Objective To systematically compare the imaging performance of cadmium zinc telluride (CZT) detectors and NaI(Tl) detectors in ¹⁷⁷Lu SPECT/CT and optimize reconstruction parameters for improved image quality.

Methods A National Electrical Manufacturers Association/International Electrotechnical Commission phantom was used. ¹⁷⁷Lu images were acquired in the 113 keV primary energy window on CZT and NaI(Tl) SPECT/CT systems. Images were reconstructed using ordered subsets expectation maximization method. The effects of different numbers of iterations (2–12) and different filters, including a Butterworth filter (normalized cut-off frequency (FC), 0.25–0.50) and a Gaussian filter (full width at half maximum (FWHM), 1.5–4.0), on image quality were analyzed, with combined corrections for attenuation, scatter, and resolution recovery. Under the phantom condition with a target to background ratio (T/B) of 16∶1, quantitative image quality metrics including the coefficient of variation (CV), contrast recovery coefficient (CRC), contrast-to-noise ratio (CNR), and signal-to-noise ratio (SNR) were used to quantitatively evaluate the images to determine the optimal reconstruction parameters. The obtained optimal reconstruction parameters were then applied to phantoms with T/B ratios of 4∶1, 8∶1, and 32∶1 to validate their generalization ability. The quantitative metrics between systems were then compared using the Wilcoxon test.

Results The quantitative study based on phantom reconstruction parameters identified the optimal reconstruction parameter combination: 2 iterations, a Gaussian filter with FWHM=3.0, and a Butterworth filter with FC=0.35 for the CZT system and FC=0.30 for the NaI(Tl) system, with the order set to 10 for both. This reconstruction parameter combination achieved relatively high image quality across multiple T/B ratios (8∶1, 16∶1, and 32∶1). Wilcoxon test results showed that at the same number of iterations and when the Butterworth filter was used, the CV, CRC, CNR, and SNR of the CZT system were all significantly higher than those of the NaI(Tl) system (all Z=−21.00, all P=0.031). When the Gaussian filter was used, the CRC, CNR, and SNR of the CZT system were all significantly higher than those of the NaI(Tl) system (all Z=−21.00, all P=0.031), and its CV was lower than that of the NaI(Tl) system (Z=21.00, P=0.031).

Conclusions CZT detectors demonstrated remarkably superior performance to NaI(Tl) detectors in ¹⁷⁷Lu SPECT/CT imaging. Validated with phantoms at multiple T/B ratios, the optimized reconstruction parameters achieved relatively good image quality in both systems, providing an important reference for precise clinical imaging.