Journal of Undergraduate Research
Volume 8, Issue 2
2 - November/December 2006

Estimates of Spongiosa Volume for Use in Patient-Specific Dosimetry

Lindsay Sinclair

ABSTRACT

Patient-specific dosimetry within the field of radionuclide therapy continues to pose a challenge due to the difficulty in predicting the toxicity of the bone marrow. An important parameter to consider in this prediction is the mass of the total active bone marrow in a patient. However, the mass of the bone marrow is very difficult to obtain due to its biological structure. It is assumed there is a linear relationship between the patient’s marrow mass and the total spongiosa volume of the individual.

In this study, two methods for prediction of the total skeletal spongiosa volume in a patient were analyzed and tested. The first method utilizes multiple regression analysis, while the second uses ratios called F-factors to predict the total spongiosa volume. Both of the methods were applied to the test subject, and the results showed the F-factors produced the most accurate spongiosa volume for the individual. However, the regression model did not have a very high percent error, and could be more advantageous to work with than the given F-factors.

INTRODUCTION

The primary objective of radionuclide therapy is to supply cancer sites with enough enough radioactive material to eradicate cancerous cells and minimize the damage done to healthy cells. Radionuclide S-values associated with the ICRP reference patient have traditionally been used to estimate the absorbed dose to the active bone marrow. In ICRP Publication 23¹ the reference adult male is 70 kg in mass and 170 cm in height. Since patients undergoing radionuclide therapy vary in both mass and height from the reference individual, many attempts at altering S-values have been made. Studies by Siegel² mentioned that in targeted radionuclide therapy, the mass of the total active marrow (m_TAM) is needed for patient-specific dosimetry. This measurement however, is extremely difficult to perform. Hence, the scaling of the m_TAM by a patient-specific parameter is necessary. Traditionally, most widely used scaler is total body mass (TBM).

Figure 1

However, Stabin et al.³ states that lean body mass would be an improved anthropometric parameter due to the fact that red marrow mass is not greatly affected by obesity. Other anthropometric parameters that have been examined are total body height and body surface area.

The anthropometric parameter that is examined in this study is that of the total skeletal spongiosa volume (TSSV). Studies by Bolch et al.⁴ and Shen et al.⁵ have concluded that the spongiosa volumes could be an effective parameter in scaling reference S-values. These studies were done in 2002 and both assumed the following to be true:

Figure 2

this implies that the ratios of the total active marrow mass at a skeletal site X is approximated by the related ratios of the spongiosa volumes at that same site. Using the spongiosa volume as the scaling factor has a great advantage to other parameters in that it is a more specific measure of the biological structure of a particular bone site. Spongiosa volumes include the bone marrow and the bone trabeculae. Since obtaining the mass of the active bone marrow is the ultimate goal of scaling, specificity will have a much more profound result.

This study further investigates two methods of obtaining total skeletal spongiosa volumes. The first method was developed by Brindle et al. (6), and its purpose was to develop a regression model for predicting the TSSV in a given patient. The TSSV included all of the major skeletal sites containing active marrow. The major skeletal sites are as follows: os coxae, lumbar, thoracic, and cervical vertebrae, mandible, ribs, scapulae, sacrum, sternum, clavicle, humerii, femoral heads and cranium.

The second method involves F-factors, which are computed by taking the average ratio of TSSV over spongiosa volume for any site, X. Through detailed manual image segmentation, TSSV and SV for any site X were computed from 20 cadavers, and an average value or F-factor was found. The F-factor can then be applied to the SV for any site X to predict the TSSV for that individual.

Manual image segmentation was performed on a particular cadaver in order to calculate the total spongiosa volume for that specific individual. Two estimates were then calculated using both models, and an analysis was done to see which of the methods would be more accurate and useful in a clinical setting.

METHODOLOGY

Cadaver-26 Analysis

The cadaver in this study will undergo several processes in order to be compared with the TSSV estimates from the multiple regression model and the F-factor model. Cadaver-26 was a 65 year old male who died from chronic obstructive pulmonary disease. His height and weight were 183 cm and 68.2 kg, respectively.

Initially this cadaver was subjected to a series of whole-body computed tomography (CT) scans with slice thickness of 2 mm. Using CT_Contours⁷ written in the Interactive Data Language software, the CT scans were manually segmented in each of the marrow-containing skeletal sites with a different tag value.

In the contour file, the number of voxels associated with a given tag value are summed up and multiplied by the voxel dimensions to give the volume of spongiosa for the region segmented. By summing all of the spongiosa volumes an estimate for the TSSV is achieved. This TSSV will be used as the “correct” TSSV to be compared with total spongiosa volume estimates from the multiple regression and F-factor models.

The Regression Model

As part of a larger National Cancer Institute study at the University of Florida, work has been done to develop a clinically useful regression multiple model that will predict the total volume of trabecular spongiosa in a given patient. This study by Brindle et al. (6), started with acquiring 20 cadavers (10 male and 10 female), which underwent a series of whole-body computed tomography (CT) scans. These scans were then manually segmented with the use of a program called CT_Contours (7)written in the Interactive Data Language software.

