Physics Contribution
A New Brain Positron Emission Tomography Scanner With Semiconductor Detectors for Target Volume Delineation and Radiotherapy Treatment Planning in Patients With Nasopharyngeal Carcinoma

https://doi.org/10.1016/j.ijrobp.2011.09.011Get rights and content

Purpose

We compared two treatment planning methods for stereotactic boost for treating nasopharyngeal carcinoma (NPC): the use of conventional whole-body bismuth germanate (BGO) scintillator positron emission tomography (PETCONVWB) versus the new brain (BR) PET system using semiconductor detectors (PETNEWBR).

Methods and Materials

Twelve patients with NPC were enrolled in this study. [18F]Fluorodeoxyglucose-PET images were acquired using both the PETNEWBR and the PETCONVWB system on the same day. Computed tomography (CT) and two PET data sets were transferred to a treatment planning system, and the PETCONVWB and PETNEWBR images were coregistered with the same set of CT images. Window width and level values for all PET images were fixed at 3000 and 300, respectively. The gross tumor volume (GTV) was visually delineated on PET images by using either PETCONVWB (GTVCONV) images or PETNEWBR (GTVNEW) images. Assuming a stereotactic radiotherapy boost of 7 ports, the prescribed dose delivered to 95% of the planning target volume (PTV) was set to 2000 cGy in 4 fractions.

Results

The average absolute volume (±standard deviation [SD]) of GTVNEW was 15.7 ml (±9.9) ml, and that of GTVCONV was 34.0 (±20.5) ml. The average GTVNEW was significantly smaller than that of GTVCONV (p = 0.0006). There was no statistically significant difference between the maximum dose (p = 0.0585) and the mean dose (p = 0.2748) of PTV. The radiotherapy treatment plan based on the new gross tumor volume (PLANNEW) significantly reduced maximum doses to the cerebrum and cerebellum (p = 0.0418) and to brain stem (p = 0.0041).

Conclusion

Results of the present study suggest that the new brain PET system using semiconductor detectors can provide more accurate tumor delineation than the conventional whole-body BGO PET system and may be an important tool for functional and molecular radiotherapy treatment planning.

Introduction

Since the advent of computed tomography (CT), sophisticated techniques in radiation treatment such as three-dimensional conformal radiotherapy, stereotactic radiotherapy, and intensity-modulated radiotherapy (IMRT) have been developed in order to focus and escalate the radiation dose to the tumor while sparing normal tissues. In these techniques, it is important to precisely determine the tumor volume. With their high anatomic resolution, CT and magnetic resonance images (MRI) have been used primarily for target volume delineation in radiotherapy treatment planning. However, when delineating the target volume, it is sometimes difficult to distinguish between tumor and nontumor tissues by using anatomical imaging alone. In the past 10 years, positron emission tomography (PET) labeled with [18F]fluorodeoxyglucose (FDG), which is able to visualize molecular information for the tumor, has been widely used in oncology for diagnosis and staging of various cancers. This functional imaging has been adopted in radiotherapy, and several studies have examined the clinical impact of PET on radiotherapy planning 1, 2, 3. However, as PET does not provide an intrinsically accurate examination, with a spatial resolution of approximately 4 to 7 mm 4, 5, 6, it is difficult to determine tumor boundaries on conventional whole-body bismuth germanate (BGO) scintillator PET images. In 2007, a new brain PET scanner with semiconductor detectors, the first in the world, was developed with Hitachi, Ltd., and was installed at our institute (7). This brain PET system is equipped with small semiconductor detectors and a depth of interaction system with sufficient sensitivity to obtain higher spatial resolution (2.3 mm at 1 cm [National Electrical Manufacturers Association (NEMA) NU 2-2001]). Semiconductor detectors also have an advantage in energy resolution. Our new semiconductor PET detectors had an energy resolution of 4.1% full-width half maximum, which is superior to the energy resolution obtained with previously available scintillation detectors (e.g., 10%–20%) 8, 9. The limited energy window set permits collection of accurate signal counts with lower noise counts. The scatter fraction of the new brain PET system was 23% (NEMA NU 2-1994), which was lower than those of other scintillation-based whole-body BGO PET scanners such as Exact HR+ (32.1%; NEMA NU 2-1994; Asahi-Siemens, Tokyo, Japan) 10, 11. In our previous study, the contrast obtained with the semiconductor brain PET scanner was 27% higher than that obtained with the conventional whole-body BGO scanner for both a cold spot phantom with 6-mm-diameter cold sphenoid defects and a dual-cylinder phantom with an adjusted concentration of 1:2 surrounded with water (7). In patients with nasopharyngeal carcinoma (NPC), the new brain PET system identified intratumoral inhomogeneity in more detail than the conventional whole-body BGO PET system, and the tumor edge was sharper on images obtained with the new brain PET system than on those obtained with the conventional whole-body BGO PET system (7). Therefore, the new brain PET system has the potential to provide high contrast and detailed images with sharper tumor edges in radiation treatment planning for NPC.

The purpose of this study was to evaluate effects of the new brain PET system for radiotherapy treatment planning for patients with NPC compared with those of a conventional whole-body BGO PET, Exact HR+.

Section snippets

Patients

Subjects in this study were 12 NPC patients who had been newly diagnosed between July 2007 and April 2009. The median age was 61 years old (range, 30–76 years old). Patient characteristics are shown in Table 1. Written informed consent was obtained from all patients.

Image acquisition and target volume delineation

Before undergoing the PET study, all patients fasted for at least 6 h. Serum glucose levels were checked in all patients before we administered [18F]FDG. The dose of [18F]FDG for each patient was 370 MBq. [18F]FDG-PET images were

Results

Absolute volumes of GTVNEW and GTVCONV are shown in Table 2. The average (±standard deviation [SD]) absolute volume of GTVNEW was 15.7 (±9.9; range, 4.9–31.6) ml, and that of GTVCONV was 34.0 ml (±20.5; range, 10.6–75.9) ml. The average absolute volume of GTVNEW was significantly smaller than that of GTVCONV (p = 0.0006). Regardless of the order in which the two [18F]FDG examinations were conducted, volumes of GTVNEW were always smaller than GTVCONV for all 12 patients.

Maximum and mean doses of

Discussion

Although PET offers better identification of tumor localization than the anatomical imaging modalities because of its higher contrast resolution, tumor boundaries are blurred on the conventional BGO PET system because of its relatively low spatial resolution due to its larger detectors and worse annihilation noncollinearity blurring because of the larger detector ring of whole-body BGO PET. Daisne et al. (12) reported thatPET-derived volumes are more accurate than CT or MRI-derived volumes for

Conclusions

Our results suggest that compared to the conventional whole-body BGO PET system, the new brain PET system using semiconductor detectors can provide better identification of tumor boundaries and more accurate tumor delineation; as such, it may be an important tool for functional and molecular radiotherapy treatment planning.

References (26)

Cited by (0)

This study was supported in part by Project for Developing Innovation Systems of the Ministry of Education, Culture, Sports, Science and Technology, the Japanese Government.

Conflict of interest: none.

View full text