Fig 1.
The image shows the measurement set-up used to acquire DR-time curves during diagnostic administrations. A) The RadEye PRD-ER device. B) The RadEye SPRD-ER device. C) RadSight software connected through bluetooth with the devices positioned on patient arms. PC screen shows real-time curves that are recorded during administration.
Fig 2.
Image A) is an 18F-FDG ideal DR-time curve acquired during the administration; Δpnor and ΔRt are defined in the paragraph Measured Parameters and Data Processing. Image B) is an ideal 177Lu-DOTATATE DR-time curve acquired during therapeutic administration. The continuous line is the signal acquired on the injection arm; the dashed line is the contralateral signal. The calculated metric is drawn in the figure.
Fig 3.
A) and B) DR-time curves acquired for normal and abnormal injections for diagnostic radiopharmaceuticals with acquired PET/CT images. The metric calculated for each curve is drawn in the figure. C) DR-time curve acquired for 177Lu-DOTATATE extravasation event with acquired planar scintigraphy 2 hours after therapeutic administration. The metric calculated is drawn in the figure.
Fig 4.
Real-time detection of extravasation events and SUV correction calculation workflow. The time required from step 1 to step 3 is 8 minutes for diagnostic injections and 15 minutes for therapeutic ones. Step 4 is different for diagnostic and therapeutic procedures. However, it is an off-line process, that requires 5 minutes with homemade software written in Python both for therapeutic and diagnostic patients.
Fig 5.
Statistical results for diagnostics.
Boxplot of ΔpinNOR and ΔRt for the two different classes of injection. The upper figure shows a box plot with the median value and interquartile range of ΔpinNOR for the two injection classes, and the lower figure shows a box plot with the median value and interquartile range of ΔRt for the two injection classes.
Fig 6.
Logistic regression for diagnostics.
An operational threshold (0.5 of the logistic curve) was identified to distinguish between normal injections and extravasation events. The upper curve refers to ΔpinNOR, the lowest to ΔRt.
Fig 7.
Statistical results for therapy.
The figure shows the trend of ΔRt with time for normal injection cases in therapy.
Table 1.
Dosimetry results for diagnostics.
Table 2.
Dosimetry results for therapy.
Fig 8.
Fitted curves of the inverse of Δpinnor (figure A) and ΔRt normalized (figure B) versus SUV correction coefficient. Ideal cases were associated with a cumulative point with an SUV correction coefficient equal to 1.