Image quality of photon counting and energy integrating chest CT – Prospective head-to-head comparison on same patients

Abstract

PurposeTo prospectively compare the image quality of high-resolution, low-dose photon-counting detector CT (PCD-CT) with standard energy-integrating-detector CT (EID) on the same patients. MethodIRB-approved, prospective study; patients received same-day non-contrast CT on EID and PCD-CT (NAEOTOM Alpha, blinded) with clinical protocols. Four blinded radiologists evaluated subsegmental bronchial wall definition, noise, and overall image quality in randomized order (0 = worst; 100 = best). Cases were quantitatively compared using the average Global-Noise-Index (GNI), Noise-Power-Spectrum average frequency (fav), NPS frequency-peak (fpeak), Task-Transfer-Function-10%-frequency (f10) an adjusted detectability index (d’adj), and applied output radiation doses (CTDIvol). ResultsSixty patients were prospectively imaged (27 men, mean age 67 ± 10 years, mean BMI 27.9 ± 6.5, 15.9–49.4 kg/m2).Subsegmental wall definition was rated significantly better for PCD-CT than EID (mean 71 [56–87] vs 60 [45–76]; P < 0.001), noise was rated higher for PCD-CT (48 [26–69] vs 34 [13–56]; P < 0.001). Overall image quality was rated significantly higher for PCD-CT than EID (66 [48–85] vs 61 [42–79], P = 0.008).Automated image quality measures showed similar differences for PCD-CT vs EID (mean GNI 70 ± 19 HU vs 26 ± 8 HU, fav 0.35 ± 0.02 vs 0.25 ± 0.02 mm−1, fpeak 0.07 ± 0.01 vs 0.09 ± 0.03 mm−1, f10 0.7 ± 0.08 vs 0.6 ± 0.1 mm−1, all p-values < 0.001). PCD-CT showed a 10% average d’adj increase (-49% min, 233% max).PCD-CT studies were acquired at significantly lower radiation doses than EID (mean CTDIvol 4.5 ± 2.1 vs 7.7 ± 3.2 mGy, P < 0.01). ConclusionThough PCD-CT had higher measured and perceived noise, it offered equivalent or better diagnostic quality compared to EID at lower radiation doses, due to its improved resolution.