A simple algorithm to assess patient suitability for Calypso-seed implantation for four-dimensional prostate localization

Abstract

To retrospectively determine the proportion of prostate cancer patients who are appropriate candidates for prostate localization with Calypso (Calypso Medical, Seattle, WA); to assess the accuracy of surface anatomy in predicting prostate depth; and to describe a simple clinical algorithm predicting patient's appropriateness for Calypso localization. Medical records and archived CT scans of all patients treated for localized prostate cancer at our institution between 2006 and 2007 were reviewed. Association between the feasibility of Calypso use, the depth of the prostate from the anterior torso, and a variety of anatomic factors were assessed (ANOVA, linear regression, and ROC). Patients were appropriate for the Calypso system in 91% of cases (localize and track, 52%; localize only, 39%). Strong correlation between greater trochanter location and the posterior prostate was seen (r 2 = 0.91, mean difference 0.6 cm). The negative predictive value of the greater trochanter measurements was 31%. Thirty-one out of forty-five patients (69%) who were deemed inappropriate for Calypso based on greater trochanter to anterior torso measurements were eligible on the basis of CT-based measurements of prostate depth. Weight, BMI, waist circumference, and hip circumference correlated with distance from the prostate to the anterior torso and were predictive of Calypso appropriateness. All patients with weight <or= 100 kg, BMI <or= 30, or waist/hip circumference <or= 100 cm, were eligible for Calypso. Most prostate cancer patients are candidates for Calypso localization +/- tracking. The greater trochanter to anterior torso distance underestimates the number of eligible patients. Weight, BMI and waist/hip circumference are good predictors for Calypso appropriateness.

Figure 1

Setup of a patient with the Calypso array in place above the torso (a). AP (b) and lateral (c) digitally reconstructed radiograph from the CT simulation of a patient. Hip circumference was taken at the level of the greater trochanter (arrow in (b)). Waist circumference was taken at the level of the anterior superior iliac spine (arrow in (c)). The largest distance from the table to the anterior torso was also recorded (arrowhead in (c)). The distance from the table top to the greater trochanter (arrow in (d), anterior prostate, and posterior prostate were measured on an axial image.

Figure 2

Mean $(\pm \text{standard deviation})$ distance from the greater trochanter, anterior prostate, and posterior prostate to the most anterior aspect of the torso (a); scatter plot of greater trochanter to anterior torso distance versus posterior prostate to anterior torso distance ${(\mathrm{r}}^2\hspace{0.17em}=\hspace{0.17em}0.91)$ (b). The slope of the line is 0.83 (95% CI 0.77 to 0.88). Based on manufacturer‐recommended surface measurements: those appropriate for localize and track (solid circles), localize only (open circles), or ineligible for Calypso (crosses) are indicated. The recommended maximum distances between the greater trochanter and anterior torso to be appropriate for tracking (17 cm) or localization (23 cm) are indicated by vertical dashes. The maximum recommended distance between implanted transponder and the array to be appropriate for tracking (21 cm) and localization (27 cm) are indicated by horizontal dashes. Patients who would be misclassified as not appropriate for “localize and track” (dark shaded region) or ineligible for Calypso use (light shaded region) are indicated.

Figure 3

Scatter plot of weight (a), BMI (b), hip circumference (c), waist circumference (d), and age (e) vs. the distance from the posterior prostate to the anterior torso. Best fit line (solid) and 95% confidence intervals (dotted lines) in addition to linear regression goodness of fit parameter $({\mathrm{r}}^2)$ are shown for each plot. Based on manufacturer‐recommended surface measurements: those appropriate for localize and track (solid circles), localize only (open circles), or ineligible for Calypso (crosses) are indicated. The maximum depths that the Calypso system can provide tracking (21 cm) and localization (27 cm) are indicated by horizontal dashed lines. Vertical dashed lines represent selected cutoff value determined by ROC analysis. ROC curves (f) for age (light long dash), weight (solid dark), BMI (solid light), waist circumference (dashed dark), and hip circumference (dashed light).