Why do anatomic backgrounds reduce lesion detectability?

Impact of Colorized Display of Mammograms on Lesion Detection

OBJECTIVE

To assess the effect of the colorized display of digital mammograms on observer detection of subtle breast lesions.

METHODS

Three separate observer studies compared detection performance using grayscale versus color display of 1) low-contrast mass-like objects in a standardized mammography phantom; 2) simulated microcalcifications in a background of normal breast parenchyma; and 3) standard-of-care clinical digital mammograms with subtle calcifications and masses. Colorization of the images was done by displaying each image pixel in blue, green, and red hues, or gray, maintaining DICOM-calibrated luminance scale and consistent luminance range. For the simulated calcifications and clinical mammogram studies, comparison of detection rates was computed using McNemar's test for paired differences.

RESULTS

For the phantom study, mass-like object detection was significantly better using a green colormap than grayscale (73.3% vs 70.8%, P = .009), with no significant improvement using blue or red colormaps (72.6% and 72.5%, respectively). For simulated microcalcifications, no significant difference was noted in detection using the green colormap, as compared with grayscale. For clinical digital screening mammograms, no significant difference was noted between gray and green colormaps for detection of microcalcifications. Green color display, however, resulted in decreased sensitivity for detection of subtle masses (63% vs 69%, P = .03).

CONCLUSION

Although modest improvement was demonstrated for a detection task using colorized display of a standard mammography phantom, no significant improvement was demonstrated using a color display for a simulated clinical detection task, and actual clinical performance was worse for colorized display of mammograms in comparison to standard grayscale display.

99mTc-HYNIC-rh-annexin-V scintigraphy: visual and quantitative evaluation of early treatment-induced apoptosis to predict treatment outcome

AIM

To determine the reliability of visual analysis of 99mTc-HYNIC-rh-annexin-V tumour uptake (ATU) compared to quantitative tracer uptake evaluation.

METHODS

Thirty-eight patients (22 male, 16 female, mean age 57) with histologically proved lymphoma (n=31), non-small cell lung cancer (NSCLC) (n=4) and head and neck squamous cell carcinoma (H&NSCC) (n=3) were examined. 99mTc-HYNIC-rh-annexin-V scintigraphy (TAS) was acquired before and within 2 days after the start of anti-cancer treatment. Maximal counts per pixel in the tumour volume (Cmax) were calculated for every target lesion. To match the quantitative and visual ATU, both were expressed as a four-grade score. Cmax as percentages of baseline values: grade 1, decrease >25%; grade 0, 1-25% decrease; grade 1, 1-25% increase; grade 2, >25% increase. Visual analysis: 0=absent, 1=weak, 2=moderate, 3=intense. Intra-observer and inter-observer variability and methodological agreement between visual and quantitative evaluation of ATU was expressed by computing Cohen's kappa statistics.

RESULTS

A statistically highly significant correlation was found between the changes in ATU and therapy outcome: r=0.97 (P<0.0001) and r=0.99 (P<0.0001) for visual and quantitative analysis, respectively. Good intra-observer reproducibility, with a high kappa of 0.82 for observer 1 and a kappa of 0.90 for observer 2, was determined. Inter-observer variability was 0.82.

CONCLUSION

Visual evaluation of ATU after image co-registration appears to be a reliable and reproducible method for preliminary assessment of early treatment-induced apoptosis.

Search for lesions in mammograms: Statistical characterization of observer responses

We investigate human performance for visually detecting simulated microcalcifications and tumors embedded in x-ray mammograms as a function of signal contrast and the number of possible signal locations. Our results show that performance degradation with an increasing number of locations is well approximated by signal detection theory (SDT) with the usual Gaussian assumption. However, more stringent statistical analysis finds a departure from Gaussian assumptions for the detection of microcalcifications. We investigated whether these departures from the SDT Gaussian model could be accounted for by an increase in human internal response correlations arising from the image-pixel correlations present in 1/f spectrum backgrounds and/or observer internal response distributions that departed from the Gaussian assumption. Results were consistent with a departure from the Gaussian response distributions and suggested that the human observer internal responses were more compact than the Gaussian distribution. Finally, we conducted a free search experiment where the signal could appear anywhere within the image. Results show that human performance in a multiple-alternative forced-choice experiment can be used to predict performance in the clinically realistic free search experiment when the investigator takes into account the search area and the observers' inherent spatial imprecision to localize the targets.

A statistical methodology for mammographic density detection

A statistical methodology is presented based on a chi-square probability analysis that allows the automated discrimination of radiolucent tissue (fat) from radiographic densities (fibroglandular tissue) in digitized mammograms. The method is based on earlier work developed at this facility that shows mammograms may be considered as evolving from a linear filtering operation where a random input field is passed through a 1/f filtering process. The filtering process is reversible which allows the solution of the input field with knowledge obtained from the raw image (the output). The input field solution is analogous to a prewhitening technique or deconvolution. This field contains all the information of the raw image in a much simplified format that can be approximated and analyzed with parametric methods. In the work presented here evidence indicates that there are two random events occurring in the input field with differing variances: (1) one relating to fat tissue with the smaller variance, and (2) the second relating to all other tissue with the larger variance. A statistical comparison of the variances is made by scanning the image with a small search window. A relaxation method allows for making a reliable estimate of the smaller variance which is considered as the global reference. If a local variance deviates significantly from the reference variance, based on chi-square analysis, it is labeled as nonfat; otherwise it is labeled as fat. This statistical test procedure results in a region by region continuous labeling of fat and nonfat tissue across the image. In the work presented here, the emphasis is on the methodology development with supporting preliminary results that are very encouraging. It is widely accepted that mammographic density is a breast cancer risk factor. An important application of this work is to incorporate density-based risk analysis into the ongoing statistical-based detection work developed at this facility. Additional applications include risk analysis dependent on either percentages or total amounts of fat or dense tissue. This work may be considered as the initial step in introducing many of the known breast cancer risk factors into the actual image data analysis.