation, the software is significantly more efficient compared with the manual scoring of metaphase images and counting DCs. A sample consisting of 1000 images can be processed in 20 (tiff) to 40 (jpg) min on a multi-core performance workstation. This software will be particularly useful in time-critical or labor-intensive situations, such as events in which multiple individuals have been exposed or were suspected to have been exposed to radiation, or where time-sensitive diagnoses and treatment decisions are critical.
Precise and accurate high throughput detection of DCs as well as dose estimation are necessary for unattended radiation assessment. Other available alternatives to the software do not fulfill both of these requirements. A user-assisted, image-based cytogenetic analysis (DCScore, Metasystems17) system requires manual verification of candidate DCs, due to a high error rate attributable to uncorrected overlaps between chromosomes, and the system does not determine radiation dose. DCScore would not be as effective as ADCI in a radiation event involving a large number of potentially exposed individuals. Large aperture microscope systems can collect images of multiple metaphase cells18, however, they do not analyze them. "CABAS"19 and "Dose Estimate"20 software can generate calibration curves and estimate dose, but do not score DCs. Other biodosimetry assays that are not based on DC analysis include H2AX fluorescence, fluorescence in situ hybridization with DNA probes targeted to specific chromosomes, gene expression, micronucleus assay, and urine and respiratory biomarkers. These methods are less specific and less sensitive for ionizing radiation, can be more costly, in some instances, are more time consuming, and have generally not been standardized across multiple reference laboratories. Most of these techniques do not detect stable radiation responses, so they cannot be used for long term assessment (>7 days post-exposure) of radiation dose. By contrast, this can evaluate individuals up to 90 days post-exposure, and can process data from any cytogenetics laboratory microscope imaging system. However, if a sample is drawn >4 weeks post-exposure, sensitivity is reduced due to the decay of dicentric aberrations1,2,3 and the software does not currently correct DC frequencies for delays in sampling exposed individuals.
This software has some limitations. Existing image selection models select mostly acceptable metaphase images, but in some instances, fail to eliminate unsatisfactory images, which can reduce the accuracy of DC detection. It is still an open question how to design a satisfactory image selection model that eliminates all unsuitable metaphase cells. The software provides accurate estimates for samples exposed to higher radiation doses (≥ 2 Gy). Despite considerable progress in reducing the number of false positive DCs16, these objects have not been eliminated. Lower quality metaphase cells at low radiation dose (especially < 1 Gy) are more prone to false positive DC detection. Therefore, low dose samples were not included when generating the calibration curve used for dose estimation of HC test samples. However, if a curve containing low dose samples is desired, a lower SVM Sigma value reduces false positive counts in low dose samples but may result in lower DC yields in high dose samples. Figure 8 compares the HC curve used for dose estimation (Sigma = 1.5) with a calibration curve fit with additional low dose samples at lower SVM sigma value (1.0). In samples with insufficient numbers of metaphase cells and/or poor quality metaphase images, it may not be possible to precisely estimate biological exposures at low dose, potentially resulting in deviations from physical dose exceeding 0.5 Gy.
The software may not accurately assess radiation types if their dose-response curves best fit a linear or near-linear model. Thus far, it has been tested only with samples exposed to X- and Gamma rays. If another radiation source is examined, users must ensure both calibration and test samples are exposed to the same type of radiation. The software uses either maximum likelihood or least squares fitting to construct a dose-response curve using a linear-quadratic model. There is currently no option to impose a strict linear curve fit, appropriate for high energy particle exposures, however such functionality will be available in the future.
Future development
Our ongoing efforts are focused on improving image selection models and accurate dose measurement, in particular of samples exposed to low radiation doses. Subsequent software versions will provide standard error measurements on dose estimates and confidence intervals on calibration curves. In addition, a high-performance-computing version of the software for the Blue Gene (BG/Q, IBM) supercomputer is under development for timely evaluation of individuals exposed in a mass-casualty radiation event. Some components of the software have already been tested and deployed on this platform11.
The authors have nothing to disclose.
We are grateful to Dr. Ruth Wilkins, Radiobiology and Protection Division at Health Canada, and Farrah Flegal, Canadian Nuclear Laboratories and their laboratory personnel for access to metaphase image data from their cytogenetic biodosimetry laboratories. This paper was supported by a contract from Build in Canada Innovation Program to CytoGnomix (Serial No. EN579-172270/001/SC). The initial version of ADCI and development of algorithms were supported by the Western Innovation Fund; the Natural Sciences and Engineering Research Council of Canada (NSERC Discovery Grant 371758-2009); US Public Health Service (DART-DOSE CMCR, 5U01AI091173-0); the Canadian Foundation for Innovation; Canada Research Chairs, and CytoGnomix Inc.
Automated Dicentric Chromosome Identifier and Dose Estimator (ADCI) | CytoGnomix | NA | ADCI software is released in a binary installation package file for Microsoft Windows 7, 8, 8.1 and 10; 235 Mb of disk storage are required for a typical installation. The software has been tested with Intel or AMD x86-64 processors; at least 1 Gb RAM is recommended. Analyses have been benchmarked on a computer configured with an Intel I7 processor and 16 Gb RAM. Operation of ADCI requires an active license and a USB-based hardware dongle, which must remain plugged in while the software is executing. The dongle encodes the software expiry date. Each time the software is started, this date is read. The software will allow access to the program if the current date and time precedes the expiration time-date stamp. Extending an expired software license can be accomplished by obtaining a new dongle or by renewing the license with an updated key at startup. |
Digital images of metaphase cell nuclei | Examples: Metasystems, Leica Microsystems | M-Search (Metasystems), Cytovision (Leica) software | High resolution TIFF format; typically >250 digital images generated with a microscope imaging capture system (minimum 63x magnification objective, 10x magnification ocular). |
MSI Leopard Pro (recommended, optional) | Micro-Star International | MSI GP62 6QF 480CA Leopard Pro | Multi-core performance workstation. |