From the image segmentation, spongiosa volumes were estimated from each of the 13 skeletal sites containing active marrow as defined by the ICRP (8). The TSSV was used in the multiple regression analysis, and is given as the sum of the 13 individual volumes. Many statistical approaches were considered until the corrected Akaike information criterion (AIC) and the Bayesian information criterion (BIC) were chosen. In addition, the os coxae width and os coxae height were the anthropometric parameters that showed to give the most accurate estimate of TSSV.

The F-Factor Model

As mentioned earlier, the F-factor is really an average value of the ratio of TSSV over spongiosa volume (SV), at a skeletal site, X, given as the formula:

Figure 3

each cadaver, and manual image segmentation gave estimates for the SV, at any of the 13 skeletal sites, X, as well as the TSSV. The F-factor was found for each of the cadavers at each skeletal site, and then an average was taken (Table 1).

Note that the F-factor for the cranium was not calculated. The standard segmentation error for all other sites was within + 5%. The segmentation percent error for the cranium was too large for it to be considered viable.

The L-5 vertebra was segmented separately from the rest of the lumbar vertebrae because of the relatively large amount of marrow found within the site.

The estimate of total skeletal spongiosa volume can be obtained by taking the value of a particular F-factor and multiplying by the spongiosa volume at that particular site.

RESULTS

From the cadaver analysis, the TSSV was calculated from the spongiosa volumes of all the segmented marrow-containing skeletal sites. After segmentation was completed the TSSV was shown to be 2246.8683 cm³.

The Regression model⁶ utilized the following formula to estimate TSSV for cadaver-26:

Figure 4

The coefficients came from Table 4 in Brindle et al.⁶, with β0 = -5874.7, β1 = -41.1 and β2 = 426.5. The Regression Model estimated the TSSV to be 2385.26 cm³. In relation to the actual TSSV, this model exhibited a -6.115 percent error.

Using the F-factor method, the results are shown in Table 3.

CONCLUSION

The purpose of this study was to analyze different methods that produce patient-specific estimates of total skeletal spongiosa volume for use in radionuclide therapy. The results of this research produced several estimates of total skeletal spongiosa volume with a relatively small error. When comparing the two methods at hand, the F-factor method exhibited a smaller error than the multiple regression model. However, the F-factor method was evaluated at 13 skeletal sites, with four of them viable. These four sites were the L-5 vertebra, the cervical vertebrae, the femoral heads, and the humerii. The humerii gave the best estimate, with a -0.79 percent error. In order for the F-factor method to be clinically useful, a CT scan would have to be performed, and manual segmentation would have to be done on one of the four skeletal sites. The formula could then be applied to produce the estimate of TSSV needed for dosimetry.

While the F-factor method showed greater accuracy, the regression model still had promising results, and could be more advantageous to work with in a clinical setting. In order to implement the regression model a CT scan would first have to be completed. Afterwards, image based measurements of the os coxae would be performed with the features found in the PacsCube™ viewing software included with the CT data sets. Making two simple measurements would be much less time-consuming than manual segmentation of a skeletal site. Hence, the regression model would be a more clinically viable method to be used with patient-specific dosimetry in the field of radionuclide therapy.

---

REFERENCES

1. ICRP. Report on the Task Group on Reference Man. ICRP Publication 23. Oxford, UK: International Commission on Radiological Protection; 1975
2. Siegel JA. Establishing a clinically meaningful predictive model of hematologic toxicity in nonmyeloablative targeted radiotherapy: practical aspects and limitations of red marrow dosimetry. Cancer Biother Radiopharm. 2005;20:126-140.
3. Stabin MG, Siegel JA, Sparks RB. Sensitivity of model-based calculations of red marrow dosimetry to changes in patient-specific parameters. Cancer Biother Radiopharm. 2002;17:535-543
4. Bolch WE, Patton PW, Shah AP, Rajon DA. Considerations of anthropomorphic, tissue volume, and tissue mass scaling for improved patient specificity of skeletal S values. Med Phys. 2002;29:1054-1070.
5. Shen S, Meredith RF, Duan J, et al. Improved prediction of myelotoxicity using a patient-specific imaging dose estimate for non-marrow-targeting (90)y-antibody therapy. J Nucl Med. 2002;43:1245-1253.
6. Brindle JM, Trindade AA, Shah AP, et al. A Linear Regression Model for predicting patient-specific total skeletal spongiosa volume for use in molecular radiotherapy dosimetry. 2006.
7. Nipper JC, Williams JL, Bolch WE. Creation of two tomographic voxel models of pediatric patients in the first year of life. Phy Med Biol. 2002;47:3143-3164.
8. ICRP. Basic anatomical and physiological data for use in radiological protection: the skeleton. ICRP Publication 70. Oxford, UK: International Commission on Radiological Protection; 1995.

---

--top--

Back to the Journal of Undergraduate Research