Increased use of CT Pulmonary angiography in suspected pulmonary embolism (PE) has driven research to minimize radiation dose while maintaining image quality and diagnostic accuracy. Following institutional review board approval, we performed a retrospective comparison study in patients with suspected PE. Patients were scanned using an ultra high pitch dual source technique (pitch = 2.6) using 120 kV (SVCTPA) (n = 54) or 100 kV (RV-CTPA) (n = 52). SV-CTPA images were reconstructed using filtered back projection (SV-wFBP) and RV-CTPA images were reconstructed using both FBP (RV-wFBP) and Iterative Reconstruction (RV-IR). Comparison of radiation dose, diagnostic ability, subjective image noise, quality, and sharpness, diagnostic agreement, signal to noise (SNR) and contrast to noise ratios (CNR) were performed. Mean effective dose was 2.56 ± 0.19 mSv for the RV protocol compared to 5.36 ± 0.60 mSv for the SV. The RV-CTPA protocol resulted in a mean DLP reduction of 52 % and mean CTDI reduction of 51 %. Pulmonary artery SNR and CNR were significantly higher on RV-IR images than SV-wFBP (p = 0.007, p = 0.003). Mean subjective image noise, quality and sharpness scores did not differ significantly between the SV-wFBP and RVIR images (p > 0.05). Subjective quality scores were significantly better for the RV-IR group compared to the RV-wFBP group (p < 0.001). Agreement between readers for presence or absence of pulmonary emboli on RV-IR images was almost perfect (κ = 0.891, p < 0.001). Iterative reconstruction complements ultra high pitch dual source CTPA examinations acquired using a reduced voltage resulting in higher mean pulmonary artery SNR and CNR when compared to both RV-wFBP and SV-CTPA.